Show More
@@ -0,0 +1,79 b'' | |||
|
1 | # Copyright (C) 2015-2024 RhodeCode GmbH | |
|
2 | # | |
|
3 | # This program is free software: you can redistribute it and/or modify | |
|
4 | # it under the terms of the GNU Affero General Public License, version 3 | |
|
5 | # (only), as published by the Free Software Foundation. | |
|
6 | # | |
|
7 | # This program is distributed in the hope that it will be useful, | |
|
8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
|
9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
|
10 | # GNU General Public License for more details. | |
|
11 | # | |
|
12 | # You should have received a copy of the GNU Affero General Public License | |
|
13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. | |
|
14 | # | |
|
15 | # This program is dual-licensed. If you wish to learn more about the | |
|
16 | # RhodeCode Enterprise Edition, including its added features, Support services, | |
|
17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ | |
|
18 | ||
|
19 | import logging | |
|
20 | ||
|
21 | from .backends.fanout_cache import FileSystemFanoutCache | |
|
22 | from .backends.objectstore_cache import ObjectStoreCache | |
|
23 | ||
|
24 | from .utils import archive_iterator # noqa | |
|
25 | from .lock import ArchiveCacheGenerationLock # noqa | |
|
26 | ||
|
27 | log = logging.getLogger(__name__) | |
|
28 | ||
|
29 | ||
|
30 | cache_meta = None | |
|
31 | ||
|
32 | ||
|
33 | def includeme(config): | |
|
34 | return # vcsserver gets its config from rhodecode on a remote call | |
|
35 | # init our cache at start | |
|
36 | settings = config.get_settings() | |
|
37 | get_archival_cache_store(settings) | |
|
38 | ||
|
39 | ||
|
40 | def get_archival_config(config): | |
|
41 | ||
|
42 | final_config = { | |
|
43 | ||
|
44 | } | |
|
45 | ||
|
46 | for k, v in config.items(): | |
|
47 | if k.startswith('archive_cache'): | |
|
48 | final_config[k] = v | |
|
49 | ||
|
50 | return final_config | |
|
51 | ||
|
52 | ||
|
53 | def get_archival_cache_store(config, always_init=False): | |
|
54 | ||
|
55 | global cache_meta | |
|
56 | if cache_meta is not None and not always_init: | |
|
57 | return cache_meta | |
|
58 | ||
|
59 | config = get_archival_config(config) | |
|
60 | backend = config['archive_cache.backend.type'] | |
|
61 | ||
|
62 | archive_cache_locking_url = config['archive_cache.locking.url'] | |
|
63 | ||
|
64 | match backend: | |
|
65 | case 'filesystem': | |
|
66 | d_cache = FileSystemFanoutCache( | |
|
67 | locking_url=archive_cache_locking_url, | |
|
68 | **config | |
|
69 | ) | |
|
70 | case 'objectstore': | |
|
71 | d_cache = ObjectStoreCache( | |
|
72 | locking_url=archive_cache_locking_url, | |
|
73 | **config | |
|
74 | ) | |
|
75 | case _: | |
|
76 | raise ValueError(f'archive_cache.backend.type only supports "filesystem" or "objectstore" got {backend} ') | |
|
77 | ||
|
78 | cache_meta = d_cache | |
|
79 | return cache_meta |
@@ -0,0 +1,17 b'' | |||
|
1 | # Copyright (C) 2015-2024 RhodeCode GmbH | |
|
2 | # | |
|
3 | # This program is free software: you can redistribute it and/or modify | |
|
4 | # it under the terms of the GNU Affero General Public License, version 3 | |
|
5 | # (only), as published by the Free Software Foundation. | |
|
6 | # | |
|
7 | # This program is distributed in the hope that it will be useful, | |
|
8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
|
9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
|
10 | # GNU General Public License for more details. | |
|
11 | # | |
|
12 | # You should have received a copy of the GNU Affero General Public License | |
|
13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. | |
|
14 | # | |
|
15 | # This program is dual-licensed. If you wish to learn more about the | |
|
16 | # RhodeCode Enterprise Edition, including its added features, Support services, | |
|
17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ |
@@ -0,0 +1,372 b'' | |||
|
1 | # Copyright (C) 2015-2024 RhodeCode GmbH | |
|
2 | # | |
|
3 | # This program is free software: you can redistribute it and/or modify | |
|
4 | # it under the terms of the GNU Affero General Public License, version 3 | |
|
5 | # (only), as published by the Free Software Foundation. | |
|
6 | # | |
|
7 | # This program is distributed in the hope that it will be useful, | |
|
8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
|
9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
|
10 | # GNU General Public License for more details. | |
|
11 | # | |
|
12 | # You should have received a copy of the GNU Affero General Public License | |
|
13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. | |
|
14 | # | |
|
15 | # This program is dual-licensed. If you wish to learn more about the | |
|
16 | # RhodeCode Enterprise Edition, including its added features, Support services, | |
|
17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ | |
|
18 | ||
|
19 | import os | |
|
20 | import functools | |
|
21 | import logging | |
|
22 | import typing | |
|
23 | import time | |
|
24 | import zlib | |
|
25 | ||
|
26 | from ...ext_json import json | |
|
27 | from ..utils import StatsDB, NOT_GIVEN, ShardFileReader, EVICTION_POLICY, format_size | |
|
28 | from ..lock import GenerationLock | |
|
29 | ||
|
30 | log = logging.getLogger(__name__) | |
|
31 | ||
|
32 | ||
|
33 | class BaseShard: | |
|
34 | storage_type: str = '' | |
|
35 | fs = None | |
|
36 | ||
|
37 | @classmethod | |
|
38 | def hash(cls, key): | |
|
39 | """Compute portable hash for `key`. | |
|
40 | ||
|
41 | :param key: key to hash | |
|
42 | :return: hash value | |
|
43 | ||
|
44 | """ | |
|
45 | mask = 0xFFFFFFFF | |
|
46 | return zlib.adler32(key.encode('utf-8')) & mask # noqa | |
|
47 | ||
|
48 | def _write_file(self, full_path, read_iterator, mode): | |
|
49 | raise NotImplementedError | |
|
50 | ||
|
51 | def _get_keyfile(self, key): | |
|
52 | raise NotImplementedError | |
|
53 | ||
|
54 | def random_filename(self): | |
|
55 | raise NotImplementedError | |
|
56 | ||
|
57 | def store(self, *args, **kwargs): | |
|
58 | raise NotImplementedError | |
|
59 | ||
|
60 | def _store(self, key, value_reader, metadata, mode): | |
|
61 | (filename, # hash-name | |
|
62 | full_path # full-path/hash-name | |
|
63 | ) = self.random_filename() | |
|
64 | ||
|
65 | key_file, key_file_path = self._get_keyfile(key) | |
|
66 | ||
|
67 | # STORE METADATA | |
|
68 | _metadata = { | |
|
69 | "version": "v1", | |
|
70 | ||
|
71 | "key_file": key_file, # this is the .key.json file storing meta | |
|
72 | "key_file_path": key_file_path, # full path to key_file | |
|
73 | "archive_key": key, # original name we stored archive under, e.g my-archive.zip | |
|
74 | "archive_filename": filename, # the actual filename we stored that file under | |
|
75 | "archive_full_path": full_path, | |
|
76 | ||
|
77 | "store_time": time.time(), | |
|
78 | "access_count": 0, | |
|
79 | "access_time": 0, | |
|
80 | ||
|
81 | "size": 0 | |
|
82 | } | |
|
83 | if metadata: | |
|
84 | _metadata.update(metadata) | |
|
85 | ||
|
86 | read_iterator = iter(functools.partial(value_reader.read, 2**22), b'') | |
|
87 | size, sha256 = self._write_file(full_path, read_iterator, mode) | |
|
88 | _metadata['size'] = size | |
|
89 | _metadata['sha256'] = sha256 | |
|
90 | ||
|
91 | # after archive is finished, we create a key to save the presence of the binary file | |
|
92 | with self.fs.open(key_file_path, 'wb') as f: | |
|
93 | f.write(json.dumps(_metadata)) | |
|
94 | ||
|
95 | return key, filename, size, _metadata | |
|
96 | ||
|
97 | def fetch(self, *args, **kwargs): | |
|
98 | raise NotImplementedError | |
|
99 | ||
|
100 | def _fetch(self, key, retry, retry_attempts, retry_backoff, | |
|
101 | presigned_url_expires: int = 0) -> tuple[ShardFileReader, dict]: | |
|
102 | if retry is NOT_GIVEN: | |
|
103 | retry = False | |
|
104 | if retry_attempts is NOT_GIVEN: | |
|
105 | retry_attempts = 0 | |
|
106 | ||
|
107 | if retry and retry_attempts > 0: | |
|
108 | for attempt in range(1, retry_attempts + 1): | |
|
109 | if key in self: | |
|
110 | break | |
|
111 | # we didn't find the key, wait retry_backoff N seconds, and re-check | |
|
112 | time.sleep(retry_backoff) | |
|
113 | ||
|
114 | if key not in self: | |
|
115 | log.exception(f'requested key={key} not found in {self} retry={retry}, attempts={retry_attempts}') | |
|
116 | raise KeyError(key) | |
|
117 | ||
|
118 | key_file, key_file_path = self._get_keyfile(key) | |
|
119 | with self.fs.open(key_file_path, 'rb') as f: | |
|
120 | metadata = json.loads(f.read()) | |
|
121 | ||
|
122 | archive_path = metadata['archive_full_path'] | |
|
123 | if presigned_url_expires and presigned_url_expires > 0: | |
|
124 | metadata['url'] = self.fs.url(archive_path, expires=presigned_url_expires) | |
|
125 | ||
|
126 | try: | |
|
127 | return ShardFileReader(self.fs.open(archive_path, 'rb')), metadata | |
|
128 | finally: | |
|
129 | # update usage stats, count and accessed | |
|
130 | metadata["access_count"] = metadata.get("access_count", 0) + 1 | |
|
131 | metadata["access_time"] = time.time() | |
|
132 | log.debug('Updated %s with access snapshot, access_count=%s access_time=%s', | |
|
133 | key_file, metadata['access_count'], metadata['access_time']) | |
|
134 | with self.fs.open(key_file_path, 'wb') as f: | |
|
135 | f.write(json.dumps(metadata)) | |
|
136 | ||
|
137 | def remove(self, *args, **kwargs): | |
|
138 | raise NotImplementedError | |
|
139 | ||
|
140 | def _remove(self, key): | |
|
141 | if key not in self: | |
|
142 | log.exception(f'requested key={key} not found in {self}') | |
|
143 | raise KeyError(key) | |
|
144 | ||
|
145 | key_file, key_file_path = self._get_keyfile(key) | |
|
146 | with self.fs.open(key_file_path, 'rb') as f: | |
|
147 | metadata = json.loads(f.read()) | |
|
148 | ||
|
149 | archive_path = metadata['archive_full_path'] | |
|
150 | self.fs.rm(archive_path) | |
|
151 | self.fs.rm(key_file_path) | |
|
152 | return 1 | |
|
153 | ||
|
154 | @property | |
|
155 | def storage_medium(self): | |
|
156 | return getattr(self, self.storage_type) | |
|
157 | ||
|
158 | @property | |
|
159 | def key_suffix(self): | |
|
160 | return 'key.json' | |
|
161 | ||
|
162 | def __contains__(self, key): | |
|
163 | """Return `True` if `key` matching item is found in cache. | |
|
164 | ||
|
165 | :param key: key matching item | |
|
166 | :return: True if key matching item | |
|
167 | ||
|
168 | """ | |
|
169 | key_file, key_file_path = self._get_keyfile(key) | |
|
170 | return self.fs.exists(key_file_path) | |
|
171 | ||
|
172 | ||
|
173 | class BaseCache: | |
|
174 | _locking_url: str = '' | |
|
175 | _storage_path: str = '' | |
|
176 | _config: dict = {} | |
|
177 | retry = False | |
|
178 | retry_attempts: int = 0 | |
|
179 | retry_backoff: int | float = 1 | |
|
180 | _shards = tuple() | |
|
181 | shard_cls = BaseShard | |
|
182 | # define the presigned url expiration, 0 == disabled | |
|
183 | presigned_url_expires: int = 0 | |
|
184 | ||
|
185 | def __contains__(self, key): | |
|
186 | """Return `True` if `key` matching item is found in cache. | |
|
187 | ||
|
188 | :param key: key matching item | |
|
189 | :return: True if key matching item | |
|
190 | ||
|
191 | """ | |
|
192 | return self.has_key(key) | |
|
193 | ||
|
194 | def __repr__(self): | |
|
195 | return f'<{self.__class__.__name__}(storage={self._storage_path})>' | |
|
196 | ||
|
197 | @classmethod | |
|
198 | def gb_to_bytes(cls, gb): | |
|
199 | return gb * (1024 ** 3) | |
|
200 | ||
|
201 | @property | |
|
202 | def storage_path(self): | |
|
203 | return self._storage_path | |
|
204 | ||
|
205 | @classmethod | |
|
206 | def get_stats_db(cls): | |
|
207 | return StatsDB() | |
|
208 | ||
|
209 | def get_conf(self, key, pop=False): | |
|
210 | if key not in self._config: | |
|
211 | raise ValueError(f"No configuration key '{key}', please make sure it exists in archive_cache config") | |
|
212 | val = self._config[key] | |
|
213 | if pop: | |
|
214 | del self._config[key] | |
|
215 | return val | |
|
216 | ||
|
217 | def _get_shard(self, key) -> shard_cls: | |
|
218 | index = self._hash(key) % self._shard_count | |
|
219 | shard = self._shards[index] | |
|
220 | return shard | |
|
221 | ||
|
222 | def _get_size(self, shard, archive_path): | |
|
223 | raise NotImplementedError | |
|
224 | ||
|
225 | def store(self, key, value_reader, metadata=None): | |
|
226 | shard = self._get_shard(key) | |
|
227 | return shard.store(key, value_reader, metadata) | |
|
228 | ||
|
229 | def fetch(self, key, retry=NOT_GIVEN, retry_attempts=NOT_GIVEN) -> tuple[typing.BinaryIO, dict]: | |
|
230 | """ | |
|
231 | Return file handle corresponding to `key` from specific shard cache. | |
|
232 | """ | |
|
233 | if retry is NOT_GIVEN: | |
|
234 | retry = self.retry | |
|
235 | if retry_attempts is NOT_GIVEN: | |
|
236 | retry_attempts = self.retry_attempts | |
|
237 | retry_backoff = self.retry_backoff | |
|
238 | presigned_url_expires = self.presigned_url_expires | |
|
239 | ||
|
240 | shard = self._get_shard(key) | |
|
241 | return shard.fetch(key, retry=retry, | |
|
242 | retry_attempts=retry_attempts, | |
|
243 | retry_backoff=retry_backoff, | |
|
244 | presigned_url_expires=presigned_url_expires) | |
|
245 | ||
|
246 | def remove(self, key): | |
|
247 | shard = self._get_shard(key) | |
|
248 | return shard.remove(key) | |
|
249 | ||
|
250 | def has_key(self, archive_key): | |
|
251 | """Return `True` if `key` matching item is found in cache. | |
|
252 | ||
|
253 | :param archive_key: key for item, this is a unique archive name we want to store data under. e.g my-archive-svn.zip | |
|
254 | :return: True if key is found | |
|
255 | ||
|
256 | """ | |
|
257 | shard = self._get_shard(archive_key) | |
|
258 | return archive_key in shard | |
|
259 | ||
|
260 | def iter_keys(self): | |
|
261 | for shard in self._shards: | |
|
262 | if shard.fs.exists(shard.storage_medium): | |
|
263 | for path, _dirs, _files in shard.fs.walk(shard.storage_medium): | |
|
264 | for key_file_path in _files: | |
|
265 | if key_file_path.endswith(shard.key_suffix): | |
|
266 | yield shard, key_file_path | |
|
267 | ||
|
268 | def get_lock(self, lock_key): | |
|
269 | return GenerationLock(lock_key, self._locking_url) | |
|
270 | ||
|
271 | def evict(self, policy=None, size_limit=None) -> dict: | |
|
272 | """ | |
|
273 | Remove old items based on the conditions | |
|
274 | ||
|
275 | ||
|
276 | explanation of this algo: | |
|
277 | iterate over each shard, then for each shard iterate over the .key files | |
|
278 | read the key files metadata stored. This gives us a full list of keys, cached_archived, their size and | |
|
279 | access data, time creation, and access counts. | |
|
280 | ||
|
281 | Store that into a memory DB in order we can run different sorting strategies easily. | |
|
282 | Summing the size is a sum sql query. | |
|
283 | ||
|
284 | Then we run a sorting strategy based on eviction policy. | |
|
285 | We iterate over sorted keys, and remove each checking if we hit the overall limit. | |
|
286 | """ | |
|
287 | removal_info = { | |
|
288 | "removed_items": 0, | |
|
289 | "removed_size": 0 | |
|
290 | } | |
|
291 | policy = policy or self._eviction_policy | |
|
292 | size_limit = size_limit or self._cache_size_limit | |
|
293 | ||
|
294 | select_policy = EVICTION_POLICY[policy]['evict'] | |
|
295 | ||
|
296 | log.debug('Running eviction policy \'%s\', and checking for size limit: %s', | |
|
297 | policy, format_size(size_limit)) | |
|
298 | ||
|
299 | if select_policy is None: | |
|
300 | return removal_info | |
|
301 | ||
|
302 | db = self.get_stats_db() | |
|
303 | ||
|
304 | data = [] | |
|
305 | cnt = 1 | |
|
306 | ||
|
307 | for shard, key_file in self.iter_keys(): | |
|
308 | with shard.fs.open(os.path.join(shard.storage_medium, key_file), 'rb') as f: | |
|
309 | metadata = json.loads(f.read()) | |
|
310 | ||
|
311 | key_file_path = os.path.join(shard.storage_medium, key_file) | |
|
312 | ||
|
313 | archive_key = metadata['archive_key'] | |
|
314 | archive_path = metadata['archive_full_path'] | |
|
315 | ||
|
316 | size = metadata.get('size') | |
|
317 | if not size: | |
|
318 | # in case we don't have size re-calc it... | |
|
319 | size = self._get_size(shard, archive_path) | |
|
320 | ||
|
321 | data.append([ | |
|
322 | cnt, | |
|
323 | key_file, | |
|
324 | key_file_path, | |
|
325 | archive_key, | |
|
326 | archive_path, | |
|
327 | metadata.get('store_time', 0), | |
|
328 | metadata.get('access_time', 0), | |
|
329 | metadata.get('access_count', 0), | |
|
330 | size, | |
|
331 | ]) | |
|
332 | cnt += 1 | |
|
333 | ||
|
334 | # Insert bulk data using executemany | |
|
335 | db.bulk_insert(data) | |
|
336 | ||
|
337 | total_size = db.get_total_size() | |
|
338 | log.debug('Analyzed %s keys, occupying: %s, running eviction to match %s', | |
|
339 | len(data), format_size(total_size), format_size(size_limit)) | |
|
340 | ||
|
341 | removed_items = 0 | |
|
342 | removed_size = 0 | |
|
343 | for key_file, archive_key, size in db.get_sorted_keys(select_policy): | |
|
344 | # simulate removal impact BEFORE removal | |
|
345 | total_size -= size | |
|
346 | ||
|
347 | if total_size <= size_limit: | |
|
348 | # we obtained what we wanted... | |
|
349 | break | |
|
350 | ||
|
351 | self.remove(archive_key) | |
|
352 | removed_items += 1 | |
|
353 | removed_size += size | |
|
354 | removal_info['removed_items'] = removed_items | |
|
355 | removal_info['removed_size'] = removed_size | |
|
356 | log.debug('Removed %s cache archives, and reduced size by: %s', | |
|
357 | removed_items, format_size(removed_size)) | |
|
358 | return removal_info | |
|
359 | ||
|
360 | def get_statistics(self): | |
|
361 | total_files = 0 | |
|
362 | total_size = 0 | |
|
363 | meta = {} | |
|
364 | ||
|
365 | for shard, key_file in self.iter_keys(): | |
|
366 | json_key = f"{shard.storage_medium}/{key_file}" | |
|
367 | with shard.fs.open(json_key, 'rb') as f: | |
|
368 | total_files += 1 | |
|
369 | metadata = json.loads(f.read()) | |
|
370 | total_size += metadata['size'] | |
|
371 | ||
|
372 | return total_files, total_size, meta |
@@ -0,0 +1,177 b'' | |||
|
1 | # Copyright (C) 2015-2024 RhodeCode GmbH | |
|
2 | # | |
|
3 | # This program is free software: you can redistribute it and/or modify | |
|
4 | # it under the terms of the GNU Affero General Public License, version 3 | |
|
5 | # (only), as published by the Free Software Foundation. | |
|
6 | # | |
|
7 | # This program is distributed in the hope that it will be useful, | |
|
8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
|
9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
|
10 | # GNU General Public License for more details. | |
|
11 | # | |
|
12 | # You should have received a copy of the GNU Affero General Public License | |
|
13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. | |
|
14 | # | |
|
15 | # This program is dual-licensed. If you wish to learn more about the | |
|
16 | # RhodeCode Enterprise Edition, including its added features, Support services, | |
|
17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ | |
|
18 | ||
|
19 | import codecs | |
|
20 | import hashlib | |
|
21 | import logging | |
|
22 | import os | |
|
23 | import typing | |
|
24 | ||
|
25 | import fsspec | |
|
26 | ||
|
27 | from .base import BaseCache, BaseShard | |
|
28 | from ..utils import ShardFileReader, NOT_GIVEN | |
|
29 | from ...type_utils import str2bool | |
|
30 | ||
|
31 | log = logging.getLogger(__name__) | |
|
32 | ||
|
33 | ||
|
34 | class FileSystemShard(BaseShard): | |
|
35 | ||
|
36 | def __init__(self, index, directory, directory_folder, fs, **settings): | |
|
37 | self._index: int = index | |
|
38 | self._directory: str = directory | |
|
39 | self._directory_folder: str = directory_folder | |
|
40 | self.storage_type: str = 'directory' | |
|
41 | ||
|
42 | self.fs = fs | |
|
43 | ||
|
44 | @property | |
|
45 | def directory(self) -> str: | |
|
46 | """Cache directory final path.""" | |
|
47 | return os.path.join(self._directory, self._directory_folder) | |
|
48 | ||
|
49 | def _get_keyfile(self, archive_key) -> tuple[str, str]: | |
|
50 | key_file: str = f'{archive_key}.{self.key_suffix}' | |
|
51 | return key_file, os.path.join(self.directory, key_file) | |
|
52 | ||
|
53 | def _get_writer(self, path, mode): | |
|
54 | for count in range(1, 11): | |
|
55 | try: | |
|
56 | # Another cache may have deleted the directory before | |
|
57 | # the file could be opened. | |
|
58 | return self.fs.open(path, mode) | |
|
59 | except OSError: | |
|
60 | if count == 10: | |
|
61 | # Give up after 10 tries to open the file. | |
|
62 | raise | |
|
63 | continue | |
|
64 | ||
|
65 | def _write_file(self, full_path, iterator, mode): | |
|
66 | ||
|
67 | # ensure dir exists | |
|
68 | destination, _ = os.path.split(full_path) | |
|
69 | if not self.fs.exists(destination): | |
|
70 | self.fs.makedirs(destination) | |
|
71 | ||
|
72 | writer = self._get_writer(full_path, mode) | |
|
73 | ||
|
74 | digest = hashlib.sha256() | |
|
75 | with writer: | |
|
76 | size = 0 | |
|
77 | for chunk in iterator: | |
|
78 | size += len(chunk) | |
|
79 | digest.update(chunk) | |
|
80 | writer.write(chunk) | |
|
81 | writer.flush() | |
|
82 | # Get the file descriptor | |
|
83 | fd = writer.fileno() | |
|
84 | ||
|
85 | # Sync the file descriptor to disk, helps with NFS cases... | |
|
86 | os.fsync(fd) | |
|
87 | sha256 = digest.hexdigest() | |
|
88 | log.debug('written new archive cache under %s, sha256: %s', full_path, sha256) | |
|
89 | return size, sha256 | |
|
90 | ||
|
91 | def store(self, key, value_reader, metadata: dict | None = None): | |
|
92 | return self._store(key, value_reader, metadata, mode='xb') | |
|
93 | ||
|
94 | def fetch(self, key, retry=NOT_GIVEN, | |
|
95 | retry_attempts=NOT_GIVEN, retry_backoff=1, **kwargs) -> tuple[ShardFileReader, dict]: | |
|
96 | return self._fetch(key, retry, retry_attempts, retry_backoff) | |
|
97 | ||
|
98 | def remove(self, key): | |
|
99 | return self._remove(key) | |
|
100 | ||
|
101 | def random_filename(self): | |
|
102 | """Return filename and full-path tuple for file storage. | |
|
103 | ||
|
104 | Filename will be a randomly generated 28 character hexadecimal string | |
|
105 | with ".archive_cache" suffixed. Two levels of sub-directories will be used to | |
|
106 | reduce the size of directories. On older filesystems, lookups in | |
|
107 | directories with many files may be slow. | |
|
108 | """ | |
|
109 | ||
|
110 | hex_name = codecs.encode(os.urandom(16), 'hex').decode('utf-8') | |
|
111 | ||
|
112 | archive_name = hex_name[4:] + '.archive_cache' | |
|
113 | filename = f"{hex_name[:2]}/{hex_name[2:4]}/{archive_name}" | |
|
114 | ||
|
115 | full_path = os.path.join(self.directory, filename) | |
|
116 | return archive_name, full_path | |
|
117 | ||
|
118 | def __repr__(self): | |
|
119 | return f'{self.__class__.__name__}(index={self._index}, dir={self.directory})' | |
|
120 | ||
|
121 | ||
|
122 | class FileSystemFanoutCache(BaseCache): | |
|
123 | shard_name: str = 'shard_{:03d}' | |
|
124 | shard_cls = FileSystemShard | |
|
125 | ||
|
126 | def __init__(self, locking_url, **settings): | |
|
127 | """ | |
|
128 | Initialize file system cache instance. | |
|
129 | ||
|
130 | :param str locking_url: redis url for a lock | |
|
131 | :param settings: settings dict | |
|
132 | ||
|
133 | """ | |
|
134 | self._locking_url = locking_url | |
|
135 | self._config = settings | |
|
136 | cache_dir = self.get_conf('archive_cache.filesystem.store_dir') | |
|
137 | directory = str(cache_dir) | |
|
138 | directory = os.path.expanduser(directory) | |
|
139 | directory = os.path.expandvars(directory) | |
|
140 | self._directory = directory | |
|
141 | self._storage_path = directory # common path for all from BaseCache | |
|
142 | ||
|
143 | self._shard_count = int(self.get_conf('archive_cache.filesystem.cache_shards', pop=True)) | |
|
144 | if self._shard_count < 1: | |
|
145 | raise ValueError('cache_shards must be 1 or more') | |
|
146 | ||
|
147 | self._eviction_policy = self.get_conf('archive_cache.filesystem.eviction_policy', pop=True) | |
|
148 | self._cache_size_limit = self.gb_to_bytes(int(self.get_conf('archive_cache.filesystem.cache_size_gb'))) | |
|
149 | ||
|
150 | self.retry = str2bool(self.get_conf('archive_cache.filesystem.retry', pop=True)) | |
|
151 | self.retry_attempts = int(self.get_conf('archive_cache.filesystem.retry_attempts', pop=True)) | |
|
152 | self.retry_backoff = int(self.get_conf('archive_cache.filesystem.retry_backoff', pop=True)) | |
|
153 | ||
|
154 | log.debug('Initializing %s archival cache instance', self) | |
|
155 | fs = fsspec.filesystem('file') | |
|
156 | # check if it's ok to write, and re-create the archive cache main dir | |
|
157 | # A directory is the virtual equivalent of a physical file cabinet. | |
|
158 | # In other words, it's a container for organizing digital data. | |
|
159 | # Unlike a folder, which can only store files, a directory can store files, | |
|
160 | # subdirectories, and other directories. | |
|
161 | if not fs.exists(self._directory): | |
|
162 | fs.makedirs(self._directory, exist_ok=True) | |
|
163 | ||
|
164 | self._shards = tuple( | |
|
165 | self.shard_cls( | |
|
166 | index=num, | |
|
167 | directory=directory, | |
|
168 | directory_folder=self.shard_name.format(num), | |
|
169 | fs=fs, | |
|
170 | **settings, | |
|
171 | ) | |
|
172 | for num in range(self._shard_count) | |
|
173 | ) | |
|
174 | self._hash = self._shards[0].hash | |
|
175 | ||
|
176 | def _get_size(self, shard, archive_path): | |
|
177 | return os.stat(archive_path).st_size |
@@ -0,0 +1,173 b'' | |||
|
1 | # Copyright (C) 2015-2024 RhodeCode GmbH | |
|
2 | # | |
|
3 | # This program is free software: you can redistribute it and/or modify | |
|
4 | # it under the terms of the GNU Affero General Public License, version 3 | |
|
5 | # (only), as published by the Free Software Foundation. | |
|
6 | # | |
|
7 | # This program is distributed in the hope that it will be useful, | |
|
8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
|
9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
|
10 | # GNU General Public License for more details. | |
|
11 | # | |
|
12 | # You should have received a copy of the GNU Affero General Public License | |
|
13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. | |
|
14 | # | |
|
15 | # This program is dual-licensed. If you wish to learn more about the | |
|
16 | # RhodeCode Enterprise Edition, including its added features, Support services, | |
|
17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ | |
|
18 | ||
|
19 | import codecs | |
|
20 | import hashlib | |
|
21 | import logging | |
|
22 | import os | |
|
23 | import typing | |
|
24 | ||
|
25 | import fsspec | |
|
26 | ||
|
27 | from .base import BaseCache, BaseShard | |
|
28 | from ..utils import ShardFileReader, NOT_GIVEN | |
|
29 | from ...type_utils import str2bool | |
|
30 | ||
|
31 | log = logging.getLogger(__name__) | |
|
32 | ||
|
33 | ||
|
34 | class S3Shard(BaseShard): | |
|
35 | ||
|
36 | def __init__(self, index, bucket, bucket_folder, fs, **settings): | |
|
37 | self._index: int = index | |
|
38 | self._bucket_folder: str = bucket_folder | |
|
39 | self.storage_type: str = 'bucket' | |
|
40 | self._bucket_main: str = bucket | |
|
41 | ||
|
42 | self.fs = fs | |
|
43 | ||
|
44 | @property | |
|
45 | def bucket(self) -> str: | |
|
46 | """Cache bucket final path.""" | |
|
47 | return os.path.join(self._bucket_main, self._bucket_folder) | |
|
48 | ||
|
49 | def _get_keyfile(self, archive_key) -> tuple[str, str]: | |
|
50 | key_file: str = f'{archive_key}-{self.key_suffix}' | |
|
51 | return key_file, os.path.join(self.bucket, key_file) | |
|
52 | ||
|
53 | def _get_writer(self, path, mode): | |
|
54 | return self.fs.open(path, 'wb') | |
|
55 | ||
|
56 | def _write_file(self, full_path, iterator, mode): | |
|
57 | ||
|
58 | # ensure folder in bucket exists | |
|
59 | destination = self.bucket | |
|
60 | if not self.fs.exists(destination): | |
|
61 | self.fs.mkdir(destination, s3_additional_kwargs={}) | |
|
62 | ||
|
63 | writer = self._get_writer(full_path, mode) | |
|
64 | ||
|
65 | digest = hashlib.sha256() | |
|
66 | with writer: | |
|
67 | size = 0 | |
|
68 | for chunk in iterator: | |
|
69 | size += len(chunk) | |
|
70 | digest.update(chunk) | |
|
71 | writer.write(chunk) | |
|
72 | ||
|
73 | sha256 = digest.hexdigest() | |
|
74 | log.debug('written new archive cache under %s, sha256: %s', full_path, sha256) | |
|
75 | return size, sha256 | |
|
76 | ||
|
77 | def store(self, key, value_reader, metadata: dict | None = None): | |
|
78 | return self._store(key, value_reader, metadata, mode='wb') | |
|
79 | ||
|
80 | def fetch(self, key, retry=NOT_GIVEN, | |
|
81 | retry_attempts=NOT_GIVEN, retry_backoff=1, | |
|
82 | presigned_url_expires: int = 0) -> tuple[ShardFileReader, dict]: | |
|
83 | return self._fetch(key, retry, retry_attempts, retry_backoff, presigned_url_expires=presigned_url_expires) | |
|
84 | ||
|
85 | def remove(self, key): | |
|
86 | return self._remove(key) | |
|
87 | ||
|
88 | def random_filename(self): | |
|
89 | """Return filename and full-path tuple for file storage. | |
|
90 | ||
|
91 | Filename will be a randomly generated 28 character hexadecimal string | |
|
92 | with ".archive_cache" suffixed. Two levels of sub-directories will be used to | |
|
93 | reduce the size of directories. On older filesystems, lookups in | |
|
94 | directories with many files may be slow. | |
|
95 | """ | |
|
96 | ||
|
97 | hex_name = codecs.encode(os.urandom(16), 'hex').decode('utf-8') | |
|
98 | ||
|
99 | archive_name = hex_name[4:] + '.archive_cache' | |
|
100 | filename = f"{hex_name[:2]}-{hex_name[2:4]}-{archive_name}" | |
|
101 | ||
|
102 | full_path = os.path.join(self.bucket, filename) | |
|
103 | return archive_name, full_path | |
|
104 | ||
|
105 | def __repr__(self): | |
|
106 | return f'{self.__class__.__name__}(index={self._index}, bucket={self.bucket})' | |
|
107 | ||
|
108 | ||
|
109 | class ObjectStoreCache(BaseCache): | |
|
110 | shard_name: str = 'shard-{:03d}' | |
|
111 | shard_cls = S3Shard | |
|
112 | ||
|
113 | def __init__(self, locking_url, **settings): | |
|
114 | """ | |
|
115 | Initialize objectstore cache instance. | |
|
116 | ||
|
117 | :param str locking_url: redis url for a lock | |
|
118 | :param settings: settings dict | |
|
119 | ||
|
120 | """ | |
|
121 | self._locking_url = locking_url | |
|
122 | self._config = settings | |
|
123 | ||
|
124 | objectstore_url = self.get_conf('archive_cache.objectstore.url') | |
|
125 | self._storage_path = objectstore_url # common path for all from BaseCache | |
|
126 | ||
|
127 | self._shard_count = int(self.get_conf('archive_cache.objectstore.bucket_shards', pop=True)) | |
|
128 | if self._shard_count < 1: | |
|
129 | raise ValueError('cache_shards must be 1 or more') | |
|
130 | ||
|
131 | self._bucket = settings.pop('archive_cache.objectstore.bucket') | |
|
132 | if not self._bucket: | |
|
133 | raise ValueError('archive_cache.objectstore.bucket needs to have a value') | |
|
134 | ||
|
135 | self._eviction_policy = self.get_conf('archive_cache.objectstore.eviction_policy', pop=True) | |
|
136 | self._cache_size_limit = self.gb_to_bytes(int(self.get_conf('archive_cache.objectstore.cache_size_gb'))) | |
|
137 | ||
|
138 | self.retry = str2bool(self.get_conf('archive_cache.objectstore.retry', pop=True)) | |
|
139 | self.retry_attempts = int(self.get_conf('archive_cache.objectstore.retry_attempts', pop=True)) | |
|
140 | self.retry_backoff = int(self.get_conf('archive_cache.objectstore.retry_backoff', pop=True)) | |
|
141 | ||
|
142 | endpoint_url = settings.pop('archive_cache.objectstore.url') | |
|
143 | key = settings.pop('archive_cache.objectstore.key') | |
|
144 | secret = settings.pop('archive_cache.objectstore.secret') | |
|
145 | region = settings.pop('archive_cache.objectstore.region') | |
|
146 | ||
|
147 | log.debug('Initializing %s archival cache instance', self) | |
|
148 | ||
|
149 | fs = fsspec.filesystem( | |
|
150 | 's3', anon=False, endpoint_url=endpoint_url, key=key, secret=secret, client_kwargs={'region_name': region} | |
|
151 | ) | |
|
152 | ||
|
153 | # init main bucket | |
|
154 | if not fs.exists(self._bucket): | |
|
155 | fs.mkdir(self._bucket) | |
|
156 | ||
|
157 | self._shards = tuple( | |
|
158 | self.shard_cls( | |
|
159 | index=num, | |
|
160 | bucket=self._bucket, | |
|
161 | bucket_folder=self.shard_name.format(num), | |
|
162 | fs=fs, | |
|
163 | **settings, | |
|
164 | ) | |
|
165 | for num in range(self._shard_count) | |
|
166 | ) | |
|
167 | self._hash = self._shards[0].hash | |
|
168 | ||
|
169 | def _get_size(self, shard, archive_path): | |
|
170 | return shard.fs.info(archive_path)['size'] | |
|
171 | ||
|
172 | def set_presigned_url_expiry(self, val: int) -> None: | |
|
173 | self.presigned_url_expires = val |
@@ -0,0 +1,62 b'' | |||
|
1 | # Copyright (C) 2015-2024 RhodeCode GmbH | |
|
2 | # | |
|
3 | # This program is free software: you can redistribute it and/or modify | |
|
4 | # it under the terms of the GNU Affero General Public License, version 3 | |
|
5 | # (only), as published by the Free Software Foundation. | |
|
6 | # | |
|
7 | # This program is distributed in the hope that it will be useful, | |
|
8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
|
9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
|
10 | # GNU General Public License for more details. | |
|
11 | # | |
|
12 | # You should have received a copy of the GNU Affero General Public License | |
|
13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. | |
|
14 | # | |
|
15 | # This program is dual-licensed. If you wish to learn more about the | |
|
16 | # RhodeCode Enterprise Edition, including its added features, Support services, | |
|
17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ | |
|
18 | ||
|
19 | import redis | |
|
20 | from .._vendor import redis_lock | |
|
21 | ||
|
22 | ||
|
23 | class ArchiveCacheGenerationLock(Exception): | |
|
24 | pass | |
|
25 | ||
|
26 | ||
|
27 | class GenerationLock: | |
|
28 | """ | |
|
29 | Locking mechanism that detects if a lock is acquired | |
|
30 | ||
|
31 | with GenerationLock(lock_key): | |
|
32 | compute_archive() | |
|
33 | """ | |
|
34 | lock_timeout = 7200 | |
|
35 | ||
|
36 | def __init__(self, lock_key, url): | |
|
37 | self.lock_key = lock_key | |
|
38 | self._create_client(url) | |
|
39 | self.lock = self.get_lock() | |
|
40 | ||
|
41 | def _create_client(self, url): | |
|
42 | connection_pool = redis.ConnectionPool.from_url(url) | |
|
43 | self.writer_client = redis.StrictRedis( | |
|
44 | connection_pool=connection_pool | |
|
45 | ) | |
|
46 | self.reader_client = self.writer_client | |
|
47 | ||
|
48 | def get_lock(self): | |
|
49 | return redis_lock.Lock( | |
|
50 | redis_client=self.writer_client, | |
|
51 | name=self.lock_key, | |
|
52 | expire=self.lock_timeout, | |
|
53 | strict=True | |
|
54 | ) | |
|
55 | ||
|
56 | def __enter__(self): | |
|
57 | acquired = self.lock.acquire(blocking=False) | |
|
58 | if not acquired: | |
|
59 | raise ArchiveCacheGenerationLock('Failed to create a lock') | |
|
60 | ||
|
61 | def __exit__(self, exc_type, exc_val, exc_tb): | |
|
62 | self.lock.release() |
@@ -0,0 +1,134 b'' | |||
|
1 | # Copyright (C) 2015-2024 RhodeCode GmbH | |
|
2 | # | |
|
3 | # This program is free software: you can redistribute it and/or modify | |
|
4 | # it under the terms of the GNU Affero General Public License, version 3 | |
|
5 | # (only), as published by the Free Software Foundation. | |
|
6 | # | |
|
7 | # This program is distributed in the hope that it will be useful, | |
|
8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
|
9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
|
10 | # GNU General Public License for more details. | |
|
11 | # | |
|
12 | # You should have received a copy of the GNU Affero General Public License | |
|
13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. | |
|
14 | # | |
|
15 | # This program is dual-licensed. If you wish to learn more about the | |
|
16 | # RhodeCode Enterprise Edition, including its added features, Support services, | |
|
17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ | |
|
18 | ||
|
19 | import sqlite3 | |
|
20 | import s3fs.core | |
|
21 | ||
|
22 | NOT_GIVEN = -917 | |
|
23 | ||
|
24 | ||
|
25 | EVICTION_POLICY = { | |
|
26 | 'none': { | |
|
27 | 'evict': None, | |
|
28 | }, | |
|
29 | 'least-recently-stored': { | |
|
30 | 'evict': 'SELECT {fields} FROM archive_cache ORDER BY store_time', | |
|
31 | }, | |
|
32 | 'least-recently-used': { | |
|
33 | 'evict': 'SELECT {fields} FROM archive_cache ORDER BY access_time', | |
|
34 | }, | |
|
35 | 'least-frequently-used': { | |
|
36 | 'evict': 'SELECT {fields} FROM archive_cache ORDER BY access_count', | |
|
37 | }, | |
|
38 | } | |
|
39 | ||
|
40 | ||
|
41 | def archive_iterator(_reader, block_size: int = 4096 * 512): | |
|
42 | # 4096 * 64 = 64KB | |
|
43 | while 1: | |
|
44 | data = _reader.read(block_size) | |
|
45 | if not data: | |
|
46 | break | |
|
47 | yield data | |
|
48 | ||
|
49 | ||
|
50 | def format_size(size): | |
|
51 | # Convert size in bytes to a human-readable format (e.g., KB, MB, GB) | |
|
52 | for unit in ['B', 'KB', 'MB', 'GB', 'TB']: | |
|
53 | if size < 1024: | |
|
54 | return f"{size:.2f} {unit}" | |
|
55 | size /= 1024 | |
|
56 | ||
|
57 | ||
|
58 | class StatsDB: | |
|
59 | ||
|
60 | def __init__(self): | |
|
61 | self.connection = sqlite3.connect(':memory:') | |
|
62 | self._init_db() | |
|
63 | ||
|
64 | def _init_db(self): | |
|
65 | qry = ''' | |
|
66 | CREATE TABLE IF NOT EXISTS archive_cache ( | |
|
67 | rowid INTEGER PRIMARY KEY, | |
|
68 | key_file TEXT, | |
|
69 | key_file_path TEXT, | |
|
70 | archive_key TEXT, | |
|
71 | archive_path TEXT, | |
|
72 | store_time REAL, | |
|
73 | access_time REAL, | |
|
74 | access_count INTEGER DEFAULT 0, | |
|
75 | size INTEGER DEFAULT 0 | |
|
76 | ) | |
|
77 | ''' | |
|
78 | ||
|
79 | self.sql(qry) | |
|
80 | self.connection.commit() | |
|
81 | ||
|
82 | @property | |
|
83 | def sql(self): | |
|
84 | return self.connection.execute | |
|
85 | ||
|
86 | def bulk_insert(self, rows): | |
|
87 | qry = ''' | |
|
88 | INSERT INTO archive_cache ( | |
|
89 | rowid, | |
|
90 | key_file, | |
|
91 | key_file_path, | |
|
92 | archive_key, | |
|
93 | archive_path, | |
|
94 | store_time, | |
|
95 | access_time, | |
|
96 | access_count, | |
|
97 | size | |
|
98 | ) | |
|
99 | VALUES ( | |
|
100 | ?, ?, ?, ?, ?, ?, ?, ?, ? | |
|
101 | ) | |
|
102 | ''' | |
|
103 | cursor = self.connection.cursor() | |
|
104 | cursor.executemany(qry, rows) | |
|
105 | self.connection.commit() | |
|
106 | ||
|
107 | def get_total_size(self): | |
|
108 | qry = 'SELECT COALESCE(SUM(size), 0) FROM archive_cache' | |
|
109 | ((total_size,),) = self.sql(qry).fetchall() | |
|
110 | return total_size | |
|
111 | ||
|
112 | def get_sorted_keys(self, select_policy): | |
|
113 | select_policy_qry = select_policy.format(fields='key_file, archive_key, size') | |
|
114 | return self.sql(select_policy_qry).fetchall() | |
|
115 | ||
|
116 | ||
|
117 | class ShardFileReader: | |
|
118 | ||
|
119 | def __init__(self, file_like_reader): | |
|
120 | self._file_like_reader = file_like_reader | |
|
121 | ||
|
122 | def __getattr__(self, item): | |
|
123 | if isinstance(self._file_like_reader, s3fs.core.S3File): | |
|
124 | match item: | |
|
125 | case 'name': | |
|
126 | # S3 FileWrapper doesn't support name attribute, and we use it | |
|
127 | return self._file_like_reader.full_name | |
|
128 | case _: | |
|
129 | return getattr(self._file_like_reader, item) | |
|
130 | else: | |
|
131 | return getattr(self._file_like_reader, item) | |
|
132 | ||
|
133 | def __repr__(self): | |
|
134 | return f'<{self.__class__.__name__}={self._file_like_reader}>' |
@@ -0,0 +1,111 b'' | |||
|
1 | # RhodeCode VCSServer provides access to different vcs backends via network. | |
|
2 | # Copyright (C) 2014-2023 RhodeCode GmbH | |
|
3 | # | |
|
4 | # This program is free software; you can redistribute it and/or modify | |
|
5 | # it under the terms of the GNU General Public License as published by | |
|
6 | # the Free Software Foundation; either version 3 of the License, or | |
|
7 | # (at your option) any later version. | |
|
8 | # | |
|
9 | # This program is distributed in the hope that it will be useful, | |
|
10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
|
11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
|
12 | # GNU General Public License for more details. | |
|
13 | # | |
|
14 | # You should have received a copy of the GNU General Public License | |
|
15 | # along with this program; if not, write to the Free Software Foundation, | |
|
16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | |
|
17 | ||
|
18 | import logging | |
|
19 | import redis | |
|
20 | ||
|
21 | from ..lib import rc_cache | |
|
22 | from ..lib.ext_json import json | |
|
23 | ||
|
24 | ||
|
25 | log = logging.getLogger(__name__) | |
|
26 | ||
|
27 | redis_client = None | |
|
28 | ||
|
29 | ||
|
30 | class RedisTxnClient: | |
|
31 | ||
|
32 | def __init__(self, url): | |
|
33 | self.url = url | |
|
34 | self._create_client(url) | |
|
35 | ||
|
36 | def _create_client(self, url): | |
|
37 | connection_pool = redis.ConnectionPool.from_url(url) | |
|
38 | self.writer_client = redis.StrictRedis( | |
|
39 | connection_pool=connection_pool | |
|
40 | ) | |
|
41 | self.reader_client = self.writer_client | |
|
42 | ||
|
43 | def set(self, key, value): | |
|
44 | self.writer_client.set(key, value) | |
|
45 | ||
|
46 | def get(self, key): | |
|
47 | return self.reader_client.get(key) | |
|
48 | ||
|
49 | def delete(self, key): | |
|
50 | self.writer_client.delete(key) | |
|
51 | ||
|
52 | ||
|
53 | def get_redis_client(url=''): | |
|
54 | ||
|
55 | global redis_client | |
|
56 | if redis_client is not None: | |
|
57 | return redis_client | |
|
58 | if not url: | |
|
59 | from vcsserver import CONFIG | |
|
60 | url = CONFIG['vcs.svn.redis_conn'] | |
|
61 | redis_client = RedisTxnClient(url) | |
|
62 | return redis_client | |
|
63 | ||
|
64 | ||
|
65 | def get_txn_id_data_key(repo_path, svn_txn_id): | |
|
66 | log.debug('svn-txn-id: %s, obtaining data path', svn_txn_id) | |
|
67 | repo_key = rc_cache.utils.compute_key_from_params(repo_path) | |
|
68 | final_key = f'{repo_key}.{svn_txn_id}.svn_txn_id' | |
|
69 | log.debug('computed final key: %s', final_key) | |
|
70 | ||
|
71 | return final_key | |
|
72 | ||
|
73 | ||
|
74 | def store_txn_id_data(repo_path, svn_txn_id, data_dict): | |
|
75 | log.debug('svn-txn-id: %s, storing data', svn_txn_id) | |
|
76 | ||
|
77 | if not svn_txn_id: | |
|
78 | log.warning('Cannot store txn_id because it is empty') | |
|
79 | return | |
|
80 | ||
|
81 | redis_conn = get_redis_client() | |
|
82 | ||
|
83 | store_key = get_txn_id_data_key(repo_path, svn_txn_id) | |
|
84 | store_data = json.dumps(data_dict) | |
|
85 | redis_conn.set(store_key, store_data) | |
|
86 | ||
|
87 | ||
|
88 | def get_txn_id_from_store(repo_path, svn_txn_id, rm_on_read=False): | |
|
89 | """ | |
|
90 | Reads txn_id from store and if present returns the data for callback manager | |
|
91 | """ | |
|
92 | log.debug('svn-txn-id: %s, retrieving data', svn_txn_id) | |
|
93 | redis_conn = get_redis_client() | |
|
94 | ||
|
95 | store_key = get_txn_id_data_key(repo_path, svn_txn_id) | |
|
96 | data = {} | |
|
97 | redis_conn.get(store_key) | |
|
98 | raw_data = 'not-set' | |
|
99 | try: | |
|
100 | raw_data = redis_conn.get(store_key) | |
|
101 | if not raw_data: | |
|
102 | raise ValueError(f'Failed to get txn_id metadata, from store: {store_key}') | |
|
103 | data = json.loads(raw_data) | |
|
104 | except Exception: | |
|
105 | log.exception('Failed to get txn_id metadata: %s', raw_data) | |
|
106 | ||
|
107 | if rm_on_read: | |
|
108 | log.debug('Cleaning up txn_id at %s', store_key) | |
|
109 | redis_conn.delete(store_key) | |
|
110 | ||
|
111 | return data |
@@ -1,5 +1,5 b'' | |||
|
1 | 1 | [bumpversion] |
|
2 |
current_version = 5. |
|
|
2 | current_version = 5.1.0 | |
|
3 | 3 | message = release: Bump version {current_version} to {new_version} |
|
4 | 4 | |
|
5 | 5 | [bumpversion:file:vcsserver/VERSION] |
@@ -1,139 +1,144 b'' | |||
|
1 | 1 | # required for pushd to work.. |
|
2 | 2 | SHELL = /bin/bash |
|
3 | 3 | |
|
4 | 4 | |
|
5 | 5 | # set by: PATH_TO_OUTDATED_PACKAGES=/some/path/outdated_packages.py |
|
6 | 6 | OUTDATED_PACKAGES = ${PATH_TO_OUTDATED_PACKAGES} |
|
7 | 7 | |
|
8 | 8 | .PHONY: clean |
|
9 | 9 | ## Cleanup compiled and cache py files |
|
10 | 10 | clean: |
|
11 | 11 | make test-clean |
|
12 | 12 | find . -type f \( -iname '*.c' -o -iname '*.pyc' -o -iname '*.so' -o -iname '*.orig' \) -exec rm '{}' ';' |
|
13 | 13 | find . -type d -name "build" -prune -exec rm -rf '{}' ';' |
|
14 | 14 | |
|
15 | 15 | |
|
16 | 16 | .PHONY: test |
|
17 | 17 | ## run test-clean and tests |
|
18 | 18 | test: |
|
19 | 19 | make test-clean |
|
20 | 20 | make test-only |
|
21 | 21 | |
|
22 | 22 | |
|
23 | 23 | .PHONY: test-clean |
|
24 | 24 | ## run test-clean and tests |
|
25 | 25 | test-clean: |
|
26 | 26 | rm -rf coverage.xml htmlcov junit.xml pylint.log result |
|
27 | 27 | find . -type d -name "__pycache__" -prune -exec rm -rf '{}' ';' |
|
28 | 28 | find . -type f \( -iname '.coverage.*' \) -exec rm '{}' ';' |
|
29 | 29 | |
|
30 | 30 | |
|
31 | 31 | .PHONY: test-only |
|
32 | 32 | ## Run tests only without cleanup |
|
33 | 33 | test-only: |
|
34 | 34 | PYTHONHASHSEED=random \ |
|
35 | 35 | py.test -x -vv -r xw -p no:sugar \ |
|
36 | 36 | --cov-report=term-missing --cov-report=html \ |
|
37 | 37 | --cov=vcsserver vcsserver |
|
38 | 38 | |
|
39 | 39 | |
|
40 | 40 | .PHONY: ruff-check |
|
41 | 41 | ## run a ruff analysis |
|
42 | 42 | ruff-check: |
|
43 | 43 | ruff check --ignore F401 --ignore I001 --ignore E402 --ignore E501 --ignore F841 --exclude rhodecode/lib/dbmigrate --exclude .eggs --exclude .dev . |
|
44 | 44 | |
|
45 | ||
|
46 | 45 | .PHONY: pip-packages |
|
47 | 46 | ## Show outdated packages |
|
48 | 47 | pip-packages: |
|
49 | 48 | python ${OUTDATED_PACKAGES} |
|
50 | 49 | |
|
51 | 50 | |
|
52 | 51 | .PHONY: build |
|
53 | 52 | ## Build sdist/egg |
|
54 | 53 | build: |
|
55 | 54 | python -m build |
|
56 | 55 | |
|
57 | 56 | |
|
58 | 57 | .PHONY: dev-sh |
|
59 | 58 | ## make dev-sh |
|
60 | 59 | dev-sh: |
|
61 | 60 | sudo echo "deb [trusted=yes] https://apt.fury.io/rsteube/ /" | sudo tee -a "/etc/apt/sources.list.d/fury.list" |
|
62 | 61 | sudo apt-get update |
|
63 | 62 | sudo apt-get install -y zsh carapace-bin |
|
64 | 63 | rm -rf /home/rhodecode/.oh-my-zsh |
|
65 | 64 | curl https://raw.githubusercontent.com/robbyrussell/oh-my-zsh/master/tools/install.sh | sh |
|
66 | echo "source <(carapace _carapace)" > /home/rhodecode/.zsrc | |
|
67 | PROMPT='%(?.%F{green}β.%F{red}?%?)%f %B%F{240}%1~%f%b %# ' zsh | |
|
65 | @echo "source <(carapace _carapace)" > /home/rhodecode/.zsrc | |
|
66 | @echo "${RC_DEV_CMD_HELP}" | |
|
67 | @PROMPT='%(?.%F{green}β.%F{red}?%?)%f %B%F{240}%1~%f%b %# ' zsh | |
|
68 | ||
|
69 | ||
|
70 | .PHONY: dev-cleanup | |
|
71 | ## Cleanup: pip freeze | grep -v "^-e" | grep -v "@" | xargs pip uninstall -y | |
|
72 | dev-cleanup: | |
|
73 | pip freeze | grep -v "^-e" | grep -v "@" | xargs pip uninstall -y | |
|
74 | rm -rf /tmp/* | |
|
68 | 75 | |
|
69 | 76 | |
|
70 | 77 | .PHONY: dev-env |
|
71 | 78 | ## make dev-env based on the requirements files and install develop of packages |
|
72 | 79 | ## Cleanup: pip freeze | grep -v "^-e" | grep -v "@" | xargs pip uninstall -y |
|
73 | 80 | dev-env: |
|
81 | sudo -u root chown rhodecode:rhodecode /home/rhodecode/.cache/pip/ | |
|
74 | 82 |
|
|
75 | 83 | pip wheel --wheel-dir=/home/rhodecode/.cache/pip/wheels -r requirements.txt -r requirements_test.txt -r requirements_debug.txt |
|
76 | 84 | pip install --no-index --find-links=/home/rhodecode/.cache/pip/wheels -r requirements.txt -r requirements_test.txt -r requirements_debug.txt |
|
77 | 85 | pip install -e . |
|
78 | 86 | |
|
79 | 87 | |
|
80 | 88 | .PHONY: sh |
|
81 | 89 | ## shortcut for make dev-sh dev-env |
|
82 | 90 | sh: |
|
83 | 91 | make dev-env |
|
84 | 92 | make dev-sh |
|
85 | 93 | |
|
86 | 94 | |
|
87 | .PHONY: dev-srv | |
|
88 | ## run develop server instance, docker exec -it $(docker ps -q --filter 'name=dev-enterprise-ce') /bin/bash | |
|
89 | dev-srv: | |
|
90 | pserve --reload .dev/dev.ini | |
|
95 | ## Allows changes of workers e.g make dev-srv-g workers=2 | |
|
96 | workers?=1 | |
|
91 | 97 | |
|
92 | ||
|
93 | .PHONY: dev-srv-g | |
|
94 | ## run gunicorn multi process workers | |
|
95 | dev-srv-g: | |
|
96 | gunicorn --workers=4 --paste .dev/dev.ini --bind=0.0.0.0:10010 --config=.dev/gunicorn_config.py | |
|
98 | .PHONY: dev-srv | |
|
99 | ## run gunicorn web server with reloader, use workers=N to set multiworker mode | |
|
100 | dev-srv: | |
|
101 | gunicorn --paste=.dev/dev.ini --bind=0.0.0.0:10010 --config=.dev/gunicorn_config.py --reload --workers=$(workers) | |
|
97 | 102 | |
|
98 | 103 | |
|
99 | 104 | # Default command on calling make |
|
100 | 105 | .DEFAULT_GOAL := show-help |
|
101 | 106 | |
|
102 | 107 | .PHONY: show-help |
|
103 | 108 | show-help: |
|
104 | 109 | @echo "$$(tput bold)Available rules:$$(tput sgr0)" |
|
105 | 110 | @echo |
|
106 | 111 | @sed -n -e "/^## / { \ |
|
107 | 112 | h; \ |
|
108 | 113 | s/.*//; \ |
|
109 | 114 | :doc" \ |
|
110 | 115 | -e "H; \ |
|
111 | 116 | n; \ |
|
112 | 117 | s/^## //; \ |
|
113 | 118 | t doc" \ |
|
114 | 119 | -e "s/:.*//; \ |
|
115 | 120 | G; \ |
|
116 | 121 | s/\\n## /---/; \ |
|
117 | 122 | s/\\n/ /g; \ |
|
118 | 123 | p; \ |
|
119 | 124 | }" ${MAKEFILE_LIST} \ |
|
120 | 125 | | LC_ALL='C' sort --ignore-case \ |
|
121 | 126 | | awk -F '---' \ |
|
122 | 127 | -v ncol=$$(tput cols) \ |
|
123 | 128 | -v indent=19 \ |
|
124 | 129 | -v col_on="$$(tput setaf 6)" \ |
|
125 | 130 | -v col_off="$$(tput sgr0)" \ |
|
126 | 131 | '{ \ |
|
127 | 132 | printf "%s%*s%s ", col_on, -indent, $$1, col_off; \ |
|
128 | 133 | n = split($$2, words, " "); \ |
|
129 | 134 | line_length = ncol - indent; \ |
|
130 | 135 | for (i = 1; i <= n; i++) { \ |
|
131 | 136 | line_length -= length(words[i]) + 1; \ |
|
132 | 137 | if (line_length <= 0) { \ |
|
133 | 138 | line_length = ncol - indent - length(words[i]) - 1; \ |
|
134 | 139 | printf "\n%*s ", -indent, " "; \ |
|
135 | 140 | } \ |
|
136 | 141 | printf "%s ", words[i]; \ |
|
137 | 142 | } \ |
|
138 | 143 | printf "\n"; \ |
|
139 | 144 | }' |
@@ -1,204 +1,191 b'' | |||
|
1 | # | |
|
2 | 1 | |
|
3 | 2 | ; ################################# |
|
4 | 3 | ; RHODECODE VCSSERVER CONFIGURATION |
|
5 | 4 | ; ################################# |
|
6 | 5 | |
|
7 | 6 | [server:main] |
|
8 | ; COMMON HOST/IP CONFIG | |
|
7 | ; COMMON HOST/IP CONFIG, This applies mostly to develop setup, | |
|
8 | ; Host port for gunicorn are controlled by gunicorn_conf.py | |
|
9 | 9 | host = 0.0.0.0 |
|
10 | 10 | port = 10010 |
|
11 | 11 | |
|
12 | ; ################################################## | |
|
13 | ; WAITRESS WSGI SERVER - Recommended for Development | |
|
14 | ; ################################################## | |
|
15 | ||
|
16 | ; use server type | |
|
17 | use = egg:waitress#main | |
|
18 | ||
|
19 | ; number of worker threads | |
|
20 | threads = 5 | |
|
21 | ||
|
22 | ; MAX BODY SIZE 100GB | |
|
23 | max_request_body_size = 107374182400 | |
|
24 | ||
|
25 | ; Use poll instead of select, fixes file descriptors limits problems. | |
|
26 | ; May not work on old windows systems. | |
|
27 | asyncore_use_poll = true | |
|
28 | ||
|
29 | 12 | |
|
30 | 13 | ; ########################### |
|
31 | 14 | ; GUNICORN APPLICATION SERVER |
|
32 | 15 | ; ########################### |
|
33 | 16 | |
|
34 |
; run with gunicorn --paste |
|
|
17 | ; run with gunicorn --config gunicorn_conf.py --paste vcsserver.ini | |
|
35 | 18 | |
|
36 | 19 | ; Module to use, this setting shouldn't be changed |
|
37 |
|
|
|
20 | use = egg:gunicorn#main | |
|
38 | 21 | |
|
39 | 22 | [app:main] |
|
40 | 23 | ; The %(here)s variable will be replaced with the absolute path of parent directory |
|
41 | 24 | ; of this file |
|
42 | 25 | ; Each option in the app:main can be override by an environmental variable |
|
43 | 26 | ; |
|
44 | 27 | ;To override an option: |
|
45 | 28 | ; |
|
46 | 29 | ;RC_<KeyName> |
|
47 | 30 | ;Everything should be uppercase, . and - should be replaced by _. |
|
48 | 31 | ;For example, if you have these configuration settings: |
|
49 | 32 | ;rc_cache.repo_object.backend = foo |
|
50 | 33 | ;can be overridden by |
|
51 | 34 | ;export RC_CACHE_REPO_OBJECT_BACKEND=foo |
|
52 | 35 | |
|
53 | 36 | use = egg:rhodecode-vcsserver |
|
54 | 37 | |
|
55 | 38 | |
|
56 | 39 | ; ############# |
|
57 | 40 | ; DEBUG OPTIONS |
|
58 | 41 | ; ############# |
|
59 | 42 | |
|
60 | 43 | # During development the we want to have the debug toolbar enabled |
|
61 | 44 | pyramid.includes = |
|
62 | 45 | pyramid_debugtoolbar |
|
63 | 46 | |
|
64 | 47 | debugtoolbar.hosts = 0.0.0.0/0 |
|
65 | 48 | debugtoolbar.exclude_prefixes = |
|
66 | 49 | /css |
|
67 | 50 | /fonts |
|
68 | 51 | /images |
|
69 | 52 | /js |
|
70 | 53 | |
|
71 | 54 | ; ################# |
|
72 | 55 | ; END DEBUG OPTIONS |
|
73 | 56 | ; ################# |
|
74 | 57 | |
|
75 | 58 | ; Pyramid default locales, we need this to be set |
|
76 | 59 | #pyramid.default_locale_name = en |
|
77 | 60 | |
|
78 | 61 | ; default locale used by VCS systems |
|
79 | 62 | #locale = en_US.UTF-8 |
|
80 | 63 | |
|
81 | ; path to binaries for vcsserver, it should be set by the installer | |
|
64 | ; path to binaries (hg,git,svn) for vcsserver, it should be set by the installer | |
|
82 | 65 | ; at installation time, e.g /home/user/.rccontrol/vcsserver-1/profile/bin |
|
83 | ; it can also be a path to nix-build output in case of development | |
|
84 |
core.binary_dir = |
|
|
66 | ; or /usr/local/bin/rhodecode_bin/vcs_bin | |
|
67 | core.binary_dir = | |
|
68 | ||
|
69 | ; Redis connection settings for svn integrations logic | |
|
70 | ; This connection string needs to be the same on ce and vcsserver | |
|
71 | vcs.svn.redis_conn = redis://redis:6379/0 | |
|
85 | 72 | |
|
86 | 73 | ; Custom exception store path, defaults to TMPDIR |
|
87 | 74 | ; This is used to store exception from RhodeCode in shared directory |
|
88 | 75 | #exception_tracker.store_path = |
|
89 | 76 | |
|
90 | 77 | ; ############# |
|
91 | 78 | ; DOGPILE CACHE |
|
92 | 79 | ; ############# |
|
93 | 80 | |
|
94 | 81 | ; Default cache dir for caches. Putting this into a ramdisk can boost performance. |
|
95 | 82 | ; eg. /tmpfs/data_ramdisk, however this directory might require large amount of space |
|
96 | 83 | #cache_dir = %(here)s/data |
|
97 | 84 | |
|
98 | 85 | ; *************************************** |
|
99 | 86 | ; `repo_object` cache, default file based |
|
100 | 87 | ; *************************************** |
|
101 | 88 | |
|
102 | 89 | ; `repo_object` cache settings for vcs methods for repositories |
|
103 | 90 | #rc_cache.repo_object.backend = dogpile.cache.rc.file_namespace |
|
104 | 91 | |
|
105 | 92 | ; cache auto-expires after N seconds |
|
106 | 93 | ; Examples: 86400 (1Day), 604800 (7Days), 1209600 (14Days), 2592000 (30days), 7776000 (90Days) |
|
107 | 94 | #rc_cache.repo_object.expiration_time = 2592000 |
|
108 | 95 | |
|
109 | 96 | ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set |
|
110 | 97 | #rc_cache.repo_object.arguments.filename = /tmp/vcsserver_cache_repo_object.db |
|
111 | 98 | |
|
112 | 99 | ; *********************************************************** |
|
113 | 100 | ; `repo_object` cache with redis backend |
|
114 | 101 | ; recommended for larger instance, and for better performance |
|
115 | 102 | ; *********************************************************** |
|
116 | 103 | |
|
117 | 104 | ; `repo_object` cache settings for vcs methods for repositories |
|
118 | 105 | #rc_cache.repo_object.backend = dogpile.cache.rc.redis_msgpack |
|
119 | 106 | |
|
120 | 107 | ; cache auto-expires after N seconds |
|
121 | 108 | ; Examples: 86400 (1Day), 604800 (7Days), 1209600 (14Days), 2592000 (30days), 7776000 (90Days) |
|
122 | 109 | #rc_cache.repo_object.expiration_time = 2592000 |
|
123 | 110 | |
|
124 | 111 | ; redis_expiration_time needs to be greater then expiration_time |
|
125 | 112 | #rc_cache.repo_object.arguments.redis_expiration_time = 3592000 |
|
126 | 113 | |
|
127 | 114 | #rc_cache.repo_object.arguments.host = localhost |
|
128 | 115 | #rc_cache.repo_object.arguments.port = 6379 |
|
129 | 116 | #rc_cache.repo_object.arguments.db = 5 |
|
130 | 117 | #rc_cache.repo_object.arguments.socket_timeout = 30 |
|
131 | 118 | ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends |
|
132 | 119 | #rc_cache.repo_object.arguments.distributed_lock = true |
|
133 | 120 | |
|
134 | 121 | ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen |
|
135 | 122 | #rc_cache.repo_object.arguments.lock_auto_renewal = true |
|
136 | 123 | |
|
137 | 124 | ; Statsd client config, this is used to send metrics to statsd |
|
138 | 125 | ; We recommend setting statsd_exported and scrape them using Promethues |
|
139 | 126 | #statsd.enabled = false |
|
140 | 127 | #statsd.statsd_host = 0.0.0.0 |
|
141 | 128 | #statsd.statsd_port = 8125 |
|
142 | 129 | #statsd.statsd_prefix = |
|
143 | 130 | #statsd.statsd_ipv6 = false |
|
144 | 131 | |
|
145 | 132 | ; configure logging automatically at server startup set to false |
|
146 | 133 | ; to use the below custom logging config. |
|
147 | 134 | ; RC_LOGGING_FORMATTER |
|
148 | 135 | ; RC_LOGGING_LEVEL |
|
149 | 136 | ; env variables can control the settings for logging in case of autoconfigure |
|
150 | 137 | |
|
151 | 138 | #logging.autoconfigure = true |
|
152 | 139 | |
|
153 | 140 | ; specify your own custom logging config file to configure logging |
|
154 | 141 | #logging.logging_conf_file = /path/to/custom_logging.ini |
|
155 | 142 | |
|
156 | 143 | ; ##################### |
|
157 | 144 | ; LOGGING CONFIGURATION |
|
158 | 145 | ; ##################### |
|
159 | 146 | |
|
160 | 147 | [loggers] |
|
161 | 148 | keys = root, vcsserver |
|
162 | 149 | |
|
163 | 150 | [handlers] |
|
164 | 151 | keys = console |
|
165 | 152 | |
|
166 | 153 | [formatters] |
|
167 | 154 | keys = generic, json |
|
168 | 155 | |
|
169 | 156 | ; ####### |
|
170 | 157 | ; LOGGERS |
|
171 | 158 | ; ####### |
|
172 | 159 | [logger_root] |
|
173 | 160 | level = NOTSET |
|
174 | 161 | handlers = console |
|
175 | 162 | |
|
176 | 163 | [logger_vcsserver] |
|
177 | 164 | level = DEBUG |
|
178 | 165 | handlers = |
|
179 | 166 | qualname = vcsserver |
|
180 | 167 | propagate = 1 |
|
181 | 168 | |
|
182 | 169 | ; ######## |
|
183 | 170 | ; HANDLERS |
|
184 | 171 | ; ######## |
|
185 | 172 | |
|
186 | 173 | [handler_console] |
|
187 | 174 | class = StreamHandler |
|
188 | 175 | args = (sys.stderr, ) |
|
189 | 176 | level = DEBUG |
|
190 | 177 | ; To enable JSON formatted logs replace 'generic' with 'json' |
|
191 | 178 | ; This allows sending properly formatted logs to grafana loki or elasticsearch |
|
192 | 179 | formatter = generic |
|
193 | 180 | |
|
194 | 181 | ; ########## |
|
195 | 182 | ; FORMATTERS |
|
196 | 183 | ; ########## |
|
197 | 184 | |
|
198 | 185 | [formatter_generic] |
|
199 | 186 | format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s |
|
200 | 187 | datefmt = %Y-%m-%d %H:%M:%S |
|
201 | 188 | |
|
202 | 189 | [formatter_json] |
|
203 | 190 | format = %(timestamp)s %(levelname)s %(name)s %(message)s %(req_id)s |
|
204 | 191 | class = vcsserver.lib._vendor.jsonlogger.JsonFormatter |
@@ -1,167 +1,171 b'' | |||
|
1 | # | |
|
2 | 1 | |
|
3 | 2 | ; ################################# |
|
4 | 3 | ; RHODECODE VCSSERVER CONFIGURATION |
|
5 | 4 | ; ################################# |
|
6 | 5 | |
|
7 | 6 | [server:main] |
|
8 | ; COMMON HOST/IP CONFIG | |
|
9 | host = 127.0.0.1 | |
|
7 | ; COMMON HOST/IP CONFIG, This applies mostly to develop setup, | |
|
8 | ; Host port for gunicorn are controlled by gunicorn_conf.py | |
|
9 | host = 0.0.0.0 | |
|
10 | 10 | port = 10010 |
|
11 | 11 | |
|
12 | 12 | |
|
13 | 13 | ; ########################### |
|
14 | 14 | ; GUNICORN APPLICATION SERVER |
|
15 | 15 | ; ########################### |
|
16 | 16 | |
|
17 |
; run with gunicorn --paste |
|
|
17 | ; run with gunicorn --config gunicorn_conf.py --paste vcsserver.ini | |
|
18 | 18 | |
|
19 | 19 | ; Module to use, this setting shouldn't be changed |
|
20 | 20 | use = egg:gunicorn#main |
|
21 | 21 | |
|
22 | 22 | [app:main] |
|
23 | 23 | ; The %(here)s variable will be replaced with the absolute path of parent directory |
|
24 | 24 | ; of this file |
|
25 | 25 | ; Each option in the app:main can be override by an environmental variable |
|
26 | 26 | ; |
|
27 | 27 | ;To override an option: |
|
28 | 28 | ; |
|
29 | 29 | ;RC_<KeyName> |
|
30 | 30 | ;Everything should be uppercase, . and - should be replaced by _. |
|
31 | 31 | ;For example, if you have these configuration settings: |
|
32 | 32 | ;rc_cache.repo_object.backend = foo |
|
33 | 33 | ;can be overridden by |
|
34 | 34 | ;export RC_CACHE_REPO_OBJECT_BACKEND=foo |
|
35 | 35 | |
|
36 | 36 | use = egg:rhodecode-vcsserver |
|
37 | 37 | |
|
38 | 38 | ; Pyramid default locales, we need this to be set |
|
39 | 39 | #pyramid.default_locale_name = en |
|
40 | 40 | |
|
41 | 41 | ; default locale used by VCS systems |
|
42 | 42 | #locale = en_US.UTF-8 |
|
43 | 43 | |
|
44 | ; path to binaries for vcsserver, it should be set by the installer | |
|
44 | ; path to binaries (hg,git,svn) for vcsserver, it should be set by the installer | |
|
45 | 45 | ; at installation time, e.g /home/user/.rccontrol/vcsserver-1/profile/bin |
|
46 | ; it can also be a path to nix-build output in case of development | |
|
47 |
core.binary_dir = |
|
|
46 | ; or /usr/local/bin/rhodecode_bin/vcs_bin | |
|
47 | core.binary_dir = | |
|
48 | ||
|
49 | ; Redis connection settings for svn integrations logic | |
|
50 | ; This connection string needs to be the same on ce and vcsserver | |
|
51 | vcs.svn.redis_conn = redis://redis:6379/0 | |
|
48 | 52 | |
|
49 | 53 | ; Custom exception store path, defaults to TMPDIR |
|
50 | 54 | ; This is used to store exception from RhodeCode in shared directory |
|
51 | 55 | #exception_tracker.store_path = |
|
52 | 56 | |
|
53 | 57 | ; ############# |
|
54 | 58 | ; DOGPILE CACHE |
|
55 | 59 | ; ############# |
|
56 | 60 | |
|
57 | 61 | ; Default cache dir for caches. Putting this into a ramdisk can boost performance. |
|
58 | 62 | ; eg. /tmpfs/data_ramdisk, however this directory might require large amount of space |
|
59 | 63 | #cache_dir = %(here)s/data |
|
60 | 64 | |
|
61 | 65 | ; *************************************** |
|
62 | 66 | ; `repo_object` cache, default file based |
|
63 | 67 | ; *************************************** |
|
64 | 68 | |
|
65 | 69 | ; `repo_object` cache settings for vcs methods for repositories |
|
66 | 70 | #rc_cache.repo_object.backend = dogpile.cache.rc.file_namespace |
|
67 | 71 | |
|
68 | 72 | ; cache auto-expires after N seconds |
|
69 | 73 | ; Examples: 86400 (1Day), 604800 (7Days), 1209600 (14Days), 2592000 (30days), 7776000 (90Days) |
|
70 | 74 | #rc_cache.repo_object.expiration_time = 2592000 |
|
71 | 75 | |
|
72 | 76 | ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set |
|
73 | 77 | #rc_cache.repo_object.arguments.filename = /tmp/vcsserver_cache_repo_object.db |
|
74 | 78 | |
|
75 | 79 | ; *********************************************************** |
|
76 | 80 | ; `repo_object` cache with redis backend |
|
77 | 81 | ; recommended for larger instance, and for better performance |
|
78 | 82 | ; *********************************************************** |
|
79 | 83 | |
|
80 | 84 | ; `repo_object` cache settings for vcs methods for repositories |
|
81 | 85 | #rc_cache.repo_object.backend = dogpile.cache.rc.redis_msgpack |
|
82 | 86 | |
|
83 | 87 | ; cache auto-expires after N seconds |
|
84 | 88 | ; Examples: 86400 (1Day), 604800 (7Days), 1209600 (14Days), 2592000 (30days), 7776000 (90Days) |
|
85 | 89 | #rc_cache.repo_object.expiration_time = 2592000 |
|
86 | 90 | |
|
87 | 91 | ; redis_expiration_time needs to be greater then expiration_time |
|
88 | 92 | #rc_cache.repo_object.arguments.redis_expiration_time = 3592000 |
|
89 | 93 | |
|
90 | 94 | #rc_cache.repo_object.arguments.host = localhost |
|
91 | 95 | #rc_cache.repo_object.arguments.port = 6379 |
|
92 | 96 | #rc_cache.repo_object.arguments.db = 5 |
|
93 | 97 | #rc_cache.repo_object.arguments.socket_timeout = 30 |
|
94 | 98 | ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends |
|
95 | 99 | #rc_cache.repo_object.arguments.distributed_lock = true |
|
96 | 100 | |
|
97 | 101 | ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen |
|
98 | 102 | #rc_cache.repo_object.arguments.lock_auto_renewal = true |
|
99 | 103 | |
|
100 | 104 | ; Statsd client config, this is used to send metrics to statsd |
|
101 | 105 | ; We recommend setting statsd_exported and scrape them using Promethues |
|
102 | 106 | #statsd.enabled = false |
|
103 | 107 | #statsd.statsd_host = 0.0.0.0 |
|
104 | 108 | #statsd.statsd_port = 8125 |
|
105 | 109 | #statsd.statsd_prefix = |
|
106 | 110 | #statsd.statsd_ipv6 = false |
|
107 | 111 | |
|
108 | 112 | ; configure logging automatically at server startup set to false |
|
109 | 113 | ; to use the below custom logging config. |
|
110 | 114 | ; RC_LOGGING_FORMATTER |
|
111 | 115 | ; RC_LOGGING_LEVEL |
|
112 | 116 | ; env variables can control the settings for logging in case of autoconfigure |
|
113 | 117 | |
|
114 | 118 | #logging.autoconfigure = true |
|
115 | 119 | |
|
116 | 120 | ; specify your own custom logging config file to configure logging |
|
117 | 121 | #logging.logging_conf_file = /path/to/custom_logging.ini |
|
118 | 122 | |
|
119 | 123 | ; ##################### |
|
120 | 124 | ; LOGGING CONFIGURATION |
|
121 | 125 | ; ##################### |
|
122 | 126 | |
|
123 | 127 | [loggers] |
|
124 | 128 | keys = root, vcsserver |
|
125 | 129 | |
|
126 | 130 | [handlers] |
|
127 | 131 | keys = console |
|
128 | 132 | |
|
129 | 133 | [formatters] |
|
130 | 134 | keys = generic, json |
|
131 | 135 | |
|
132 | 136 | ; ####### |
|
133 | 137 | ; LOGGERS |
|
134 | 138 | ; ####### |
|
135 | 139 | [logger_root] |
|
136 | 140 | level = NOTSET |
|
137 | 141 | handlers = console |
|
138 | 142 | |
|
139 | 143 | [logger_vcsserver] |
|
140 | 144 | level = INFO |
|
141 | 145 | handlers = |
|
142 | 146 | qualname = vcsserver |
|
143 | 147 | propagate = 1 |
|
144 | 148 | |
|
145 | 149 | ; ######## |
|
146 | 150 | ; HANDLERS |
|
147 | 151 | ; ######## |
|
148 | 152 | |
|
149 | 153 | [handler_console] |
|
150 | 154 | class = StreamHandler |
|
151 | 155 | args = (sys.stderr, ) |
|
152 | 156 | level = INFO |
|
153 | 157 | ; To enable JSON formatted logs replace 'generic' with 'json' |
|
154 | 158 | ; This allows sending properly formatted logs to grafana loki or elasticsearch |
|
155 | 159 | formatter = generic |
|
156 | 160 | |
|
157 | 161 | ; ########## |
|
158 | 162 | ; FORMATTERS |
|
159 | 163 | ; ########## |
|
160 | 164 | |
|
161 | 165 | [formatter_generic] |
|
162 | 166 | format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s |
|
163 | 167 | datefmt = %Y-%m-%d %H:%M:%S |
|
164 | 168 | |
|
165 | 169 | [formatter_json] |
|
166 | 170 | format = %(timestamp)s %(levelname)s %(name)s %(message)s %(req_id)s |
|
167 | 171 | class = vcsserver.lib._vendor.jsonlogger.JsonFormatter |
@@ -1,77 +1,102 b'' | |||
|
1 | 1 | # deps, generated via pipdeptree --exclude setuptools,wheel,pipdeptree,pip -f | tr '[:upper:]' '[:lower:]' |
|
2 | 2 | |
|
3 | 3 | async-timeout==4.0.3 |
|
4 | 4 | atomicwrites==1.4.1 |
|
5 | 5 | celery==5.3.6 |
|
6 | 6 | billiard==4.2.0 |
|
7 | 7 | click==8.1.3 |
|
8 | 8 | click-didyoumean==0.3.0 |
|
9 | 9 | click==8.1.3 |
|
10 | 10 | click-plugins==1.1.1 |
|
11 | 11 | click==8.1.3 |
|
12 | 12 | click-repl==0.2.0 |
|
13 | 13 | click==8.1.3 |
|
14 | 14 | prompt-toolkit==3.0.38 |
|
15 | 15 | wcwidth==0.2.6 |
|
16 | 16 | six==1.16.0 |
|
17 | 17 | kombu==5.3.5 |
|
18 | 18 | amqp==5.2.0 |
|
19 | 19 | vine==5.1.0 |
|
20 | 20 | vine==5.1.0 |
|
21 | 21 | python-dateutil==2.8.2 |
|
22 | 22 | six==1.16.0 |
|
23 |
tzdata==202 |
|
|
23 | tzdata==2024.1 | |
|
24 | 24 | vine==5.1.0 |
|
25 | 25 | contextlib2==21.6.0 |
|
26 | cov-core==1.15.0 | |
|
27 | coverage==7.2.3 | |
|
28 | diskcache==5.6.3 | |
|
29 | dogpile.cache==1.3.0 | |
|
26 | dogpile.cache==1.3.3 | |
|
30 | 27 | decorator==5.1.1 |
|
31 | 28 | stevedore==5.1.0 |
|
32 | 29 | pbr==5.11.1 |
|
33 | 30 | dulwich==0.21.6 |
|
34 | 31 | urllib3==1.26.14 |
|
32 | fsspec==2024.6.0 | |
|
35 | 33 | gunicorn==21.2.0 |
|
36 |
packaging==2 |
|
|
37 |
hg-evolve==11. |
|
|
34 | packaging==24.0 | |
|
35 | hg-evolve==11.1.3 | |
|
38 | 36 | importlib-metadata==6.0.0 |
|
39 | 37 | zipp==3.15.0 |
|
40 |
mercurial==6. |
|
|
41 | mock==5.0.2 | |
|
38 | mercurial==6.7.4 | |
|
42 | 39 | more-itertools==9.1.0 |
|
43 |
msgpack==1.0. |
|
|
44 |
orjson==3. |
|
|
40 | msgpack==1.0.8 | |
|
41 | orjson==3.10.3 | |
|
45 | 42 | psutil==5.9.8 |
|
46 | 43 | py==1.11.0 |
|
47 | 44 | pygit2==1.13.3 |
|
48 | 45 | cffi==1.16.0 |
|
49 | 46 | pycparser==2.21 |
|
50 | 47 | pygments==2.15.1 |
|
51 | 48 | pyparsing==3.1.1 |
|
52 | 49 | pyramid==2.0.2 |
|
53 | 50 | hupper==1.12 |
|
54 | 51 | plaster==1.1.2 |
|
55 | 52 | plaster-pastedeploy==1.0.1 |
|
56 | 53 | pastedeploy==3.1.0 |
|
57 | 54 | plaster==1.1.2 |
|
58 | 55 | translationstring==1.4 |
|
59 | 56 | venusian==3.0.0 |
|
60 | 57 | webob==1.8.7 |
|
61 | 58 | zope.deprecation==5.0.0 |
|
62 |
zope.interface==6. |
|
|
63 |
redis==5.0. |
|
|
59 | zope.interface==6.3.0 | |
|
60 | redis==5.0.4 | |
|
64 | 61 | async-timeout==4.0.3 |
|
65 | 62 | repoze.lru==0.7 |
|
63 | s3fs==2024.6.0 | |
|
64 | aiobotocore==2.13.0 | |
|
65 | aiohttp==3.9.5 | |
|
66 | aiosignal==1.3.1 | |
|
67 | frozenlist==1.4.1 | |
|
68 | attrs==22.2.0 | |
|
69 | frozenlist==1.4.1 | |
|
70 | multidict==6.0.5 | |
|
71 | yarl==1.9.4 | |
|
72 | idna==3.4 | |
|
73 | multidict==6.0.5 | |
|
74 | aioitertools==0.11.0 | |
|
75 | botocore==1.34.106 | |
|
76 | jmespath==1.0.1 | |
|
77 | python-dateutil==2.8.2 | |
|
78 | six==1.16.0 | |
|
79 | urllib3==1.26.14 | |
|
80 | wrapt==1.16.0 | |
|
81 | aiohttp==3.9.5 | |
|
82 | aiosignal==1.3.1 | |
|
83 | frozenlist==1.4.1 | |
|
84 | attrs==22.2.0 | |
|
85 | frozenlist==1.4.1 | |
|
86 | multidict==6.0.5 | |
|
87 | yarl==1.9.4 | |
|
88 | idna==3.4 | |
|
89 | multidict==6.0.5 | |
|
90 | fsspec==2024.6.0 | |
|
66 | 91 | scandir==1.10.0 |
|
67 | 92 | setproctitle==1.3.3 |
|
68 | 93 | subvertpy==0.11.0 |
|
69 | 94 | waitress==3.0.0 |
|
70 | 95 | wcwidth==0.2.6 |
|
71 | 96 | |
|
72 | 97 | |
|
73 | 98 | ## test related requirements |
|
74 | 99 | #-r requirements_test.txt |
|
75 | 100 | |
|
76 | 101 | ## uncomment to add the debug libraries |
|
77 | 102 | #-r requirements_debug.txt |
@@ -1,45 +1,48 b'' | |||
|
1 | 1 | # test related requirements |
|
2 | ||
|
3 | cov-core==1.15.0 | |
|
4 |
coverage==7. |
|
|
5 | mock==5.0.2 | |
|
6 | py==1.11.0 | |
|
7 | pytest-cov==4.0.0 | |
|
8 | coverage==7.2.3 | |
|
9 | pytest==7.3.1 | |
|
10 | attrs==22.2.0 | |
|
2 | mock==5.1.0 | |
|
3 | pytest-cov==4.1.0 | |
|
4 | coverage==7.4.3 | |
|
5 | pytest==8.1.1 | |
|
11 | 6 | iniconfig==2.0.0 |
|
12 |
packaging==2 |
|
|
13 |
pluggy==1. |
|
|
7 | packaging==24.0 | |
|
8 | pluggy==1.4.0 | |
|
9 | pytest-env==1.1.3 | |
|
10 | pytest==8.1.1 | |
|
11 | iniconfig==2.0.0 | |
|
12 | packaging==24.0 | |
|
13 | pluggy==1.4.0 | |
|
14 | 14 | pytest-profiling==1.7.0 |
|
15 | 15 | gprof2dot==2022.7.29 |
|
16 |
pytest== |
|
|
17 | attrs==22.2.0 | |
|
16 | pytest==8.1.1 | |
|
18 | 17 | iniconfig==2.0.0 |
|
19 |
packaging==2 |
|
|
20 |
pluggy==1. |
|
|
18 | packaging==24.0 | |
|
19 | pluggy==1.4.0 | |
|
21 | 20 | six==1.16.0 |
|
22 |
pytest-r |
|
|
23 | pytest-sugar==0.9.7 | |
|
24 | packaging==23.1 | |
|
25 | pytest==7.3.1 | |
|
26 | attrs==22.2.0 | |
|
21 | pytest-rerunfailures==13.0 | |
|
22 | packaging==24.0 | |
|
23 | pytest==8.1.1 | |
|
27 | 24 | iniconfig==2.0.0 |
|
28 |
packaging==2 |
|
|
29 |
pluggy==1. |
|
|
30 | termcolor==2.3.0 | |
|
31 |
pytest- |
|
|
32 | pytest==7.3.1 | |
|
33 | attrs==22.2.0 | |
|
25 | packaging==24.0 | |
|
26 | pluggy==1.4.0 | |
|
27 | pytest-runner==6.0.1 | |
|
28 | pytest-sugar==1.0.0 | |
|
29 | packaging==24.0 | |
|
30 | pytest==8.1.1 | |
|
34 | 31 | iniconfig==2.0.0 |
|
35 |
packaging==2 |
|
|
36 |
pluggy==1. |
|
|
32 | packaging==24.0 | |
|
33 | pluggy==1.4.0 | |
|
34 | termcolor==2.4.0 | |
|
35 | pytest-timeout==2.3.1 | |
|
36 | pytest==8.1.1 | |
|
37 | iniconfig==2.0.0 | |
|
38 | packaging==24.0 | |
|
39 | pluggy==1.4.0 | |
|
37 | 40 | webtest==3.0.0 |
|
38 |
beautifulsoup4==4.1 |
|
|
39 |
soupsieve==2. |
|
|
41 | beautifulsoup4==4.12.3 | |
|
42 | soupsieve==2.5 | |
|
40 | 43 | waitress==3.0.0 |
|
41 | 44 | webob==1.8.7 |
|
42 | 45 | |
|
43 | 46 | # RhodeCode test-data |
|
44 | 47 | rc_testdata @ https://code.rhodecode.com/upstream/rc-testdata-dist/raw/77378e9097f700b4c1b9391b56199fe63566b5c9/rc_testdata-0.11.0.tar.gz#egg=rc_testdata |
|
45 | 48 | rc_testdata==0.11.0 |
@@ -1,193 +1,187 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | import os |
|
18 | 18 | import sys |
|
19 | 19 | import tempfile |
|
20 | 20 | import logging |
|
21 | 21 | import urllib.parse |
|
22 | 22 | |
|
23 |
from vcsserver.lib. |
|
|
23 | from vcsserver.lib.archive_cache import get_archival_cache_store | |
|
24 | 24 | |
|
25 | 25 | from vcsserver import exceptions |
|
26 | 26 | from vcsserver.exceptions import NoContentException |
|
27 | 27 | from vcsserver.hgcompat import archival |
|
28 | from vcsserver.str_utils import safe_bytes | |
|
28 | from vcsserver.lib.str_utils import safe_bytes | |
|
29 | 29 | from vcsserver.lib.exc_tracking import format_exc |
|
30 | 30 | log = logging.getLogger(__name__) |
|
31 | 31 | |
|
32 | 32 | |
|
33 | 33 | class RepoFactory: |
|
34 | 34 | """ |
|
35 | 35 | Utility to create instances of repository |
|
36 | 36 | |
|
37 | 37 | It provides internal caching of the `repo` object based on |
|
38 | 38 | the :term:`call context`. |
|
39 | 39 | """ |
|
40 | 40 | repo_type = None |
|
41 | 41 | |
|
42 | 42 | def __init__(self): |
|
43 | 43 | pass |
|
44 | 44 | |
|
45 | 45 | def _create_config(self, path, config): |
|
46 | 46 | config = {} |
|
47 | 47 | return config |
|
48 | 48 | |
|
49 | 49 | def _create_repo(self, wire, create): |
|
50 | 50 | raise NotImplementedError() |
|
51 | 51 | |
|
52 | 52 | def repo(self, wire, create=False): |
|
53 | 53 | raise NotImplementedError() |
|
54 | 54 | |
|
55 | 55 | |
|
56 | 56 | def obfuscate_qs(query_string): |
|
57 | 57 | if query_string is None: |
|
58 | 58 | return None |
|
59 | 59 | |
|
60 | 60 | parsed = [] |
|
61 | 61 | for k, v in urllib.parse.parse_qsl(query_string, keep_blank_values=True): |
|
62 | 62 | if k in ['auth_token', 'api_key']: |
|
63 | 63 | v = "*****" |
|
64 | 64 | parsed.append((k, v)) |
|
65 | 65 | |
|
66 | 66 | return '&'.join('{}{}'.format( |
|
67 | 67 | k, f'={v}' if v else '') for k, v in parsed) |
|
68 | 68 | |
|
69 | 69 | |
|
70 | 70 | def raise_from_original(new_type, org_exc: Exception): |
|
71 | 71 | """ |
|
72 | 72 | Raise a new exception type with original args and traceback. |
|
73 | 73 | """ |
|
74 | 74 | exc_info = sys.exc_info() |
|
75 | 75 | exc_type, exc_value, exc_traceback = exc_info |
|
76 | 76 | new_exc = new_type(*exc_value.args) |
|
77 | 77 | |
|
78 | 78 | # store the original traceback into the new exc |
|
79 | 79 | new_exc._org_exc_tb = format_exc(exc_info) |
|
80 | 80 | |
|
81 | 81 | try: |
|
82 | 82 | raise new_exc.with_traceback(exc_traceback) |
|
83 | 83 | finally: |
|
84 | 84 | del exc_traceback |
|
85 | 85 | |
|
86 | 86 | |
|
87 | 87 | class ArchiveNode: |
|
88 | 88 | def __init__(self, path, mode, is_link, raw_bytes): |
|
89 | 89 | self.path = path |
|
90 | 90 | self.mode = mode |
|
91 | 91 | self.is_link = is_link |
|
92 | 92 | self.raw_bytes = raw_bytes |
|
93 | 93 | |
|
94 | 94 | |
|
95 | 95 | def store_archive_in_cache(node_walker, archive_key, kind, mtime, archive_at_path, archive_dir_name, |
|
96 | 96 | commit_id, write_metadata=True, extra_metadata=None, cache_config=None): |
|
97 | 97 | """ |
|
98 | 98 | Function that would store generate archive and send it to a dedicated backend store |
|
99 | 99 | In here we use diskcache |
|
100 | 100 | |
|
101 | 101 | :param node_walker: a generator returning nodes to add to archive |
|
102 | 102 | :param archive_key: key used to store the path |
|
103 | 103 | :param kind: archive kind |
|
104 | 104 | :param mtime: time of creation |
|
105 | 105 | :param archive_at_path: default '/' the path at archive was started. |
|
106 | 106 | If this is not '/' it means it's a partial archive |
|
107 | 107 | :param archive_dir_name: inside dir name when creating an archive |
|
108 | 108 | :param commit_id: commit sha of revision archive was created at |
|
109 | 109 | :param write_metadata: |
|
110 | 110 | :param extra_metadata: |
|
111 | 111 | :param cache_config: |
|
112 | 112 | |
|
113 | 113 | walker should be a file walker, for example, |
|
114 | 114 | def node_walker(): |
|
115 | 115 | for file_info in files: |
|
116 | 116 | yield ArchiveNode(fn, mode, is_link, ctx[fn].data) |
|
117 | 117 | """ |
|
118 | 118 | extra_metadata = extra_metadata or {} |
|
119 | 119 | |
|
120 | 120 | d_cache = get_archival_cache_store(config=cache_config) |
|
121 | 121 | |
|
122 | 122 | if archive_key in d_cache: |
|
123 | with d_cache as d_cache_reader: | |
|
124 | reader, tag = d_cache_reader.get(archive_key, read=True, tag=True, retry=True) | |
|
125 | return reader.name | |
|
123 | reader, metadata = d_cache.fetch(archive_key) | |
|
124 | return reader.name | |
|
126 | 125 | |
|
127 | 126 | archive_tmp_path = safe_bytes(tempfile.mkstemp()[1]) |
|
128 | 127 | log.debug('Creating new temp archive in %s', archive_tmp_path) |
|
129 | 128 | |
|
130 | 129 | if kind == "tgz": |
|
131 | 130 | archiver = archival.tarit(archive_tmp_path, mtime, b"gz") |
|
132 | 131 | elif kind == "tbz2": |
|
133 | 132 | archiver = archival.tarit(archive_tmp_path, mtime, b"bz2") |
|
134 | 133 | elif kind == 'zip': |
|
135 | 134 | archiver = archival.zipit(archive_tmp_path, mtime) |
|
136 | 135 | else: |
|
137 | 136 | raise exceptions.ArchiveException()( |
|
138 | 137 | f'Remote does not support: "{kind}" archive type.') |
|
139 | 138 | |
|
140 | 139 | for f in node_walker(commit_id, archive_at_path): |
|
141 | 140 | f_path = os.path.join(safe_bytes(archive_dir_name), safe_bytes(f.path).lstrip(b'/')) |
|
141 | ||
|
142 | 142 | try: |
|
143 | 143 | archiver.addfile(f_path, f.mode, f.is_link, f.raw_bytes()) |
|
144 | 144 | except NoContentException: |
|
145 | 145 | # NOTE(marcink): this is a special case for SVN so we can create "empty" |
|
146 | 146 | # directories which are not supported by archiver |
|
147 | 147 | archiver.addfile(os.path.join(f_path, b'.dir'), f.mode, f.is_link, b'') |
|
148 | 148 | |
|
149 | metadata = dict([ | |
|
150 | ('commit_id', commit_id), | |
|
151 | ('mtime', mtime), | |
|
152 | ]) | |
|
153 | metadata.update(extra_metadata) | |
|
149 | 154 | if write_metadata: |
|
150 | metadata = dict([ | |
|
151 | ('commit_id', commit_id), | |
|
152 | ('mtime', mtime), | |
|
153 | ]) | |
|
154 | metadata.update(extra_metadata) | |
|
155 | ||
|
156 | 155 | meta = [safe_bytes(f"{f_name}:{value}") for f_name, value in metadata.items()] |
|
157 | 156 | f_path = os.path.join(safe_bytes(archive_dir_name), b'.archival.txt') |
|
158 | 157 | archiver.addfile(f_path, 0o644, False, b'\n'.join(meta)) |
|
159 | 158 | |
|
160 | 159 | archiver.done() |
|
161 | 160 | |
|
162 | # ensure set & get are atomic | |
|
163 | with d_cache.transact(): | |
|
164 | ||
|
165 | with open(archive_tmp_path, 'rb') as archive_file: | |
|
166 | add_result = d_cache.set(archive_key, archive_file, read=True, tag='db-name', retry=True) | |
|
167 | if not add_result: | |
|
168 | log.error('Failed to store cache for key=%s', archive_key) | |
|
161 | with open(archive_tmp_path, 'rb') as archive_file: | |
|
162 | add_result = d_cache.store(archive_key, archive_file, metadata=metadata) | |
|
163 | if not add_result: | |
|
164 | log.error('Failed to store cache for key=%s', archive_key) | |
|
169 | 165 | |
|
170 |
|
|
|
166 | os.remove(archive_tmp_path) | |
|
171 | 167 | |
|
172 |
|
|
|
173 | if not reader: | |
|
174 | raise AssertionError(f'empty reader on key={archive_key} added={add_result}') | |
|
168 | reader, metadata = d_cache.fetch(archive_key) | |
|
175 | 169 | |
|
176 |
|
|
|
170 | return reader.name | |
|
177 | 171 | |
|
178 | 172 | |
|
179 | 173 | class BinaryEnvelope: |
|
180 | 174 | def __init__(self, val): |
|
181 | 175 | self.val = val |
|
182 | 176 | |
|
183 | 177 | |
|
184 | 178 | class BytesEnvelope(bytes): |
|
185 | 179 | def __new__(cls, content): |
|
186 | 180 | if isinstance(content, bytes): |
|
187 | 181 | return super().__new__(cls, content) |
|
188 | 182 | else: |
|
189 | 183 | raise TypeError('BytesEnvelope content= param must be bytes. Use BinaryEnvelope to wrap other types') |
|
190 | 184 | |
|
191 | 185 | |
|
192 | 186 | class BinaryBytesEnvelope(BytesEnvelope): |
|
193 | 187 | pass |
@@ -1,185 +1,185 b'' | |||
|
1 | 1 | # Copyright (C) 2010-2023 RhodeCode GmbH |
|
2 | 2 | # |
|
3 | 3 | # This program is free software: you can redistribute it and/or modify |
|
4 | 4 | # it under the terms of the GNU Affero General Public License, version 3 |
|
5 | 5 | # (only), as published by the Free Software Foundation. |
|
6 | 6 | # |
|
7 | 7 | # This program is distributed in the hope that it will be useful, |
|
8 | 8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
9 | 9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
10 | 10 | # GNU General Public License for more details. |
|
11 | 11 | # |
|
12 | 12 | # You should have received a copy of the GNU Affero General Public License |
|
13 | 13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. |
|
14 | 14 | # |
|
15 | 15 | # This program is dual-licensed. If you wish to learn more about the |
|
16 | 16 | # RhodeCode Enterprise Edition, including its added features, Support services, |
|
17 | 17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ |
|
18 | 18 | |
|
19 | 19 | import os |
|
20 | 20 | import textwrap |
|
21 | 21 | import string |
|
22 | 22 | import functools |
|
23 | 23 | import logging |
|
24 | 24 | import tempfile |
|
25 | 25 | import logging.config |
|
26 | 26 | |
|
27 | from vcsserver.type_utils import str2bool, aslist | |
|
27 | from vcsserver.lib.type_utils import str2bool, aslist | |
|
28 | 28 | |
|
29 | 29 | log = logging.getLogger(__name__) |
|
30 | 30 | |
|
31 | 31 | |
|
32 | 32 | # skip keys, that are set here, so we don't double process those |
|
33 | 33 | set_keys = { |
|
34 | 34 | '__file__': '' |
|
35 | 35 | } |
|
36 | 36 | |
|
37 | 37 | |
|
38 | 38 | class SettingsMaker: |
|
39 | 39 | |
|
40 | 40 | def __init__(self, app_settings): |
|
41 | 41 | self.settings = app_settings |
|
42 | 42 | |
|
43 | 43 | @classmethod |
|
44 | 44 | def _bool_func(cls, input_val): |
|
45 | 45 | if isinstance(input_val, bytes): |
|
46 | 46 | # decode to str |
|
47 | 47 | input_val = input_val.decode('utf8') |
|
48 | 48 | return str2bool(input_val) |
|
49 | 49 | |
|
50 | 50 | @classmethod |
|
51 | 51 | def _int_func(cls, input_val): |
|
52 | 52 | return int(input_val) |
|
53 | 53 | |
|
54 | 54 | @classmethod |
|
55 | 55 | def _float_func(cls, input_val): |
|
56 | 56 | return float(input_val) |
|
57 | 57 | |
|
58 | 58 | @classmethod |
|
59 | 59 | def _list_func(cls, input_val, sep=','): |
|
60 | 60 | return aslist(input_val, sep=sep) |
|
61 | 61 | |
|
62 | 62 | @classmethod |
|
63 | 63 | def _string_func(cls, input_val, lower=True): |
|
64 | 64 | if lower: |
|
65 | 65 | input_val = input_val.lower() |
|
66 | 66 | return input_val |
|
67 | 67 | |
|
68 | 68 | @classmethod |
|
69 | 69 | def _string_no_quote_func(cls, input_val, lower=True): |
|
70 | 70 | """ |
|
71 | 71 | Special case string function that detects if value is set to empty quote string |
|
72 | 72 | e.g. |
|
73 | 73 | |
|
74 | 74 | core.binary_dir = "" |
|
75 | 75 | """ |
|
76 | 76 | |
|
77 | 77 | input_val = cls._string_func(input_val, lower=lower) |
|
78 | 78 | if input_val in ['""', "''"]: |
|
79 | 79 | return '' |
|
80 | 80 | return input_val |
|
81 | 81 | |
|
82 | 82 | @classmethod |
|
83 | 83 | def _dir_func(cls, input_val, ensure_dir=False, mode=0o755): |
|
84 | 84 | |
|
85 | 85 | # ensure we have our dir created |
|
86 | 86 | if not os.path.isdir(input_val) and ensure_dir: |
|
87 | 87 | os.makedirs(input_val, mode=mode, exist_ok=True) |
|
88 | 88 | |
|
89 | 89 | if not os.path.isdir(input_val): |
|
90 | 90 | raise Exception(f'Dir at {input_val} does not exist') |
|
91 | 91 | return input_val |
|
92 | 92 | |
|
93 | 93 | @classmethod |
|
94 | 94 | def _file_path_func(cls, input_val, ensure_dir=False, mode=0o755): |
|
95 | 95 | dirname = os.path.dirname(input_val) |
|
96 | 96 | cls._dir_func(dirname, ensure_dir=ensure_dir) |
|
97 | 97 | return input_val |
|
98 | 98 | |
|
99 | 99 | @classmethod |
|
100 | 100 | def _key_transformator(cls, key): |
|
101 | 101 | return "{}_{}".format('RC'.upper(), key.upper().replace('.', '_').replace('-', '_')) |
|
102 | 102 | |
|
103 | 103 | def maybe_env_key(self, key): |
|
104 | 104 | # now maybe we have this KEY in env, search and use the value with higher priority. |
|
105 | 105 | transformed_key = self._key_transformator(key) |
|
106 | 106 | envvar_value = os.environ.get(transformed_key) |
|
107 | 107 | if envvar_value: |
|
108 | 108 | log.debug('using `%s` key instead of `%s` key for config', transformed_key, key) |
|
109 | 109 | |
|
110 | 110 | return envvar_value |
|
111 | 111 | |
|
112 | 112 | def env_expand(self): |
|
113 | 113 | replaced = {} |
|
114 | 114 | for k, v in self.settings.items(): |
|
115 | 115 | if k not in set_keys: |
|
116 | 116 | envvar_value = self.maybe_env_key(k) |
|
117 | 117 | if envvar_value: |
|
118 | 118 | replaced[k] = envvar_value |
|
119 | 119 | set_keys[k] = envvar_value |
|
120 | 120 | |
|
121 | 121 | # replace ALL keys updated |
|
122 | 122 | self.settings.update(replaced) |
|
123 | 123 | |
|
124 | 124 | def enable_logging(self, logging_conf=None, level='INFO', formatter='generic'): |
|
125 | 125 | """ |
|
126 | 126 | Helper to enable debug on running instance |
|
127 | 127 | :return: |
|
128 | 128 | """ |
|
129 | 129 | |
|
130 | 130 | if not str2bool(self.settings.get('logging.autoconfigure')): |
|
131 | 131 | log.info('logging configuration based on main .ini file') |
|
132 | 132 | return |
|
133 | 133 | |
|
134 | 134 | if logging_conf is None: |
|
135 | 135 | logging_conf = self.settings.get('logging.logging_conf_file') or '' |
|
136 | 136 | |
|
137 | 137 | if not os.path.isfile(logging_conf): |
|
138 | 138 | log.error('Unable to setup logging based on %s, ' |
|
139 | 139 | 'file does not exist.... specify path using logging.logging_conf_file= config setting. ', logging_conf) |
|
140 | 140 | return |
|
141 | 141 | |
|
142 | 142 | with open(logging_conf, 'rt') as f: |
|
143 | 143 | ini_template = textwrap.dedent(f.read()) |
|
144 | 144 | ini_template = string.Template(ini_template).safe_substitute( |
|
145 | 145 | RC_LOGGING_LEVEL=os.environ.get('RC_LOGGING_LEVEL', '') or level, |
|
146 | 146 | RC_LOGGING_FORMATTER=os.environ.get('RC_LOGGING_FORMATTER', '') or formatter |
|
147 | 147 | ) |
|
148 | 148 | |
|
149 | 149 | with tempfile.NamedTemporaryFile(prefix='rc_logging_', suffix='.ini', delete=False) as f: |
|
150 | 150 | log.info('Saved Temporary LOGGING config at %s', f.name) |
|
151 | 151 | f.write(ini_template) |
|
152 | 152 | |
|
153 | 153 | logging.config.fileConfig(f.name) |
|
154 | 154 | os.remove(f.name) |
|
155 | 155 | |
|
156 | 156 | def make_setting(self, key, default, lower=False, default_when_empty=False, parser=None): |
|
157 | 157 | input_val = self.settings.get(key, default) |
|
158 | 158 | |
|
159 | 159 | if default_when_empty and not input_val: |
|
160 | 160 | # use default value when value is set in the config but it is empty |
|
161 | 161 | input_val = default |
|
162 | 162 | |
|
163 | 163 | parser_func = { |
|
164 | 164 | 'bool': self._bool_func, |
|
165 | 165 | 'int': self._int_func, |
|
166 | 166 | 'float': self._float_func, |
|
167 | 167 | 'list': self._list_func, |
|
168 | 168 | 'list:newline': functools.partial(self._list_func, sep='/n'), |
|
169 | 169 | 'list:spacesep': functools.partial(self._list_func, sep=' '), |
|
170 | 170 | 'string': functools.partial(self._string_func, lower=lower), |
|
171 | 171 | 'string:noquote': functools.partial(self._string_no_quote_func, lower=lower), |
|
172 | 172 | 'dir': self._dir_func, |
|
173 | 173 | 'dir:ensured': functools.partial(self._dir_func, ensure_dir=True), |
|
174 | 174 | 'file': self._file_path_func, |
|
175 | 175 | 'file:ensured': functools.partial(self._file_path_func, ensure_dir=True), |
|
176 | 176 | None: lambda i: i |
|
177 | 177 | }[parser] |
|
178 | 178 | |
|
179 | 179 | envvar_value = self.maybe_env_key(key) |
|
180 | 180 | if envvar_value: |
|
181 | 181 | input_val = envvar_value |
|
182 | 182 | set_keys[key] = input_val |
|
183 | 183 | |
|
184 | 184 | self.settings[key] = parser_func(input_val) |
|
185 | 185 | return self.settings[key] |
@@ -1,296 +1,296 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import re |
|
19 | 19 | import logging |
|
20 | 20 | |
|
21 | 21 | from pyramid.config import Configurator |
|
22 | 22 | from pyramid.response import Response, FileIter |
|
23 | 23 | from pyramid.httpexceptions import ( |
|
24 | 24 | HTTPBadRequest, HTTPNotImplemented, HTTPNotFound, HTTPForbidden, |
|
25 | 25 | HTTPUnprocessableEntity) |
|
26 | 26 | |
|
27 |
from vcsserver.lib. |
|
|
27 | from vcsserver.lib.ext_json import json | |
|
28 | 28 | from vcsserver.git_lfs.lib import OidHandler, LFSOidStore |
|
29 | 29 | from vcsserver.git_lfs.utils import safe_result, get_cython_compat_decorator |
|
30 | from vcsserver.str_utils import safe_int | |
|
30 | from vcsserver.lib.str_utils import safe_int | |
|
31 | 31 | |
|
32 | 32 | log = logging.getLogger(__name__) |
|
33 | 33 | |
|
34 | 34 | |
|
35 | 35 | GIT_LFS_CONTENT_TYPE = 'application/vnd.git-lfs' # +json ? |
|
36 | 36 | GIT_LFS_PROTO_PAT = re.compile(r'^/(.+)/(info/lfs/(.+))') |
|
37 | 37 | |
|
38 | 38 | |
|
39 | 39 | def write_response_error(http_exception, text=None): |
|
40 | 40 | content_type = GIT_LFS_CONTENT_TYPE + '+json' |
|
41 | 41 | _exception = http_exception(content_type=content_type) |
|
42 | 42 | _exception.content_type = content_type |
|
43 | 43 | if text: |
|
44 | 44 | _exception.body = json.dumps({'message': text}) |
|
45 | 45 | log.debug('LFS: writing response of type %s to client with text:%s', |
|
46 | 46 | http_exception, text) |
|
47 | 47 | return _exception |
|
48 | 48 | |
|
49 | 49 | |
|
50 | 50 | class AuthHeaderRequired: |
|
51 | 51 | """ |
|
52 | 52 | Decorator to check if request has proper auth-header |
|
53 | 53 | """ |
|
54 | 54 | |
|
55 | 55 | def __call__(self, func): |
|
56 | 56 | return get_cython_compat_decorator(self.__wrapper, func) |
|
57 | 57 | |
|
58 | 58 | def __wrapper(self, func, *fargs, **fkwargs): |
|
59 | 59 | request = fargs[1] |
|
60 | 60 | auth = request.authorization |
|
61 | 61 | if not auth: |
|
62 | 62 | return write_response_error(HTTPForbidden) |
|
63 | 63 | return func(*fargs[1:], **fkwargs) |
|
64 | 64 | |
|
65 | 65 | |
|
66 | 66 | # views |
|
67 | 67 | |
|
68 | 68 | def lfs_objects(request): |
|
69 | 69 | # indicate not supported, V1 API |
|
70 | 70 | log.warning('LFS: v1 api not supported, reporting it back to client') |
|
71 | 71 | return write_response_error(HTTPNotImplemented, 'LFS: v1 api not supported') |
|
72 | 72 | |
|
73 | 73 | |
|
74 | 74 | @AuthHeaderRequired() |
|
75 | 75 | def lfs_objects_batch(request): |
|
76 | 76 | """ |
|
77 | 77 | The client sends the following information to the Batch endpoint to transfer some objects: |
|
78 | 78 | |
|
79 | 79 | operation - Should be download or upload. |
|
80 | 80 | transfers - An optional Array of String identifiers for transfer |
|
81 | 81 | adapters that the client has configured. If omitted, the basic |
|
82 | 82 | transfer adapter MUST be assumed by the server. |
|
83 | 83 | objects - An Array of objects to download. |
|
84 | 84 | oid - String OID of the LFS object. |
|
85 | 85 | size - Integer byte size of the LFS object. Must be at least zero. |
|
86 | 86 | """ |
|
87 | 87 | request.response.content_type = GIT_LFS_CONTENT_TYPE + '+json' |
|
88 | 88 | auth = request.authorization |
|
89 | 89 | repo = request.matchdict.get('repo') |
|
90 | 90 | data = request.json |
|
91 | 91 | operation = data.get('operation') |
|
92 | 92 | http_scheme = request.registry.git_lfs_http_scheme |
|
93 | 93 | |
|
94 | 94 | if operation not in ('download', 'upload'): |
|
95 | 95 | log.debug('LFS: unsupported operation:%s', operation) |
|
96 | 96 | return write_response_error( |
|
97 | 97 | HTTPBadRequest, f'unsupported operation mode: `{operation}`') |
|
98 | 98 | |
|
99 | 99 | if 'objects' not in data: |
|
100 | 100 | log.debug('LFS: missing objects data') |
|
101 | 101 | return write_response_error( |
|
102 | 102 | HTTPBadRequest, 'missing objects data') |
|
103 | 103 | |
|
104 | 104 | log.debug('LFS: handling operation of type: %s', operation) |
|
105 | 105 | |
|
106 | 106 | objects = [] |
|
107 | 107 | for o in data['objects']: |
|
108 | 108 | try: |
|
109 | 109 | oid = o['oid'] |
|
110 | 110 | obj_size = o['size'] |
|
111 | 111 | except KeyError: |
|
112 | 112 | log.exception('LFS, failed to extract data') |
|
113 | 113 | return write_response_error( |
|
114 | 114 | HTTPBadRequest, 'unsupported data in objects') |
|
115 | 115 | |
|
116 | 116 | obj_data = {'oid': oid} |
|
117 | 117 | if http_scheme == 'http': |
|
118 | 118 | # Note(marcink): when using http, we might have a custom port |
|
119 | 119 | # so we skip setting it to http, url dispatch then wont generate a port in URL |
|
120 | 120 | # for development we need this |
|
121 | 121 | http_scheme = None |
|
122 | 122 | |
|
123 | 123 | obj_href = request.route_url('lfs_objects_oid', repo=repo, oid=oid, |
|
124 | 124 | _scheme=http_scheme) |
|
125 | 125 | obj_verify_href = request.route_url('lfs_objects_verify', repo=repo, |
|
126 | 126 | _scheme=http_scheme) |
|
127 | 127 | store = LFSOidStore( |
|
128 | 128 | oid, repo, store_location=request.registry.git_lfs_store_path) |
|
129 | 129 | handler = OidHandler( |
|
130 | 130 | store, repo, auth, oid, obj_size, obj_data, |
|
131 | 131 | obj_href, obj_verify_href) |
|
132 | 132 | |
|
133 | 133 | # this verifies also OIDs |
|
134 | 134 | actions, errors = handler.exec_operation(operation) |
|
135 | 135 | if errors: |
|
136 | 136 | log.warning('LFS: got following errors: %s', errors) |
|
137 | 137 | obj_data['errors'] = errors |
|
138 | 138 | |
|
139 | 139 | if actions: |
|
140 | 140 | obj_data['actions'] = actions |
|
141 | 141 | |
|
142 | 142 | obj_data['size'] = obj_size |
|
143 | 143 | obj_data['authenticated'] = True |
|
144 | 144 | objects.append(obj_data) |
|
145 | 145 | |
|
146 | 146 | result = {'objects': objects, 'transfer': 'basic'} |
|
147 | 147 | log.debug('LFS Response %s', safe_result(result)) |
|
148 | 148 | |
|
149 | 149 | return result |
|
150 | 150 | |
|
151 | 151 | |
|
152 | 152 | def lfs_objects_oid_upload(request): |
|
153 | 153 | request.response.content_type = GIT_LFS_CONTENT_TYPE + '+json' |
|
154 | 154 | repo = request.matchdict.get('repo') |
|
155 | 155 | oid = request.matchdict.get('oid') |
|
156 | 156 | store = LFSOidStore( |
|
157 | 157 | oid, repo, store_location=request.registry.git_lfs_store_path) |
|
158 | 158 | engine = store.get_engine(mode='wb') |
|
159 | 159 | log.debug('LFS: starting chunked write of LFS oid: %s to storage', oid) |
|
160 | 160 | |
|
161 | 161 | body = request.environ['wsgi.input'] |
|
162 | 162 | |
|
163 | 163 | with engine as f: |
|
164 | 164 | blksize = 64 * 1024 # 64kb |
|
165 | 165 | while True: |
|
166 | 166 | # read in chunks as stream comes in from Gunicorn |
|
167 | 167 | # this is a specific Gunicorn support function. |
|
168 | 168 | # might work differently on waitress |
|
169 | 169 | chunk = body.read(blksize) |
|
170 | 170 | if not chunk: |
|
171 | 171 | break |
|
172 | 172 | f.write(chunk) |
|
173 | 173 | |
|
174 | 174 | return {'upload': 'ok'} |
|
175 | 175 | |
|
176 | 176 | |
|
177 | 177 | def lfs_objects_oid_download(request): |
|
178 | 178 | repo = request.matchdict.get('repo') |
|
179 | 179 | oid = request.matchdict.get('oid') |
|
180 | 180 | |
|
181 | 181 | store = LFSOidStore( |
|
182 | 182 | oid, repo, store_location=request.registry.git_lfs_store_path) |
|
183 | 183 | if not store.has_oid(): |
|
184 | 184 | log.debug('LFS: oid %s does not exists in store', oid) |
|
185 | 185 | return write_response_error( |
|
186 | 186 | HTTPNotFound, f'requested file with oid `{oid}` not found in store') |
|
187 | 187 | |
|
188 | 188 | # TODO(marcink): support range header ? |
|
189 | 189 | # Range: bytes=0-, `bytes=(\d+)\-.*` |
|
190 | 190 | |
|
191 | 191 | f = open(store.oid_path, 'rb') |
|
192 | 192 | response = Response( |
|
193 | 193 | content_type='application/octet-stream', app_iter=FileIter(f)) |
|
194 | 194 | response.headers.add('X-RC-LFS-Response-Oid', str(oid)) |
|
195 | 195 | return response |
|
196 | 196 | |
|
197 | 197 | |
|
198 | 198 | def lfs_objects_verify(request): |
|
199 | 199 | request.response.content_type = GIT_LFS_CONTENT_TYPE + '+json' |
|
200 | 200 | repo = request.matchdict.get('repo') |
|
201 | 201 | |
|
202 | 202 | data = request.json |
|
203 | 203 | oid = data.get('oid') |
|
204 | 204 | size = safe_int(data.get('size')) |
|
205 | 205 | |
|
206 | 206 | if not (oid and size): |
|
207 | 207 | return write_response_error( |
|
208 | 208 | HTTPBadRequest, 'missing oid and size in request data') |
|
209 | 209 | |
|
210 | 210 | store = LFSOidStore( |
|
211 | 211 | oid, repo, store_location=request.registry.git_lfs_store_path) |
|
212 | 212 | if not store.has_oid(): |
|
213 | 213 | log.debug('LFS: oid %s does not exists in store', oid) |
|
214 | 214 | return write_response_error( |
|
215 | 215 | HTTPNotFound, f'oid `{oid}` does not exists in store') |
|
216 | 216 | |
|
217 | 217 | store_size = store.size_oid() |
|
218 | 218 | if store_size != size: |
|
219 | 219 | msg = 'requested file size mismatch store size:{} requested:{}'.format( |
|
220 | 220 | store_size, size) |
|
221 | 221 | return write_response_error( |
|
222 | 222 | HTTPUnprocessableEntity, msg) |
|
223 | 223 | |
|
224 | 224 | return {'message': {'size': 'ok', 'in_store': 'ok'}} |
|
225 | 225 | |
|
226 | 226 | |
|
227 | 227 | def lfs_objects_lock(request): |
|
228 | 228 | return write_response_error( |
|
229 | 229 | HTTPNotImplemented, 'GIT LFS locking api not supported') |
|
230 | 230 | |
|
231 | 231 | |
|
232 | 232 | def not_found(request): |
|
233 | 233 | return write_response_error( |
|
234 | 234 | HTTPNotFound, 'request path not found') |
|
235 | 235 | |
|
236 | 236 | |
|
237 | 237 | def lfs_disabled(request): |
|
238 | 238 | return write_response_error( |
|
239 | 239 | HTTPNotImplemented, 'GIT LFS disabled for this repo') |
|
240 | 240 | |
|
241 | 241 | |
|
242 | 242 | def git_lfs_app(config): |
|
243 | 243 | |
|
244 | 244 | # v1 API deprecation endpoint |
|
245 | 245 | config.add_route('lfs_objects', |
|
246 | 246 | '/{repo:.*?[^/]}/info/lfs/objects') |
|
247 | 247 | config.add_view(lfs_objects, route_name='lfs_objects', |
|
248 | 248 | request_method='POST', renderer='json') |
|
249 | 249 | |
|
250 | 250 | # locking API |
|
251 | 251 | config.add_route('lfs_objects_lock', |
|
252 | 252 | '/{repo:.*?[^/]}/info/lfs/locks') |
|
253 | 253 | config.add_view(lfs_objects_lock, route_name='lfs_objects_lock', |
|
254 | 254 | request_method=('POST', 'GET'), renderer='json') |
|
255 | 255 | |
|
256 | 256 | config.add_route('lfs_objects_lock_verify', |
|
257 | 257 | '/{repo:.*?[^/]}/info/lfs/locks/verify') |
|
258 | 258 | config.add_view(lfs_objects_lock, route_name='lfs_objects_lock_verify', |
|
259 | 259 | request_method=('POST', 'GET'), renderer='json') |
|
260 | 260 | |
|
261 | 261 | # batch API |
|
262 | 262 | config.add_route('lfs_objects_batch', |
|
263 | 263 | '/{repo:.*?[^/]}/info/lfs/objects/batch') |
|
264 | 264 | config.add_view(lfs_objects_batch, route_name='lfs_objects_batch', |
|
265 | 265 | request_method='POST', renderer='json') |
|
266 | 266 | |
|
267 | 267 | # oid upload/download API |
|
268 | 268 | config.add_route('lfs_objects_oid', |
|
269 | 269 | '/{repo:.*?[^/]}/info/lfs/objects/{oid}') |
|
270 | 270 | config.add_view(lfs_objects_oid_upload, route_name='lfs_objects_oid', |
|
271 | 271 | request_method='PUT', renderer='json') |
|
272 | 272 | config.add_view(lfs_objects_oid_download, route_name='lfs_objects_oid', |
|
273 | 273 | request_method='GET', renderer='json') |
|
274 | 274 | |
|
275 | 275 | # verification API |
|
276 | 276 | config.add_route('lfs_objects_verify', |
|
277 | 277 | '/{repo:.*?[^/]}/info/lfs/verify') |
|
278 | 278 | config.add_view(lfs_objects_verify, route_name='lfs_objects_verify', |
|
279 | 279 | request_method='POST', renderer='json') |
|
280 | 280 | |
|
281 | 281 | # not found handler for API |
|
282 | 282 | config.add_notfound_view(not_found, renderer='json') |
|
283 | 283 | |
|
284 | 284 | |
|
285 | 285 | def create_app(git_lfs_enabled, git_lfs_store_path, git_lfs_http_scheme): |
|
286 | 286 | config = Configurator() |
|
287 | 287 | if git_lfs_enabled: |
|
288 | 288 | config.include(git_lfs_app) |
|
289 | 289 | config.registry.git_lfs_store_path = git_lfs_store_path |
|
290 | 290 | config.registry.git_lfs_http_scheme = git_lfs_http_scheme |
|
291 | 291 | else: |
|
292 | 292 | # not found handler for API, reporting disabled LFS support |
|
293 | 293 | config.add_notfound_view(lfs_disabled, renderer='json') |
|
294 | 294 | |
|
295 | 295 | app = config.make_wsgi_app() |
|
296 | 296 | return app |
@@ -1,274 +1,274 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import os |
|
19 | 19 | import pytest |
|
20 | 20 | from webtest.app import TestApp as WebObTestApp |
|
21 | 21 | |
|
22 |
from vcsserver.lib. |
|
|
23 | from vcsserver.str_utils import safe_bytes | |
|
22 | from vcsserver.lib.ext_json import json | |
|
23 | from vcsserver.lib.str_utils import safe_bytes | |
|
24 | 24 | from vcsserver.git_lfs.app import create_app |
|
25 | 25 | from vcsserver.git_lfs.lib import LFSOidStore |
|
26 | 26 | |
|
27 | 27 | |
|
28 | 28 | @pytest.fixture(scope='function') |
|
29 | 29 | def git_lfs_app(tmpdir): |
|
30 | 30 | custom_app = WebObTestApp(create_app( |
|
31 | 31 | git_lfs_enabled=True, git_lfs_store_path=str(tmpdir), |
|
32 | 32 | git_lfs_http_scheme='http')) |
|
33 | 33 | custom_app._store = str(tmpdir) |
|
34 | 34 | return custom_app |
|
35 | 35 | |
|
36 | 36 | |
|
37 | 37 | @pytest.fixture(scope='function') |
|
38 | 38 | def git_lfs_https_app(tmpdir): |
|
39 | 39 | custom_app = WebObTestApp(create_app( |
|
40 | 40 | git_lfs_enabled=True, git_lfs_store_path=str(tmpdir), |
|
41 | 41 | git_lfs_http_scheme='https')) |
|
42 | 42 | custom_app._store = str(tmpdir) |
|
43 | 43 | return custom_app |
|
44 | 44 | |
|
45 | 45 | |
|
46 | 46 | @pytest.fixture() |
|
47 | 47 | def http_auth(): |
|
48 | 48 | return {'HTTP_AUTHORIZATION': "Basic XXXXX"} |
|
49 | 49 | |
|
50 | 50 | |
|
51 | 51 | class TestLFSApplication: |
|
52 | 52 | |
|
53 | 53 | def test_app_wrong_path(self, git_lfs_app): |
|
54 | 54 | git_lfs_app.get('/repo/info/lfs/xxx', status=404) |
|
55 | 55 | |
|
56 | 56 | def test_app_deprecated_endpoint(self, git_lfs_app): |
|
57 | 57 | response = git_lfs_app.post('/repo/info/lfs/objects', status=501) |
|
58 | 58 | assert response.status_code == 501 |
|
59 | 59 | assert json.loads(response.text) == {'message': 'LFS: v1 api not supported'} |
|
60 | 60 | |
|
61 | 61 | def test_app_lock_verify_api_not_available(self, git_lfs_app): |
|
62 | 62 | response = git_lfs_app.post('/repo/info/lfs/locks/verify', status=501) |
|
63 | 63 | assert response.status_code == 501 |
|
64 | 64 | assert json.loads(response.text) == { |
|
65 | 65 | 'message': 'GIT LFS locking api not supported'} |
|
66 | 66 | |
|
67 | 67 | def test_app_lock_api_not_available(self, git_lfs_app): |
|
68 | 68 | response = git_lfs_app.post('/repo/info/lfs/locks', status=501) |
|
69 | 69 | assert response.status_code == 501 |
|
70 | 70 | assert json.loads(response.text) == { |
|
71 | 71 | 'message': 'GIT LFS locking api not supported'} |
|
72 | 72 | |
|
73 | 73 | def test_app_batch_api_missing_auth(self, git_lfs_app): |
|
74 | 74 | git_lfs_app.post_json( |
|
75 | 75 | '/repo/info/lfs/objects/batch', params={}, status=403) |
|
76 | 76 | |
|
77 | 77 | def test_app_batch_api_unsupported_operation(self, git_lfs_app, http_auth): |
|
78 | 78 | response = git_lfs_app.post_json( |
|
79 | 79 | '/repo/info/lfs/objects/batch', params={}, status=400, |
|
80 | 80 | extra_environ=http_auth) |
|
81 | 81 | assert json.loads(response.text) == { |
|
82 | 82 | 'message': 'unsupported operation mode: `None`'} |
|
83 | 83 | |
|
84 | 84 | def test_app_batch_api_missing_objects(self, git_lfs_app, http_auth): |
|
85 | 85 | response = git_lfs_app.post_json( |
|
86 | 86 | '/repo/info/lfs/objects/batch', params={'operation': 'download'}, |
|
87 | 87 | status=400, extra_environ=http_auth) |
|
88 | 88 | assert json.loads(response.text) == { |
|
89 | 89 | 'message': 'missing objects data'} |
|
90 | 90 | |
|
91 | 91 | def test_app_batch_api_unsupported_data_in_objects( |
|
92 | 92 | self, git_lfs_app, http_auth): |
|
93 | 93 | params = {'operation': 'download', |
|
94 | 94 | 'objects': [{}]} |
|
95 | 95 | response = git_lfs_app.post_json( |
|
96 | 96 | '/repo/info/lfs/objects/batch', params=params, status=400, |
|
97 | 97 | extra_environ=http_auth) |
|
98 | 98 | assert json.loads(response.text) == { |
|
99 | 99 | 'message': 'unsupported data in objects'} |
|
100 | 100 | |
|
101 | 101 | def test_app_batch_api_download_missing_object( |
|
102 | 102 | self, git_lfs_app, http_auth): |
|
103 | 103 | params = {'operation': 'download', |
|
104 | 104 | 'objects': [{'oid': '123', 'size': '1024'}]} |
|
105 | 105 | response = git_lfs_app.post_json( |
|
106 | 106 | '/repo/info/lfs/objects/batch', params=params, |
|
107 | 107 | extra_environ=http_auth) |
|
108 | 108 | |
|
109 | 109 | expected_objects = [ |
|
110 | 110 | {'authenticated': True, |
|
111 | 111 | 'errors': {'error': { |
|
112 | 112 | 'code': 404, |
|
113 | 113 | 'message': 'object: 123 does not exist in store'}}, |
|
114 | 114 | 'oid': '123', |
|
115 | 115 | 'size': '1024'} |
|
116 | 116 | ] |
|
117 | 117 | assert json.loads(response.text) == { |
|
118 | 118 | 'objects': expected_objects, 'transfer': 'basic'} |
|
119 | 119 | |
|
120 | 120 | def test_app_batch_api_download(self, git_lfs_app, http_auth): |
|
121 | 121 | oid = '456' |
|
122 | 122 | oid_path = LFSOidStore(oid=oid, repo=None, store_location=git_lfs_app._store).oid_path |
|
123 | 123 | if not os.path.isdir(os.path.dirname(oid_path)): |
|
124 | 124 | os.makedirs(os.path.dirname(oid_path)) |
|
125 | 125 | with open(oid_path, 'wb') as f: |
|
126 | 126 | f.write(safe_bytes('OID_CONTENT')) |
|
127 | 127 | |
|
128 | 128 | params = {'operation': 'download', |
|
129 | 129 | 'objects': [{'oid': oid, 'size': '1024'}]} |
|
130 | 130 | response = git_lfs_app.post_json( |
|
131 | 131 | '/repo/info/lfs/objects/batch', params=params, |
|
132 | 132 | extra_environ=http_auth) |
|
133 | 133 | |
|
134 | 134 | expected_objects = [ |
|
135 | 135 | {'authenticated': True, |
|
136 | 136 | 'actions': { |
|
137 | 137 | 'download': { |
|
138 | 138 | 'header': {'Authorization': 'Basic XXXXX'}, |
|
139 | 139 | 'href': 'http://localhost/repo/info/lfs/objects/456'}, |
|
140 | 140 | }, |
|
141 | 141 | 'oid': '456', |
|
142 | 142 | 'size': '1024'} |
|
143 | 143 | ] |
|
144 | 144 | assert json.loads(response.text) == { |
|
145 | 145 | 'objects': expected_objects, 'transfer': 'basic'} |
|
146 | 146 | |
|
147 | 147 | def test_app_batch_api_upload(self, git_lfs_app, http_auth): |
|
148 | 148 | params = {'operation': 'upload', |
|
149 | 149 | 'objects': [{'oid': '123', 'size': '1024'}]} |
|
150 | 150 | response = git_lfs_app.post_json( |
|
151 | 151 | '/repo/info/lfs/objects/batch', params=params, |
|
152 | 152 | extra_environ=http_auth) |
|
153 | 153 | expected_objects = [ |
|
154 | 154 | {'authenticated': True, |
|
155 | 155 | 'actions': { |
|
156 | 156 | 'upload': { |
|
157 | 157 | 'header': {'Authorization': 'Basic XXXXX', |
|
158 | 158 | 'Transfer-Encoding': 'chunked'}, |
|
159 | 159 | 'href': 'http://localhost/repo/info/lfs/objects/123'}, |
|
160 | 160 | 'verify': { |
|
161 | 161 | 'header': {'Authorization': 'Basic XXXXX'}, |
|
162 | 162 | 'href': 'http://localhost/repo/info/lfs/verify'} |
|
163 | 163 | }, |
|
164 | 164 | 'oid': '123', |
|
165 | 165 | 'size': '1024'} |
|
166 | 166 | ] |
|
167 | 167 | assert json.loads(response.text) == { |
|
168 | 168 | 'objects': expected_objects, 'transfer': 'basic'} |
|
169 | 169 | |
|
170 | 170 | def test_app_batch_api_upload_for_https(self, git_lfs_https_app, http_auth): |
|
171 | 171 | params = {'operation': 'upload', |
|
172 | 172 | 'objects': [{'oid': '123', 'size': '1024'}]} |
|
173 | 173 | response = git_lfs_https_app.post_json( |
|
174 | 174 | '/repo/info/lfs/objects/batch', params=params, |
|
175 | 175 | extra_environ=http_auth) |
|
176 | 176 | expected_objects = [ |
|
177 | 177 | {'authenticated': True, |
|
178 | 178 | 'actions': { |
|
179 | 179 | 'upload': { |
|
180 | 180 | 'header': {'Authorization': 'Basic XXXXX', |
|
181 | 181 | 'Transfer-Encoding': 'chunked'}, |
|
182 | 182 | 'href': 'https://localhost/repo/info/lfs/objects/123'}, |
|
183 | 183 | 'verify': { |
|
184 | 184 | 'header': {'Authorization': 'Basic XXXXX'}, |
|
185 | 185 | 'href': 'https://localhost/repo/info/lfs/verify'} |
|
186 | 186 | }, |
|
187 | 187 | 'oid': '123', |
|
188 | 188 | 'size': '1024'} |
|
189 | 189 | ] |
|
190 | 190 | assert json.loads(response.text) == { |
|
191 | 191 | 'objects': expected_objects, 'transfer': 'basic'} |
|
192 | 192 | |
|
193 | 193 | def test_app_verify_api_missing_data(self, git_lfs_app): |
|
194 | 194 | params = {'oid': 'missing'} |
|
195 | 195 | response = git_lfs_app.post_json( |
|
196 | 196 | '/repo/info/lfs/verify', params=params, |
|
197 | 197 | status=400) |
|
198 | 198 | |
|
199 | 199 | assert json.loads(response.text) == { |
|
200 | 200 | 'message': 'missing oid and size in request data'} |
|
201 | 201 | |
|
202 | 202 | def test_app_verify_api_missing_obj(self, git_lfs_app): |
|
203 | 203 | params = {'oid': 'missing', 'size': '1024'} |
|
204 | 204 | response = git_lfs_app.post_json( |
|
205 | 205 | '/repo/info/lfs/verify', params=params, |
|
206 | 206 | status=404) |
|
207 | 207 | |
|
208 | 208 | assert json.loads(response.text) == { |
|
209 | 209 | 'message': 'oid `missing` does not exists in store'} |
|
210 | 210 | |
|
211 | 211 | def test_app_verify_api_size_mismatch(self, git_lfs_app): |
|
212 | 212 | oid = 'existing' |
|
213 | 213 | oid_path = LFSOidStore(oid=oid, repo=None, store_location=git_lfs_app._store).oid_path |
|
214 | 214 | if not os.path.isdir(os.path.dirname(oid_path)): |
|
215 | 215 | os.makedirs(os.path.dirname(oid_path)) |
|
216 | 216 | with open(oid_path, 'wb') as f: |
|
217 | 217 | f.write(safe_bytes('OID_CONTENT')) |
|
218 | 218 | |
|
219 | 219 | params = {'oid': oid, 'size': '1024'} |
|
220 | 220 | response = git_lfs_app.post_json( |
|
221 | 221 | '/repo/info/lfs/verify', params=params, status=422) |
|
222 | 222 | |
|
223 | 223 | assert json.loads(response.text) == { |
|
224 | 224 | 'message': 'requested file size mismatch ' |
|
225 | 225 | 'store size:11 requested:1024'} |
|
226 | 226 | |
|
227 | 227 | def test_app_verify_api(self, git_lfs_app): |
|
228 | 228 | oid = 'existing' |
|
229 | 229 | oid_path = LFSOidStore(oid=oid, repo=None, store_location=git_lfs_app._store).oid_path |
|
230 | 230 | if not os.path.isdir(os.path.dirname(oid_path)): |
|
231 | 231 | os.makedirs(os.path.dirname(oid_path)) |
|
232 | 232 | with open(oid_path, 'wb') as f: |
|
233 | 233 | f.write(safe_bytes('OID_CONTENT')) |
|
234 | 234 | |
|
235 | 235 | params = {'oid': oid, 'size': 11} |
|
236 | 236 | response = git_lfs_app.post_json( |
|
237 | 237 | '/repo/info/lfs/verify', params=params) |
|
238 | 238 | |
|
239 | 239 | assert json.loads(response.text) == { |
|
240 | 240 | 'message': {'size': 'ok', 'in_store': 'ok'}} |
|
241 | 241 | |
|
242 | 242 | def test_app_download_api_oid_not_existing(self, git_lfs_app): |
|
243 | 243 | oid = 'missing' |
|
244 | 244 | |
|
245 | 245 | response = git_lfs_app.get( |
|
246 | 246 | '/repo/info/lfs/objects/{oid}'.format(oid=oid), status=404) |
|
247 | 247 | |
|
248 | 248 | assert json.loads(response.text) == { |
|
249 | 249 | 'message': 'requested file with oid `missing` not found in store'} |
|
250 | 250 | |
|
251 | 251 | def test_app_download_api(self, git_lfs_app): |
|
252 | 252 | oid = 'existing' |
|
253 | 253 | oid_path = LFSOidStore(oid=oid, repo=None, store_location=git_lfs_app._store).oid_path |
|
254 | 254 | if not os.path.isdir(os.path.dirname(oid_path)): |
|
255 | 255 | os.makedirs(os.path.dirname(oid_path)) |
|
256 | 256 | with open(oid_path, 'wb') as f: |
|
257 | 257 | f.write(safe_bytes('OID_CONTENT')) |
|
258 | 258 | |
|
259 | 259 | response = git_lfs_app.get( |
|
260 | 260 | '/repo/info/lfs/objects/{oid}'.format(oid=oid)) |
|
261 | 261 | assert response |
|
262 | 262 | |
|
263 | 263 | def test_app_upload(self, git_lfs_app): |
|
264 | 264 | oid = 'uploaded' |
|
265 | 265 | |
|
266 | 266 | response = git_lfs_app.put( |
|
267 | 267 | '/repo/info/lfs/objects/{oid}'.format(oid=oid), params='CONTENT') |
|
268 | 268 | |
|
269 | 269 | assert json.loads(response.text) == {'upload': 'ok'} |
|
270 | 270 | |
|
271 | 271 | # verify that we actually wrote that OID |
|
272 | 272 | oid_path = LFSOidStore(oid=oid, repo=None, store_location=git_lfs_app._store).oid_path |
|
273 | 273 | assert os.path.isfile(oid_path) |
|
274 | 274 | assert 'CONTENT' == open(oid_path).read() |
@@ -1,142 +1,142 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import os |
|
19 | 19 | import pytest |
|
20 | from vcsserver.str_utils import safe_bytes | |
|
20 | from vcsserver.lib.str_utils import safe_bytes | |
|
21 | 21 | from vcsserver.git_lfs.lib import OidHandler, LFSOidStore |
|
22 | 22 | |
|
23 | 23 | |
|
24 | 24 | @pytest.fixture() |
|
25 | 25 | def lfs_store(tmpdir): |
|
26 | 26 | repo = 'test' |
|
27 | 27 | oid = '123456789' |
|
28 | 28 | store = LFSOidStore(oid=oid, repo=repo, store_location=str(tmpdir)) |
|
29 | 29 | return store |
|
30 | 30 | |
|
31 | 31 | |
|
32 | 32 | @pytest.fixture() |
|
33 | 33 | def oid_handler(lfs_store): |
|
34 | 34 | store = lfs_store |
|
35 | 35 | repo = store.repo |
|
36 | 36 | oid = store.oid |
|
37 | 37 | |
|
38 | 38 | oid_handler = OidHandler( |
|
39 | 39 | store=store, repo_name=repo, auth=('basic', 'xxxx'), |
|
40 | 40 | oid=oid, |
|
41 | 41 | obj_size='1024', obj_data={}, obj_href='http://localhost/handle_oid', |
|
42 | 42 | obj_verify_href='http://localhost/verify') |
|
43 | 43 | return oid_handler |
|
44 | 44 | |
|
45 | 45 | |
|
46 | 46 | class TestOidHandler: |
|
47 | 47 | |
|
48 | 48 | @pytest.mark.parametrize('exec_action', [ |
|
49 | 49 | 'download', |
|
50 | 50 | 'upload', |
|
51 | 51 | ]) |
|
52 | 52 | def test_exec_action(self, exec_action, oid_handler): |
|
53 | 53 | handler = oid_handler.exec_operation(exec_action) |
|
54 | 54 | assert handler |
|
55 | 55 | |
|
56 | 56 | def test_exec_action_undefined(self, oid_handler): |
|
57 | 57 | with pytest.raises(AttributeError): |
|
58 | 58 | oid_handler.exec_operation('wrong') |
|
59 | 59 | |
|
60 | 60 | def test_download_oid_not_existing(self, oid_handler): |
|
61 | 61 | response, has_errors = oid_handler.exec_operation('download') |
|
62 | 62 | |
|
63 | 63 | assert response is None |
|
64 | 64 | assert has_errors['error'] == { |
|
65 | 65 | 'code': 404, |
|
66 | 66 | 'message': 'object: 123456789 does not exist in store'} |
|
67 | 67 | |
|
68 | 68 | def test_download_oid(self, oid_handler): |
|
69 | 69 | store = oid_handler.get_store() |
|
70 | 70 | if not os.path.isdir(os.path.dirname(store.oid_path)): |
|
71 | 71 | os.makedirs(os.path.dirname(store.oid_path)) |
|
72 | 72 | |
|
73 | 73 | with open(store.oid_path, 'wb') as f: |
|
74 | 74 | f.write(safe_bytes('CONTENT')) |
|
75 | 75 | |
|
76 | 76 | response, has_errors = oid_handler.exec_operation('download') |
|
77 | 77 | |
|
78 | 78 | assert has_errors is None |
|
79 | 79 | assert response['download'] == { |
|
80 | 80 | 'header': {'Authorization': 'basic xxxx'}, |
|
81 | 81 | 'href': 'http://localhost/handle_oid' |
|
82 | 82 | } |
|
83 | 83 | |
|
84 | 84 | def test_upload_oid_that_exists(self, oid_handler): |
|
85 | 85 | store = oid_handler.get_store() |
|
86 | 86 | if not os.path.isdir(os.path.dirname(store.oid_path)): |
|
87 | 87 | os.makedirs(os.path.dirname(store.oid_path)) |
|
88 | 88 | |
|
89 | 89 | with open(store.oid_path, 'wb') as f: |
|
90 | 90 | f.write(safe_bytes('CONTENT')) |
|
91 | 91 | oid_handler.obj_size = 7 |
|
92 | 92 | response, has_errors = oid_handler.exec_operation('upload') |
|
93 | 93 | assert has_errors is None |
|
94 | 94 | assert response is None |
|
95 | 95 | |
|
96 | 96 | def test_upload_oid_that_exists_but_has_wrong_size(self, oid_handler): |
|
97 | 97 | store = oid_handler.get_store() |
|
98 | 98 | if not os.path.isdir(os.path.dirname(store.oid_path)): |
|
99 | 99 | os.makedirs(os.path.dirname(store.oid_path)) |
|
100 | 100 | |
|
101 | 101 | with open(store.oid_path, 'wb') as f: |
|
102 | 102 | f.write(safe_bytes('CONTENT')) |
|
103 | 103 | |
|
104 | 104 | oid_handler.obj_size = 10240 |
|
105 | 105 | response, has_errors = oid_handler.exec_operation('upload') |
|
106 | 106 | assert has_errors is None |
|
107 | 107 | assert response['upload'] == { |
|
108 | 108 | 'header': {'Authorization': 'basic xxxx', |
|
109 | 109 | 'Transfer-Encoding': 'chunked'}, |
|
110 | 110 | 'href': 'http://localhost/handle_oid', |
|
111 | 111 | } |
|
112 | 112 | |
|
113 | 113 | def test_upload_oid(self, oid_handler): |
|
114 | 114 | response, has_errors = oid_handler.exec_operation('upload') |
|
115 | 115 | assert has_errors is None |
|
116 | 116 | assert response['upload'] == { |
|
117 | 117 | 'header': {'Authorization': 'basic xxxx', |
|
118 | 118 | 'Transfer-Encoding': 'chunked'}, |
|
119 | 119 | 'href': 'http://localhost/handle_oid' |
|
120 | 120 | } |
|
121 | 121 | |
|
122 | 122 | |
|
123 | 123 | class TestLFSStore: |
|
124 | 124 | def test_write_oid(self, lfs_store): |
|
125 | 125 | oid_location = lfs_store.oid_path |
|
126 | 126 | |
|
127 | 127 | assert not os.path.isfile(oid_location) |
|
128 | 128 | |
|
129 | 129 | engine = lfs_store.get_engine(mode='wb') |
|
130 | 130 | with engine as f: |
|
131 | 131 | f.write(safe_bytes('CONTENT')) |
|
132 | 132 | |
|
133 | 133 | assert os.path.isfile(oid_location) |
|
134 | 134 | |
|
135 | 135 | def test_detect_has_oid(self, lfs_store): |
|
136 | 136 | |
|
137 | 137 | assert lfs_store.has_oid() is False |
|
138 | 138 | engine = lfs_store.get_engine(mode='wb') |
|
139 | 139 | with engine as f: |
|
140 | 140 | f.write(safe_bytes('CONTENT')) |
|
141 | 141 | |
|
142 | 142 | assert lfs_store.has_oid() is True |
@@ -1,92 +1,92 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | """ |
|
19 | 19 | Mercurial libs compatibility |
|
20 | 20 | """ |
|
21 | 21 | |
|
22 | 22 | import mercurial |
|
23 | 23 | from mercurial import demandimport |
|
24 | 24 | |
|
25 | 25 | # patch demandimport, due to bug in mercurial when it always triggers |
|
26 | 26 | # demandimport.enable() |
|
27 | from vcsserver.str_utils import safe_bytes | |
|
27 | from vcsserver.lib.str_utils import safe_bytes | |
|
28 | 28 | |
|
29 | 29 | demandimport.enable = lambda *args, **kwargs: 1 |
|
30 | 30 | |
|
31 | 31 | from mercurial import ui |
|
32 | 32 | from mercurial import patch |
|
33 | 33 | from mercurial import config |
|
34 | 34 | from mercurial import extensions |
|
35 | 35 | from mercurial import scmutil |
|
36 | 36 | from mercurial import archival |
|
37 | 37 | from mercurial import discovery |
|
38 | 38 | from mercurial import unionrepo |
|
39 | 39 | from mercurial import localrepo |
|
40 | 40 | from mercurial import merge as hg_merge |
|
41 | 41 | from mercurial import subrepo |
|
42 | 42 | from mercurial import subrepoutil |
|
43 | 43 | from mercurial import tags as hg_tag |
|
44 | 44 | from mercurial import util as hgutil |
|
45 | 45 | from mercurial.commands import clone, pull |
|
46 | 46 | from mercurial.node import nullid |
|
47 | 47 | from mercurial.context import memctx, memfilectx |
|
48 | 48 | from mercurial.error import ( |
|
49 | 49 | LookupError, RepoError, RepoLookupError, Abort, InterventionRequired, |
|
50 | 50 | RequirementError, ProgrammingError) |
|
51 | 51 | from mercurial.hgweb import hgweb_mod |
|
52 | 52 | from mercurial.localrepo import instance |
|
53 | 53 | from mercurial.match import match, alwaysmatcher, patternmatcher |
|
54 | 54 | from mercurial.mdiff import diffopts |
|
55 | 55 | from mercurial.node import bin, hex |
|
56 | 56 | from mercurial.encoding import tolocal |
|
57 | 57 | from mercurial.discovery import findcommonoutgoing |
|
58 | 58 | from mercurial.hg import peer |
|
59 | from mercurial.httppeer import makepeer | |
|
59 | from mercurial.httppeer import make_peer | |
|
60 | 60 | from mercurial.utils.urlutil import url as hg_url |
|
61 | 61 | from mercurial.scmutil import revrange, revsymbol |
|
62 | 62 | from mercurial.node import nullrev |
|
63 | 63 | from mercurial import exchange |
|
64 | 64 | from hgext import largefiles |
|
65 | 65 | |
|
66 | 66 | # those authnadlers are patched for python 2.6.5 bug an |
|
67 | 67 | # infinit looping when given invalid resources |
|
68 | 68 | from mercurial.url import httpbasicauthhandler, httpdigestauthhandler |
|
69 | 69 | |
|
70 | 70 | # hg strip is in core now |
|
71 | 71 | from mercurial import strip as hgext_strip |
|
72 | 72 | |
|
73 | 73 | |
|
74 | 74 | def get_ctx(repo, ref): |
|
75 | 75 | if not isinstance(ref, int): |
|
76 | 76 | ref = safe_bytes(ref) |
|
77 | 77 | |
|
78 | 78 | try: |
|
79 | 79 | ctx = repo[ref] |
|
80 | 80 | return ctx |
|
81 | 81 | except (ProgrammingError, TypeError): |
|
82 | 82 | # we're unable to find the rev using a regular lookup, we fallback |
|
83 | 83 | # to slower, but backward compat revsymbol usage |
|
84 | 84 | pass |
|
85 | 85 | except (LookupError, RepoLookupError): |
|
86 | 86 | # Similar case as above but only for refs that are not numeric |
|
87 | 87 | if isinstance(ref, int): |
|
88 | 88 | raise |
|
89 | 89 | |
|
90 | 90 | ctx = revsymbol(repo, ref) |
|
91 | 91 | |
|
92 | 92 | return ctx |
@@ -1,230 +1,238 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | |
|
19 | 19 | import re |
|
20 | 20 | import os |
|
21 | 21 | import sys |
|
22 | 22 | import datetime |
|
23 | 23 | import logging |
|
24 | 24 | import pkg_resources |
|
25 | 25 | |
|
26 | 26 | import vcsserver |
|
27 | 27 | import vcsserver.settings |
|
28 | from vcsserver.str_utils import safe_bytes | |
|
28 | from vcsserver.lib.str_utils import safe_bytes | |
|
29 | 29 | |
|
30 | 30 | log = logging.getLogger(__name__) |
|
31 | 31 | |
|
32 | 32 | HOOKS_DIR_MODE = 0o755 |
|
33 | 33 | HOOKS_FILE_MODE = 0o755 |
|
34 | 34 | |
|
35 | 35 | |
|
36 | 36 | def set_permissions_if_needed(path_to_check, perms: oct): |
|
37 | 37 | # Get current permissions |
|
38 | 38 | current_permissions = os.stat(path_to_check).st_mode & 0o777 # Extract permission bits |
|
39 | 39 | |
|
40 | 40 | # Check if current permissions are lower than required |
|
41 | 41 | if current_permissions < int(perms): |
|
42 | 42 | # Change the permissions if they are lower than required |
|
43 | 43 | os.chmod(path_to_check, perms) |
|
44 | 44 | |
|
45 | 45 | |
|
46 | 46 | def get_git_hooks_path(repo_path, bare): |
|
47 | 47 | hooks_path = os.path.join(repo_path, 'hooks') |
|
48 | 48 | if not bare: |
|
49 | 49 | hooks_path = os.path.join(repo_path, '.git', 'hooks') |
|
50 | 50 | |
|
51 | 51 | return hooks_path |
|
52 | 52 | |
|
53 | 53 | |
|
54 | 54 | def install_git_hooks(repo_path, bare, executable=None, force_create=False): |
|
55 | 55 | """ |
|
56 | 56 | Creates a RhodeCode hook inside a git repository |
|
57 | 57 | |
|
58 | 58 | :param repo_path: path to repository |
|
59 | 59 | :param bare: defines if repository is considered a bare git repo |
|
60 | 60 | :param executable: binary executable to put in the hooks |
|
61 | 61 | :param force_create: Creates even if the same name hook exists |
|
62 | 62 | """ |
|
63 | 63 | executable = executable or sys.executable |
|
64 | 64 | hooks_path = get_git_hooks_path(repo_path, bare) |
|
65 | 65 | |
|
66 | 66 | # we always call it to ensure dir exists and it has a proper mode |
|
67 | 67 | if not os.path.exists(hooks_path): |
|
68 | 68 | # If it doesn't exist, create a new directory with the specified mode |
|
69 | 69 | os.makedirs(hooks_path, mode=HOOKS_DIR_MODE, exist_ok=True) |
|
70 | 70 | # If it exists, change the directory's mode to the specified mode |
|
71 | 71 | set_permissions_if_needed(hooks_path, perms=HOOKS_DIR_MODE) |
|
72 | 72 | |
|
73 | 73 | tmpl_post = pkg_resources.resource_string( |
|
74 | 74 | 'vcsserver', '/'.join( |
|
75 | 75 | ('hook_utils', 'hook_templates', 'git_post_receive.py.tmpl'))) |
|
76 | 76 | tmpl_pre = pkg_resources.resource_string( |
|
77 | 77 | 'vcsserver', '/'.join( |
|
78 | 78 | ('hook_utils', 'hook_templates', 'git_pre_receive.py.tmpl'))) |
|
79 | 79 | |
|
80 | 80 | path = '' # not used for now |
|
81 | 81 | timestamp = datetime.datetime.utcnow().isoformat() |
|
82 | 82 | |
|
83 | 83 | for h_type, template in [('pre', tmpl_pre), ('post', tmpl_post)]: |
|
84 | 84 | log.debug('Installing git hook in repo %s', repo_path) |
|
85 | 85 | _hook_file = os.path.join(hooks_path, f'{h_type}-receive') |
|
86 | 86 | _rhodecode_hook = check_rhodecode_hook(_hook_file) |
|
87 | 87 | |
|
88 | 88 | if _rhodecode_hook or force_create: |
|
89 | 89 | log.debug('writing git %s hook file at %s !', h_type, _hook_file) |
|
90 | env_expand = str([ | |
|
91 | ('RC_INI_FILE', vcsserver.CONFIG['__file__']), | |
|
92 | ('RC_CORE_BINARY_DIR', vcsserver.settings.BINARY_DIR), | |
|
93 | ('RC_GIT_EXECUTABLE', vcsserver.settings.GIT_EXECUTABLE()), | |
|
94 | ('RC_SVN_EXECUTABLE', vcsserver.settings.SVN_EXECUTABLE()), | |
|
95 | ('RC_SVNLOOK_EXECUTABLE', vcsserver.settings.SVNLOOK_EXECUTABLE()), | |
|
96 | ]) | |
|
90 | 97 | try: |
|
91 | 98 | with open(_hook_file, 'wb') as f: |
|
99 | template = template.replace(b'_OS_EXPAND_', safe_bytes(env_expand)) | |
|
92 | 100 | template = template.replace(b'_TMPL_', safe_bytes(vcsserver.get_version())) |
|
93 | 101 | template = template.replace(b'_DATE_', safe_bytes(timestamp)) |
|
94 | 102 | template = template.replace(b'_ENV_', safe_bytes(executable)) |
|
95 | 103 | template = template.replace(b'_PATH_', safe_bytes(path)) |
|
96 | 104 | f.write(template) |
|
97 | 105 | set_permissions_if_needed(_hook_file, perms=HOOKS_FILE_MODE) |
|
98 | 106 | except OSError: |
|
99 | 107 | log.exception('error writing hook file %s', _hook_file) |
|
100 | 108 | else: |
|
101 | 109 | log.debug('skipping writing hook file') |
|
102 | 110 | |
|
103 | 111 | return True |
|
104 | 112 | |
|
105 | 113 | |
|
106 | 114 | def get_svn_hooks_path(repo_path): |
|
107 | 115 | hooks_path = os.path.join(repo_path, 'hooks') |
|
108 | 116 | |
|
109 | 117 | return hooks_path |
|
110 | 118 | |
|
111 | 119 | |
|
112 | 120 | def install_svn_hooks(repo_path, executable=None, force_create=False): |
|
113 | 121 | """ |
|
114 | 122 | Creates RhodeCode hooks inside a svn repository |
|
115 | 123 | |
|
116 | 124 | :param repo_path: path to repository |
|
117 | 125 | :param executable: binary executable to put in the hooks |
|
118 | 126 | :param force_create: Create even if same name hook exists |
|
119 | 127 | """ |
|
120 | 128 | executable = executable or sys.executable |
|
121 | 129 | hooks_path = get_svn_hooks_path(repo_path) |
|
122 | 130 | if not os.path.isdir(hooks_path): |
|
123 | 131 | os.makedirs(hooks_path, mode=0o777, exist_ok=True) |
|
124 | 132 | |
|
125 | 133 | tmpl_post = pkg_resources.resource_string( |
|
126 | 134 | 'vcsserver', '/'.join( |
|
127 | 135 | ('hook_utils', 'hook_templates', 'svn_post_commit_hook.py.tmpl'))) |
|
128 | 136 | tmpl_pre = pkg_resources.resource_string( |
|
129 | 137 | 'vcsserver', '/'.join( |
|
130 | 138 | ('hook_utils', 'hook_templates', 'svn_pre_commit_hook.py.tmpl'))) |
|
131 | 139 | |
|
132 | 140 | path = '' # not used for now |
|
133 | 141 | timestamp = datetime.datetime.utcnow().isoformat() |
|
134 | 142 | |
|
135 | 143 | for h_type, template in [('pre', tmpl_pre), ('post', tmpl_post)]: |
|
136 | 144 | log.debug('Installing svn hook in repo %s', repo_path) |
|
137 | 145 | _hook_file = os.path.join(hooks_path, f'{h_type}-commit') |
|
138 | 146 | _rhodecode_hook = check_rhodecode_hook(_hook_file) |
|
139 | 147 | |
|
140 | 148 | if _rhodecode_hook or force_create: |
|
141 | 149 | log.debug('writing svn %s hook file at %s !', h_type, _hook_file) |
|
142 | 150 | |
|
143 | 151 | env_expand = str([ |
|
152 | ('RC_INI_FILE', vcsserver.CONFIG['__file__']), | |
|
144 | 153 | ('RC_CORE_BINARY_DIR', vcsserver.settings.BINARY_DIR), |
|
145 | 154 | ('RC_GIT_EXECUTABLE', vcsserver.settings.GIT_EXECUTABLE()), |
|
146 | 155 | ('RC_SVN_EXECUTABLE', vcsserver.settings.SVN_EXECUTABLE()), |
|
147 | 156 | ('RC_SVNLOOK_EXECUTABLE', vcsserver.settings.SVNLOOK_EXECUTABLE()), |
|
148 | ||
|
149 | 157 | ]) |
|
150 | 158 | try: |
|
151 | 159 | with open(_hook_file, 'wb') as f: |
|
160 | template = template.replace(b'_OS_EXPAND_', safe_bytes(env_expand)) | |
|
152 | 161 | template = template.replace(b'_TMPL_', safe_bytes(vcsserver.get_version())) |
|
153 | 162 | template = template.replace(b'_DATE_', safe_bytes(timestamp)) |
|
154 | template = template.replace(b'_OS_EXPAND_', safe_bytes(env_expand)) | |
|
155 | 163 | template = template.replace(b'_ENV_', safe_bytes(executable)) |
|
156 | 164 | template = template.replace(b'_PATH_', safe_bytes(path)) |
|
157 | 165 | |
|
158 | 166 | f.write(template) |
|
159 | 167 | os.chmod(_hook_file, 0o755) |
|
160 | 168 | except OSError: |
|
161 | 169 | log.exception('error writing hook file %s', _hook_file) |
|
162 | 170 | else: |
|
163 | 171 | log.debug('skipping writing hook file') |
|
164 | 172 | |
|
165 | 173 | return True |
|
166 | 174 | |
|
167 | 175 | |
|
168 | 176 | def get_version_from_hook(hook_path): |
|
169 | 177 | version = b'' |
|
170 | 178 | hook_content = read_hook_content(hook_path) |
|
171 | 179 | matches = re.search(rb'RC_HOOK_VER\s*=\s*(.*)', hook_content) |
|
172 | 180 | if matches: |
|
173 | 181 | try: |
|
174 | 182 | version = matches.groups()[0] |
|
175 | 183 | log.debug('got version %s from hooks.', version) |
|
176 | 184 | except Exception: |
|
177 | 185 | log.exception("Exception while reading the hook version.") |
|
178 | 186 | return version.replace(b"'", b"") |
|
179 | 187 | |
|
180 | 188 | |
|
181 | 189 | def check_rhodecode_hook(hook_path): |
|
182 | 190 | """ |
|
183 | 191 | Check if the hook was created by RhodeCode |
|
184 | 192 | """ |
|
185 | 193 | if not os.path.exists(hook_path): |
|
186 | 194 | return True |
|
187 | 195 | |
|
188 | 196 | log.debug('hook exists, checking if it is from RhodeCode') |
|
189 | 197 | |
|
190 | 198 | version = get_version_from_hook(hook_path) |
|
191 | 199 | if version: |
|
192 | 200 | return True |
|
193 | 201 | |
|
194 | 202 | return False |
|
195 | 203 | |
|
196 | 204 | |
|
197 | 205 | def read_hook_content(hook_path) -> bytes: |
|
198 | 206 | content = b'' |
|
199 | 207 | if os.path.isfile(hook_path): |
|
200 | 208 | with open(hook_path, 'rb') as f: |
|
201 | 209 | content = f.read() |
|
202 | 210 | return content |
|
203 | 211 | |
|
204 | 212 | |
|
205 | 213 | def get_git_pre_hook_version(repo_path, bare): |
|
206 | 214 | hooks_path = get_git_hooks_path(repo_path, bare) |
|
207 | 215 | _hook_file = os.path.join(hooks_path, 'pre-receive') |
|
208 | 216 | version = get_version_from_hook(_hook_file) |
|
209 | 217 | return version |
|
210 | 218 | |
|
211 | 219 | |
|
212 | 220 | def get_git_post_hook_version(repo_path, bare): |
|
213 | 221 | hooks_path = get_git_hooks_path(repo_path, bare) |
|
214 | 222 | _hook_file = os.path.join(hooks_path, 'post-receive') |
|
215 | 223 | version = get_version_from_hook(_hook_file) |
|
216 | 224 | return version |
|
217 | 225 | |
|
218 | 226 | |
|
219 | 227 | def get_svn_pre_hook_version(repo_path): |
|
220 | 228 | hooks_path = get_svn_hooks_path(repo_path) |
|
221 | 229 | _hook_file = os.path.join(hooks_path, 'pre-commit') |
|
222 | 230 | version = get_version_from_hook(_hook_file) |
|
223 | 231 | return version |
|
224 | 232 | |
|
225 | 233 | |
|
226 | 234 | def get_svn_post_hook_version(repo_path): |
|
227 | 235 | hooks_path = get_svn_hooks_path(repo_path) |
|
228 | 236 | _hook_file = os.path.join(hooks_path, 'post-commit') |
|
229 | 237 | version = get_version_from_hook(_hook_file) |
|
230 | 238 | return version |
@@ -1,51 +1,59 b'' | |||
|
1 | 1 | #!_ENV_ |
|
2 | ||
|
2 | 3 | import os |
|
3 | 4 | import sys |
|
4 | 5 | path_adjust = [_PATH_] |
|
5 | 6 | |
|
6 | 7 | if path_adjust: |
|
7 | 8 | sys.path = path_adjust |
|
8 | 9 | |
|
10 | # special trick to pass in some information from rc to hooks | |
|
11 | # mod_dav strips ALL env vars and we can't even access things like PATH | |
|
12 | for env_k, env_v in _OS_EXPAND_: | |
|
13 | os.environ[env_k] = env_v | |
|
14 | ||
|
9 | 15 | try: |
|
10 | 16 | from vcsserver import hooks |
|
11 | 17 | except ImportError: |
|
12 | 18 | if os.environ.get('RC_DEBUG_GIT_HOOK'): |
|
13 | 19 | import traceback |
|
14 | 20 | print(traceback.format_exc()) |
|
15 | 21 | hooks = None |
|
16 | 22 | |
|
17 | 23 | |
|
18 | 24 | # TIMESTAMP: _DATE_ |
|
19 | 25 | RC_HOOK_VER = '_TMPL_' |
|
20 | 26 | |
|
21 | 27 | |
|
22 | 28 | def main(): |
|
23 | 29 | if hooks is None: |
|
24 | 30 | # exit with success if we cannot import vcsserver.hooks !! |
|
25 | 31 | # this allows simply push to this repo even without rhodecode |
|
26 | 32 | sys.exit(0) |
|
27 | 33 | |
|
28 | 34 | if os.environ.get('RC_SKIP_HOOKS') or os.environ.get('RC_SKIP_GIT_HOOKS'): |
|
29 | 35 | sys.exit(0) |
|
30 | 36 | |
|
31 | 37 | repo_path = os.getcwd() |
|
32 | 38 | push_data = sys.stdin.readlines() |
|
33 | os.environ['RC_HOOK_VER'] = RC_HOOK_VER | |
|
39 | ||
|
34 | 40 | # os.environ is modified here by a subprocess call that |
|
35 | 41 | # runs git and later git executes this hook. |
|
36 | 42 | # Environ gets some additional info from rhodecode system |
|
37 | 43 | # like IP or username from basic-auth |
|
44 | ||
|
45 | os.environ['RC_HOOK_VER'] = RC_HOOK_VER | |
|
38 | 46 | try: |
|
39 | 47 | result = hooks.git_post_receive(repo_path, push_data, os.environ) |
|
40 | 48 | sys.exit(result) |
|
41 | 49 | except Exception as error: |
|
42 | 50 | # TODO: johbo: Improve handling of this special case |
|
43 | 51 | if not getattr(error, '_vcs_kind', None) == 'repo_locked': |
|
44 | 52 | raise |
|
45 | 53 | print(f'ERROR: {error}') |
|
46 | 54 | sys.exit(1) |
|
47 | 55 | sys.exit(0) |
|
48 | 56 | |
|
49 | 57 | |
|
50 | 58 | if __name__ == '__main__': |
|
51 | 59 | main() |
@@ -1,51 +1,59 b'' | |||
|
1 | 1 | #!_ENV_ |
|
2 | ||
|
2 | 3 | import os |
|
3 | 4 | import sys |
|
4 | 5 | path_adjust = [_PATH_] |
|
5 | 6 | |
|
6 | 7 | if path_adjust: |
|
7 | 8 | sys.path = path_adjust |
|
8 | 9 | |
|
10 | # special trick to pass in some information from rc to hooks | |
|
11 | # mod_dav strips ALL env vars and we can't even access things like PATH | |
|
12 | for env_k, env_v in _OS_EXPAND_: | |
|
13 | os.environ[env_k] = env_v | |
|
14 | ||
|
9 | 15 | try: |
|
10 | 16 | from vcsserver import hooks |
|
11 | 17 | except ImportError: |
|
12 | 18 | if os.environ.get('RC_DEBUG_GIT_HOOK'): |
|
13 | 19 | import traceback |
|
14 | 20 | print(traceback.format_exc()) |
|
15 | 21 | hooks = None |
|
16 | 22 | |
|
17 | 23 | |
|
18 | 24 | # TIMESTAMP: _DATE_ |
|
19 | 25 | RC_HOOK_VER = '_TMPL_' |
|
20 | 26 | |
|
21 | 27 | |
|
22 | 28 | def main(): |
|
23 | 29 | if hooks is None: |
|
24 | 30 | # exit with success if we cannot import vcsserver.hooks !! |
|
25 | 31 | # this allows simply push to this repo even without rhodecode |
|
26 | 32 | sys.exit(0) |
|
27 | 33 | |
|
28 | 34 | if os.environ.get('RC_SKIP_HOOKS') or os.environ.get('RC_SKIP_GIT_HOOKS'): |
|
29 | 35 | sys.exit(0) |
|
30 | 36 | |
|
31 | 37 | repo_path = os.getcwd() |
|
32 | 38 | push_data = sys.stdin.readlines() |
|
33 | os.environ['RC_HOOK_VER'] = RC_HOOK_VER | |
|
39 | ||
|
34 | 40 | # os.environ is modified here by a subprocess call that |
|
35 | 41 | # runs git and later git executes this hook. |
|
36 | 42 | # Environ gets some additional info from rhodecode system |
|
37 | 43 | # like IP or username from basic-auth |
|
44 | ||
|
45 | os.environ['RC_HOOK_VER'] = RC_HOOK_VER | |
|
38 | 46 | try: |
|
39 | 47 | result = hooks.git_pre_receive(repo_path, push_data, os.environ) |
|
40 | 48 | sys.exit(result) |
|
41 | 49 | except Exception as error: |
|
42 | 50 | # TODO: johbo: Improve handling of this special case |
|
43 | 51 | if not getattr(error, '_vcs_kind', None) == 'repo_locked': |
|
44 | 52 | raise |
|
45 | 53 | print(f'ERROR: {error}') |
|
46 | 54 | sys.exit(1) |
|
47 | 55 | sys.exit(0) |
|
48 | 56 | |
|
49 | 57 | |
|
50 | 58 | if __name__ == '__main__': |
|
51 | 59 | main() |
@@ -1,54 +1,59 b'' | |||
|
1 | 1 | #!_ENV_ |
|
2 | 2 | |
|
3 | 3 | import os |
|
4 | 4 | import sys |
|
5 | 5 | path_adjust = [_PATH_] |
|
6 | 6 | |
|
7 | 7 | if path_adjust: |
|
8 | 8 | sys.path = path_adjust |
|
9 | 9 | |
|
10 | # special trick to pass in some information from rc to hooks | |
|
11 | # mod_dav strips ALL env vars and we can't even access things like PATH | |
|
12 | for env_k, env_v in _OS_EXPAND_: | |
|
13 | os.environ[env_k] = env_v | |
|
14 | ||
|
10 | 15 | try: |
|
11 | 16 | from vcsserver import hooks |
|
12 | 17 | except ImportError: |
|
13 | 18 | if os.environ.get('RC_DEBUG_SVN_HOOK'): |
|
14 | 19 | import traceback |
|
15 | 20 | print(traceback.format_exc()) |
|
16 | 21 | hooks = None |
|
17 | 22 | |
|
18 | 23 | |
|
19 | 24 | # TIMESTAMP: _DATE_ |
|
20 | 25 | RC_HOOK_VER = '_TMPL_' |
|
21 | 26 | |
|
22 | 27 | |
|
23 | 28 | # special trick to pass in some information from rc to hooks |
|
24 | 29 | # mod_dav strips ALL env vars and we can't even access things like PATH |
|
25 | 30 | for env_k, env_v in _OS_EXPAND_: |
|
26 | 31 | os.environ[env_k] = env_v |
|
27 | 32 | |
|
28 | 33 | def main(): |
|
29 | 34 | if hooks is None: |
|
30 | 35 | # exit with success if we cannot import vcsserver.hooks !! |
|
31 | 36 | # this allows simply push to this repo even without rhodecode |
|
32 | 37 | sys.exit(0) |
|
33 | 38 | |
|
34 | 39 | if os.environ.get('RC_SKIP_HOOKS') or os.environ.get('RC_SKIP_SVN_HOOKS'): |
|
35 | 40 | sys.exit(0) |
|
36 | repo_path = os.getcwd() | |
|
41 | cwd_repo_path = os.getcwd() | |
|
37 | 42 | push_data = sys.argv[1:] |
|
38 | 43 | |
|
39 | 44 | os.environ['RC_HOOK_VER'] = RC_HOOK_VER |
|
40 | 45 | |
|
41 | 46 | try: |
|
42 | result = hooks.svn_post_commit(repo_path, push_data, os.environ) | |
|
47 | result = hooks.svn_post_commit(cwd_repo_path, push_data, os.environ) | |
|
43 | 48 | sys.exit(result) |
|
44 | 49 | except Exception as error: |
|
45 | 50 | # TODO: johbo: Improve handling of this special case |
|
46 | 51 | if not getattr(error, '_vcs_kind', None) == 'repo_locked': |
|
47 | 52 | raise |
|
48 | 53 | print(f'ERROR: {error}') |
|
49 | 54 | sys.exit(1) |
|
50 | 55 | sys.exit(0) |
|
51 | 56 | |
|
52 | 57 | |
|
53 | 58 | if __name__ == '__main__': |
|
54 | 59 | main() |
@@ -1,58 +1,62 b'' | |||
|
1 | 1 | #!_ENV_ |
|
2 | 2 | |
|
3 | 3 | import os |
|
4 | 4 | import sys |
|
5 | 5 | path_adjust = [_PATH_] |
|
6 | 6 | |
|
7 | 7 | if path_adjust: |
|
8 | 8 | sys.path = path_adjust |
|
9 | 9 | |
|
10 | # special trick to pass in some information from rc to hooks | |
|
11 | # mod_dav strips ALL env vars and we can't even access things like PATH | |
|
12 | for env_k, env_v in _OS_EXPAND_: | |
|
13 | os.environ[env_k] = env_v | |
|
14 | ||
|
10 | 15 | try: |
|
11 | 16 | from vcsserver import hooks |
|
12 | 17 | except ImportError: |
|
13 | 18 | if os.environ.get('RC_DEBUG_SVN_HOOK'): |
|
14 | 19 | import traceback |
|
15 | 20 | print(traceback.format_exc()) |
|
16 | 21 | hooks = None |
|
17 | 22 | |
|
18 | 23 | |
|
19 | 24 | # TIMESTAMP: _DATE_ |
|
20 | 25 | RC_HOOK_VER = '_TMPL_' |
|
21 | 26 | |
|
22 | 27 | |
|
23 | 28 | # special trick to pass in some information from rc to hooks |
|
24 | 29 | # mod_dav strips ALL env vars and we can't even access things like PATH |
|
25 | 30 | for env_k, env_v in _OS_EXPAND_: |
|
26 | 31 | os.environ[env_k] = env_v |
|
27 | 32 | |
|
28 | 33 | def main(): |
|
29 | 34 | if os.environ.get('SSH_READ_ONLY') == '1': |
|
30 | 35 | sys.stderr.write('Only read-only access is allowed') |
|
31 | 36 | sys.exit(1) |
|
32 | 37 | |
|
33 | 38 | if hooks is None: |
|
34 | 39 | # exit with success if we cannot import vcsserver.hooks !! |
|
35 | 40 | # this allows simply push to this repo even without rhodecode |
|
36 | 41 | sys.exit(0) |
|
37 | 42 | |
|
38 | 43 | if os.environ.get('RC_SKIP_HOOKS') or os.environ.get('RC_SKIP_SVN_HOOKS'): |
|
39 | 44 | sys.exit(0) |
|
40 | repo_path = os.getcwd() | |
|
45 | cwd_repo_path = os.getcwd() | |
|
41 | 46 | push_data = sys.argv[1:] |
|
42 | 47 | |
|
43 | 48 | os.environ['RC_HOOK_VER'] = RC_HOOK_VER |
|
44 | ||
|
45 | 49 | try: |
|
46 | result = hooks.svn_pre_commit(repo_path, push_data, os.environ) | |
|
50 | result = hooks.svn_pre_commit(cwd_repo_path, push_data, os.environ) | |
|
47 | 51 | sys.exit(result) |
|
48 | 52 | except Exception as error: |
|
49 | 53 | # TODO: johbo: Improve handling of this special case |
|
50 | 54 | if not getattr(error, '_vcs_kind', None) == 'repo_locked': |
|
51 | 55 | raise |
|
52 | 56 | print(f'ERROR: {error}') |
|
53 | 57 | sys.exit(1) |
|
54 | 58 | sys.exit(0) |
|
55 | 59 | |
|
56 | 60 | |
|
57 | 61 | if __name__ == '__main__': |
|
58 | 62 | main() |
@@ -1,826 +1,822 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import io |
|
19 | 19 | import os |
|
20 | 20 | import sys |
|
21 | 21 | import logging |
|
22 | 22 | import collections |
|
23 | 23 | import base64 |
|
24 | 24 | import msgpack |
|
25 | 25 | import dataclasses |
|
26 | 26 | import pygit2 |
|
27 | 27 | |
|
28 | 28 | import http.client |
|
29 | 29 | from celery import Celery |
|
30 | 30 | |
|
31 | 31 | import mercurial.scmutil |
|
32 | 32 | import mercurial.node |
|
33 | 33 | |
|
34 | from vcsserver.lib.rc_json import json | |
|
35 | 34 | from vcsserver import exceptions, subprocessio, settings |
|
36 | from vcsserver.str_utils import ascii_str, safe_str | |
|
35 | from vcsserver.lib.ext_json import json | |
|
36 | from vcsserver.lib.str_utils import ascii_str, safe_str | |
|
37 | from vcsserver.lib.svn_txn_utils import get_txn_id_from_store | |
|
37 | 38 | from vcsserver.remote.git_remote import Repository |
|
38 | 39 | |
|
39 | 40 | celery_app = Celery('__vcsserver__') |
|
40 | 41 | log = logging.getLogger(__name__) |
|
41 | 42 | |
|
42 | 43 | |
|
43 | 44 | class HooksHttpClient: |
|
44 | 45 | proto = 'msgpack.v1' |
|
45 | 46 | connection = None |
|
46 | 47 | |
|
47 | 48 | def __init__(self, hooks_uri): |
|
48 | 49 | self.hooks_uri = hooks_uri |
|
49 | 50 | |
|
50 | 51 | def __repr__(self): |
|
51 | 52 | return f'{self.__class__}(hook_uri={self.hooks_uri}, proto={self.proto})' |
|
52 | 53 | |
|
53 | 54 | def __call__(self, method, extras): |
|
54 | 55 | connection = http.client.HTTPConnection(self.hooks_uri) |
|
55 | 56 | # binary msgpack body |
|
56 | 57 | headers, body = self._serialize(method, extras) |
|
57 | 58 | log.debug('Doing a new hooks call using HTTPConnection to %s', self.hooks_uri) |
|
58 | 59 | |
|
59 | 60 | try: |
|
60 | 61 | try: |
|
61 | 62 | connection.request('POST', '/', body, headers) |
|
62 | 63 | except Exception as error: |
|
63 | 64 | log.error('Hooks calling Connection failed on %s, org error: %s', connection.__dict__, error) |
|
64 | 65 | raise |
|
65 | 66 | |
|
66 | 67 | response = connection.getresponse() |
|
67 | 68 | try: |
|
68 | 69 | return msgpack.load(response) |
|
69 | 70 | except Exception: |
|
70 | 71 | response_data = response.read() |
|
71 | 72 | log.exception('Failed to decode hook response json data. ' |
|
72 | 73 | 'response_code:%s, raw_data:%s', |
|
73 | 74 | response.status, response_data) |
|
74 | 75 | raise |
|
75 | 76 | finally: |
|
76 | 77 | connection.close() |
|
77 | 78 | |
|
78 | 79 | @classmethod |
|
79 | 80 | def _serialize(cls, hook_name, extras): |
|
80 | 81 | data = { |
|
81 | 82 | 'method': hook_name, |
|
82 | 83 | 'extras': extras |
|
83 | 84 | } |
|
84 | 85 | headers = { |
|
85 | 86 | "rc-hooks-protocol": cls.proto, |
|
86 | 87 | "Connection": "keep-alive" |
|
87 | 88 | } |
|
88 | 89 | return headers, msgpack.packb(data) |
|
89 | 90 | |
|
90 | 91 | |
|
91 | 92 | class HooksCeleryClient: |
|
92 | 93 | TASK_TIMEOUT = 60 # time in seconds |
|
93 | 94 | |
|
94 | 95 | def __init__(self, queue, backend): |
|
95 | 96 | celery_app.config_from_object({ |
|
96 | 97 | 'broker_url': queue, 'result_backend': backend, |
|
97 | 98 | 'broker_connection_retry_on_startup': True, |
|
98 |
'task_serializer': ' |
|
|
99 | 'task_serializer': 'json', | |
|
99 | 100 | 'accept_content': ['json', 'msgpack'], |
|
100 |
'result_serializer': ' |
|
|
101 | 'result_serializer': 'json', | |
|
101 | 102 | 'result_accept_content': ['json', 'msgpack'] |
|
102 | 103 | }) |
|
103 | 104 | self.celery_app = celery_app |
|
104 | 105 | |
|
105 | 106 | def __call__(self, method, extras): |
|
106 | 107 | inquired_task = self.celery_app.signature( |
|
107 | 108 | f'rhodecode.lib.celerylib.tasks.{method}' |
|
108 | 109 | ) |
|
109 | 110 | return inquired_task.delay(extras).get(timeout=self.TASK_TIMEOUT) |
|
110 | 111 | |
|
111 | 112 | |
|
112 | 113 | class HooksShadowRepoClient: |
|
113 | 114 | |
|
114 | 115 | def __call__(self, hook_name, extras): |
|
115 | 116 | return {'output': '', 'status': 0} |
|
116 | 117 | |
|
117 | 118 | |
|
118 | 119 | class RemoteMessageWriter: |
|
119 | 120 | """Writer base class.""" |
|
120 | 121 | def write(self, message): |
|
121 | 122 | raise NotImplementedError() |
|
122 | 123 | |
|
123 | 124 | |
|
124 | 125 | class HgMessageWriter(RemoteMessageWriter): |
|
125 | 126 | """Writer that knows how to send messages to mercurial clients.""" |
|
126 | 127 | |
|
127 | 128 | def __init__(self, ui): |
|
128 | 129 | self.ui = ui |
|
129 | 130 | |
|
130 | 131 | def write(self, message: str): |
|
131 | 132 | # TODO: Check why the quiet flag is set by default. |
|
132 | 133 | old = self.ui.quiet |
|
133 | 134 | self.ui.quiet = False |
|
134 | 135 | self.ui.status(message.encode('utf-8')) |
|
135 | 136 | self.ui.quiet = old |
|
136 | 137 | |
|
137 | 138 | |
|
138 | 139 | class GitMessageWriter(RemoteMessageWriter): |
|
139 | 140 | """Writer that knows how to send messages to git clients.""" |
|
140 | 141 | |
|
141 | 142 | def __init__(self, stdout=None): |
|
142 | 143 | self.stdout = stdout or sys.stdout |
|
143 | 144 | |
|
144 | 145 | def write(self, message: str): |
|
145 | 146 | self.stdout.write(message) |
|
146 | 147 | |
|
147 | 148 | |
|
148 | 149 | class SvnMessageWriter(RemoteMessageWriter): |
|
149 | 150 | """Writer that knows how to send messages to svn clients.""" |
|
150 | 151 | |
|
151 | 152 | def __init__(self, stderr=None): |
|
152 | 153 | # SVN needs data sent to stderr for back-to-client messaging |
|
153 | 154 | self.stderr = stderr or sys.stderr |
|
154 | 155 | |
|
155 | 156 | def write(self, message): |
|
156 | 157 | self.stderr.write(message) |
|
157 | 158 | |
|
158 | 159 | |
|
159 | 160 | def _handle_exception(result): |
|
160 | 161 | exception_class = result.get('exception') |
|
161 | 162 | exception_traceback = result.get('exception_traceback') |
|
162 | 163 | log.debug('Handling hook-call exception: %s', exception_class) |
|
163 | 164 | |
|
164 | 165 | if exception_traceback: |
|
165 | 166 | log.error('Got traceback from remote call:%s', exception_traceback) |
|
166 | 167 | |
|
167 | 168 | if exception_class == 'HTTPLockedRC': |
|
168 | 169 | raise exceptions.RepositoryLockedException()(*result['exception_args']) |
|
169 | 170 | elif exception_class == 'HTTPBranchProtected': |
|
170 | 171 | raise exceptions.RepositoryBranchProtectedException()(*result['exception_args']) |
|
171 | 172 | elif exception_class == 'RepositoryError': |
|
172 | 173 | raise exceptions.VcsException()(*result['exception_args']) |
|
173 | 174 | elif exception_class: |
|
174 | 175 | raise Exception( |
|
175 | 176 | f"""Got remote exception "{exception_class}" with args "{result['exception_args']}" """ |
|
176 | 177 | ) |
|
177 | 178 | |
|
178 | 179 | |
|
179 | 180 | def _get_hooks_client(extras): |
|
180 | 181 | hooks_uri = extras.get('hooks_uri') |
|
181 | 182 | task_queue = extras.get('task_queue') |
|
182 | 183 | task_backend = extras.get('task_backend') |
|
183 | 184 | is_shadow_repo = extras.get('is_shadow_repo') |
|
184 | 185 | |
|
185 | 186 | if hooks_uri: |
|
186 | 187 | return HooksHttpClient(hooks_uri) |
|
187 | 188 | elif task_queue and task_backend: |
|
188 | 189 | return HooksCeleryClient(task_queue, task_backend) |
|
189 | 190 | elif is_shadow_repo: |
|
190 | 191 | return HooksShadowRepoClient() |
|
191 | 192 | else: |
|
192 | 193 | raise Exception("Hooks client not found!") |
|
193 | 194 | |
|
194 | 195 | |
|
195 | 196 | def _call_hook(hook_name, extras, writer): |
|
196 | 197 | hooks_client = _get_hooks_client(extras) |
|
197 | 198 | log.debug('Hooks, using client:%s', hooks_client) |
|
198 | 199 | result = hooks_client(hook_name, extras) |
|
199 | 200 | log.debug('Hooks got result: %s', result) |
|
200 | 201 | _handle_exception(result) |
|
201 | 202 | writer.write(result['output']) |
|
202 | 203 | |
|
203 | 204 | return result['status'] |
|
204 | 205 | |
|
205 | 206 | |
|
206 | 207 | def _extras_from_ui(ui): |
|
207 | 208 | hook_data = ui.config(b'rhodecode', b'RC_SCM_DATA') |
|
208 | 209 | if not hook_data: |
|
209 | 210 | # maybe it's inside environ ? |
|
210 | 211 | env_hook_data = os.environ.get('RC_SCM_DATA') |
|
211 | 212 | if env_hook_data: |
|
212 | 213 | hook_data = env_hook_data |
|
213 | 214 | |
|
214 | 215 | extras = {} |
|
215 | 216 | if hook_data: |
|
216 | 217 | extras = json.loads(hook_data) |
|
217 | 218 | return extras |
|
218 | 219 | |
|
219 | 220 | |
|
220 | 221 | def _rev_range_hash(repo, node, check_heads=False): |
|
221 | 222 | from vcsserver.hgcompat import get_ctx |
|
222 | 223 | |
|
223 | 224 | commits = [] |
|
224 | 225 | revs = [] |
|
225 | 226 | start = get_ctx(repo, node).rev() |
|
226 | 227 | end = len(repo) |
|
227 | 228 | for rev in range(start, end): |
|
228 | 229 | revs.append(rev) |
|
229 | 230 | ctx = get_ctx(repo, rev) |
|
230 | 231 | commit_id = ascii_str(mercurial.node.hex(ctx.node())) |
|
231 | 232 | branch = safe_str(ctx.branch()) |
|
232 | 233 | commits.append((commit_id, branch)) |
|
233 | 234 | |
|
234 | 235 | parent_heads = [] |
|
235 | 236 | if check_heads: |
|
236 | 237 | parent_heads = _check_heads(repo, start, end, revs) |
|
237 | 238 | return commits, parent_heads |
|
238 | 239 | |
|
239 | 240 | |
|
240 | 241 | def _check_heads(repo, start, end, commits): |
|
241 | 242 | from vcsserver.hgcompat import get_ctx |
|
242 | 243 | changelog = repo.changelog |
|
243 | 244 | parents = set() |
|
244 | 245 | |
|
245 | 246 | for new_rev in commits: |
|
246 | 247 | for p in changelog.parentrevs(new_rev): |
|
247 | 248 | if p == mercurial.node.nullrev: |
|
248 | 249 | continue |
|
249 | 250 | if p < start: |
|
250 | 251 | parents.add(p) |
|
251 | 252 | |
|
252 | 253 | for p in parents: |
|
253 | 254 | branch = get_ctx(repo, p).branch() |
|
254 | 255 | # The heads descending from that parent, on the same branch |
|
255 | 256 | parent_heads = {p} |
|
256 | 257 | reachable = {p} |
|
257 | 258 | for x in range(p + 1, end): |
|
258 | 259 | if get_ctx(repo, x).branch() != branch: |
|
259 | 260 | continue |
|
260 | 261 | for pp in changelog.parentrevs(x): |
|
261 | 262 | if pp in reachable: |
|
262 | 263 | reachable.add(x) |
|
263 | 264 | parent_heads.discard(pp) |
|
264 | 265 | parent_heads.add(x) |
|
265 | 266 | # More than one head? Suggest merging |
|
266 | 267 | if len(parent_heads) > 1: |
|
267 | 268 | return list(parent_heads) |
|
268 | 269 | |
|
269 | 270 | return [] |
|
270 | 271 | |
|
271 | 272 | |
|
272 | 273 | def _get_git_env(): |
|
273 | 274 | env = {} |
|
274 | 275 | for k, v in os.environ.items(): |
|
275 | 276 | if k.startswith('GIT'): |
|
276 | 277 | env[k] = v |
|
277 | 278 | |
|
278 | 279 | # serialized version |
|
279 | 280 | return [(k, v) for k, v in env.items()] |
|
280 | 281 | |
|
281 | 282 | |
|
282 | 283 | def _get_hg_env(old_rev, new_rev, txnid, repo_path): |
|
283 | 284 | env = {} |
|
284 | 285 | for k, v in os.environ.items(): |
|
285 | 286 | if k.startswith('HG'): |
|
286 | 287 | env[k] = v |
|
287 | 288 | |
|
288 | 289 | env['HG_NODE'] = old_rev |
|
289 | 290 | env['HG_NODE_LAST'] = new_rev |
|
290 | 291 | env['HG_TXNID'] = txnid |
|
291 | 292 | env['HG_PENDING'] = repo_path |
|
292 | 293 | |
|
293 | 294 | return [(k, v) for k, v in env.items()] |
|
294 | 295 | |
|
295 | 296 | |
|
297 | def _get_ini_settings(ini_file): | |
|
298 | from vcsserver.http_main import sanitize_settings_and_apply_defaults | |
|
299 | from vcsserver.lib.config_utils import get_app_config_lightweight, configure_and_store_settings | |
|
300 | ||
|
301 | global_config = {'__file__': ini_file} | |
|
302 | ini_settings = get_app_config_lightweight(ini_file) | |
|
303 | sanitize_settings_and_apply_defaults(global_config, ini_settings) | |
|
304 | configure_and_store_settings(global_config, ini_settings) | |
|
305 | ||
|
306 | return ini_settings | |
|
307 | ||
|
308 | ||
|
296 | 309 | def _fix_hooks_executables(ini_path=''): |
|
297 | 310 | """ |
|
298 | 311 | This is a trick to set proper settings.EXECUTABLE paths for certain execution patterns |
|
299 | 312 | especially for subversion where hooks strip entire env, and calling just 'svn' command will most likely fail |
|
300 | 313 | because svn is not on PATH |
|
301 | 314 | """ |
|
302 | from vcsserver.http_main import sanitize_settings_and_apply_defaults | |
|
303 | from vcsserver.lib.config_utils import get_app_config_lightweight | |
|
304 | ||
|
315 | # set defaults, in case we can't read from ini_file | |
|
305 | 316 | core_binary_dir = settings.BINARY_DIR or '/usr/local/bin/rhodecode_bin/vcs_bin' |
|
306 | 317 | if ini_path: |
|
307 | ||
|
308 | ini_settings = get_app_config_lightweight(ini_path) | |
|
309 | ini_settings = sanitize_settings_and_apply_defaults({'__file__': ini_path}, ini_settings) | |
|
318 | ini_settings = _get_ini_settings(ini_path) | |
|
310 | 319 | core_binary_dir = ini_settings['core.binary_dir'] |
|
311 | 320 | |
|
312 | 321 | settings.BINARY_DIR = core_binary_dir |
|
313 | 322 | |
|
314 | 323 | |
|
315 | 324 | def repo_size(ui, repo, **kwargs): |
|
316 | 325 | extras = _extras_from_ui(ui) |
|
317 | 326 | return _call_hook('repo_size', extras, HgMessageWriter(ui)) |
|
318 | 327 | |
|
319 | 328 | |
|
320 | 329 | def pre_pull(ui, repo, **kwargs): |
|
321 | 330 | extras = _extras_from_ui(ui) |
|
322 | 331 | return _call_hook('pre_pull', extras, HgMessageWriter(ui)) |
|
323 | 332 | |
|
324 | 333 | |
|
325 | 334 | def pre_pull_ssh(ui, repo, **kwargs): |
|
326 | 335 | extras = _extras_from_ui(ui) |
|
327 | 336 | if extras and extras.get('SSH'): |
|
328 | 337 | return pre_pull(ui, repo, **kwargs) |
|
329 | 338 | return 0 |
|
330 | 339 | |
|
331 | 340 | |
|
332 | 341 | def post_pull(ui, repo, **kwargs): |
|
333 | 342 | extras = _extras_from_ui(ui) |
|
334 | 343 | return _call_hook('post_pull', extras, HgMessageWriter(ui)) |
|
335 | 344 | |
|
336 | 345 | |
|
337 | 346 | def post_pull_ssh(ui, repo, **kwargs): |
|
338 | 347 | extras = _extras_from_ui(ui) |
|
339 | 348 | if extras and extras.get('SSH'): |
|
340 | 349 | return post_pull(ui, repo, **kwargs) |
|
341 | 350 | return 0 |
|
342 | 351 | |
|
343 | 352 | |
|
344 | 353 | def pre_push(ui, repo, node=None, **kwargs): |
|
345 | 354 | """ |
|
346 | 355 | Mercurial pre_push hook |
|
347 | 356 | """ |
|
348 | 357 | extras = _extras_from_ui(ui) |
|
349 | 358 | detect_force_push = extras.get('detect_force_push') |
|
350 | 359 | |
|
351 | 360 | rev_data = [] |
|
352 | 361 | hook_type: str = safe_str(kwargs.get('hooktype')) |
|
353 | 362 | |
|
354 | 363 | if node and hook_type == 'pretxnchangegroup': |
|
355 | 364 | branches = collections.defaultdict(list) |
|
356 | 365 | commits, _heads = _rev_range_hash(repo, node, check_heads=detect_force_push) |
|
357 | 366 | for commit_id, branch in commits: |
|
358 | 367 | branches[branch].append(commit_id) |
|
359 | 368 | |
|
360 | 369 | for branch, commits in branches.items(): |
|
361 | 370 | old_rev = ascii_str(kwargs.get('node_last')) or commits[0] |
|
362 | 371 | rev_data.append({ |
|
363 | 372 | 'total_commits': len(commits), |
|
364 | 373 | 'old_rev': old_rev, |
|
365 | 374 | 'new_rev': commits[-1], |
|
366 | 375 | 'ref': '', |
|
367 | 376 | 'type': 'branch', |
|
368 | 377 | 'name': branch, |
|
369 | 378 | }) |
|
370 | 379 | |
|
371 | 380 | for push_ref in rev_data: |
|
372 | 381 | push_ref['multiple_heads'] = _heads |
|
373 | 382 | |
|
374 | 383 | repo_path = os.path.join( |
|
375 | 384 | extras.get('repo_store', ''), extras.get('repository', '')) |
|
376 | 385 | push_ref['hg_env'] = _get_hg_env( |
|
377 | 386 | old_rev=push_ref['old_rev'], |
|
378 | 387 | new_rev=push_ref['new_rev'], txnid=ascii_str(kwargs.get('txnid')), |
|
379 | 388 | repo_path=repo_path) |
|
380 | 389 | |
|
381 | 390 | extras['hook_type'] = hook_type or 'pre_push' |
|
382 | 391 | extras['commit_ids'] = rev_data |
|
383 | 392 | |
|
384 | 393 | return _call_hook('pre_push', extras, HgMessageWriter(ui)) |
|
385 | 394 | |
|
386 | 395 | |
|
387 | 396 | def pre_push_ssh(ui, repo, node=None, **kwargs): |
|
388 | 397 | extras = _extras_from_ui(ui) |
|
389 | 398 | if extras.get('SSH'): |
|
390 | 399 | return pre_push(ui, repo, node, **kwargs) |
|
391 | 400 | |
|
392 | 401 | return 0 |
|
393 | 402 | |
|
394 | 403 | |
|
395 | 404 | def pre_push_ssh_auth(ui, repo, node=None, **kwargs): |
|
396 | 405 | """ |
|
397 | 406 | Mercurial pre_push hook for SSH |
|
398 | 407 | """ |
|
399 | 408 | extras = _extras_from_ui(ui) |
|
400 | 409 | if extras.get('SSH'): |
|
401 | 410 | permission = extras['SSH_PERMISSIONS'] |
|
402 | 411 | |
|
403 | 412 | if 'repository.write' == permission or 'repository.admin' == permission: |
|
404 | 413 | return 0 |
|
405 | 414 | |
|
406 | 415 | # non-zero ret code |
|
407 | 416 | return 1 |
|
408 | 417 | |
|
409 | 418 | return 0 |
|
410 | 419 | |
|
411 | 420 | |
|
412 | 421 | def post_push(ui, repo, node, **kwargs): |
|
413 | 422 | """ |
|
414 | 423 | Mercurial post_push hook |
|
415 | 424 | """ |
|
416 | 425 | extras = _extras_from_ui(ui) |
|
417 | 426 | |
|
418 | 427 | commit_ids = [] |
|
419 | 428 | branches = [] |
|
420 | 429 | bookmarks = [] |
|
421 | 430 | tags = [] |
|
422 | 431 | hook_type: str = safe_str(kwargs.get('hooktype')) |
|
423 | 432 | |
|
424 | 433 | commits, _heads = _rev_range_hash(repo, node) |
|
425 | 434 | for commit_id, branch in commits: |
|
426 | 435 | commit_ids.append(commit_id) |
|
427 | 436 | if branch not in branches: |
|
428 | 437 | branches.append(branch) |
|
429 | 438 | |
|
430 | 439 | if hasattr(ui, '_rc_pushkey_bookmarks'): |
|
431 | 440 | bookmarks = ui._rc_pushkey_bookmarks |
|
432 | 441 | |
|
433 | 442 | extras['hook_type'] = hook_type or 'post_push' |
|
434 | 443 | extras['commit_ids'] = commit_ids |
|
435 | 444 | |
|
436 | 445 | extras['new_refs'] = { |
|
437 | 446 | 'branches': branches, |
|
438 | 447 | 'bookmarks': bookmarks, |
|
439 | 448 | 'tags': tags |
|
440 | 449 | } |
|
441 | 450 | |
|
442 | 451 | return _call_hook('post_push', extras, HgMessageWriter(ui)) |
|
443 | 452 | |
|
444 | 453 | |
|
445 | 454 | def post_push_ssh(ui, repo, node, **kwargs): |
|
446 | 455 | """ |
|
447 | 456 | Mercurial post_push hook for SSH |
|
448 | 457 | """ |
|
449 | 458 | if _extras_from_ui(ui).get('SSH'): |
|
450 | 459 | return post_push(ui, repo, node, **kwargs) |
|
451 | 460 | return 0 |
|
452 | 461 | |
|
453 | 462 | |
|
454 | 463 | def key_push(ui, repo, **kwargs): |
|
455 | 464 | from vcsserver.hgcompat import get_ctx |
|
456 | 465 | |
|
457 | 466 | if kwargs['new'] != b'0' and kwargs['namespace'] == b'bookmarks': |
|
458 | 467 | # store new bookmarks in our UI object propagated later to post_push |
|
459 | 468 | ui._rc_pushkey_bookmarks = get_ctx(repo, kwargs['key']).bookmarks() |
|
460 | 469 | return |
|
461 | 470 | |
|
462 | 471 | |
|
463 | 472 | # backward compat |
|
464 | 473 | log_pull_action = post_pull |
|
465 | 474 | |
|
466 | 475 | # backward compat |
|
467 | 476 | log_push_action = post_push |
|
468 | 477 | |
|
469 | 478 | |
|
470 | 479 | def handle_git_pre_receive(unused_repo_path, unused_revs, unused_env): |
|
471 | 480 | """ |
|
472 | 481 | Old hook name: keep here for backward compatibility. |
|
473 | 482 | |
|
474 | 483 | This is only required when the installed git hooks are not upgraded. |
|
475 | 484 | """ |
|
476 | 485 | pass |
|
477 | 486 | |
|
478 | 487 | |
|
479 | 488 | def handle_git_post_receive(unused_repo_path, unused_revs, unused_env): |
|
480 | 489 | """ |
|
481 | 490 | Old hook name: keep here for backward compatibility. |
|
482 | 491 | |
|
483 | 492 | This is only required when the installed git hooks are not upgraded. |
|
484 | 493 | """ |
|
485 | 494 | pass |
|
486 | 495 | |
|
487 | 496 | |
|
488 | 497 | @dataclasses.dataclass |
|
489 | 498 | class HookResponse: |
|
490 | 499 | status: int |
|
491 | 500 | output: str |
|
492 | 501 | |
|
493 | 502 | |
|
494 | 503 | def git_pre_pull(extras) -> HookResponse: |
|
495 | 504 | """ |
|
496 | 505 | Pre pull hook. |
|
497 | 506 | |
|
498 | 507 | :param extras: dictionary containing the keys defined in simplevcs |
|
499 | 508 | :type extras: dict |
|
500 | 509 | |
|
501 | 510 | :return: status code of the hook. 0 for success. |
|
502 | 511 | :rtype: int |
|
503 | 512 | """ |
|
504 | 513 | |
|
505 | 514 | if 'pull' not in extras['hooks']: |
|
506 | 515 | return HookResponse(0, '') |
|
507 | 516 | |
|
508 | 517 | stdout = io.StringIO() |
|
509 | 518 | try: |
|
510 | 519 | status_code = _call_hook('pre_pull', extras, GitMessageWriter(stdout)) |
|
511 | 520 | |
|
512 | 521 | except Exception as error: |
|
513 | 522 | log.exception('Failed to call pre_pull hook') |
|
514 | 523 | status_code = 128 |
|
515 | 524 | stdout.write(f'ERROR: {error}\n') |
|
516 | 525 | |
|
517 | 526 | return HookResponse(status_code, stdout.getvalue()) |
|
518 | 527 | |
|
519 | 528 | |
|
520 | 529 | def git_post_pull(extras) -> HookResponse: |
|
521 | 530 | """ |
|
522 | 531 | Post pull hook. |
|
523 | 532 | |
|
524 | 533 | :param extras: dictionary containing the keys defined in simplevcs |
|
525 | 534 | :type extras: dict |
|
526 | 535 | |
|
527 | 536 | :return: status code of the hook. 0 for success. |
|
528 | 537 | :rtype: int |
|
529 | 538 | """ |
|
530 | 539 | if 'pull' not in extras['hooks']: |
|
531 | 540 | return HookResponse(0, '') |
|
532 | 541 | |
|
533 | 542 | stdout = io.StringIO() |
|
534 | 543 | try: |
|
535 | 544 | status = _call_hook('post_pull', extras, GitMessageWriter(stdout)) |
|
536 | 545 | except Exception as error: |
|
537 | 546 | status = 128 |
|
538 | 547 | stdout.write(f'ERROR: {error}\n') |
|
539 | 548 | |
|
540 | 549 | return HookResponse(status, stdout.getvalue()) |
|
541 | 550 | |
|
542 | 551 | |
|
543 | 552 | def _parse_git_ref_lines(revision_lines): |
|
544 | 553 | rev_data = [] |
|
545 | 554 | for revision_line in revision_lines or []: |
|
546 | 555 | old_rev, new_rev, ref = revision_line.strip().split(' ') |
|
547 | 556 | ref_data = ref.split('/', 2) |
|
548 | 557 | if ref_data[1] in ('tags', 'heads'): |
|
549 | 558 | rev_data.append({ |
|
550 | 559 | # NOTE(marcink): |
|
551 | 560 | # we're unable to tell total_commits for git at this point |
|
552 | 561 | # but we set the variable for consistency with GIT |
|
553 | 562 | 'total_commits': -1, |
|
554 | 563 | 'old_rev': old_rev, |
|
555 | 564 | 'new_rev': new_rev, |
|
556 | 565 | 'ref': ref, |
|
557 | 566 | 'type': ref_data[1], |
|
558 | 567 | 'name': ref_data[2], |
|
559 | 568 | }) |
|
560 | 569 | return rev_data |
|
561 | 570 | |
|
562 | 571 | |
|
563 | 572 | def git_pre_receive(unused_repo_path, revision_lines, env) -> int: |
|
564 | 573 | """ |
|
565 | 574 | Pre push hook. |
|
566 | 575 | |
|
567 | 576 | :return: status code of the hook. 0 for success. |
|
568 | 577 | """ |
|
569 | 578 | extras = json.loads(env['RC_SCM_DATA']) |
|
570 | 579 | rev_data = _parse_git_ref_lines(revision_lines) |
|
571 | 580 | if 'push' not in extras['hooks']: |
|
572 | 581 | return 0 |
|
573 | _fix_hooks_executables() | |
|
582 | _fix_hooks_executables(env.get('RC_INI_FILE')) | |
|
574 | 583 | |
|
575 | 584 | empty_commit_id = '0' * 40 |
|
576 | 585 | |
|
577 | 586 | detect_force_push = extras.get('detect_force_push') |
|
578 | 587 | |
|
579 | 588 | for push_ref in rev_data: |
|
580 | 589 | # store our git-env which holds the temp store |
|
581 | 590 | push_ref['git_env'] = _get_git_env() |
|
582 | 591 | push_ref['pruned_sha'] = '' |
|
583 | 592 | if not detect_force_push: |
|
584 | 593 | # don't check for forced-push when we don't need to |
|
585 | 594 | continue |
|
586 | 595 | |
|
587 | 596 | type_ = push_ref['type'] |
|
588 | 597 | new_branch = push_ref['old_rev'] == empty_commit_id |
|
589 | 598 | delete_branch = push_ref['new_rev'] == empty_commit_id |
|
590 | 599 | if type_ == 'heads' and not (new_branch or delete_branch): |
|
591 | 600 | old_rev = push_ref['old_rev'] |
|
592 | 601 | new_rev = push_ref['new_rev'] |
|
593 | 602 | cmd = [settings.GIT_EXECUTABLE(), 'rev-list', old_rev, f'^{new_rev}'] |
|
594 | 603 | stdout, stderr = subprocessio.run_command( |
|
595 | 604 | cmd, env=os.environ.copy()) |
|
596 | 605 | # means we're having some non-reachable objects, this forced push was used |
|
597 | 606 | if stdout: |
|
598 | 607 | push_ref['pruned_sha'] = stdout.splitlines() |
|
599 | 608 | |
|
600 | 609 | extras['hook_type'] = 'pre_receive' |
|
601 | 610 | extras['commit_ids'] = rev_data |
|
602 | 611 | |
|
603 | 612 | stdout = sys.stdout |
|
604 | 613 | status_code = _call_hook('pre_push', extras, GitMessageWriter(stdout)) |
|
605 | 614 | |
|
606 | 615 | return status_code |
|
607 | 616 | |
|
608 | 617 | |
|
609 | 618 | def git_post_receive(unused_repo_path, revision_lines, env) -> int: |
|
610 | 619 | """ |
|
611 | 620 | Post push hook. |
|
612 | 621 | |
|
613 | 622 | :return: status code of the hook. 0 for success. |
|
614 | 623 | """ |
|
615 | 624 | extras = json.loads(env['RC_SCM_DATA']) |
|
616 | 625 | if 'push' not in extras['hooks']: |
|
617 | 626 | return 0 |
|
618 | 627 | |
|
619 | _fix_hooks_executables() | |
|
628 | _fix_hooks_executables(env.get('RC_INI_FILE')) | |
|
620 | 629 | |
|
621 | 630 | rev_data = _parse_git_ref_lines(revision_lines) |
|
622 | 631 | |
|
623 | 632 | git_revs = [] |
|
624 | 633 | |
|
625 | 634 | # N.B.(skreft): it is ok to just call git, as git before calling a |
|
626 | 635 | # subcommand sets the PATH environment variable so that it point to the |
|
627 | 636 | # correct version of the git executable. |
|
628 | 637 | empty_commit_id = '0' * 40 |
|
629 | 638 | branches = [] |
|
630 | 639 | tags = [] |
|
631 | 640 | for push_ref in rev_data: |
|
632 | 641 | type_ = push_ref['type'] |
|
633 | 642 | |
|
634 | 643 | if type_ == 'heads': |
|
635 | 644 | # starting new branch case |
|
636 | 645 | if push_ref['old_rev'] == empty_commit_id: |
|
637 | 646 | push_ref_name = push_ref['name'] |
|
638 | 647 | |
|
639 | 648 | if push_ref_name not in branches: |
|
640 | 649 | branches.append(push_ref_name) |
|
641 | 650 | |
|
642 | 651 | need_head_set = '' |
|
643 | 652 | with Repository(os.getcwd()) as repo: |
|
644 | 653 | try: |
|
645 | 654 | repo.head |
|
646 | 655 | except pygit2.GitError: |
|
647 | 656 | need_head_set = f'refs/heads/{push_ref_name}' |
|
648 | 657 | |
|
649 | 658 | if need_head_set: |
|
650 | 659 | repo.set_head(need_head_set) |
|
651 | 660 | print(f"Setting default branch to {push_ref_name}") |
|
652 | 661 | |
|
653 | 662 | cmd = [settings.GIT_EXECUTABLE(), 'for-each-ref', '--format=%(refname)', 'refs/heads/*'] |
|
654 | 663 | stdout, stderr = subprocessio.run_command( |
|
655 | 664 | cmd, env=os.environ.copy()) |
|
656 | 665 | heads = safe_str(stdout) |
|
657 | 666 | heads = heads.replace(push_ref['ref'], '') |
|
658 | 667 | heads = ' '.join(head for head |
|
659 | 668 | in heads.splitlines() if head) or '.' |
|
660 | 669 | cmd = [settings.GIT_EXECUTABLE(), 'log', '--reverse', |
|
661 | 670 | '--pretty=format:%H', '--', push_ref['new_rev'], |
|
662 | 671 | '--not', heads] |
|
663 | 672 | stdout, stderr = subprocessio.run_command( |
|
664 | 673 | cmd, env=os.environ.copy()) |
|
665 | 674 | git_revs.extend(list(map(ascii_str, stdout.splitlines()))) |
|
666 | 675 | |
|
667 | 676 | # delete branch case |
|
668 | 677 | elif push_ref['new_rev'] == empty_commit_id: |
|
669 | 678 | git_revs.append(f'delete_branch=>{push_ref["name"]}') |
|
670 | 679 | else: |
|
671 | 680 | if push_ref['name'] not in branches: |
|
672 | 681 | branches.append(push_ref['name']) |
|
673 | 682 | |
|
674 | 683 | cmd = [settings.GIT_EXECUTABLE(), 'log', |
|
675 | 684 | f'{push_ref["old_rev"]}..{push_ref["new_rev"]}', |
|
676 | 685 | '--reverse', '--pretty=format:%H'] |
|
677 | 686 | stdout, stderr = subprocessio.run_command( |
|
678 | 687 | cmd, env=os.environ.copy()) |
|
679 | 688 | # we get bytes from stdout, we need str to be consistent |
|
680 | 689 | log_revs = list(map(ascii_str, stdout.splitlines())) |
|
681 | 690 | git_revs.extend(log_revs) |
|
682 | 691 | |
|
683 | 692 | # Pure pygit2 impl. but still 2-3x slower :/ |
|
684 | 693 | # results = [] |
|
685 | 694 | # |
|
686 | 695 | # with Repository(os.getcwd()) as repo: |
|
687 | 696 | # repo_new_rev = repo[push_ref['new_rev']] |
|
688 | 697 | # repo_old_rev = repo[push_ref['old_rev']] |
|
689 | 698 | # walker = repo.walk(repo_new_rev.id, pygit2.GIT_SORT_TOPOLOGICAL) |
|
690 | 699 | # |
|
691 | 700 | # for commit in walker: |
|
692 | 701 | # if commit.id == repo_old_rev.id: |
|
693 | 702 | # break |
|
694 | 703 | # results.append(commit.id.hex) |
|
695 | 704 | # # reverse the order, can't use GIT_SORT_REVERSE |
|
696 | 705 | # log_revs = results[::-1] |
|
697 | 706 | |
|
698 | 707 | elif type_ == 'tags': |
|
699 | 708 | if push_ref['name'] not in tags: |
|
700 | 709 | tags.append(push_ref['name']) |
|
701 | 710 | git_revs.append(f'tag=>{push_ref["name"]}') |
|
702 | 711 | |
|
703 | 712 | extras['hook_type'] = 'post_receive' |
|
704 | 713 | extras['commit_ids'] = git_revs |
|
705 | 714 | extras['new_refs'] = { |
|
706 | 715 | 'branches': branches, |
|
707 | 716 | 'bookmarks': [], |
|
708 | 717 | 'tags': tags, |
|
709 | 718 | } |
|
710 | 719 | |
|
711 | 720 | stdout = sys.stdout |
|
712 | 721 | |
|
713 | 722 | if 'repo_size' in extras['hooks']: |
|
714 | 723 | try: |
|
715 | 724 | _call_hook('repo_size', extras, GitMessageWriter(stdout)) |
|
716 | 725 | except Exception: |
|
717 | 726 | pass |
|
718 | 727 | |
|
719 | 728 | status_code = _call_hook('post_push', extras, GitMessageWriter(stdout)) |
|
720 | 729 | return status_code |
|
721 | 730 | |
|
722 | 731 | |
|
723 |
def |
|
|
724 | _fix_hooks_executables() | |
|
725 | ||
|
726 | extras = {} | |
|
727 | try: | |
|
728 | cmd = [settings.SVNLOOK_EXECUTABLE(), 'pget', | |
|
729 | '-t', txn_id, | |
|
730 | '--revprop', path, 'rc-scm-extras'] | |
|
731 | stdout, stderr = subprocessio.run_command( | |
|
732 | cmd, env=os.environ.copy()) | |
|
733 | extras = json.loads(base64.urlsafe_b64decode(stdout)) | |
|
734 | except Exception: | |
|
735 | log.exception('Failed to extract extras info from txn_id') | |
|
736 | ||
|
737 | return extras | |
|
738 | ||
|
739 | ||
|
740 | def _get_extras_from_commit_id(commit_id, path): | |
|
741 | _fix_hooks_executables() | |
|
742 | ||
|
743 | extras = {} | |
|
744 | try: | |
|
745 | cmd = [settings.SVNLOOK_EXECUTABLE(), 'pget', | |
|
746 | '-r', commit_id, | |
|
747 | '--revprop', path, 'rc-scm-extras'] | |
|
748 | stdout, stderr = subprocessio.run_command( | |
|
749 | cmd, env=os.environ.copy()) | |
|
750 | extras = json.loads(base64.urlsafe_b64decode(stdout)) | |
|
751 | except Exception: | |
|
752 | log.exception('Failed to extract extras info from commit_id') | |
|
753 | ||
|
732 | def get_extras_from_txn_id(repo_path, txn_id): | |
|
733 | extras = get_txn_id_from_store(repo_path, txn_id) | |
|
754 | 734 | return extras |
|
755 | 735 | |
|
756 | 736 | |
|
757 | 737 | def svn_pre_commit(repo_path, commit_data, env): |
|
758 | 738 | |
|
759 | 739 | path, txn_id = commit_data |
|
760 | 740 | branches = [] |
|
761 | 741 | tags = [] |
|
762 | 742 | |
|
763 | 743 | if env.get('RC_SCM_DATA'): |
|
764 | 744 | extras = json.loads(env['RC_SCM_DATA']) |
|
765 | 745 | else: |
|
746 | ini_path = env.get('RC_INI_FILE') | |
|
747 | if ini_path: | |
|
748 | _get_ini_settings(ini_path) | |
|
766 | 749 | # fallback method to read from TXN-ID stored data |
|
767 |
extras = |
|
|
768 | if not extras: | |
|
769 | return 0 | |
|
750 | extras = get_extras_from_txn_id(path, txn_id) | |
|
751 | ||
|
752 | if not extras: | |
|
753 | raise ValueError('SVN-PRE-COMMIT: Failed to extract context data in called extras for hook execution') | |
|
754 | ||
|
755 | if extras.get('rc_internal_commit'): | |
|
756 | # special marker for internal commit, we don't call hooks client | |
|
757 | return 0 | |
|
770 | 758 | |
|
771 | 759 | extras['hook_type'] = 'pre_commit' |
|
772 | 760 | extras['commit_ids'] = [txn_id] |
|
773 | 761 | extras['txn_id'] = txn_id |
|
774 | 762 | extras['new_refs'] = { |
|
775 | 763 | 'total_commits': 1, |
|
776 | 764 | 'branches': branches, |
|
777 | 765 | 'bookmarks': [], |
|
778 | 766 | 'tags': tags, |
|
779 | 767 | } |
|
780 | 768 | |
|
781 | 769 | return _call_hook('pre_push', extras, SvnMessageWriter()) |
|
782 | 770 | |
|
783 | 771 | |
|
784 | 772 | def svn_post_commit(repo_path, commit_data, env): |
|
785 | 773 | """ |
|
786 | 774 | commit_data is path, rev, txn_id |
|
787 | 775 | """ |
|
788 | 776 | |
|
789 | 777 | if len(commit_data) == 3: |
|
790 | 778 | path, commit_id, txn_id = commit_data |
|
791 | 779 | elif len(commit_data) == 2: |
|
792 | 780 | log.error('Failed to extract txn_id from commit_data using legacy method. ' |
|
793 | 781 | 'Some functionality might be limited') |
|
794 | 782 | path, commit_id = commit_data |
|
795 | 783 | txn_id = None |
|
796 | 784 | else: |
|
797 | 785 | return 0 |
|
798 | 786 | |
|
799 | 787 | branches = [] |
|
800 | 788 | tags = [] |
|
801 | 789 | |
|
802 | 790 | if env.get('RC_SCM_DATA'): |
|
803 | 791 | extras = json.loads(env['RC_SCM_DATA']) |
|
804 | 792 | else: |
|
793 | ini_path = env.get('RC_INI_FILE') | |
|
794 | if ini_path: | |
|
795 | _get_ini_settings(ini_path) | |
|
805 | 796 | # fallback method to read from TXN-ID stored data |
|
806 |
extras = |
|
|
807 | if not extras: | |
|
808 | return 0 | |
|
797 | extras = get_extras_from_txn_id(path, txn_id) | |
|
798 | ||
|
799 | if not extras and txn_id: | |
|
800 | raise ValueError('SVN-POST-COMMIT: Failed to extract context data in called extras for hook execution') | |
|
801 | ||
|
802 | if extras.get('rc_internal_commit'): | |
|
803 | # special marker for internal commit, we don't call hooks client | |
|
804 | return 0 | |
|
809 | 805 | |
|
810 | 806 | extras['hook_type'] = 'post_commit' |
|
811 | 807 | extras['commit_ids'] = [commit_id] |
|
812 | 808 | extras['txn_id'] = txn_id |
|
813 | 809 | extras['new_refs'] = { |
|
814 | 810 | 'branches': branches, |
|
815 | 811 | 'bookmarks': [], |
|
816 | 812 | 'tags': tags, |
|
817 | 813 | 'total_commits': 1, |
|
818 | 814 | } |
|
819 | 815 | |
|
820 | 816 | if 'repo_size' in extras['hooks']: |
|
821 | 817 | try: |
|
822 | 818 | _call_hook('repo_size', extras, SvnMessageWriter()) |
|
823 | 819 | except Exception: |
|
824 | 820 | pass |
|
825 | 821 | |
|
826 | 822 | return _call_hook('post_push', extras, SvnMessageWriter()) |
@@ -1,774 +1,763 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import io |
|
19 | 19 | import os |
|
20 | 20 | import platform |
|
21 | 21 | import sys |
|
22 | 22 | import locale |
|
23 | 23 | import logging |
|
24 | 24 | import uuid |
|
25 | 25 | import time |
|
26 | 26 | import wsgiref.util |
|
27 | 27 | import tempfile |
|
28 | 28 | import psutil |
|
29 | 29 | |
|
30 | 30 | from itertools import chain |
|
31 | 31 | |
|
32 | 32 | import msgpack |
|
33 | 33 | import configparser |
|
34 | 34 | |
|
35 | 35 | from pyramid.config import Configurator |
|
36 | 36 | from pyramid.wsgi import wsgiapp |
|
37 | 37 | from pyramid.response import Response |
|
38 | 38 | |
|
39 | 39 | from vcsserver.base import BytesEnvelope, BinaryEnvelope |
|
40 | from vcsserver.lib.rc_json import json | |
|
40 | ||
|
41 | 41 | from vcsserver.config.settings_maker import SettingsMaker |
|
42 | from vcsserver.str_utils import safe_int | |
|
43 | from vcsserver.lib.statsd_client import StatsdClient | |
|
42 | ||
|
44 | 43 | from vcsserver.tweens.request_wrapper import get_headers_call_context |
|
45 | 44 | |
|
46 | import vcsserver | |
|
47 | from vcsserver import remote_wsgi, scm_app, settings, hgpatches | |
|
45 | from vcsserver import remote_wsgi, scm_app, hgpatches | |
|
46 | from vcsserver.server import VcsServer | |
|
48 | 47 | from vcsserver.git_lfs.app import GIT_LFS_CONTENT_TYPE, GIT_LFS_PROTO_PAT |
|
49 | 48 | from vcsserver.echo_stub import remote_wsgi as remote_wsgi_stub |
|
50 | 49 | from vcsserver.echo_stub.echo_app import EchoApp |
|
51 | 50 | from vcsserver.exceptions import HTTPRepoLocked, HTTPRepoBranchProtected |
|
52 | 51 | from vcsserver.lib.exc_tracking import store_exception, format_exc |
|
53 |
from vcsserver. |
|
|
52 | from vcsserver.lib.str_utils import safe_int | |
|
53 | from vcsserver.lib.statsd_client import StatsdClient | |
|
54 | from vcsserver.lib.ext_json import json | |
|
55 | from vcsserver.lib.config_utils import configure_and_store_settings | |
|
56 | ||
|
54 | 57 | |
|
55 | 58 | strict_vcs = True |
|
56 | 59 | |
|
57 | 60 | git_import_err = None |
|
58 | 61 | try: |
|
59 | 62 | from vcsserver.remote.git_remote import GitFactory, GitRemote |
|
60 | 63 | except ImportError as e: |
|
61 | 64 | GitFactory = None |
|
62 | 65 | GitRemote = None |
|
63 | 66 | git_import_err = e |
|
64 | 67 | if strict_vcs: |
|
65 | 68 | raise |
|
66 | 69 | |
|
67 | 70 | |
|
68 | 71 | hg_import_err = None |
|
69 | 72 | try: |
|
70 | 73 | from vcsserver.remote.hg_remote import MercurialFactory, HgRemote |
|
71 | 74 | except ImportError as e: |
|
72 | 75 | MercurialFactory = None |
|
73 | 76 | HgRemote = None |
|
74 | 77 | hg_import_err = e |
|
75 | 78 | if strict_vcs: |
|
76 | 79 | raise |
|
77 | 80 | |
|
78 | 81 | |
|
79 | 82 | svn_import_err = None |
|
80 | 83 | try: |
|
81 | 84 | from vcsserver.remote.svn_remote import SubversionFactory, SvnRemote |
|
82 | 85 | except ImportError as e: |
|
83 | 86 | SubversionFactory = None |
|
84 | 87 | SvnRemote = None |
|
85 | 88 | svn_import_err = e |
|
86 | 89 | if strict_vcs: |
|
87 | 90 | raise |
|
88 | 91 | |
|
89 | 92 | log = logging.getLogger(__name__) |
|
90 | 93 | |
|
91 | 94 | # due to Mercurial/glibc2.27 problems we need to detect if locale settings are |
|
92 | 95 | # causing problems and "fix" it in case they do and fallback to LC_ALL = C |
|
93 | 96 | |
|
94 | 97 | try: |
|
95 | 98 | locale.setlocale(locale.LC_ALL, '') |
|
96 | 99 | except locale.Error as e: |
|
97 | log.error( | |
|
98 | 'LOCALE ERROR: failed to set LC_ALL, fallback to LC_ALL=C, org error: %s', e) | |
|
100 | log.error('LOCALE ERROR: failed to set LC_ALL, fallback to LC_ALL=C, org error: %s', e) | |
|
99 | 101 | os.environ['LC_ALL'] = 'C' |
|
100 | 102 | |
|
101 | 103 | |
|
102 | 104 | def _is_request_chunked(environ): |
|
103 | 105 | stream = environ.get('HTTP_TRANSFER_ENCODING', '') == 'chunked' |
|
104 | 106 | return stream |
|
105 | 107 | |
|
106 | 108 | |
|
107 | 109 | def log_max_fd(): |
|
108 | 110 | try: |
|
109 | 111 | maxfd = psutil.Process().rlimit(psutil.RLIMIT_NOFILE)[1] |
|
110 | 112 | log.info('Max file descriptors value: %s', maxfd) |
|
111 | 113 | except Exception: |
|
112 | 114 | pass |
|
113 | 115 | |
|
114 | 116 | |
|
115 | 117 | class VCS: |
|
116 | 118 | def __init__(self, locale_conf=None, cache_config=None): |
|
117 | 119 | self.locale = locale_conf |
|
118 | 120 | self.cache_config = cache_config |
|
119 | 121 | self._configure_locale() |
|
120 | 122 | |
|
121 | 123 | log_max_fd() |
|
122 | 124 | |
|
123 | 125 | if GitFactory and GitRemote: |
|
124 | 126 | git_factory = GitFactory() |
|
125 | 127 | self._git_remote = GitRemote(git_factory) |
|
126 | 128 | else: |
|
127 | 129 | log.error("Git client import failed: %s", git_import_err) |
|
128 | 130 | |
|
129 | 131 | if MercurialFactory and HgRemote: |
|
130 | 132 | hg_factory = MercurialFactory() |
|
131 | 133 | self._hg_remote = HgRemote(hg_factory) |
|
132 | 134 | else: |
|
133 | 135 | log.error("Mercurial client import failed: %s", hg_import_err) |
|
134 | 136 | |
|
135 | 137 | if SubversionFactory and SvnRemote: |
|
136 | 138 | svn_factory = SubversionFactory() |
|
137 | 139 | |
|
138 | 140 | # hg factory is used for svn url validation |
|
139 | 141 | hg_factory = MercurialFactory() |
|
140 | 142 | self._svn_remote = SvnRemote(svn_factory, hg_factory=hg_factory) |
|
141 | 143 | else: |
|
142 | 144 | log.error("Subversion client import failed: %s", svn_import_err) |
|
143 | 145 | |
|
144 | 146 | self._vcsserver = VcsServer() |
|
145 | 147 | |
|
146 | 148 | def _configure_locale(self): |
|
147 | 149 | if self.locale: |
|
148 | 150 | log.info('Settings locale: `LC_ALL` to %s', self.locale) |
|
149 | 151 | else: |
|
150 | 152 | log.info('Configuring locale subsystem based on environment variables') |
|
151 | 153 | try: |
|
152 | 154 | # If self.locale is the empty string, then the locale |
|
153 | 155 | # module will use the environment variables. See the |
|
154 | 156 | # documentation of the package `locale`. |
|
155 | 157 | locale.setlocale(locale.LC_ALL, self.locale) |
|
156 | 158 | |
|
157 | 159 | language_code, encoding = locale.getlocale() |
|
158 | 160 | log.info( |
|
159 | 161 | 'Locale set to language code "%s" with encoding "%s".', |
|
160 | 162 | language_code, encoding) |
|
161 | 163 | except locale.Error: |
|
162 | 164 | log.exception('Cannot set locale, not configuring the locale system') |
|
163 | 165 | |
|
164 | 166 | |
|
165 | 167 | class WsgiProxy: |
|
166 | 168 | def __init__(self, wsgi): |
|
167 | 169 | self.wsgi = wsgi |
|
168 | 170 | |
|
169 | 171 | def __call__(self, environ, start_response): |
|
170 | 172 | input_data = environ['wsgi.input'].read() |
|
171 | 173 | input_data = msgpack.unpackb(input_data) |
|
172 | 174 | |
|
173 | 175 | error = None |
|
174 | 176 | try: |
|
175 | 177 | data, status, headers = self.wsgi.handle( |
|
176 | 178 | input_data['environment'], input_data['input_data'], |
|
177 | 179 | *input_data['args'], **input_data['kwargs']) |
|
178 | 180 | except Exception as e: |
|
179 | 181 | data, status, headers = [], None, None |
|
180 | 182 | error = { |
|
181 | 183 | 'message': str(e), |
|
182 | 184 | '_vcs_kind': getattr(e, '_vcs_kind', None) |
|
183 | 185 | } |
|
184 | 186 | |
|
185 | 187 | start_response(200, {}) |
|
186 | 188 | return self._iterator(error, status, headers, data) |
|
187 | 189 | |
|
188 | 190 | def _iterator(self, error, status, headers, data): |
|
189 | 191 | initial_data = [ |
|
190 | 192 | error, |
|
191 | 193 | status, |
|
192 | 194 | headers, |
|
193 | 195 | ] |
|
194 | 196 | |
|
195 | 197 | for d in chain(initial_data, data): |
|
196 | 198 | yield msgpack.packb(d) |
|
197 | 199 | |
|
198 | 200 | |
|
199 | 201 | def not_found(request): |
|
200 | 202 | return {'status': '404 NOT FOUND'} |
|
201 | 203 | |
|
202 | 204 | |
|
203 | 205 | class VCSViewPredicate: |
|
204 | 206 | def __init__(self, val, config): |
|
205 | 207 | self.remotes = val |
|
206 | 208 | |
|
207 | 209 | def text(self): |
|
208 | 210 | return f'vcs view method = {list(self.remotes.keys())}' |
|
209 | 211 | |
|
210 | 212 | phash = text |
|
211 | 213 | |
|
212 | 214 | def __call__(self, context, request): |
|
213 | 215 | """ |
|
214 | 216 | View predicate that returns true if given backend is supported by |
|
215 | 217 | defined remotes. |
|
216 | 218 | """ |
|
217 | 219 | backend = request.matchdict.get('backend') |
|
218 | 220 | return backend in self.remotes |
|
219 | 221 | |
|
220 | 222 | |
|
221 | 223 | class HTTPApplication: |
|
222 | 224 | ALLOWED_EXCEPTIONS = ('KeyError', 'URLError') |
|
223 | 225 | |
|
224 | 226 | remote_wsgi = remote_wsgi |
|
225 | 227 | _use_echo_app = False |
|
226 | 228 | |
|
227 | 229 | def __init__(self, settings=None, global_config=None): |
|
228 | 230 | |
|
229 | 231 | self.config = Configurator(settings=settings) |
|
230 | 232 | # Init our statsd at very start |
|
231 | 233 | self.config.registry.statsd = StatsdClient.statsd |
|
232 | 234 | self.config.registry.vcs_call_context = {} |
|
233 | 235 | |
|
234 | 236 | self.global_config = global_config |
|
235 | 237 | self.config.include('vcsserver.lib.rc_cache') |
|
236 |
self.config.include('vcsserver.lib. |
|
|
238 | self.config.include('vcsserver.lib.archive_cache') | |
|
237 | 239 | |
|
238 | 240 | settings_locale = settings.get('locale', '') or 'en_US.UTF-8' |
|
239 | 241 | vcs = VCS(locale_conf=settings_locale, cache_config=settings) |
|
240 | 242 | self._remotes = { |
|
241 | 243 | 'hg': vcs._hg_remote, |
|
242 | 244 | 'git': vcs._git_remote, |
|
243 | 245 | 'svn': vcs._svn_remote, |
|
244 | 246 | 'server': vcs._vcsserver, |
|
245 | 247 | } |
|
246 | 248 | if settings.get('dev.use_echo_app', 'false').lower() == 'true': |
|
247 | 249 | self._use_echo_app = True |
|
248 | 250 | log.warning("Using EchoApp for VCS operations.") |
|
249 | 251 | self.remote_wsgi = remote_wsgi_stub |
|
250 | 252 | |
|
251 |
|
|
|
253 | configure_and_store_settings(global_config, settings) | |
|
252 | 254 | |
|
253 | 255 | self._configure() |
|
254 | 256 | |
|
255 | def _configure_settings(self, global_config, app_settings): | |
|
256 | """ | |
|
257 | Configure the settings module. | |
|
258 | """ | |
|
259 | settings_merged = global_config.copy() | |
|
260 | settings_merged.update(app_settings) | |
|
261 | ||
|
262 | binary_dir = app_settings['core.binary_dir'] | |
|
263 | ||
|
264 | settings.BINARY_DIR = binary_dir | |
|
265 | ||
|
266 | # Store the settings to make them available to other modules. | |
|
267 | vcsserver.PYRAMID_SETTINGS = settings_merged | |
|
268 | vcsserver.CONFIG = settings_merged | |
|
269 | ||
|
270 | 257 | def _configure(self): |
|
271 | 258 | self.config.add_renderer(name='msgpack', factory=self._msgpack_renderer_factory) |
|
272 | 259 | |
|
273 | 260 | self.config.add_route('service', '/_service') |
|
274 | 261 | self.config.add_route('status', '/status') |
|
275 | 262 | self.config.add_route('hg_proxy', '/proxy/hg') |
|
276 | 263 | self.config.add_route('git_proxy', '/proxy/git') |
|
277 | 264 | |
|
278 | 265 | # rpc methods |
|
279 | 266 | self.config.add_route('vcs', '/{backend}') |
|
280 | 267 | |
|
281 | 268 | # streaming rpc remote methods |
|
282 | 269 | self.config.add_route('vcs_stream', '/{backend}/stream') |
|
283 | 270 | |
|
284 | 271 | # vcs operations clone/push as streaming |
|
285 | 272 | self.config.add_route('stream_git', '/stream/git/*repo_name') |
|
286 | 273 | self.config.add_route('stream_hg', '/stream/hg/*repo_name') |
|
287 | 274 | |
|
288 | 275 | self.config.add_view(self.status_view, route_name='status', renderer='json') |
|
289 | 276 | self.config.add_view(self.service_view, route_name='service', renderer='msgpack') |
|
290 | 277 | |
|
291 | 278 | self.config.add_view(self.hg_proxy(), route_name='hg_proxy') |
|
292 | 279 | self.config.add_view(self.git_proxy(), route_name='git_proxy') |
|
293 | 280 | self.config.add_view(self.vcs_view, route_name='vcs', renderer='msgpack', |
|
294 | 281 | vcs_view=self._remotes) |
|
295 | 282 | self.config.add_view(self.vcs_stream_view, route_name='vcs_stream', |
|
296 | 283 | vcs_view=self._remotes) |
|
297 | 284 | |
|
298 | 285 | self.config.add_view(self.hg_stream(), route_name='stream_hg') |
|
299 | 286 | self.config.add_view(self.git_stream(), route_name='stream_git') |
|
300 | 287 | |
|
301 | 288 | self.config.add_view_predicate('vcs_view', VCSViewPredicate) |
|
302 | 289 | |
|
303 | 290 | self.config.add_notfound_view(not_found, renderer='json') |
|
304 | 291 | |
|
305 | 292 | self.config.add_view(self.handle_vcs_exception, context=Exception) |
|
306 | 293 | |
|
307 | 294 | self.config.add_tween( |
|
308 | 295 | 'vcsserver.tweens.request_wrapper.RequestWrapperTween', |
|
309 | 296 | ) |
|
310 | 297 | self.config.add_request_method( |
|
311 | 298 | 'vcsserver.lib.request_counter.get_request_counter', |
|
312 | 299 | 'request_count') |
|
313 | 300 | |
|
314 | 301 | def wsgi_app(self): |
|
315 | 302 | return self.config.make_wsgi_app() |
|
316 | 303 | |
|
317 | 304 | def _vcs_view_params(self, request): |
|
318 | 305 | remote = self._remotes[request.matchdict['backend']] |
|
319 | 306 | payload = msgpack.unpackb(request.body, use_list=True) |
|
320 | 307 | |
|
321 | 308 | method = payload.get('method') |
|
322 | 309 | params = payload['params'] |
|
323 | 310 | wire = params.get('wire') |
|
324 | 311 | args = params.get('args') |
|
325 | 312 | kwargs = params.get('kwargs') |
|
326 | 313 | context_uid = None |
|
327 | 314 | |
|
328 | 315 | request.registry.vcs_call_context = { |
|
329 | 316 | 'method': method, |
|
330 | 317 | 'repo_name': payload.get('_repo_name'), |
|
331 | 318 | } |
|
332 | 319 | |
|
333 | 320 | if wire: |
|
334 | 321 | try: |
|
335 | 322 | wire['context'] = context_uid = uuid.UUID(wire['context']) |
|
336 | 323 | except KeyError: |
|
337 | 324 | pass |
|
338 | 325 | args.insert(0, wire) |
|
339 | 326 | repo_state_uid = wire.get('repo_state_uid') if wire else None |
|
340 | 327 | |
|
341 | 328 | # NOTE(marcink): trading complexity for slight performance |
|
342 | 329 | if log.isEnabledFor(logging.DEBUG): |
|
343 | 330 | # also we SKIP printing out any of those methods args since they maybe excessive |
|
344 | 331 | just_args_methods = { |
|
345 | 332 | 'commitctx': ('content', 'removed', 'updated'), |
|
346 | 333 | 'commit': ('content', 'removed', 'updated') |
|
347 | 334 | } |
|
348 | 335 | if method in just_args_methods: |
|
349 | 336 | skip_args = just_args_methods[method] |
|
350 | 337 | call_args = '' |
|
351 | 338 | call_kwargs = {} |
|
352 | 339 | for k in kwargs: |
|
353 | 340 | if k in skip_args: |
|
354 | 341 | # replace our skip key with dummy |
|
355 | 342 | call_kwargs[k] = f'RemovedParam({k})' |
|
356 | 343 | else: |
|
357 | 344 | call_kwargs[k] = kwargs[k] |
|
358 | 345 | else: |
|
359 | 346 | call_args = args[1:] |
|
360 | 347 | call_kwargs = kwargs |
|
361 | 348 | |
|
362 | 349 | log.debug('Method requested:`%s` with args:%s kwargs:%s context_uid: %s, repo_state_uid:%s', |
|
363 | 350 | method, call_args, call_kwargs, context_uid, repo_state_uid) |
|
364 | 351 | |
|
365 | 352 | statsd = request.registry.statsd |
|
366 | 353 | if statsd: |
|
367 | 354 | statsd.incr( |
|
368 | 355 | 'vcsserver_method_total', tags=[ |
|
369 | 356 | f"method:{method}", |
|
370 | 357 | ]) |
|
371 | 358 | return payload, remote, method, args, kwargs |
|
372 | 359 | |
|
373 | 360 | def vcs_view(self, request): |
|
374 | 361 | |
|
375 | 362 | payload, remote, method, args, kwargs = self._vcs_view_params(request) |
|
376 | 363 | payload_id = payload.get('id') |
|
377 | 364 | |
|
378 | 365 | try: |
|
379 | 366 | resp = getattr(remote, method)(*args, **kwargs) |
|
380 | 367 | except Exception as e: |
|
381 | 368 | exc_info = list(sys.exc_info()) |
|
382 | 369 | exc_type, exc_value, exc_traceback = exc_info |
|
383 | 370 | |
|
384 | 371 | org_exc = getattr(e, '_org_exc', None) |
|
385 | 372 | org_exc_name = None |
|
386 | 373 | org_exc_tb = '' |
|
387 | 374 | if org_exc: |
|
388 | 375 | org_exc_name = org_exc.__class__.__name__ |
|
389 | 376 | org_exc_tb = getattr(e, '_org_exc_tb', '') |
|
390 | 377 | # replace our "faked" exception with our org |
|
391 | 378 | exc_info[0] = org_exc.__class__ |
|
392 | 379 | exc_info[1] = org_exc |
|
393 | 380 | |
|
394 | 381 | should_store_exc = True |
|
395 | 382 | if org_exc: |
|
396 | 383 | def get_exc_fqn(_exc_obj): |
|
397 | 384 | module_name = getattr(org_exc.__class__, '__module__', 'UNKNOWN') |
|
398 | 385 | return module_name + '.' + org_exc_name |
|
399 | 386 | |
|
400 | 387 | exc_fqn = get_exc_fqn(org_exc) |
|
401 | 388 | |
|
402 | 389 | if exc_fqn in ['mercurial.error.RepoLookupError', |
|
403 | 390 | 'vcsserver.exceptions.RefNotFoundException']: |
|
404 | 391 | should_store_exc = False |
|
405 | 392 | |
|
406 | 393 | if should_store_exc: |
|
407 | 394 | store_exception(id(exc_info), exc_info, request_path=request.path) |
|
408 | 395 | |
|
409 | 396 | tb_info = format_exc(exc_info) |
|
410 | 397 | |
|
411 | 398 | type_ = e.__class__.__name__ |
|
412 | 399 | if type_ not in self.ALLOWED_EXCEPTIONS: |
|
413 | 400 | type_ = None |
|
414 | 401 | |
|
415 | 402 | resp = { |
|
416 | 403 | 'id': payload_id, |
|
417 | 404 | 'error': { |
|
418 | 405 | 'message': str(e), |
|
419 | 406 | 'traceback': tb_info, |
|
420 | 407 | 'org_exc': org_exc_name, |
|
421 | 408 | 'org_exc_tb': org_exc_tb, |
|
422 | 409 | 'type': type_ |
|
423 | 410 | } |
|
424 | 411 | } |
|
425 | 412 | |
|
426 | 413 | try: |
|
427 | 414 | resp['error']['_vcs_kind'] = getattr(e, '_vcs_kind', None) |
|
428 | 415 | except AttributeError: |
|
429 | 416 | pass |
|
430 | 417 | else: |
|
431 | 418 | resp = { |
|
432 | 419 | 'id': payload_id, |
|
433 | 420 | 'result': resp |
|
434 | 421 | } |
|
435 | 422 | log.debug('Serving data for method %s', method) |
|
436 | 423 | return resp |
|
437 | 424 | |
|
438 | 425 | def vcs_stream_view(self, request): |
|
439 | 426 | payload, remote, method, args, kwargs = self._vcs_view_params(request) |
|
440 | 427 | # this method has a stream: marker we remove it here |
|
441 | 428 | method = method.split('stream:')[-1] |
|
442 | 429 | chunk_size = safe_int(payload.get('chunk_size')) or 4096 |
|
443 | 430 | |
|
444 | 431 | resp = getattr(remote, method)(*args, **kwargs) |
|
445 | 432 | |
|
446 | 433 | def get_chunked_data(method_resp): |
|
447 | 434 | stream = io.BytesIO(method_resp) |
|
448 | 435 | while 1: |
|
449 | 436 | chunk = stream.read(chunk_size) |
|
450 | 437 | if not chunk: |
|
451 | 438 | break |
|
452 | 439 | yield chunk |
|
453 | 440 | |
|
454 | 441 | response = Response(app_iter=get_chunked_data(resp)) |
|
455 | 442 | response.content_type = 'application/octet-stream' |
|
456 | 443 | |
|
457 | 444 | return response |
|
458 | 445 | |
|
459 | 446 | def status_view(self, request): |
|
460 | 447 | import vcsserver |
|
461 | 448 | _platform_id = platform.uname()[1] or 'instance' |
|
462 | 449 | |
|
463 | 450 | return { |
|
464 | 451 | "status": "OK", |
|
465 | 452 | "vcsserver_version": vcsserver.get_version(), |
|
466 | 453 | "platform": _platform_id, |
|
467 | 454 | "pid": os.getpid(), |
|
468 | 455 | } |
|
469 | 456 | |
|
470 | 457 | def service_view(self, request): |
|
471 | 458 | import vcsserver |
|
472 | 459 | |
|
473 | 460 | payload = msgpack.unpackb(request.body, use_list=True) |
|
474 | 461 | server_config, app_config = {}, {} |
|
475 | 462 | |
|
476 | 463 | try: |
|
477 | 464 | path = self.global_config['__file__'] |
|
478 | 465 | config = configparser.RawConfigParser() |
|
479 | 466 | |
|
480 | 467 | config.read(path) |
|
481 | 468 | |
|
482 | 469 | if config.has_section('server:main'): |
|
483 | 470 | server_config = dict(config.items('server:main')) |
|
484 | 471 | if config.has_section('app:main'): |
|
485 | 472 | app_config = dict(config.items('app:main')) |
|
486 | 473 | |
|
487 | 474 | except Exception: |
|
488 | 475 | log.exception('Failed to read .ini file for display') |
|
489 | 476 | |
|
490 | 477 | environ = list(os.environ.items()) |
|
491 | 478 | |
|
492 | 479 | resp = { |
|
493 | 480 | 'id': payload.get('id'), |
|
494 | 481 | 'result': dict( |
|
495 | 482 | version=vcsserver.get_version(), |
|
496 | 483 | config=server_config, |
|
497 | 484 | app_config=app_config, |
|
498 | 485 | environ=environ, |
|
499 | 486 | payload=payload, |
|
500 | 487 | ) |
|
501 | 488 | } |
|
502 | 489 | return resp |
|
503 | 490 | |
|
504 | 491 | def _msgpack_renderer_factory(self, info): |
|
505 | 492 | |
|
506 | 493 | def _render(value, system): |
|
507 | 494 | bin_type = False |
|
508 | 495 | res = value.get('result') |
|
509 | 496 | if isinstance(res, BytesEnvelope): |
|
510 | 497 | log.debug('Result is wrapped in BytesEnvelope type') |
|
511 | 498 | bin_type = True |
|
512 | 499 | elif isinstance(res, BinaryEnvelope): |
|
513 | 500 | log.debug('Result is wrapped in BinaryEnvelope type') |
|
514 | 501 | value['result'] = res.val |
|
515 | 502 | bin_type = True |
|
516 | 503 | |
|
517 | 504 | request = system.get('request') |
|
518 | 505 | if request is not None: |
|
519 | 506 | response = request.response |
|
520 | 507 | ct = response.content_type |
|
521 | 508 | if ct == response.default_content_type: |
|
522 | 509 | response.content_type = 'application/x-msgpack' |
|
523 | 510 | if bin_type: |
|
524 | 511 | response.content_type = 'application/x-msgpack-bin' |
|
525 | 512 | |
|
526 | 513 | return msgpack.packb(value, use_bin_type=bin_type) |
|
527 | 514 | return _render |
|
528 | 515 | |
|
529 | 516 | def set_env_from_config(self, environ, config): |
|
530 | 517 | dict_conf = {} |
|
531 | 518 | try: |
|
532 | 519 | for elem in config: |
|
533 | 520 | if elem[0] == 'rhodecode': |
|
534 | 521 | dict_conf = json.loads(elem[2]) |
|
535 | 522 | break |
|
536 | 523 | except Exception: |
|
537 | 524 | log.exception('Failed to fetch SCM CONFIG') |
|
538 | 525 | return |
|
539 | 526 | |
|
540 | 527 | username = dict_conf.get('username') |
|
541 | 528 | if username: |
|
542 | 529 | environ['REMOTE_USER'] = username |
|
543 | 530 | # mercurial specific, some extension api rely on this |
|
544 | 531 | environ['HGUSER'] = username |
|
545 | 532 | |
|
546 | 533 | ip = dict_conf.get('ip') |
|
547 | 534 | if ip: |
|
548 | 535 | environ['REMOTE_HOST'] = ip |
|
549 | 536 | |
|
550 | 537 | if _is_request_chunked(environ): |
|
551 | 538 | # set the compatibility flag for webob |
|
552 | 539 | environ['wsgi.input_terminated'] = True |
|
553 | 540 | |
|
554 | 541 | def hg_proxy(self): |
|
555 | 542 | @wsgiapp |
|
556 | 543 | def _hg_proxy(environ, start_response): |
|
557 | 544 | app = WsgiProxy(self.remote_wsgi.HgRemoteWsgi()) |
|
558 | 545 | return app(environ, start_response) |
|
559 | 546 | return _hg_proxy |
|
560 | 547 | |
|
561 | 548 | def git_proxy(self): |
|
562 | 549 | @wsgiapp |
|
563 | 550 | def _git_proxy(environ, start_response): |
|
564 | 551 | app = WsgiProxy(self.remote_wsgi.GitRemoteWsgi()) |
|
565 | 552 | return app(environ, start_response) |
|
566 | 553 | return _git_proxy |
|
567 | 554 | |
|
568 | 555 | def hg_stream(self): |
|
569 | 556 | if self._use_echo_app: |
|
570 | 557 | @wsgiapp |
|
571 | 558 | def _hg_stream(environ, start_response): |
|
572 | 559 | app = EchoApp('fake_path', 'fake_name', None) |
|
573 | 560 | return app(environ, start_response) |
|
574 | 561 | return _hg_stream |
|
575 | 562 | else: |
|
576 | 563 | @wsgiapp |
|
577 | 564 | def _hg_stream(environ, start_response): |
|
578 | 565 | log.debug('http-app: handling hg stream') |
|
579 | 566 | call_context = get_headers_call_context(environ) |
|
580 | 567 | |
|
581 | 568 | repo_path = call_context['repo_path'] |
|
582 | 569 | repo_name = call_context['repo_name'] |
|
583 | 570 | config = call_context['repo_config'] |
|
584 | 571 | |
|
585 | 572 | app = scm_app.create_hg_wsgi_app( |
|
586 | 573 | repo_path, repo_name, config) |
|
587 | 574 | |
|
588 | 575 | # Consistent path information for hgweb |
|
589 | 576 | environ['PATH_INFO'] = call_context['path_info'] |
|
590 | 577 | environ['REPO_NAME'] = repo_name |
|
591 | 578 | self.set_env_from_config(environ, config) |
|
592 | 579 | |
|
593 | 580 | log.debug('http-app: starting app handler ' |
|
594 | 581 | 'with %s and process request', app) |
|
595 | 582 | return app(environ, ResponseFilter(start_response)) |
|
596 | 583 | return _hg_stream |
|
597 | 584 | |
|
598 | 585 | def git_stream(self): |
|
599 | 586 | if self._use_echo_app: |
|
600 | 587 | @wsgiapp |
|
601 | 588 | def _git_stream(environ, start_response): |
|
602 | 589 | app = EchoApp('fake_path', 'fake_name', None) |
|
603 | 590 | return app(environ, start_response) |
|
604 | 591 | return _git_stream |
|
605 | 592 | else: |
|
606 | 593 | @wsgiapp |
|
607 | 594 | def _git_stream(environ, start_response): |
|
608 | 595 | log.debug('http-app: handling git stream') |
|
609 | 596 | |
|
610 | 597 | call_context = get_headers_call_context(environ) |
|
611 | 598 | |
|
612 | 599 | repo_path = call_context['repo_path'] |
|
613 | 600 | repo_name = call_context['repo_name'] |
|
614 | 601 | config = call_context['repo_config'] |
|
615 | 602 | |
|
616 | 603 | environ['PATH_INFO'] = call_context['path_info'] |
|
617 | 604 | self.set_env_from_config(environ, config) |
|
618 | 605 | |
|
619 | 606 | content_type = environ.get('CONTENT_TYPE', '') |
|
620 | 607 | |
|
621 | 608 | path = environ['PATH_INFO'] |
|
622 | 609 | is_lfs_request = GIT_LFS_CONTENT_TYPE in content_type |
|
623 | 610 | log.debug( |
|
624 | 611 | 'LFS: Detecting if request `%s` is LFS server path based ' |
|
625 | 612 | 'on content type:`%s`, is_lfs:%s', |
|
626 | 613 | path, content_type, is_lfs_request) |
|
627 | 614 | |
|
628 | 615 | if not is_lfs_request: |
|
629 | 616 | # fallback detection by path |
|
630 | 617 | if GIT_LFS_PROTO_PAT.match(path): |
|
631 | 618 | is_lfs_request = True |
|
632 | 619 | log.debug( |
|
633 | 620 | 'LFS: fallback detection by path of: `%s`, is_lfs:%s', |
|
634 | 621 | path, is_lfs_request) |
|
635 | 622 | |
|
636 | 623 | if is_lfs_request: |
|
637 | 624 | app = scm_app.create_git_lfs_wsgi_app( |
|
638 | 625 | repo_path, repo_name, config) |
|
639 | 626 | else: |
|
640 | 627 | app = scm_app.create_git_wsgi_app( |
|
641 | 628 | repo_path, repo_name, config) |
|
642 | 629 | |
|
643 | 630 | log.debug('http-app: starting app handler ' |
|
644 | 631 | 'with %s and process request', app) |
|
645 | 632 | |
|
646 | 633 | return app(environ, start_response) |
|
647 | 634 | |
|
648 | 635 | return _git_stream |
|
649 | 636 | |
|
650 | 637 | def handle_vcs_exception(self, exception, request): |
|
651 | 638 | _vcs_kind = getattr(exception, '_vcs_kind', '') |
|
652 | 639 | |
|
653 | 640 | if _vcs_kind == 'repo_locked': |
|
654 | 641 | headers_call_context = get_headers_call_context(request.environ) |
|
655 | 642 | status_code = safe_int(headers_call_context['locked_status_code']) |
|
656 | 643 | |
|
657 | 644 | return HTTPRepoLocked( |
|
658 | 645 | title=str(exception), status_code=status_code, headers=[('X-Rc-Locked', '1')]) |
|
659 | 646 | |
|
660 | 647 | elif _vcs_kind == 'repo_branch_protected': |
|
661 | 648 | # Get custom repo-branch-protected status code if present. |
|
662 | 649 | return HTTPRepoBranchProtected( |
|
663 | 650 | title=str(exception), headers=[('X-Rc-Branch-Protection', '1')]) |
|
664 | 651 | |
|
665 | 652 | exc_info = request.exc_info |
|
666 | 653 | store_exception(id(exc_info), exc_info) |
|
667 | 654 | |
|
668 | 655 | traceback_info = 'unavailable' |
|
669 | 656 | if request.exc_info: |
|
670 | 657 | traceback_info = format_exc(request.exc_info) |
|
671 | 658 | |
|
672 | 659 | log.error( |
|
673 | 660 | 'error occurred handling this request for path: %s, \n%s', |
|
674 | 661 | request.path, traceback_info) |
|
675 | 662 | |
|
676 | 663 | statsd = request.registry.statsd |
|
677 | 664 | if statsd: |
|
678 | 665 | exc_type = f"{exception.__class__.__module__}.{exception.__class__.__name__}" |
|
679 | 666 | statsd.incr('vcsserver_exception_total', |
|
680 | 667 | tags=[f"type:{exc_type}"]) |
|
681 | 668 | raise exception |
|
682 | 669 | |
|
683 | 670 | |
|
684 | 671 | class ResponseFilter: |
|
685 | 672 | |
|
686 | 673 | def __init__(self, start_response): |
|
687 | 674 | self._start_response = start_response |
|
688 | 675 | |
|
689 | 676 | def __call__(self, status, response_headers, exc_info=None): |
|
690 | 677 | headers = tuple( |
|
691 | 678 | (h, v) for h, v in response_headers |
|
692 | 679 | if not wsgiref.util.is_hop_by_hop(h)) |
|
693 | 680 | return self._start_response(status, headers, exc_info) |
|
694 | 681 | |
|
695 | 682 | |
|
696 | 683 | def sanitize_settings_and_apply_defaults(global_config, settings): |
|
697 | 684 | _global_settings_maker = SettingsMaker(global_config) |
|
698 | 685 | settings_maker = SettingsMaker(settings) |
|
699 | 686 | |
|
700 | 687 | settings_maker.make_setting('logging.autoconfigure', False, parser='bool') |
|
701 | 688 | |
|
702 | 689 | logging_conf = os.path.join(os.path.dirname(global_config.get('__file__')), 'logging.ini') |
|
703 | 690 | settings_maker.enable_logging(logging_conf) |
|
704 | 691 | |
|
705 | 692 | # Default includes, possible to change as a user |
|
706 | 693 | pyramid_includes = settings_maker.make_setting('pyramid.includes', [], parser='list:newline') |
|
707 | 694 | log.debug("Using the following pyramid.includes: %s", pyramid_includes) |
|
708 | 695 | |
|
709 | 696 | settings_maker.make_setting('__file__', global_config.get('__file__')) |
|
710 | 697 | |
|
711 | 698 | settings_maker.make_setting('pyramid.default_locale_name', 'en') |
|
712 | 699 | settings_maker.make_setting('locale', 'en_US.UTF-8') |
|
713 | 700 | |
|
714 | 701 | settings_maker.make_setting( |
|
715 | 702 | 'core.binary_dir', '/usr/local/bin/rhodecode_bin/vcs_bin', |
|
716 | 703 | default_when_empty=True, parser='string:noquote') |
|
717 | 704 | |
|
705 | settings_maker.make_setting('vcs.svn.redis_conn', 'redis://redis:6379/0') | |
|
706 | ||
|
718 | 707 | temp_store = tempfile.gettempdir() |
|
719 | 708 | default_cache_dir = os.path.join(temp_store, 'rc_cache') |
|
720 | 709 | # save default, cache dir, and use it for all backends later. |
|
721 | 710 | default_cache_dir = settings_maker.make_setting( |
|
722 | 711 | 'cache_dir', |
|
723 | 712 | default=default_cache_dir, default_when_empty=True, |
|
724 | 713 | parser='dir:ensured') |
|
725 | 714 | |
|
726 | 715 | # exception store cache |
|
727 | 716 | settings_maker.make_setting( |
|
728 | 717 | 'exception_tracker.store_path', |
|
729 | 718 | default=os.path.join(default_cache_dir, 'exc_store'), default_when_empty=True, |
|
730 | 719 | parser='dir:ensured' |
|
731 | 720 | ) |
|
732 | 721 | |
|
733 | 722 | # repo_object cache defaults |
|
734 | 723 | settings_maker.make_setting( |
|
735 | 724 | 'rc_cache.repo_object.backend', |
|
736 | 725 | default='dogpile.cache.rc.file_namespace', |
|
737 | 726 | parser='string') |
|
738 | 727 | settings_maker.make_setting( |
|
739 | 728 | 'rc_cache.repo_object.expiration_time', |
|
740 | 729 | default=30 * 24 * 60 * 60, # 30days |
|
741 | 730 | parser='int') |
|
742 | 731 | settings_maker.make_setting( |
|
743 | 732 | 'rc_cache.repo_object.arguments.filename', |
|
744 | 733 | default=os.path.join(default_cache_dir, 'vcsserver_cache_repo_object.db'), |
|
745 | 734 | parser='string') |
|
746 | 735 | |
|
747 | 736 | # statsd |
|
748 | 737 | settings_maker.make_setting('statsd.enabled', False, parser='bool') |
|
749 | 738 | settings_maker.make_setting('statsd.statsd_host', 'statsd-exporter', parser='string') |
|
750 | 739 | settings_maker.make_setting('statsd.statsd_port', 9125, parser='int') |
|
751 | 740 | settings_maker.make_setting('statsd.statsd_prefix', '') |
|
752 | 741 | settings_maker.make_setting('statsd.statsd_ipv6', False, parser='bool') |
|
753 | 742 | |
|
754 | 743 | settings_maker.env_expand() |
|
755 | 744 | |
|
756 | 745 | |
|
757 | 746 | def main(global_config, **settings): |
|
758 | 747 | start_time = time.time() |
|
759 | 748 | log.info('Pyramid app config starting') |
|
760 | 749 | |
|
761 | 750 | if MercurialFactory: |
|
762 | 751 | hgpatches.patch_largefiles_capabilities() |
|
763 | 752 | hgpatches.patch_subrepo_type_mapping() |
|
764 | 753 | |
|
765 | 754 | # Fill in and sanitize the defaults & do ENV expansion |
|
766 | 755 | sanitize_settings_and_apply_defaults(global_config, settings) |
|
767 | 756 | |
|
768 | 757 | # init and bootstrap StatsdClient |
|
769 | 758 | StatsdClient.setup(settings) |
|
770 | 759 | |
|
771 | 760 | pyramid_app = HTTPApplication(settings=settings, global_config=global_config).wsgi_app() |
|
772 | 761 | total_time = time.time() - start_time |
|
773 | 762 | log.info('Pyramid app created and configured in %.2fs', total_time) |
|
774 | 763 | return pyramid_app |
@@ -1,243 +1,237 b'' | |||
|
1 | 1 | ''' |
|
2 | 2 | This library is provided to allow standard python logging |
|
3 | 3 | to output log data as JSON formatted strings |
|
4 | 4 | ''' |
|
5 | 5 | import logging |
|
6 | import json | |
|
7 | 6 | import re |
|
8 | 7 | from datetime import date, datetime, time, tzinfo, timedelta |
|
9 | 8 | import traceback |
|
10 | 9 | import importlib |
|
11 | 10 | |
|
12 | 11 | from inspect import istraceback |
|
13 | 12 | |
|
14 | 13 | from collections import OrderedDict |
|
15 | 14 | |
|
16 | ||
|
17 | def _inject_req_id(record, *args, **kwargs): | |
|
18 | return record | |
|
19 | ||
|
20 | ||
|
21 | ExceptionAwareFormatter = logging.Formatter | |
|
22 | ||
|
15 | from ...logging_formatter import _inject_req_id, ExceptionAwareFormatter | |
|
16 | from ...ext_json import sjson as json | |
|
23 | 17 | |
|
24 | 18 | ZERO = timedelta(0) |
|
25 | 19 | HOUR = timedelta(hours=1) |
|
26 | 20 | |
|
27 | 21 | |
|
28 | 22 | class UTC(tzinfo): |
|
29 | 23 | """UTC""" |
|
30 | 24 | |
|
31 | 25 | def utcoffset(self, dt): |
|
32 | 26 | return ZERO |
|
33 | 27 | |
|
34 | 28 | def tzname(self, dt): |
|
35 | 29 | return "UTC" |
|
36 | 30 | |
|
37 | 31 | def dst(self, dt): |
|
38 | 32 | return ZERO |
|
39 | 33 | |
|
40 | 34 | utc = UTC() |
|
41 | 35 | |
|
42 | 36 | |
|
43 | 37 | # skip natural LogRecord attributes |
|
44 | 38 | # http://docs.python.org/library/logging.html#logrecord-attributes |
|
45 | 39 | RESERVED_ATTRS = ( |
|
46 | 40 | 'args', 'asctime', 'created', 'exc_info', 'exc_text', 'filename', |
|
47 | 41 | 'funcName', 'levelname', 'levelno', 'lineno', 'module', |
|
48 | 42 | 'msecs', 'message', 'msg', 'name', 'pathname', 'process', |
|
49 | 43 | 'processName', 'relativeCreated', 'stack_info', 'thread', 'threadName') |
|
50 | 44 | |
|
51 | 45 | |
|
52 | 46 | def merge_record_extra(record, target, reserved): |
|
53 | 47 | """ |
|
54 | 48 | Merges extra attributes from LogRecord object into target dictionary |
|
55 | 49 | |
|
56 | 50 | :param record: logging.LogRecord |
|
57 | 51 | :param target: dict to update |
|
58 | 52 | :param reserved: dict or list with reserved keys to skip |
|
59 | 53 | """ |
|
60 | 54 | for key, value in record.__dict__.items(): |
|
61 | 55 | # this allows to have numeric keys |
|
62 | 56 | if (key not in reserved |
|
63 | 57 | and not (hasattr(key, "startswith") |
|
64 | 58 | and key.startswith('_'))): |
|
65 | 59 | target[key] = value |
|
66 | 60 | return target |
|
67 | 61 | |
|
68 | 62 | |
|
69 | 63 | class JsonEncoder(json.JSONEncoder): |
|
70 | 64 | """ |
|
71 | 65 | A custom encoder extending the default JSONEncoder |
|
72 | 66 | """ |
|
73 | 67 | |
|
74 | 68 | def default(self, obj): |
|
75 | 69 | if isinstance(obj, (date, datetime, time)): |
|
76 | 70 | return self.format_datetime_obj(obj) |
|
77 | 71 | |
|
78 | 72 | elif istraceback(obj): |
|
79 | 73 | return ''.join(traceback.format_tb(obj)).strip() |
|
80 | 74 | |
|
81 | 75 | elif type(obj) == Exception \ |
|
82 | 76 | or isinstance(obj, Exception) \ |
|
83 | 77 | or type(obj) == type: |
|
84 | 78 | return str(obj) |
|
85 | 79 | |
|
86 | 80 | try: |
|
87 | 81 | return super().default(obj) |
|
88 | 82 | |
|
89 | 83 | except TypeError: |
|
90 | 84 | try: |
|
91 | 85 | return str(obj) |
|
92 | 86 | |
|
93 | 87 | except Exception: |
|
94 | 88 | return None |
|
95 | 89 | |
|
96 | 90 | def format_datetime_obj(self, obj): |
|
97 | 91 | return obj.isoformat() |
|
98 | 92 | |
|
99 | 93 | |
|
100 | 94 | class JsonFormatter(ExceptionAwareFormatter): |
|
101 | 95 | """ |
|
102 | 96 | A custom formatter to format logging records as json strings. |
|
103 | 97 | Extra values will be formatted as str() if not supported by |
|
104 | 98 | json default encoder |
|
105 | 99 | """ |
|
106 | 100 | |
|
107 | 101 | def __init__(self, *args, **kwargs): |
|
108 | 102 | """ |
|
109 | 103 | :param json_default: a function for encoding non-standard objects |
|
110 | 104 | as outlined in http://docs.python.org/2/library/json.html |
|
111 | 105 | :param json_encoder: optional custom encoder |
|
112 | 106 | :param json_serializer: a :meth:`json.dumps`-compatible callable |
|
113 | 107 | that will be used to serialize the log record. |
|
114 | 108 | :param json_indent: an optional :meth:`json.dumps`-compatible numeric value |
|
115 | 109 | that will be used to customize the indent of the output json. |
|
116 | 110 | :param prefix: an optional string prefix added at the beginning of |
|
117 | 111 | the formatted string |
|
118 | 112 | :param json_indent: indent parameter for json.dumps |
|
119 | 113 | :param json_ensure_ascii: ensure_ascii parameter for json.dumps |
|
120 | 114 | :param reserved_attrs: an optional list of fields that will be skipped when |
|
121 | 115 | outputting json log record. Defaults to all log record attributes: |
|
122 | 116 | http://docs.python.org/library/logging.html#logrecord-attributes |
|
123 | 117 | :param timestamp: an optional string/boolean field to add a timestamp when |
|
124 | 118 | outputting the json log record. If string is passed, timestamp will be added |
|
125 | 119 | to log record using string as key. If True boolean is passed, timestamp key |
|
126 | 120 | will be "timestamp". Defaults to False/off. |
|
127 | 121 | """ |
|
128 | 122 | self.json_default = self._str_to_fn(kwargs.pop("json_default", None)) |
|
129 | 123 | self.json_encoder = self._str_to_fn(kwargs.pop("json_encoder", None)) |
|
130 | 124 | self.json_serializer = self._str_to_fn(kwargs.pop("json_serializer", json.dumps)) |
|
131 | 125 | self.json_indent = kwargs.pop("json_indent", None) |
|
132 | 126 | self.json_ensure_ascii = kwargs.pop("json_ensure_ascii", True) |
|
133 | 127 | self.prefix = kwargs.pop("prefix", "") |
|
134 | 128 | reserved_attrs = kwargs.pop("reserved_attrs", RESERVED_ATTRS) |
|
135 | 129 | self.reserved_attrs = dict(list(zip(reserved_attrs, reserved_attrs))) |
|
136 | 130 | self.timestamp = kwargs.pop("timestamp", True) |
|
137 | 131 | |
|
138 | 132 | # super(JsonFormatter, self).__init__(*args, **kwargs) |
|
139 | 133 | logging.Formatter.__init__(self, *args, **kwargs) |
|
140 | 134 | if not self.json_encoder and not self.json_default: |
|
141 | 135 | self.json_encoder = JsonEncoder |
|
142 | 136 | |
|
143 | 137 | self._required_fields = self.parse() |
|
144 | 138 | self._skip_fields = dict(list(zip(self._required_fields, |
|
145 | 139 | self._required_fields))) |
|
146 | 140 | self._skip_fields.update(self.reserved_attrs) |
|
147 | 141 | |
|
148 | 142 | def _str_to_fn(self, fn_as_str): |
|
149 | 143 | """ |
|
150 | 144 | If the argument is not a string, return whatever was passed in. |
|
151 | 145 | Parses a string such as package.module.function, imports the module |
|
152 | 146 | and returns the function. |
|
153 | 147 | |
|
154 | 148 | :param fn_as_str: The string to parse. If not a string, return it. |
|
155 | 149 | """ |
|
156 | 150 | if not isinstance(fn_as_str, str): |
|
157 | 151 | return fn_as_str |
|
158 | 152 | |
|
159 | 153 | path, _, function = fn_as_str.rpartition('.') |
|
160 | 154 | module = importlib.import_module(path) |
|
161 | 155 | return getattr(module, function) |
|
162 | 156 | |
|
163 | 157 | def parse(self): |
|
164 | 158 | """ |
|
165 | 159 | Parses format string looking for substitutions |
|
166 | 160 | |
|
167 | 161 | This method is responsible for returning a list of fields (as strings) |
|
168 | 162 | to include in all log messages. |
|
169 | 163 | """ |
|
170 | 164 | standard_formatters = re.compile(r'\((.+?)\)', re.IGNORECASE) |
|
171 | 165 | return standard_formatters.findall(self._fmt) |
|
172 | 166 | |
|
173 | 167 | def add_fields(self, log_record, record, message_dict): |
|
174 | 168 | """ |
|
175 | 169 | Override this method to implement custom logic for adding fields. |
|
176 | 170 | """ |
|
177 | 171 | for field in self._required_fields: |
|
178 | 172 | log_record[field] = record.__dict__.get(field) |
|
179 | 173 | log_record.update(message_dict) |
|
180 | 174 | merge_record_extra(record, log_record, reserved=self._skip_fields) |
|
181 | 175 | |
|
182 | 176 | if self.timestamp: |
|
183 | 177 | key = self.timestamp if type(self.timestamp) == str else 'timestamp' |
|
184 | 178 | log_record[key] = datetime.fromtimestamp(record.created, tz=utc) |
|
185 | 179 | |
|
186 | 180 | def process_log_record(self, log_record): |
|
187 | 181 | """ |
|
188 | 182 | Override this method to implement custom logic |
|
189 | 183 | on the possibly ordered dictionary. |
|
190 | 184 | """ |
|
191 | 185 | return log_record |
|
192 | 186 | |
|
193 | 187 | def jsonify_log_record(self, log_record): |
|
194 | 188 | """Returns a json string of the log record.""" |
|
195 | 189 | return self.json_serializer(log_record, |
|
196 | 190 | default=self.json_default, |
|
197 | 191 | cls=self.json_encoder, |
|
198 | 192 | indent=self.json_indent, |
|
199 | 193 | ensure_ascii=self.json_ensure_ascii) |
|
200 | 194 | |
|
201 | 195 | def serialize_log_record(self, log_record): |
|
202 | 196 | """Returns the final representation of the log record.""" |
|
203 | 197 | return "{}{}".format(self.prefix, self.jsonify_log_record(log_record)) |
|
204 | 198 | |
|
205 | 199 | def format(self, record): |
|
206 | 200 | """Formats a log record and serializes to json""" |
|
207 | 201 | message_dict = {} |
|
208 | 202 | # FIXME: logging.LogRecord.msg and logging.LogRecord.message in typeshed |
|
209 | 203 | # are always type of str. We shouldn't need to override that. |
|
210 | 204 | if isinstance(record.msg, dict): |
|
211 | 205 | message_dict = record.msg |
|
212 | 206 | record.message = None |
|
213 | 207 | else: |
|
214 | 208 | record.message = record.getMessage() |
|
215 | 209 | # only format time if needed |
|
216 | 210 | if "asctime" in self._required_fields: |
|
217 | 211 | record.asctime = self.formatTime(record, self.datefmt) |
|
218 | 212 | |
|
219 | 213 | # Display formatted exception, but allow overriding it in the |
|
220 | 214 | # user-supplied dict. |
|
221 | 215 | if record.exc_info and not message_dict.get('exc_info'): |
|
222 | 216 | message_dict['exc_info'] = self.formatException(record.exc_info) |
|
223 | 217 | if not message_dict.get('exc_info') and record.exc_text: |
|
224 | 218 | message_dict['exc_info'] = record.exc_text |
|
225 | 219 | # Display formatted record of stack frames |
|
226 | 220 | # default format is a string returned from :func:`traceback.print_stack` |
|
227 | 221 | try: |
|
228 | 222 | if record.stack_info and not message_dict.get('stack_info'): |
|
229 | 223 | message_dict['stack_info'] = self.formatStack(record.stack_info) |
|
230 | 224 | except AttributeError: |
|
231 | 225 | # Python2.7 doesn't have stack_info. |
|
232 | 226 | pass |
|
233 | 227 | |
|
234 | 228 | try: |
|
235 | 229 | log_record = OrderedDict() |
|
236 | 230 | except NameError: |
|
237 | 231 | log_record = {} |
|
238 | 232 | |
|
239 | 233 | _inject_req_id(record, with_prefix=False) |
|
240 | 234 | self.add_fields(log_record, record, message_dict) |
|
241 | 235 | log_record = self.process_log_record(log_record) |
|
242 | 236 | |
|
243 | 237 | return self.serialize_log_record(log_record) |
@@ -1,394 +1,402 b'' | |||
|
1 | 1 | |
|
2 | 2 | import threading |
|
3 | 3 | import weakref |
|
4 | 4 | from base64 import b64encode |
|
5 | 5 | from logging import getLogger |
|
6 | 6 | from os import urandom |
|
7 | 7 | from typing import Union |
|
8 | 8 | |
|
9 | 9 | from redis import StrictRedis |
|
10 | 10 | |
|
11 | 11 | __version__ = '4.0.0' |
|
12 | 12 | |
|
13 | 13 | loggers = { |
|
14 | 14 | k: getLogger("vcsserver." + ".".join((__name__, k))) |
|
15 | 15 | for k in [ |
|
16 | 16 | "acquire", |
|
17 | 17 | "refresh.thread.start", |
|
18 | 18 | "refresh.thread.stop", |
|
19 | 19 | "refresh.thread.exit", |
|
20 | 20 | "refresh.start", |
|
21 | 21 | "refresh.shutdown", |
|
22 | 22 | "refresh.exit", |
|
23 | 23 | "release", |
|
24 | 24 | ] |
|
25 | 25 | } |
|
26 | 26 | |
|
27 | 27 | text_type = str |
|
28 | 28 | binary_type = bytes |
|
29 | 29 | |
|
30 | 30 | |
|
31 | 31 | # Check if the id match. If not, return an error code. |
|
32 | 32 | UNLOCK_SCRIPT = b""" |
|
33 | 33 | if redis.call("get", KEYS[1]) ~= ARGV[1] then |
|
34 | 34 | return 1 |
|
35 | 35 | else |
|
36 | 36 | redis.call("del", KEYS[2]) |
|
37 | 37 | redis.call("lpush", KEYS[2], 1) |
|
38 | 38 | redis.call("pexpire", KEYS[2], ARGV[2]) |
|
39 | 39 | redis.call("del", KEYS[1]) |
|
40 | 40 | return 0 |
|
41 | 41 | end |
|
42 | 42 | """ |
|
43 | 43 | |
|
44 | 44 | # Covers both cases when key doesn't exist and doesn't equal to lock's id |
|
45 | 45 | EXTEND_SCRIPT = b""" |
|
46 | 46 | if redis.call("get", KEYS[1]) ~= ARGV[1] then |
|
47 | 47 | return 1 |
|
48 | 48 | elseif redis.call("ttl", KEYS[1]) < 0 then |
|
49 | 49 | return 2 |
|
50 | 50 | else |
|
51 | 51 | redis.call("expire", KEYS[1], ARGV[2]) |
|
52 | 52 | return 0 |
|
53 | 53 | end |
|
54 | 54 | """ |
|
55 | 55 | |
|
56 | 56 | RESET_SCRIPT = b""" |
|
57 | 57 | redis.call('del', KEYS[2]) |
|
58 | 58 | redis.call('lpush', KEYS[2], 1) |
|
59 | 59 | redis.call('pexpire', KEYS[2], ARGV[2]) |
|
60 | 60 | return redis.call('del', KEYS[1]) |
|
61 | 61 | """ |
|
62 | 62 | |
|
63 | 63 | RESET_ALL_SCRIPT = b""" |
|
64 | 64 | local locks = redis.call('keys', 'lock:*') |
|
65 | 65 | local signal |
|
66 | 66 | for _, lock in pairs(locks) do |
|
67 | 67 | signal = 'lock-signal:' .. string.sub(lock, 6) |
|
68 | 68 | redis.call('del', signal) |
|
69 | 69 | redis.call('lpush', signal, 1) |
|
70 | 70 | redis.call('expire', signal, 1) |
|
71 | 71 | redis.call('del', lock) |
|
72 | 72 | end |
|
73 | 73 | return #locks |
|
74 | 74 | """ |
|
75 | 75 | |
|
76 | 76 | |
|
77 | 77 | class AlreadyAcquired(RuntimeError): |
|
78 | 78 | pass |
|
79 | 79 | |
|
80 | 80 | |
|
81 | 81 | class NotAcquired(RuntimeError): |
|
82 | 82 | pass |
|
83 | 83 | |
|
84 | 84 | |
|
85 | 85 | class AlreadyStarted(RuntimeError): |
|
86 | 86 | pass |
|
87 | 87 | |
|
88 | 88 | |
|
89 | 89 | class TimeoutNotUsable(RuntimeError): |
|
90 | 90 | pass |
|
91 | 91 | |
|
92 | 92 | |
|
93 | 93 | class InvalidTimeout(RuntimeError): |
|
94 | 94 | pass |
|
95 | 95 | |
|
96 | 96 | |
|
97 | 97 | class TimeoutTooLarge(RuntimeError): |
|
98 | 98 | pass |
|
99 | 99 | |
|
100 | 100 | |
|
101 | 101 | class NotExpirable(RuntimeError): |
|
102 | 102 | pass |
|
103 | 103 | |
|
104 | 104 | |
|
105 | 105 | class Lock: |
|
106 | 106 | """ |
|
107 | 107 | A Lock context manager implemented via redis SETNX/BLPOP. |
|
108 | 108 | """ |
|
109 | 109 | |
|
110 | 110 | unlock_script = None |
|
111 | 111 | extend_script = None |
|
112 | 112 | reset_script = None |
|
113 | 113 | reset_all_script = None |
|
114 | blocking = None | |
|
114 | 115 | |
|
115 | 116 | _lock_renewal_interval: float |
|
116 | 117 | _lock_renewal_thread: Union[threading.Thread, None] |
|
117 | 118 | |
|
118 | def __init__(self, redis_client, name, expire=None, id=None, auto_renewal=False, strict=True, signal_expire=1000): | |
|
119 | def __init__(self, redis_client, name, expire=None, id=None, auto_renewal=False, strict=True, signal_expire=1000, blocking=True): | |
|
119 | 120 | """ |
|
120 | 121 | :param redis_client: |
|
121 | 122 | An instance of :class:`~StrictRedis`. |
|
122 | 123 | :param name: |
|
123 | 124 | The name (redis key) the lock should have. |
|
124 | 125 | :param expire: |
|
125 | 126 | The lock expiry time in seconds. If left at the default (None) |
|
126 | 127 | the lock will not expire. |
|
127 | 128 | :param id: |
|
128 | 129 | The ID (redis value) the lock should have. A random value is |
|
129 | 130 | generated when left at the default. |
|
130 | 131 | |
|
131 | 132 | Note that if you specify this then the lock is marked as "held". Acquires |
|
132 | 133 | won't be possible. |
|
133 | 134 | :param auto_renewal: |
|
134 | 135 | If set to ``True``, Lock will automatically renew the lock so that it |
|
135 | 136 | doesn't expire for as long as the lock is held (acquire() called |
|
136 | 137 | or running in a context manager). |
|
137 | 138 | |
|
138 | 139 | Implementation note: Renewal will happen using a daemon thread with |
|
139 | 140 | an interval of ``expire*2/3``. If wishing to use a different renewal |
|
140 | 141 | time, subclass Lock, call ``super().__init__()`` then set |
|
141 | 142 | ``self._lock_renewal_interval`` to your desired interval. |
|
142 | 143 | :param strict: |
|
143 | 144 | If set ``True`` then the ``redis_client`` needs to be an instance of ``redis.StrictRedis``. |
|
144 | 145 | :param signal_expire: |
|
145 | 146 | Advanced option to override signal list expiration in milliseconds. Increase it for very slow clients. Default: ``1000``. |
|
147 | :param blocking: | |
|
148 | Boolean value specifying whether lock should be blocking or not. | |
|
149 | Used in `__enter__` method. | |
|
146 | 150 | """ |
|
147 | 151 | if strict and not isinstance(redis_client, StrictRedis): |
|
148 | 152 | raise ValueError("redis_client must be instance of StrictRedis. " |
|
149 | 153 | "Use strict=False if you know what you're doing.") |
|
150 | 154 | if auto_renewal and expire is None: |
|
151 | 155 | raise ValueError("Expire may not be None when auto_renewal is set") |
|
152 | 156 | |
|
153 | 157 | self._client = redis_client |
|
154 | 158 | |
|
155 | 159 | if expire: |
|
156 | 160 | expire = int(expire) |
|
157 | 161 | if expire < 0: |
|
158 | 162 | raise ValueError("A negative expire is not acceptable.") |
|
159 | 163 | else: |
|
160 | 164 | expire = None |
|
161 | 165 | self._expire = expire |
|
162 | 166 | |
|
163 | 167 | self._signal_expire = signal_expire |
|
164 | 168 | if id is None: |
|
165 | 169 | self._id = b64encode(urandom(18)).decode('ascii') |
|
166 | 170 | elif isinstance(id, binary_type): |
|
167 | 171 | try: |
|
168 | 172 | self._id = id.decode('ascii') |
|
169 | 173 | except UnicodeDecodeError: |
|
170 | 174 | self._id = b64encode(id).decode('ascii') |
|
171 | 175 | elif isinstance(id, text_type): |
|
172 | 176 | self._id = id |
|
173 | 177 | else: |
|
174 | 178 | raise TypeError(f"Incorrect type for `id`. Must be bytes/str not {type(id)}.") |
|
175 | 179 | self._name = 'lock:' + name |
|
176 | 180 | self._signal = 'lock-signal:' + name |
|
177 | 181 | self._lock_renewal_interval = (float(expire) * 2 / 3 |
|
178 | 182 | if auto_renewal |
|
179 | 183 | else None) |
|
180 | 184 | self._lock_renewal_thread = None |
|
181 | 185 | |
|
186 | self.blocking = blocking | |
|
187 | ||
|
182 | 188 | self.register_scripts(redis_client) |
|
183 | 189 | |
|
184 | 190 | @classmethod |
|
185 | 191 | def register_scripts(cls, redis_client): |
|
186 | 192 | global reset_all_script |
|
187 | 193 | if reset_all_script is None: |
|
188 | 194 | cls.unlock_script = redis_client.register_script(UNLOCK_SCRIPT) |
|
189 | 195 | cls.extend_script = redis_client.register_script(EXTEND_SCRIPT) |
|
190 | 196 | cls.reset_script = redis_client.register_script(RESET_SCRIPT) |
|
191 | 197 | cls.reset_all_script = redis_client.register_script(RESET_ALL_SCRIPT) |
|
192 | 198 | reset_all_script = redis_client.register_script(RESET_ALL_SCRIPT) |
|
193 | 199 | |
|
194 | 200 | @property |
|
195 | 201 | def _held(self): |
|
196 | 202 | return self.id == self.get_owner_id() |
|
197 | 203 | |
|
198 | 204 | def reset(self): |
|
199 | 205 | """ |
|
200 | 206 | Forcibly deletes the lock. Use this with care. |
|
201 | 207 | """ |
|
202 | 208 | self.reset_script(client=self._client, keys=(self._name, self._signal), args=(self.id, self._signal_expire)) |
|
203 | 209 | |
|
204 | 210 | @property |
|
205 | 211 | def id(self): |
|
206 | 212 | return self._id |
|
207 | 213 | |
|
208 | 214 | def get_owner_id(self): |
|
209 | 215 | owner_id = self._client.get(self._name) |
|
210 | 216 | if isinstance(owner_id, binary_type): |
|
211 | 217 | owner_id = owner_id.decode('ascii', 'replace') |
|
212 | 218 | return owner_id |
|
213 | 219 | |
|
214 | 220 | def acquire(self, blocking=True, timeout=None): |
|
215 | 221 | """ |
|
216 | 222 | :param blocking: |
|
217 | 223 | Boolean value specifying whether lock should be blocking or not. |
|
218 | 224 | :param timeout: |
|
219 | 225 | An integer value specifying the maximum number of seconds to block. |
|
220 | 226 | """ |
|
221 | 227 | logger = loggers["acquire"] |
|
222 | 228 | |
|
223 | 229 | logger.debug("Getting blocking: %s acquire on %r ...", blocking, self._name) |
|
224 | 230 | |
|
225 | 231 | if self._held: |
|
226 | 232 | owner_id = self.get_owner_id() |
|
227 | 233 | raise AlreadyAcquired("Already acquired from this Lock instance. Lock id: {}".format(owner_id)) |
|
228 | 234 | |
|
229 | 235 | if not blocking and timeout is not None: |
|
230 | 236 | raise TimeoutNotUsable("Timeout cannot be used if blocking=False") |
|
231 | 237 | |
|
232 | 238 | if timeout: |
|
233 | 239 | timeout = int(timeout) |
|
234 | 240 | if timeout < 0: |
|
235 | 241 | raise InvalidTimeout(f"Timeout ({timeout}) cannot be less than or equal to 0") |
|
236 | 242 | |
|
237 | 243 | if self._expire and not self._lock_renewal_interval and timeout > self._expire: |
|
238 | 244 | raise TimeoutTooLarge(f"Timeout ({timeout}) cannot be greater than expire ({self._expire})") |
|
239 | 245 | |
|
240 | 246 | busy = True |
|
241 | 247 | blpop_timeout = timeout or self._expire or 0 |
|
242 | 248 | timed_out = False |
|
243 | 249 | while busy: |
|
244 | 250 | busy = not self._client.set(self._name, self._id, nx=True, ex=self._expire) |
|
245 | 251 | if busy: |
|
246 | 252 | if timed_out: |
|
247 | 253 | return False |
|
248 | 254 | elif blocking: |
|
249 | 255 | timed_out = not self._client.blpop(self._signal, blpop_timeout) and timeout |
|
250 | 256 | else: |
|
251 | 257 | logger.warning("Failed to acquire Lock(%r).", self._name) |
|
252 | 258 | return False |
|
253 | 259 | |
|
254 | 260 | logger.debug("Acquired Lock(%r).", self._name) |
|
255 | 261 | if self._lock_renewal_interval is not None: |
|
256 | 262 | self._start_lock_renewer() |
|
257 | 263 | return True |
|
258 | 264 | |
|
259 | 265 | def extend(self, expire=None): |
|
260 | 266 | """ |
|
261 | 267 | Extends expiration time of the lock. |
|
262 | 268 | |
|
263 | 269 | :param expire: |
|
264 | 270 | New expiration time. If ``None`` - `expire` provided during |
|
265 | 271 | lock initialization will be taken. |
|
266 | 272 | """ |
|
267 | 273 | if expire: |
|
268 | 274 | expire = int(expire) |
|
269 | 275 | if expire < 0: |
|
270 | 276 | raise ValueError("A negative expire is not acceptable.") |
|
271 | 277 | elif self._expire is not None: |
|
272 | 278 | expire = self._expire |
|
273 | 279 | else: |
|
274 | 280 | raise TypeError( |
|
275 | 281 | "To extend a lock 'expire' must be provided as an " |
|
276 | 282 | "argument to extend() method or at initialization time." |
|
277 | 283 | ) |
|
278 | 284 | |
|
279 | 285 | error = self.extend_script(client=self._client, keys=(self._name, self._signal), args=(self._id, expire)) |
|
280 | 286 | if error == 1: |
|
281 | 287 | raise NotAcquired(f"Lock {self._name} is not acquired or it already expired.") |
|
282 | 288 | elif error == 2: |
|
283 | 289 | raise NotExpirable(f"Lock {self._name} has no assigned expiration time") |
|
284 | 290 | elif error: |
|
285 | 291 | raise RuntimeError(f"Unsupported error code {error} from EXTEND script") |
|
286 | 292 | |
|
287 | 293 | @staticmethod |
|
288 | 294 | def _lock_renewer(name, lockref, interval, stop): |
|
289 | 295 | """ |
|
290 | 296 | Renew the lock key in redis every `interval` seconds for as long |
|
291 | 297 | as `self._lock_renewal_thread.should_exit` is False. |
|
292 | 298 | """ |
|
293 | 299 | while not stop.wait(timeout=interval): |
|
294 | 300 | loggers["refresh.thread.start"].debug("Refreshing Lock(%r).", name) |
|
295 | 301 | lock: "Lock" = lockref() |
|
296 | 302 | if lock is None: |
|
297 | 303 | loggers["refresh.thread.stop"].debug( |
|
298 | 304 | "Stopping loop because Lock(%r) was garbage collected.", name |
|
299 | 305 | ) |
|
300 | 306 | break |
|
301 | 307 | lock.extend(expire=lock._expire) |
|
302 | 308 | del lock |
|
303 | 309 | loggers["refresh.thread.exit"].debug("Exiting renewal thread for Lock(%r).", name) |
|
304 | 310 | |
|
305 | 311 | def _start_lock_renewer(self): |
|
306 | 312 | """ |
|
307 | 313 | Starts the lock refresher thread. |
|
308 | 314 | """ |
|
309 | 315 | if self._lock_renewal_thread is not None: |
|
310 | 316 | raise AlreadyStarted("Lock refresh thread already started") |
|
311 | 317 | |
|
312 | 318 | loggers["refresh.start"].debug( |
|
313 | 319 | "Starting renewal thread for Lock(%r). Refresh interval: %s seconds.", |
|
314 | 320 | self._name, self._lock_renewal_interval |
|
315 | 321 | ) |
|
316 | 322 | self._lock_renewal_stop = threading.Event() |
|
317 | 323 | self._lock_renewal_thread = threading.Thread( |
|
318 | 324 | group=None, |
|
319 | 325 | target=self._lock_renewer, |
|
320 | 326 | kwargs={ |
|
321 | 327 | 'name': self._name, |
|
322 | 328 | 'lockref': weakref.ref(self), |
|
323 | 329 | 'interval': self._lock_renewal_interval, |
|
324 | 330 | 'stop': self._lock_renewal_stop, |
|
325 | 331 | }, |
|
326 | 332 | ) |
|
327 | 333 | self._lock_renewal_thread.daemon = True |
|
328 | 334 | self._lock_renewal_thread.start() |
|
329 | 335 | |
|
330 | 336 | def _stop_lock_renewer(self): |
|
331 | 337 | """ |
|
332 | 338 | Stop the lock renewer. |
|
333 | 339 | |
|
334 | 340 | This signals the renewal thread and waits for its exit. |
|
335 | 341 | """ |
|
336 | 342 | if self._lock_renewal_thread is None or not self._lock_renewal_thread.is_alive(): |
|
337 | 343 | return |
|
338 | 344 | loggers["refresh.shutdown"].debug("Signaling renewal thread for Lock(%r) to exit.", self._name) |
|
339 | 345 | self._lock_renewal_stop.set() |
|
340 | 346 | self._lock_renewal_thread.join() |
|
341 | 347 | self._lock_renewal_thread = None |
|
342 | 348 | loggers["refresh.exit"].debug("Renewal thread for Lock(%r) exited.", self._name) |
|
343 | 349 | |
|
344 | 350 | def __enter__(self): |
|
345 |
acquired = self.acquire(blocking= |
|
|
351 | acquired = self.acquire(blocking=self.blocking) | |
|
346 | 352 | if not acquired: |
|
347 | raise AssertionError(f"Lock({self._name}) wasn't acquired, but blocking=True was used!") | |
|
353 | if self.blocking: | |
|
354 | raise AssertionError(f"Lock({self._name}) wasn't acquired, but blocking=True was used!") | |
|
355 | raise NotAcquired(f"Lock({self._name}) is not acquired or it already expired.") | |
|
348 | 356 | return self |
|
349 | 357 | |
|
350 | 358 | def __exit__(self, exc_type=None, exc_value=None, traceback=None): |
|
351 | 359 | self.release() |
|
352 | 360 | |
|
353 | 361 | def release(self): |
|
354 | 362 | """Releases the lock, that was acquired with the same object. |
|
355 | 363 | |
|
356 | 364 | .. note:: |
|
357 | 365 | |
|
358 | 366 | If you want to release a lock that you acquired in a different place you have two choices: |
|
359 | 367 | |
|
360 | 368 | * Use ``Lock("name", id=id_from_other_place).release()`` |
|
361 | 369 | * Use ``Lock("name").reset()`` |
|
362 | 370 | """ |
|
363 | 371 | if self._lock_renewal_thread is not None: |
|
364 | 372 | self._stop_lock_renewer() |
|
365 | 373 | loggers["release"].debug("Releasing Lock(%r).", self._name) |
|
366 | 374 | error = self.unlock_script(client=self._client, keys=(self._name, self._signal), args=(self._id, self._signal_expire)) |
|
367 | 375 | if error == 1: |
|
368 | 376 | raise NotAcquired(f"Lock({self._name}) is not acquired or it already expired.") |
|
369 | 377 | elif error: |
|
370 | 378 | raise RuntimeError(f"Unsupported error code {error} from EXTEND script.") |
|
371 | 379 | |
|
372 | 380 | def locked(self): |
|
373 | 381 | """ |
|
374 | 382 | Return true if the lock is acquired. |
|
375 | 383 | |
|
376 | 384 | Checks that lock with same name already exists. This method returns true, even if |
|
377 | 385 | lock have another id. |
|
378 | 386 | """ |
|
379 | 387 | return self._client.exists(self._name) == 1 |
|
380 | 388 | |
|
381 | 389 | |
|
382 | 390 | reset_all_script = None |
|
383 | 391 | |
|
384 | 392 | |
|
385 | 393 | def reset_all(redis_client): |
|
386 | 394 | """ |
|
387 | 395 | Forcibly deletes all locks if its remains (like a crash reason). Use this with care. |
|
388 | 396 | |
|
389 | 397 | :param redis_client: |
|
390 | 398 | An instance of :class:`~StrictRedis`. |
|
391 | 399 | """ |
|
392 | 400 | Lock.register_scripts(redis_client) |
|
393 | 401 | |
|
394 | 402 | reset_all_script(client=redis_client) # noqa |
@@ -1,50 +1,51 b'' | |||
|
1 | 1 | import logging |
|
2 | 2 | |
|
3 | 3 | from .stream import TCPStatsClient, UnixSocketStatsClient # noqa |
|
4 | 4 | from .udp import StatsClient # noqa |
|
5 | 5 | |
|
6 | 6 | HOST = 'localhost' |
|
7 | 7 | PORT = 8125 |
|
8 | 8 | IPV6 = False |
|
9 | 9 | PREFIX = None |
|
10 | 10 | MAXUDPSIZE = 512 |
|
11 | 11 | |
|
12 | 12 | log = logging.getLogger('rhodecode.statsd') |
|
13 | 13 | |
|
14 | 14 | |
|
15 | 15 | def statsd_config(config, prefix='statsd.'): |
|
16 | 16 | _config = {} |
|
17 | for key in config.keys(): | |
|
17 | for key in list(config.keys()): | |
|
18 | 18 | if key.startswith(prefix): |
|
19 | 19 | _config[key[len(prefix):]] = config[key] |
|
20 | 20 | return _config |
|
21 | 21 | |
|
22 | 22 | |
|
23 | 23 | def client_from_config(configuration, prefix='statsd.', **kwargs): |
|
24 | 24 | from pyramid.settings import asbool |
|
25 | 25 | |
|
26 | 26 | _config = statsd_config(configuration, prefix) |
|
27 | statsd_flag = _config.get('enabled') | |
|
27 | 28 | statsd_enabled = asbool(_config.pop('enabled', False)) |
|
28 | 29 | if not statsd_enabled: |
|
29 | log.debug('statsd client not enabled by statsd.enabled = flag, skipping...') | |
|
30 | log.debug('statsd client not enabled by statsd.enabled = %s flag, skipping...', statsd_flag) | |
|
30 | 31 | return |
|
31 | 32 | |
|
32 | 33 | host = _config.pop('statsd_host', HOST) |
|
33 | 34 | port = _config.pop('statsd_port', PORT) |
|
34 | 35 | prefix = _config.pop('statsd_prefix', PREFIX) |
|
35 | 36 | maxudpsize = _config.pop('statsd_maxudpsize', MAXUDPSIZE) |
|
36 | 37 | ipv6 = asbool(_config.pop('statsd_ipv6', IPV6)) |
|
37 | 38 | log.debug('configured statsd client %s:%s', host, port) |
|
38 | 39 | |
|
39 | 40 | try: |
|
40 | 41 | client = StatsClient( |
|
41 | 42 | host=host, port=port, prefix=prefix, maxudpsize=maxudpsize, ipv6=ipv6) |
|
42 | 43 | except Exception: |
|
43 | 44 | log.exception('StatsD is enabled, but failed to connect to statsd server, fallback: disable statsd') |
|
44 | 45 | client = None |
|
45 | 46 | |
|
46 | 47 | return client |
|
47 | 48 | |
|
48 | 49 | |
|
49 | 50 | def get_statsd_client(request): |
|
50 | 51 | return client_from_config(request.registry.settings) |
@@ -1,40 +1,58 b'' | |||
|
1 | 1 | # Copyright (C) 2010-2023 RhodeCode GmbH |
|
2 | 2 | # |
|
3 | 3 | # This program is free software: you can redistribute it and/or modify |
|
4 | 4 | # it under the terms of the GNU Affero General Public License, version 3 |
|
5 | 5 | # (only), as published by the Free Software Foundation. |
|
6 | 6 | # |
|
7 | 7 | # This program is distributed in the hope that it will be useful, |
|
8 | 8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
9 | 9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
10 | 10 | # GNU General Public License for more details. |
|
11 | 11 | # |
|
12 | 12 | # You should have received a copy of the GNU Affero General Public License |
|
13 | 13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. |
|
14 | 14 | # |
|
15 | 15 | # This program is dual-licensed. If you wish to learn more about the |
|
16 | 16 | # RhodeCode Enterprise Edition, including its added features, Support services, |
|
17 | 17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ |
|
18 | 18 | import os |
|
19 | import vcsserver | |
|
20 | import vcsserver.settings | |
|
19 | 21 | |
|
20 | 22 | |
|
21 | 23 | def get_config(ini_path, **kwargs): |
|
22 | 24 | import configparser |
|
23 | 25 | parser = configparser.ConfigParser(**kwargs) |
|
24 | 26 | parser.read(ini_path) |
|
25 | 27 | return parser |
|
26 | 28 | |
|
27 | 29 | |
|
28 | 30 | def get_app_config_lightweight(ini_path): |
|
29 | 31 | parser = get_config(ini_path) |
|
30 | 32 | parser.set('app:main', 'here', os.getcwd()) |
|
31 | 33 | parser.set('app:main', '__file__', ini_path) |
|
32 | 34 | return dict(parser.items('app:main')) |
|
33 | 35 | |
|
34 | 36 | |
|
35 | 37 | def get_app_config(ini_path): |
|
36 | 38 | """ |
|
37 | 39 | This loads the app context and provides a heavy type iniliaziation of config |
|
38 | 40 | """ |
|
39 | 41 | from paste.deploy.loadwsgi import appconfig |
|
40 | 42 | return appconfig(f'config:{ini_path}', relative_to=os.getcwd()) |
|
43 | ||
|
44 | ||
|
45 | def configure_and_store_settings(global_config, app_settings): | |
|
46 | """ | |
|
47 | Configure the settings module. | |
|
48 | """ | |
|
49 | settings_merged = global_config.copy() | |
|
50 | settings_merged.update(app_settings) | |
|
51 | ||
|
52 | binary_dir = app_settings['core.binary_dir'] | |
|
53 | ||
|
54 | vcsserver.settings.BINARY_DIR = binary_dir | |
|
55 | ||
|
56 | # Store the settings to make them available to other modules. | |
|
57 | vcsserver.PYRAMID_SETTINGS = settings_merged | |
|
58 | vcsserver.CONFIG = settings_merged |
@@ -1,2 +1,14 b'' | |||
|
1 | # use orjson by default | |
|
2 | import orjson as json | |
|
1 | import json as stdlib_json | |
|
2 | ||
|
3 | try: | |
|
4 | # we keep simplejson for having dump functionality still | |
|
5 | # orjson doesn't support it | |
|
6 | import simplejson as sjson | |
|
7 | except ImportError: | |
|
8 | sjson = stdlib_json | |
|
9 | ||
|
10 | try: | |
|
11 | import orjson | |
|
12 | import orjson as json | |
|
13 | except ImportError: | |
|
14 | json = stdlib_json |
@@ -1,53 +1,63 b'' | |||
|
1 | 1 | # Copyright (C) 2010-2023 RhodeCode GmbH |
|
2 | 2 | # |
|
3 | 3 | # This program is free software: you can redistribute it and/or modify |
|
4 | 4 | # it under the terms of the GNU Affero General Public License, version 3 |
|
5 | 5 | # (only), as published by the Free Software Foundation. |
|
6 | 6 | # |
|
7 | 7 | # This program is distributed in the hope that it will be useful, |
|
8 | 8 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
9 | 9 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
10 | 10 | # GNU General Public License for more details. |
|
11 | 11 | # |
|
12 | 12 | # You should have received a copy of the GNU Affero General Public License |
|
13 | 13 | # along with this program. If not, see <http://www.gnu.org/licenses/>. |
|
14 | 14 | # |
|
15 | 15 | # This program is dual-licensed. If you wish to learn more about the |
|
16 | 16 | # RhodeCode Enterprise Edition, including its added features, Support services, |
|
17 | 17 | # and proprietary license terms, please see https://rhodecode.com/licenses/ |
|
18 | 18 | |
|
19 | 19 | import sys |
|
20 | 20 | import logging |
|
21 | 21 | |
|
22 | 22 | |
|
23 | 23 | BLACK, RED, GREEN, YELLOW, BLUE, MAGENTA, CYAN, WHITE = list(range(30, 38)) |
|
24 | 24 | |
|
25 | 25 | # Sequences |
|
26 | 26 | RESET_SEQ = "\033[0m" |
|
27 | 27 | COLOR_SEQ = "\033[0;%dm" |
|
28 | 28 | BOLD_SEQ = "\033[1m" |
|
29 | 29 | |
|
30 | 30 | COLORS = { |
|
31 | 31 | 'CRITICAL': MAGENTA, |
|
32 | 32 | 'ERROR': RED, |
|
33 | 33 | 'WARNING': CYAN, |
|
34 | 34 | 'INFO': GREEN, |
|
35 | 35 | 'DEBUG': BLUE, |
|
36 | 36 | 'SQL': YELLOW |
|
37 | 37 | } |
|
38 | 38 | |
|
39 | 39 | |
|
40 | def _inject_req_id(record, *args, **kwargs): | |
|
41 | return record | |
|
42 | ||
|
43 | ||
|
44 | class ExceptionAwareFormatter(logging.Formatter): | |
|
45 | pass | |
|
46 | ||
|
47 | ||
|
40 | 48 | class ColorFormatter(logging.Formatter): |
|
41 | 49 | |
|
42 | 50 | def format(self, record): |
|
43 | 51 | """ |
|
44 | Change record's levelname to use with COLORS enum | |
|
52 | Changes record's levelname to use with COLORS enum | |
|
45 | 53 | """ |
|
46 | 54 | def_record = super().format(record) |
|
47 | 55 | |
|
48 | 56 | levelname = record.levelname |
|
49 | 57 | start = COLOR_SEQ % (COLORS[levelname]) |
|
50 | 58 | end = RESET_SEQ |
|
51 | 59 | |
|
52 | 60 | colored_record = ''.join([start, def_record, end]) |
|
53 | 61 | return colored_record |
|
62 | ||
|
63 |
@@ -1,63 +1,63 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | |
|
19 | 19 | import logging |
|
20 | 20 | |
|
21 | 21 | from repoze.lru import LRUCache |
|
22 | 22 | |
|
23 | from vcsserver.str_utils import safe_str | |
|
23 | from vcsserver.lib.str_utils import safe_str | |
|
24 | 24 | |
|
25 | 25 | log = logging.getLogger(__name__) |
|
26 | 26 | |
|
27 | 27 | |
|
28 | 28 | class LRUDict(LRUCache): |
|
29 | 29 | """ |
|
30 | 30 | Wrapper to provide partial dict access |
|
31 | 31 | """ |
|
32 | 32 | |
|
33 | 33 | def __setitem__(self, key, value): |
|
34 | 34 | return self.put(key, value) |
|
35 | 35 | |
|
36 | 36 | def __getitem__(self, key): |
|
37 | 37 | return self.get(key) |
|
38 | 38 | |
|
39 | 39 | def __contains__(self, key): |
|
40 | 40 | return bool(self.get(key)) |
|
41 | 41 | |
|
42 | 42 | def __delitem__(self, key): |
|
43 | 43 | del self.data[key] |
|
44 | 44 | |
|
45 | 45 | def keys(self): |
|
46 | 46 | return list(self.data.keys()) |
|
47 | 47 | |
|
48 | 48 | |
|
49 | 49 | class LRUDictDebug(LRUDict): |
|
50 | 50 | """ |
|
51 | 51 | Wrapper to provide some debug options |
|
52 | 52 | """ |
|
53 | 53 | def _report_keys(self): |
|
54 | 54 | elems_cnt = f'{len(list(self.keys()))}/{self.size}' |
|
55 | 55 | # trick for pformat print it more nicely |
|
56 | 56 | fmt = '\n' |
|
57 | 57 | for cnt, elem in enumerate(self.keys()): |
|
58 | 58 | fmt += f'{cnt+1} - {safe_str(elem)}\n' |
|
59 | 59 | log.debug('current LRU keys (%s):%s', elems_cnt, fmt) |
|
60 | 60 | |
|
61 | 61 | def __getitem__(self, key): |
|
62 | 62 | self._report_keys() |
|
63 | 63 | return self.get(key) |
@@ -1,303 +1,303 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | #import errno |
|
19 | 19 | import fcntl |
|
20 | 20 | import functools |
|
21 | 21 | import logging |
|
22 | 22 | import os |
|
23 | 23 | import pickle |
|
24 | 24 | #import time |
|
25 | 25 | |
|
26 | 26 | #import gevent |
|
27 | 27 | import msgpack |
|
28 | 28 | import redis |
|
29 | 29 | |
|
30 | 30 | flock_org = fcntl.flock |
|
31 | 31 | from typing import Union |
|
32 | 32 | |
|
33 | 33 | from dogpile.cache.api import Deserializer, Serializer |
|
34 | 34 | from dogpile.cache.backends import file as file_backend |
|
35 | 35 | from dogpile.cache.backends import memory as memory_backend |
|
36 | 36 | from dogpile.cache.backends import redis as redis_backend |
|
37 | 37 | from dogpile.cache.backends.file import FileLock |
|
38 | 38 | from dogpile.cache.util import memoized_property |
|
39 | 39 | |
|
40 | 40 | from vcsserver.lib.memory_lru_dict import LRUDict, LRUDictDebug |
|
41 | from vcsserver.str_utils import safe_bytes, safe_str | |
|
42 | from vcsserver.type_utils import str2bool | |
|
41 | from vcsserver.lib.str_utils import safe_bytes, safe_str | |
|
42 | from vcsserver.lib.type_utils import str2bool | |
|
43 | 43 | |
|
44 | 44 | _default_max_size = 1024 |
|
45 | 45 | |
|
46 | 46 | log = logging.getLogger(__name__) |
|
47 | 47 | |
|
48 | 48 | |
|
49 | 49 | class LRUMemoryBackend(memory_backend.MemoryBackend): |
|
50 | 50 | key_prefix = 'lru_mem_backend' |
|
51 | 51 | pickle_values = False |
|
52 | 52 | |
|
53 | 53 | def __init__(self, arguments): |
|
54 | 54 | self.max_size = arguments.pop('max_size', _default_max_size) |
|
55 | 55 | |
|
56 | 56 | LRUDictClass = LRUDict |
|
57 | 57 | if arguments.pop('log_key_count', None): |
|
58 | 58 | LRUDictClass = LRUDictDebug |
|
59 | 59 | |
|
60 | 60 | arguments['cache_dict'] = LRUDictClass(self.max_size) |
|
61 | 61 | super().__init__(arguments) |
|
62 | 62 | |
|
63 | 63 | def __repr__(self): |
|
64 | 64 | return f'{self.__class__}(maxsize=`{self.max_size}`)' |
|
65 | 65 | |
|
66 | 66 | def __str__(self): |
|
67 | 67 | return self.__repr__() |
|
68 | 68 | |
|
69 | 69 | def delete(self, key): |
|
70 | 70 | try: |
|
71 | 71 | del self._cache[key] |
|
72 | 72 | except KeyError: |
|
73 | 73 | # we don't care if key isn't there at deletion |
|
74 | 74 | pass |
|
75 | 75 | |
|
76 | 76 | def list_keys(self, prefix): |
|
77 | 77 | return list(self._cache.keys()) |
|
78 | 78 | |
|
79 | 79 | def delete_multi(self, keys): |
|
80 | 80 | for key in keys: |
|
81 | 81 | self.delete(key) |
|
82 | 82 | |
|
83 | 83 | def delete_multi_by_prefix(self, prefix): |
|
84 | 84 | cache_keys = self.list_keys(prefix=prefix) |
|
85 | 85 | num_affected_keys = len(cache_keys) |
|
86 | 86 | if num_affected_keys: |
|
87 | 87 | self.delete_multi(cache_keys) |
|
88 | 88 | return num_affected_keys |
|
89 | 89 | |
|
90 | 90 | |
|
91 | 91 | class PickleSerializer: |
|
92 | 92 | serializer: None | Serializer = staticmethod( # type: ignore |
|
93 | 93 | functools.partial(pickle.dumps, protocol=pickle.HIGHEST_PROTOCOL) |
|
94 | 94 | ) |
|
95 | 95 | deserializer: None | Deserializer = staticmethod( # type: ignore |
|
96 | 96 | functools.partial(pickle.loads) |
|
97 | 97 | ) |
|
98 | 98 | |
|
99 | 99 | |
|
100 | 100 | class MsgPackSerializer: |
|
101 | 101 | serializer: None | Serializer = staticmethod( # type: ignore |
|
102 | 102 | msgpack.packb |
|
103 | 103 | ) |
|
104 | 104 | deserializer: None | Deserializer = staticmethod( # type: ignore |
|
105 | 105 | functools.partial(msgpack.unpackb, use_list=False) |
|
106 | 106 | ) |
|
107 | 107 | |
|
108 | 108 | |
|
109 | 109 | class CustomLockFactory(FileLock): |
|
110 | 110 | |
|
111 | 111 | pass |
|
112 | 112 | |
|
113 | 113 | |
|
114 | 114 | class FileNamespaceBackend(PickleSerializer, file_backend.DBMBackend): |
|
115 | 115 | key_prefix = 'file_backend' |
|
116 | 116 | |
|
117 | 117 | def __init__(self, arguments): |
|
118 | 118 | arguments['lock_factory'] = CustomLockFactory |
|
119 | 119 | db_file = arguments.get('filename') |
|
120 | 120 | |
|
121 | 121 | log.debug('initialing cache-backend=%s db in %s', self.__class__.__name__, db_file) |
|
122 | 122 | db_file_dir = os.path.dirname(db_file) |
|
123 | 123 | if not os.path.isdir(db_file_dir): |
|
124 | 124 | os.makedirs(db_file_dir) |
|
125 | 125 | |
|
126 | 126 | try: |
|
127 | 127 | super().__init__(arguments) |
|
128 | 128 | except Exception: |
|
129 | 129 | log.exception('Failed to initialize db at: %s', db_file) |
|
130 | 130 | raise |
|
131 | 131 | |
|
132 | 132 | def __repr__(self): |
|
133 | 133 | return f'{self.__class__}(file=`{self.filename}`)' |
|
134 | 134 | |
|
135 | 135 | def __str__(self): |
|
136 | 136 | return self.__repr__() |
|
137 | 137 | |
|
138 | 138 | def _get_keys_pattern(self, prefix: bytes = b''): |
|
139 | 139 | return b'%b:%b' % (safe_bytes(self.key_prefix), safe_bytes(prefix)) |
|
140 | 140 | |
|
141 | 141 | def list_keys(self, prefix: bytes = b''): |
|
142 | 142 | prefix = self._get_keys_pattern(prefix) |
|
143 | 143 | |
|
144 | 144 | def cond(dbm_key: bytes): |
|
145 | 145 | if not prefix: |
|
146 | 146 | return True |
|
147 | 147 | |
|
148 | 148 | if dbm_key.startswith(prefix): |
|
149 | 149 | return True |
|
150 | 150 | return False |
|
151 | 151 | |
|
152 | 152 | with self._dbm_file(True) as dbm: |
|
153 | 153 | try: |
|
154 | 154 | return list(filter(cond, dbm.keys())) |
|
155 | 155 | except Exception: |
|
156 | 156 | log.error('Failed to fetch DBM keys from DB: %s', self.get_store()) |
|
157 | 157 | raise |
|
158 | 158 | |
|
159 | 159 | def delete_multi_by_prefix(self, prefix): |
|
160 | 160 | cache_keys = self.list_keys(prefix=prefix) |
|
161 | 161 | num_affected_keys = len(cache_keys) |
|
162 | 162 | if num_affected_keys: |
|
163 | 163 | self.delete_multi(cache_keys) |
|
164 | 164 | return num_affected_keys |
|
165 | 165 | |
|
166 | 166 | def get_store(self): |
|
167 | 167 | return self.filename |
|
168 | 168 | |
|
169 | 169 | |
|
170 | 170 | class BaseRedisBackend(redis_backend.RedisBackend): |
|
171 | 171 | key_prefix = '' |
|
172 | 172 | |
|
173 | 173 | def __init__(self, arguments): |
|
174 | 174 | self.db_conn = arguments.get('host', '') or arguments.get('url', '') or 'redis-host' |
|
175 | 175 | super().__init__(arguments) |
|
176 | 176 | |
|
177 | 177 | self._lock_timeout = self.lock_timeout |
|
178 | 178 | self._lock_auto_renewal = str2bool(arguments.pop("lock_auto_renewal", True)) |
|
179 | 179 | |
|
180 | 180 | if self._lock_auto_renewal and not self._lock_timeout: |
|
181 | 181 | # set default timeout for auto_renewal |
|
182 | 182 | self._lock_timeout = 30 |
|
183 | 183 | |
|
184 | 184 | def __repr__(self): |
|
185 | 185 | return f'{self.__class__}(conn=`{self.db_conn}`)' |
|
186 | 186 | |
|
187 | 187 | def __str__(self): |
|
188 | 188 | return self.__repr__() |
|
189 | 189 | |
|
190 | 190 | def _create_client(self): |
|
191 | 191 | args = {} |
|
192 | 192 | |
|
193 | 193 | if self.url is not None: |
|
194 | 194 | args.update(url=self.url) |
|
195 | 195 | |
|
196 | 196 | else: |
|
197 | 197 | args.update( |
|
198 | 198 | host=self.host, password=self.password, |
|
199 | 199 | port=self.port, db=self.db |
|
200 | 200 | ) |
|
201 | 201 | |
|
202 | 202 | connection_pool = redis.ConnectionPool(**args) |
|
203 | 203 | self.writer_client = redis.StrictRedis( |
|
204 | 204 | connection_pool=connection_pool |
|
205 | 205 | ) |
|
206 | 206 | self.reader_client = self.writer_client |
|
207 | 207 | |
|
208 | 208 | def _get_keys_pattern(self, prefix: bytes = b''): |
|
209 | 209 | return b'%b:%b*' % (safe_bytes(self.key_prefix), safe_bytes(prefix)) |
|
210 | 210 | |
|
211 | 211 | def list_keys(self, prefix: bytes = b''): |
|
212 | 212 | prefix = self._get_keys_pattern(prefix) |
|
213 | 213 | return self.reader_client.keys(prefix) |
|
214 | 214 | |
|
215 | 215 | def delete_multi_by_prefix(self, prefix, use_lua=False): |
|
216 | 216 | if use_lua: |
|
217 | 217 | # high efficient LUA script to delete ALL keys by prefix... |
|
218 | 218 | lua = """local keys = redis.call('keys', ARGV[1]) |
|
219 | 219 | for i=1,#keys,5000 do |
|
220 | 220 | redis.call('del', unpack(keys, i, math.min(i+(5000-1), #keys))) |
|
221 | 221 | end |
|
222 | 222 | return #keys""" |
|
223 | 223 | num_affected_keys = self.writer_client.eval( |
|
224 | 224 | lua, |
|
225 | 225 | 0, |
|
226 | 226 | f"{prefix}*") |
|
227 | 227 | else: |
|
228 | 228 | cache_keys = self.list_keys(prefix=prefix) |
|
229 | 229 | num_affected_keys = len(cache_keys) |
|
230 | 230 | if num_affected_keys: |
|
231 | 231 | self.delete_multi(cache_keys) |
|
232 | 232 | return num_affected_keys |
|
233 | 233 | |
|
234 | 234 | def get_store(self): |
|
235 | 235 | return self.reader_client.connection_pool |
|
236 | 236 | |
|
237 | 237 | def get_mutex(self, key): |
|
238 | 238 | if self.distributed_lock: |
|
239 | 239 | lock_key = f'_lock_{safe_str(key)}' |
|
240 | 240 | return get_mutex_lock( |
|
241 | 241 | self.writer_client, lock_key, |
|
242 | 242 | self._lock_timeout, |
|
243 | 243 | auto_renewal=self._lock_auto_renewal |
|
244 | 244 | ) |
|
245 | 245 | else: |
|
246 | 246 | return None |
|
247 | 247 | |
|
248 | 248 | |
|
249 | 249 | class RedisPickleBackend(PickleSerializer, BaseRedisBackend): |
|
250 | 250 | key_prefix = 'redis_pickle_backend' |
|
251 | 251 | pass |
|
252 | 252 | |
|
253 | 253 | |
|
254 | 254 | class RedisMsgPackBackend(MsgPackSerializer, BaseRedisBackend): |
|
255 | 255 | key_prefix = 'redis_msgpack_backend' |
|
256 | 256 | pass |
|
257 | 257 | |
|
258 | 258 | |
|
259 | 259 | def get_mutex_lock(client, lock_key, lock_timeout, auto_renewal=False): |
|
260 | 260 | from vcsserver.lib._vendor import redis_lock |
|
261 | 261 | |
|
262 | 262 | class _RedisLockWrapper: |
|
263 | 263 | """LockWrapper for redis_lock""" |
|
264 | 264 | |
|
265 | 265 | @classmethod |
|
266 | 266 | def get_lock(cls): |
|
267 | 267 | return redis_lock.Lock( |
|
268 | 268 | redis_client=client, |
|
269 | 269 | name=lock_key, |
|
270 | 270 | expire=lock_timeout, |
|
271 | 271 | auto_renewal=auto_renewal, |
|
272 | 272 | strict=True, |
|
273 | 273 | ) |
|
274 | 274 | |
|
275 | 275 | def __repr__(self): |
|
276 | 276 | return f"{self.__class__.__name__}:{lock_key}" |
|
277 | 277 | |
|
278 | 278 | def __str__(self): |
|
279 | 279 | return f"{self.__class__.__name__}:{lock_key}" |
|
280 | 280 | |
|
281 | 281 | def __init__(self): |
|
282 | 282 | self.lock = self.get_lock() |
|
283 | 283 | self.lock_key = lock_key |
|
284 | 284 | |
|
285 | 285 | def acquire(self, wait=True): |
|
286 | 286 | log.debug('Trying to acquire Redis lock for key %s', self.lock_key) |
|
287 | 287 | try: |
|
288 | 288 | acquired = self.lock.acquire(wait) |
|
289 | 289 | log.debug('Got lock for key %s, %s', self.lock_key, acquired) |
|
290 | 290 | return acquired |
|
291 | 291 | except redis_lock.AlreadyAcquired: |
|
292 | 292 | return False |
|
293 | 293 | except redis_lock.AlreadyStarted: |
|
294 | 294 | # refresh thread exists, but it also means we acquired the lock |
|
295 | 295 | return True |
|
296 | 296 | |
|
297 | 297 | def release(self): |
|
298 | 298 | try: |
|
299 | 299 | self.lock.release() |
|
300 | 300 | except redis_lock.NotAcquired: |
|
301 | 301 | pass |
|
302 | 302 | |
|
303 | 303 | return _RedisLockWrapper() |
@@ -1,245 +1,245 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import functools |
|
19 | 19 | import logging |
|
20 | 20 | import os |
|
21 | 21 | import threading |
|
22 | 22 | import time |
|
23 | 23 | |
|
24 | 24 | import decorator |
|
25 | 25 | from dogpile.cache import CacheRegion |
|
26 | 26 | |
|
27 | 27 | |
|
28 | 28 | from vcsserver.utils import sha1 |
|
29 | from vcsserver.str_utils import safe_bytes | |
|
30 | from vcsserver.type_utils import str2bool # noqa :required by imports from .utils | |
|
29 | from vcsserver.lib.str_utils import safe_bytes | |
|
30 | from vcsserver.lib.type_utils import str2bool # noqa :required by imports from .utils | |
|
31 | 31 | |
|
32 | 32 | from . import region_meta |
|
33 | 33 | |
|
34 | 34 | log = logging.getLogger(__name__) |
|
35 | 35 | |
|
36 | 36 | |
|
37 | 37 | class RhodeCodeCacheRegion(CacheRegion): |
|
38 | 38 | |
|
39 | 39 | def __repr__(self): |
|
40 | 40 | return f'`{self.__class__.__name__}(name={self.name}, backend={self.backend.__class__})`' |
|
41 | 41 | |
|
42 | 42 | def conditional_cache_on_arguments( |
|
43 | 43 | self, namespace=None, |
|
44 | 44 | expiration_time=None, |
|
45 | 45 | should_cache_fn=None, |
|
46 | 46 | to_str=str, |
|
47 | 47 | function_key_generator=None, |
|
48 | 48 | condition=True): |
|
49 | 49 | """ |
|
50 | 50 | Custom conditional decorator, that will not touch any dogpile internals if |
|
51 | 51 | condition isn't meet. This works a bit different from should_cache_fn |
|
52 | 52 | And it's faster in cases we don't ever want to compute cached values |
|
53 | 53 | """ |
|
54 | 54 | expiration_time_is_callable = callable(expiration_time) |
|
55 | 55 | if not namespace: |
|
56 | 56 | namespace = getattr(self, '_default_namespace', None) |
|
57 | 57 | |
|
58 | 58 | if function_key_generator is None: |
|
59 | 59 | function_key_generator = self.function_key_generator |
|
60 | 60 | |
|
61 | 61 | def get_or_create_for_user_func(func_key_generator, user_func, *arg, **kw): |
|
62 | 62 | |
|
63 | 63 | if not condition: |
|
64 | 64 | log.debug('Calling un-cached method:%s', user_func.__name__) |
|
65 | 65 | start = time.time() |
|
66 | 66 | result = user_func(*arg, **kw) |
|
67 | 67 | total = time.time() - start |
|
68 | 68 | log.debug('un-cached method:%s took %.4fs', user_func.__name__, total) |
|
69 | 69 | return result |
|
70 | 70 | |
|
71 | 71 | key = func_key_generator(*arg, **kw) |
|
72 | 72 | |
|
73 | 73 | timeout = expiration_time() if expiration_time_is_callable \ |
|
74 | 74 | else expiration_time |
|
75 | 75 | |
|
76 | 76 | log.debug('Calling cached method:`%s`', user_func.__name__) |
|
77 | 77 | return self.get_or_create(key, user_func, timeout, should_cache_fn, (arg, kw)) |
|
78 | 78 | |
|
79 | 79 | def cache_decorator(user_func): |
|
80 | 80 | if to_str is str: |
|
81 | 81 | # backwards compatible |
|
82 | 82 | key_generator = function_key_generator(namespace, user_func) |
|
83 | 83 | else: |
|
84 | 84 | key_generator = function_key_generator(namespace, user_func, to_str=to_str) |
|
85 | 85 | |
|
86 | 86 | def refresh(*arg, **kw): |
|
87 | 87 | """ |
|
88 | 88 | Like invalidate, but regenerates the value instead |
|
89 | 89 | """ |
|
90 | 90 | key = key_generator(*arg, **kw) |
|
91 | 91 | value = user_func(*arg, **kw) |
|
92 | 92 | self.set(key, value) |
|
93 | 93 | return value |
|
94 | 94 | |
|
95 | 95 | def invalidate(*arg, **kw): |
|
96 | 96 | key = key_generator(*arg, **kw) |
|
97 | 97 | self.delete(key) |
|
98 | 98 | |
|
99 | 99 | def set_(value, *arg, **kw): |
|
100 | 100 | key = key_generator(*arg, **kw) |
|
101 | 101 | self.set(key, value) |
|
102 | 102 | |
|
103 | 103 | def get(*arg, **kw): |
|
104 | 104 | key = key_generator(*arg, **kw) |
|
105 | 105 | return self.get(key) |
|
106 | 106 | |
|
107 | 107 | user_func.set = set_ |
|
108 | 108 | user_func.invalidate = invalidate |
|
109 | 109 | user_func.get = get |
|
110 | 110 | user_func.refresh = refresh |
|
111 | 111 | user_func.key_generator = key_generator |
|
112 | 112 | user_func.original = user_func |
|
113 | 113 | |
|
114 | 114 | # Use `decorate` to preserve the signature of :param:`user_func`. |
|
115 | 115 | return decorator.decorate(user_func, functools.partial( |
|
116 | 116 | get_or_create_for_user_func, key_generator)) |
|
117 | 117 | |
|
118 | 118 | return cache_decorator |
|
119 | 119 | |
|
120 | 120 | |
|
121 | 121 | def make_region(*arg, **kw): |
|
122 | 122 | return RhodeCodeCacheRegion(*arg, **kw) |
|
123 | 123 | |
|
124 | 124 | |
|
125 | 125 | def get_default_cache_settings(settings, prefixes=None): |
|
126 | 126 | prefixes = prefixes or [] |
|
127 | 127 | cache_settings = {} |
|
128 | 128 | for key in settings.keys(): |
|
129 | 129 | for prefix in prefixes: |
|
130 | 130 | if key.startswith(prefix): |
|
131 | 131 | name = key.split(prefix)[1].strip() |
|
132 | 132 | val = settings[key] |
|
133 | 133 | if isinstance(val, str): |
|
134 | 134 | val = val.strip() |
|
135 | 135 | cache_settings[name] = val |
|
136 | 136 | return cache_settings |
|
137 | 137 | |
|
138 | 138 | |
|
139 | 139 | def compute_key_from_params(*args): |
|
140 | 140 | """ |
|
141 | 141 | Helper to compute key from given params to be used in cache manager |
|
142 | 142 | """ |
|
143 | 143 | return sha1(safe_bytes("_".join(map(str, args)))) |
|
144 | 144 | |
|
145 | 145 | |
|
146 | 146 | def custom_key_generator(backend, namespace, fn): |
|
147 | 147 | func_name = fn.__name__ |
|
148 | 148 | |
|
149 | 149 | def generate_key(*args): |
|
150 | 150 | backend_pref = getattr(backend, 'key_prefix', None) or 'backend_prefix' |
|
151 | 151 | namespace_pref = namespace or 'default_namespace' |
|
152 | 152 | arg_key = compute_key_from_params(*args) |
|
153 | 153 | final_key = f"{backend_pref}:{namespace_pref}:{func_name}_{arg_key}" |
|
154 | 154 | |
|
155 | 155 | return final_key |
|
156 | 156 | |
|
157 | 157 | return generate_key |
|
158 | 158 | |
|
159 | 159 | |
|
160 | 160 | def backend_key_generator(backend): |
|
161 | 161 | """ |
|
162 | 162 | Special wrapper that also sends over the backend to the key generator |
|
163 | 163 | """ |
|
164 | 164 | def wrapper(namespace, fn): |
|
165 | 165 | return custom_key_generator(backend, namespace, fn) |
|
166 | 166 | return wrapper |
|
167 | 167 | |
|
168 | 168 | |
|
169 | 169 | def get_or_create_region(region_name, region_namespace: str = None, use_async_runner=False): |
|
170 | 170 | from .backends import FileNamespaceBackend |
|
171 | 171 | from . import async_creation_runner |
|
172 | 172 | |
|
173 | 173 | region_obj = region_meta.dogpile_cache_regions.get(region_name) |
|
174 | 174 | if not region_obj: |
|
175 | 175 | reg_keys = list(region_meta.dogpile_cache_regions.keys()) |
|
176 | 176 | raise OSError(f'Region `{region_name}` not in configured: {reg_keys}.') |
|
177 | 177 | |
|
178 | 178 | region_uid_name = f'{region_name}:{region_namespace}' |
|
179 | 179 | |
|
180 | 180 | # Special case for ONLY the FileNamespaceBackend backend. We register one-file-per-region |
|
181 | 181 | if isinstance(region_obj.actual_backend, FileNamespaceBackend): |
|
182 | 182 | if not region_namespace: |
|
183 | 183 | raise ValueError(f'{FileNamespaceBackend} used requires to specify region_namespace param') |
|
184 | 184 | |
|
185 | 185 | region_exist = region_meta.dogpile_cache_regions.get(region_namespace) |
|
186 | 186 | if region_exist: |
|
187 | 187 | log.debug('Using already configured region: %s', region_namespace) |
|
188 | 188 | return region_exist |
|
189 | 189 | |
|
190 | 190 | expiration_time = region_obj.expiration_time |
|
191 | 191 | |
|
192 | 192 | cache_dir = region_meta.dogpile_config_defaults['cache_dir'] |
|
193 | 193 | namespace_cache_dir = cache_dir |
|
194 | 194 | |
|
195 | 195 | # we default the namespace_cache_dir to our default cache dir. |
|
196 | 196 | # however, if this backend is configured with filename= param, we prioritize that |
|
197 | 197 | # so all caches within that particular region, even those namespaced end up in the same path |
|
198 | 198 | if region_obj.actual_backend.filename: |
|
199 | 199 | namespace_cache_dir = os.path.dirname(region_obj.actual_backend.filename) |
|
200 | 200 | |
|
201 | 201 | if not os.path.isdir(namespace_cache_dir): |
|
202 | 202 | os.makedirs(namespace_cache_dir) |
|
203 | 203 | new_region = make_region( |
|
204 | 204 | name=region_uid_name, |
|
205 | 205 | function_key_generator=backend_key_generator(region_obj.actual_backend) |
|
206 | 206 | ) |
|
207 | 207 | |
|
208 | 208 | namespace_filename = os.path.join( |
|
209 | 209 | namespace_cache_dir, f"{region_name}_{region_namespace}.cache_db") |
|
210 | 210 | # special type that allows 1db per namespace |
|
211 | 211 | new_region.configure( |
|
212 | 212 | backend='dogpile.cache.rc.file_namespace', |
|
213 | 213 | expiration_time=expiration_time, |
|
214 | 214 | arguments={"filename": namespace_filename} |
|
215 | 215 | ) |
|
216 | 216 | |
|
217 | 217 | # create and save in region caches |
|
218 | 218 | log.debug('configuring new region: %s', region_uid_name) |
|
219 | 219 | region_obj = region_meta.dogpile_cache_regions[region_namespace] = new_region |
|
220 | 220 | |
|
221 | 221 | region_obj._default_namespace = region_namespace |
|
222 | 222 | if use_async_runner: |
|
223 | 223 | region_obj.async_creation_runner = async_creation_runner |
|
224 | 224 | return region_obj |
|
225 | 225 | |
|
226 | 226 | |
|
227 | 227 | def clear_cache_namespace(cache_region: str | RhodeCodeCacheRegion, cache_namespace_uid: str, method: str) -> int: |
|
228 | 228 | from . import CLEAR_DELETE, CLEAR_INVALIDATE |
|
229 | 229 | |
|
230 | 230 | if not isinstance(cache_region, RhodeCodeCacheRegion): |
|
231 | 231 | cache_region = get_or_create_region(cache_region, cache_namespace_uid) |
|
232 | 232 | log.debug('clearing cache region: %s [prefix:%s] with method=%s', |
|
233 | 233 | cache_region, cache_namespace_uid, method) |
|
234 | 234 | |
|
235 | 235 | num_affected_keys = 0 |
|
236 | 236 | |
|
237 | 237 | if method == CLEAR_INVALIDATE: |
|
238 | 238 | # NOTE: The CacheRegion.invalidate() methodβs default mode of |
|
239 | 239 | # operation is to set a timestamp local to this CacheRegion in this Python process only. |
|
240 | 240 | # It does not impact other Python processes or regions as the timestamp is only stored locally in memory. |
|
241 | 241 | cache_region.invalidate(hard=True) |
|
242 | 242 | |
|
243 | 243 | if method == CLEAR_DELETE: |
|
244 | 244 | num_affected_keys = cache_region.backend.delete_multi_by_prefix(prefix=cache_namespace_uid) |
|
245 | 245 | return num_affected_keys |
@@ -1,158 +1,158 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import typing |
|
19 | 19 | import base64 |
|
20 | 20 | import logging |
|
21 | 21 | |
|
22 | 22 | |
|
23 | 23 | log = logging.getLogger(__name__) |
|
24 | 24 | |
|
25 | 25 | |
|
26 | 26 | def safe_int(val, default=None) -> int: |
|
27 | 27 | """ |
|
28 | 28 | Returns int() of val if val is not convertable to int use default |
|
29 | 29 | instead |
|
30 | 30 | |
|
31 | 31 | :param val: |
|
32 | 32 | :param default: |
|
33 | 33 | """ |
|
34 | 34 | |
|
35 | 35 | try: |
|
36 | 36 | val = int(val) |
|
37 | 37 | except (ValueError, TypeError): |
|
38 | 38 | val = default |
|
39 | 39 | |
|
40 | 40 | return val |
|
41 | 41 | |
|
42 | 42 | |
|
43 | 43 | def base64_to_str(text) -> str: |
|
44 | 44 | return safe_str(base64.encodebytes(safe_bytes(text))).strip() |
|
45 | 45 | |
|
46 | 46 | |
|
47 | 47 | def get_default_encodings() -> list[str]: |
|
48 | 48 | return ['utf8'] |
|
49 | 49 | |
|
50 | 50 | |
|
51 | 51 | def safe_str(str_, to_encoding=None) -> str: |
|
52 | 52 | """ |
|
53 | 53 | safe str function. Does few trick to turn unicode_ into string |
|
54 | 54 | |
|
55 | 55 | :param str_: str to encode |
|
56 | 56 | :param to_encoding: encode to this type UTF8 default |
|
57 | 57 | """ |
|
58 | 58 | if isinstance(str_, str): |
|
59 | 59 | return str_ |
|
60 | 60 | |
|
61 | 61 | # if it's bytes cast to str |
|
62 | 62 | if not isinstance(str_, bytes): |
|
63 | 63 | return str(str_) |
|
64 | 64 | |
|
65 | 65 | to_encoding = to_encoding or get_default_encodings() |
|
66 | 66 | if not isinstance(to_encoding, (list, tuple)): |
|
67 | 67 | to_encoding = [to_encoding] |
|
68 | 68 | |
|
69 | 69 | for enc in to_encoding: |
|
70 | 70 | try: |
|
71 | 71 | return str(str_, enc) |
|
72 | 72 | except UnicodeDecodeError: |
|
73 | 73 | pass |
|
74 | 74 | |
|
75 | 75 | return str(str_, to_encoding[0], 'replace') |
|
76 | 76 | |
|
77 | 77 | |
|
78 | 78 | def safe_bytes(str_, from_encoding=None) -> bytes: |
|
79 | 79 | """ |
|
80 | 80 | safe bytes function. Does few trick to turn str_ into bytes string: |
|
81 | 81 | |
|
82 | 82 | :param str_: string to decode |
|
83 | 83 | :param from_encoding: encode from this type UTF8 default |
|
84 | 84 | """ |
|
85 | 85 | if isinstance(str_, bytes): |
|
86 | 86 | return str_ |
|
87 | 87 | |
|
88 | 88 | if not isinstance(str_, str): |
|
89 | 89 | raise ValueError(f'safe_bytes cannot convert other types than str: got: {type(str_)}') |
|
90 | 90 | |
|
91 | 91 | from_encoding = from_encoding or get_default_encodings() |
|
92 | 92 | if not isinstance(from_encoding, (list, tuple)): |
|
93 | 93 | from_encoding = [from_encoding] |
|
94 | 94 | |
|
95 | 95 | for enc in from_encoding: |
|
96 | 96 | try: |
|
97 | 97 | return str_.encode(enc) |
|
98 | 98 | except UnicodeDecodeError: |
|
99 | 99 | pass |
|
100 | 100 | |
|
101 | 101 | return str_.encode(from_encoding[0], 'replace') |
|
102 | 102 | |
|
103 | 103 | |
|
104 | 104 | def ascii_bytes(str_, allow_bytes=False) -> bytes: |
|
105 | 105 | """ |
|
106 | 106 | Simple conversion from str to bytes, with assumption that str_ is pure ASCII. |
|
107 | 107 | Fails with UnicodeError on invalid input. |
|
108 | 108 | This should be used where encoding and "safe" ambiguity should be avoided. |
|
109 | 109 | Where strings already have been encoded in other ways but still are unicode |
|
110 | 110 | string - for example to hex, base64, json, urlencoding, or are known to be |
|
111 | 111 | identifiers. |
|
112 | 112 | """ |
|
113 | 113 | if allow_bytes and isinstance(str_, bytes): |
|
114 | 114 | return str_ |
|
115 | 115 | |
|
116 | 116 | if not isinstance(str_, str): |
|
117 | 117 | raise ValueError(f'ascii_bytes cannot convert other types than str: got: {type(str_)}') |
|
118 | 118 | return str_.encode('ascii') |
|
119 | 119 | |
|
120 | 120 | |
|
121 | 121 | def ascii_str(str_) -> str: |
|
122 | 122 | """ |
|
123 | 123 | Simple conversion from bytes to str, with assumption that str_ is pure ASCII. |
|
124 | 124 | Fails with UnicodeError on invalid input. |
|
125 | 125 | This should be used where encoding and "safe" ambiguity should be avoided. |
|
126 | 126 | Where strings are encoded but also in other ways are known to be ASCII, and |
|
127 | 127 | where a unicode string is wanted without caring about encoding. For example |
|
128 | 128 | to hex, base64, urlencoding, or are known to be identifiers. |
|
129 | 129 | """ |
|
130 | 130 | |
|
131 | 131 | if not isinstance(str_, bytes): |
|
132 | 132 | raise ValueError(f'ascii_str cannot convert other types than bytes: got: {type(str_)}') |
|
133 | 133 | return str_.decode('ascii') |
|
134 | 134 | |
|
135 | 135 | |
|
136 | 136 | def convert_to_str(data): |
|
137 | 137 | if isinstance(data, bytes): |
|
138 | 138 | return safe_str(data) |
|
139 | 139 | elif isinstance(data, tuple): |
|
140 | 140 | return tuple(convert_to_str(item) for item in data) |
|
141 | 141 | elif isinstance(data, list): |
|
142 | 142 | return list(convert_to_str(item) for item in data) |
|
143 | 143 | else: |
|
144 | 144 | return data |
|
145 | 145 | |
|
146 | 146 | |
|
147 | 147 | def splitnewlines(text: bytes): |
|
148 | 148 | """ |
|
149 | 149 | like splitlines, but only split on newlines. |
|
150 | 150 | """ |
|
151 | 151 | |
|
152 | 152 | lines = [_l + b'\n' for _l in text.split(b'\n')] |
|
153 | 153 | if lines: |
|
154 | 154 | if lines[-1] == b'\n': |
|
155 | 155 | lines.pop() |
|
156 | 156 | else: |
|
157 | 157 | lines[-1] = lines[-1][:-1] |
|
158 | return lines No newline at end of file | |
|
158 | return lines |
@@ -1,160 +1,160 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import os |
|
19 | 19 | import tempfile |
|
20 | 20 | |
|
21 | 21 | from svn import client |
|
22 | 22 | from svn import core |
|
23 | 23 | from svn import ra |
|
24 | 24 | |
|
25 | 25 | from mercurial import error |
|
26 | 26 | |
|
27 | from vcsserver.str_utils import safe_bytes | |
|
27 | from vcsserver.lib.str_utils import safe_bytes | |
|
28 | 28 | |
|
29 | 29 | core.svn_config_ensure(None) |
|
30 | 30 | svn_config = core.svn_config_get_config(None) |
|
31 | 31 | |
|
32 | 32 | |
|
33 | 33 | class RaCallbacks(ra.Callbacks): |
|
34 | 34 | @staticmethod |
|
35 | 35 | def open_tmp_file(pool): # pragma: no cover |
|
36 | 36 | (fd, fn) = tempfile.mkstemp() |
|
37 | 37 | os.close(fd) |
|
38 | 38 | return fn |
|
39 | 39 | |
|
40 | 40 | @staticmethod |
|
41 | 41 | def get_client_string(pool): |
|
42 | 42 | return b'RhodeCode-subversion-url-checker' |
|
43 | 43 | |
|
44 | 44 | |
|
45 | 45 | class SubversionException(Exception): |
|
46 | 46 | pass |
|
47 | 47 | |
|
48 | 48 | |
|
49 | 49 | class SubversionConnectionException(SubversionException): |
|
50 | 50 | """Exception raised when a generic error occurs when connecting to a repository.""" |
|
51 | 51 | |
|
52 | 52 | |
|
53 | 53 | def normalize_url(url): |
|
54 | 54 | if not url: |
|
55 | 55 | return url |
|
56 | 56 | if url.startswith(b'svn+http://') or url.startswith(b'svn+https://'): |
|
57 | 57 | url = url[4:] |
|
58 | 58 | url = url.rstrip(b'/') |
|
59 | 59 | return url |
|
60 | 60 | |
|
61 | 61 | |
|
62 | 62 | def _create_auth_baton(pool): |
|
63 | 63 | """Create a Subversion authentication baton. """ |
|
64 | 64 | # Give the client context baton a suite of authentication |
|
65 | 65 | # providers.h |
|
66 | 66 | platform_specific = [ |
|
67 | 67 | 'svn_auth_get_gnome_keyring_simple_provider', |
|
68 | 68 | 'svn_auth_get_gnome_keyring_ssl_client_cert_pw_provider', |
|
69 | 69 | 'svn_auth_get_keychain_simple_provider', |
|
70 | 70 | 'svn_auth_get_keychain_ssl_client_cert_pw_provider', |
|
71 | 71 | 'svn_auth_get_kwallet_simple_provider', |
|
72 | 72 | 'svn_auth_get_kwallet_ssl_client_cert_pw_provider', |
|
73 | 73 | 'svn_auth_get_ssl_client_cert_file_provider', |
|
74 | 74 | 'svn_auth_get_windows_simple_provider', |
|
75 | 75 | 'svn_auth_get_windows_ssl_server_trust_provider', |
|
76 | 76 | ] |
|
77 | 77 | |
|
78 | 78 | providers = [] |
|
79 | 79 | |
|
80 | 80 | for p in platform_specific: |
|
81 | 81 | if getattr(core, p, None) is not None: |
|
82 | 82 | try: |
|
83 | 83 | providers.append(getattr(core, p)()) |
|
84 | 84 | except RuntimeError: |
|
85 | 85 | pass |
|
86 | 86 | |
|
87 | 87 | providers += [ |
|
88 | 88 | client.get_simple_provider(), |
|
89 | 89 | client.get_username_provider(), |
|
90 | 90 | client.get_ssl_client_cert_file_provider(), |
|
91 | 91 | client.get_ssl_client_cert_pw_file_provider(), |
|
92 | 92 | client.get_ssl_server_trust_file_provider(), |
|
93 | 93 | ] |
|
94 | 94 | |
|
95 | 95 | return core.svn_auth_open(providers, pool) |
|
96 | 96 | |
|
97 | 97 | |
|
98 | 98 | class SubversionRepo: |
|
99 | 99 | """Wrapper for a Subversion repository. |
|
100 | 100 | |
|
101 | 101 | It uses the SWIG Python bindings, see above for requirements. |
|
102 | 102 | """ |
|
103 | 103 | def __init__(self, svn_url: bytes = b'', username: bytes = b'', password: bytes = b''): |
|
104 | 104 | |
|
105 | 105 | self.username = username |
|
106 | 106 | self.password = password |
|
107 | 107 | self.svn_url = core.svn_path_canonicalize(svn_url) |
|
108 | 108 | |
|
109 | 109 | self.auth_baton_pool = core.Pool() |
|
110 | 110 | self.auth_baton = _create_auth_baton(self.auth_baton_pool) |
|
111 | 111 | # self.init_ra_and_client() assumes that a pool already exists |
|
112 | 112 | self.pool = core.Pool() |
|
113 | 113 | |
|
114 | 114 | self.ra = self.init_ra_and_client() |
|
115 | 115 | self.uuid = ra.get_uuid(self.ra, self.pool) |
|
116 | 116 | |
|
117 | 117 | def init_ra_and_client(self): |
|
118 | 118 | """Initializes the RA and client layers, because sometimes getting |
|
119 | 119 | unified diffs runs the remote server out of open files. |
|
120 | 120 | """ |
|
121 | 121 | |
|
122 | 122 | if self.username: |
|
123 | 123 | core.svn_auth_set_parameter(self.auth_baton, |
|
124 | 124 | core.SVN_AUTH_PARAM_DEFAULT_USERNAME, |
|
125 | 125 | self.username) |
|
126 | 126 | if self.password: |
|
127 | 127 | core.svn_auth_set_parameter(self.auth_baton, |
|
128 | 128 | core.SVN_AUTH_PARAM_DEFAULT_PASSWORD, |
|
129 | 129 | self.password) |
|
130 | 130 | |
|
131 | 131 | callbacks = RaCallbacks() |
|
132 | 132 | callbacks.auth_baton = self.auth_baton |
|
133 | 133 | |
|
134 | 134 | try: |
|
135 | 135 | return ra.open2(self.svn_url, callbacks, svn_config, self.pool) |
|
136 | 136 | except SubversionException as e: |
|
137 | 137 | # e.child contains a detailed error messages |
|
138 | 138 | msglist = [] |
|
139 | 139 | svn_exc = e |
|
140 | 140 | while svn_exc: |
|
141 | 141 | if svn_exc.args[0]: |
|
142 | 142 | msglist.append(svn_exc.args[0]) |
|
143 | 143 | svn_exc = svn_exc.child |
|
144 | 144 | msg = '\n'.join(msglist) |
|
145 | 145 | raise SubversionConnectionException(msg) |
|
146 | 146 | |
|
147 | 147 | |
|
148 | 148 | class svnremoterepo: |
|
149 | 149 | """ the dumb wrapper for actual Subversion repositories """ |
|
150 | 150 | |
|
151 | 151 | def __init__(self, username: bytes = b'', password: bytes = b'', svn_url: bytes = b''): |
|
152 | 152 | self.username = username or b'' |
|
153 | 153 | self.password = password or b'' |
|
154 | 154 | self.path = normalize_url(svn_url) |
|
155 | 155 | |
|
156 | 156 | def svn(self): |
|
157 | 157 | try: |
|
158 | 158 | return SubversionRepo(self.path, self.username, self.password) |
|
159 | 159 | except SubversionConnectionException as e: |
|
160 | 160 | raise error.Abort(safe_bytes(e)) |
|
1 | NO CONTENT: file renamed from vcsserver/type_utils.py to vcsserver/lib/type_utils.py |
@@ -1,417 +1,417 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | """Handles the Git smart protocol.""" |
|
19 | 19 | |
|
20 | 20 | import os |
|
21 | 21 | import socket |
|
22 | 22 | import logging |
|
23 | 23 | |
|
24 | 24 | import dulwich.protocol |
|
25 | 25 | from dulwich.protocol import CAPABILITY_SIDE_BAND, CAPABILITY_SIDE_BAND_64K |
|
26 | 26 | from webob import Request, Response, exc |
|
27 | 27 | |
|
28 |
from vcsserver.lib. |
|
|
28 | from vcsserver.lib.ext_json import json | |
|
29 | 29 | from vcsserver import hooks, subprocessio |
|
30 | from vcsserver.str_utils import ascii_bytes | |
|
30 | from vcsserver.lib.str_utils import ascii_bytes | |
|
31 | 31 | |
|
32 | 32 | |
|
33 | 33 | log = logging.getLogger(__name__) |
|
34 | 34 | |
|
35 | 35 | |
|
36 | 36 | class FileWrapper: |
|
37 | 37 | """File wrapper that ensures how much data is read from it.""" |
|
38 | 38 | |
|
39 | 39 | def __init__(self, fd, content_length): |
|
40 | 40 | self.fd = fd |
|
41 | 41 | self.content_length = content_length |
|
42 | 42 | self.remain = content_length |
|
43 | 43 | |
|
44 | 44 | def read(self, size): |
|
45 | 45 | if size <= self.remain: |
|
46 | 46 | try: |
|
47 | 47 | data = self.fd.read(size) |
|
48 | 48 | except socket.error: |
|
49 | 49 | raise IOError(self) |
|
50 | 50 | self.remain -= size |
|
51 | 51 | elif self.remain: |
|
52 | 52 | data = self.fd.read(self.remain) |
|
53 | 53 | self.remain = 0 |
|
54 | 54 | else: |
|
55 | 55 | data = None |
|
56 | 56 | return data |
|
57 | 57 | |
|
58 | 58 | def __repr__(self): |
|
59 | 59 | return '<FileWrapper {} len: {}, read: {}>'.format( |
|
60 | 60 | self.fd, self.content_length, self.content_length - self.remain |
|
61 | 61 | ) |
|
62 | 62 | |
|
63 | 63 | |
|
64 | 64 | class GitRepository: |
|
65 | 65 | """WSGI app for handling Git smart protocol endpoints.""" |
|
66 | 66 | |
|
67 | 67 | git_folder_signature = frozenset(('config', 'head', 'info', 'objects', 'refs')) |
|
68 | 68 | commands = frozenset(('git-upload-pack', 'git-receive-pack')) |
|
69 | 69 | valid_accepts = frozenset(f'application/x-{c}-result' for c in commands) |
|
70 | 70 | |
|
71 | 71 | # The last bytes are the SHA1 of the first 12 bytes. |
|
72 | 72 | EMPTY_PACK = ( |
|
73 | 73 | b'PACK\x00\x00\x00\x02\x00\x00\x00\x00\x02\x9d\x08' + |
|
74 | 74 | b'\x82;\xd8\xa8\xea\xb5\x10\xadj\xc7\\\x82<\xfd>\xd3\x1e' |
|
75 | 75 | ) |
|
76 | 76 | FLUSH_PACKET = b"0000" |
|
77 | 77 | |
|
78 | 78 | SIDE_BAND_CAPS = frozenset((CAPABILITY_SIDE_BAND, CAPABILITY_SIDE_BAND_64K)) |
|
79 | 79 | |
|
80 | 80 | def __init__(self, repo_name, content_path, git_path, update_server_info, extras): |
|
81 | 81 | files = frozenset(f.lower() for f in os.listdir(content_path)) |
|
82 | 82 | valid_dir_signature = self.git_folder_signature.issubset(files) |
|
83 | 83 | |
|
84 | 84 | if not valid_dir_signature: |
|
85 | 85 | raise OSError(f'{content_path} missing git signature') |
|
86 | 86 | |
|
87 | 87 | self.content_path = content_path |
|
88 | 88 | self.repo_name = repo_name |
|
89 | 89 | self.extras = extras |
|
90 | 90 | self.git_path = git_path |
|
91 | 91 | self.update_server_info = update_server_info |
|
92 | 92 | |
|
93 | 93 | def _get_fixedpath(self, path): |
|
94 | 94 | """ |
|
95 | 95 | Small fix for repo_path |
|
96 | 96 | |
|
97 | 97 | :param path: |
|
98 | 98 | """ |
|
99 | 99 | path = path.split(self.repo_name, 1)[-1] |
|
100 | 100 | if path.startswith('.git'): |
|
101 | 101 | # for bare repos we still get the .git prefix inside, we skip it |
|
102 | 102 | # here, and remove from the service command |
|
103 | 103 | path = path[4:] |
|
104 | 104 | |
|
105 | 105 | return path.strip('/') |
|
106 | 106 | |
|
107 | 107 | def inforefs(self, request, unused_environ): |
|
108 | 108 | """ |
|
109 | 109 | WSGI Response producer for HTTP GET Git Smart |
|
110 | 110 | HTTP /info/refs request. |
|
111 | 111 | """ |
|
112 | 112 | |
|
113 | 113 | git_command = request.GET.get('service') |
|
114 | 114 | if git_command not in self.commands: |
|
115 | 115 | log.debug('command %s not allowed', git_command) |
|
116 | 116 | return exc.HTTPForbidden() |
|
117 | 117 | |
|
118 | 118 | # please, resist the urge to add '\n' to git capture and increment |
|
119 | 119 | # line count by 1. |
|
120 | 120 | # by git docs: Documentation/technical/http-protocol.txt#L214 \n is |
|
121 | 121 | # a part of protocol. |
|
122 | 122 | # The code in Git client not only does NOT need '\n', but actually |
|
123 | 123 | # blows up if you sprinkle "flush" (0000) as "0001\n". |
|
124 | 124 | # It reads binary, per number of bytes specified. |
|
125 | 125 | # if you do add '\n' as part of data, count it. |
|
126 | 126 | server_advert = f'# service={git_command}\n' |
|
127 | 127 | packet_len = hex(len(server_advert) + 4)[2:].rjust(4, '0').lower() |
|
128 | 128 | try: |
|
129 | 129 | gitenv = dict(os.environ) |
|
130 | 130 | # forget all configs |
|
131 | 131 | gitenv['RC_SCM_DATA'] = json.dumps(self.extras) |
|
132 | 132 | command = [self.git_path, git_command[4:], '--stateless-rpc', |
|
133 | 133 | '--advertise-refs', self.content_path] |
|
134 | 134 | out = subprocessio.SubprocessIOChunker( |
|
135 | 135 | command, |
|
136 | 136 | env=gitenv, |
|
137 | 137 | starting_values=[ascii_bytes(packet_len + server_advert) + self.FLUSH_PACKET], |
|
138 | 138 | shell=False |
|
139 | 139 | ) |
|
140 | 140 | except OSError: |
|
141 | 141 | log.exception('Error processing command') |
|
142 | 142 | raise exc.HTTPExpectationFailed() |
|
143 | 143 | |
|
144 | 144 | resp = Response() |
|
145 | 145 | resp.content_type = f'application/x-{git_command}-advertisement' |
|
146 | 146 | resp.charset = None |
|
147 | 147 | resp.app_iter = out |
|
148 | 148 | |
|
149 | 149 | return resp |
|
150 | 150 | |
|
151 | 151 | def _get_want_capabilities(self, request): |
|
152 | 152 | """Read the capabilities found in the first want line of the request.""" |
|
153 | 153 | pos = request.body_file_seekable.tell() |
|
154 | 154 | first_line = request.body_file_seekable.readline() |
|
155 | 155 | request.body_file_seekable.seek(pos) |
|
156 | 156 | |
|
157 | 157 | return frozenset( |
|
158 | 158 | dulwich.protocol.extract_want_line_capabilities(first_line)[1]) |
|
159 | 159 | |
|
160 | 160 | def _build_failed_pre_pull_response(self, capabilities, pre_pull_messages): |
|
161 | 161 | """ |
|
162 | 162 | Construct a response with an empty PACK file. |
|
163 | 163 | |
|
164 | 164 | We use an empty PACK file, as that would trigger the failure of the pull |
|
165 | 165 | or clone command. |
|
166 | 166 | |
|
167 | 167 | We also print in the error output a message explaining why the command |
|
168 | 168 | was aborted. |
|
169 | 169 | |
|
170 | 170 | If additionally, the user is accepting messages we send them the output |
|
171 | 171 | of the pre-pull hook. |
|
172 | 172 | |
|
173 | 173 | Note that for clients not supporting side-band we just send them the |
|
174 | 174 | emtpy PACK file. |
|
175 | 175 | """ |
|
176 | 176 | |
|
177 | 177 | if self.SIDE_BAND_CAPS.intersection(capabilities): |
|
178 | 178 | response = [] |
|
179 | 179 | proto = dulwich.protocol.Protocol(None, response.append) |
|
180 | 180 | proto.write_pkt_line(dulwich.protocol.NAK_LINE) |
|
181 | 181 | |
|
182 | 182 | self._write_sideband_to_proto(proto, ascii_bytes(pre_pull_messages, allow_bytes=True), capabilities) |
|
183 | 183 | # N.B.(skreft): Do not change the sideband channel to 3, as that |
|
184 | 184 | # produces a fatal error in the client: |
|
185 | 185 | # fatal: error in sideband demultiplexer |
|
186 | 186 | proto.write_sideband( |
|
187 | 187 | dulwich.protocol.SIDE_BAND_CHANNEL_PROGRESS, |
|
188 | 188 | ascii_bytes('Pre pull hook failed: aborting\n', allow_bytes=True)) |
|
189 | 189 | proto.write_sideband( |
|
190 | 190 | dulwich.protocol.SIDE_BAND_CHANNEL_DATA, |
|
191 | 191 | ascii_bytes(self.EMPTY_PACK, allow_bytes=True)) |
|
192 | 192 | |
|
193 | 193 | # writes b"0000" as default |
|
194 | 194 | proto.write_pkt_line(None) |
|
195 | 195 | |
|
196 | 196 | return response |
|
197 | 197 | else: |
|
198 | 198 | return [ascii_bytes(self.EMPTY_PACK, allow_bytes=True)] |
|
199 | 199 | |
|
200 | 200 | def _build_post_pull_response(self, response, capabilities, start_message, end_message): |
|
201 | 201 | """ |
|
202 | 202 | Given a list response we inject the post-pull messages. |
|
203 | 203 | |
|
204 | 204 | We only inject the messages if the client supports sideband, and the |
|
205 | 205 | response has the format: |
|
206 | 206 | 0008NAK\n...0000 |
|
207 | 207 | |
|
208 | 208 | Note that we do not check the no-progress capability as by default, git |
|
209 | 209 | sends it, which effectively would block all messages. |
|
210 | 210 | """ |
|
211 | 211 | |
|
212 | 212 | if not self.SIDE_BAND_CAPS.intersection(capabilities): |
|
213 | 213 | return response |
|
214 | 214 | |
|
215 | 215 | if not start_message and not end_message: |
|
216 | 216 | return response |
|
217 | 217 | |
|
218 | 218 | try: |
|
219 | 219 | iter(response) |
|
220 | 220 | # iterator probably will work, we continue |
|
221 | 221 | except TypeError: |
|
222 | 222 | raise TypeError(f'response must be an iterator: got {type(response)}') |
|
223 | 223 | if isinstance(response, (list, tuple)): |
|
224 | 224 | raise TypeError(f'response must be an iterator: got {type(response)}') |
|
225 | 225 | |
|
226 | 226 | def injected_response(): |
|
227 | 227 | |
|
228 | 228 | do_loop = 1 |
|
229 | 229 | header_injected = 0 |
|
230 | 230 | next_item = None |
|
231 | 231 | has_item = False |
|
232 | 232 | item = b'' |
|
233 | 233 | |
|
234 | 234 | while do_loop: |
|
235 | 235 | |
|
236 | 236 | try: |
|
237 | 237 | next_item = next(response) |
|
238 | 238 | except StopIteration: |
|
239 | 239 | do_loop = 0 |
|
240 | 240 | |
|
241 | 241 | if has_item: |
|
242 | 242 | # last item ! alter it now |
|
243 | 243 | if do_loop == 0 and item.endswith(self.FLUSH_PACKET): |
|
244 | 244 | new_response = [item[:-4]] |
|
245 | 245 | new_response.extend(self._get_messages(end_message, capabilities)) |
|
246 | 246 | new_response.append(self.FLUSH_PACKET) |
|
247 | 247 | item = b''.join(new_response) |
|
248 | 248 | |
|
249 | 249 | yield item |
|
250 | 250 | |
|
251 | 251 | has_item = True |
|
252 | 252 | item = next_item |
|
253 | 253 | |
|
254 | 254 | # alter item if it's the initial chunk |
|
255 | 255 | if not header_injected and item.startswith(b'0008NAK\n'): |
|
256 | 256 | new_response = [b'0008NAK\n'] |
|
257 | 257 | new_response.extend(self._get_messages(start_message, capabilities)) |
|
258 | 258 | new_response.append(item[8:]) |
|
259 | 259 | item = b''.join(new_response) |
|
260 | 260 | header_injected = 1 |
|
261 | 261 | |
|
262 | 262 | return injected_response() |
|
263 | 263 | |
|
264 | 264 | def _write_sideband_to_proto(self, proto, data, capabilities): |
|
265 | 265 | """ |
|
266 | 266 | Write the data to the proto's sideband number 2 == SIDE_BAND_CHANNEL_PROGRESS |
|
267 | 267 | |
|
268 | 268 | We do not use dulwich's write_sideband directly as it only supports |
|
269 | 269 | side-band-64k. |
|
270 | 270 | """ |
|
271 | 271 | if not data: |
|
272 | 272 | return |
|
273 | 273 | |
|
274 | 274 | # N.B.(skreft): The values below are explained in the pack protocol |
|
275 | 275 | # documentation, section Packfile Data. |
|
276 | 276 | # https://github.com/git/git/blob/master/Documentation/technical/pack-protocol.txt |
|
277 | 277 | if CAPABILITY_SIDE_BAND_64K in capabilities: |
|
278 | 278 | chunk_size = 65515 |
|
279 | 279 | elif CAPABILITY_SIDE_BAND in capabilities: |
|
280 | 280 | chunk_size = 995 |
|
281 | 281 | else: |
|
282 | 282 | return |
|
283 | 283 | |
|
284 | 284 | chunker = (data[i:i + chunk_size] for i in range(0, len(data), chunk_size)) |
|
285 | 285 | |
|
286 | 286 | for chunk in chunker: |
|
287 | 287 | proto.write_sideband(dulwich.protocol.SIDE_BAND_CHANNEL_PROGRESS, ascii_bytes(chunk, allow_bytes=True)) |
|
288 | 288 | |
|
289 | 289 | def _get_messages(self, data, capabilities): |
|
290 | 290 | """Return a list with packets for sending data in sideband number 2.""" |
|
291 | 291 | response = [] |
|
292 | 292 | proto = dulwich.protocol.Protocol(None, response.append) |
|
293 | 293 | |
|
294 | 294 | self._write_sideband_to_proto(proto, data, capabilities) |
|
295 | 295 | |
|
296 | 296 | return response |
|
297 | 297 | |
|
298 | 298 | def backend(self, request, environ): |
|
299 | 299 | """ |
|
300 | 300 | WSGI Response producer for HTTP POST Git Smart HTTP requests. |
|
301 | 301 | Reads commands and data from HTTP POST's body. |
|
302 | 302 | returns an iterator obj with contents of git command's |
|
303 | 303 | response to stdout |
|
304 | 304 | """ |
|
305 | 305 | # TODO(skreft): think how we could detect an HTTPLockedException, as |
|
306 | 306 | # we probably want to have the same mechanism used by mercurial and |
|
307 | 307 | # simplevcs. |
|
308 | 308 | # For that we would need to parse the output of the command looking for |
|
309 | 309 | # some signs of the HTTPLockedError, parse the data and reraise it in |
|
310 | 310 | # pygrack. However, that would interfere with the streaming. |
|
311 | 311 | # |
|
312 | 312 | # Now the output of a blocked push is: |
|
313 | 313 | # Pushing to http://test_regular:test12@127.0.0.1:5001/vcs_test_git |
|
314 | 314 | # POST git-receive-pack (1047 bytes) |
|
315 | 315 | # remote: ERROR: Repository `vcs_test_git` locked by user `test_admin`. Reason:`lock_auto` |
|
316 | 316 | # To http://test_regular:test12@127.0.0.1:5001/vcs_test_git |
|
317 | 317 | # ! [remote rejected] master -> master (pre-receive hook declined) |
|
318 | 318 | # error: failed to push some refs to 'http://test_regular:test12@127.0.0.1:5001/vcs_test_git' |
|
319 | 319 | |
|
320 | 320 | git_command = self._get_fixedpath(request.path_info) |
|
321 | 321 | if git_command not in self.commands: |
|
322 | 322 | log.debug('command %s not allowed', git_command) |
|
323 | 323 | return exc.HTTPForbidden() |
|
324 | 324 | |
|
325 | 325 | capabilities = None |
|
326 | 326 | if git_command == 'git-upload-pack': |
|
327 | 327 | capabilities = self._get_want_capabilities(request) |
|
328 | 328 | |
|
329 | 329 | if 'CONTENT_LENGTH' in environ: |
|
330 | 330 | inputstream = FileWrapper(request.body_file_seekable, |
|
331 | 331 | request.content_length) |
|
332 | 332 | else: |
|
333 | 333 | inputstream = request.body_file_seekable |
|
334 | 334 | |
|
335 | 335 | resp = Response() |
|
336 | 336 | resp.content_type = f'application/x-{git_command}-result' |
|
337 | 337 | resp.charset = None |
|
338 | 338 | |
|
339 | 339 | pre_pull_messages = '' |
|
340 | 340 | # Upload-pack == clone |
|
341 | 341 | if git_command == 'git-upload-pack': |
|
342 | 342 | hook_response = hooks.git_pre_pull(self.extras) |
|
343 | 343 | if hook_response.status != 0: |
|
344 | 344 | pre_pull_messages = hook_response.output |
|
345 | 345 | resp.app_iter = self._build_failed_pre_pull_response( |
|
346 | 346 | capabilities, pre_pull_messages) |
|
347 | 347 | return resp |
|
348 | 348 | |
|
349 | 349 | gitenv = dict(os.environ) |
|
350 | 350 | # forget all configs |
|
351 | 351 | gitenv['GIT_CONFIG_NOGLOBAL'] = '1' |
|
352 | 352 | gitenv['RC_SCM_DATA'] = json.dumps(self.extras) |
|
353 | 353 | cmd = [self.git_path, git_command[4:], '--stateless-rpc', |
|
354 | 354 | self.content_path] |
|
355 | 355 | log.debug('handling cmd %s', cmd) |
|
356 | 356 | |
|
357 | 357 | out = subprocessio.SubprocessIOChunker( |
|
358 | 358 | cmd, |
|
359 | 359 | input_stream=inputstream, |
|
360 | 360 | env=gitenv, |
|
361 | 361 | cwd=self.content_path, |
|
362 | 362 | shell=False, |
|
363 | 363 | fail_on_stderr=False, |
|
364 | 364 | fail_on_return_code=False |
|
365 | 365 | ) |
|
366 | 366 | |
|
367 | 367 | if self.update_server_info and git_command == 'git-receive-pack': |
|
368 | 368 | # We need to fully consume the iterator here, as the |
|
369 | 369 | # update-server-info command needs to be run after the push. |
|
370 | 370 | out = list(out) |
|
371 | 371 | |
|
372 | 372 | # Updating refs manually after each push. |
|
373 | 373 | # This is required as some clients are exposing Git repos internally |
|
374 | 374 | # with the dumb protocol. |
|
375 | 375 | cmd = [self.git_path, 'update-server-info'] |
|
376 | 376 | log.debug('handling cmd %s', cmd) |
|
377 | 377 | output = subprocessio.SubprocessIOChunker( |
|
378 | 378 | cmd, |
|
379 | 379 | input_stream=inputstream, |
|
380 | 380 | env=gitenv, |
|
381 | 381 | cwd=self.content_path, |
|
382 | 382 | shell=False, |
|
383 | 383 | fail_on_stderr=False, |
|
384 | 384 | fail_on_return_code=False |
|
385 | 385 | ) |
|
386 | 386 | # Consume all the output so the subprocess finishes |
|
387 | 387 | for _ in output: |
|
388 | 388 | pass |
|
389 | 389 | |
|
390 | 390 | # Upload-pack == clone |
|
391 | 391 | if git_command == 'git-upload-pack': |
|
392 | 392 | hook_response = hooks.git_post_pull(self.extras) |
|
393 | 393 | post_pull_messages = hook_response.output |
|
394 | 394 | resp.app_iter = self._build_post_pull_response(out, capabilities, pre_pull_messages, post_pull_messages) |
|
395 | 395 | else: |
|
396 | 396 | resp.app_iter = out |
|
397 | 397 | |
|
398 | 398 | return resp |
|
399 | 399 | |
|
400 | 400 | def __call__(self, environ, start_response): |
|
401 | 401 | request = Request(environ) |
|
402 | 402 | _path = self._get_fixedpath(request.path_info) |
|
403 | 403 | if _path.startswith('info/refs'): |
|
404 | 404 | app = self.inforefs |
|
405 | 405 | else: |
|
406 | 406 | app = self.backend |
|
407 | 407 | |
|
408 | 408 | try: |
|
409 | 409 | resp = app(request, environ) |
|
410 | 410 | except exc.HTTPException as error: |
|
411 | 411 | log.exception('HTTP Error') |
|
412 | 412 | resp = error |
|
413 | 413 | except Exception: |
|
414 | 414 | log.exception('Unknown error') |
|
415 | 415 | resp = exc.HTTPInternalServerError() |
|
416 | 416 | |
|
417 | 417 | return resp(environ, start_response) |
@@ -1,1519 +1,1526 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import collections |
|
19 | 19 | import logging |
|
20 | 20 | import os |
|
21 | 21 | import re |
|
22 | 22 | import stat |
|
23 | 23 | import traceback |
|
24 | 24 | import urllib.request |
|
25 | 25 | import urllib.parse |
|
26 | 26 | import urllib.error |
|
27 | 27 | from functools import wraps |
|
28 | 28 | |
|
29 | 29 | import more_itertools |
|
30 | 30 | import pygit2 |
|
31 | 31 | from pygit2 import Repository as LibGit2Repo |
|
32 | 32 | from pygit2 import index as LibGit2Index |
|
33 | 33 | from dulwich import index, objects |
|
34 | 34 | from dulwich.client import HttpGitClient, LocalGitClient, FetchPackResult |
|
35 | 35 | from dulwich.errors import ( |
|
36 | 36 | NotGitRepository, ChecksumMismatch, WrongObjectException, |
|
37 | 37 | MissingCommitError, ObjectMissing, HangupException, |
|
38 | 38 | UnexpectedCommandError) |
|
39 | 39 | from dulwich.repo import Repo as DulwichRepo |
|
40 | 40 | |
|
41 | 41 | import rhodecode |
|
42 | 42 | from vcsserver import exceptions, settings, subprocessio |
|
43 | from vcsserver.str_utils import safe_str, safe_int, safe_bytes, ascii_bytes, convert_to_str, splitnewlines | |
|
43 | from vcsserver.lib.str_utils import safe_str, safe_int, safe_bytes, ascii_bytes, convert_to_str, splitnewlines | |
|
44 | 44 | from vcsserver.base import RepoFactory, obfuscate_qs, ArchiveNode, store_archive_in_cache, BytesEnvelope, BinaryEnvelope |
|
45 | 45 | from vcsserver.hgcompat import ( |
|
46 | 46 | hg_url as url_parser, httpbasicauthhandler, httpdigestauthhandler) |
|
47 | 47 | from vcsserver.git_lfs.lib import LFSOidStore |
|
48 | 48 | from vcsserver.vcs_base import RemoteBase |
|
49 | 49 | |
|
50 | 50 | DIR_STAT = stat.S_IFDIR |
|
51 | 51 | FILE_MODE = stat.S_IFMT |
|
52 | 52 | GIT_LINK = objects.S_IFGITLINK |
|
53 | 53 | PEELED_REF_MARKER = b'^{}' |
|
54 | 54 | HEAD_MARKER = b'HEAD' |
|
55 | 55 | |
|
56 | 56 | log = logging.getLogger(__name__) |
|
57 | 57 | |
|
58 | 58 | |
|
59 | 59 | def reraise_safe_exceptions(func): |
|
60 | 60 | """Converts Dulwich exceptions to something neutral.""" |
|
61 | 61 | |
|
62 | 62 | @wraps(func) |
|
63 | 63 | def wrapper(*args, **kwargs): |
|
64 | 64 | try: |
|
65 | 65 | return func(*args, **kwargs) |
|
66 | 66 | except (ChecksumMismatch, WrongObjectException, MissingCommitError, ObjectMissing,) as e: |
|
67 | 67 | exc = exceptions.LookupException(org_exc=e) |
|
68 | 68 | raise exc(safe_str(e)) |
|
69 | 69 | except (HangupException, UnexpectedCommandError) as e: |
|
70 | 70 | exc = exceptions.VcsException(org_exc=e) |
|
71 | 71 | raise exc(safe_str(e)) |
|
72 | 72 | except Exception: |
|
73 | 73 | # NOTE(marcink): because of how dulwich handles some exceptions |
|
74 | 74 | # (KeyError on empty repos), we cannot track this and catch all |
|
75 | 75 | # exceptions, it's an exceptions from other handlers |
|
76 | 76 | #if not hasattr(e, '_vcs_kind'): |
|
77 | 77 | #log.exception("Unhandled exception in git remote call") |
|
78 | 78 | #raise_from_original(exceptions.UnhandledException) |
|
79 | 79 | raise |
|
80 | 80 | return wrapper |
|
81 | 81 | |
|
82 | 82 | |
|
83 | 83 | class Repo(DulwichRepo): |
|
84 | 84 | """ |
|
85 | 85 | A wrapper for dulwich Repo class. |
|
86 | 86 | |
|
87 | 87 | Since dulwich is sometimes keeping .idx file descriptors open, it leads to |
|
88 | 88 | "Too many open files" error. We need to close all opened file descriptors |
|
89 | 89 | once the repo object is destroyed. |
|
90 | 90 | """ |
|
91 | 91 | def __del__(self): |
|
92 | 92 | if hasattr(self, 'object_store'): |
|
93 | 93 | self.close() |
|
94 | 94 | |
|
95 | 95 | |
|
96 | 96 | class Repository(LibGit2Repo): |
|
97 | 97 | |
|
98 | 98 | def __enter__(self): |
|
99 | 99 | return self |
|
100 | 100 | |
|
101 | 101 | def __exit__(self, exc_type, exc_val, exc_tb): |
|
102 | 102 | self.free() |
|
103 | 103 | |
|
104 | 104 | |
|
105 | 105 | class GitFactory(RepoFactory): |
|
106 | 106 | repo_type = 'git' |
|
107 | 107 | |
|
108 | 108 | def _create_repo(self, wire, create, use_libgit2=False): |
|
109 | 109 | if use_libgit2: |
|
110 | 110 | repo = Repository(safe_bytes(wire['path'])) |
|
111 | 111 | else: |
|
112 | 112 | # dulwich mode |
|
113 | 113 | repo_path = safe_str(wire['path'], to_encoding=settings.WIRE_ENCODING) |
|
114 | 114 | repo = Repo(repo_path) |
|
115 | 115 | |
|
116 | 116 | log.debug('repository created: got GIT object: %s', repo) |
|
117 | 117 | return repo |
|
118 | 118 | |
|
119 | 119 | def repo(self, wire, create=False, use_libgit2=False): |
|
120 | 120 | """ |
|
121 | 121 | Get a repository instance for the given path. |
|
122 | 122 | """ |
|
123 | 123 | return self._create_repo(wire, create, use_libgit2) |
|
124 | 124 | |
|
125 | 125 | def repo_libgit2(self, wire): |
|
126 | 126 | return self.repo(wire, use_libgit2=True) |
|
127 | 127 | |
|
128 | 128 | |
|
129 | 129 | def create_signature_from_string(author_str, **kwargs): |
|
130 | 130 | """ |
|
131 | 131 | Creates a pygit2.Signature object from a string of the format 'Name <email>'. |
|
132 | 132 | |
|
133 | 133 | :param author_str: String of the format 'Name <email>' |
|
134 | 134 | :return: pygit2.Signature object |
|
135 | 135 | """ |
|
136 | 136 | match = re.match(r'^(.+) <(.+)>$', author_str) |
|
137 | 137 | if match is None: |
|
138 | 138 | raise ValueError(f"Invalid format: {author_str}") |
|
139 | 139 | |
|
140 | 140 | name, email = match.groups() |
|
141 | 141 | return pygit2.Signature(name, email, **kwargs) |
|
142 | 142 | |
|
143 | 143 | |
|
144 | 144 | def get_obfuscated_url(url_obj): |
|
145 | 145 | url_obj.passwd = b'*****' if url_obj.passwd else url_obj.passwd |
|
146 | 146 | url_obj.query = obfuscate_qs(url_obj.query) |
|
147 | 147 | obfuscated_uri = str(url_obj) |
|
148 | 148 | return obfuscated_uri |
|
149 | 149 | |
|
150 | 150 | |
|
151 | 151 | class GitRemote(RemoteBase): |
|
152 | 152 | |
|
153 | 153 | def __init__(self, factory): |
|
154 | 154 | self._factory = factory |
|
155 | 155 | self._bulk_methods = { |
|
156 | 156 | "date": self.date, |
|
157 | 157 | "author": self.author, |
|
158 | 158 | "branch": self.branch, |
|
159 | 159 | "message": self.message, |
|
160 | 160 | "parents": self.parents, |
|
161 | 161 | "_commit": self.revision, |
|
162 | 162 | } |
|
163 | 163 | self._bulk_file_methods = { |
|
164 | 164 | "size": self.get_node_size, |
|
165 | 165 | "data": self.get_node_data, |
|
166 | 166 | "flags": self.get_node_flags, |
|
167 | 167 | "is_binary": self.get_node_is_binary, |
|
168 | 168 | "md5": self.md5_hash |
|
169 | 169 | } |
|
170 | 170 | |
|
171 | 171 | def _wire_to_config(self, wire): |
|
172 | 172 | if 'config' in wire: |
|
173 | 173 | return {x[0] + '_' + x[1]: x[2] for x in wire['config']} |
|
174 | 174 | return {} |
|
175 | 175 | |
|
176 | 176 | def _remote_conf(self, config): |
|
177 | 177 | params = [ |
|
178 | 178 | '-c', 'core.askpass=""', |
|
179 | 179 | ] |
|
180 | 180 | config_attrs = { |
|
181 | 181 | 'vcs_ssl_dir': 'http.sslCAinfo={}', |
|
182 | 182 | 'vcs_git_lfs_store_location': 'lfs.storage={}' |
|
183 | 183 | } |
|
184 | 184 | for key, param in config_attrs.items(): |
|
185 | 185 | if value := config.get(key): |
|
186 | 186 | params.extend(['-c', param.format(value)]) |
|
187 | 187 | return params |
|
188 | 188 | |
|
189 | 189 | @reraise_safe_exceptions |
|
190 | 190 | def discover_git_version(self): |
|
191 | 191 | stdout, _ = self.run_git_command( |
|
192 | 192 | {}, ['--version'], _bare=True, _safe=True) |
|
193 | 193 | prefix = b'git version' |
|
194 | 194 | if stdout.startswith(prefix): |
|
195 | 195 | stdout = stdout[len(prefix):] |
|
196 | 196 | return safe_str(stdout.strip()) |
|
197 | 197 | |
|
198 | 198 | @reraise_safe_exceptions |
|
199 | 199 | def is_empty(self, wire): |
|
200 | 200 | repo_init = self._factory.repo_libgit2(wire) |
|
201 | 201 | with repo_init as repo: |
|
202 | 202 | try: |
|
203 | 203 | has_head = repo.head.name |
|
204 | 204 | if has_head: |
|
205 | 205 | return False |
|
206 | 206 | |
|
207 | 207 | # NOTE(marcink): check again using more expensive method |
|
208 | 208 | return repo.is_empty |
|
209 | 209 | except Exception: |
|
210 | 210 | pass |
|
211 | 211 | |
|
212 | 212 | return True |
|
213 | 213 | |
|
214 | 214 | @reraise_safe_exceptions |
|
215 | 215 | def assert_correct_path(self, wire): |
|
216 | 216 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
217 | 217 | region = self._region(wire) |
|
218 | 218 | |
|
219 | 219 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
220 | 220 | def _assert_correct_path(_context_uid, _repo_id, fast_check): |
|
221 | 221 | if fast_check: |
|
222 | 222 | path = safe_str(wire['path']) |
|
223 | 223 | if pygit2.discover_repository(path): |
|
224 | 224 | return True |
|
225 | 225 | return False |
|
226 | 226 | else: |
|
227 | 227 | try: |
|
228 | 228 | repo_init = self._factory.repo_libgit2(wire) |
|
229 | 229 | with repo_init: |
|
230 | 230 | pass |
|
231 | 231 | except pygit2.GitError: |
|
232 | 232 | path = wire.get('path') |
|
233 | 233 | tb = traceback.format_exc() |
|
234 | 234 | log.debug("Invalid Git path `%s`, tb: %s", path, tb) |
|
235 | 235 | return False |
|
236 | 236 | return True |
|
237 | 237 | |
|
238 | 238 | return _assert_correct_path(context_uid, repo_id, True) |
|
239 | 239 | |
|
240 | 240 | @reraise_safe_exceptions |
|
241 | 241 | def bare(self, wire): |
|
242 | 242 | repo_init = self._factory.repo_libgit2(wire) |
|
243 | 243 | with repo_init as repo: |
|
244 | 244 | return repo.is_bare |
|
245 | 245 | |
|
246 | 246 | @reraise_safe_exceptions |
|
247 | 247 | def get_node_data(self, wire, commit_id, path): |
|
248 | 248 | repo_init = self._factory.repo_libgit2(wire) |
|
249 | 249 | with repo_init as repo: |
|
250 | 250 | commit = repo[commit_id] |
|
251 | 251 | blob_obj = commit.tree[path] |
|
252 | 252 | |
|
253 | 253 | if blob_obj.type != pygit2.GIT_OBJ_BLOB: |
|
254 | 254 | raise exceptions.LookupException()( |
|
255 | 255 | f'Tree for commit_id:{commit_id} is not a blob: {blob_obj.type_str}') |
|
256 | 256 | |
|
257 | 257 | return BytesEnvelope(blob_obj.data) |
|
258 | 258 | |
|
259 | 259 | @reraise_safe_exceptions |
|
260 | 260 | def get_node_size(self, wire, commit_id, path): |
|
261 | 261 | repo_init = self._factory.repo_libgit2(wire) |
|
262 | 262 | with repo_init as repo: |
|
263 | 263 | commit = repo[commit_id] |
|
264 | 264 | blob_obj = commit.tree[path] |
|
265 | 265 | |
|
266 | 266 | if blob_obj.type != pygit2.GIT_OBJ_BLOB: |
|
267 | 267 | raise exceptions.LookupException()( |
|
268 | 268 | f'Tree for commit_id:{commit_id} is not a blob: {blob_obj.type_str}') |
|
269 | 269 | |
|
270 | 270 | return blob_obj.size |
|
271 | 271 | |
|
272 | 272 | @reraise_safe_exceptions |
|
273 | 273 | def get_node_flags(self, wire, commit_id, path): |
|
274 | 274 | repo_init = self._factory.repo_libgit2(wire) |
|
275 | 275 | with repo_init as repo: |
|
276 | 276 | commit = repo[commit_id] |
|
277 | 277 | blob_obj = commit.tree[path] |
|
278 | 278 | |
|
279 | 279 | if blob_obj.type != pygit2.GIT_OBJ_BLOB: |
|
280 | 280 | raise exceptions.LookupException()( |
|
281 | 281 | f'Tree for commit_id:{commit_id} is not a blob: {blob_obj.type_str}') |
|
282 | 282 | |
|
283 | 283 | return blob_obj.filemode |
|
284 | 284 | |
|
285 | 285 | @reraise_safe_exceptions |
|
286 | 286 | def get_node_is_binary(self, wire, commit_id, path): |
|
287 | 287 | repo_init = self._factory.repo_libgit2(wire) |
|
288 | 288 | with repo_init as repo: |
|
289 | 289 | commit = repo[commit_id] |
|
290 | 290 | blob_obj = commit.tree[path] |
|
291 | 291 | |
|
292 | 292 | if blob_obj.type != pygit2.GIT_OBJ_BLOB: |
|
293 | 293 | raise exceptions.LookupException()( |
|
294 | 294 | f'Tree for commit_id:{commit_id} is not a blob: {blob_obj.type_str}') |
|
295 | 295 | |
|
296 | 296 | return blob_obj.is_binary |
|
297 | 297 | |
|
298 | 298 | @reraise_safe_exceptions |
|
299 | 299 | def blob_as_pretty_string(self, wire, sha): |
|
300 | 300 | repo_init = self._factory.repo_libgit2(wire) |
|
301 | 301 | with repo_init as repo: |
|
302 | 302 | blob_obj = repo[sha] |
|
303 | 303 | return BytesEnvelope(blob_obj.data) |
|
304 | 304 | |
|
305 | 305 | @reraise_safe_exceptions |
|
306 | 306 | def blob_raw_length(self, wire, sha): |
|
307 | 307 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
308 | 308 | region = self._region(wire) |
|
309 | 309 | |
|
310 | 310 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
311 | 311 | def _blob_raw_length(_repo_id, _sha): |
|
312 | 312 | |
|
313 | 313 | repo_init = self._factory.repo_libgit2(wire) |
|
314 | 314 | with repo_init as repo: |
|
315 | 315 | blob = repo[sha] |
|
316 | 316 | return blob.size |
|
317 | 317 | |
|
318 | 318 | return _blob_raw_length(repo_id, sha) |
|
319 | 319 | |
|
320 | 320 | def _parse_lfs_pointer(self, raw_content): |
|
321 | 321 | spec_string = b'version https://git-lfs.github.com/spec' |
|
322 | 322 | if raw_content and raw_content.startswith(spec_string): |
|
323 | 323 | |
|
324 | 324 | pattern = re.compile(rb""" |
|
325 | 325 | (?:\n)? |
|
326 | 326 | ^version[ ]https://git-lfs\.github\.com/spec/(?P<spec_ver>v\d+)\n |
|
327 | 327 | ^oid[ ] sha256:(?P<oid_hash>[0-9a-f]{64})\n |
|
328 | 328 | ^size[ ](?P<oid_size>[0-9]+)\n |
|
329 | 329 | (?:\n)? |
|
330 | 330 | """, re.VERBOSE | re.MULTILINE) |
|
331 | 331 | match = pattern.match(raw_content) |
|
332 | 332 | if match: |
|
333 | 333 | return match.groupdict() |
|
334 | 334 | |
|
335 | 335 | return {} |
|
336 | 336 | |
|
337 | 337 | @reraise_safe_exceptions |
|
338 | 338 | def is_large_file(self, wire, commit_id): |
|
339 | 339 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
340 | 340 | region = self._region(wire) |
|
341 | 341 | |
|
342 | 342 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
343 | 343 | def _is_large_file(_repo_id, _sha): |
|
344 | 344 | repo_init = self._factory.repo_libgit2(wire) |
|
345 | 345 | with repo_init as repo: |
|
346 | 346 | blob = repo[commit_id] |
|
347 | 347 | if blob.is_binary: |
|
348 | 348 | return {} |
|
349 | 349 | |
|
350 | 350 | return self._parse_lfs_pointer(blob.data) |
|
351 | 351 | |
|
352 | 352 | return _is_large_file(repo_id, commit_id) |
|
353 | 353 | |
|
354 | 354 | @reraise_safe_exceptions |
|
355 | 355 | def is_binary(self, wire, tree_id): |
|
356 | 356 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
357 | 357 | region = self._region(wire) |
|
358 | 358 | |
|
359 | 359 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
360 | 360 | def _is_binary(_repo_id, _tree_id): |
|
361 | 361 | repo_init = self._factory.repo_libgit2(wire) |
|
362 | 362 | with repo_init as repo: |
|
363 | 363 | blob_obj = repo[tree_id] |
|
364 | 364 | return blob_obj.is_binary |
|
365 | 365 | |
|
366 | 366 | return _is_binary(repo_id, tree_id) |
|
367 | 367 | |
|
368 | 368 | @reraise_safe_exceptions |
|
369 | 369 | def md5_hash(self, wire, commit_id, path): |
|
370 | 370 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
371 | 371 | region = self._region(wire) |
|
372 | 372 | |
|
373 | 373 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
374 | 374 | def _md5_hash(_repo_id, _commit_id, _path): |
|
375 | 375 | repo_init = self._factory.repo_libgit2(wire) |
|
376 | 376 | with repo_init as repo: |
|
377 | 377 | commit = repo[_commit_id] |
|
378 | 378 | blob_obj = commit.tree[_path] |
|
379 | 379 | |
|
380 | 380 | if blob_obj.type != pygit2.GIT_OBJ_BLOB: |
|
381 | 381 | raise exceptions.LookupException()( |
|
382 | 382 | f'Tree for commit_id:{_commit_id} is not a blob: {blob_obj.type_str}') |
|
383 | 383 | |
|
384 | 384 | return '' |
|
385 | 385 | |
|
386 | 386 | return _md5_hash(repo_id, commit_id, path) |
|
387 | 387 | |
|
388 | 388 | @reraise_safe_exceptions |
|
389 | 389 | def in_largefiles_store(self, wire, oid): |
|
390 | 390 | conf = self._wire_to_config(wire) |
|
391 | 391 | repo_init = self._factory.repo_libgit2(wire) |
|
392 | 392 | with repo_init as repo: |
|
393 | 393 | repo_name = repo.path |
|
394 | 394 | |
|
395 | 395 | store_location = conf.get('vcs_git_lfs_store_location') |
|
396 | 396 | if store_location: |
|
397 | 397 | |
|
398 | 398 | store = LFSOidStore( |
|
399 | 399 | oid=oid, repo=repo_name, store_location=store_location) |
|
400 | 400 | return store.has_oid() |
|
401 | 401 | |
|
402 | 402 | return False |
|
403 | 403 | |
|
404 | 404 | @reraise_safe_exceptions |
|
405 | 405 | def store_path(self, wire, oid): |
|
406 | 406 | conf = self._wire_to_config(wire) |
|
407 | 407 | repo_init = self._factory.repo_libgit2(wire) |
|
408 | 408 | with repo_init as repo: |
|
409 | 409 | repo_name = repo.path |
|
410 | 410 | |
|
411 | 411 | store_location = conf.get('vcs_git_lfs_store_location') |
|
412 | 412 | if store_location: |
|
413 | 413 | store = LFSOidStore( |
|
414 | 414 | oid=oid, repo=repo_name, store_location=store_location) |
|
415 | 415 | return store.oid_path |
|
416 | 416 | raise ValueError(f'Unable to fetch oid with path {oid}') |
|
417 | 417 | |
|
418 | 418 | @reraise_safe_exceptions |
|
419 | 419 | def bulk_request(self, wire, rev, pre_load): |
|
420 | 420 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
421 | 421 | region = self._region(wire) |
|
422 | 422 | |
|
423 | 423 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
424 | 424 | def _bulk_request(_repo_id, _rev, _pre_load): |
|
425 | 425 | result = {} |
|
426 | 426 | for attr in pre_load: |
|
427 | 427 | try: |
|
428 | 428 | method = self._bulk_methods[attr] |
|
429 | 429 | wire.update({'cache': False}) # disable cache for bulk calls so we don't double cache |
|
430 | 430 | args = [wire, rev] |
|
431 | 431 | result[attr] = method(*args) |
|
432 | 432 | except KeyError as e: |
|
433 | 433 | raise exceptions.VcsException(e)(f"Unknown bulk attribute: {attr}") |
|
434 | 434 | return result |
|
435 | 435 | |
|
436 | 436 | return _bulk_request(repo_id, rev, sorted(pre_load)) |
|
437 | 437 | |
|
438 | 438 | @reraise_safe_exceptions |
|
439 | 439 | def bulk_file_request(self, wire, commit_id, path, pre_load): |
|
440 | 440 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
441 | 441 | region = self._region(wire) |
|
442 | 442 | |
|
443 | 443 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
444 | 444 | def _bulk_file_request(_repo_id, _commit_id, _path, _pre_load): |
|
445 | 445 | result = {} |
|
446 | 446 | for attr in pre_load: |
|
447 | 447 | try: |
|
448 | 448 | method = self._bulk_file_methods[attr] |
|
449 | 449 | wire.update({'cache': False}) # disable cache for bulk calls so we don't double cache |
|
450 | 450 | result[attr] = method(wire, _commit_id, _path) |
|
451 | 451 | except KeyError as e: |
|
452 | 452 | raise exceptions.VcsException(e)(f'Unknown bulk attribute: "{attr}"') |
|
453 | 453 | return result |
|
454 | 454 | |
|
455 | 455 | return BinaryEnvelope(_bulk_file_request(repo_id, commit_id, path, sorted(pre_load))) |
|
456 | 456 | |
|
457 | 457 | def _build_opener(self, url: str): |
|
458 | 458 | handlers = [] |
|
459 | 459 | url_obj = url_parser(safe_bytes(url)) |
|
460 | 460 | authinfo = url_obj.authinfo()[1] |
|
461 | 461 | |
|
462 | 462 | if authinfo: |
|
463 | 463 | # create a password manager |
|
464 | 464 | passmgr = urllib.request.HTTPPasswordMgrWithDefaultRealm() |
|
465 | 465 | passmgr.add_password(*convert_to_str(authinfo)) |
|
466 | 466 | |
|
467 | 467 | handlers.extend((httpbasicauthhandler(passmgr), |
|
468 | 468 | httpdigestauthhandler(passmgr))) |
|
469 | 469 | |
|
470 | 470 | return urllib.request.build_opener(*handlers) |
|
471 | 471 | |
|
472 | 472 | @reraise_safe_exceptions |
|
473 | 473 | def check_url(self, url, config): |
|
474 | 474 | url_obj = url_parser(safe_bytes(url)) |
|
475 | 475 | |
|
476 | 476 | test_uri = safe_str(url_obj.authinfo()[0]) |
|
477 | 477 | obfuscated_uri = get_obfuscated_url(url_obj) |
|
478 | 478 | |
|
479 | 479 | log.info("Checking URL for remote cloning/import: %s", obfuscated_uri) |
|
480 | 480 | |
|
481 | 481 | if not test_uri.endswith('info/refs'): |
|
482 | 482 | test_uri = test_uri.rstrip('/') + '/info/refs' |
|
483 | 483 | |
|
484 | 484 | o = self._build_opener(url=url) |
|
485 | 485 | o.addheaders = [('User-Agent', 'git/1.7.8.0')] # fake some git |
|
486 | 486 | |
|
487 | 487 | q = {"service": 'git-upload-pack'} |
|
488 | 488 | qs = f'?{urllib.parse.urlencode(q)}' |
|
489 | 489 | cu = f"{test_uri}{qs}" |
|
490 | 490 | |
|
491 | 491 | try: |
|
492 | 492 | req = urllib.request.Request(cu, None, {}) |
|
493 | 493 | log.debug("Trying to open URL %s", obfuscated_uri) |
|
494 | 494 | resp = o.open(req) |
|
495 | 495 | if resp.code != 200: |
|
496 | 496 | raise exceptions.URLError()('Return Code is not 200') |
|
497 | 497 | except Exception as e: |
|
498 | 498 | log.warning("URL cannot be opened: %s", obfuscated_uri, exc_info=True) |
|
499 | 499 | # means it cannot be cloned |
|
500 | 500 | raise exceptions.URLError(e)(f"[{obfuscated_uri}] org_exc: {e}") |
|
501 | 501 | |
|
502 | 502 | # now detect if it's proper git repo |
|
503 | 503 | gitdata: bytes = resp.read() |
|
504 | 504 | |
|
505 | 505 | if b'service=git-upload-pack' in gitdata: |
|
506 | 506 | pass |
|
507 | 507 | elif re.findall(br'[0-9a-fA-F]{40}\s+refs', gitdata): |
|
508 | 508 | # old style git can return some other format! |
|
509 | 509 | pass |
|
510 | 510 | else: |
|
511 | 511 | e = None |
|
512 | 512 | raise exceptions.URLError(e)( |
|
513 | 513 | f"url [{obfuscated_uri}] does not look like an hg repo org_exc: {e}") |
|
514 | 514 | |
|
515 | 515 | return True |
|
516 | 516 | |
|
517 | 517 | @reraise_safe_exceptions |
|
518 | 518 | def clone(self, wire, url, deferred, valid_refs, update_after_clone): |
|
519 | 519 | # TODO(marcink): deprecate this method. Last i checked we don't use it anymore |
|
520 | 520 | remote_refs = self.pull(wire, url, apply_refs=False) |
|
521 | 521 | repo = self._factory.repo(wire) |
|
522 | 522 | if isinstance(valid_refs, list): |
|
523 | 523 | valid_refs = tuple(valid_refs) |
|
524 | 524 | |
|
525 | 525 | for k in remote_refs: |
|
526 | 526 | # only parse heads/tags and skip so called deferred tags |
|
527 | 527 | if k.startswith(valid_refs) and not k.endswith(deferred): |
|
528 | 528 | repo[k] = remote_refs[k] |
|
529 | 529 | |
|
530 | 530 | if update_after_clone: |
|
531 | 531 | # we want to checkout HEAD |
|
532 | 532 | repo["HEAD"] = remote_refs["HEAD"] |
|
533 | 533 | index.build_index_from_tree(repo.path, repo.index_path(), |
|
534 | 534 | repo.object_store, repo["HEAD"].tree) |
|
535 | 535 | |
|
536 | 536 | @reraise_safe_exceptions |
|
537 | 537 | def branch(self, wire, commit_id): |
|
538 | 538 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
539 | 539 | region = self._region(wire) |
|
540 | 540 | |
|
541 | 541 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
542 | 542 | def _branch(_context_uid, _repo_id, _commit_id): |
|
543 | 543 | regex = re.compile('^refs/heads') |
|
544 | 544 | |
|
545 | 545 | def filter_with(ref): |
|
546 | 546 | return regex.match(ref[0]) and ref[1] == _commit_id |
|
547 | 547 | |
|
548 | 548 | branches = list(filter(filter_with, list(self.get_refs(wire).items()))) |
|
549 | 549 | return [x[0].split('refs/heads/')[-1] for x in branches] |
|
550 | 550 | |
|
551 | 551 | return _branch(context_uid, repo_id, commit_id) |
|
552 | 552 | |
|
553 | 553 | @reraise_safe_exceptions |
|
554 | def delete_branch(self, wire, branch_name): | |
|
555 | repo_init = self._factory.repo_libgit2(wire) | |
|
556 | with repo_init as repo: | |
|
557 | if branch := repo.lookup_branch(branch_name): | |
|
558 | branch.delete() | |
|
559 | ||
|
560 | @reraise_safe_exceptions | |
|
554 | 561 | def commit_branches(self, wire, commit_id): |
|
555 | 562 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
556 | 563 | region = self._region(wire) |
|
557 | 564 | |
|
558 | 565 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
559 | 566 | def _commit_branches(_context_uid, _repo_id, _commit_id): |
|
560 | 567 | repo_init = self._factory.repo_libgit2(wire) |
|
561 | 568 | with repo_init as repo: |
|
562 | 569 | branches = [x for x in repo.branches.with_commit(_commit_id)] |
|
563 | 570 | return branches |
|
564 | 571 | |
|
565 | 572 | return _commit_branches(context_uid, repo_id, commit_id) |
|
566 | 573 | |
|
567 | 574 | @reraise_safe_exceptions |
|
568 | 575 | def add_object(self, wire, content): |
|
569 | 576 | repo_init = self._factory.repo_libgit2(wire) |
|
570 | 577 | with repo_init as repo: |
|
571 | 578 | blob = objects.Blob() |
|
572 | 579 | blob.set_raw_string(content) |
|
573 | 580 | repo.object_store.add_object(blob) |
|
574 | 581 | return blob.id |
|
575 | 582 | |
|
576 | 583 | @reraise_safe_exceptions |
|
577 | 584 | def create_commit(self, wire, author, committer, message, branch, new_tree_id, |
|
578 | 585 | date_args: list[int, int] = None, |
|
579 | 586 | parents: list | None = None): |
|
580 | 587 | |
|
581 | 588 | repo_init = self._factory.repo_libgit2(wire) |
|
582 | 589 | with repo_init as repo: |
|
583 | 590 | |
|
584 | 591 | if date_args: |
|
585 | 592 | current_time, offset = date_args |
|
586 | 593 | |
|
587 | 594 | kw = { |
|
588 | 595 | 'time': current_time, |
|
589 | 596 | 'offset': offset |
|
590 | 597 | } |
|
591 | 598 | author = create_signature_from_string(author, **kw) |
|
592 | 599 | committer = create_signature_from_string(committer, **kw) |
|
593 | 600 | |
|
594 | 601 | tree = new_tree_id |
|
595 | 602 | if isinstance(tree, (bytes, str)): |
|
596 | 603 | # validate this tree is in the repo... |
|
597 | 604 | tree = repo[safe_str(tree)].id |
|
598 | 605 | |
|
599 | 606 | if parents: |
|
600 | 607 | # run via sha's and validate them in repo |
|
601 | 608 | parents = [repo[c].id for c in parents] |
|
602 | 609 | else: |
|
603 | 610 | parents = [] |
|
604 | 611 | # ensure we COMMIT on top of given branch head |
|
605 | 612 | # check if this repo has ANY branches, otherwise it's a new branch case we need to make |
|
606 | 613 | if branch in repo.branches.local: |
|
607 | 614 | parents += [repo.branches[branch].target] |
|
608 | 615 | elif [x for x in repo.branches.local]: |
|
609 | 616 | parents += [repo.head.target] |
|
610 | 617 | #else: |
|
611 | 618 | # in case we want to commit on new branch we create it on top of HEAD |
|
612 | 619 | #repo.branches.local.create(branch, repo.revparse_single('HEAD')) |
|
613 | 620 | |
|
614 | 621 | # # Create a new commit |
|
615 | 622 | commit_oid = repo.create_commit( |
|
616 | 623 | f'refs/heads/{branch}', # the name of the reference to update |
|
617 | 624 | author, # the author of the commit |
|
618 | 625 | committer, # the committer of the commit |
|
619 | 626 | message, # the commit message |
|
620 | 627 | tree, # the tree produced by the index |
|
621 | 628 | parents # list of parents for the new commit, usually just one, |
|
622 | 629 | ) |
|
623 | 630 | |
|
624 | 631 | new_commit_id = safe_str(commit_oid) |
|
625 | 632 | |
|
626 | 633 | return new_commit_id |
|
627 | 634 | |
|
628 | 635 | @reraise_safe_exceptions |
|
629 | 636 | def commit(self, wire, commit_data, branch, commit_tree, updated, removed): |
|
630 | 637 | |
|
631 | 638 | def mode2pygit(mode): |
|
632 | 639 | """ |
|
633 | 640 | git only supports two filemode 644 and 755 |
|
634 | 641 | |
|
635 | 642 | 0o100755 -> 33261 |
|
636 | 643 | 0o100644 -> 33188 |
|
637 | 644 | """ |
|
638 | 645 | return { |
|
639 | 646 | 0o100644: pygit2.GIT_FILEMODE_BLOB, |
|
640 | 647 | 0o100755: pygit2.GIT_FILEMODE_BLOB_EXECUTABLE, |
|
641 | 648 | 0o120000: pygit2.GIT_FILEMODE_LINK |
|
642 | 649 | }.get(mode) or pygit2.GIT_FILEMODE_BLOB |
|
643 | 650 | |
|
644 | 651 | repo_init = self._factory.repo_libgit2(wire) |
|
645 | 652 | with repo_init as repo: |
|
646 | 653 | repo_index = repo.index |
|
647 | 654 | |
|
648 | 655 | commit_parents = None |
|
649 | 656 | if commit_tree and commit_data['parents']: |
|
650 | 657 | commit_parents = commit_data['parents'] |
|
651 | 658 | parent_commit = repo[commit_parents[0]] |
|
652 | 659 | repo_index.read_tree(parent_commit.tree) |
|
653 | 660 | |
|
654 | 661 | for pathspec in updated: |
|
655 | 662 | blob_id = repo.create_blob(pathspec['content']) |
|
656 | 663 | ie = pygit2.IndexEntry(pathspec['path'], blob_id, mode2pygit(pathspec['mode'])) |
|
657 | 664 | repo_index.add(ie) |
|
658 | 665 | |
|
659 | 666 | for pathspec in removed: |
|
660 | 667 | repo_index.remove(pathspec) |
|
661 | 668 | |
|
662 | 669 | # Write changes to the index |
|
663 | 670 | repo_index.write() |
|
664 | 671 | |
|
665 | 672 | # Create a tree from the updated index |
|
666 | 673 | written_commit_tree = repo_index.write_tree() |
|
667 | 674 | |
|
668 | 675 | new_tree_id = written_commit_tree |
|
669 | 676 | |
|
670 | 677 | author = commit_data['author'] |
|
671 | 678 | committer = commit_data['committer'] |
|
672 | 679 | message = commit_data['message'] |
|
673 | 680 | |
|
674 | 681 | date_args = [int(commit_data['commit_time']), int(commit_data['commit_timezone'])] |
|
675 | 682 | |
|
676 | 683 | new_commit_id = self.create_commit(wire, author, committer, message, branch, |
|
677 | 684 | new_tree_id, date_args=date_args, parents=commit_parents) |
|
678 | 685 | |
|
679 | 686 | # libgit2, ensure the branch is there and exists |
|
680 | 687 | self.create_branch(wire, branch, new_commit_id) |
|
681 | 688 | |
|
682 | 689 | # libgit2, set new ref to this created commit |
|
683 | 690 | self.set_refs(wire, f'refs/heads/{branch}', new_commit_id) |
|
684 | 691 | |
|
685 | 692 | return new_commit_id |
|
686 | 693 | |
|
687 | 694 | @reraise_safe_exceptions |
|
688 | 695 | def pull(self, wire, url, apply_refs=True, refs=None, update_after=False): |
|
689 | 696 | if url != 'default' and '://' not in url: |
|
690 | 697 | client = LocalGitClient(url) |
|
691 | 698 | else: |
|
692 | 699 | url_obj = url_parser(safe_bytes(url)) |
|
693 | 700 | o = self._build_opener(url) |
|
694 | 701 | url = url_obj.authinfo()[0] |
|
695 | 702 | client = HttpGitClient(base_url=url, opener=o) |
|
696 | 703 | repo = self._factory.repo(wire) |
|
697 | 704 | |
|
698 | 705 | determine_wants = repo.object_store.determine_wants_all |
|
699 | 706 | |
|
700 | 707 | if refs: |
|
701 | 708 | refs: list[bytes] = [ascii_bytes(x) for x in refs] |
|
702 | 709 | |
|
703 | 710 | def determine_wants_requested(_remote_refs): |
|
704 | 711 | determined = [] |
|
705 | 712 | for ref_name, ref_hash in _remote_refs.items(): |
|
706 | 713 | bytes_ref_name = safe_bytes(ref_name) |
|
707 | 714 | |
|
708 | 715 | if bytes_ref_name in refs: |
|
709 | 716 | bytes_ref_hash = safe_bytes(ref_hash) |
|
710 | 717 | determined.append(bytes_ref_hash) |
|
711 | 718 | return determined |
|
712 | 719 | |
|
713 | 720 | # swap with our custom requested wants |
|
714 | 721 | determine_wants = determine_wants_requested |
|
715 | 722 | |
|
716 | 723 | try: |
|
717 | 724 | remote_refs = client.fetch( |
|
718 | 725 | path=url, target=repo, determine_wants=determine_wants) |
|
719 | 726 | |
|
720 | 727 | except NotGitRepository as e: |
|
721 | 728 | log.warning( |
|
722 | 729 | 'Trying to fetch from "%s" failed, not a Git repository.', url) |
|
723 | 730 | # Exception can contain unicode which we convert |
|
724 | 731 | raise exceptions.AbortException(e)(repr(e)) |
|
725 | 732 | |
|
726 | 733 | # mikhail: client.fetch() returns all the remote refs, but fetches only |
|
727 | 734 | # refs filtered by `determine_wants` function. We need to filter result |
|
728 | 735 | # as well |
|
729 | 736 | if refs: |
|
730 | 737 | remote_refs = {k: remote_refs[k] for k in remote_refs if k in refs} |
|
731 | 738 | |
|
732 | 739 | if apply_refs: |
|
733 | 740 | # TODO: johbo: Needs proper test coverage with a git repository |
|
734 | 741 | # that contains a tag object, so that we would end up with |
|
735 | 742 | # a peeled ref at this point. |
|
736 | 743 | for k in remote_refs: |
|
737 | 744 | if k.endswith(PEELED_REF_MARKER): |
|
738 | 745 | log.debug("Skipping peeled reference %s", k) |
|
739 | 746 | continue |
|
740 | 747 | repo[k] = remote_refs[k] |
|
741 | 748 | |
|
742 | 749 | if refs and not update_after: |
|
743 | 750 | # update to ref |
|
744 | 751 | # mikhail: explicitly set the head to the last ref. |
|
745 | 752 | update_to_ref = refs[-1] |
|
746 | 753 | if isinstance(update_after, str): |
|
747 | 754 | update_to_ref = update_after |
|
748 | 755 | |
|
749 | 756 | repo[HEAD_MARKER] = remote_refs[update_to_ref] |
|
750 | 757 | |
|
751 | 758 | if update_after: |
|
752 | 759 | # we want to check out HEAD |
|
753 | 760 | repo[HEAD_MARKER] = remote_refs[HEAD_MARKER] |
|
754 | 761 | index.build_index_from_tree(repo.path, repo.index_path(), |
|
755 | 762 | repo.object_store, repo[HEAD_MARKER].tree) |
|
756 | 763 | |
|
757 | 764 | if isinstance(remote_refs, FetchPackResult): |
|
758 | 765 | return remote_refs.refs |
|
759 | 766 | return remote_refs |
|
760 | 767 | |
|
761 | 768 | @reraise_safe_exceptions |
|
762 | 769 | def sync_fetch(self, wire, url, refs=None, all_refs=False, **kwargs): |
|
763 | 770 | self._factory.repo(wire) |
|
764 | 771 | if refs and not isinstance(refs, (list, tuple)): |
|
765 | 772 | refs = [refs] |
|
766 | 773 | |
|
767 | 774 | config = self._wire_to_config(wire) |
|
768 | 775 | # get all remote refs we'll use to fetch later |
|
769 | 776 | cmd = ['ls-remote'] |
|
770 | 777 | if not all_refs: |
|
771 | 778 | cmd += ['--heads', '--tags'] |
|
772 | 779 | cmd += [url] |
|
773 | 780 | output, __ = self.run_git_command( |
|
774 | 781 | wire, cmd, fail_on_stderr=False, |
|
775 | 782 | _copts=self._remote_conf(config), |
|
776 | 783 | extra_env={'GIT_TERMINAL_PROMPT': '0'}) |
|
777 | 784 | |
|
778 | 785 | remote_refs = collections.OrderedDict() |
|
779 | 786 | fetch_refs = [] |
|
780 | 787 | |
|
781 | 788 | for ref_line in output.splitlines(): |
|
782 | 789 | sha, ref = ref_line.split(b'\t') |
|
783 | 790 | sha = sha.strip() |
|
784 | 791 | if ref in remote_refs: |
|
785 | 792 | # duplicate, skip |
|
786 | 793 | continue |
|
787 | 794 | if ref.endswith(PEELED_REF_MARKER): |
|
788 | 795 | log.debug("Skipping peeled reference %s", ref) |
|
789 | 796 | continue |
|
790 | 797 | # don't sync HEAD |
|
791 | 798 | if ref in [HEAD_MARKER]: |
|
792 | 799 | continue |
|
793 | 800 | |
|
794 | 801 | remote_refs[ref] = sha |
|
795 | 802 | |
|
796 | 803 | if refs and sha in refs: |
|
797 | 804 | # we filter fetch using our specified refs |
|
798 | 805 | fetch_refs.append(f'{safe_str(ref)}:{safe_str(ref)}') |
|
799 | 806 | elif not refs: |
|
800 | 807 | fetch_refs.append(f'{safe_str(ref)}:{safe_str(ref)}') |
|
801 | 808 | log.debug('Finished obtaining fetch refs, total: %s', len(fetch_refs)) |
|
802 | 809 | |
|
803 | 810 | if fetch_refs: |
|
804 | 811 | for chunk in more_itertools.chunked(fetch_refs, 128): |
|
805 | 812 | fetch_refs_chunks = list(chunk) |
|
806 | 813 | log.debug('Fetching %s refs from import url', len(fetch_refs_chunks)) |
|
807 | 814 | self.run_git_command( |
|
808 | 815 | wire, ['fetch', url, '--force', '--prune', '--'] + fetch_refs_chunks, |
|
809 | 816 | fail_on_stderr=False, |
|
810 | 817 | _copts=self._remote_conf(config), |
|
811 | 818 | extra_env={'GIT_TERMINAL_PROMPT': '0'}) |
|
812 | 819 | if kwargs.get('sync_large_objects'): |
|
813 | 820 | self.run_git_command( |
|
814 | 821 | wire, ['lfs', 'fetch', url, '--all'], |
|
815 | 822 | fail_on_stderr=False, |
|
816 | 823 | _copts=self._remote_conf(config), |
|
817 | 824 | ) |
|
818 | 825 | |
|
819 | 826 | return remote_refs |
|
820 | 827 | |
|
821 | 828 | @reraise_safe_exceptions |
|
822 | 829 | def sync_push(self, wire, url, refs=None, **kwargs): |
|
823 | 830 | if not self.check_url(url, wire): |
|
824 | 831 | return |
|
825 | 832 | config = self._wire_to_config(wire) |
|
826 | 833 | self._factory.repo(wire) |
|
827 | 834 | self.run_git_command( |
|
828 | 835 | wire, ['push', url, '--mirror'], fail_on_stderr=False, |
|
829 | 836 | _copts=self._remote_conf(config), |
|
830 | 837 | extra_env={'GIT_TERMINAL_PROMPT': '0'}) |
|
831 | 838 | if kwargs.get('sync_large_objects'): |
|
832 | 839 | self.run_git_command( |
|
833 | 840 | wire, ['lfs', 'push', url, '--all'], |
|
834 | 841 | fail_on_stderr=False, |
|
835 | 842 | _copts=self._remote_conf(config), |
|
836 | 843 | ) |
|
837 | 844 | |
|
838 | 845 | @reraise_safe_exceptions |
|
839 | 846 | def get_remote_refs(self, wire, url): |
|
840 | 847 | repo = Repo(url) |
|
841 | 848 | return repo.get_refs() |
|
842 | 849 | |
|
843 | 850 | @reraise_safe_exceptions |
|
844 | 851 | def get_description(self, wire): |
|
845 | 852 | repo = self._factory.repo(wire) |
|
846 | 853 | return repo.get_description() |
|
847 | 854 | |
|
848 | 855 | @reraise_safe_exceptions |
|
849 | 856 | def get_missing_revs(self, wire, rev1, rev2, other_repo_path): |
|
850 | 857 | origin_repo_path = wire['path'] |
|
851 | 858 | repo = self._factory.repo(wire) |
|
852 | 859 | # fetch from other_repo_path to our origin repo |
|
853 | 860 | LocalGitClient(thin_packs=False).fetch(other_repo_path, repo) |
|
854 | 861 | |
|
855 | 862 | wire_remote = wire.copy() |
|
856 | 863 | wire_remote['path'] = other_repo_path |
|
857 | 864 | repo_remote = self._factory.repo(wire_remote) |
|
858 | 865 | |
|
859 | 866 | # fetch from origin_repo_path to our remote repo |
|
860 | 867 | LocalGitClient(thin_packs=False).fetch(origin_repo_path, repo_remote) |
|
861 | 868 | |
|
862 | 869 | revs = [ |
|
863 | 870 | x.commit.id |
|
864 | 871 | for x in repo_remote.get_walker(include=[safe_bytes(rev2)], exclude=[safe_bytes(rev1)])] |
|
865 | 872 | return revs |
|
866 | 873 | |
|
867 | 874 | @reraise_safe_exceptions |
|
868 | 875 | def get_object(self, wire, sha, maybe_unreachable=False): |
|
869 | 876 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
870 | 877 | region = self._region(wire) |
|
871 | 878 | |
|
872 | 879 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
873 | 880 | def _get_object(_context_uid, _repo_id, _sha): |
|
874 | 881 | repo_init = self._factory.repo_libgit2(wire) |
|
875 | 882 | with repo_init as repo: |
|
876 | 883 | |
|
877 | 884 | missing_commit_err = 'Commit {} does not exist for `{}`'.format(sha, wire['path']) |
|
878 | 885 | try: |
|
879 | 886 | commit = repo.revparse_single(sha) |
|
880 | 887 | except KeyError: |
|
881 | 888 | # NOTE(marcink): KeyError doesn't give us any meaningful information |
|
882 | 889 | # here, we instead give something more explicit |
|
883 | 890 | e = exceptions.RefNotFoundException('SHA: %s not found', sha) |
|
884 | 891 | raise exceptions.LookupException(e)(missing_commit_err) |
|
885 | 892 | except ValueError as e: |
|
886 | 893 | raise exceptions.LookupException(e)(missing_commit_err) |
|
887 | 894 | |
|
888 | 895 | is_tag = False |
|
889 | 896 | if isinstance(commit, pygit2.Tag): |
|
890 | 897 | commit = repo.get(commit.target) |
|
891 | 898 | is_tag = True |
|
892 | 899 | |
|
893 | 900 | check_dangling = True |
|
894 | 901 | if is_tag: |
|
895 | 902 | check_dangling = False |
|
896 | 903 | |
|
897 | 904 | if check_dangling and maybe_unreachable: |
|
898 | 905 | check_dangling = False |
|
899 | 906 | |
|
900 | 907 | # we used a reference and it parsed means we're not having a dangling commit |
|
901 | 908 | if sha != commit.hex: |
|
902 | 909 | check_dangling = False |
|
903 | 910 | |
|
904 | 911 | if check_dangling: |
|
905 | 912 | # check for dangling commit |
|
906 | 913 | for branch in repo.branches.with_commit(commit.hex): |
|
907 | 914 | if branch: |
|
908 | 915 | break |
|
909 | 916 | else: |
|
910 | 917 | # NOTE(marcink): Empty error doesn't give us any meaningful information |
|
911 | 918 | # here, we instead give something more explicit |
|
912 | 919 | e = exceptions.RefNotFoundException('SHA: %s not found in branches', sha) |
|
913 | 920 | raise exceptions.LookupException(e)(missing_commit_err) |
|
914 | 921 | |
|
915 | 922 | commit_id = commit.hex |
|
916 | 923 | type_str = commit.type_str |
|
917 | 924 | |
|
918 | 925 | return { |
|
919 | 926 | 'id': commit_id, |
|
920 | 927 | 'type': type_str, |
|
921 | 928 | 'commit_id': commit_id, |
|
922 | 929 | 'idx': 0 |
|
923 | 930 | } |
|
924 | 931 | |
|
925 | 932 | return _get_object(context_uid, repo_id, sha) |
|
926 | 933 | |
|
927 | 934 | @reraise_safe_exceptions |
|
928 | 935 | def get_refs(self, wire): |
|
929 | 936 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
930 | 937 | region = self._region(wire) |
|
931 | 938 | |
|
932 | 939 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
933 | 940 | def _get_refs(_context_uid, _repo_id): |
|
934 | 941 | |
|
935 | 942 | repo_init = self._factory.repo_libgit2(wire) |
|
936 | 943 | with repo_init as repo: |
|
937 | 944 | regex = re.compile('^refs/(heads|tags)/') |
|
938 | 945 | return {x.name: x.target.hex for x in |
|
939 | 946 | [ref for ref in repo.listall_reference_objects() if regex.match(ref.name)]} |
|
940 | 947 | |
|
941 | 948 | return _get_refs(context_uid, repo_id) |
|
942 | 949 | |
|
943 | 950 | @reraise_safe_exceptions |
|
944 | 951 | def get_branch_pointers(self, wire): |
|
945 | 952 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
946 | 953 | region = self._region(wire) |
|
947 | 954 | |
|
948 | 955 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
949 | 956 | def _get_branch_pointers(_context_uid, _repo_id): |
|
950 | 957 | |
|
951 | 958 | repo_init = self._factory.repo_libgit2(wire) |
|
952 | 959 | regex = re.compile('^refs/heads') |
|
953 | 960 | with repo_init as repo: |
|
954 | 961 | branches = [ref for ref in repo.listall_reference_objects() if regex.match(ref.name)] |
|
955 | 962 | return {x.target.hex: x.shorthand for x in branches} |
|
956 | 963 | |
|
957 | 964 | return _get_branch_pointers(context_uid, repo_id) |
|
958 | 965 | |
|
959 | 966 | @reraise_safe_exceptions |
|
960 | 967 | def head(self, wire, show_exc=True): |
|
961 | 968 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
962 | 969 | region = self._region(wire) |
|
963 | 970 | |
|
964 | 971 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
965 | 972 | def _head(_context_uid, _repo_id, _show_exc): |
|
966 | 973 | repo_init = self._factory.repo_libgit2(wire) |
|
967 | 974 | with repo_init as repo: |
|
968 | 975 | try: |
|
969 | 976 | return repo.head.peel().hex |
|
970 | 977 | except Exception: |
|
971 | 978 | if show_exc: |
|
972 | 979 | raise |
|
973 | 980 | return _head(context_uid, repo_id, show_exc) |
|
974 | 981 | |
|
975 | 982 | @reraise_safe_exceptions |
|
976 | 983 | def init(self, wire): |
|
977 | 984 | repo_path = safe_str(wire['path']) |
|
978 | 985 | os.makedirs(repo_path, mode=0o755) |
|
979 | 986 | pygit2.init_repository(repo_path, bare=False) |
|
980 | 987 | |
|
981 | 988 | @reraise_safe_exceptions |
|
982 | 989 | def init_bare(self, wire): |
|
983 | 990 | repo_path = safe_str(wire['path']) |
|
984 | 991 | os.makedirs(repo_path, mode=0o755) |
|
985 | 992 | pygit2.init_repository(repo_path, bare=True) |
|
986 | 993 | |
|
987 | 994 | @reraise_safe_exceptions |
|
988 | 995 | def revision(self, wire, rev): |
|
989 | 996 | |
|
990 | 997 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
991 | 998 | region = self._region(wire) |
|
992 | 999 | |
|
993 | 1000 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
994 | 1001 | def _revision(_context_uid, _repo_id, _rev): |
|
995 | 1002 | repo_init = self._factory.repo_libgit2(wire) |
|
996 | 1003 | with repo_init as repo: |
|
997 | 1004 | commit = repo[rev] |
|
998 | 1005 | obj_data = { |
|
999 | 1006 | 'id': commit.id.hex, |
|
1000 | 1007 | } |
|
1001 | 1008 | # tree objects itself don't have tree_id attribute |
|
1002 | 1009 | if hasattr(commit, 'tree_id'): |
|
1003 | 1010 | obj_data['tree'] = commit.tree_id.hex |
|
1004 | 1011 | |
|
1005 | 1012 | return obj_data |
|
1006 | 1013 | return _revision(context_uid, repo_id, rev) |
|
1007 | 1014 | |
|
1008 | 1015 | @reraise_safe_exceptions |
|
1009 | 1016 | def date(self, wire, commit_id): |
|
1010 | 1017 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
1011 | 1018 | region = self._region(wire) |
|
1012 | 1019 | |
|
1013 | 1020 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1014 | 1021 | def _date(_repo_id, _commit_id): |
|
1015 | 1022 | repo_init = self._factory.repo_libgit2(wire) |
|
1016 | 1023 | with repo_init as repo: |
|
1017 | 1024 | commit = repo[commit_id] |
|
1018 | 1025 | |
|
1019 | 1026 | if hasattr(commit, 'commit_time'): |
|
1020 | 1027 | commit_time, commit_time_offset = commit.commit_time, commit.commit_time_offset |
|
1021 | 1028 | else: |
|
1022 | 1029 | commit = commit.get_object() |
|
1023 | 1030 | commit_time, commit_time_offset = commit.commit_time, commit.commit_time_offset |
|
1024 | 1031 | |
|
1025 | 1032 | # TODO(marcink): check dulwich difference of offset vs timezone |
|
1026 | 1033 | return [commit_time, commit_time_offset] |
|
1027 | 1034 | return _date(repo_id, commit_id) |
|
1028 | 1035 | |
|
1029 | 1036 | @reraise_safe_exceptions |
|
1030 | 1037 | def author(self, wire, commit_id): |
|
1031 | 1038 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
1032 | 1039 | region = self._region(wire) |
|
1033 | 1040 | |
|
1034 | 1041 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1035 | 1042 | def _author(_repo_id, _commit_id): |
|
1036 | 1043 | repo_init = self._factory.repo_libgit2(wire) |
|
1037 | 1044 | with repo_init as repo: |
|
1038 | 1045 | commit = repo[commit_id] |
|
1039 | 1046 | |
|
1040 | 1047 | if hasattr(commit, 'author'): |
|
1041 | 1048 | author = commit.author |
|
1042 | 1049 | else: |
|
1043 | 1050 | author = commit.get_object().author |
|
1044 | 1051 | |
|
1045 | 1052 | if author.email: |
|
1046 | 1053 | return f"{author.name} <{author.email}>" |
|
1047 | 1054 | |
|
1048 | 1055 | try: |
|
1049 | 1056 | return f"{author.name}" |
|
1050 | 1057 | except Exception: |
|
1051 | 1058 | return f"{safe_str(author.raw_name)}" |
|
1052 | 1059 | |
|
1053 | 1060 | return _author(repo_id, commit_id) |
|
1054 | 1061 | |
|
1055 | 1062 | @reraise_safe_exceptions |
|
1056 | 1063 | def message(self, wire, commit_id): |
|
1057 | 1064 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
1058 | 1065 | region = self._region(wire) |
|
1059 | 1066 | |
|
1060 | 1067 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1061 | 1068 | def _message(_repo_id, _commit_id): |
|
1062 | 1069 | repo_init = self._factory.repo_libgit2(wire) |
|
1063 | 1070 | with repo_init as repo: |
|
1064 | 1071 | commit = repo[commit_id] |
|
1065 | 1072 | return commit.message |
|
1066 | 1073 | return _message(repo_id, commit_id) |
|
1067 | 1074 | |
|
1068 | 1075 | @reraise_safe_exceptions |
|
1069 | 1076 | def parents(self, wire, commit_id): |
|
1070 | 1077 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
1071 | 1078 | region = self._region(wire) |
|
1072 | 1079 | |
|
1073 | 1080 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1074 | 1081 | def _parents(_repo_id, _commit_id): |
|
1075 | 1082 | repo_init = self._factory.repo_libgit2(wire) |
|
1076 | 1083 | with repo_init as repo: |
|
1077 | 1084 | commit = repo[commit_id] |
|
1078 | 1085 | if hasattr(commit, 'parent_ids'): |
|
1079 | 1086 | parent_ids = commit.parent_ids |
|
1080 | 1087 | else: |
|
1081 | 1088 | parent_ids = commit.get_object().parent_ids |
|
1082 | 1089 | |
|
1083 | 1090 | return [x.hex for x in parent_ids] |
|
1084 | 1091 | return _parents(repo_id, commit_id) |
|
1085 | 1092 | |
|
1086 | 1093 | @reraise_safe_exceptions |
|
1087 | 1094 | def children(self, wire, commit_id): |
|
1088 | 1095 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
1089 | 1096 | region = self._region(wire) |
|
1090 | 1097 | |
|
1091 | 1098 | head = self.head(wire) |
|
1092 | 1099 | |
|
1093 | 1100 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1094 | 1101 | def _children(_repo_id, _commit_id): |
|
1095 | 1102 | |
|
1096 | 1103 | output, __ = self.run_git_command( |
|
1097 | 1104 | wire, ['rev-list', '--all', '--children', f'{commit_id}^..{head}']) |
|
1098 | 1105 | |
|
1099 | 1106 | child_ids = [] |
|
1100 | 1107 | pat = re.compile(fr'^{commit_id}') |
|
1101 | 1108 | for line in output.splitlines(): |
|
1102 | 1109 | line = safe_str(line) |
|
1103 | 1110 | if pat.match(line): |
|
1104 | 1111 | found_ids = line.split(' ')[1:] |
|
1105 | 1112 | child_ids.extend(found_ids) |
|
1106 | 1113 | break |
|
1107 | 1114 | |
|
1108 | 1115 | return child_ids |
|
1109 | 1116 | return _children(repo_id, commit_id) |
|
1110 | 1117 | |
|
1111 | 1118 | @reraise_safe_exceptions |
|
1112 | 1119 | def set_refs(self, wire, key, value): |
|
1113 | 1120 | repo_init = self._factory.repo_libgit2(wire) |
|
1114 | 1121 | with repo_init as repo: |
|
1115 | 1122 | repo.references.create(key, value, force=True) |
|
1116 | 1123 | |
|
1117 | 1124 | @reraise_safe_exceptions |
|
1118 | 1125 | def update_refs(self, wire, key, value): |
|
1119 | 1126 | repo_init = self._factory.repo_libgit2(wire) |
|
1120 | 1127 | with repo_init as repo: |
|
1121 | 1128 | if key not in repo.references: |
|
1122 | 1129 | raise ValueError(f'Reference {key} not found in the repository') |
|
1123 | 1130 | repo.references.create(key, value, force=True) |
|
1124 | 1131 | |
|
1125 | 1132 | @reraise_safe_exceptions |
|
1126 | 1133 | def create_branch(self, wire, branch_name, commit_id, force=False): |
|
1127 | 1134 | repo_init = self._factory.repo_libgit2(wire) |
|
1128 | 1135 | with repo_init as repo: |
|
1129 | 1136 | if commit_id: |
|
1130 | 1137 | commit = repo[commit_id] |
|
1131 | 1138 | else: |
|
1132 | 1139 | # if commit is not given just use the HEAD |
|
1133 | 1140 | commit = repo.head() |
|
1134 | 1141 | |
|
1135 | 1142 | if force: |
|
1136 | 1143 | repo.branches.local.create(branch_name, commit, force=force) |
|
1137 | 1144 | elif not repo.branches.get(branch_name): |
|
1138 | 1145 | # create only if that branch isn't existing |
|
1139 | 1146 | repo.branches.local.create(branch_name, commit, force=force) |
|
1140 | 1147 | |
|
1141 | 1148 | @reraise_safe_exceptions |
|
1142 | 1149 | def remove_ref(self, wire, key): |
|
1143 | 1150 | repo_init = self._factory.repo_libgit2(wire) |
|
1144 | 1151 | with repo_init as repo: |
|
1145 | 1152 | repo.references.delete(key) |
|
1146 | 1153 | |
|
1147 | 1154 | @reraise_safe_exceptions |
|
1148 | 1155 | def tag_remove(self, wire, tag_name): |
|
1149 | 1156 | repo_init = self._factory.repo_libgit2(wire) |
|
1150 | 1157 | with repo_init as repo: |
|
1151 | 1158 | key = f'refs/tags/{tag_name}' |
|
1152 | 1159 | repo.references.delete(key) |
|
1153 | 1160 | |
|
1154 | 1161 | @reraise_safe_exceptions |
|
1155 | 1162 | def tree_changes(self, wire, source_id, target_id): |
|
1156 | 1163 | repo = self._factory.repo(wire) |
|
1157 | 1164 | # source can be empty |
|
1158 | 1165 | source_id = safe_bytes(source_id if source_id else b'') |
|
1159 | 1166 | target_id = safe_bytes(target_id) |
|
1160 | 1167 | |
|
1161 | 1168 | source = repo[source_id].tree if source_id else None |
|
1162 | 1169 | target = repo[target_id].tree |
|
1163 | 1170 | result = repo.object_store.tree_changes(source, target) |
|
1164 | 1171 | |
|
1165 | 1172 | added = set() |
|
1166 | 1173 | modified = set() |
|
1167 | 1174 | deleted = set() |
|
1168 | 1175 | for (old_path, new_path), (_, _), (_, _) in list(result): |
|
1169 | 1176 | if new_path and old_path: |
|
1170 | 1177 | modified.add(new_path) |
|
1171 | 1178 | elif new_path and not old_path: |
|
1172 | 1179 | added.add(new_path) |
|
1173 | 1180 | elif not new_path and old_path: |
|
1174 | 1181 | deleted.add(old_path) |
|
1175 | 1182 | |
|
1176 | 1183 | return list(added), list(modified), list(deleted) |
|
1177 | 1184 | |
|
1178 | 1185 | @reraise_safe_exceptions |
|
1179 | 1186 | def tree_and_type_for_path(self, wire, commit_id, path): |
|
1180 | 1187 | |
|
1181 | 1188 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
1182 | 1189 | region = self._region(wire) |
|
1183 | 1190 | |
|
1184 | 1191 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1185 | 1192 | def _tree_and_type_for_path(_context_uid, _repo_id, _commit_id, _path): |
|
1186 | 1193 | repo_init = self._factory.repo_libgit2(wire) |
|
1187 | 1194 | |
|
1188 | 1195 | with repo_init as repo: |
|
1189 | 1196 | commit = repo[commit_id] |
|
1190 | 1197 | try: |
|
1191 | 1198 | tree = commit.tree[path] |
|
1192 | 1199 | except KeyError: |
|
1193 | 1200 | return None, None, None |
|
1194 | 1201 | |
|
1195 | 1202 | return tree.id.hex, tree.type_str, tree.filemode |
|
1196 | 1203 | return _tree_and_type_for_path(context_uid, repo_id, commit_id, path) |
|
1197 | 1204 | |
|
1198 | 1205 | @reraise_safe_exceptions |
|
1199 | 1206 | def tree_items(self, wire, tree_id): |
|
1200 | 1207 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
1201 | 1208 | region = self._region(wire) |
|
1202 | 1209 | |
|
1203 | 1210 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1204 | 1211 | def _tree_items(_repo_id, _tree_id): |
|
1205 | 1212 | |
|
1206 | 1213 | repo_init = self._factory.repo_libgit2(wire) |
|
1207 | 1214 | with repo_init as repo: |
|
1208 | 1215 | try: |
|
1209 | 1216 | tree = repo[tree_id] |
|
1210 | 1217 | except KeyError: |
|
1211 | 1218 | raise ObjectMissing(f'No tree with id: {tree_id}') |
|
1212 | 1219 | |
|
1213 | 1220 | result = [] |
|
1214 | 1221 | for item in tree: |
|
1215 | 1222 | item_sha = item.hex |
|
1216 | 1223 | item_mode = item.filemode |
|
1217 | 1224 | item_type = item.type_str |
|
1218 | 1225 | |
|
1219 | 1226 | if item_type == 'commit': |
|
1220 | 1227 | # NOTE(marcink): submodules we translate to 'link' for backward compat |
|
1221 | 1228 | item_type = 'link' |
|
1222 | 1229 | |
|
1223 | 1230 | result.append((item.name, item_mode, item_sha, item_type)) |
|
1224 | 1231 | return result |
|
1225 | 1232 | return _tree_items(repo_id, tree_id) |
|
1226 | 1233 | |
|
1227 | 1234 | @reraise_safe_exceptions |
|
1228 | 1235 | def diff_2(self, wire, commit_id_1, commit_id_2, file_filter, opt_ignorews, context): |
|
1229 | 1236 | """ |
|
1230 | 1237 | Old version that uses subprocess to call diff |
|
1231 | 1238 | """ |
|
1232 | 1239 | |
|
1233 | 1240 | flags = [ |
|
1234 | 1241 | f'-U{context}', '--patch', |
|
1235 | 1242 | '--binary', |
|
1236 | 1243 | '--find-renames', |
|
1237 | 1244 | '--no-indent-heuristic', |
|
1238 | 1245 | # '--indent-heuristic', |
|
1239 | 1246 | #'--full-index', |
|
1240 | 1247 | #'--abbrev=40' |
|
1241 | 1248 | ] |
|
1242 | 1249 | |
|
1243 | 1250 | if opt_ignorews: |
|
1244 | 1251 | flags.append('--ignore-all-space') |
|
1245 | 1252 | |
|
1246 | 1253 | if commit_id_1 == self.EMPTY_COMMIT: |
|
1247 | 1254 | cmd = ['show'] + flags + [commit_id_2] |
|
1248 | 1255 | else: |
|
1249 | 1256 | cmd = ['diff'] + flags + [commit_id_1, commit_id_2] |
|
1250 | 1257 | |
|
1251 | 1258 | if file_filter: |
|
1252 | 1259 | cmd.extend(['--', file_filter]) |
|
1253 | 1260 | |
|
1254 | 1261 | diff, __ = self.run_git_command(wire, cmd) |
|
1255 | 1262 | # If we used 'show' command, strip first few lines (until actual diff |
|
1256 | 1263 | # starts) |
|
1257 | 1264 | if commit_id_1 == self.EMPTY_COMMIT: |
|
1258 | 1265 | lines = diff.splitlines() |
|
1259 | 1266 | x = 0 |
|
1260 | 1267 | for line in lines: |
|
1261 | 1268 | if line.startswith(b'diff'): |
|
1262 | 1269 | break |
|
1263 | 1270 | x += 1 |
|
1264 | 1271 | # Append new line just like 'diff' command do |
|
1265 | 1272 | diff = '\n'.join(lines[x:]) + '\n' |
|
1266 | 1273 | return diff |
|
1267 | 1274 | |
|
1268 | 1275 | @reraise_safe_exceptions |
|
1269 | 1276 | def diff(self, wire, commit_id_1, commit_id_2, file_filter, opt_ignorews, context): |
|
1270 | 1277 | repo_init = self._factory.repo_libgit2(wire) |
|
1271 | 1278 | |
|
1272 | 1279 | with repo_init as repo: |
|
1273 | 1280 | swap = True |
|
1274 | 1281 | flags = 0 |
|
1275 | 1282 | flags |= pygit2.GIT_DIFF_SHOW_BINARY |
|
1276 | 1283 | |
|
1277 | 1284 | if opt_ignorews: |
|
1278 | 1285 | flags |= pygit2.GIT_DIFF_IGNORE_WHITESPACE |
|
1279 | 1286 | |
|
1280 | 1287 | if commit_id_1 == self.EMPTY_COMMIT: |
|
1281 | 1288 | comm1 = repo[commit_id_2] |
|
1282 | 1289 | diff_obj = comm1.tree.diff_to_tree( |
|
1283 | 1290 | flags=flags, context_lines=context, swap=swap) |
|
1284 | 1291 | |
|
1285 | 1292 | else: |
|
1286 | 1293 | comm1 = repo[commit_id_2] |
|
1287 | 1294 | comm2 = repo[commit_id_1] |
|
1288 | 1295 | diff_obj = comm1.tree.diff_to_tree( |
|
1289 | 1296 | comm2.tree, flags=flags, context_lines=context, swap=swap) |
|
1290 | 1297 | similar_flags = 0 |
|
1291 | 1298 | similar_flags |= pygit2.GIT_DIFF_FIND_RENAMES |
|
1292 | 1299 | diff_obj.find_similar(flags=similar_flags) |
|
1293 | 1300 | |
|
1294 | 1301 | if file_filter: |
|
1295 | 1302 | for p in diff_obj: |
|
1296 | 1303 | if p.delta.old_file.path == file_filter: |
|
1297 | 1304 | return BytesEnvelope(p.data) or BytesEnvelope(b'') |
|
1298 | 1305 | # fo matching path == no diff |
|
1299 | 1306 | return BytesEnvelope(b'') |
|
1300 | 1307 | |
|
1301 | 1308 | return BytesEnvelope(safe_bytes(diff_obj.patch)) or BytesEnvelope(b'') |
|
1302 | 1309 | |
|
1303 | 1310 | @reraise_safe_exceptions |
|
1304 | 1311 | def node_history(self, wire, commit_id, path, limit): |
|
1305 | 1312 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
1306 | 1313 | region = self._region(wire) |
|
1307 | 1314 | |
|
1308 | 1315 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1309 | 1316 | def _node_history(_context_uid, _repo_id, _commit_id, _path, _limit): |
|
1310 | 1317 | # optimize for n==1, rev-list is much faster for that use-case |
|
1311 | 1318 | if limit == 1: |
|
1312 | 1319 | cmd = ['rev-list', '-1', commit_id, '--', path] |
|
1313 | 1320 | else: |
|
1314 | 1321 | cmd = ['log'] |
|
1315 | 1322 | if limit: |
|
1316 | 1323 | cmd.extend(['-n', str(safe_int(limit, 0))]) |
|
1317 | 1324 | cmd.extend(['--pretty=format: %H', '-s', commit_id, '--', path]) |
|
1318 | 1325 | |
|
1319 | 1326 | output, __ = self.run_git_command(wire, cmd) |
|
1320 | 1327 | commit_ids = re.findall(rb'[0-9a-fA-F]{40}', output) |
|
1321 | 1328 | |
|
1322 | 1329 | return [x for x in commit_ids] |
|
1323 | 1330 | return _node_history(context_uid, repo_id, commit_id, path, limit) |
|
1324 | 1331 | |
|
1325 | 1332 | @reraise_safe_exceptions |
|
1326 | 1333 | def node_annotate_legacy(self, wire, commit_id, path): |
|
1327 | 1334 | # note: replaced by pygit2 implementation |
|
1328 | 1335 | cmd = ['blame', '-l', '--root', '-r', commit_id, '--', path] |
|
1329 | 1336 | # -l ==> outputs long shas (and we need all 40 characters) |
|
1330 | 1337 | # --root ==> doesn't put '^' character for boundaries |
|
1331 | 1338 | # -r commit_id ==> blames for the given commit |
|
1332 | 1339 | output, __ = self.run_git_command(wire, cmd) |
|
1333 | 1340 | |
|
1334 | 1341 | result = [] |
|
1335 | 1342 | for i, blame_line in enumerate(output.splitlines()[:-1]): |
|
1336 | 1343 | line_no = i + 1 |
|
1337 | 1344 | blame_commit_id, line = re.split(rb' ', blame_line, 1) |
|
1338 | 1345 | result.append((line_no, blame_commit_id, line)) |
|
1339 | 1346 | |
|
1340 | 1347 | return result |
|
1341 | 1348 | |
|
1342 | 1349 | @reraise_safe_exceptions |
|
1343 | 1350 | def node_annotate(self, wire, commit_id, path): |
|
1344 | 1351 | |
|
1345 | 1352 | result_libgit = [] |
|
1346 | 1353 | repo_init = self._factory.repo_libgit2(wire) |
|
1347 | 1354 | with repo_init as repo: |
|
1348 | 1355 | commit = repo[commit_id] |
|
1349 | 1356 | blame_obj = repo.blame(path, newest_commit=commit_id) |
|
1350 | 1357 | file_content = commit.tree[path].data |
|
1351 | 1358 | for i, line in enumerate(splitnewlines(file_content)): |
|
1352 | 1359 | line_no = i + 1 |
|
1353 | 1360 | hunk = blame_obj.for_line(line_no) |
|
1354 | 1361 | blame_commit_id = hunk.final_commit_id.hex |
|
1355 | 1362 | |
|
1356 | 1363 | result_libgit.append((line_no, blame_commit_id, line)) |
|
1357 | 1364 | |
|
1358 | 1365 | return BinaryEnvelope(result_libgit) |
|
1359 | 1366 | |
|
1360 | 1367 | @reraise_safe_exceptions |
|
1361 | 1368 | def update_server_info(self, wire, force=False): |
|
1362 | 1369 | cmd = ['update-server-info'] |
|
1363 | 1370 | if force: |
|
1364 | 1371 | cmd += ['--force'] |
|
1365 | 1372 | output, __ = self.run_git_command(wire, cmd) |
|
1366 | 1373 | return output.splitlines() |
|
1367 | 1374 | |
|
1368 | 1375 | @reraise_safe_exceptions |
|
1369 | 1376 | def get_all_commit_ids(self, wire): |
|
1370 | 1377 | |
|
1371 | 1378 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
1372 | 1379 | region = self._region(wire) |
|
1373 | 1380 | |
|
1374 | 1381 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1375 | 1382 | def _get_all_commit_ids(_context_uid, _repo_id): |
|
1376 | 1383 | |
|
1377 | 1384 | cmd = ['rev-list', '--reverse', '--date-order', '--branches', '--tags'] |
|
1378 | 1385 | try: |
|
1379 | 1386 | output, __ = self.run_git_command(wire, cmd) |
|
1380 | 1387 | return output.splitlines() |
|
1381 | 1388 | except Exception: |
|
1382 | 1389 | # Can be raised for empty repositories |
|
1383 | 1390 | return [] |
|
1384 | 1391 | |
|
1385 | 1392 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
1386 | 1393 | def _get_all_commit_ids_pygit2(_context_uid, _repo_id): |
|
1387 | 1394 | repo_init = self._factory.repo_libgit2(wire) |
|
1388 | 1395 | from pygit2 import GIT_SORT_REVERSE, GIT_SORT_TIME, GIT_BRANCH_ALL |
|
1389 | 1396 | results = [] |
|
1390 | 1397 | with repo_init as repo: |
|
1391 | 1398 | for commit in repo.walk(repo.head.target, GIT_SORT_TIME | GIT_BRANCH_ALL | GIT_SORT_REVERSE): |
|
1392 | 1399 | results.append(commit.id.hex) |
|
1393 | 1400 | |
|
1394 | 1401 | return _get_all_commit_ids(context_uid, repo_id) |
|
1395 | 1402 | |
|
1396 | 1403 | @reraise_safe_exceptions |
|
1397 | 1404 | def run_git_command(self, wire, cmd, **opts): |
|
1398 | 1405 | path = wire.get('path', None) |
|
1399 | 1406 | debug_mode = rhodecode.ConfigGet().get_bool('debug') |
|
1400 | 1407 | |
|
1401 | 1408 | if path and os.path.isdir(path): |
|
1402 | 1409 | opts['cwd'] = path |
|
1403 | 1410 | |
|
1404 | 1411 | if '_bare' in opts: |
|
1405 | 1412 | _copts = [] |
|
1406 | 1413 | del opts['_bare'] |
|
1407 | 1414 | else: |
|
1408 | 1415 | _copts = ['-c', 'core.quotepath=false', '-c', 'advice.diverging=false'] |
|
1409 | 1416 | safe_call = False |
|
1410 | 1417 | if '_safe' in opts: |
|
1411 | 1418 | # no exc on failure |
|
1412 | 1419 | del opts['_safe'] |
|
1413 | 1420 | safe_call = True |
|
1414 | 1421 | |
|
1415 | 1422 | if '_copts' in opts: |
|
1416 | 1423 | _copts.extend(opts['_copts'] or []) |
|
1417 | 1424 | del opts['_copts'] |
|
1418 | 1425 | |
|
1419 | 1426 | gitenv = os.environ.copy() |
|
1420 | 1427 | gitenv.update(opts.pop('extra_env', {})) |
|
1421 | 1428 | # need to clean fix GIT_DIR ! |
|
1422 | 1429 | if 'GIT_DIR' in gitenv: |
|
1423 | 1430 | del gitenv['GIT_DIR'] |
|
1424 | 1431 | gitenv['GIT_CONFIG_NOGLOBAL'] = '1' |
|
1425 | 1432 | gitenv['GIT_DISCOVERY_ACROSS_FILESYSTEM'] = '1' |
|
1426 | 1433 | |
|
1427 | 1434 | cmd = [settings.GIT_EXECUTABLE()] + _copts + cmd |
|
1428 | 1435 | _opts = {'env': gitenv, 'shell': False} |
|
1429 | 1436 | |
|
1430 | 1437 | proc = None |
|
1431 | 1438 | try: |
|
1432 | 1439 | _opts.update(opts) |
|
1433 | 1440 | proc = subprocessio.SubprocessIOChunker(cmd, **_opts) |
|
1434 | 1441 | |
|
1435 | 1442 | return b''.join(proc), b''.join(proc.stderr) |
|
1436 | 1443 | except OSError as err: |
|
1437 | 1444 | cmd = ' '.join(map(safe_str, cmd)) # human friendly CMD |
|
1438 | 1445 | call_opts = {} |
|
1439 | 1446 | if debug_mode: |
|
1440 | 1447 | call_opts = _opts |
|
1441 | 1448 | |
|
1442 | 1449 | tb_err = ("Couldn't run git command ({}).\n" |
|
1443 | 1450 | "Original error was:{}\n" |
|
1444 | 1451 | "Call options:{}\n" |
|
1445 | 1452 | .format(cmd, err, call_opts)) |
|
1446 | 1453 | log.exception(tb_err) |
|
1447 | 1454 | if safe_call: |
|
1448 | 1455 | return '', err |
|
1449 | 1456 | else: |
|
1450 | 1457 | raise exceptions.VcsException()(tb_err) |
|
1451 | 1458 | finally: |
|
1452 | 1459 | if proc: |
|
1453 | 1460 | proc.close() |
|
1454 | 1461 | |
|
1455 | 1462 | @reraise_safe_exceptions |
|
1456 | 1463 | def install_hooks(self, wire, force=False): |
|
1457 | 1464 | from vcsserver.hook_utils import install_git_hooks |
|
1458 | 1465 | bare = self.bare(wire) |
|
1459 | 1466 | path = wire['path'] |
|
1460 | 1467 | binary_dir = settings.BINARY_DIR |
|
1461 | 1468 | if binary_dir: |
|
1462 | 1469 | os.path.join(binary_dir, 'python3') |
|
1463 | 1470 | return install_git_hooks(path, bare, force_create=force) |
|
1464 | 1471 | |
|
1465 | 1472 | @reraise_safe_exceptions |
|
1466 | 1473 | def get_hooks_info(self, wire): |
|
1467 | 1474 | from vcsserver.hook_utils import ( |
|
1468 | 1475 | get_git_pre_hook_version, get_git_post_hook_version) |
|
1469 | 1476 | bare = self.bare(wire) |
|
1470 | 1477 | path = wire['path'] |
|
1471 | 1478 | return { |
|
1472 | 1479 | 'pre_version': get_git_pre_hook_version(path, bare), |
|
1473 | 1480 | 'post_version': get_git_post_hook_version(path, bare), |
|
1474 | 1481 | } |
|
1475 | 1482 | |
|
1476 | 1483 | @reraise_safe_exceptions |
|
1477 | 1484 | def set_head_ref(self, wire, head_name): |
|
1478 | 1485 | log.debug('Setting refs/head to `%s`', head_name) |
|
1479 | 1486 | repo_init = self._factory.repo_libgit2(wire) |
|
1480 | 1487 | with repo_init as repo: |
|
1481 | 1488 | repo.set_head(f'refs/heads/{head_name}') |
|
1482 | 1489 | |
|
1483 | 1490 | return [head_name] + [f'set HEAD to refs/heads/{head_name}'] |
|
1484 | 1491 | |
|
1485 | 1492 | @reraise_safe_exceptions |
|
1486 | 1493 | def archive_repo(self, wire, archive_name_key, kind, mtime, archive_at_path, |
|
1487 | 1494 | archive_dir_name, commit_id, cache_config): |
|
1488 | 1495 | |
|
1489 | 1496 | def file_walker(_commit_id, path): |
|
1490 | 1497 | repo_init = self._factory.repo_libgit2(wire) |
|
1491 | 1498 | |
|
1492 | 1499 | with repo_init as repo: |
|
1493 | 1500 | commit = repo[commit_id] |
|
1494 | 1501 | |
|
1495 | 1502 | if path in ['', '/']: |
|
1496 | 1503 | tree = commit.tree |
|
1497 | 1504 | else: |
|
1498 | 1505 | tree = commit.tree[path.rstrip('/')] |
|
1499 | 1506 | tree_id = tree.id.hex |
|
1500 | 1507 | try: |
|
1501 | 1508 | tree = repo[tree_id] |
|
1502 | 1509 | except KeyError: |
|
1503 | 1510 | raise ObjectMissing(f'No tree with id: {tree_id}') |
|
1504 | 1511 | |
|
1505 | 1512 | index = LibGit2Index.Index() |
|
1506 | 1513 | index.read_tree(tree) |
|
1507 | 1514 | file_iter = index |
|
1508 | 1515 | |
|
1509 | 1516 | for file_node in file_iter: |
|
1510 | 1517 | file_path = file_node.path |
|
1511 | 1518 | mode = file_node.mode |
|
1512 | 1519 | is_link = stat.S_ISLNK(mode) |
|
1513 | 1520 | if mode == pygit2.GIT_FILEMODE_COMMIT: |
|
1514 | 1521 | log.debug('Skipping path %s as a commit node', file_path) |
|
1515 | 1522 | continue |
|
1516 | 1523 | yield ArchiveNode(file_path, mode, is_link, repo[file_node.hex].read_raw) |
|
1517 | 1524 | |
|
1518 | 1525 | return store_archive_in_cache( |
|
1519 | 1526 | file_walker, archive_name_key, kind, mtime, archive_at_path, archive_dir_name, commit_id, cache_config=cache_config) |
@@ -1,1213 +1,1217 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import binascii |
|
19 | 19 | import io |
|
20 | 20 | import logging |
|
21 | 21 | import stat |
|
22 | 22 | import sys |
|
23 | 23 | import urllib.request |
|
24 | 24 | import urllib.parse |
|
25 | 25 | import hashlib |
|
26 | 26 | |
|
27 | 27 | from hgext import largefiles, rebase |
|
28 | 28 | |
|
29 | 29 | from mercurial import commands |
|
30 | 30 | from mercurial import unionrepo |
|
31 | 31 | from mercurial import verify |
|
32 | 32 | from mercurial import repair |
|
33 | 33 | from mercurial.error import AmbiguousPrefixLookupError |
|
34 | from mercurial.utils.urlutil import path as hg_path | |
|
34 | 35 | |
|
35 | 36 | import vcsserver |
|
36 | 37 | from vcsserver import exceptions |
|
37 | 38 | from vcsserver.base import ( |
|
38 | 39 | RepoFactory, |
|
39 | 40 | obfuscate_qs, |
|
40 | 41 | raise_from_original, |
|
41 | 42 | store_archive_in_cache, |
|
42 | 43 | ArchiveNode, |
|
43 | 44 | BytesEnvelope, |
|
44 | 45 | BinaryEnvelope, |
|
45 | 46 | ) |
|
46 | 47 | from vcsserver.hgcompat import ( |
|
47 | 48 | archival, |
|
48 | 49 | bin, |
|
49 | 50 | clone, |
|
50 | 51 | config as hgconfig, |
|
51 | 52 | diffopts, |
|
52 | 53 | hex, |
|
53 | 54 | get_ctx, |
|
54 | 55 | hg_url as url_parser, |
|
55 | 56 | httpbasicauthhandler, |
|
56 | 57 | httpdigestauthhandler, |
|
57 | makepeer, | |
|
58 | make_peer, | |
|
58 | 59 | instance, |
|
59 | 60 | match, |
|
60 | 61 | memctx, |
|
61 | 62 | exchange, |
|
62 | 63 | memfilectx, |
|
63 | 64 | nullrev, |
|
64 | 65 | hg_merge, |
|
65 | 66 | patch, |
|
66 | 67 | peer, |
|
67 | 68 | revrange, |
|
68 | 69 | ui, |
|
69 | 70 | hg_tag, |
|
70 | 71 | Abort, |
|
71 | 72 | LookupError, |
|
72 | 73 | RepoError, |
|
73 | 74 | RepoLookupError, |
|
74 | 75 | InterventionRequired, |
|
75 | 76 | RequirementError, |
|
76 | 77 | alwaysmatcher, |
|
77 | 78 | patternmatcher, |
|
78 | 79 | hgext_strip, |
|
79 | 80 | ) |
|
80 | from vcsserver.str_utils import ascii_bytes, ascii_str, safe_str, safe_bytes, convert_to_str | |
|
81 | from vcsserver.lib.str_utils import ascii_bytes, ascii_str, safe_str, safe_bytes, convert_to_str | |
|
81 | 82 | from vcsserver.vcs_base import RemoteBase |
|
82 | 83 | from vcsserver.config import hooks as hooks_config |
|
83 | 84 | from vcsserver.lib.exc_tracking import format_exc |
|
84 | 85 | |
|
85 | 86 | log = logging.getLogger(__name__) |
|
86 | 87 | |
|
87 | 88 | |
|
88 | def make_ui_from_config(repo_config): | |
|
89 | def make_ui_from_config(repo_config, interactive=True): | |
|
89 | 90 | |
|
90 | 91 | class LoggingUI(ui.ui): |
|
91 | 92 | |
|
92 | 93 | def status(self, *msg, **opts): |
|
93 | 94 | str_msg = map(safe_str, msg) |
|
94 | 95 | log.info(' '.join(str_msg).rstrip('\n')) |
|
95 | 96 | #super(LoggingUI, self).status(*msg, **opts) |
|
96 | 97 | |
|
97 | 98 | def warn(self, *msg, **opts): |
|
98 | 99 | str_msg = map(safe_str, msg) |
|
99 | 100 | log.warning('ui_logger:'+' '.join(str_msg).rstrip('\n')) |
|
100 | 101 | #super(LoggingUI, self).warn(*msg, **opts) |
|
101 | 102 | |
|
102 | 103 | def error(self, *msg, **opts): |
|
103 | 104 | str_msg = map(safe_str, msg) |
|
104 | 105 | log.error('ui_logger:'+' '.join(str_msg).rstrip('\n')) |
|
105 | 106 | #super(LoggingUI, self).error(*msg, **opts) |
|
106 | 107 | |
|
107 | 108 | def note(self, *msg, **opts): |
|
108 | 109 | str_msg = map(safe_str, msg) |
|
109 | 110 | log.info('ui_logger:'+' '.join(str_msg).rstrip('\n')) |
|
110 | 111 | #super(LoggingUI, self).note(*msg, **opts) |
|
111 | 112 | |
|
112 | 113 | def debug(self, *msg, **opts): |
|
113 | 114 | str_msg = map(safe_str, msg) |
|
114 | 115 | log.debug('ui_logger:'+' '.join(str_msg).rstrip('\n')) |
|
115 | 116 | #super(LoggingUI, self).debug(*msg, **opts) |
|
116 | 117 | |
|
117 | 118 | baseui = LoggingUI() |
|
118 | 119 | |
|
119 | 120 | # clean the baseui object |
|
120 | 121 | baseui._ocfg = hgconfig.config() |
|
121 | 122 | baseui._ucfg = hgconfig.config() |
|
122 | 123 | baseui._tcfg = hgconfig.config() |
|
123 | 124 | |
|
124 | 125 | for section, option, value in repo_config: |
|
125 | 126 | baseui.setconfig(ascii_bytes(section), ascii_bytes(option), ascii_bytes(value)) |
|
126 | 127 | |
|
127 | 128 | # make our hgweb quiet so it doesn't print output |
|
128 | 129 | baseui.setconfig(b'ui', b'quiet', b'true') |
|
129 | 130 | |
|
130 | 131 | baseui.setconfig(b'ui', b'paginate', b'never') |
|
131 | 132 | # for better Error reporting of Mercurial |
|
132 | 133 | baseui.setconfig(b'ui', b'message-output', b'stderr') |
|
133 | 134 | |
|
134 | 135 | # force mercurial to only use 1 thread, otherwise it may try to set a |
|
135 | 136 | # signal in a non-main thread, thus generating a ValueError. |
|
136 | 137 | baseui.setconfig(b'worker', b'numcpus', 1) |
|
137 | 138 | |
|
138 | 139 | # If there is no config for the largefiles extension, we explicitly disable |
|
139 | 140 | # it here. This overrides settings from repositories hgrc file. Recent |
|
140 | 141 | # mercurial versions enable largefiles in hgrc on clone from largefile |
|
141 | 142 | # repo. |
|
142 | 143 | if not baseui.hasconfig(b'extensions', b'largefiles'): |
|
143 | 144 | log.debug('Explicitly disable largefiles extension for repo.') |
|
144 | 145 | baseui.setconfig(b'extensions', b'largefiles', b'!') |
|
145 | 146 | |
|
147 | baseui.setconfig(b'ui', b'interactive', b'true' if interactive else b'false') | |
|
146 | 148 | return baseui |
|
147 | 149 | |
|
148 | 150 | |
|
149 | 151 | def reraise_safe_exceptions(func): |
|
150 | 152 | """Decorator for converting mercurial exceptions to something neutral.""" |
|
151 | 153 | |
|
152 | 154 | def wrapper(*args, **kwargs): |
|
153 | 155 | try: |
|
154 | 156 | return func(*args, **kwargs) |
|
155 | 157 | except (Abort, InterventionRequired) as e: |
|
156 | 158 | raise_from_original(exceptions.AbortException(e), e) |
|
157 | 159 | except RepoLookupError as e: |
|
158 | 160 | raise_from_original(exceptions.LookupException(e), e) |
|
159 | 161 | except RequirementError as e: |
|
160 | 162 | raise_from_original(exceptions.RequirementException(e), e) |
|
161 | 163 | except RepoError as e: |
|
162 | 164 | raise_from_original(exceptions.VcsException(e), e) |
|
163 | 165 | except LookupError as e: |
|
164 | 166 | raise_from_original(exceptions.LookupException(e), e) |
|
165 | 167 | except Exception as e: |
|
166 | 168 | if not hasattr(e, '_vcs_kind'): |
|
167 | 169 | log.exception("Unhandled exception in hg remote call") |
|
168 | 170 | raise_from_original(exceptions.UnhandledException(e), e) |
|
169 | 171 | |
|
170 | 172 | raise |
|
171 | 173 | return wrapper |
|
172 | 174 | |
|
173 | 175 | |
|
174 | 176 | class MercurialFactory(RepoFactory): |
|
175 | 177 | repo_type = 'hg' |
|
176 | 178 | |
|
177 | 179 | def _create_config(self, config, hooks=True): |
|
178 | 180 | if not hooks: |
|
179 | 181 | |
|
180 | 182 | hooks_to_clean = { |
|
181 | 183 | |
|
182 | 184 | hooks_config.HOOK_REPO_SIZE, |
|
183 | 185 | hooks_config.HOOK_PRE_PULL, |
|
184 | 186 | hooks_config.HOOK_PULL, |
|
185 | 187 | |
|
186 | 188 | hooks_config.HOOK_PRE_PUSH, |
|
187 | 189 | # TODO: what about PRETXT, this was disabled in pre 5.0.0 |
|
188 | 190 | hooks_config.HOOK_PRETX_PUSH, |
|
189 | 191 | |
|
190 | 192 | } |
|
191 | 193 | new_config = [] |
|
192 | 194 | for section, option, value in config: |
|
193 | 195 | if section == 'hooks' and option in hooks_to_clean: |
|
194 | 196 | continue |
|
195 | 197 | new_config.append((section, option, value)) |
|
196 | 198 | config = new_config |
|
197 | 199 | |
|
198 | 200 | baseui = make_ui_from_config(config) |
|
199 | 201 | return baseui |
|
200 | 202 | |
|
201 | 203 | def _create_repo(self, wire, create): |
|
202 | 204 | baseui = self._create_config(wire["config"]) |
|
203 | 205 | repo = instance(baseui, safe_bytes(wire["path"]), create) |
|
204 | 206 | log.debug('repository created: got HG object: %s', repo) |
|
205 | 207 | return repo |
|
206 | 208 | |
|
207 | 209 | def repo(self, wire, create=False): |
|
208 | 210 | """ |
|
209 | 211 | Get a repository instance for the given path. |
|
210 | 212 | """ |
|
211 | 213 | return self._create_repo(wire, create) |
|
212 | 214 | |
|
213 | 215 | |
|
214 | 216 | def patch_ui_message_output(baseui): |
|
215 | 217 | baseui.setconfig(b'ui', b'quiet', b'false') |
|
216 | 218 | output = io.BytesIO() |
|
217 | 219 | |
|
218 | 220 | def write(data, **unused_kwargs): |
|
219 | 221 | output.write(data) |
|
220 | 222 | |
|
221 | 223 | baseui.status = write |
|
222 | 224 | baseui.write = write |
|
223 | 225 | baseui.warn = write |
|
224 | 226 | baseui.debug = write |
|
225 | 227 | |
|
226 | 228 | return baseui, output |
|
227 | 229 | |
|
228 | 230 | |
|
229 | 231 | def get_obfuscated_url(url_obj): |
|
230 | 232 | url_obj.passwd = b'*****' if url_obj.passwd else url_obj.passwd |
|
231 | 233 | url_obj.query = obfuscate_qs(url_obj.query) |
|
232 | 234 | obfuscated_uri = str(url_obj) |
|
233 | 235 | return obfuscated_uri |
|
234 | 236 | |
|
235 | 237 | |
|
236 | 238 | def normalize_url_for_hg(url: str): |
|
237 | 239 | _proto = None |
|
238 | 240 | |
|
239 | 241 | if '+' in url[:url.find('://')]: |
|
240 | 242 | _proto = url[0:url.find('+')] |
|
241 | 243 | url = url[url.find('+') + 1:] |
|
242 | 244 | return url, _proto |
|
243 | 245 | |
|
244 | 246 | |
|
245 | 247 | class HgRemote(RemoteBase): |
|
246 | 248 | |
|
247 | 249 | def __init__(self, factory): |
|
248 | 250 | self._factory = factory |
|
249 | 251 | self._bulk_methods = { |
|
250 | 252 | "affected_files": self.ctx_files, |
|
251 | 253 | "author": self.ctx_user, |
|
252 | 254 | "branch": self.ctx_branch, |
|
253 | 255 | "children": self.ctx_children, |
|
254 | 256 | "date": self.ctx_date, |
|
255 | 257 | "message": self.ctx_description, |
|
256 | 258 | "parents": self.ctx_parents, |
|
257 | 259 | "status": self.ctx_status, |
|
258 | 260 | "obsolete": self.ctx_obsolete, |
|
259 | 261 | "phase": self.ctx_phase, |
|
260 | 262 | "hidden": self.ctx_hidden, |
|
261 | 263 | "_file_paths": self.ctx_list, |
|
262 | 264 | } |
|
263 | 265 | self._bulk_file_methods = { |
|
264 | 266 | "size": self.fctx_size, |
|
265 | 267 | "data": self.fctx_node_data, |
|
266 | 268 | "flags": self.fctx_flags, |
|
267 | 269 | "is_binary": self.is_binary, |
|
268 | 270 | "md5": self.md5_hash, |
|
269 | 271 | } |
|
270 | 272 | |
|
271 | 273 | def _get_ctx(self, repo, ref): |
|
272 | 274 | return get_ctx(repo, ref) |
|
273 | 275 | |
|
274 | 276 | @reraise_safe_exceptions |
|
275 | 277 | def discover_hg_version(self): |
|
276 | 278 | from mercurial import util |
|
277 | 279 | return safe_str(util.version()) |
|
278 | 280 | |
|
279 | 281 | @reraise_safe_exceptions |
|
280 | 282 | def is_empty(self, wire): |
|
281 | 283 | repo = self._factory.repo(wire) |
|
282 | 284 | |
|
283 | 285 | try: |
|
284 | 286 | return len(repo) == 0 |
|
285 | 287 | except Exception: |
|
286 | 288 | log.exception("failed to read object_store") |
|
287 | 289 | return False |
|
288 | 290 | |
|
289 | 291 | @reraise_safe_exceptions |
|
290 | 292 | def bookmarks(self, wire): |
|
291 | 293 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
292 | 294 | region = self._region(wire) |
|
293 | 295 | |
|
294 | 296 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
295 | 297 | def _bookmarks(_context_uid, _repo_id): |
|
296 | 298 | repo = self._factory.repo(wire) |
|
297 | 299 | return {safe_str(name): ascii_str(hex(sha)) for name, sha in repo._bookmarks.items()} |
|
298 | 300 | |
|
299 | 301 | return _bookmarks(context_uid, repo_id) |
|
300 | 302 | |
|
301 | 303 | @reraise_safe_exceptions |
|
302 | 304 | def branches(self, wire, normal, closed): |
|
303 | 305 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
304 | 306 | region = self._region(wire) |
|
305 | 307 | |
|
306 | 308 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
307 | 309 | def _branches(_context_uid, _repo_id, _normal, _closed): |
|
308 | 310 | repo = self._factory.repo(wire) |
|
309 | 311 | iter_branches = repo.branchmap().iterbranches() |
|
310 | 312 | bt = {} |
|
311 | 313 | for branch_name, _heads, tip_node, is_closed in iter_branches: |
|
312 | 314 | if normal and not is_closed: |
|
313 | 315 | bt[safe_str(branch_name)] = ascii_str(hex(tip_node)) |
|
314 | 316 | if closed and is_closed: |
|
315 | 317 | bt[safe_str(branch_name)] = ascii_str(hex(tip_node)) |
|
316 | 318 | |
|
317 | 319 | return bt |
|
318 | 320 | |
|
319 | 321 | return _branches(context_uid, repo_id, normal, closed) |
|
320 | 322 | |
|
321 | 323 | @reraise_safe_exceptions |
|
322 | 324 | def bulk_request(self, wire, commit_id, pre_load): |
|
323 | 325 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
324 | 326 | region = self._region(wire) |
|
325 | 327 | |
|
326 | 328 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
327 | 329 | def _bulk_request(_repo_id, _commit_id, _pre_load): |
|
328 | 330 | result = {} |
|
329 | 331 | for attr in pre_load: |
|
330 | 332 | try: |
|
331 | 333 | method = self._bulk_methods[attr] |
|
332 | 334 | wire.update({'cache': False}) # disable cache for bulk calls so we don't double cache |
|
333 | 335 | result[attr] = method(wire, commit_id) |
|
334 | 336 | except KeyError as e: |
|
335 | 337 | raise exceptions.VcsException(e)( |
|
336 | 338 | f'Unknown bulk attribute: "{attr}"') |
|
337 | 339 | return result |
|
338 | 340 | |
|
339 | 341 | return _bulk_request(repo_id, commit_id, sorted(pre_load)) |
|
340 | 342 | |
|
341 | 343 | @reraise_safe_exceptions |
|
342 | 344 | def ctx_branch(self, wire, commit_id): |
|
343 | 345 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
344 | 346 | region = self._region(wire) |
|
345 | 347 | |
|
346 | 348 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
347 | 349 | def _ctx_branch(_repo_id, _commit_id): |
|
348 | 350 | repo = self._factory.repo(wire) |
|
349 | 351 | ctx = self._get_ctx(repo, commit_id) |
|
350 | 352 | return ctx.branch() |
|
351 | 353 | return _ctx_branch(repo_id, commit_id) |
|
352 | 354 | |
|
353 | 355 | @reraise_safe_exceptions |
|
354 | 356 | def ctx_date(self, wire, commit_id): |
|
355 | 357 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
356 | 358 | region = self._region(wire) |
|
357 | 359 | |
|
358 | 360 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
359 | 361 | def _ctx_date(_repo_id, _commit_id): |
|
360 | 362 | repo = self._factory.repo(wire) |
|
361 | 363 | ctx = self._get_ctx(repo, commit_id) |
|
362 | 364 | return ctx.date() |
|
363 | 365 | return _ctx_date(repo_id, commit_id) |
|
364 | 366 | |
|
365 | 367 | @reraise_safe_exceptions |
|
366 | 368 | def ctx_description(self, wire, revision): |
|
367 | 369 | repo = self._factory.repo(wire) |
|
368 | 370 | ctx = self._get_ctx(repo, revision) |
|
369 | 371 | return ctx.description() |
|
370 | 372 | |
|
371 | 373 | @reraise_safe_exceptions |
|
372 | 374 | def ctx_files(self, wire, commit_id): |
|
373 | 375 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
374 | 376 | region = self._region(wire) |
|
375 | 377 | |
|
376 | 378 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
377 | 379 | def _ctx_files(_repo_id, _commit_id): |
|
378 | 380 | repo = self._factory.repo(wire) |
|
379 | 381 | ctx = self._get_ctx(repo, commit_id) |
|
380 | 382 | return ctx.files() |
|
381 | 383 | |
|
382 | 384 | return _ctx_files(repo_id, commit_id) |
|
383 | 385 | |
|
384 | 386 | @reraise_safe_exceptions |
|
385 | 387 | def ctx_list(self, path, revision): |
|
386 | 388 | repo = self._factory.repo(path) |
|
387 | 389 | ctx = self._get_ctx(repo, revision) |
|
388 | 390 | return list(ctx) |
|
389 | 391 | |
|
390 | 392 | @reraise_safe_exceptions |
|
391 | 393 | def ctx_parents(self, wire, commit_id): |
|
392 | 394 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
393 | 395 | region = self._region(wire) |
|
394 | 396 | |
|
395 | 397 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
396 | 398 | def _ctx_parents(_repo_id, _commit_id): |
|
397 | 399 | repo = self._factory.repo(wire) |
|
398 | 400 | ctx = self._get_ctx(repo, commit_id) |
|
399 | 401 | return [parent.hex() for parent in ctx.parents() |
|
400 | 402 | if not (parent.hidden() or parent.obsolete())] |
|
401 | 403 | |
|
402 | 404 | return _ctx_parents(repo_id, commit_id) |
|
403 | 405 | |
|
404 | 406 | @reraise_safe_exceptions |
|
405 | 407 | def ctx_children(self, wire, commit_id): |
|
406 | 408 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
407 | 409 | region = self._region(wire) |
|
408 | 410 | |
|
409 | 411 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
410 | 412 | def _ctx_children(_repo_id, _commit_id): |
|
411 | 413 | repo = self._factory.repo(wire) |
|
412 | 414 | ctx = self._get_ctx(repo, commit_id) |
|
413 | 415 | return [child.hex() for child in ctx.children() |
|
414 | 416 | if not (child.hidden() or child.obsolete())] |
|
415 | 417 | |
|
416 | 418 | return _ctx_children(repo_id, commit_id) |
|
417 | 419 | |
|
418 | 420 | @reraise_safe_exceptions |
|
419 | 421 | def ctx_phase(self, wire, commit_id): |
|
420 | 422 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
421 | 423 | region = self._region(wire) |
|
422 | 424 | |
|
423 | 425 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
424 | 426 | def _ctx_phase(_context_uid, _repo_id, _commit_id): |
|
425 | 427 | repo = self._factory.repo(wire) |
|
426 | 428 | ctx = self._get_ctx(repo, commit_id) |
|
427 | 429 | # public=0, draft=1, secret=3 |
|
428 | 430 | return ctx.phase() |
|
429 | 431 | return _ctx_phase(context_uid, repo_id, commit_id) |
|
430 | 432 | |
|
431 | 433 | @reraise_safe_exceptions |
|
432 | 434 | def ctx_obsolete(self, wire, commit_id): |
|
433 | 435 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
434 | 436 | region = self._region(wire) |
|
435 | 437 | |
|
436 | 438 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
437 | 439 | def _ctx_obsolete(_context_uid, _repo_id, _commit_id): |
|
438 | 440 | repo = self._factory.repo(wire) |
|
439 | 441 | ctx = self._get_ctx(repo, commit_id) |
|
440 | 442 | return ctx.obsolete() |
|
441 | 443 | return _ctx_obsolete(context_uid, repo_id, commit_id) |
|
442 | 444 | |
|
443 | 445 | @reraise_safe_exceptions |
|
444 | 446 | def ctx_hidden(self, wire, commit_id): |
|
445 | 447 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
446 | 448 | region = self._region(wire) |
|
447 | 449 | |
|
448 | 450 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
449 | 451 | def _ctx_hidden(_context_uid, _repo_id, _commit_id): |
|
450 | 452 | repo = self._factory.repo(wire) |
|
451 | 453 | ctx = self._get_ctx(repo, commit_id) |
|
452 | 454 | return ctx.hidden() |
|
453 | 455 | return _ctx_hidden(context_uid, repo_id, commit_id) |
|
454 | 456 | |
|
455 | 457 | @reraise_safe_exceptions |
|
456 | 458 | def ctx_substate(self, wire, revision): |
|
457 | 459 | repo = self._factory.repo(wire) |
|
458 | 460 | ctx = self._get_ctx(repo, revision) |
|
459 | 461 | return ctx.substate |
|
460 | 462 | |
|
461 | 463 | @reraise_safe_exceptions |
|
462 | 464 | def ctx_status(self, wire, revision): |
|
463 | 465 | repo = self._factory.repo(wire) |
|
464 | 466 | ctx = self._get_ctx(repo, revision) |
|
465 | 467 | status = repo[ctx.p1().node()].status(other=ctx.node()) |
|
466 | 468 | # object of status (odd, custom named tuple in mercurial) is not |
|
467 | 469 | # correctly serializable, we make it a list, as the underling |
|
468 | 470 | # API expects this to be a list |
|
469 | 471 | return list(status) |
|
470 | 472 | |
|
471 | 473 | @reraise_safe_exceptions |
|
472 | 474 | def ctx_user(self, wire, revision): |
|
473 | 475 | repo = self._factory.repo(wire) |
|
474 | 476 | ctx = self._get_ctx(repo, revision) |
|
475 | 477 | return ctx.user() |
|
476 | 478 | |
|
477 | 479 | @reraise_safe_exceptions |
|
478 | 480 | def check_url(self, url, config): |
|
479 | 481 | url, _proto = normalize_url_for_hg(url) |
|
480 | 482 | url_obj = url_parser(safe_bytes(url)) |
|
481 | 483 | |
|
482 | 484 | test_uri = safe_str(url_obj.authinfo()[0]) |
|
483 | 485 | authinfo = url_obj.authinfo()[1] |
|
484 | 486 | obfuscated_uri = get_obfuscated_url(url_obj) |
|
485 | 487 | log.info("Checking URL for remote cloning/import: %s", obfuscated_uri) |
|
486 | 488 | |
|
487 | 489 | handlers = [] |
|
488 | 490 | if authinfo: |
|
489 | 491 | # create a password manager |
|
490 | 492 | passmgr = urllib.request.HTTPPasswordMgrWithDefaultRealm() |
|
491 | 493 | passmgr.add_password(*convert_to_str(authinfo)) |
|
492 | 494 | |
|
493 | 495 | handlers.extend((httpbasicauthhandler(passmgr), |
|
494 | 496 | httpdigestauthhandler(passmgr))) |
|
495 | 497 | |
|
496 | 498 | o = urllib.request.build_opener(*handlers) |
|
497 | 499 | o.addheaders = [('Content-Type', 'application/mercurial-0.1'), |
|
498 | 500 | ('Accept', 'application/mercurial-0.1')] |
|
499 | 501 | |
|
500 | 502 | q = {"cmd": 'between'} |
|
501 | 503 | q.update({'pairs': "{}-{}".format('0' * 40, '0' * 40)}) |
|
502 | 504 | qs = f'?{urllib.parse.urlencode(q)}' |
|
503 | 505 | cu = f"{test_uri}{qs}" |
|
504 | 506 | |
|
505 | 507 | try: |
|
506 | 508 | req = urllib.request.Request(cu, None, {}) |
|
507 | 509 | log.debug("Trying to open URL %s", obfuscated_uri) |
|
508 | 510 | resp = o.open(req) |
|
509 | 511 | if resp.code != 200: |
|
510 | 512 | raise exceptions.URLError()('Return Code is not 200') |
|
511 | 513 | except Exception as e: |
|
512 | 514 | log.warning("URL cannot be opened: %s", obfuscated_uri, exc_info=True) |
|
513 | 515 | # means it cannot be cloned |
|
514 | 516 | raise exceptions.URLError(e)(f"[{obfuscated_uri}] org_exc: {e}") |
|
515 | 517 | |
|
516 | 518 | # now check if it's a proper hg repo, but don't do it for svn |
|
517 | 519 | try: |
|
518 | 520 | if _proto == 'svn': |
|
519 | 521 | pass |
|
520 | 522 | else: |
|
521 | 523 | # check for pure hg repos |
|
522 | 524 | log.debug( |
|
523 | 525 | "Verifying if URL is a Mercurial repository: %s", obfuscated_uri) |
|
524 | ui = make_ui_from_config(config) | |
|
525 | peer_checker = makepeer(ui, safe_bytes(url)) | |
|
526 | # Create repo path with custom mercurial path object | |
|
527 | ui = make_ui_from_config(config, interactive=False) | |
|
528 | repo_path = hg_path(ui=ui, rawloc=safe_bytes(url)) | |
|
529 | peer_checker = make_peer(ui, repo_path, False) | |
|
526 | 530 | peer_checker.lookup(b'tip') |
|
527 | 531 | except Exception as e: |
|
528 | 532 | log.warning("URL is not a valid Mercurial repository: %s", |
|
529 | 533 | obfuscated_uri) |
|
530 | 534 | raise exceptions.URLError(e)( |
|
531 | 535 | f"url [{obfuscated_uri}] does not look like an hg repo org_exc: {e}") |
|
532 | 536 | |
|
533 | 537 | log.info("URL is a valid Mercurial repository: %s", obfuscated_uri) |
|
534 | 538 | return True |
|
535 | 539 | |
|
536 | 540 | @reraise_safe_exceptions |
|
537 | 541 | def diff(self, wire, commit_id_1, commit_id_2, file_filter, opt_git, opt_ignorews, context): |
|
538 | 542 | repo = self._factory.repo(wire) |
|
539 | 543 | |
|
540 | 544 | if file_filter: |
|
541 | 545 | # unpack the file-filter |
|
542 | 546 | repo_path, node_path = file_filter |
|
543 | 547 | match_filter = match(safe_bytes(repo_path), b'', [safe_bytes(node_path)]) |
|
544 | 548 | else: |
|
545 | 549 | match_filter = file_filter |
|
546 | 550 | opts = diffopts(git=opt_git, ignorews=opt_ignorews, context=context, showfunc=1) |
|
547 | 551 | |
|
548 | 552 | try: |
|
549 | 553 | diff_iter = patch.diff( |
|
550 | 554 | repo, node1=commit_id_1, node2=commit_id_2, match=match_filter, opts=opts) |
|
551 | 555 | return BytesEnvelope(b"".join(diff_iter)) |
|
552 | 556 | except RepoLookupError as e: |
|
553 | 557 | raise exceptions.LookupException(e)() |
|
554 | 558 | |
|
555 | 559 | @reraise_safe_exceptions |
|
556 | 560 | def node_history(self, wire, revision, path, limit): |
|
557 | 561 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
558 | 562 | region = self._region(wire) |
|
559 | 563 | |
|
560 | 564 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
561 | 565 | def _node_history(_context_uid, _repo_id, _revision, _path, _limit): |
|
562 | 566 | repo = self._factory.repo(wire) |
|
563 | 567 | |
|
564 | 568 | ctx = self._get_ctx(repo, revision) |
|
565 | 569 | fctx = ctx.filectx(safe_bytes(path)) |
|
566 | 570 | |
|
567 | 571 | def history_iter(): |
|
568 | 572 | limit_rev = fctx.rev() |
|
569 | 573 | |
|
570 | 574 | for fctx_candidate in reversed(list(fctx.filelog())): |
|
571 | 575 | f_obj = fctx.filectx(fctx_candidate) |
|
572 | 576 | |
|
573 | 577 | # NOTE: This can be problematic...we can hide ONLY history node resulting in empty history |
|
574 | 578 | _ctx = f_obj.changectx() |
|
575 | 579 | if _ctx.hidden() or _ctx.obsolete(): |
|
576 | 580 | continue |
|
577 | 581 | |
|
578 | 582 | if limit_rev >= f_obj.rev(): |
|
579 | 583 | yield f_obj |
|
580 | 584 | |
|
581 | 585 | history = [] |
|
582 | 586 | for cnt, obj in enumerate(history_iter()): |
|
583 | 587 | if limit and cnt >= limit: |
|
584 | 588 | break |
|
585 | 589 | history.append(hex(obj.node())) |
|
586 | 590 | |
|
587 | 591 | return [x for x in history] |
|
588 | 592 | return _node_history(context_uid, repo_id, revision, path, limit) |
|
589 | 593 | |
|
590 | 594 | @reraise_safe_exceptions |
|
591 | 595 | def node_history_until(self, wire, revision, path, limit): |
|
592 | 596 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
593 | 597 | region = self._region(wire) |
|
594 | 598 | |
|
595 | 599 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
596 | 600 | def _node_history_until(_context_uid, _repo_id): |
|
597 | 601 | repo = self._factory.repo(wire) |
|
598 | 602 | ctx = self._get_ctx(repo, revision) |
|
599 | 603 | fctx = ctx.filectx(safe_bytes(path)) |
|
600 | 604 | |
|
601 | 605 | file_log = list(fctx.filelog()) |
|
602 | 606 | if limit: |
|
603 | 607 | # Limit to the last n items |
|
604 | 608 | file_log = file_log[-limit:] |
|
605 | 609 | |
|
606 | 610 | return [hex(fctx.filectx(cs).node()) for cs in reversed(file_log)] |
|
607 | 611 | return _node_history_until(context_uid, repo_id, revision, path, limit) |
|
608 | 612 | |
|
609 | 613 | @reraise_safe_exceptions |
|
610 | 614 | def bulk_file_request(self, wire, commit_id, path, pre_load): |
|
611 | 615 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
612 | 616 | region = self._region(wire) |
|
613 | 617 | |
|
614 | 618 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
615 | 619 | def _bulk_file_request(_repo_id, _commit_id, _path, _pre_load): |
|
616 | 620 | result = {} |
|
617 | 621 | for attr in pre_load: |
|
618 | 622 | try: |
|
619 | 623 | method = self._bulk_file_methods[attr] |
|
620 | 624 | wire.update({'cache': False}) # disable cache for bulk calls so we don't double cache |
|
621 | 625 | result[attr] = method(wire, _commit_id, _path) |
|
622 | 626 | except KeyError as e: |
|
623 | 627 | raise exceptions.VcsException(e)(f'Unknown bulk attribute: "{attr}"') |
|
624 | 628 | return result |
|
625 | 629 | |
|
626 | 630 | return BinaryEnvelope(_bulk_file_request(repo_id, commit_id, path, sorted(pre_load))) |
|
627 | 631 | |
|
628 | 632 | @reraise_safe_exceptions |
|
629 | 633 | def fctx_annotate(self, wire, revision, path): |
|
630 | 634 | repo = self._factory.repo(wire) |
|
631 | 635 | ctx = self._get_ctx(repo, revision) |
|
632 | 636 | fctx = ctx.filectx(safe_bytes(path)) |
|
633 | 637 | |
|
634 | 638 | result = [] |
|
635 | 639 | for i, annotate_obj in enumerate(fctx.annotate(), 1): |
|
636 | 640 | ln_no = i |
|
637 | 641 | sha = hex(annotate_obj.fctx.node()) |
|
638 | 642 | content = annotate_obj.text |
|
639 | 643 | result.append((ln_no, ascii_str(sha), content)) |
|
640 | 644 | return BinaryEnvelope(result) |
|
641 | 645 | |
|
642 | 646 | @reraise_safe_exceptions |
|
643 | 647 | def fctx_node_data(self, wire, revision, path): |
|
644 | 648 | repo = self._factory.repo(wire) |
|
645 | 649 | ctx = self._get_ctx(repo, revision) |
|
646 | 650 | fctx = ctx.filectx(safe_bytes(path)) |
|
647 | 651 | return BytesEnvelope(fctx.data()) |
|
648 | 652 | |
|
649 | 653 | @reraise_safe_exceptions |
|
650 | 654 | def fctx_flags(self, wire, commit_id, path): |
|
651 | 655 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
652 | 656 | region = self._region(wire) |
|
653 | 657 | |
|
654 | 658 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
655 | 659 | def _fctx_flags(_repo_id, _commit_id, _path): |
|
656 | 660 | repo = self._factory.repo(wire) |
|
657 | 661 | ctx = self._get_ctx(repo, commit_id) |
|
658 | 662 | fctx = ctx.filectx(safe_bytes(path)) |
|
659 | 663 | return fctx.flags() |
|
660 | 664 | |
|
661 | 665 | return _fctx_flags(repo_id, commit_id, path) |
|
662 | 666 | |
|
663 | 667 | @reraise_safe_exceptions |
|
664 | 668 | def fctx_size(self, wire, commit_id, path): |
|
665 | 669 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
666 | 670 | region = self._region(wire) |
|
667 | 671 | |
|
668 | 672 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
669 | 673 | def _fctx_size(_repo_id, _revision, _path): |
|
670 | 674 | repo = self._factory.repo(wire) |
|
671 | 675 | ctx = self._get_ctx(repo, commit_id) |
|
672 | 676 | fctx = ctx.filectx(safe_bytes(path)) |
|
673 | 677 | return fctx.size() |
|
674 | 678 | return _fctx_size(repo_id, commit_id, path) |
|
675 | 679 | |
|
676 | 680 | @reraise_safe_exceptions |
|
677 | 681 | def get_all_commit_ids(self, wire, name): |
|
678 | 682 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
679 | 683 | region = self._region(wire) |
|
680 | 684 | |
|
681 | 685 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
682 | 686 | def _get_all_commit_ids(_context_uid, _repo_id, _name): |
|
683 | 687 | repo = self._factory.repo(wire) |
|
684 | 688 | revs = [ascii_str(repo[x].hex()) for x in repo.filtered(b'visible').changelog.revs()] |
|
685 | 689 | return revs |
|
686 | 690 | return _get_all_commit_ids(context_uid, repo_id, name) |
|
687 | 691 | |
|
688 | 692 | @reraise_safe_exceptions |
|
689 | 693 | def get_config_value(self, wire, section, name, untrusted=False): |
|
690 | 694 | repo = self._factory.repo(wire) |
|
691 | 695 | return repo.ui.config(ascii_bytes(section), ascii_bytes(name), untrusted=untrusted) |
|
692 | 696 | |
|
693 | 697 | @reraise_safe_exceptions |
|
694 | 698 | def is_large_file(self, wire, commit_id, path): |
|
695 | 699 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
696 | 700 | region = self._region(wire) |
|
697 | 701 | |
|
698 | 702 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
699 | 703 | def _is_large_file(_context_uid, _repo_id, _commit_id, _path): |
|
700 | 704 | return largefiles.lfutil.isstandin(safe_bytes(path)) |
|
701 | 705 | |
|
702 | 706 | return _is_large_file(context_uid, repo_id, commit_id, path) |
|
703 | 707 | |
|
704 | 708 | @reraise_safe_exceptions |
|
705 | 709 | def is_binary(self, wire, revision, path): |
|
706 | 710 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
707 | 711 | region = self._region(wire) |
|
708 | 712 | |
|
709 | 713 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
710 | 714 | def _is_binary(_repo_id, _sha, _path): |
|
711 | 715 | repo = self._factory.repo(wire) |
|
712 | 716 | ctx = self._get_ctx(repo, revision) |
|
713 | 717 | fctx = ctx.filectx(safe_bytes(path)) |
|
714 | 718 | return fctx.isbinary() |
|
715 | 719 | |
|
716 | 720 | return _is_binary(repo_id, revision, path) |
|
717 | 721 | |
|
718 | 722 | @reraise_safe_exceptions |
|
719 | 723 | def md5_hash(self, wire, revision, path): |
|
720 | 724 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
721 | 725 | region = self._region(wire) |
|
722 | 726 | |
|
723 | 727 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
724 | 728 | def _md5_hash(_repo_id, _sha, _path): |
|
725 | 729 | repo = self._factory.repo(wire) |
|
726 | 730 | ctx = self._get_ctx(repo, revision) |
|
727 | 731 | fctx = ctx.filectx(safe_bytes(path)) |
|
728 | 732 | return hashlib.md5(fctx.data()).hexdigest() |
|
729 | 733 | |
|
730 | 734 | return _md5_hash(repo_id, revision, path) |
|
731 | 735 | |
|
732 | 736 | @reraise_safe_exceptions |
|
733 | 737 | def in_largefiles_store(self, wire, sha): |
|
734 | 738 | repo = self._factory.repo(wire) |
|
735 | 739 | return largefiles.lfutil.instore(repo, sha) |
|
736 | 740 | |
|
737 | 741 | @reraise_safe_exceptions |
|
738 | 742 | def in_user_cache(self, wire, sha): |
|
739 | 743 | repo = self._factory.repo(wire) |
|
740 | 744 | return largefiles.lfutil.inusercache(repo.ui, sha) |
|
741 | 745 | |
|
742 | 746 | @reraise_safe_exceptions |
|
743 | 747 | def store_path(self, wire, sha): |
|
744 | 748 | repo = self._factory.repo(wire) |
|
745 | 749 | return largefiles.lfutil.storepath(repo, sha) |
|
746 | 750 | |
|
747 | 751 | @reraise_safe_exceptions |
|
748 | 752 | def link(self, wire, sha, path): |
|
749 | 753 | repo = self._factory.repo(wire) |
|
750 | 754 | largefiles.lfutil.link( |
|
751 | 755 | largefiles.lfutil.usercachepath(repo.ui, sha), path) |
|
752 | 756 | |
|
753 | 757 | @reraise_safe_exceptions |
|
754 | 758 | def localrepository(self, wire, create=False): |
|
755 | 759 | self._factory.repo(wire, create=create) |
|
756 | 760 | |
|
757 | 761 | @reraise_safe_exceptions |
|
758 | 762 | def lookup(self, wire, revision, both): |
|
759 | 763 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
760 | 764 | region = self._region(wire) |
|
761 | 765 | |
|
762 | 766 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
763 | 767 | def _lookup(_context_uid, _repo_id, _revision, _both): |
|
764 | 768 | repo = self._factory.repo(wire) |
|
765 | 769 | rev = _revision |
|
766 | 770 | if isinstance(rev, int): |
|
767 | 771 | # NOTE(marcink): |
|
768 | 772 | # since Mercurial doesn't support negative indexes properly |
|
769 | 773 | # we need to shift accordingly by one to get proper index, e.g |
|
770 | 774 | # repo[-1] => repo[-2] |
|
771 | 775 | # repo[0] => repo[-1] |
|
772 | 776 | if rev <= 0: |
|
773 | 777 | rev = rev + -1 |
|
774 | 778 | try: |
|
775 | 779 | ctx = self._get_ctx(repo, rev) |
|
776 | 780 | except AmbiguousPrefixLookupError: |
|
777 | 781 | e = RepoLookupError(rev) |
|
778 | 782 | e._org_exc_tb = format_exc(sys.exc_info()) |
|
779 | 783 | raise exceptions.LookupException(e)(rev) |
|
780 | 784 | except (TypeError, RepoLookupError, binascii.Error) as e: |
|
781 | 785 | e._org_exc_tb = format_exc(sys.exc_info()) |
|
782 | 786 | raise exceptions.LookupException(e)(rev) |
|
783 | 787 | except LookupError as e: |
|
784 | 788 | e._org_exc_tb = format_exc(sys.exc_info()) |
|
785 | 789 | raise exceptions.LookupException(e)(e.name) |
|
786 | 790 | |
|
787 | 791 | if not both: |
|
788 | 792 | return ctx.hex() |
|
789 | 793 | |
|
790 | 794 | ctx = repo[ctx.hex()] |
|
791 | 795 | return ctx.hex(), ctx.rev() |
|
792 | 796 | |
|
793 | 797 | return _lookup(context_uid, repo_id, revision, both) |
|
794 | 798 | |
|
795 | 799 | @reraise_safe_exceptions |
|
796 | 800 | def sync_push(self, wire, url): |
|
797 | 801 | if not self.check_url(url, wire['config']): |
|
798 | 802 | return |
|
799 | 803 | |
|
800 | 804 | repo = self._factory.repo(wire) |
|
801 | 805 | |
|
802 | 806 | # Disable any prompts for this repo |
|
803 |
repo.ui.setconfig(b'ui', b'interactive', b' |
|
|
807 | repo.ui.setconfig(b'ui', b'interactive', b'false', b'-y') | |
|
804 | 808 | |
|
805 | 809 | bookmarks = list(dict(repo._bookmarks).keys()) |
|
806 | 810 | remote = peer(repo, {}, safe_bytes(url)) |
|
807 | 811 | # Disable any prompts for this remote |
|
808 |
remote.ui.setconfig(b'ui', b'interactive', b' |
|
|
812 | remote.ui.setconfig(b'ui', b'interactive', b'false', b'-y') | |
|
809 | 813 | |
|
810 | 814 | return exchange.push( |
|
811 | 815 | repo, remote, newbranch=True, bookmarks=bookmarks).cgresult |
|
812 | 816 | |
|
813 | 817 | @reraise_safe_exceptions |
|
814 | 818 | def revision(self, wire, rev): |
|
815 | 819 | repo = self._factory.repo(wire) |
|
816 | 820 | ctx = self._get_ctx(repo, rev) |
|
817 | 821 | return ctx.rev() |
|
818 | 822 | |
|
819 | 823 | @reraise_safe_exceptions |
|
820 | 824 | def rev_range(self, wire, commit_filter): |
|
821 | 825 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
822 | 826 | region = self._region(wire) |
|
823 | 827 | |
|
824 | 828 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
825 | 829 | def _rev_range(_context_uid, _repo_id, _filter): |
|
826 | 830 | repo = self._factory.repo(wire) |
|
827 | 831 | revisions = [ |
|
828 | 832 | ascii_str(repo[rev].hex()) |
|
829 | 833 | for rev in revrange(repo, list(map(ascii_bytes, commit_filter))) |
|
830 | 834 | ] |
|
831 | 835 | return revisions |
|
832 | 836 | |
|
833 | 837 | return _rev_range(context_uid, repo_id, sorted(commit_filter)) |
|
834 | 838 | |
|
835 | 839 | @reraise_safe_exceptions |
|
836 | 840 | def rev_range_hash(self, wire, node): |
|
837 | 841 | repo = self._factory.repo(wire) |
|
838 | 842 | |
|
839 | 843 | def get_revs(repo, rev_opt): |
|
840 | 844 | if rev_opt: |
|
841 | 845 | revs = revrange(repo, rev_opt) |
|
842 | 846 | if len(revs) == 0: |
|
843 | 847 | return (nullrev, nullrev) |
|
844 | 848 | return max(revs), min(revs) |
|
845 | 849 | else: |
|
846 | 850 | return len(repo) - 1, 0 |
|
847 | 851 | |
|
848 | 852 | stop, start = get_revs(repo, [node + ':']) |
|
849 | 853 | revs = [ascii_str(repo[r].hex()) for r in range(start, stop + 1)] |
|
850 | 854 | return revs |
|
851 | 855 | |
|
852 | 856 | @reraise_safe_exceptions |
|
853 | 857 | def revs_from_revspec(self, wire, rev_spec, *args, **kwargs): |
|
854 | 858 | org_path = safe_bytes(wire["path"]) |
|
855 | 859 | other_path = safe_bytes(kwargs.pop('other_path', '')) |
|
856 | 860 | |
|
857 | 861 | # case when we want to compare two independent repositories |
|
858 | 862 | if other_path and other_path != wire["path"]: |
|
859 | 863 | baseui = self._factory._create_config(wire["config"]) |
|
860 | 864 | repo = unionrepo.makeunionrepository(baseui, other_path, org_path) |
|
861 | 865 | else: |
|
862 | 866 | repo = self._factory.repo(wire) |
|
863 | 867 | return list(repo.revs(rev_spec, *args)) |
|
864 | 868 | |
|
865 | 869 | @reraise_safe_exceptions |
|
866 | 870 | def verify(self, wire,): |
|
867 | 871 | repo = self._factory.repo(wire) |
|
868 | 872 | baseui = self._factory._create_config(wire['config']) |
|
869 | 873 | |
|
870 | 874 | baseui, output = patch_ui_message_output(baseui) |
|
871 | 875 | |
|
872 | 876 | repo.ui = baseui |
|
873 | 877 | verify.verify(repo) |
|
874 | 878 | return output.getvalue() |
|
875 | 879 | |
|
876 | 880 | @reraise_safe_exceptions |
|
877 | 881 | def hg_update_cache(self, wire,): |
|
878 | 882 | repo = self._factory.repo(wire) |
|
879 | 883 | baseui = self._factory._create_config(wire['config']) |
|
880 | 884 | baseui, output = patch_ui_message_output(baseui) |
|
881 | 885 | |
|
882 | 886 | repo.ui = baseui |
|
883 | 887 | with repo.wlock(), repo.lock(): |
|
884 | 888 | repo.updatecaches(full=True) |
|
885 | 889 | |
|
886 | 890 | return output.getvalue() |
|
887 | 891 | |
|
888 | 892 | @reraise_safe_exceptions |
|
889 | 893 | def hg_rebuild_fn_cache(self, wire,): |
|
890 | 894 | repo = self._factory.repo(wire) |
|
891 | 895 | baseui = self._factory._create_config(wire['config']) |
|
892 | 896 | baseui, output = patch_ui_message_output(baseui) |
|
893 | 897 | |
|
894 | 898 | repo.ui = baseui |
|
895 | 899 | |
|
896 | 900 | repair.rebuildfncache(baseui, repo) |
|
897 | 901 | |
|
898 | 902 | return output.getvalue() |
|
899 | 903 | |
|
900 | 904 | @reraise_safe_exceptions |
|
901 | 905 | def tags(self, wire): |
|
902 | 906 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
903 | 907 | region = self._region(wire) |
|
904 | 908 | |
|
905 | 909 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
906 | 910 | def _tags(_context_uid, _repo_id): |
|
907 | 911 | repo = self._factory.repo(wire) |
|
908 | 912 | return {safe_str(name): ascii_str(hex(sha)) for name, sha in repo.tags().items()} |
|
909 | 913 | |
|
910 | 914 | return _tags(context_uid, repo_id) |
|
911 | 915 | |
|
912 | 916 | @reraise_safe_exceptions |
|
913 | 917 | def update(self, wire, node='', clean=False): |
|
914 | 918 | repo = self._factory.repo(wire) |
|
915 | 919 | baseui = self._factory._create_config(wire['config']) |
|
916 | 920 | node = safe_bytes(node) |
|
917 | 921 | |
|
918 | 922 | commands.update(baseui, repo, node=node, clean=clean) |
|
919 | 923 | |
|
920 | 924 | @reraise_safe_exceptions |
|
921 | 925 | def identify(self, wire): |
|
922 | 926 | repo = self._factory.repo(wire) |
|
923 | 927 | baseui = self._factory._create_config(wire['config']) |
|
924 | 928 | output = io.BytesIO() |
|
925 | 929 | baseui.write = output.write |
|
926 | 930 | # This is required to get a full node id |
|
927 | 931 | baseui.debugflag = True |
|
928 | 932 | commands.identify(baseui, repo, id=True) |
|
929 | 933 | |
|
930 | 934 | return output.getvalue() |
|
931 | 935 | |
|
932 | 936 | @reraise_safe_exceptions |
|
933 | 937 | def heads(self, wire, branch=None): |
|
934 | 938 | repo = self._factory.repo(wire) |
|
935 | 939 | baseui = self._factory._create_config(wire['config']) |
|
936 | 940 | output = io.BytesIO() |
|
937 | 941 | |
|
938 | 942 | def write(data, **unused_kwargs): |
|
939 | 943 | output.write(data) |
|
940 | 944 | |
|
941 | 945 | baseui.write = write |
|
942 | 946 | if branch: |
|
943 | 947 | args = [safe_bytes(branch)] |
|
944 | 948 | else: |
|
945 | 949 | args = [] |
|
946 | 950 | commands.heads(baseui, repo, template=b'{node} ', *args) |
|
947 | 951 | |
|
948 | 952 | return output.getvalue() |
|
949 | 953 | |
|
950 | 954 | @reraise_safe_exceptions |
|
951 | 955 | def ancestor(self, wire, revision1, revision2): |
|
952 | 956 | repo = self._factory.repo(wire) |
|
953 | 957 | changelog = repo.changelog |
|
954 | 958 | lookup = repo.lookup |
|
955 | 959 | a = changelog.ancestor(lookup(safe_bytes(revision1)), lookup(safe_bytes(revision2))) |
|
956 | 960 | return hex(a) |
|
957 | 961 | |
|
958 | 962 | @reraise_safe_exceptions |
|
959 | 963 | def clone(self, wire, source, dest, update_after_clone=False, hooks=True): |
|
960 | 964 | baseui = self._factory._create_config(wire["config"], hooks=hooks) |
|
961 | 965 | clone(baseui, safe_bytes(source), safe_bytes(dest), noupdate=not update_after_clone) |
|
962 | 966 | |
|
963 | 967 | @reraise_safe_exceptions |
|
964 | 968 | def commitctx(self, wire, message, parents, commit_time, commit_timezone, user, files, extra, removed, updated): |
|
965 | 969 | |
|
966 | 970 | repo = self._factory.repo(wire) |
|
967 | 971 | baseui = self._factory._create_config(wire['config']) |
|
968 | 972 | publishing = baseui.configbool(b'phases', b'publish') |
|
969 | 973 | |
|
970 | 974 | def _filectxfn(_repo, ctx, path: bytes): |
|
971 | 975 | """ |
|
972 | 976 | Marks given path as added/changed/removed in a given _repo. This is |
|
973 | 977 | for internal mercurial commit function. |
|
974 | 978 | """ |
|
975 | 979 | |
|
976 | 980 | # check if this path is removed |
|
977 | 981 | if safe_str(path) in removed: |
|
978 | 982 | # returning None is a way to mark node for removal |
|
979 | 983 | return None |
|
980 | 984 | |
|
981 | 985 | # check if this path is added |
|
982 | 986 | for node in updated: |
|
983 | 987 | if safe_bytes(node['path']) == path: |
|
984 | 988 | return memfilectx( |
|
985 | 989 | _repo, |
|
986 | 990 | changectx=ctx, |
|
987 | 991 | path=safe_bytes(node['path']), |
|
988 | 992 | data=safe_bytes(node['content']), |
|
989 | 993 | islink=False, |
|
990 | 994 | isexec=bool(node['mode'] & stat.S_IXUSR), |
|
991 | 995 | copysource=False) |
|
992 | 996 | abort_exc = exceptions.AbortException() |
|
993 | 997 | raise abort_exc(f"Given path haven't been marked as added, changed or removed ({path})") |
|
994 | 998 | |
|
995 | 999 | if publishing: |
|
996 | 1000 | new_commit_phase = b'public' |
|
997 | 1001 | else: |
|
998 | 1002 | new_commit_phase = b'draft' |
|
999 | 1003 | with repo.ui.configoverride({(b'phases', b'new-commit'): new_commit_phase}): |
|
1000 | 1004 | kwargs = {safe_bytes(k): safe_bytes(v) for k, v in extra.items()} |
|
1001 | 1005 | commit_ctx = memctx( |
|
1002 | 1006 | repo=repo, |
|
1003 | 1007 | parents=parents, |
|
1004 | 1008 | text=safe_bytes(message), |
|
1005 | 1009 | files=[safe_bytes(x) for x in files], |
|
1006 | 1010 | filectxfn=_filectxfn, |
|
1007 | 1011 | user=safe_bytes(user), |
|
1008 | 1012 | date=(commit_time, commit_timezone), |
|
1009 | 1013 | extra=kwargs) |
|
1010 | 1014 | |
|
1011 | 1015 | n = repo.commitctx(commit_ctx) |
|
1012 | 1016 | new_id = hex(n) |
|
1013 | 1017 | |
|
1014 | 1018 | return new_id |
|
1015 | 1019 | |
|
1016 | 1020 | @reraise_safe_exceptions |
|
1017 | 1021 | def pull(self, wire, url, commit_ids=None): |
|
1018 | 1022 | repo = self._factory.repo(wire) |
|
1019 | 1023 | # Disable any prompts for this repo |
|
1020 |
repo.ui.setconfig(b'ui', b'interactive', b' |
|
|
1024 | repo.ui.setconfig(b'ui', b'interactive', b'false', b'-y') | |
|
1021 | 1025 | |
|
1022 | 1026 | remote = peer(repo, {}, safe_bytes(url)) |
|
1023 | 1027 | # Disable any prompts for this remote |
|
1024 |
remote.ui.setconfig(b'ui', b'interactive', b' |
|
|
1028 | remote.ui.setconfig(b'ui', b'interactive', b'false', b'-y') | |
|
1025 | 1029 | |
|
1026 | 1030 | if commit_ids: |
|
1027 | 1031 | commit_ids = [bin(commit_id) for commit_id in commit_ids] |
|
1028 | 1032 | |
|
1029 | 1033 | return exchange.pull( |
|
1030 | 1034 | repo, remote, heads=commit_ids, force=None).cgresult |
|
1031 | 1035 | |
|
1032 | 1036 | @reraise_safe_exceptions |
|
1033 | 1037 | def pull_cmd(self, wire, source, bookmark='', branch='', revision='', hooks=True): |
|
1034 | 1038 | repo = self._factory.repo(wire) |
|
1035 | 1039 | baseui = self._factory._create_config(wire['config'], hooks=hooks) |
|
1036 | 1040 | |
|
1037 | 1041 | source = safe_bytes(source) |
|
1038 | 1042 | |
|
1039 | 1043 | # Mercurial internally has a lot of logic that checks ONLY if |
|
1040 | 1044 | # option is defined, we just pass those if they are defined then |
|
1041 | opts = {} | |
|
1045 | opts = {"remote_hidden": False} | |
|
1042 | 1046 | |
|
1043 | 1047 | if bookmark: |
|
1044 | 1048 | opts['bookmark'] = [safe_bytes(x) for x in bookmark] \ |
|
1045 | 1049 | if isinstance(bookmark, list) else safe_bytes(bookmark) |
|
1046 | 1050 | |
|
1047 | 1051 | if branch: |
|
1048 | 1052 | opts['branch'] = [safe_bytes(x) for x in branch] \ |
|
1049 | 1053 | if isinstance(branch, list) else safe_bytes(branch) |
|
1050 | 1054 | |
|
1051 | 1055 | if revision: |
|
1052 | 1056 | opts['rev'] = [safe_bytes(x) for x in revision] \ |
|
1053 | 1057 | if isinstance(revision, list) else safe_bytes(revision) |
|
1054 | 1058 | |
|
1055 | 1059 | commands.pull(baseui, repo, source, **opts) |
|
1056 | 1060 | |
|
1057 | 1061 | @reraise_safe_exceptions |
|
1058 | 1062 | def push(self, wire, revisions, dest_path, hooks: bool = True, push_branches: bool = False): |
|
1059 | 1063 | repo = self._factory.repo(wire) |
|
1060 | 1064 | baseui = self._factory._create_config(wire['config'], hooks=hooks) |
|
1061 | 1065 | |
|
1062 | 1066 | revisions = [safe_bytes(x) for x in revisions] \ |
|
1063 | 1067 | if isinstance(revisions, list) else safe_bytes(revisions) |
|
1064 | 1068 | |
|
1065 | 1069 | commands.push(baseui, repo, safe_bytes(dest_path), |
|
1066 | 1070 | rev=revisions, |
|
1067 | 1071 | new_branch=push_branches) |
|
1068 | 1072 | |
|
1069 | 1073 | @reraise_safe_exceptions |
|
1070 | 1074 | def strip(self, wire, revision, update, backup): |
|
1071 | 1075 | repo = self._factory.repo(wire) |
|
1072 | 1076 | ctx = self._get_ctx(repo, revision) |
|
1073 | 1077 | hgext_strip.strip( |
|
1074 | 1078 | repo.baseui, repo, ctx.node(), update=update, backup=backup) |
|
1075 | 1079 | |
|
1076 | 1080 | @reraise_safe_exceptions |
|
1077 | 1081 | def get_unresolved_files(self, wire): |
|
1078 | 1082 | repo = self._factory.repo(wire) |
|
1079 | 1083 | |
|
1080 | 1084 | log.debug('Calculating unresolved files for repo: %s', repo) |
|
1081 | 1085 | output = io.BytesIO() |
|
1082 | 1086 | |
|
1083 | 1087 | def write(data, **unused_kwargs): |
|
1084 | 1088 | output.write(data) |
|
1085 | 1089 | |
|
1086 | 1090 | baseui = self._factory._create_config(wire['config']) |
|
1087 | 1091 | baseui.write = write |
|
1088 | 1092 | |
|
1089 | 1093 | commands.resolve(baseui, repo, list=True) |
|
1090 | 1094 | unresolved = output.getvalue().splitlines(0) |
|
1091 | 1095 | return unresolved |
|
1092 | 1096 | |
|
1093 | 1097 | @reraise_safe_exceptions |
|
1094 | 1098 | def merge(self, wire, revision): |
|
1095 | 1099 | repo = self._factory.repo(wire) |
|
1096 | 1100 | baseui = self._factory._create_config(wire['config']) |
|
1097 | 1101 | repo.ui.setconfig(b'ui', b'merge', b'internal:dump') |
|
1098 | 1102 | |
|
1099 | 1103 | # In case of sub repositories are used mercurial prompts the user in |
|
1100 | 1104 | # case of merge conflicts or different sub repository sources. By |
|
1101 | 1105 | # setting the interactive flag to `False` mercurial doesn't prompt the |
|
1102 | 1106 | # used but instead uses a default value. |
|
1103 |
repo.ui.setconfig(b'ui', b'interactive', |
|
|
1107 | repo.ui.setconfig(b'ui', b'interactive', b'false') | |
|
1104 | 1108 | commands.merge(baseui, repo, rev=safe_bytes(revision)) |
|
1105 | 1109 | |
|
1106 | 1110 | @reraise_safe_exceptions |
|
1107 | 1111 | def merge_state(self, wire): |
|
1108 | 1112 | repo = self._factory.repo(wire) |
|
1109 | 1113 | repo.ui.setconfig(b'ui', b'merge', b'internal:dump') |
|
1110 | 1114 | |
|
1111 | 1115 | # In case of sub repositories are used mercurial prompts the user in |
|
1112 | 1116 | # case of merge conflicts or different sub repository sources. By |
|
1113 | 1117 | # setting the interactive flag to `False` mercurial doesn't prompt the |
|
1114 | 1118 | # used but instead uses a default value. |
|
1115 |
repo.ui.setconfig(b'ui', b'interactive', |
|
|
1119 | repo.ui.setconfig(b'ui', b'interactive', b'false') | |
|
1116 | 1120 | ms = hg_merge.mergestate(repo) |
|
1117 | 1121 | return [x for x in ms.unresolved()] |
|
1118 | 1122 | |
|
1119 | 1123 | @reraise_safe_exceptions |
|
1120 | 1124 | def commit(self, wire, message, username, close_branch=False): |
|
1121 | 1125 | repo = self._factory.repo(wire) |
|
1122 | 1126 | baseui = self._factory._create_config(wire['config']) |
|
1123 | 1127 | repo.ui.setconfig(b'ui', b'username', safe_bytes(username)) |
|
1124 | 1128 | commands.commit(baseui, repo, message=safe_bytes(message), close_branch=close_branch) |
|
1125 | 1129 | |
|
1126 | 1130 | @reraise_safe_exceptions |
|
1127 | 1131 | def rebase(self, wire, source='', dest='', abort=False): |
|
1128 | 1132 | |
|
1129 | 1133 | repo = self._factory.repo(wire) |
|
1130 | 1134 | baseui = self._factory._create_config(wire['config']) |
|
1131 | 1135 | repo.ui.setconfig(b'ui', b'merge', b'internal:dump') |
|
1132 | 1136 | # In case of sub repositories are used mercurial prompts the user in |
|
1133 | 1137 | # case of merge conflicts or different sub repository sources. By |
|
1134 | 1138 | # setting the interactive flag to `False` mercurial doesn't prompt the |
|
1135 | 1139 | # used but instead uses a default value. |
|
1136 |
repo.ui.setconfig(b'ui', b'interactive', |
|
|
1140 | repo.ui.setconfig(b'ui', b'interactive', b'false') | |
|
1137 | 1141 | |
|
1138 | 1142 | rebase_kws = dict( |
|
1139 | 1143 | keep=not abort, |
|
1140 | 1144 | abort=abort |
|
1141 | 1145 | ) |
|
1142 | 1146 | |
|
1143 | 1147 | if source: |
|
1144 | 1148 | source = repo[source] |
|
1145 | 1149 | rebase_kws['base'] = [source.hex()] |
|
1146 | 1150 | if dest: |
|
1147 | 1151 | dest = repo[dest] |
|
1148 | 1152 | rebase_kws['dest'] = dest.hex() |
|
1149 | 1153 | |
|
1150 | 1154 | rebase.rebase(baseui, repo, **rebase_kws) |
|
1151 | 1155 | |
|
1152 | 1156 | @reraise_safe_exceptions |
|
1153 | 1157 | def tag(self, wire, name, revision, message, local, user, tag_time, tag_timezone): |
|
1154 | 1158 | repo = self._factory.repo(wire) |
|
1155 | 1159 | ctx = self._get_ctx(repo, revision) |
|
1156 | 1160 | node = ctx.node() |
|
1157 | 1161 | |
|
1158 | 1162 | date = (tag_time, tag_timezone) |
|
1159 | 1163 | try: |
|
1160 | 1164 | hg_tag.tag(repo, safe_bytes(name), node, safe_bytes(message), local, safe_bytes(user), date) |
|
1161 | 1165 | except Abort as e: |
|
1162 | 1166 | log.exception("Tag operation aborted") |
|
1163 | 1167 | # Exception can contain unicode which we convert |
|
1164 | 1168 | raise exceptions.AbortException(e)(repr(e)) |
|
1165 | 1169 | |
|
1166 | 1170 | @reraise_safe_exceptions |
|
1167 | 1171 | def bookmark(self, wire, bookmark, revision=''): |
|
1168 | 1172 | repo = self._factory.repo(wire) |
|
1169 | 1173 | baseui = self._factory._create_config(wire['config']) |
|
1170 | 1174 | revision = revision or '' |
|
1171 | 1175 | commands.bookmark(baseui, repo, safe_bytes(bookmark), rev=safe_bytes(revision), force=True) |
|
1172 | 1176 | |
|
1173 | 1177 | @reraise_safe_exceptions |
|
1174 | 1178 | def install_hooks(self, wire, force=False): |
|
1175 | 1179 | # we don't need any special hooks for Mercurial |
|
1176 | 1180 | pass |
|
1177 | 1181 | |
|
1178 | 1182 | @reraise_safe_exceptions |
|
1179 | 1183 | def get_hooks_info(self, wire): |
|
1180 | 1184 | return { |
|
1181 | 1185 | 'pre_version': vcsserver.get_version(), |
|
1182 | 1186 | 'post_version': vcsserver.get_version(), |
|
1183 | 1187 | } |
|
1184 | 1188 | |
|
1185 | 1189 | @reraise_safe_exceptions |
|
1186 | 1190 | def set_head_ref(self, wire, head_name): |
|
1187 | 1191 | pass |
|
1188 | 1192 | |
|
1189 | 1193 | @reraise_safe_exceptions |
|
1190 | 1194 | def archive_repo(self, wire, archive_name_key, kind, mtime, archive_at_path, |
|
1191 | 1195 | archive_dir_name, commit_id, cache_config): |
|
1192 | 1196 | |
|
1193 | 1197 | def file_walker(_commit_id, path): |
|
1194 | 1198 | repo = self._factory.repo(wire) |
|
1195 | 1199 | ctx = repo[_commit_id] |
|
1196 | 1200 | is_root = path in ['', '/'] |
|
1197 | 1201 | if is_root: |
|
1198 | 1202 | matcher = alwaysmatcher(badfn=None) |
|
1199 | 1203 | else: |
|
1200 | 1204 | matcher = patternmatcher('', [(b'glob', safe_bytes(path)+b'/**', b'')], badfn=None) |
|
1201 | 1205 | file_iter = ctx.manifest().walk(matcher) |
|
1202 | 1206 | |
|
1203 | 1207 | for fn in file_iter: |
|
1204 | 1208 | file_path = fn |
|
1205 | 1209 | flags = ctx.flags(fn) |
|
1206 | 1210 | mode = b'x' in flags and 0o755 or 0o644 |
|
1207 | 1211 | is_link = b'l' in flags |
|
1208 | 1212 | |
|
1209 | 1213 | yield ArchiveNode(file_path, mode, is_link, ctx[fn].data) |
|
1210 | 1214 | |
|
1211 | 1215 | return store_archive_in_cache( |
|
1212 | 1216 | file_walker, archive_name_key, kind, mtime, archive_at_path, archive_dir_name, commit_id, cache_config=cache_config) |
|
1213 | 1217 |
@@ -1,954 +1,959 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | |
|
19 | 19 | import os |
|
20 | 20 | import subprocess |
|
21 | 21 | from urllib.error import URLError |
|
22 | 22 | import urllib.parse |
|
23 | 23 | import logging |
|
24 | 24 | import posixpath as vcspath |
|
25 | 25 | import io |
|
26 | 26 | import urllib.request |
|
27 | 27 | import urllib.parse |
|
28 | 28 | import urllib.error |
|
29 | 29 | import traceback |
|
30 | 30 | |
|
31 | ||
|
32 | 31 | import svn.client # noqa |
|
33 | 32 | import svn.core # noqa |
|
34 | 33 | import svn.delta # noqa |
|
35 | 34 | import svn.diff # noqa |
|
36 | 35 | import svn.fs # noqa |
|
37 | 36 | import svn.repos # noqa |
|
38 | 37 | |
|
39 | 38 | import rhodecode |
|
40 | 39 | from vcsserver import svn_diff, exceptions, subprocessio, settings |
|
41 | 40 | from vcsserver.base import ( |
|
42 | 41 | RepoFactory, |
|
43 | 42 | raise_from_original, |
|
44 | 43 | ArchiveNode, |
|
45 | 44 | store_archive_in_cache, |
|
46 | 45 | BytesEnvelope, |
|
47 | 46 | BinaryEnvelope, |
|
48 | 47 | ) |
|
49 | 48 | from vcsserver.exceptions import NoContentException |
|
50 | from vcsserver.str_utils import safe_str, safe_bytes | |
|
51 | from vcsserver.type_utils import assert_bytes | |
|
52 | 49 | from vcsserver.vcs_base import RemoteBase |
|
50 | from vcsserver.lib.str_utils import safe_str, safe_bytes | |
|
51 | from vcsserver.lib.type_utils import assert_bytes | |
|
53 | 52 | from vcsserver.lib.svnremoterepo import svnremoterepo |
|
53 | from vcsserver.lib.svn_txn_utils import store_txn_id_data | |
|
54 | 54 | |
|
55 | 55 | log = logging.getLogger(__name__) |
|
56 | 56 | |
|
57 | 57 | |
|
58 | 58 | svn_compatible_versions_map = { |
|
59 | 59 | 'pre-1.4-compatible': '1.3', |
|
60 | 60 | 'pre-1.5-compatible': '1.4', |
|
61 | 61 | 'pre-1.6-compatible': '1.5', |
|
62 | 62 | 'pre-1.8-compatible': '1.7', |
|
63 | 63 | 'pre-1.9-compatible': '1.8', |
|
64 | 64 | } |
|
65 | 65 | |
|
66 | 66 | current_compatible_version = '1.14' |
|
67 | 67 | |
|
68 | 68 | |
|
69 | 69 | def reraise_safe_exceptions(func): |
|
70 | 70 | """Decorator for converting svn exceptions to something neutral.""" |
|
71 | 71 | def wrapper(*args, **kwargs): |
|
72 | 72 | try: |
|
73 | 73 | return func(*args, **kwargs) |
|
74 | 74 | except Exception as e: |
|
75 | 75 | if not hasattr(e, '_vcs_kind'): |
|
76 | 76 | log.exception("Unhandled exception in svn remote call") |
|
77 | 77 | raise_from_original(exceptions.UnhandledException(e), e) |
|
78 | 78 | raise |
|
79 | 79 | return wrapper |
|
80 | 80 | |
|
81 | 81 | |
|
82 | 82 | class SubversionFactory(RepoFactory): |
|
83 | 83 | repo_type = 'svn' |
|
84 | 84 | |
|
85 | 85 | def _create_repo(self, wire, create, compatible_version): |
|
86 | 86 | path = svn.core.svn_path_canonicalize(wire['path']) |
|
87 | 87 | if create: |
|
88 | 88 | fs_config = {'compatible-version': current_compatible_version} |
|
89 | 89 | if compatible_version: |
|
90 | 90 | |
|
91 | 91 | compatible_version_string = \ |
|
92 | 92 | svn_compatible_versions_map.get(compatible_version) \ |
|
93 | 93 | or compatible_version |
|
94 | 94 | fs_config['compatible-version'] = compatible_version_string |
|
95 | 95 | |
|
96 | 96 | log.debug('Create SVN repo with config `%s`', fs_config) |
|
97 | 97 | repo = svn.repos.create(path, "", "", None, fs_config) |
|
98 | 98 | else: |
|
99 | 99 | repo = svn.repos.open(path) |
|
100 | 100 | |
|
101 | 101 | log.debug('repository created: got SVN object: %s', repo) |
|
102 | 102 | return repo |
|
103 | 103 | |
|
104 | 104 | def repo(self, wire, create=False, compatible_version=None): |
|
105 | 105 | """ |
|
106 | 106 | Get a repository instance for the given path. |
|
107 | 107 | """ |
|
108 | 108 | return self._create_repo(wire, create, compatible_version) |
|
109 | 109 | |
|
110 | 110 | |
|
111 | 111 | NODE_TYPE_MAPPING = { |
|
112 | 112 | svn.core.svn_node_file: 'file', |
|
113 | 113 | svn.core.svn_node_dir: 'dir', |
|
114 | 114 | } |
|
115 | 115 | |
|
116 | 116 | |
|
117 | 117 | class SvnRemote(RemoteBase): |
|
118 | 118 | |
|
119 | 119 | def __init__(self, factory, hg_factory=None): |
|
120 | 120 | self._factory = factory |
|
121 | 121 | |
|
122 | 122 | self._bulk_methods = { |
|
123 | 123 | # NOT supported in SVN ATM... |
|
124 | 124 | } |
|
125 | 125 | self._bulk_file_methods = { |
|
126 | 126 | "size": self.get_file_size, |
|
127 | 127 | "data": self.get_file_content, |
|
128 | 128 | "flags": self.get_node_type, |
|
129 | 129 | "is_binary": self.is_binary, |
|
130 | 130 | "md5": self.md5_hash |
|
131 | 131 | } |
|
132 | 132 | |
|
133 | 133 | @reraise_safe_exceptions |
|
134 | 134 | def bulk_file_request(self, wire, commit_id, path, pre_load): |
|
135 | 135 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
136 | 136 | region = self._region(wire) |
|
137 | 137 | |
|
138 | 138 | # since we use unified API, we need to cast from str to in for SVN |
|
139 | 139 | commit_id = int(commit_id) |
|
140 | 140 | |
|
141 | 141 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
142 | 142 | def _bulk_file_request(_repo_id, _commit_id, _path, _pre_load): |
|
143 | 143 | result = {} |
|
144 | 144 | for attr in pre_load: |
|
145 | 145 | try: |
|
146 | 146 | method = self._bulk_file_methods[attr] |
|
147 | 147 | wire.update({'cache': False}) # disable cache for bulk calls so we don't double cache |
|
148 | 148 | result[attr] = method(wire, _commit_id, _path) |
|
149 | 149 | except KeyError as e: |
|
150 | 150 | raise exceptions.VcsException(e)(f'Unknown bulk attribute: "{attr}"') |
|
151 | 151 | return result |
|
152 | 152 | |
|
153 | 153 | return BinaryEnvelope(_bulk_file_request(repo_id, commit_id, path, sorted(pre_load))) |
|
154 | 154 | |
|
155 | 155 | @reraise_safe_exceptions |
|
156 | 156 | def discover_svn_version(self): |
|
157 | 157 | try: |
|
158 | 158 | import svn.core |
|
159 | 159 | svn_ver = svn.core.SVN_VERSION |
|
160 | 160 | except ImportError: |
|
161 | 161 | svn_ver = None |
|
162 | 162 | return safe_str(svn_ver) |
|
163 | 163 | |
|
164 | 164 | @reraise_safe_exceptions |
|
165 | 165 | def is_empty(self, wire): |
|
166 | 166 | try: |
|
167 | 167 | return self.lookup(wire, -1) == 0 |
|
168 | 168 | except Exception: |
|
169 | 169 | log.exception("failed to read object_store") |
|
170 | 170 | return False |
|
171 | 171 | |
|
172 | 172 | def check_url(self, url, config): |
|
173 | 173 | |
|
174 | 174 | # uuid function gets only valid UUID from proper repo, else |
|
175 | 175 | # throws exception |
|
176 | 176 | username, password, src_url = self.get_url_and_credentials(url) |
|
177 | 177 | try: |
|
178 | 178 | svnremoterepo(safe_bytes(username), safe_bytes(password), safe_bytes(src_url)).svn().uuid |
|
179 | 179 | except Exception: |
|
180 | 180 | tb = traceback.format_exc() |
|
181 | 181 | log.debug("Invalid Subversion url: `%s`, tb: %s", url, tb) |
|
182 | 182 | raise URLError(f'"{url}" is not a valid Subversion source url.') |
|
183 | 183 | return True |
|
184 | 184 | |
|
185 | 185 | def is_path_valid_repository(self, wire, path): |
|
186 | 186 | # NOTE(marcink): short circuit the check for SVN repo |
|
187 | 187 | # the repos.open might be expensive to check, but we have one cheap |
|
188 | 188 | # pre-condition that we can use, to check for 'format' file |
|
189 | 189 | if not os.path.isfile(os.path.join(path, 'format')): |
|
190 | 190 | return False |
|
191 | 191 | |
|
192 | 192 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
193 | 193 | region = self._region(wire) |
|
194 | 194 | |
|
195 | 195 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
196 | 196 | def _assert_correct_path(_context_uid, _repo_id, fast_check): |
|
197 | 197 | |
|
198 | 198 | try: |
|
199 | 199 | svn.repos.open(path) |
|
200 | 200 | except svn.core.SubversionException: |
|
201 | 201 | tb = traceback.format_exc() |
|
202 | 202 | log.debug("Invalid Subversion path `%s`, tb: %s", path, tb) |
|
203 | 203 | return False |
|
204 | 204 | return True |
|
205 | 205 | |
|
206 | 206 | return _assert_correct_path(context_uid, repo_id, True) |
|
207 | 207 | |
|
208 | 208 | @reraise_safe_exceptions |
|
209 | 209 | def verify(self, wire,): |
|
210 | 210 | repo_path = wire['path'] |
|
211 | 211 | if not self.is_path_valid_repository(wire, repo_path): |
|
212 | 212 | raise Exception( |
|
213 | 213 | f"Path {repo_path} is not a valid Subversion repository.") |
|
214 | 214 | |
|
215 | 215 | cmd = ['svnadmin', 'info', repo_path] |
|
216 | 216 | stdout, stderr = subprocessio.run_command(cmd) |
|
217 | 217 | return stdout |
|
218 | 218 | |
|
219 | 219 | @reraise_safe_exceptions |
|
220 | 220 | def lookup(self, wire, revision): |
|
221 | 221 | if revision not in [-1, None, 'HEAD']: |
|
222 | 222 | raise NotImplementedError |
|
223 | 223 | repo = self._factory.repo(wire) |
|
224 | 224 | fs_ptr = svn.repos.fs(repo) |
|
225 | 225 | head = svn.fs.youngest_rev(fs_ptr) |
|
226 | 226 | return head |
|
227 | 227 | |
|
228 | 228 | @reraise_safe_exceptions |
|
229 | 229 | def lookup_interval(self, wire, start_ts, end_ts): |
|
230 | 230 | repo = self._factory.repo(wire) |
|
231 | 231 | fsobj = svn.repos.fs(repo) |
|
232 | 232 | start_rev = None |
|
233 | 233 | end_rev = None |
|
234 | 234 | if start_ts: |
|
235 | 235 | start_ts_svn = apr_time_t(start_ts) |
|
236 | 236 | start_rev = svn.repos.dated_revision(repo, start_ts_svn) + 1 |
|
237 | 237 | else: |
|
238 | 238 | start_rev = 1 |
|
239 | 239 | if end_ts: |
|
240 | 240 | end_ts_svn = apr_time_t(end_ts) |
|
241 | 241 | end_rev = svn.repos.dated_revision(repo, end_ts_svn) |
|
242 | 242 | else: |
|
243 | 243 | end_rev = svn.fs.youngest_rev(fsobj) |
|
244 | 244 | return start_rev, end_rev |
|
245 | 245 | |
|
246 | 246 | @reraise_safe_exceptions |
|
247 | 247 | def revision_properties(self, wire, revision): |
|
248 | 248 | |
|
249 | 249 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
250 | 250 | region = self._region(wire) |
|
251 | 251 | |
|
252 | 252 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
253 | 253 | def _revision_properties(_repo_id, _revision): |
|
254 | 254 | repo = self._factory.repo(wire) |
|
255 | 255 | fs_ptr = svn.repos.fs(repo) |
|
256 | 256 | return svn.fs.revision_proplist(fs_ptr, revision) |
|
257 | 257 | return _revision_properties(repo_id, revision) |
|
258 | 258 | |
|
259 | 259 | def revision_changes(self, wire, revision): |
|
260 | 260 | |
|
261 | 261 | repo = self._factory.repo(wire) |
|
262 | 262 | fsobj = svn.repos.fs(repo) |
|
263 | 263 | rev_root = svn.fs.revision_root(fsobj, revision) |
|
264 | 264 | |
|
265 | 265 | editor = svn.repos.ChangeCollector(fsobj, rev_root) |
|
266 | 266 | editor_ptr, editor_baton = svn.delta.make_editor(editor) |
|
267 | 267 | base_dir = "" |
|
268 | 268 | send_deltas = False |
|
269 | 269 | svn.repos.replay2( |
|
270 | 270 | rev_root, base_dir, svn.core.SVN_INVALID_REVNUM, send_deltas, |
|
271 | 271 | editor_ptr, editor_baton, None) |
|
272 | 272 | |
|
273 | 273 | added = [] |
|
274 | 274 | changed = [] |
|
275 | 275 | removed = [] |
|
276 | 276 | |
|
277 | 277 | # TODO: CHANGE_ACTION_REPLACE: Figure out where it belongs |
|
278 | 278 | for path, change in editor.changes.items(): |
|
279 | 279 | # TODO: Decide what to do with directory nodes. Subversion can add |
|
280 | 280 | # empty directories. |
|
281 | 281 | |
|
282 | 282 | if change.item_kind == svn.core.svn_node_dir: |
|
283 | 283 | continue |
|
284 | 284 | if change.action in [svn.repos.CHANGE_ACTION_ADD]: |
|
285 | 285 | added.append(path) |
|
286 | 286 | elif change.action in [svn.repos.CHANGE_ACTION_MODIFY, |
|
287 | 287 | svn.repos.CHANGE_ACTION_REPLACE]: |
|
288 | 288 | changed.append(path) |
|
289 | 289 | elif change.action in [svn.repos.CHANGE_ACTION_DELETE]: |
|
290 | 290 | removed.append(path) |
|
291 | 291 | else: |
|
292 | 292 | raise NotImplementedError( |
|
293 | 293 | "Action {} not supported on path {}".format( |
|
294 | 294 | change.action, path)) |
|
295 | 295 | |
|
296 | 296 | changes = { |
|
297 | 297 | 'added': added, |
|
298 | 298 | 'changed': changed, |
|
299 | 299 | 'removed': removed, |
|
300 | 300 | } |
|
301 | 301 | return changes |
|
302 | 302 | |
|
303 | 303 | @reraise_safe_exceptions |
|
304 | 304 | def node_history(self, wire, path, revision, limit): |
|
305 | 305 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
306 | 306 | region = self._region(wire) |
|
307 | 307 | |
|
308 | 308 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
309 | 309 | def _assert_correct_path(_context_uid, _repo_id, _path, _revision, _limit): |
|
310 | 310 | cross_copies = False |
|
311 | 311 | repo = self._factory.repo(wire) |
|
312 | 312 | fsobj = svn.repos.fs(repo) |
|
313 | 313 | rev_root = svn.fs.revision_root(fsobj, revision) |
|
314 | 314 | |
|
315 | 315 | history_revisions = [] |
|
316 | 316 | history = svn.fs.node_history(rev_root, path) |
|
317 | 317 | history = svn.fs.history_prev(history, cross_copies) |
|
318 | 318 | while history: |
|
319 | 319 | __, node_revision = svn.fs.history_location(history) |
|
320 | 320 | history_revisions.append(node_revision) |
|
321 | 321 | if limit and len(history_revisions) >= limit: |
|
322 | 322 | break |
|
323 | 323 | history = svn.fs.history_prev(history, cross_copies) |
|
324 | 324 | return history_revisions |
|
325 | 325 | return _assert_correct_path(context_uid, repo_id, path, revision, limit) |
|
326 | 326 | |
|
327 | 327 | @reraise_safe_exceptions |
|
328 | 328 | def node_properties(self, wire, path, revision): |
|
329 | 329 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
330 | 330 | region = self._region(wire) |
|
331 | 331 | |
|
332 | 332 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
333 | 333 | def _node_properties(_repo_id, _path, _revision): |
|
334 | 334 | repo = self._factory.repo(wire) |
|
335 | 335 | fsobj = svn.repos.fs(repo) |
|
336 | 336 | rev_root = svn.fs.revision_root(fsobj, revision) |
|
337 | 337 | return svn.fs.node_proplist(rev_root, path) |
|
338 | 338 | return _node_properties(repo_id, path, revision) |
|
339 | 339 | |
|
340 | 340 | def file_annotate(self, wire, path, revision): |
|
341 | 341 | abs_path = 'file://' + urllib.request.pathname2url( |
|
342 | 342 | vcspath.join(wire['path'], path)) |
|
343 | 343 | file_uri = svn.core.svn_path_canonicalize(abs_path) |
|
344 | 344 | |
|
345 | 345 | start_rev = svn_opt_revision_value_t(0) |
|
346 | 346 | peg_rev = svn_opt_revision_value_t(revision) |
|
347 | 347 | end_rev = peg_rev |
|
348 | 348 | |
|
349 | 349 | annotations = [] |
|
350 | 350 | |
|
351 | 351 | def receiver(line_no, revision, author, date, line, pool): |
|
352 | 352 | annotations.append((line_no, revision, line)) |
|
353 | 353 | |
|
354 | 354 | # TODO: Cannot use blame5, missing typemap function in the swig code |
|
355 | 355 | try: |
|
356 | 356 | svn.client.blame2( |
|
357 | 357 | file_uri, peg_rev, start_rev, end_rev, |
|
358 | 358 | receiver, svn.client.create_context()) |
|
359 | 359 | except svn.core.SubversionException as exc: |
|
360 | 360 | log.exception("Error during blame operation.") |
|
361 | 361 | raise Exception( |
|
362 | 362 | f"Blame not supported or file does not exist at path {path}. " |
|
363 | 363 | f"Error {exc}.") |
|
364 | 364 | |
|
365 | 365 | return BinaryEnvelope(annotations) |
|
366 | 366 | |
|
367 | 367 | @reraise_safe_exceptions |
|
368 | 368 | def get_node_type(self, wire, revision=None, path=''): |
|
369 | 369 | |
|
370 | 370 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
371 | 371 | region = self._region(wire) |
|
372 | 372 | |
|
373 | 373 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
374 | 374 | def _get_node_type(_repo_id, _revision, _path): |
|
375 | 375 | repo = self._factory.repo(wire) |
|
376 | 376 | fs_ptr = svn.repos.fs(repo) |
|
377 | 377 | if _revision is None: |
|
378 | 378 | _revision = svn.fs.youngest_rev(fs_ptr) |
|
379 | 379 | root = svn.fs.revision_root(fs_ptr, _revision) |
|
380 | 380 | node = svn.fs.check_path(root, path) |
|
381 | 381 | return NODE_TYPE_MAPPING.get(node, None) |
|
382 | 382 | return _get_node_type(repo_id, revision, path) |
|
383 | 383 | |
|
384 | 384 | @reraise_safe_exceptions |
|
385 | 385 | def get_nodes(self, wire, revision=None, path=''): |
|
386 | 386 | |
|
387 | 387 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
388 | 388 | region = self._region(wire) |
|
389 | 389 | |
|
390 | 390 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
391 | 391 | def _get_nodes(_repo_id, _path, _revision): |
|
392 | 392 | repo = self._factory.repo(wire) |
|
393 | 393 | fsobj = svn.repos.fs(repo) |
|
394 | 394 | if _revision is None: |
|
395 | 395 | _revision = svn.fs.youngest_rev(fsobj) |
|
396 | 396 | root = svn.fs.revision_root(fsobj, _revision) |
|
397 | 397 | entries = svn.fs.dir_entries(root, path) |
|
398 | 398 | result = [] |
|
399 | 399 | for entry_path, entry_info in entries.items(): |
|
400 | 400 | result.append( |
|
401 | 401 | (entry_path, NODE_TYPE_MAPPING.get(entry_info.kind, None))) |
|
402 | 402 | return result |
|
403 | 403 | return _get_nodes(repo_id, path, revision) |
|
404 | 404 | |
|
405 | 405 | @reraise_safe_exceptions |
|
406 | 406 | def get_file_content(self, wire, rev=None, path=''): |
|
407 | 407 | repo = self._factory.repo(wire) |
|
408 | 408 | fsobj = svn.repos.fs(repo) |
|
409 | 409 | |
|
410 | 410 | if rev is None: |
|
411 | 411 | rev = svn.fs.youngest_rev(fsobj) |
|
412 | 412 | |
|
413 | 413 | root = svn.fs.revision_root(fsobj, rev) |
|
414 | 414 | content = svn.core.Stream(svn.fs.file_contents(root, path)) |
|
415 | 415 | return BytesEnvelope(content.read()) |
|
416 | 416 | |
|
417 | 417 | @reraise_safe_exceptions |
|
418 | 418 | def get_file_size(self, wire, revision=None, path=''): |
|
419 | 419 | |
|
420 | 420 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
421 | 421 | region = self._region(wire) |
|
422 | 422 | |
|
423 | 423 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
424 | 424 | def _get_file_size(_repo_id, _revision, _path): |
|
425 | 425 | repo = self._factory.repo(wire) |
|
426 | 426 | fsobj = svn.repos.fs(repo) |
|
427 | 427 | if _revision is None: |
|
428 | 428 | _revision = svn.fs.youngest_revision(fsobj) |
|
429 | 429 | root = svn.fs.revision_root(fsobj, _revision) |
|
430 | 430 | size = svn.fs.file_length(root, path) |
|
431 | 431 | return size |
|
432 | 432 | return _get_file_size(repo_id, revision, path) |
|
433 | 433 | |
|
434 | 434 | def create_repository(self, wire, compatible_version=None): |
|
435 | 435 | log.info('Creating Subversion repository in path "%s"', wire['path']) |
|
436 | 436 | self._factory.repo(wire, create=True, |
|
437 | 437 | compatible_version=compatible_version) |
|
438 | 438 | |
|
439 | 439 | def get_url_and_credentials(self, src_url) -> tuple[str, str, str]: |
|
440 | 440 | obj = urllib.parse.urlparse(src_url) |
|
441 | 441 | username = obj.username or '' |
|
442 | 442 | password = obj.password or '' |
|
443 | 443 | return username, password, src_url |
|
444 | 444 | |
|
445 | 445 | def import_remote_repository(self, wire, src_url): |
|
446 | 446 | repo_path = wire['path'] |
|
447 | 447 | if not self.is_path_valid_repository(wire, repo_path): |
|
448 | 448 | raise Exception( |
|
449 | 449 | f"Path {repo_path} is not a valid Subversion repository.") |
|
450 | 450 | |
|
451 | 451 | username, password, src_url = self.get_url_and_credentials(src_url) |
|
452 | 452 | rdump_cmd = ['svnrdump', 'dump', '--non-interactive', |
|
453 | 453 | '--trust-server-cert-failures=unknown-ca'] |
|
454 | 454 | if username and password: |
|
455 | 455 | rdump_cmd += ['--username', username, '--password', password] |
|
456 | 456 | rdump_cmd += [src_url] |
|
457 | 457 | |
|
458 | 458 | rdump = subprocess.Popen( |
|
459 | 459 | rdump_cmd, |
|
460 | 460 | stdout=subprocess.PIPE, stderr=subprocess.PIPE) |
|
461 | 461 | load = subprocess.Popen( |
|
462 | 462 | ['svnadmin', 'load', repo_path], stdin=rdump.stdout) |
|
463 | 463 | |
|
464 | 464 | # TODO: johbo: This can be a very long operation, might be better |
|
465 | 465 | # to track some kind of status and provide an api to check if the |
|
466 | 466 | # import is done. |
|
467 | 467 | rdump.wait() |
|
468 | 468 | load.wait() |
|
469 | 469 | |
|
470 | 470 | log.debug('Return process ended with code: %s', rdump.returncode) |
|
471 | 471 | if rdump.returncode != 0: |
|
472 | 472 | errors = rdump.stderr.read() |
|
473 | 473 | log.error('svnrdump dump failed: statuscode %s: message: %s', rdump.returncode, errors) |
|
474 | 474 | |
|
475 | 475 | reason = 'UNKNOWN' |
|
476 | 476 | if b'svnrdump: E230001:' in errors: |
|
477 | 477 | reason = 'INVALID_CERTIFICATE' |
|
478 | 478 | |
|
479 | 479 | if reason == 'UNKNOWN': |
|
480 | 480 | reason = f'UNKNOWN:{safe_str(errors)}' |
|
481 | 481 | |
|
482 | 482 | raise Exception( |
|
483 | 483 | 'Failed to dump the remote repository from {}. Reason:{}'.format( |
|
484 | 484 | src_url, reason)) |
|
485 | 485 | if load.returncode != 0: |
|
486 | 486 | raise Exception( |
|
487 | 487 | f'Failed to load the dump of remote repository from {src_url}.') |
|
488 | 488 | |
|
489 | 489 | def commit(self, wire, message, author, timestamp, updated, removed): |
|
490 | 490 | |
|
491 | 491 | message = safe_bytes(message) |
|
492 | 492 | author = safe_bytes(author) |
|
493 | 493 | |
|
494 | 494 | repo = self._factory.repo(wire) |
|
495 | 495 | fsobj = svn.repos.fs(repo) |
|
496 | 496 | |
|
497 | 497 | rev = svn.fs.youngest_rev(fsobj) |
|
498 | 498 | txn = svn.repos.fs_begin_txn_for_commit(repo, rev, author, message) |
|
499 | 499 | txn_root = svn.fs.txn_root(txn) |
|
500 | 500 | |
|
501 | 501 | for node in updated: |
|
502 | 502 | TxnNodeProcessor(node, txn_root).update() |
|
503 | 503 | for node in removed: |
|
504 | 504 | TxnNodeProcessor(node, txn_root).remove() |
|
505 | 505 | |
|
506 | svn_txn_id = safe_str(svn.fs.svn_fs_txn_name(txn)) | |
|
507 | full_repo_path = wire['path'] | |
|
508 | txn_id_data = {'svn_txn_id': svn_txn_id, 'rc_internal_commit': True} | |
|
509 | ||
|
510 | store_txn_id_data(full_repo_path, svn_txn_id, txn_id_data) | |
|
506 | 511 | commit_id = svn.repos.fs_commit_txn(repo, txn) |
|
507 | 512 | |
|
508 | 513 | if timestamp: |
|
509 | 514 | apr_time = apr_time_t(timestamp) |
|
510 | 515 | ts_formatted = svn.core.svn_time_to_cstring(apr_time) |
|
511 | 516 | svn.fs.change_rev_prop(fsobj, commit_id, 'svn:date', ts_formatted) |
|
512 | 517 | |
|
513 | 518 | log.debug('Committed revision "%s" to "%s".', commit_id, wire['path']) |
|
514 | 519 | return commit_id |
|
515 | 520 | |
|
516 | 521 | @reraise_safe_exceptions |
|
517 | 522 | def diff(self, wire, rev1, rev2, path1=None, path2=None, |
|
518 | 523 | ignore_whitespace=False, context=3): |
|
519 | 524 | |
|
520 | 525 | wire.update(cache=False) |
|
521 | 526 | repo = self._factory.repo(wire) |
|
522 | 527 | diff_creator = SvnDiffer( |
|
523 | 528 | repo, rev1, path1, rev2, path2, ignore_whitespace, context) |
|
524 | 529 | try: |
|
525 | 530 | return BytesEnvelope(diff_creator.generate_diff()) |
|
526 | 531 | except svn.core.SubversionException as e: |
|
527 | 532 | log.exception( |
|
528 | 533 | "Error during diff operation operation. " |
|
529 | 534 | "Path might not exist %s, %s", path1, path2) |
|
530 | 535 | return BytesEnvelope(b'') |
|
531 | 536 | |
|
532 | 537 | @reraise_safe_exceptions |
|
533 | 538 | def is_large_file(self, wire, path): |
|
534 | 539 | return False |
|
535 | 540 | |
|
536 | 541 | @reraise_safe_exceptions |
|
537 | 542 | def is_binary(self, wire, rev, path): |
|
538 | 543 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
539 | 544 | region = self._region(wire) |
|
540 | 545 | |
|
541 | 546 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
542 | 547 | def _is_binary(_repo_id, _rev, _path): |
|
543 | 548 | raw_bytes = self.get_file_content(wire, rev, path) |
|
544 | 549 | if not raw_bytes: |
|
545 | 550 | return False |
|
546 | 551 | return b'\0' in raw_bytes |
|
547 | 552 | |
|
548 | 553 | return _is_binary(repo_id, rev, path) |
|
549 | 554 | |
|
550 | 555 | @reraise_safe_exceptions |
|
551 | 556 | def md5_hash(self, wire, rev, path): |
|
552 | 557 | cache_on, context_uid, repo_id = self._cache_on(wire) |
|
553 | 558 | region = self._region(wire) |
|
554 | 559 | |
|
555 | 560 | @region.conditional_cache_on_arguments(condition=cache_on) |
|
556 | 561 | def _md5_hash(_repo_id, _rev, _path): |
|
557 | 562 | return '' |
|
558 | 563 | |
|
559 | 564 | return _md5_hash(repo_id, rev, path) |
|
560 | 565 | |
|
561 | 566 | @reraise_safe_exceptions |
|
562 | 567 | def run_svn_command(self, wire, cmd, **opts): |
|
563 | 568 | path = wire.get('path', None) |
|
564 | 569 | debug_mode = rhodecode.ConfigGet().get_bool('debug') |
|
565 | 570 | |
|
566 | 571 | if path and os.path.isdir(path): |
|
567 | 572 | opts['cwd'] = path |
|
568 | 573 | |
|
569 | 574 | safe_call = opts.pop('_safe', False) |
|
570 | 575 | |
|
571 | 576 | svnenv = os.environ.copy() |
|
572 | 577 | svnenv.update(opts.pop('extra_env', {})) |
|
573 | 578 | |
|
574 | 579 | _opts = {'env': svnenv, 'shell': False} |
|
575 | 580 | |
|
576 | 581 | try: |
|
577 | 582 | _opts.update(opts) |
|
578 | 583 | proc = subprocessio.SubprocessIOChunker(cmd, **_opts) |
|
579 | 584 | |
|
580 | 585 | return b''.join(proc), b''.join(proc.stderr) |
|
581 | 586 | except OSError as err: |
|
582 | 587 | if safe_call: |
|
583 | 588 | return '', safe_str(err).strip() |
|
584 | 589 | else: |
|
585 | 590 | cmd = ' '.join(map(safe_str, cmd)) # human friendly CMD |
|
586 | 591 | call_opts = {} |
|
587 | 592 | if debug_mode: |
|
588 | 593 | call_opts = _opts |
|
589 | 594 | |
|
590 | 595 | tb_err = ("Couldn't run svn command ({}).\n" |
|
591 | 596 | "Original error was:{}\n" |
|
592 | 597 | "Call options:{}\n" |
|
593 | 598 | .format(cmd, err, call_opts)) |
|
594 | 599 | log.exception(tb_err) |
|
595 | 600 | raise exceptions.VcsException()(tb_err) |
|
596 | 601 | |
|
597 | 602 | @reraise_safe_exceptions |
|
598 | 603 | def install_hooks(self, wire, force=False): |
|
599 | 604 | from vcsserver.hook_utils import install_svn_hooks |
|
600 | 605 | repo_path = wire['path'] |
|
601 | 606 | binary_dir = settings.BINARY_DIR |
|
602 | 607 | executable = None |
|
603 | 608 | if binary_dir: |
|
604 | 609 | executable = os.path.join(binary_dir, 'python3') |
|
605 | 610 | return install_svn_hooks(repo_path, force_create=force) |
|
606 | 611 | |
|
607 | 612 | @reraise_safe_exceptions |
|
608 | 613 | def get_hooks_info(self, wire): |
|
609 | 614 | from vcsserver.hook_utils import ( |
|
610 | 615 | get_svn_pre_hook_version, get_svn_post_hook_version) |
|
611 | 616 | repo_path = wire['path'] |
|
612 | 617 | return { |
|
613 | 618 | 'pre_version': get_svn_pre_hook_version(repo_path), |
|
614 | 619 | 'post_version': get_svn_post_hook_version(repo_path), |
|
615 | 620 | } |
|
616 | 621 | |
|
617 | 622 | @reraise_safe_exceptions |
|
618 | 623 | def set_head_ref(self, wire, head_name): |
|
619 | 624 | pass |
|
620 | 625 | |
|
621 | 626 | @reraise_safe_exceptions |
|
622 | 627 | def archive_repo(self, wire, archive_name_key, kind, mtime, archive_at_path, |
|
623 | 628 | archive_dir_name, commit_id, cache_config): |
|
624 | 629 | |
|
625 | 630 | def walk_tree(root, root_dir, _commit_id): |
|
626 | 631 | """ |
|
627 | 632 | Special recursive svn repo walker |
|
628 | 633 | """ |
|
629 | 634 | root_dir = safe_bytes(root_dir) |
|
630 | 635 | |
|
631 | 636 | filemode_default = 0o100644 |
|
632 | 637 | filemode_executable = 0o100755 |
|
633 | 638 | |
|
634 | 639 | file_iter = svn.fs.dir_entries(root, root_dir) |
|
635 | 640 | for f_name in file_iter: |
|
636 | 641 | f_type = NODE_TYPE_MAPPING.get(file_iter[f_name].kind, None) |
|
637 | 642 | |
|
638 | 643 | if f_type == 'dir': |
|
639 | 644 | # return only DIR, and then all entries in that dir |
|
640 | 645 | yield os.path.join(root_dir, f_name), {'mode': filemode_default}, f_type |
|
641 | 646 | new_root = os.path.join(root_dir, f_name) |
|
642 | 647 | yield from walk_tree(root, new_root, _commit_id) |
|
643 | 648 | else: |
|
644 | 649 | |
|
645 | 650 | f_path = os.path.join(root_dir, f_name).rstrip(b'/') |
|
646 | 651 | prop_list = svn.fs.node_proplist(root, f_path) |
|
647 | 652 | |
|
648 | 653 | f_mode = filemode_default |
|
649 | 654 | if prop_list.get('svn:executable'): |
|
650 | 655 | f_mode = filemode_executable |
|
651 | 656 | |
|
652 | 657 | f_is_link = False |
|
653 | 658 | if prop_list.get('svn:special'): |
|
654 | 659 | f_is_link = True |
|
655 | 660 | |
|
656 | 661 | data = { |
|
657 | 662 | 'is_link': f_is_link, |
|
658 | 663 | 'mode': f_mode, |
|
659 | 664 | 'content_stream': svn.core.Stream(svn.fs.file_contents(root, f_path)).read |
|
660 | 665 | } |
|
661 | 666 | |
|
662 | 667 | yield f_path, data, f_type |
|
663 | 668 | |
|
664 | 669 | def file_walker(_commit_id, path): |
|
665 | 670 | repo = self._factory.repo(wire) |
|
666 | 671 | root = svn.fs.revision_root(svn.repos.fs(repo), int(commit_id)) |
|
667 | 672 | |
|
668 | 673 | def no_content(): |
|
669 | 674 | raise NoContentException() |
|
670 | 675 | |
|
671 | 676 | for f_name, f_data, f_type in walk_tree(root, path, _commit_id): |
|
672 | 677 | file_path = f_name |
|
673 | 678 | |
|
674 | 679 | if f_type == 'dir': |
|
675 | 680 | mode = f_data['mode'] |
|
676 | 681 | yield ArchiveNode(file_path, mode, False, no_content) |
|
677 | 682 | else: |
|
678 | 683 | mode = f_data['mode'] |
|
679 | 684 | is_link = f_data['is_link'] |
|
680 | 685 | data_stream = f_data['content_stream'] |
|
681 | 686 | yield ArchiveNode(file_path, mode, is_link, data_stream) |
|
682 | 687 | |
|
683 | 688 | return store_archive_in_cache( |
|
684 | 689 | file_walker, archive_name_key, kind, mtime, archive_at_path, archive_dir_name, commit_id, cache_config=cache_config) |
|
685 | 690 | |
|
686 | 691 | |
|
687 | 692 | class SvnDiffer: |
|
688 | 693 | """ |
|
689 | 694 | Utility to create diffs based on difflib and the Subversion api |
|
690 | 695 | """ |
|
691 | 696 | |
|
692 | 697 | binary_content = False |
|
693 | 698 | |
|
694 | 699 | def __init__( |
|
695 | 700 | self, repo, src_rev, src_path, tgt_rev, tgt_path, |
|
696 | 701 | ignore_whitespace, context): |
|
697 | 702 | self.repo = repo |
|
698 | 703 | self.ignore_whitespace = ignore_whitespace |
|
699 | 704 | self.context = context |
|
700 | 705 | |
|
701 | 706 | fsobj = svn.repos.fs(repo) |
|
702 | 707 | |
|
703 | 708 | self.tgt_rev = tgt_rev |
|
704 | 709 | self.tgt_path = tgt_path or '' |
|
705 | 710 | self.tgt_root = svn.fs.revision_root(fsobj, tgt_rev) |
|
706 | 711 | self.tgt_kind = svn.fs.check_path(self.tgt_root, self.tgt_path) |
|
707 | 712 | |
|
708 | 713 | self.src_rev = src_rev |
|
709 | 714 | self.src_path = src_path or self.tgt_path |
|
710 | 715 | self.src_root = svn.fs.revision_root(fsobj, src_rev) |
|
711 | 716 | self.src_kind = svn.fs.check_path(self.src_root, self.src_path) |
|
712 | 717 | |
|
713 | 718 | self._validate() |
|
714 | 719 | |
|
715 | 720 | def _validate(self): |
|
716 | 721 | if (self.tgt_kind != svn.core.svn_node_none and |
|
717 | 722 | self.src_kind != svn.core.svn_node_none and |
|
718 | 723 | self.src_kind != self.tgt_kind): |
|
719 | 724 | # TODO: johbo: proper error handling |
|
720 | 725 | raise Exception( |
|
721 | 726 | "Source and target are not compatible for diff generation. " |
|
722 | 727 | "Source type: %s, target type: %s" % |
|
723 | 728 | (self.src_kind, self.tgt_kind)) |
|
724 | 729 | |
|
725 | 730 | def generate_diff(self) -> bytes: |
|
726 | 731 | buf = io.BytesIO() |
|
727 | 732 | if self.tgt_kind == svn.core.svn_node_dir: |
|
728 | 733 | self._generate_dir_diff(buf) |
|
729 | 734 | else: |
|
730 | 735 | self._generate_file_diff(buf) |
|
731 | 736 | return buf.getvalue() |
|
732 | 737 | |
|
733 | 738 | def _generate_dir_diff(self, buf: io.BytesIO): |
|
734 | 739 | editor = DiffChangeEditor() |
|
735 | 740 | editor_ptr, editor_baton = svn.delta.make_editor(editor) |
|
736 | 741 | svn.repos.dir_delta2( |
|
737 | 742 | self.src_root, |
|
738 | 743 | self.src_path, |
|
739 | 744 | '', # src_entry |
|
740 | 745 | self.tgt_root, |
|
741 | 746 | self.tgt_path, |
|
742 | 747 | editor_ptr, editor_baton, |
|
743 | 748 | authorization_callback_allow_all, |
|
744 | 749 | False, # text_deltas |
|
745 | 750 | svn.core.svn_depth_infinity, # depth |
|
746 | 751 | False, # entry_props |
|
747 | 752 | False, # ignore_ancestry |
|
748 | 753 | ) |
|
749 | 754 | |
|
750 | 755 | for path, __, change in sorted(editor.changes): |
|
751 | 756 | self._generate_node_diff( |
|
752 | 757 | buf, change, path, self.tgt_path, path, self.src_path) |
|
753 | 758 | |
|
754 | 759 | def _generate_file_diff(self, buf: io.BytesIO): |
|
755 | 760 | change = None |
|
756 | 761 | if self.src_kind == svn.core.svn_node_none: |
|
757 | 762 | change = "add" |
|
758 | 763 | elif self.tgt_kind == svn.core.svn_node_none: |
|
759 | 764 | change = "delete" |
|
760 | 765 | tgt_base, tgt_path = vcspath.split(self.tgt_path) |
|
761 | 766 | src_base, src_path = vcspath.split(self.src_path) |
|
762 | 767 | self._generate_node_diff( |
|
763 | 768 | buf, change, tgt_path, tgt_base, src_path, src_base) |
|
764 | 769 | |
|
765 | 770 | def _generate_node_diff( |
|
766 | 771 | self, buf: io.BytesIO, change, tgt_path, tgt_base, src_path, src_base): |
|
767 | 772 | |
|
768 | 773 | tgt_path_bytes = safe_bytes(tgt_path) |
|
769 | 774 | tgt_path = safe_str(tgt_path) |
|
770 | 775 | |
|
771 | 776 | src_path_bytes = safe_bytes(src_path) |
|
772 | 777 | src_path = safe_str(src_path) |
|
773 | 778 | |
|
774 | 779 | if self.src_rev == self.tgt_rev and tgt_base == src_base: |
|
775 | 780 | # makes consistent behaviour with git/hg to return empty diff if |
|
776 | 781 | # we compare same revisions |
|
777 | 782 | return |
|
778 | 783 | |
|
779 | 784 | tgt_full_path = vcspath.join(tgt_base, tgt_path) |
|
780 | 785 | src_full_path = vcspath.join(src_base, src_path) |
|
781 | 786 | |
|
782 | 787 | self.binary_content = False |
|
783 | 788 | mime_type = self._get_mime_type(tgt_full_path) |
|
784 | 789 | |
|
785 | 790 | if mime_type and not mime_type.startswith(b'text'): |
|
786 | 791 | self.binary_content = True |
|
787 | 792 | buf.write(b"=" * 67 + b'\n') |
|
788 | 793 | buf.write(b"Cannot display: file marked as a binary type.\n") |
|
789 | 794 | buf.write(b"svn:mime-type = %s\n" % mime_type) |
|
790 | 795 | buf.write(b"Index: %b\n" % tgt_path_bytes) |
|
791 | 796 | buf.write(b"=" * 67 + b'\n') |
|
792 | 797 | buf.write(b"diff --git a/%b b/%b\n" % (tgt_path_bytes, tgt_path_bytes)) |
|
793 | 798 | |
|
794 | 799 | if change == 'add': |
|
795 | 800 | # TODO: johbo: SVN is missing a zero here compared to git |
|
796 | 801 | buf.write(b"new file mode 10644\n") |
|
797 | 802 | |
|
798 | 803 | # TODO(marcink): intro to binary detection of svn patches |
|
799 | 804 | # if self.binary_content: |
|
800 | 805 | # buf.write(b'GIT binary patch\n') |
|
801 | 806 | |
|
802 | 807 | buf.write(b"--- /dev/null\t(revision 0)\n") |
|
803 | 808 | src_lines = [] |
|
804 | 809 | else: |
|
805 | 810 | if change == 'delete': |
|
806 | 811 | buf.write(b"deleted file mode 10644\n") |
|
807 | 812 | |
|
808 | 813 | # TODO(marcink): intro to binary detection of svn patches |
|
809 | 814 | # if self.binary_content: |
|
810 | 815 | # buf.write('GIT binary patch\n') |
|
811 | 816 | |
|
812 | 817 | buf.write(b"--- a/%b\t(revision %d)\n" % (src_path_bytes, self.src_rev)) |
|
813 | 818 | src_lines = self._svn_readlines(self.src_root, src_full_path) |
|
814 | 819 | |
|
815 | 820 | if change == 'delete': |
|
816 | 821 | buf.write(b"+++ /dev/null\t(revision %d)\n" % self.tgt_rev) |
|
817 | 822 | tgt_lines = [] |
|
818 | 823 | else: |
|
819 | 824 | buf.write(b"+++ b/%b\t(revision %d)\n" % (tgt_path_bytes, self.tgt_rev)) |
|
820 | 825 | tgt_lines = self._svn_readlines(self.tgt_root, tgt_full_path) |
|
821 | 826 | |
|
822 | 827 | # we made our diff header, time to generate the diff content into our buffer |
|
823 | 828 | |
|
824 | 829 | if not self.binary_content: |
|
825 | 830 | udiff = svn_diff.unified_diff( |
|
826 | 831 | src_lines, tgt_lines, context=self.context, |
|
827 | 832 | ignore_blank_lines=self.ignore_whitespace, |
|
828 | 833 | ignore_case=False, |
|
829 | 834 | ignore_space_changes=self.ignore_whitespace) |
|
830 | 835 | |
|
831 | 836 | buf.writelines(udiff) |
|
832 | 837 | |
|
833 | 838 | def _get_mime_type(self, path) -> bytes: |
|
834 | 839 | try: |
|
835 | 840 | mime_type = svn.fs.node_prop( |
|
836 | 841 | self.tgt_root, path, svn.core.SVN_PROP_MIME_TYPE) |
|
837 | 842 | except svn.core.SubversionException: |
|
838 | 843 | mime_type = svn.fs.node_prop( |
|
839 | 844 | self.src_root, path, svn.core.SVN_PROP_MIME_TYPE) |
|
840 | 845 | return mime_type |
|
841 | 846 | |
|
842 | 847 | def _svn_readlines(self, fs_root, node_path): |
|
843 | 848 | if self.binary_content: |
|
844 | 849 | return [] |
|
845 | 850 | node_kind = svn.fs.check_path(fs_root, node_path) |
|
846 | 851 | if node_kind not in ( |
|
847 | 852 | svn.core.svn_node_file, svn.core.svn_node_symlink): |
|
848 | 853 | return [] |
|
849 | 854 | content = svn.core.Stream( |
|
850 | 855 | svn.fs.file_contents(fs_root, node_path)).read() |
|
851 | 856 | |
|
852 | 857 | return content.splitlines(True) |
|
853 | 858 | |
|
854 | 859 | |
|
855 | 860 | class DiffChangeEditor(svn.delta.Editor): |
|
856 | 861 | """ |
|
857 | 862 | Records changes between two given revisions |
|
858 | 863 | """ |
|
859 | 864 | |
|
860 | 865 | def __init__(self): |
|
861 | 866 | self.changes = [] |
|
862 | 867 | |
|
863 | 868 | def delete_entry(self, path, revision, parent_baton, pool=None): |
|
864 | 869 | self.changes.append((path, None, 'delete')) |
|
865 | 870 | |
|
866 | 871 | def add_file( |
|
867 | 872 | self, path, parent_baton, copyfrom_path, copyfrom_revision, |
|
868 | 873 | file_pool=None): |
|
869 | 874 | self.changes.append((path, 'file', 'add')) |
|
870 | 875 | |
|
871 | 876 | def open_file(self, path, parent_baton, base_revision, file_pool=None): |
|
872 | 877 | self.changes.append((path, 'file', 'change')) |
|
873 | 878 | |
|
874 | 879 | |
|
875 | 880 | def authorization_callback_allow_all(root, path, pool): |
|
876 | 881 | return True |
|
877 | 882 | |
|
878 | 883 | |
|
879 | 884 | class TxnNodeProcessor: |
|
880 | 885 | """ |
|
881 | 886 | Utility to process the change of one node within a transaction root. |
|
882 | 887 | |
|
883 | 888 | It encapsulates the knowledge of how to add, update or remove |
|
884 | 889 | a node for a given transaction root. The purpose is to support the method |
|
885 | 890 | `SvnRemote.commit`. |
|
886 | 891 | """ |
|
887 | 892 | |
|
888 | 893 | def __init__(self, node, txn_root): |
|
889 | 894 | assert_bytes(node['path']) |
|
890 | 895 | |
|
891 | 896 | self.node = node |
|
892 | 897 | self.txn_root = txn_root |
|
893 | 898 | |
|
894 | 899 | def update(self): |
|
895 | 900 | self._ensure_parent_dirs() |
|
896 | 901 | self._add_file_if_node_does_not_exist() |
|
897 | 902 | self._update_file_content() |
|
898 | 903 | self._update_file_properties() |
|
899 | 904 | |
|
900 | 905 | def remove(self): |
|
901 | 906 | svn.fs.delete(self.txn_root, self.node['path']) |
|
902 | 907 | # TODO: Clean up directory if empty |
|
903 | 908 | |
|
904 | 909 | def _ensure_parent_dirs(self): |
|
905 | 910 | curdir = vcspath.dirname(self.node['path']) |
|
906 | 911 | dirs_to_create = [] |
|
907 | 912 | while not self._svn_path_exists(curdir): |
|
908 | 913 | dirs_to_create.append(curdir) |
|
909 | 914 | curdir = vcspath.dirname(curdir) |
|
910 | 915 | |
|
911 | 916 | for curdir in reversed(dirs_to_create): |
|
912 | 917 | log.debug('Creating missing directory "%s"', curdir) |
|
913 | 918 | svn.fs.make_dir(self.txn_root, curdir) |
|
914 | 919 | |
|
915 | 920 | def _svn_path_exists(self, path): |
|
916 | 921 | path_status = svn.fs.check_path(self.txn_root, path) |
|
917 | 922 | return path_status != svn.core.svn_node_none |
|
918 | 923 | |
|
919 | 924 | def _add_file_if_node_does_not_exist(self): |
|
920 | 925 | kind = svn.fs.check_path(self.txn_root, self.node['path']) |
|
921 | 926 | if kind == svn.core.svn_node_none: |
|
922 | 927 | svn.fs.make_file(self.txn_root, self.node['path']) |
|
923 | 928 | |
|
924 | 929 | def _update_file_content(self): |
|
925 | 930 | assert_bytes(self.node['content']) |
|
926 | 931 | |
|
927 | 932 | handler, baton = svn.fs.apply_textdelta( |
|
928 | 933 | self.txn_root, self.node['path'], None, None) |
|
929 | 934 | svn.delta.svn_txdelta_send_string(self.node['content'], handler, baton) |
|
930 | 935 | |
|
931 | 936 | def _update_file_properties(self): |
|
932 | 937 | properties = self.node.get('properties', {}) |
|
933 | 938 | for key, value in properties.items(): |
|
934 | 939 | svn.fs.change_node_prop( |
|
935 | 940 | self.txn_root, self.node['path'], safe_bytes(key), safe_bytes(value)) |
|
936 | 941 | |
|
937 | 942 | |
|
938 | 943 | def apr_time_t(timestamp): |
|
939 | 944 | """ |
|
940 | 945 | Convert a Python timestamp into APR timestamp type apr_time_t |
|
941 | 946 | """ |
|
942 | 947 | return int(timestamp * 1E6) |
|
943 | 948 | |
|
944 | 949 | |
|
945 | 950 | def svn_opt_revision_value_t(num): |
|
946 | 951 | """ |
|
947 | 952 | Put `num` into a `svn_opt_revision_value_t` structure. |
|
948 | 953 | """ |
|
949 | 954 | value = svn.core.svn_opt_revision_value_t() |
|
950 | 955 | value.number = num |
|
951 | 956 | revision = svn.core.svn_opt_revision_t() |
|
952 | 957 | revision.kind = svn.core.svn_opt_revision_number |
|
953 | 958 | revision.value = value |
|
954 | 959 | return revision |
@@ -1,255 +1,258 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import os |
|
19 | 19 | import logging |
|
20 | 20 | import itertools |
|
21 | 21 | |
|
22 | 22 | import mercurial |
|
23 | 23 | import mercurial.error |
|
24 | 24 | import mercurial.wireprotoserver |
|
25 | 25 | import mercurial.hgweb.common |
|
26 | 26 | import mercurial.hgweb.hgweb_mod |
|
27 | 27 | import webob.exc |
|
28 | 28 | |
|
29 | 29 | from vcsserver import pygrack, exceptions, settings, git_lfs |
|
30 | from vcsserver.str_utils import ascii_bytes, safe_bytes | |
|
30 | from vcsserver.lib.str_utils import ascii_bytes, safe_bytes | |
|
31 | 31 | |
|
32 | 32 | log = logging.getLogger(__name__) |
|
33 | 33 | |
|
34 | 34 | |
|
35 | 35 | # propagated from mercurial documentation |
|
36 | 36 | HG_UI_SECTIONS = [ |
|
37 | 37 | 'alias', 'auth', 'decode/encode', 'defaults', 'diff', 'email', 'extensions', |
|
38 | 38 | 'format', 'merge-patterns', 'merge-tools', 'hooks', 'http_proxy', 'smtp', |
|
39 | 39 | 'patch', 'paths', 'profiling', 'server', 'trusted', 'ui', 'web', |
|
40 | 40 | ] |
|
41 | 41 | |
|
42 | 42 | |
|
43 | 43 | class HgWeb(mercurial.hgweb.hgweb_mod.hgweb): |
|
44 | 44 | """Extension of hgweb that simplifies some functions.""" |
|
45 | 45 | |
|
46 | 46 | def _get_view(self, repo): |
|
47 | 47 | """Views are not supported.""" |
|
48 | 48 | return repo |
|
49 | 49 | |
|
50 | 50 | def loadsubweb(self): |
|
51 | 51 | """The result is only used in the templater method which is not used.""" |
|
52 | 52 | return None |
|
53 | 53 | |
|
54 | 54 | def run(self): |
|
55 | 55 | """Unused function so raise an exception if accidentally called.""" |
|
56 | 56 | raise NotImplementedError |
|
57 | 57 | |
|
58 | 58 | def templater(self, req): |
|
59 | 59 | """Function used in an unreachable code path. |
|
60 | 60 | |
|
61 | 61 | This code is unreachable because we guarantee that the HTTP request, |
|
62 | 62 | corresponds to a Mercurial command. See the is_hg method. So, we are |
|
63 | 63 | never going to get a user-visible url. |
|
64 | 64 | """ |
|
65 | 65 | raise NotImplementedError |
|
66 | 66 | |
|
67 | 67 | def archivelist(self, nodeid): |
|
68 | 68 | """Unused function so raise an exception if accidentally called.""" |
|
69 | 69 | raise NotImplementedError |
|
70 | 70 | |
|
71 | 71 | def __call__(self, environ, start_response): |
|
72 | 72 | """Run the WSGI application. |
|
73 | 73 | |
|
74 | 74 | This may be called by multiple threads. |
|
75 | 75 | """ |
|
76 | 76 | from mercurial.hgweb import request as requestmod |
|
77 | 77 | req = requestmod.parserequestfromenv(environ) |
|
78 | 78 | res = requestmod.wsgiresponse(req, start_response) |
|
79 | 79 | gen = self.run_wsgi(req, res) |
|
80 | 80 | |
|
81 | 81 | first_chunk = None |
|
82 | 82 | |
|
83 | 83 | try: |
|
84 | 84 | data = next(gen) |
|
85 | 85 | |
|
86 | 86 | def first_chunk(): |
|
87 | 87 | yield data |
|
88 | 88 | except StopIteration: |
|
89 | 89 | pass |
|
90 | 90 | |
|
91 | 91 | if first_chunk: |
|
92 | 92 | return itertools.chain(first_chunk(), gen) |
|
93 | 93 | return gen |
|
94 | 94 | |
|
95 | 95 | def _runwsgi(self, req, res, repo): |
|
96 | 96 | |
|
97 | 97 | cmd = req.qsparams.get(b'cmd', '') |
|
98 | 98 | if not mercurial.wireprotoserver.iscmd(cmd): |
|
99 | 99 | # NOTE(marcink): for unsupported commands, we return bad request |
|
100 | 100 | # internally from HG |
|
101 | 101 | log.warning('cmd: `%s` is not supported by the mercurial wireprotocol v1', cmd) |
|
102 | 102 | from mercurial.hgweb.common import statusmessage |
|
103 | 103 | res.status = statusmessage(mercurial.hgweb.common.HTTP_BAD_REQUEST) |
|
104 | 104 | res.setbodybytes(b'') |
|
105 | 105 | return res.sendresponse() |
|
106 | 106 | |
|
107 | 107 | return super()._runwsgi(req, res, repo) |
|
108 | 108 | |
|
109 | 109 | |
|
110 | 110 | def sanitize_hg_ui(baseui): |
|
111 | 111 | # NOTE(marcink): since python3 hgsubversion is deprecated. |
|
112 | 112 | # From old installations we might still have this set enabled |
|
113 | 113 | # we explicitly remove this now here to make sure it wont propagate further |
|
114 | 114 | |
|
115 | 115 | if baseui.config(b'extensions', b'hgsubversion') is not None: |
|
116 | 116 | for cfg in (baseui._ocfg, baseui._tcfg, baseui._ucfg): |
|
117 | 117 | if b'extensions' in cfg: |
|
118 | 118 | if b'hgsubversion' in cfg[b'extensions']: |
|
119 | 119 | del cfg[b'extensions'][b'hgsubversion'] |
|
120 | 120 | |
|
121 | 121 | |
|
122 | 122 | def make_hg_ui_from_config(repo_config): |
|
123 | 123 | baseui = mercurial.ui.ui() |
|
124 | 124 | |
|
125 | 125 | # clean the baseui object |
|
126 | 126 | baseui._ocfg = mercurial.config.config() |
|
127 | 127 | baseui._ucfg = mercurial.config.config() |
|
128 | 128 | baseui._tcfg = mercurial.config.config() |
|
129 | 129 | |
|
130 | 130 | for section, option, value in repo_config: |
|
131 | 131 | baseui.setconfig( |
|
132 | 132 | ascii_bytes(section, allow_bytes=True), |
|
133 | 133 | ascii_bytes(option, allow_bytes=True), |
|
134 | 134 | ascii_bytes(value, allow_bytes=True)) |
|
135 | 135 | |
|
136 | 136 | # make our hgweb quiet so it doesn't print output |
|
137 | 137 | baseui.setconfig(b'ui', b'quiet', b'true') |
|
138 | 138 | |
|
139 | # use POST requests with args instead of GET with headers - fixes issues with big repos with lots of branches | |
|
140 | baseui.setconfig(b'experimental', b'httppostargs', b'true') | |
|
141 | ||
|
139 | 142 | return baseui |
|
140 | 143 | |
|
141 | 144 | |
|
142 | 145 | def update_hg_ui_from_hgrc(baseui, repo_path): |
|
143 | 146 | path = os.path.join(repo_path, '.hg', 'hgrc') |
|
144 | 147 | |
|
145 | 148 | if not os.path.isfile(path): |
|
146 | 149 | log.debug('hgrc file is not present at %s, skipping...', path) |
|
147 | 150 | return |
|
148 | 151 | log.debug('reading hgrc from %s', path) |
|
149 | 152 | cfg = mercurial.config.config() |
|
150 | 153 | cfg.read(ascii_bytes(path)) |
|
151 | 154 | for section in HG_UI_SECTIONS: |
|
152 | 155 | for k, v in cfg.items(section): |
|
153 | 156 | log.debug('settings ui from file: [%s] %s=%s', section, k, v) |
|
154 | 157 | baseui.setconfig( |
|
155 | 158 | ascii_bytes(section, allow_bytes=True), |
|
156 | 159 | ascii_bytes(k, allow_bytes=True), |
|
157 | 160 | ascii_bytes(v, allow_bytes=True)) |
|
158 | 161 | |
|
159 | 162 | |
|
160 | 163 | def create_hg_wsgi_app(repo_path, repo_name, config): |
|
161 | 164 | """ |
|
162 | 165 | Prepares a WSGI application to handle Mercurial requests. |
|
163 | 166 | |
|
164 | 167 | :param config: is a list of 3-item tuples representing a ConfigObject |
|
165 | 168 | (it is the serialized version of the config object). |
|
166 | 169 | """ |
|
167 | 170 | log.debug("Creating Mercurial WSGI application") |
|
168 | 171 | |
|
169 | 172 | baseui = make_hg_ui_from_config(config) |
|
170 | 173 | update_hg_ui_from_hgrc(baseui, repo_path) |
|
171 | 174 | sanitize_hg_ui(baseui) |
|
172 | 175 | |
|
173 | 176 | try: |
|
174 | 177 | return HgWeb(safe_bytes(repo_path), name=safe_bytes(repo_name), baseui=baseui) |
|
175 | 178 | except mercurial.error.RequirementError as e: |
|
176 | 179 | raise exceptions.RequirementException(e)(e) |
|
177 | 180 | |
|
178 | 181 | |
|
179 | 182 | class GitHandler: |
|
180 | 183 | """ |
|
181 | 184 | Handler for Git operations like push/pull etc |
|
182 | 185 | """ |
|
183 | 186 | def __init__(self, repo_location, repo_name, git_path, update_server_info, |
|
184 | 187 | extras): |
|
185 | 188 | if not os.path.isdir(repo_location): |
|
186 | 189 | raise OSError(repo_location) |
|
187 | 190 | self.content_path = repo_location |
|
188 | 191 | self.repo_name = repo_name |
|
189 | 192 | self.repo_location = repo_location |
|
190 | 193 | self.extras = extras |
|
191 | 194 | self.git_path = git_path |
|
192 | 195 | self.update_server_info = update_server_info |
|
193 | 196 | |
|
194 | 197 | def __call__(self, environ, start_response): |
|
195 | 198 | app = webob.exc.HTTPNotFound() |
|
196 | 199 | candidate_paths = ( |
|
197 | 200 | self.content_path, os.path.join(self.content_path, '.git')) |
|
198 | 201 | |
|
199 | 202 | for content_path in candidate_paths: |
|
200 | 203 | try: |
|
201 | 204 | app = pygrack.GitRepository( |
|
202 | 205 | self.repo_name, content_path, self.git_path, |
|
203 | 206 | self.update_server_info, self.extras) |
|
204 | 207 | break |
|
205 | 208 | except OSError: |
|
206 | 209 | continue |
|
207 | 210 | |
|
208 | 211 | return app(environ, start_response) |
|
209 | 212 | |
|
210 | 213 | |
|
211 | 214 | def create_git_wsgi_app(repo_path, repo_name, config): |
|
212 | 215 | """ |
|
213 | 216 | Creates a WSGI application to handle Git requests. |
|
214 | 217 | |
|
215 | 218 | :param config: is a dictionary holding the extras. |
|
216 | 219 | """ |
|
217 | 220 | git_path = settings.GIT_EXECUTABLE() |
|
218 | 221 | update_server_info = config.pop('git_update_server_info') |
|
219 | 222 | app = GitHandler( |
|
220 | 223 | repo_path, repo_name, git_path, update_server_info, config) |
|
221 | 224 | |
|
222 | 225 | return app |
|
223 | 226 | |
|
224 | 227 | |
|
225 | 228 | class GitLFSHandler: |
|
226 | 229 | """ |
|
227 | 230 | Handler for Git LFS operations |
|
228 | 231 | """ |
|
229 | 232 | |
|
230 | 233 | def __init__(self, repo_location, repo_name, git_path, update_server_info, |
|
231 | 234 | extras): |
|
232 | 235 | if not os.path.isdir(repo_location): |
|
233 | 236 | raise OSError(repo_location) |
|
234 | 237 | self.content_path = repo_location |
|
235 | 238 | self.repo_name = repo_name |
|
236 | 239 | self.repo_location = repo_location |
|
237 | 240 | self.extras = extras |
|
238 | 241 | self.git_path = git_path |
|
239 | 242 | self.update_server_info = update_server_info |
|
240 | 243 | |
|
241 | 244 | def get_app(self, git_lfs_enabled, git_lfs_store_path, git_lfs_http_scheme): |
|
242 | 245 | app = git_lfs.create_app(git_lfs_enabled, git_lfs_store_path, git_lfs_http_scheme) |
|
243 | 246 | return app |
|
244 | 247 | |
|
245 | 248 | |
|
246 | 249 | def create_git_lfs_wsgi_app(repo_path, repo_name, config): |
|
247 | 250 | git_path = settings.GIT_EXECUTABLE() |
|
248 | 251 | update_server_info = config.pop('git_update_server_info') |
|
249 | 252 | git_lfs_enabled = config.pop('git_lfs_enabled') |
|
250 | 253 | git_lfs_store_path = config.pop('git_lfs_store_path') |
|
251 | 254 | git_lfs_http_scheme = config.pop('git_lfs_http_scheme', 'http') |
|
252 | 255 | app = GitLFSHandler( |
|
253 | 256 | repo_path, repo_name, git_path, update_server_info, config) |
|
254 | 257 | |
|
255 | 258 | return app.get_app(git_lfs_enabled, git_lfs_store_path, git_lfs_http_scheme) |
@@ -1,563 +1,563 b'' | |||
|
1 | 1 | """ |
|
2 | 2 | Module provides a class allowing to wrap communication over subprocess.Popen |
|
3 | 3 | input, output, error streams into a meaningfull, non-blocking, concurrent |
|
4 | 4 | stream processor exposing the output data as an iterator fitting to be a |
|
5 | 5 | return value passed by a WSGI applicaiton to a WSGI server per PEP 3333. |
|
6 | 6 | |
|
7 | 7 | Copyright (c) 2011 Daniel Dotsenko <dotsa[at]hotmail.com> |
|
8 | 8 | |
|
9 | 9 | This file is part of git_http_backend.py Project. |
|
10 | 10 | |
|
11 | 11 | git_http_backend.py Project is free software: you can redistribute it and/or |
|
12 | 12 | modify it under the terms of the GNU Lesser General Public License as |
|
13 | 13 | published by the Free Software Foundation, either version 2.1 of the License, |
|
14 | 14 | or (at your option) any later version. |
|
15 | 15 | |
|
16 | 16 | git_http_backend.py Project is distributed in the hope that it will be useful, |
|
17 | 17 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
18 | 18 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
19 | 19 | GNU Lesser General Public License for more details. |
|
20 | 20 | |
|
21 | 21 | You should have received a copy of the GNU Lesser General Public License |
|
22 | 22 | along with git_http_backend.py Project. |
|
23 | 23 | If not, see <http://www.gnu.org/licenses/>. |
|
24 | 24 | """ |
|
25 | 25 | import os |
|
26 | 26 | import collections |
|
27 | 27 | import logging |
|
28 | 28 | import subprocess |
|
29 | 29 | import threading |
|
30 | 30 | |
|
31 | from vcsserver.str_utils import safe_str | |
|
31 | from vcsserver.lib.str_utils import safe_str | |
|
32 | 32 | |
|
33 | 33 | log = logging.getLogger(__name__) |
|
34 | 34 | |
|
35 | 35 | |
|
36 | 36 | class StreamFeeder(threading.Thread): |
|
37 | 37 | """ |
|
38 | 38 | Normal writing into pipe-like is blocking once the buffer is filled. |
|
39 | 39 | This thread allows a thread to seep data from a file-like into a pipe |
|
40 | 40 | without blocking the main thread. |
|
41 | 41 | We close inpipe once the end of the source stream is reached. |
|
42 | 42 | """ |
|
43 | 43 | |
|
44 | 44 | def __init__(self, source): |
|
45 | 45 | super().__init__() |
|
46 | 46 | self.daemon = True |
|
47 | 47 | filelike = False |
|
48 | 48 | self.bytes = b'' |
|
49 | 49 | if type(source) in (str, bytes, bytearray): # string-like |
|
50 | 50 | self.bytes = bytes(source) |
|
51 | 51 | else: # can be either file pointer or file-like |
|
52 | 52 | if isinstance(source, int): # file pointer it is |
|
53 | 53 | # converting file descriptor (int) stdin into file-like |
|
54 | 54 | source = os.fdopen(source, 'rb', 16384) |
|
55 | 55 | # let's see if source is file-like by now |
|
56 | 56 | filelike = hasattr(source, 'read') |
|
57 | 57 | if not filelike and not self.bytes: |
|
58 | 58 | raise TypeError("StreamFeeder's source object must be a readable " |
|
59 | 59 | "file-like, a file descriptor, or a string-like.") |
|
60 | 60 | self.source = source |
|
61 | 61 | self.readiface, self.writeiface = os.pipe() |
|
62 | 62 | |
|
63 | 63 | def run(self): |
|
64 | 64 | writer = self.writeiface |
|
65 | 65 | try: |
|
66 | 66 | if self.bytes: |
|
67 | 67 | os.write(writer, self.bytes) |
|
68 | 68 | else: |
|
69 | 69 | s = self.source |
|
70 | 70 | |
|
71 | 71 | while 1: |
|
72 | 72 | _bytes = s.read(4096) |
|
73 | 73 | if not _bytes: |
|
74 | 74 | break |
|
75 | 75 | os.write(writer, _bytes) |
|
76 | 76 | |
|
77 | 77 | finally: |
|
78 | 78 | os.close(writer) |
|
79 | 79 | |
|
80 | 80 | @property |
|
81 | 81 | def output(self): |
|
82 | 82 | return self.readiface |
|
83 | 83 | |
|
84 | 84 | |
|
85 | 85 | class InputStreamChunker(threading.Thread): |
|
86 | 86 | def __init__(self, source, target, buffer_size, chunk_size): |
|
87 | 87 | |
|
88 | 88 | super().__init__() |
|
89 | 89 | |
|
90 | 90 | self.daemon = True # die die die. |
|
91 | 91 | |
|
92 | 92 | self.source = source |
|
93 | 93 | self.target = target |
|
94 | 94 | self.chunk_count_max = int(buffer_size / chunk_size) + 1 |
|
95 | 95 | self.chunk_size = chunk_size |
|
96 | 96 | |
|
97 | 97 | self.data_added = threading.Event() |
|
98 | 98 | self.data_added.clear() |
|
99 | 99 | |
|
100 | 100 | self.keep_reading = threading.Event() |
|
101 | 101 | self.keep_reading.set() |
|
102 | 102 | |
|
103 | 103 | self.EOF = threading.Event() |
|
104 | 104 | self.EOF.clear() |
|
105 | 105 | |
|
106 | 106 | self.go = threading.Event() |
|
107 | 107 | self.go.set() |
|
108 | 108 | |
|
109 | 109 | def stop(self): |
|
110 | 110 | self.go.clear() |
|
111 | 111 | self.EOF.set() |
|
112 | 112 | try: |
|
113 | 113 | # this is not proper, but is done to force the reader thread let |
|
114 | 114 | # go of the input because, if successful, .close() will send EOF |
|
115 | 115 | # down the pipe. |
|
116 | 116 | self.source.close() |
|
117 | 117 | except Exception: |
|
118 | 118 | pass |
|
119 | 119 | |
|
120 | 120 | def run(self): |
|
121 | 121 | s = self.source |
|
122 | 122 | t = self.target |
|
123 | 123 | cs = self.chunk_size |
|
124 | 124 | chunk_count_max = self.chunk_count_max |
|
125 | 125 | keep_reading = self.keep_reading |
|
126 | 126 | da = self.data_added |
|
127 | 127 | go = self.go |
|
128 | 128 | |
|
129 | 129 | try: |
|
130 | 130 | b = s.read(cs) |
|
131 | 131 | except ValueError: |
|
132 | 132 | b = '' |
|
133 | 133 | |
|
134 | 134 | timeout_input = 20 |
|
135 | 135 | while b and go.is_set(): |
|
136 | 136 | if len(t) > chunk_count_max: |
|
137 | 137 | keep_reading.clear() |
|
138 | 138 | keep_reading.wait(timeout_input) |
|
139 | 139 | if len(t) > chunk_count_max + timeout_input: |
|
140 | 140 | log.error("Timed out while waiting for input from subprocess.") |
|
141 | 141 | os._exit(-1) # this will cause the worker to recycle itself |
|
142 | 142 | |
|
143 | 143 | t.append(b) |
|
144 | 144 | da.set() |
|
145 | 145 | |
|
146 | 146 | try: |
|
147 | 147 | b = s.read(cs) |
|
148 | 148 | except ValueError: # probably "I/O operation on closed file" |
|
149 | 149 | b = '' |
|
150 | 150 | |
|
151 | 151 | self.EOF.set() |
|
152 | 152 | da.set() # for cases when done but there was no input. |
|
153 | 153 | |
|
154 | 154 | |
|
155 | 155 | class BufferedGenerator: |
|
156 | 156 | """ |
|
157 | 157 | Class behaves as a non-blocking, buffered pipe reader. |
|
158 | 158 | Reads chunks of data (through a thread) |
|
159 | 159 | from a blocking pipe, and attaches these to an array (Deque) of chunks. |
|
160 | 160 | Reading is halted in the thread when max chunks is internally buffered. |
|
161 | 161 | The .next() may operate in blocking or non-blocking fashion by yielding |
|
162 | 162 | '' if no data is ready |
|
163 | 163 | to be sent or by not returning until there is some data to send |
|
164 | 164 | When we get EOF from underlying source pipe we raise the marker to raise |
|
165 | 165 | StopIteration after the last chunk of data is yielded. |
|
166 | 166 | """ |
|
167 | 167 | |
|
168 | 168 | def __init__(self, name, source, buffer_size=65536, chunk_size=4096, |
|
169 | 169 | starting_values=None, bottomless=False): |
|
170 | 170 | starting_values = starting_values or [] |
|
171 | 171 | self.name = name |
|
172 | 172 | self.buffer_size = buffer_size |
|
173 | 173 | self.chunk_size = chunk_size |
|
174 | 174 | |
|
175 | 175 | if bottomless: |
|
176 | 176 | maxlen = int(buffer_size / chunk_size) |
|
177 | 177 | else: |
|
178 | 178 | maxlen = None |
|
179 | 179 | |
|
180 | 180 | self.data_queue = collections.deque(starting_values, maxlen) |
|
181 | 181 | self.worker = InputStreamChunker(source, self.data_queue, buffer_size, chunk_size) |
|
182 | 182 | if starting_values: |
|
183 | 183 | self.worker.data_added.set() |
|
184 | 184 | self.worker.start() |
|
185 | 185 | |
|
186 | 186 | #################### |
|
187 | 187 | # Generator's methods |
|
188 | 188 | #################### |
|
189 | 189 | def __str__(self): |
|
190 | 190 | return f'BufferedGenerator(name={self.name} chunk: {self.chunk_size} on buffer: {self.buffer_size})' |
|
191 | 191 | |
|
192 | 192 | def __iter__(self): |
|
193 | 193 | return self |
|
194 | 194 | |
|
195 | 195 | def __next__(self): |
|
196 | 196 | |
|
197 | 197 | while not self.length and not self.worker.EOF.is_set(): |
|
198 | 198 | self.worker.data_added.clear() |
|
199 | 199 | self.worker.data_added.wait(0.2) |
|
200 | 200 | |
|
201 | 201 | if self.length: |
|
202 | 202 | self.worker.keep_reading.set() |
|
203 | 203 | return bytes(self.data_queue.popleft()) |
|
204 | 204 | elif self.worker.EOF.is_set(): |
|
205 | 205 | raise StopIteration |
|
206 | 206 | |
|
207 | 207 | def throw(self, exc_type, value=None, traceback=None): |
|
208 | 208 | if not self.worker.EOF.is_set(): |
|
209 | 209 | raise exc_type(value) |
|
210 | 210 | |
|
211 | 211 | def start(self): |
|
212 | 212 | self.worker.start() |
|
213 | 213 | |
|
214 | 214 | def stop(self): |
|
215 | 215 | self.worker.stop() |
|
216 | 216 | |
|
217 | 217 | def close(self): |
|
218 | 218 | try: |
|
219 | 219 | self.worker.stop() |
|
220 | 220 | self.throw(GeneratorExit) |
|
221 | 221 | except (GeneratorExit, StopIteration): |
|
222 | 222 | pass |
|
223 | 223 | |
|
224 | 224 | #################### |
|
225 | 225 | # Threaded reader's infrastructure. |
|
226 | 226 | #################### |
|
227 | 227 | @property |
|
228 | 228 | def input(self): |
|
229 | 229 | return self.worker.w |
|
230 | 230 | |
|
231 | 231 | @property |
|
232 | 232 | def data_added_event(self): |
|
233 | 233 | return self.worker.data_added |
|
234 | 234 | |
|
235 | 235 | @property |
|
236 | 236 | def data_added(self): |
|
237 | 237 | return self.worker.data_added.is_set() |
|
238 | 238 | |
|
239 | 239 | @property |
|
240 | 240 | def reading_paused(self): |
|
241 | 241 | return not self.worker.keep_reading.is_set() |
|
242 | 242 | |
|
243 | 243 | @property |
|
244 | 244 | def done_reading_event(self): |
|
245 | 245 | """ |
|
246 | 246 | Done_reding does not mean that the iterator's buffer is empty. |
|
247 | 247 | Iterator might have done reading from underlying source, but the read |
|
248 | 248 | chunks might still be available for serving through .next() method. |
|
249 | 249 | |
|
250 | 250 | :returns: An Event class instance. |
|
251 | 251 | """ |
|
252 | 252 | return self.worker.EOF |
|
253 | 253 | |
|
254 | 254 | @property |
|
255 | 255 | def done_reading(self): |
|
256 | 256 | """ |
|
257 | 257 | Done_reading does not mean that the iterator's buffer is empty. |
|
258 | 258 | Iterator might have done reading from underlying source, but the read |
|
259 | 259 | chunks might still be available for serving through .next() method. |
|
260 | 260 | |
|
261 | 261 | :returns: An Bool value. |
|
262 | 262 | """ |
|
263 | 263 | return self.worker.EOF.is_set() |
|
264 | 264 | |
|
265 | 265 | @property |
|
266 | 266 | def length(self): |
|
267 | 267 | """ |
|
268 | 268 | returns int. |
|
269 | 269 | |
|
270 | 270 | This is the length of the queue of chunks, not the length of |
|
271 | 271 | the combined contents in those chunks. |
|
272 | 272 | |
|
273 | 273 | __len__() cannot be meaningfully implemented because this |
|
274 | 274 | reader is just flying through a bottomless pit content and |
|
275 | 275 | can only know the length of what it already saw. |
|
276 | 276 | |
|
277 | 277 | If __len__() on WSGI server per PEP 3333 returns a value, |
|
278 | 278 | the response's length will be set to that. In order not to |
|
279 | 279 | confuse WSGI PEP3333 servers, we will not implement __len__ |
|
280 | 280 | at all. |
|
281 | 281 | """ |
|
282 | 282 | return len(self.data_queue) |
|
283 | 283 | |
|
284 | 284 | def prepend(self, x): |
|
285 | 285 | self.data_queue.appendleft(x) |
|
286 | 286 | |
|
287 | 287 | def append(self, x): |
|
288 | 288 | self.data_queue.append(x) |
|
289 | 289 | |
|
290 | 290 | def extend(self, o): |
|
291 | 291 | self.data_queue.extend(o) |
|
292 | 292 | |
|
293 | 293 | def __getitem__(self, i): |
|
294 | 294 | return self.data_queue[i] |
|
295 | 295 | |
|
296 | 296 | |
|
297 | 297 | class SubprocessIOChunker: |
|
298 | 298 | """ |
|
299 | 299 | Processor class wrapping handling of subprocess IO. |
|
300 | 300 | |
|
301 | 301 | .. important:: |
|
302 | 302 | |
|
303 | 303 | Watch out for the method `__del__` on this class. If this object |
|
304 | 304 | is deleted, it will kill the subprocess, so avoid to |
|
305 | 305 | return the `output` attribute or usage of it like in the following |
|
306 | 306 | example:: |
|
307 | 307 | |
|
308 | 308 | # `args` expected to run a program that produces a lot of output |
|
309 | 309 | output = ''.join(SubprocessIOChunker( |
|
310 | 310 | args, shell=False, inputstream=inputstream, env=environ).output) |
|
311 | 311 | |
|
312 | 312 | # `output` will not contain all the data, because the __del__ method |
|
313 | 313 | # has already killed the subprocess in this case before all output |
|
314 | 314 | # has been consumed. |
|
315 | 315 | |
|
316 | 316 | |
|
317 | 317 | |
|
318 | 318 | In a way, this is a "communicate()" replacement with a twist. |
|
319 | 319 | |
|
320 | 320 | - We are multithreaded. Writing in and reading out, err are all sep threads. |
|
321 | 321 | - We support concurrent (in and out) stream processing. |
|
322 | 322 | - The output is not a stream. It's a queue of read string (bytes, not str) |
|
323 | 323 | chunks. The object behaves as an iterable. You can "for chunk in obj:" us. |
|
324 | 324 | - We are non-blocking in more respects than communicate() |
|
325 | 325 | (reading from subprocess out pauses when internal buffer is full, but |
|
326 | 326 | does not block the parent calling code. On the flip side, reading from |
|
327 | 327 | slow-yielding subprocess may block the iteration until data shows up. This |
|
328 | 328 | does not block the parallel inpipe reading occurring parallel thread.) |
|
329 | 329 | |
|
330 | 330 | The purpose of the object is to allow us to wrap subprocess interactions into |
|
331 | 331 | an iterable that can be passed to a WSGI server as the application's return |
|
332 | 332 | value. Because of stream-processing-ability, WSGI does not have to read ALL |
|
333 | 333 | of the subprocess's output and buffer it, before handing it to WSGI server for |
|
334 | 334 | HTTP response. Instead, the class initializer reads just a bit of the stream |
|
335 | 335 | to figure out if error occurred or likely to occur and if not, just hands the |
|
336 | 336 | further iteration over subprocess output to the server for completion of HTTP |
|
337 | 337 | response. |
|
338 | 338 | |
|
339 | 339 | The real or perceived subprocess error is trapped and raised as one of |
|
340 | 340 | OSError family of exceptions |
|
341 | 341 | |
|
342 | 342 | Example usage: |
|
343 | 343 | # try: |
|
344 | 344 | # answer = SubprocessIOChunker( |
|
345 | 345 | # cmd, |
|
346 | 346 | # input, |
|
347 | 347 | # buffer_size = 65536, |
|
348 | 348 | # chunk_size = 4096 |
|
349 | 349 | # ) |
|
350 | 350 | # except (OSError) as e: |
|
351 | 351 | # print str(e) |
|
352 | 352 | # raise e |
|
353 | 353 | # |
|
354 | 354 | # return answer |
|
355 | 355 | |
|
356 | 356 | |
|
357 | 357 | """ |
|
358 | 358 | |
|
359 | 359 | # TODO: johbo: This is used to make sure that the open end of the PIPE |
|
360 | 360 | # is closed in the end. It would be way better to wrap this into an |
|
361 | 361 | # object, so that it is closed automatically once it is consumed or |
|
362 | 362 | # something similar. |
|
363 | 363 | _close_input_fd = None |
|
364 | 364 | |
|
365 | 365 | _closed = False |
|
366 | 366 | _stdout = None |
|
367 | 367 | _stderr = None |
|
368 | 368 | |
|
369 | 369 | def __init__(self, cmd, input_stream=None, buffer_size=65536, |
|
370 | 370 | chunk_size=4096, starting_values=None, fail_on_stderr=True, |
|
371 | 371 | fail_on_return_code=True, **kwargs): |
|
372 | 372 | """ |
|
373 | 373 | Initializes SubprocessIOChunker |
|
374 | 374 | |
|
375 | 375 | :param cmd: A Subprocess.Popen style "cmd". Can be string or array of strings |
|
376 | 376 | :param input_stream: (Default: None) A file-like, string, or file pointer. |
|
377 | 377 | :param buffer_size: (Default: 65536) A size of total buffer per stream in bytes. |
|
378 | 378 | :param chunk_size: (Default: 4096) A max size of a chunk. Actual chunk may be smaller. |
|
379 | 379 | :param starting_values: (Default: []) An array of strings to put in front of output que. |
|
380 | 380 | :param fail_on_stderr: (Default: True) Whether to raise an exception in |
|
381 | 381 | case something is written to stderr. |
|
382 | 382 | :param fail_on_return_code: (Default: True) Whether to raise an |
|
383 | 383 | exception if the return code is not 0. |
|
384 | 384 | """ |
|
385 | 385 | |
|
386 | 386 | kwargs['shell'] = kwargs.get('shell', True) |
|
387 | 387 | |
|
388 | 388 | starting_values = starting_values or [] |
|
389 | 389 | if input_stream: |
|
390 | 390 | input_streamer = StreamFeeder(input_stream) |
|
391 | 391 | input_streamer.start() |
|
392 | 392 | input_stream = input_streamer.output |
|
393 | 393 | self._close_input_fd = input_stream |
|
394 | 394 | |
|
395 | 395 | self._fail_on_stderr = fail_on_stderr |
|
396 | 396 | self._fail_on_return_code = fail_on_return_code |
|
397 | 397 | self.cmd = cmd |
|
398 | 398 | |
|
399 | 399 | _p = subprocess.Popen(cmd, bufsize=-1, stdin=input_stream, stdout=subprocess.PIPE, stderr=subprocess.PIPE, |
|
400 | 400 | **kwargs) |
|
401 | 401 | self.process = _p |
|
402 | 402 | |
|
403 | 403 | bg_out = BufferedGenerator('stdout', _p.stdout, buffer_size, chunk_size, starting_values) |
|
404 | 404 | bg_err = BufferedGenerator('stderr', _p.stderr, 10240, 1, bottomless=True) |
|
405 | 405 | |
|
406 | 406 | while not bg_out.done_reading and not bg_out.reading_paused and not bg_err.length: |
|
407 | 407 | # doing this until we reach either end of file, or end of buffer. |
|
408 | 408 | bg_out.data_added_event.wait(0.2) |
|
409 | 409 | bg_out.data_added_event.clear() |
|
410 | 410 | |
|
411 | 411 | # at this point it's still ambiguous if we are done reading or just full buffer. |
|
412 | 412 | # Either way, if error (returned by ended process, or implied based on |
|
413 | 413 | # presence of stuff in stderr output) we error out. |
|
414 | 414 | # Else, we are happy. |
|
415 | 415 | return_code = _p.poll() |
|
416 | 416 | ret_code_ok = return_code in [None, 0] |
|
417 | 417 | ret_code_fail = return_code is not None and return_code != 0 |
|
418 | 418 | if ( |
|
419 | 419 | (ret_code_fail and fail_on_return_code) or |
|
420 | 420 | (ret_code_ok and fail_on_stderr and bg_err.length) |
|
421 | 421 | ): |
|
422 | 422 | |
|
423 | 423 | try: |
|
424 | 424 | _p.terminate() |
|
425 | 425 | except Exception: |
|
426 | 426 | pass |
|
427 | 427 | |
|
428 | 428 | bg_out.stop() |
|
429 | 429 | out = b''.join(bg_out) |
|
430 | 430 | self._stdout = out |
|
431 | 431 | |
|
432 | 432 | bg_err.stop() |
|
433 | 433 | err = b''.join(bg_err) |
|
434 | 434 | self._stderr = err |
|
435 | 435 | |
|
436 | 436 | # code from https://github.com/schacon/grack/pull/7 |
|
437 | 437 | if err.strip() == b'fatal: The remote end hung up unexpectedly' and out.startswith(b'0034shallow '): |
|
438 | 438 | bg_out = iter([out]) |
|
439 | 439 | _p = None |
|
440 | 440 | elif err and fail_on_stderr: |
|
441 | 441 | text_err = err.decode() |
|
442 | 442 | raise OSError( |
|
443 | 443 | f"Subprocess exited due to an error:\n{text_err}") |
|
444 | 444 | |
|
445 | 445 | if ret_code_fail and fail_on_return_code: |
|
446 | 446 | text_err = err.decode() |
|
447 | 447 | if not err: |
|
448 | 448 | # maybe get empty stderr, try stdout instead |
|
449 | 449 | # in many cases git reports the errors on stdout too |
|
450 | 450 | text_err = out.decode() |
|
451 | 451 | raise OSError( |
|
452 | 452 | f"Subprocess exited with non 0 ret code:{return_code}: stderr:{text_err}") |
|
453 | 453 | |
|
454 | 454 | self.stdout = bg_out |
|
455 | 455 | self.stderr = bg_err |
|
456 | 456 | self.inputstream = input_stream |
|
457 | 457 | |
|
458 | 458 | def __str__(self): |
|
459 | 459 | proc = getattr(self, 'process', 'NO_PROCESS') |
|
460 | 460 | return f'SubprocessIOChunker: {proc}' |
|
461 | 461 | |
|
462 | 462 | def __iter__(self): |
|
463 | 463 | return self |
|
464 | 464 | |
|
465 | 465 | def __next__(self): |
|
466 | 466 | # Note: mikhail: We need to be sure that we are checking the return |
|
467 | 467 | # code after the stdout stream is closed. Some processes, e.g. git |
|
468 | 468 | # are doing some magic in between closing stdout and terminating the |
|
469 | 469 | # process and, as a result, we are not getting return code on "slow" |
|
470 | 470 | # systems. |
|
471 | 471 | result = None |
|
472 | 472 | stop_iteration = None |
|
473 | 473 | try: |
|
474 | 474 | result = next(self.stdout) |
|
475 | 475 | except StopIteration as e: |
|
476 | 476 | stop_iteration = e |
|
477 | 477 | |
|
478 | 478 | if self.process: |
|
479 | 479 | return_code = self.process.poll() |
|
480 | 480 | ret_code_fail = return_code is not None and return_code != 0 |
|
481 | 481 | if ret_code_fail and self._fail_on_return_code: |
|
482 | 482 | self.stop_streams() |
|
483 | 483 | err = self.get_stderr() |
|
484 | 484 | raise OSError( |
|
485 | 485 | f"Subprocess exited (exit_code:{return_code}) due to an error during iteration:\n{err}") |
|
486 | 486 | |
|
487 | 487 | if stop_iteration: |
|
488 | 488 | raise stop_iteration |
|
489 | 489 | return result |
|
490 | 490 | |
|
491 | 491 | def throw(self, exc_type, value=None, traceback=None): |
|
492 | 492 | if self.stdout.length or not self.stdout.done_reading: |
|
493 | 493 | raise exc_type(value) |
|
494 | 494 | |
|
495 | 495 | def close(self): |
|
496 | 496 | if self._closed: |
|
497 | 497 | return |
|
498 | 498 | |
|
499 | 499 | try: |
|
500 | 500 | self.process.terminate() |
|
501 | 501 | except Exception: |
|
502 | 502 | pass |
|
503 | 503 | if self._close_input_fd: |
|
504 | 504 | os.close(self._close_input_fd) |
|
505 | 505 | try: |
|
506 | 506 | self.stdout.close() |
|
507 | 507 | except Exception: |
|
508 | 508 | pass |
|
509 | 509 | try: |
|
510 | 510 | self.stderr.close() |
|
511 | 511 | except Exception: |
|
512 | 512 | pass |
|
513 | 513 | try: |
|
514 | 514 | os.close(self.inputstream) |
|
515 | 515 | except Exception: |
|
516 | 516 | pass |
|
517 | 517 | |
|
518 | 518 | self._closed = True |
|
519 | 519 | |
|
520 | 520 | def stop_streams(self): |
|
521 | 521 | getattr(self.stdout, 'stop', lambda: None)() |
|
522 | 522 | getattr(self.stderr, 'stop', lambda: None)() |
|
523 | 523 | |
|
524 | 524 | def get_stdout(self): |
|
525 | 525 | if self._stdout: |
|
526 | 526 | return self._stdout |
|
527 | 527 | else: |
|
528 | 528 | return b''.join(self.stdout) |
|
529 | 529 | |
|
530 | 530 | def get_stderr(self): |
|
531 | 531 | if self._stderr: |
|
532 | 532 | return self._stderr |
|
533 | 533 | else: |
|
534 | 534 | return b''.join(self.stderr) |
|
535 | 535 | |
|
536 | 536 | |
|
537 | 537 | def run_command(arguments, env=None): |
|
538 | 538 | """ |
|
539 | 539 | Run the specified command and return the stdout. |
|
540 | 540 | |
|
541 | 541 | :param arguments: sequence of program arguments (including the program name) |
|
542 | 542 | :type arguments: list[str] |
|
543 | 543 | """ |
|
544 | 544 | |
|
545 | 545 | cmd = arguments |
|
546 | 546 | log.debug('Running subprocessio command %s', cmd) |
|
547 | 547 | proc = None |
|
548 | 548 | try: |
|
549 | 549 | _opts = {'shell': False, 'fail_on_stderr': False} |
|
550 | 550 | if env: |
|
551 | 551 | _opts.update({'env': env}) |
|
552 | 552 | proc = SubprocessIOChunker(cmd, **_opts) |
|
553 | 553 | return b''.join(proc), b''.join(proc.stderr) |
|
554 | 554 | except OSError as err: |
|
555 | 555 | cmd = ' '.join(map(safe_str, cmd)) # human friendly CMD |
|
556 | 556 | tb_err = ("Couldn't run subprocessio command (%s).\n" |
|
557 | 557 | "Original error was:%s\n" % (cmd, err)) |
|
558 | 558 | log.exception(tb_err) |
|
559 | 559 | raise Exception(tb_err) |
|
560 | 560 | finally: |
|
561 | 561 | if proc: |
|
562 | 562 | proc.close() |
|
563 | 563 |
@@ -1,257 +1,257 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import threading |
|
19 | 19 | import msgpack |
|
20 | 20 | |
|
21 | 21 | from http.server import BaseHTTPRequestHandler |
|
22 | 22 | from socketserver import TCPServer |
|
23 | 23 | |
|
24 | 24 | import mercurial.ui |
|
25 | 25 | import mock |
|
26 | 26 | import pytest |
|
27 | 27 | |
|
28 | 28 | from vcsserver.hooks import HooksHttpClient |
|
29 |
from vcsserver.lib. |
|
|
29 | from vcsserver.lib.ext_json import json | |
|
30 | 30 | from vcsserver import hooks |
|
31 | 31 | |
|
32 | 32 | |
|
33 | 33 | def get_hg_ui(extras=None): |
|
34 | 34 | """Create a Config object with a valid RC_SCM_DATA entry.""" |
|
35 | 35 | extras = extras or {} |
|
36 | 36 | required_extras = { |
|
37 | 37 | 'username': '', |
|
38 | 38 | 'repository': '', |
|
39 | 39 | 'locked_by': '', |
|
40 | 40 | 'scm': '', |
|
41 | 41 | 'make_lock': '', |
|
42 | 42 | 'action': '', |
|
43 | 43 | 'ip': '', |
|
44 | 44 | 'hooks_uri': 'fake_hooks_uri', |
|
45 | 45 | } |
|
46 | 46 | required_extras.update(extras) |
|
47 | 47 | hg_ui = mercurial.ui.ui() |
|
48 | 48 | hg_ui.setconfig(b'rhodecode', b'RC_SCM_DATA', json.dumps(required_extras)) |
|
49 | 49 | |
|
50 | 50 | return hg_ui |
|
51 | 51 | |
|
52 | 52 | |
|
53 | 53 | def test_git_pre_receive_is_disabled(): |
|
54 | 54 | extras = {'hooks': ['pull']} |
|
55 | 55 | response = hooks.git_pre_receive(None, None, |
|
56 | 56 | {'RC_SCM_DATA': json.dumps(extras)}) |
|
57 | 57 | |
|
58 | 58 | assert response == 0 |
|
59 | 59 | |
|
60 | 60 | |
|
61 | 61 | def test_git_post_receive_is_disabled(): |
|
62 | 62 | extras = {'hooks': ['pull']} |
|
63 | 63 | response = hooks.git_post_receive(None, '', |
|
64 | 64 | {'RC_SCM_DATA': json.dumps(extras)}) |
|
65 | 65 | |
|
66 | 66 | assert response == 0 |
|
67 | 67 | |
|
68 | 68 | |
|
69 | 69 | def test_git_post_receive_calls_repo_size(): |
|
70 | 70 | extras = {'hooks': ['push', 'repo_size']} |
|
71 | 71 | |
|
72 | 72 | with mock.patch.object(hooks, '_call_hook') as call_hook_mock: |
|
73 | 73 | hooks.git_post_receive( |
|
74 | 74 | None, '', {'RC_SCM_DATA': json.dumps(extras)}) |
|
75 | 75 | extras.update({'commit_ids': [], 'hook_type': 'post_receive', |
|
76 | 76 | 'new_refs': {'bookmarks': [], 'branches': [], 'tags': []}}) |
|
77 | 77 | expected_calls = [ |
|
78 | 78 | mock.call('repo_size', extras, mock.ANY), |
|
79 | 79 | mock.call('post_push', extras, mock.ANY), |
|
80 | 80 | ] |
|
81 | 81 | assert call_hook_mock.call_args_list == expected_calls |
|
82 | 82 | |
|
83 | 83 | |
|
84 | 84 | def test_git_post_receive_does_not_call_disabled_repo_size(): |
|
85 | 85 | extras = {'hooks': ['push']} |
|
86 | 86 | |
|
87 | 87 | with mock.patch.object(hooks, '_call_hook') as call_hook_mock: |
|
88 | 88 | hooks.git_post_receive( |
|
89 | 89 | None, '', {'RC_SCM_DATA': json.dumps(extras)}) |
|
90 | 90 | extras.update({'commit_ids': [], 'hook_type': 'post_receive', |
|
91 | 91 | 'new_refs': {'bookmarks': [], 'branches': [], 'tags': []}}) |
|
92 | 92 | expected_calls = [ |
|
93 | 93 | mock.call('post_push', extras, mock.ANY) |
|
94 | 94 | ] |
|
95 | 95 | assert call_hook_mock.call_args_list == expected_calls |
|
96 | 96 | |
|
97 | 97 | |
|
98 | 98 | def test_repo_size_exception_does_not_affect_git_post_receive(): |
|
99 | 99 | extras = {'hooks': ['push', 'repo_size']} |
|
100 | 100 | status = 0 |
|
101 | 101 | |
|
102 | 102 | def side_effect(name, *args, **kwargs): |
|
103 | 103 | if name == 'repo_size': |
|
104 | 104 | raise Exception('Fake exception') |
|
105 | 105 | else: |
|
106 | 106 | return status |
|
107 | 107 | |
|
108 | 108 | with mock.patch.object(hooks, '_call_hook') as call_hook_mock: |
|
109 | 109 | call_hook_mock.side_effect = side_effect |
|
110 | 110 | result = hooks.git_post_receive( |
|
111 | 111 | None, '', {'RC_SCM_DATA': json.dumps(extras)}) |
|
112 | 112 | assert result == status |
|
113 | 113 | |
|
114 | 114 | |
|
115 | 115 | def test_git_pre_pull_is_disabled(): |
|
116 | 116 | assert hooks.git_pre_pull({'hooks': ['push']}) == hooks.HookResponse(0, '') |
|
117 | 117 | |
|
118 | 118 | |
|
119 | 119 | def test_git_post_pull_is_disabled(): |
|
120 | 120 | assert ( |
|
121 | 121 | hooks.git_post_pull({'hooks': ['push']}) == hooks.HookResponse(0, '')) |
|
122 | 122 | |
|
123 | 123 | |
|
124 | 124 | class TestGetHooksClient: |
|
125 | 125 | |
|
126 | 126 | def test_returns_http_client_when_protocol_matches(self): |
|
127 | 127 | hooks_uri = 'localhost:8000' |
|
128 | 128 | result = hooks._get_hooks_client({ |
|
129 | 129 | 'hooks_uri': hooks_uri, |
|
130 | 130 | 'hooks_protocol': 'http' |
|
131 | 131 | }) |
|
132 | 132 | assert isinstance(result, hooks.HooksHttpClient) |
|
133 | 133 | assert result.hooks_uri == hooks_uri |
|
134 | 134 | |
|
135 | 135 | def test_return_celery_client_when_queue_and_backend_provided(self): |
|
136 | 136 | task_queue = 'redis://task_queue:0' |
|
137 | 137 | task_backend = task_queue |
|
138 | 138 | result = hooks._get_hooks_client({ |
|
139 | 139 | 'task_queue': task_queue, |
|
140 | 140 | 'task_backend': task_backend |
|
141 | 141 | }) |
|
142 | 142 | assert isinstance(result, hooks.HooksCeleryClient) |
|
143 | 143 | |
|
144 | 144 | |
|
145 | 145 | class TestHooksHttpClient: |
|
146 | 146 | def test_init_sets_hooks_uri(self): |
|
147 | 147 | uri = 'localhost:3000' |
|
148 | 148 | client = hooks.HooksHttpClient(uri) |
|
149 | 149 | assert client.hooks_uri == uri |
|
150 | 150 | |
|
151 | 151 | def test_serialize_returns_serialized_string(self): |
|
152 | 152 | client = hooks.HooksHttpClient('localhost:3000') |
|
153 | 153 | hook_name = 'test' |
|
154 | 154 | extras = { |
|
155 | 155 | 'first': 1, |
|
156 | 156 | 'second': 'two' |
|
157 | 157 | } |
|
158 | 158 | hooks_proto, result = client._serialize(hook_name, extras) |
|
159 | 159 | expected_result = msgpack.packb({ |
|
160 | 160 | 'method': hook_name, |
|
161 | 161 | 'extras': extras, |
|
162 | 162 | }) |
|
163 | 163 | assert hooks_proto == {'rc-hooks-protocol': 'msgpack.v1', 'Connection': 'keep-alive'} |
|
164 | 164 | assert result == expected_result |
|
165 | 165 | |
|
166 | 166 | def test_call_queries_http_server(self, http_mirror): |
|
167 | 167 | client = hooks.HooksHttpClient(http_mirror.uri) |
|
168 | 168 | hook_name = 'test' |
|
169 | 169 | extras = { |
|
170 | 170 | 'first': 1, |
|
171 | 171 | 'second': 'two' |
|
172 | 172 | } |
|
173 | 173 | result = client(hook_name, extras) |
|
174 | 174 | expected_result = msgpack.unpackb(msgpack.packb({ |
|
175 | 175 | 'method': hook_name, |
|
176 | 176 | 'extras': extras |
|
177 | 177 | }), raw=False) |
|
178 | 178 | assert result == expected_result |
|
179 | 179 | |
|
180 | 180 | |
|
181 | 181 | @pytest.fixture |
|
182 | 182 | def http_mirror(request): |
|
183 | 183 | server = MirrorHttpServer() |
|
184 | 184 | request.addfinalizer(server.stop) |
|
185 | 185 | return server |
|
186 | 186 | |
|
187 | 187 | |
|
188 | 188 | class MirrorHttpHandler(BaseHTTPRequestHandler): |
|
189 | 189 | |
|
190 | 190 | def do_POST(self): |
|
191 | 191 | length = int(self.headers['Content-Length']) |
|
192 | 192 | body = self.rfile.read(length) |
|
193 | 193 | self.send_response(200) |
|
194 | 194 | self.end_headers() |
|
195 | 195 | self.wfile.write(body) |
|
196 | 196 | |
|
197 | 197 | |
|
198 | 198 | class MirrorHttpServer: |
|
199 | 199 | ip_address = '127.0.0.1' |
|
200 | 200 | port = 0 |
|
201 | 201 | |
|
202 | 202 | def __init__(self): |
|
203 | 203 | self._daemon = TCPServer((self.ip_address, 0), MirrorHttpHandler) |
|
204 | 204 | _, self.port = self._daemon.server_address |
|
205 | 205 | self._thread = threading.Thread(target=self._daemon.serve_forever) |
|
206 | 206 | self._thread.daemon = True |
|
207 | 207 | self._thread.start() |
|
208 | 208 | |
|
209 | 209 | def stop(self): |
|
210 | 210 | self._daemon.shutdown() |
|
211 | 211 | self._thread.join() |
|
212 | 212 | self._daemon = None |
|
213 | 213 | self._thread = None |
|
214 | 214 | |
|
215 | 215 | @property |
|
216 | 216 | def uri(self): |
|
217 | 217 | return '{}:{}'.format(self.ip_address, self.port) |
|
218 | 218 | |
|
219 | 219 | |
|
220 | 220 | def test_hooks_http_client_init(): |
|
221 | 221 | hooks_uri = 'http://localhost:8000' |
|
222 | 222 | client = HooksHttpClient(hooks_uri) |
|
223 | 223 | assert client.hooks_uri == hooks_uri |
|
224 | 224 | |
|
225 | 225 | |
|
226 | 226 | def test_hooks_http_client_call(): |
|
227 | 227 | hooks_uri = 'http://localhost:8000' |
|
228 | 228 | |
|
229 | 229 | method = 'test_method' |
|
230 | 230 | extras = {'key': 'value'} |
|
231 | 231 | |
|
232 | 232 | with \ |
|
233 | 233 | mock.patch('http.client.HTTPConnection') as mock_connection,\ |
|
234 | 234 | mock.patch('msgpack.load') as mock_load: |
|
235 | 235 | |
|
236 | 236 | client = HooksHttpClient(hooks_uri) |
|
237 | 237 | |
|
238 | 238 | mock_load.return_value = {'result': 'success'} |
|
239 | 239 | response = mock.MagicMock() |
|
240 | 240 | response.status = 200 |
|
241 | 241 | mock_connection.request.side_effect = None |
|
242 | 242 | mock_connection.getresponse.return_value = response |
|
243 | 243 | |
|
244 | 244 | result = client(method, extras) |
|
245 | 245 | |
|
246 | 246 | mock_connection.assert_called_with(hooks_uri) |
|
247 | 247 | mock_connection.return_value.request.assert_called_once() |
|
248 | 248 | assert result == {'result': 'success'} |
|
249 | 249 | |
|
250 | 250 | |
|
251 | 251 | def test_hooks_http_client_serialize(): |
|
252 | 252 | method = 'test_method' |
|
253 | 253 | extras = {'key': 'value'} |
|
254 | 254 | headers, body = HooksHttpClient._serialize(method, extras) |
|
255 | 255 | |
|
256 | 256 | assert headers == {'rc-hooks-protocol': HooksHttpClient.proto, 'Connection': 'keep-alive'} |
|
257 | 257 | assert msgpack.unpackb(body) == {'method': method, 'extras': extras} |
@@ -1,289 +1,289 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import os |
|
19 | 19 | import sys |
|
20 | 20 | import stat |
|
21 | 21 | import pytest |
|
22 | 22 | import vcsserver |
|
23 | 23 | import tempfile |
|
24 | 24 | from vcsserver import hook_utils |
|
25 | 25 | from vcsserver.hook_utils import set_permissions_if_needed, HOOKS_DIR_MODE, HOOKS_FILE_MODE |
|
26 | 26 | from vcsserver.tests.fixture import no_newline_id_generator |
|
27 | from vcsserver.str_utils import safe_bytes | |
|
27 | from vcsserver.lib.str_utils import safe_bytes | |
|
28 | 28 | from vcsserver.utils import AttributeDict |
|
29 | 29 | |
|
30 | 30 | |
|
31 | 31 | class TestCheckRhodecodeHook: |
|
32 | 32 | |
|
33 | 33 | def test_returns_false_when_hook_file_is_wrong_found(self, tmpdir): |
|
34 | 34 | hook = os.path.join(str(tmpdir), 'fake_hook_file.py') |
|
35 | 35 | with open(hook, 'wb') as f: |
|
36 | 36 | f.write(b'dummy test') |
|
37 | 37 | result = hook_utils.check_rhodecode_hook(hook) |
|
38 | 38 | assert result is False |
|
39 | 39 | |
|
40 | 40 | def test_returns_true_when_no_hook_file_found(self, tmpdir): |
|
41 | 41 | hook = os.path.join(str(tmpdir), 'fake_hook_file_not_existing.py') |
|
42 | 42 | result = hook_utils.check_rhodecode_hook(hook) |
|
43 | 43 | assert result |
|
44 | 44 | |
|
45 | 45 | @pytest.mark.parametrize("file_content, expected_result", [ |
|
46 | 46 | ("RC_HOOK_VER = '3.3.3'\n", True), |
|
47 | 47 | ("RC_HOOK = '3.3.3'\n", False), |
|
48 | 48 | ], ids=no_newline_id_generator) |
|
49 | 49 | def test_signatures(self, file_content, expected_result, tmpdir): |
|
50 | 50 | hook = os.path.join(str(tmpdir), 'fake_hook_file_1.py') |
|
51 | 51 | with open(hook, 'wb') as f: |
|
52 | 52 | f.write(safe_bytes(file_content)) |
|
53 | 53 | |
|
54 | 54 | result = hook_utils.check_rhodecode_hook(hook) |
|
55 | 55 | |
|
56 | 56 | assert result is expected_result |
|
57 | 57 | |
|
58 | 58 | |
|
59 | 59 | class BaseInstallHooks: |
|
60 | 60 | HOOK_FILES = () |
|
61 | 61 | |
|
62 | 62 | def _check_hook_file_dir_mode(self, file_path): |
|
63 | 63 | dir_path = os.path.dirname(file_path) |
|
64 | 64 | assert os.path.exists(dir_path), f'dir {file_path} missing' |
|
65 | 65 | stat_info = os.stat(dir_path) |
|
66 | 66 | |
|
67 | 67 | file_mode = stat.S_IMODE(stat_info.st_mode) |
|
68 | 68 | expected_mode = int(HOOKS_DIR_MODE) |
|
69 | 69 | assert expected_mode == file_mode, f'expected mode: {oct(expected_mode)} got: {oct(file_mode)} for {dir_path}' |
|
70 | 70 | |
|
71 | 71 | def _check_hook_file_mode(self, file_path): |
|
72 | 72 | assert os.path.exists(file_path), f'path {file_path} missing' |
|
73 | 73 | stat_info = os.stat(file_path) |
|
74 | 74 | |
|
75 | 75 | file_mode = stat.S_IMODE(stat_info.st_mode) |
|
76 | 76 | expected_mode = int(HOOKS_FILE_MODE) |
|
77 | 77 | assert expected_mode == file_mode, f'expected mode: {oct(expected_mode)} got: {oct(file_mode)} for {file_path}' |
|
78 | 78 | |
|
79 | 79 | def _check_hook_file_content(self, file_path, executable): |
|
80 | 80 | executable = executable or sys.executable |
|
81 | 81 | with open(file_path, 'rt') as hook_file: |
|
82 | 82 | content = hook_file.read() |
|
83 | 83 | |
|
84 | 84 | expected_env = '#!{}'.format(executable) |
|
85 | 85 | expected_rc_version = "\nRC_HOOK_VER = '{}'\n".format(vcsserver.get_version()) |
|
86 | 86 | assert content.strip().startswith(expected_env) |
|
87 | 87 | assert expected_rc_version in content |
|
88 | 88 | |
|
89 | 89 | def _create_fake_hook(self, file_path, content): |
|
90 | 90 | with open(file_path, 'w') as hook_file: |
|
91 | 91 | hook_file.write(content) |
|
92 | 92 | |
|
93 | 93 | def create_dummy_repo(self, repo_type): |
|
94 | 94 | tmpdir = tempfile.mkdtemp() |
|
95 | 95 | repo = AttributeDict() |
|
96 | 96 | if repo_type == 'git': |
|
97 | 97 | repo.path = os.path.join(tmpdir, 'test_git_hooks_installation_repo') |
|
98 | 98 | os.makedirs(repo.path) |
|
99 | 99 | os.makedirs(os.path.join(repo.path, 'hooks')) |
|
100 | 100 | repo.bare = True |
|
101 | 101 | |
|
102 | 102 | elif repo_type == 'svn': |
|
103 | 103 | repo.path = os.path.join(tmpdir, 'test_svn_hooks_installation_repo') |
|
104 | 104 | os.makedirs(repo.path) |
|
105 | 105 | os.makedirs(os.path.join(repo.path, 'hooks')) |
|
106 | 106 | |
|
107 | 107 | return repo |
|
108 | 108 | |
|
109 | 109 | def check_hooks(self, repo_path, repo_bare=True): |
|
110 | 110 | for file_name in self.HOOK_FILES: |
|
111 | 111 | if repo_bare: |
|
112 | 112 | file_path = os.path.join(repo_path, 'hooks', file_name) |
|
113 | 113 | else: |
|
114 | 114 | file_path = os.path.join(repo_path, '.git', 'hooks', file_name) |
|
115 | 115 | |
|
116 | 116 | self._check_hook_file_dir_mode(file_path) |
|
117 | 117 | self._check_hook_file_mode(file_path) |
|
118 | 118 | self._check_hook_file_content(file_path, sys.executable) |
|
119 | 119 | |
|
120 | 120 | |
|
121 | 121 | class TestInstallGitHooks(BaseInstallHooks): |
|
122 | 122 | HOOK_FILES = ('pre-receive', 'post-receive') |
|
123 | 123 | |
|
124 | 124 | def test_hooks_are_installed(self): |
|
125 | 125 | repo = self.create_dummy_repo('git') |
|
126 | 126 | result = hook_utils.install_git_hooks(repo.path, repo.bare) |
|
127 | 127 | assert result |
|
128 | 128 | self.check_hooks(repo.path, repo.bare) |
|
129 | 129 | |
|
130 | 130 | def test_hooks_are_replaced(self): |
|
131 | 131 | repo = self.create_dummy_repo('git') |
|
132 | 132 | hooks_path = os.path.join(repo.path, 'hooks') |
|
133 | 133 | for file_path in [os.path.join(hooks_path, f) for f in self.HOOK_FILES]: |
|
134 | 134 | self._create_fake_hook( |
|
135 | 135 | file_path, content="RC_HOOK_VER = 'abcde'\n") |
|
136 | 136 | |
|
137 | 137 | result = hook_utils.install_git_hooks(repo.path, repo.bare) |
|
138 | 138 | assert result |
|
139 | 139 | self.check_hooks(repo.path, repo.bare) |
|
140 | 140 | |
|
141 | 141 | def test_non_rc_hooks_are_not_replaced(self): |
|
142 | 142 | repo = self.create_dummy_repo('git') |
|
143 | 143 | hooks_path = os.path.join(repo.path, 'hooks') |
|
144 | 144 | non_rc_content = 'echo "non rc hook"\n' |
|
145 | 145 | for file_path in [os.path.join(hooks_path, f) for f in self.HOOK_FILES]: |
|
146 | 146 | self._create_fake_hook( |
|
147 | 147 | file_path, content=non_rc_content) |
|
148 | 148 | |
|
149 | 149 | result = hook_utils.install_git_hooks(repo.path, repo.bare) |
|
150 | 150 | assert result |
|
151 | 151 | |
|
152 | 152 | for file_path in [os.path.join(hooks_path, f) for f in self.HOOK_FILES]: |
|
153 | 153 | with open(file_path, 'rt') as hook_file: |
|
154 | 154 | content = hook_file.read() |
|
155 | 155 | assert content == non_rc_content |
|
156 | 156 | |
|
157 | 157 | def test_non_rc_hooks_are_replaced_with_force_flag(self): |
|
158 | 158 | repo = self.create_dummy_repo('git') |
|
159 | 159 | hooks_path = os.path.join(repo.path, 'hooks') |
|
160 | 160 | non_rc_content = 'echo "non rc hook"\n' |
|
161 | 161 | for file_path in [os.path.join(hooks_path, f) for f in self.HOOK_FILES]: |
|
162 | 162 | self._create_fake_hook( |
|
163 | 163 | file_path, content=non_rc_content) |
|
164 | 164 | |
|
165 | 165 | result = hook_utils.install_git_hooks( |
|
166 | 166 | repo.path, repo.bare, force_create=True) |
|
167 | 167 | assert result |
|
168 | 168 | self.check_hooks(repo.path, repo.bare) |
|
169 | 169 | |
|
170 | 170 | |
|
171 | 171 | class TestInstallSvnHooks(BaseInstallHooks): |
|
172 | 172 | HOOK_FILES = ('pre-commit', 'post-commit') |
|
173 | 173 | |
|
174 | 174 | def test_hooks_are_installed(self): |
|
175 | 175 | repo = self.create_dummy_repo('svn') |
|
176 | 176 | result = hook_utils.install_svn_hooks(repo.path) |
|
177 | 177 | assert result |
|
178 | 178 | self.check_hooks(repo.path) |
|
179 | 179 | |
|
180 | 180 | def test_hooks_are_replaced(self): |
|
181 | 181 | repo = self.create_dummy_repo('svn') |
|
182 | 182 | hooks_path = os.path.join(repo.path, 'hooks') |
|
183 | 183 | for file_path in [os.path.join(hooks_path, f) for f in self.HOOK_FILES]: |
|
184 | 184 | self._create_fake_hook( |
|
185 | 185 | file_path, content="RC_HOOK_VER = 'abcde'\n") |
|
186 | 186 | |
|
187 | 187 | result = hook_utils.install_svn_hooks(repo.path) |
|
188 | 188 | assert result |
|
189 | 189 | self.check_hooks(repo.path) |
|
190 | 190 | |
|
191 | 191 | def test_non_rc_hooks_are_not_replaced(self): |
|
192 | 192 | repo = self.create_dummy_repo('svn') |
|
193 | 193 | hooks_path = os.path.join(repo.path, 'hooks') |
|
194 | 194 | non_rc_content = 'echo "non rc hook"\n' |
|
195 | 195 | for file_path in [os.path.join(hooks_path, f) for f in self.HOOK_FILES]: |
|
196 | 196 | self._create_fake_hook( |
|
197 | 197 | file_path, content=non_rc_content) |
|
198 | 198 | |
|
199 | 199 | result = hook_utils.install_svn_hooks(repo.path) |
|
200 | 200 | assert result |
|
201 | 201 | |
|
202 | 202 | for file_path in [os.path.join(hooks_path, f) for f in self.HOOK_FILES]: |
|
203 | 203 | with open(file_path, 'rt') as hook_file: |
|
204 | 204 | content = hook_file.read() |
|
205 | 205 | assert content == non_rc_content |
|
206 | 206 | |
|
207 | 207 | def test_non_rc_hooks_are_replaced_with_force_flag(self): |
|
208 | 208 | repo = self.create_dummy_repo('svn') |
|
209 | 209 | hooks_path = os.path.join(repo.path, 'hooks') |
|
210 | 210 | non_rc_content = 'echo "non rc hook"\n' |
|
211 | 211 | for file_path in [os.path.join(hooks_path, f) for f in self.HOOK_FILES]: |
|
212 | 212 | self._create_fake_hook( |
|
213 | 213 | file_path, content=non_rc_content) |
|
214 | 214 | |
|
215 | 215 | result = hook_utils.install_svn_hooks( |
|
216 | 216 | repo.path, force_create=True) |
|
217 | 217 | assert result |
|
218 | 218 | self.check_hooks(repo.path, ) |
|
219 | 219 | |
|
220 | 220 | |
|
221 | 221 | def create_test_file(filename): |
|
222 | 222 | """Utility function to create a test file.""" |
|
223 | 223 | with open(filename, 'w') as f: |
|
224 | 224 | f.write("Test file") |
|
225 | 225 | |
|
226 | 226 | |
|
227 | 227 | def remove_test_file(filename): |
|
228 | 228 | """Utility function to remove a test file.""" |
|
229 | 229 | if os.path.exists(filename): |
|
230 | 230 | os.remove(filename) |
|
231 | 231 | |
|
232 | 232 | |
|
233 | 233 | @pytest.fixture |
|
234 | 234 | def test_file(): |
|
235 | 235 | filename = 'test_file.txt' |
|
236 | 236 | create_test_file(filename) |
|
237 | 237 | yield filename |
|
238 | 238 | remove_test_file(filename) |
|
239 | 239 | |
|
240 | 240 | |
|
241 | 241 | def test_increase_permissions(test_file): |
|
242 | 242 | # Set initial lower permissions |
|
243 | 243 | initial_perms = 0o644 |
|
244 | 244 | os.chmod(test_file, initial_perms) |
|
245 | 245 | |
|
246 | 246 | # Set higher permissions |
|
247 | 247 | new_perms = 0o666 |
|
248 | 248 | set_permissions_if_needed(test_file, new_perms) |
|
249 | 249 | |
|
250 | 250 | # Check if permissions were updated |
|
251 | 251 | assert (os.stat(test_file).st_mode & 0o777) == new_perms |
|
252 | 252 | |
|
253 | 253 | |
|
254 | 254 | def test_no_permission_change_needed(test_file): |
|
255 | 255 | # Set initial permissions |
|
256 | 256 | initial_perms = 0o666 |
|
257 | 257 | os.chmod(test_file, initial_perms) |
|
258 | 258 | |
|
259 | 259 | # Attempt to set the same permissions |
|
260 | 260 | set_permissions_if_needed(test_file, initial_perms) |
|
261 | 261 | |
|
262 | 262 | # Check if permissions were unchanged |
|
263 | 263 | assert (os.stat(test_file).st_mode & 0o777) == initial_perms |
|
264 | 264 | |
|
265 | 265 | |
|
266 | 266 | def test_no_permission_reduction(test_file): |
|
267 | 267 | # Set initial higher permissions |
|
268 | 268 | initial_perms = 0o666 |
|
269 | 269 | os.chmod(test_file, initial_perms) |
|
270 | 270 | |
|
271 | 271 | # Attempt to set lower permissions |
|
272 | 272 | lower_perms = 0o644 |
|
273 | 273 | set_permissions_if_needed(test_file, lower_perms) |
|
274 | 274 | |
|
275 | 275 | # Check if permissions were not reduced |
|
276 | 276 | assert (os.stat(test_file).st_mode & 0o777) == initial_perms |
|
277 | 277 | |
|
278 | 278 | |
|
279 | 279 | def test_no_permission_reduction_when_on_777(test_file): |
|
280 | 280 | # Set initial higher permissions |
|
281 | 281 | initial_perms = 0o777 |
|
282 | 282 | os.chmod(test_file, initial_perms) |
|
283 | 283 | |
|
284 | 284 | # Attempt to set lower permissions |
|
285 | 285 | lower_perms = 0o755 |
|
286 | 286 | set_permissions_if_needed(test_file, lower_perms) |
|
287 | 287 | |
|
288 | 288 | # Check if permissions were not reduced |
|
289 | 289 | assert (os.stat(test_file).st_mode & 0o777) == initial_perms |
@@ -1,295 +1,295 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import io |
|
19 | 19 | import more_itertools |
|
20 | 20 | |
|
21 | 21 | import dulwich.protocol |
|
22 | 22 | import mock |
|
23 | 23 | import pytest |
|
24 | 24 | import webob |
|
25 | 25 | import webtest |
|
26 | 26 | |
|
27 | 27 | from vcsserver import hooks, pygrack |
|
28 | 28 | |
|
29 | from vcsserver.str_utils import ascii_bytes | |
|
29 | from vcsserver.lib.str_utils import ascii_bytes | |
|
30 | 30 | |
|
31 | 31 | |
|
32 | 32 | @pytest.fixture() |
|
33 | 33 | def pygrack_instance(tmpdir): |
|
34 | 34 | """ |
|
35 | 35 | Creates a pygrack app instance. |
|
36 | 36 | |
|
37 | 37 | Right now, it does not much helpful regarding the passed directory. |
|
38 | 38 | It just contains the required folders to pass the signature test. |
|
39 | 39 | """ |
|
40 | 40 | for dir_name in ('config', 'head', 'info', 'objects', 'refs'): |
|
41 | 41 | tmpdir.mkdir(dir_name) |
|
42 | 42 | |
|
43 | 43 | return pygrack.GitRepository('repo_name', str(tmpdir), 'git', False, {}) |
|
44 | 44 | |
|
45 | 45 | |
|
46 | 46 | @pytest.fixture() |
|
47 | 47 | def pygrack_app(pygrack_instance): |
|
48 | 48 | """ |
|
49 | 49 | Creates a pygrack app wrapped in webtest.TestApp. |
|
50 | 50 | """ |
|
51 | 51 | return webtest.TestApp(pygrack_instance) |
|
52 | 52 | |
|
53 | 53 | |
|
54 | 54 | def test_invalid_service_info_refs_returns_403(pygrack_app): |
|
55 | 55 | response = pygrack_app.get('/info/refs?service=git-upload-packs', |
|
56 | 56 | expect_errors=True) |
|
57 | 57 | |
|
58 | 58 | assert response.status_int == 403 |
|
59 | 59 | |
|
60 | 60 | |
|
61 | 61 | def test_invalid_endpoint_returns_403(pygrack_app): |
|
62 | 62 | response = pygrack_app.post('/git-upload-packs', expect_errors=True) |
|
63 | 63 | |
|
64 | 64 | assert response.status_int == 403 |
|
65 | 65 | |
|
66 | 66 | |
|
67 | 67 | @pytest.mark.parametrize('sideband', [ |
|
68 | 68 | 'side-band-64k', |
|
69 | 69 | 'side-band', |
|
70 | 70 | 'side-band no-progress', |
|
71 | 71 | ]) |
|
72 | 72 | def test_pre_pull_hook_fails_with_sideband(pygrack_app, sideband): |
|
73 | 73 | request = ''.join([ |
|
74 | 74 | '0054want 74730d410fcb6603ace96f1dc55ea6196122532d ', |
|
75 | 75 | f'multi_ack {sideband} ofs-delta\n', |
|
76 | 76 | '0000', |
|
77 | 77 | '0009done\n', |
|
78 | 78 | ]) |
|
79 | 79 | with mock.patch('vcsserver.hooks.git_pre_pull', return_value=hooks.HookResponse(1, 'foo')): |
|
80 | 80 | response = pygrack_app.post( |
|
81 | 81 | '/git-upload-pack', params=request, |
|
82 | 82 | content_type='application/x-git-upload-pack') |
|
83 | 83 | |
|
84 | 84 | data = io.BytesIO(response.body) |
|
85 | 85 | proto = dulwich.protocol.Protocol(data.read, None) |
|
86 | 86 | packets = list(proto.read_pkt_seq()) |
|
87 | 87 | |
|
88 | 88 | expected_packets = [ |
|
89 | 89 | b'NAK\n', b'\x02foo', b'\x02Pre pull hook failed: aborting\n', |
|
90 | 90 | b'\x01' + pygrack.GitRepository.EMPTY_PACK, |
|
91 | 91 | ] |
|
92 | 92 | assert packets == expected_packets |
|
93 | 93 | |
|
94 | 94 | |
|
95 | 95 | def test_pre_pull_hook_fails_no_sideband(pygrack_app): |
|
96 | 96 | request = ''.join([ |
|
97 | 97 | '0054want 74730d410fcb6603ace96f1dc55ea6196122532d ' + |
|
98 | 98 | 'multi_ack ofs-delta\n' |
|
99 | 99 | '0000', |
|
100 | 100 | '0009done\n', |
|
101 | 101 | ]) |
|
102 | 102 | with mock.patch('vcsserver.hooks.git_pre_pull', |
|
103 | 103 | return_value=hooks.HookResponse(1, 'foo')): |
|
104 | 104 | response = pygrack_app.post( |
|
105 | 105 | '/git-upload-pack', params=request, |
|
106 | 106 | content_type='application/x-git-upload-pack') |
|
107 | 107 | |
|
108 | 108 | assert response.body == pygrack.GitRepository.EMPTY_PACK |
|
109 | 109 | |
|
110 | 110 | |
|
111 | 111 | def test_pull_has_hook_messages(pygrack_app): |
|
112 | 112 | request = ''.join([ |
|
113 | 113 | '0054want 74730d410fcb6603ace96f1dc55ea6196122532d ' + |
|
114 | 114 | 'multi_ack side-band-64k ofs-delta\n' |
|
115 | 115 | '0000', |
|
116 | 116 | '0009done\n', |
|
117 | 117 | ]) |
|
118 | 118 | |
|
119 | 119 | pre_pull = 'pre_pull_output' |
|
120 | 120 | post_pull = 'post_pull_output' |
|
121 | 121 | |
|
122 | 122 | with mock.patch('vcsserver.hooks.git_pre_pull', |
|
123 | 123 | return_value=hooks.HookResponse(0, pre_pull)): |
|
124 | 124 | with mock.patch('vcsserver.hooks.git_post_pull', |
|
125 | 125 | return_value=hooks.HookResponse(1, post_pull)): |
|
126 | 126 | with mock.patch('vcsserver.subprocessio.SubprocessIOChunker', |
|
127 | 127 | return_value=more_itertools.always_iterable([b'0008NAK\n0009subp\n0000'])): |
|
128 | 128 | response = pygrack_app.post( |
|
129 | 129 | '/git-upload-pack', params=request, |
|
130 | 130 | content_type='application/x-git-upload-pack') |
|
131 | 131 | |
|
132 | 132 | data = io.BytesIO(response.body) |
|
133 | 133 | proto = dulwich.protocol.Protocol(data.read, None) |
|
134 | 134 | packets = list(proto.read_pkt_seq()) |
|
135 | 135 | |
|
136 | 136 | assert packets == [b'NAK\n', |
|
137 | 137 | # pre-pull only outputs if IT FAILS as in != 0 ret code |
|
138 | 138 | #b'\x02pre_pull_output', |
|
139 | 139 | b'subp\n', |
|
140 | 140 | b'\x02post_pull_output'] |
|
141 | 141 | |
|
142 | 142 | |
|
143 | 143 | def test_get_want_capabilities(pygrack_instance): |
|
144 | 144 | data = io.BytesIO( |
|
145 | 145 | b'0054want 74730d410fcb6603ace96f1dc55ea6196122532d ' + |
|
146 | 146 | b'multi_ack side-band-64k ofs-delta\n00000009done\n') |
|
147 | 147 | |
|
148 | 148 | request = webob.Request({ |
|
149 | 149 | 'wsgi.input': data, |
|
150 | 150 | 'REQUEST_METHOD': 'POST', |
|
151 | 151 | 'webob.is_body_seekable': True |
|
152 | 152 | }) |
|
153 | 153 | |
|
154 | 154 | capabilities = pygrack_instance._get_want_capabilities(request) |
|
155 | 155 | |
|
156 | 156 | assert capabilities == frozenset( |
|
157 | 157 | (b'ofs-delta', b'multi_ack', b'side-band-64k')) |
|
158 | 158 | assert data.tell() == 0 |
|
159 | 159 | |
|
160 | 160 | |
|
161 | 161 | @pytest.mark.parametrize('data,capabilities,expected', [ |
|
162 | 162 | ('foo', [], []), |
|
163 | 163 | ('', [pygrack.CAPABILITY_SIDE_BAND_64K], []), |
|
164 | 164 | ('', [pygrack.CAPABILITY_SIDE_BAND], []), |
|
165 | 165 | ('foo', [pygrack.CAPABILITY_SIDE_BAND_64K], [b'0008\x02foo']), |
|
166 | 166 | ('foo', [pygrack.CAPABILITY_SIDE_BAND], [b'0008\x02foo']), |
|
167 | 167 | ('f'*1000, [pygrack.CAPABILITY_SIDE_BAND_64K], [b'03ed\x02' + b'f' * 1000]), |
|
168 | 168 | ('f'*1000, [pygrack.CAPABILITY_SIDE_BAND], [b'03e8\x02' + b'f' * 995, b'000a\x02fffff']), |
|
169 | 169 | ('f'*65520, [pygrack.CAPABILITY_SIDE_BAND_64K], [b'fff0\x02' + b'f' * 65515, b'000a\x02fffff']), |
|
170 | 170 | ('f'*65520, [pygrack.CAPABILITY_SIDE_BAND], [b'03e8\x02' + b'f' * 995] * 65 + [b'0352\x02' + b'f' * 845]), |
|
171 | 171 | ], ids=[ |
|
172 | 172 | 'foo-empty', |
|
173 | 173 | 'empty-64k', 'empty', |
|
174 | 174 | 'foo-64k', 'foo', |
|
175 | 175 | 'f-1000-64k', 'f-1000', |
|
176 | 176 | 'f-65520-64k', 'f-65520']) |
|
177 | 177 | def test_get_messages(pygrack_instance, data, capabilities, expected): |
|
178 | 178 | messages = pygrack_instance._get_messages(data, capabilities) |
|
179 | 179 | |
|
180 | 180 | assert messages == expected |
|
181 | 181 | |
|
182 | 182 | |
|
183 | 183 | @pytest.mark.parametrize('response,capabilities,pre_pull_messages,post_pull_messages', [ |
|
184 | 184 | # Unexpected response |
|
185 | 185 | ([b'unexpected_response[no_initial_header]'], [pygrack.CAPABILITY_SIDE_BAND_64K], 'foo', 'bar'), |
|
186 | 186 | # No sideband |
|
187 | 187 | ([b'no-sideband'], [], 'foo', 'bar'), |
|
188 | 188 | # No messages |
|
189 | 189 | ([b'no-messages'], [pygrack.CAPABILITY_SIDE_BAND_64K], '', ''), |
|
190 | 190 | ]) |
|
191 | 191 | def test_inject_messages_to_response_nothing_to_do( |
|
192 | 192 | pygrack_instance, response, capabilities, pre_pull_messages, post_pull_messages): |
|
193 | 193 | |
|
194 | 194 | new_response = pygrack_instance._build_post_pull_response( |
|
195 | 195 | more_itertools.always_iterable(response), capabilities, pre_pull_messages, post_pull_messages) |
|
196 | 196 | |
|
197 | 197 | assert list(new_response) == response |
|
198 | 198 | |
|
199 | 199 | |
|
200 | 200 | @pytest.mark.parametrize('capabilities', [ |
|
201 | 201 | [pygrack.CAPABILITY_SIDE_BAND], |
|
202 | 202 | [pygrack.CAPABILITY_SIDE_BAND_64K], |
|
203 | 203 | ]) |
|
204 | 204 | def test_inject_messages_to_response_single_element(pygrack_instance, capabilities): |
|
205 | 205 | response = [b'0008NAK\n0009subp\n0000'] |
|
206 | 206 | new_response = pygrack_instance._build_post_pull_response( |
|
207 | 207 | more_itertools.always_iterable(response), capabilities, 'foo', 'bar') |
|
208 | 208 | |
|
209 | 209 | expected_response = b''.join([ |
|
210 | 210 | b'0008NAK\n', |
|
211 | 211 | b'0008\x02foo', |
|
212 | 212 | b'0009subp\n', |
|
213 | 213 | b'0008\x02bar', |
|
214 | 214 | b'0000']) |
|
215 | 215 | |
|
216 | 216 | assert b''.join(new_response) == expected_response |
|
217 | 217 | |
|
218 | 218 | |
|
219 | 219 | @pytest.mark.parametrize('capabilities', [ |
|
220 | 220 | [pygrack.CAPABILITY_SIDE_BAND], |
|
221 | 221 | [pygrack.CAPABILITY_SIDE_BAND_64K], |
|
222 | 222 | ]) |
|
223 | 223 | def test_inject_messages_to_response_multi_element(pygrack_instance, capabilities): |
|
224 | 224 | response = more_itertools.always_iterable([ |
|
225 | 225 | b'0008NAK\n000asubp1\n', b'000asubp2\n', b'000asubp3\n', b'000asubp4\n0000' |
|
226 | 226 | ]) |
|
227 | 227 | new_response = pygrack_instance._build_post_pull_response(response, capabilities, 'foo', 'bar') |
|
228 | 228 | |
|
229 | 229 | expected_response = b''.join([ |
|
230 | 230 | b'0008NAK\n', |
|
231 | 231 | b'0008\x02foo', |
|
232 | 232 | b'000asubp1\n', b'000asubp2\n', b'000asubp3\n', b'000asubp4\n', |
|
233 | 233 | b'0008\x02bar', |
|
234 | 234 | b'0000' |
|
235 | 235 | ]) |
|
236 | 236 | |
|
237 | 237 | assert b''.join(new_response) == expected_response |
|
238 | 238 | |
|
239 | 239 | |
|
240 | 240 | def test_build_failed_pre_pull_response_no_sideband(pygrack_instance): |
|
241 | 241 | response = pygrack_instance._build_failed_pre_pull_response([], 'foo') |
|
242 | 242 | |
|
243 | 243 | assert response == [pygrack.GitRepository.EMPTY_PACK] |
|
244 | 244 | |
|
245 | 245 | |
|
246 | 246 | @pytest.mark.parametrize('capabilities', [ |
|
247 | 247 | [pygrack.CAPABILITY_SIDE_BAND], |
|
248 | 248 | [pygrack.CAPABILITY_SIDE_BAND_64K], |
|
249 | 249 | [pygrack.CAPABILITY_SIDE_BAND_64K, b'no-progress'], |
|
250 | 250 | ]) |
|
251 | 251 | def test_build_failed_pre_pull_response(pygrack_instance, capabilities): |
|
252 | 252 | response = pygrack_instance._build_failed_pre_pull_response(capabilities, 'foo') |
|
253 | 253 | |
|
254 | 254 | expected_response = [ |
|
255 | 255 | b'0008NAK\n', b'0008\x02foo', b'0024\x02Pre pull hook failed: aborting\n', |
|
256 | 256 | b'%04x\x01%s' % (len(pygrack.GitRepository.EMPTY_PACK) + 5, pygrack.GitRepository.EMPTY_PACK), |
|
257 | 257 | pygrack.GitRepository.FLUSH_PACKET, |
|
258 | 258 | ] |
|
259 | 259 | |
|
260 | 260 | assert response == expected_response |
|
261 | 261 | |
|
262 | 262 | |
|
263 | 263 | def test_inject_messages_to_response_generator(pygrack_instance): |
|
264 | 264 | |
|
265 | 265 | def response_generator(): |
|
266 | 266 | response = [ |
|
267 | 267 | # protocol start |
|
268 | 268 | b'0008NAK\n', |
|
269 | 269 | ] |
|
270 | 270 | response += [ascii_bytes(f'000asubp{x}\n') for x in range(1000)] |
|
271 | 271 | response += [ |
|
272 | 272 | # protocol end |
|
273 | 273 | pygrack.GitRepository.FLUSH_PACKET |
|
274 | 274 | ] |
|
275 | 275 | for elem in response: |
|
276 | 276 | yield elem |
|
277 | 277 | |
|
278 | 278 | new_response = pygrack_instance._build_post_pull_response( |
|
279 | 279 | response_generator(), [pygrack.CAPABILITY_SIDE_BAND_64K, b'no-progress'], 'PRE_PULL_MSG\n', 'POST_PULL_MSG\n') |
|
280 | 280 | |
|
281 | 281 | assert iter(new_response) |
|
282 | 282 | |
|
283 | 283 | expected_response = b''.join([ |
|
284 | 284 | # start |
|
285 | 285 | b'0008NAK\n0012\x02PRE_PULL_MSG\n', |
|
286 | 286 | ] + [ |
|
287 | 287 | # ... rest |
|
288 | 288 | ascii_bytes(f'000asubp{x}\n') for x in range(1000) |
|
289 | 289 | ] + [ |
|
290 | 290 | # final message, |
|
291 | 291 | b'0013\x02POST_PULL_MSG\n0000', |
|
292 | 292 | |
|
293 | 293 | ]) |
|
294 | 294 | |
|
295 | 295 | assert b''.join(new_response) == expected_response |
@@ -1,87 +1,87 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import os |
|
19 | 19 | |
|
20 | 20 | import mercurial.hg |
|
21 | 21 | import mercurial.ui |
|
22 | 22 | import mercurial.error |
|
23 | 23 | import mock |
|
24 | 24 | import pytest |
|
25 | 25 | import webtest |
|
26 | 26 | |
|
27 | 27 | from vcsserver import scm_app |
|
28 | from vcsserver.str_utils import ascii_bytes | |
|
28 | from vcsserver.lib.str_utils import ascii_bytes | |
|
29 | 29 | |
|
30 | 30 | |
|
31 | 31 | def test_hg_does_not_accept_invalid_cmd(tmpdir): |
|
32 | 32 | repo = mercurial.hg.repository(mercurial.ui.ui(), ascii_bytes(str(tmpdir)), create=True) |
|
33 | 33 | app = webtest.TestApp(scm_app.HgWeb(repo)) |
|
34 | 34 | |
|
35 | 35 | response = app.get('/repo?cmd=invalidcmd', expect_errors=True) |
|
36 | 36 | |
|
37 | 37 | assert response.status_int == 400 |
|
38 | 38 | |
|
39 | 39 | |
|
40 | 40 | def test_create_hg_wsgi_app_requirement_error(tmpdir): |
|
41 | 41 | repo = mercurial.hg.repository(mercurial.ui.ui(), ascii_bytes(str(tmpdir)), create=True) |
|
42 | 42 | config = ( |
|
43 | 43 | ('paths', 'default', ''), |
|
44 | 44 | ) |
|
45 | 45 | with mock.patch('vcsserver.scm_app.HgWeb') as hgweb_mock: |
|
46 | 46 | hgweb_mock.side_effect = mercurial.error.RequirementError() |
|
47 | 47 | with pytest.raises(Exception): |
|
48 | 48 | scm_app.create_hg_wsgi_app(str(tmpdir), repo, config) |
|
49 | 49 | |
|
50 | 50 | |
|
51 | 51 | def test_git_returns_not_found(tmpdir): |
|
52 | 52 | app = webtest.TestApp( |
|
53 | 53 | scm_app.GitHandler(str(tmpdir), 'repo_name', 'git', False, {})) |
|
54 | 54 | |
|
55 | 55 | response = app.get('/repo_name/inforefs?service=git-upload-pack', |
|
56 | 56 | expect_errors=True) |
|
57 | 57 | |
|
58 | 58 | assert response.status_int == 404 |
|
59 | 59 | |
|
60 | 60 | |
|
61 | 61 | def test_git(tmpdir): |
|
62 | 62 | for dir_name in ('config', 'head', 'info', 'objects', 'refs'): |
|
63 | 63 | tmpdir.mkdir(dir_name) |
|
64 | 64 | |
|
65 | 65 | app = webtest.TestApp( |
|
66 | 66 | scm_app.GitHandler(str(tmpdir), 'repo_name', 'git', False, {})) |
|
67 | 67 | |
|
68 | 68 | # We set service to git-upload-packs to trigger a 403 |
|
69 | 69 | response = app.get('/repo_name/inforefs?service=git-upload-packs', |
|
70 | 70 | expect_errors=True) |
|
71 | 71 | |
|
72 | 72 | assert response.status_int == 403 |
|
73 | 73 | |
|
74 | 74 | |
|
75 | 75 | def test_git_fallbacks_to_git_folder(tmpdir): |
|
76 | 76 | tmpdir.mkdir('.git') |
|
77 | 77 | for dir_name in ('config', 'head', 'info', 'objects', 'refs'): |
|
78 | 78 | tmpdir.mkdir(os.path.join('.git', dir_name)) |
|
79 | 79 | |
|
80 | 80 | app = webtest.TestApp( |
|
81 | 81 | scm_app.GitHandler(str(tmpdir), 'repo_name', 'git', False, {})) |
|
82 | 82 | |
|
83 | 83 | # We set service to git-upload-packs to trigger a 403 |
|
84 | 84 | response = app.get('/repo_name/inforefs?service=git-upload-packs', |
|
85 | 85 | expect_errors=True) |
|
86 | 86 | |
|
87 | 87 | assert response.status_int == 403 |
@@ -1,155 +1,155 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import io |
|
19 | 19 | import os |
|
20 | 20 | import sys |
|
21 | 21 | |
|
22 | 22 | import pytest |
|
23 | 23 | |
|
24 | 24 | from vcsserver import subprocessio |
|
25 | from vcsserver.str_utils import ascii_bytes | |
|
25 | from vcsserver.lib.str_utils import ascii_bytes | |
|
26 | 26 | |
|
27 | 27 | |
|
28 | 28 | class FileLikeObj: # pragma: no cover |
|
29 | 29 | |
|
30 | 30 | def __init__(self, data: bytes, size): |
|
31 | 31 | chunks = size // len(data) |
|
32 | 32 | |
|
33 | 33 | self.stream = self._get_stream(data, chunks) |
|
34 | 34 | |
|
35 | 35 | def _get_stream(self, data, chunks): |
|
36 | 36 | for x in range(chunks): |
|
37 | 37 | yield data |
|
38 | 38 | |
|
39 | 39 | def read(self, n): |
|
40 | 40 | |
|
41 | 41 | buffer_stream = b'' |
|
42 | 42 | for chunk in self.stream: |
|
43 | 43 | buffer_stream += chunk |
|
44 | 44 | if len(buffer_stream) >= n: |
|
45 | 45 | break |
|
46 | 46 | |
|
47 | 47 | # self.stream = self.bytes[n:] |
|
48 | 48 | return buffer_stream |
|
49 | 49 | |
|
50 | 50 | |
|
51 | 51 | @pytest.fixture(scope='module') |
|
52 | 52 | def environ(): |
|
53 | 53 | """Delete coverage variables, as they make the tests fail.""" |
|
54 | 54 | env = dict(os.environ) |
|
55 | 55 | for key in list(env.keys()): |
|
56 | 56 | if key.startswith('COV_CORE_'): |
|
57 | 57 | del env[key] |
|
58 | 58 | |
|
59 | 59 | return env |
|
60 | 60 | |
|
61 | 61 | |
|
62 | 62 | def _get_python_args(script): |
|
63 | 63 | return [sys.executable, '-c', 'import sys; import time; import shutil; ' + script] |
|
64 | 64 | |
|
65 | 65 | |
|
66 | 66 | def test_raise_exception_on_non_zero_return_code(environ): |
|
67 | 67 | call_args = _get_python_args('raise ValueError("fail")') |
|
68 | 68 | with pytest.raises(OSError): |
|
69 | 69 | b''.join(subprocessio.SubprocessIOChunker(call_args, shell=False, env=environ)) |
|
70 | 70 | |
|
71 | 71 | |
|
72 | 72 | def test_does_not_fail_on_non_zero_return_code(environ): |
|
73 | 73 | call_args = _get_python_args('sys.stdout.write("hello"); sys.exit(1)') |
|
74 | 74 | proc = subprocessio.SubprocessIOChunker(call_args, shell=False, fail_on_return_code=False, env=environ) |
|
75 | 75 | output = b''.join(proc) |
|
76 | 76 | |
|
77 | 77 | assert output == b'hello' |
|
78 | 78 | |
|
79 | 79 | |
|
80 | 80 | def test_raise_exception_on_stderr(environ): |
|
81 | 81 | call_args = _get_python_args('sys.stderr.write("WRITE_TO_STDERR"); time.sleep(1);') |
|
82 | 82 | |
|
83 | 83 | with pytest.raises(OSError) as excinfo: |
|
84 | 84 | b''.join(subprocessio.SubprocessIOChunker(call_args, shell=False, env=environ)) |
|
85 | 85 | |
|
86 | 86 | assert 'exited due to an error:\nWRITE_TO_STDERR' in str(excinfo.value) |
|
87 | 87 | |
|
88 | 88 | |
|
89 | 89 | def test_does_not_fail_on_stderr(environ): |
|
90 | 90 | call_args = _get_python_args('sys.stderr.write("WRITE_TO_STDERR"); sys.stderr.flush; time.sleep(2);') |
|
91 | 91 | proc = subprocessio.SubprocessIOChunker(call_args, shell=False, fail_on_stderr=False, env=environ) |
|
92 | 92 | output = b''.join(proc) |
|
93 | 93 | |
|
94 | 94 | assert output == b'' |
|
95 | 95 | |
|
96 | 96 | |
|
97 | 97 | @pytest.mark.parametrize('size', [ |
|
98 | 98 | 1, |
|
99 | 99 | 10 ** 5 |
|
100 | 100 | ]) |
|
101 | 101 | def test_output_with_no_input(size, environ): |
|
102 | 102 | call_args = _get_python_args(f'sys.stdout.write("X" * {size});') |
|
103 | 103 | proc = subprocessio.SubprocessIOChunker(call_args, shell=False, env=environ) |
|
104 | 104 | output = b''.join(proc) |
|
105 | 105 | |
|
106 | 106 | assert output == ascii_bytes("X" * size) |
|
107 | 107 | |
|
108 | 108 | |
|
109 | 109 | @pytest.mark.parametrize('size', [ |
|
110 | 110 | 1, |
|
111 | 111 | 10 ** 5 |
|
112 | 112 | ]) |
|
113 | 113 | def test_output_with_no_input_does_not_fail(size, environ): |
|
114 | 114 | |
|
115 | 115 | call_args = _get_python_args(f'sys.stdout.write("X" * {size}); sys.exit(1)') |
|
116 | 116 | proc = subprocessio.SubprocessIOChunker(call_args, shell=False, fail_on_return_code=False, env=environ) |
|
117 | 117 | output = b''.join(proc) |
|
118 | 118 | |
|
119 | 119 | assert output == ascii_bytes("X" * size) |
|
120 | 120 | |
|
121 | 121 | |
|
122 | 122 | @pytest.mark.parametrize('size', [ |
|
123 | 123 | 1, |
|
124 | 124 | 10 ** 5 |
|
125 | 125 | ]) |
|
126 | 126 | def test_output_with_input(size, environ): |
|
127 | 127 | data_len = size |
|
128 | 128 | inputstream = FileLikeObj(b'X', size) |
|
129 | 129 | |
|
130 | 130 | # This acts like the cat command. |
|
131 | 131 | call_args = _get_python_args('shutil.copyfileobj(sys.stdin, sys.stdout)') |
|
132 | 132 | # note: in this tests we explicitly don't assign chunker to a variable and let it stream directly |
|
133 | 133 | output = b''.join( |
|
134 | 134 | subprocessio.SubprocessIOChunker(call_args, shell=False, input_stream=inputstream, env=environ) |
|
135 | 135 | ) |
|
136 | 136 | |
|
137 | 137 | assert len(output) == data_len |
|
138 | 138 | |
|
139 | 139 | |
|
140 | 140 | @pytest.mark.parametrize('size', [ |
|
141 | 141 | 1, |
|
142 | 142 | 10 ** 5 |
|
143 | 143 | ]) |
|
144 | 144 | def test_output_with_input_skipping_iterator(size, environ): |
|
145 | 145 | data_len = size |
|
146 | 146 | inputstream = FileLikeObj(b'X', size) |
|
147 | 147 | |
|
148 | 148 | # This acts like the cat command. |
|
149 | 149 | call_args = _get_python_args('shutil.copyfileobj(sys.stdin, sys.stdout)') |
|
150 | 150 | |
|
151 | 151 | # Note: assigning the chunker makes sure that it is not deleted too early |
|
152 | 152 | proc = subprocessio.SubprocessIOChunker(call_args, shell=False, input_stream=inputstream, env=environ) |
|
153 | 153 | output = b''.join(proc.stdout) |
|
154 | 154 | |
|
155 | 155 | assert len(output) == data_len |
@@ -1,103 +1,103 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import io |
|
19 | 19 | import mock |
|
20 | 20 | import pytest |
|
21 | 21 | import sys |
|
22 | 22 | |
|
23 | from vcsserver.str_utils import ascii_bytes | |
|
23 | from vcsserver.lib.str_utils import ascii_bytes | |
|
24 | 24 | |
|
25 | 25 | |
|
26 | 26 | class MockPopen: |
|
27 | 27 | def __init__(self, stderr): |
|
28 | 28 | self.stdout = io.BytesIO(b'') |
|
29 | 29 | self.stderr = io.BytesIO(stderr) |
|
30 | 30 | self.returncode = 1 |
|
31 | 31 | |
|
32 | 32 | def wait(self): |
|
33 | 33 | pass |
|
34 | 34 | |
|
35 | 35 | |
|
36 | 36 | INVALID_CERTIFICATE_STDERR = '\n'.join([ |
|
37 | 37 | 'svnrdump: E230001: Unable to connect to a repository at URL url', |
|
38 | 38 | 'svnrdump: E230001: Server SSL certificate verification failed: issuer is not trusted', |
|
39 | 39 | ]) |
|
40 | 40 | |
|
41 | 41 | |
|
42 | 42 | @pytest.mark.parametrize('stderr,expected_reason', [ |
|
43 | 43 | (INVALID_CERTIFICATE_STDERR, 'INVALID_CERTIFICATE'), |
|
44 | 44 | ('svnrdump: E123456', 'UNKNOWN:svnrdump: E123456'), |
|
45 | 45 | ], ids=['invalid-cert-stderr', 'svnrdump-err-123456']) |
|
46 | 46 | @pytest.mark.xfail(sys.platform == "cygwin", |
|
47 | 47 | reason="SVN not packaged for Cygwin") |
|
48 | 48 | def test_import_remote_repository_certificate_error(stderr, expected_reason): |
|
49 | 49 | from vcsserver.remote import svn_remote |
|
50 | 50 | factory = mock.Mock() |
|
51 | 51 | factory.repo = mock.Mock(return_value=mock.Mock()) |
|
52 | 52 | |
|
53 | 53 | remote = svn_remote.SvnRemote(factory) |
|
54 | 54 | remote.is_path_valid_repository = lambda wire, path: True |
|
55 | 55 | |
|
56 | 56 | with mock.patch('subprocess.Popen', |
|
57 | 57 | return_value=MockPopen(ascii_bytes(stderr))): |
|
58 | 58 | with pytest.raises(Exception) as excinfo: |
|
59 | 59 | remote.import_remote_repository({'path': 'path'}, 'url') |
|
60 | 60 | |
|
61 | 61 | expected_error_args = 'Failed to dump the remote repository from url. Reason:{}'.format(expected_reason) |
|
62 | 62 | |
|
63 | 63 | assert excinfo.value.args[0] == expected_error_args |
|
64 | 64 | |
|
65 | 65 | |
|
66 | 66 | def test_svn_libraries_can_be_imported(): |
|
67 | 67 | import svn.client # noqa |
|
68 | 68 | assert svn.client is not None |
|
69 | 69 | |
|
70 | 70 | |
|
71 | 71 | @pytest.mark.parametrize('example_url, parts', [ |
|
72 | 72 | ('http://server.com', ('', '', 'http://server.com')), |
|
73 | 73 | ('http://user@server.com', ('user', '', 'http://user@server.com')), |
|
74 | 74 | ('http://user:pass@server.com', ('user', 'pass', 'http://user:pass@server.com')), |
|
75 | 75 | ('<script>', ('', '', '<script>')), |
|
76 | 76 | ('http://', ('', '', 'http://')), |
|
77 | 77 | ]) |
|
78 | 78 | def test_username_password_extraction_from_url(example_url, parts): |
|
79 | 79 | from vcsserver.remote import svn_remote |
|
80 | 80 | |
|
81 | 81 | factory = mock.Mock() |
|
82 | 82 | factory.repo = mock.Mock(return_value=mock.Mock()) |
|
83 | 83 | |
|
84 | 84 | remote = svn_remote.SvnRemote(factory) |
|
85 | 85 | remote.is_path_valid_repository = lambda wire, path: True |
|
86 | 86 | |
|
87 | 87 | assert remote.get_url_and_credentials(example_url) == parts |
|
88 | 88 | |
|
89 | 89 | |
|
90 | 90 | @pytest.mark.parametrize('call_url', [ |
|
91 | 91 | b'https://svn.code.sf.net/p/svnbook/source/trunk/', |
|
92 | 92 | b'https://marcink@svn.code.sf.net/p/svnbook/source/trunk/', |
|
93 | 93 | b'https://marcink:qweqwe@svn.code.sf.net/p/svnbook/source/trunk/', |
|
94 | 94 | ]) |
|
95 | 95 | def test_check_url(call_url): |
|
96 | 96 | from vcsserver.remote import svn_remote |
|
97 | 97 | factory = mock.Mock() |
|
98 | 98 | factory.repo = mock.Mock(return_value=mock.Mock()) |
|
99 | 99 | |
|
100 | 100 | remote = svn_remote.SvnRemote(factory) |
|
101 | 101 | remote.is_path_valid_repository = lambda wire, path: True |
|
102 | 102 | assert remote.check_url(call_url, {'dummy': 'config'}) |
|
103 | 103 |
@@ -1,69 +1,69 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import pytest |
|
19 | from vcsserver.str_utils import ascii_bytes, ascii_str, convert_to_str | |
|
19 | from vcsserver.lib.str_utils import ascii_bytes, ascii_str, convert_to_str | |
|
20 | 20 | |
|
21 | 21 | |
|
22 | 22 | @pytest.mark.parametrize('given, expected', [ |
|
23 | 23 | ('a', b'a'), |
|
24 | 24 | ('a', b'a'), |
|
25 | 25 | ]) |
|
26 | 26 | def test_ascii_bytes(given, expected): |
|
27 | 27 | assert ascii_bytes(given) == expected |
|
28 | 28 | |
|
29 | 29 | |
|
30 | 30 | @pytest.mark.parametrize('given', [ |
|
31 | 31 | 'Γ₯', |
|
32 | 32 | 'Γ₯'.encode('utf8') |
|
33 | 33 | ]) |
|
34 | 34 | def test_ascii_bytes_raises(given): |
|
35 | 35 | with pytest.raises(ValueError): |
|
36 | 36 | ascii_bytes(given) |
|
37 | 37 | |
|
38 | 38 | |
|
39 | 39 | @pytest.mark.parametrize('given, expected', [ |
|
40 | 40 | (b'a', 'a'), |
|
41 | 41 | ]) |
|
42 | 42 | def test_ascii_str(given, expected): |
|
43 | 43 | assert ascii_str(given) == expected |
|
44 | 44 | |
|
45 | 45 | |
|
46 | 46 | @pytest.mark.parametrize('given', [ |
|
47 | 47 | 'a', |
|
48 | 48 | 'Γ₯'.encode('utf8'), |
|
49 | 49 | 'Γ₯' |
|
50 | 50 | ]) |
|
51 | 51 | def test_ascii_str_raises(given): |
|
52 | 52 | with pytest.raises(ValueError): |
|
53 | 53 | ascii_str(given) |
|
54 | 54 | |
|
55 | 55 | |
|
56 | 56 | @pytest.mark.parametrize('given, expected', [ |
|
57 | 57 | ('a', 'a'), |
|
58 | 58 | (b'a', 'a'), |
|
59 | 59 | # tuple |
|
60 | 60 | (('a', b'b', b'c'), ('a', 'b', 'c')), |
|
61 | 61 | # nested tuple |
|
62 | 62 | (('a', b'b', (b'd', b'e')), ('a', 'b', ('d', 'e'))), |
|
63 | 63 | # list |
|
64 | 64 | (['a', b'b', b'c'], ['a', 'b', 'c']), |
|
65 | 65 | # mixed |
|
66 | 66 | (['a', b'b', b'c', (b'b1', b'b2')], ['a', 'b', 'c', ('b1', 'b2')]) |
|
67 | 67 | ]) |
|
68 | 68 | def test_convert_to_str(given, expected): |
|
69 | 69 | assert convert_to_str(given) == expected |
@@ -1,98 +1,98 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | import wsgiref.simple_server |
|
19 | 19 | import wsgiref.validate |
|
20 | 20 | |
|
21 | 21 | from vcsserver import wsgi_app_caller |
|
22 | from vcsserver.str_utils import ascii_bytes, safe_str | |
|
22 | from vcsserver.lib.str_utils import ascii_bytes, safe_str | |
|
23 | 23 | |
|
24 | 24 | |
|
25 | 25 | @wsgiref.validate.validator |
|
26 | 26 | def demo_app(environ, start_response): |
|
27 | 27 | """WSGI app used for testing.""" |
|
28 | 28 | |
|
29 | 29 | input_data = safe_str(environ['wsgi.input'].read(1024)) |
|
30 | 30 | |
|
31 | 31 | data = [ |
|
32 | 32 | 'Hello World!\n', |
|
33 | 33 | f'input_data={input_data}\n', |
|
34 | 34 | ] |
|
35 | 35 | for key, value in sorted(environ.items()): |
|
36 | 36 | data.append(f'{key}={value}\n') |
|
37 | 37 | |
|
38 | 38 | write = start_response("200 OK", [('Content-Type', 'text/plain')]) |
|
39 | 39 | write(b'Old school write method\n') |
|
40 | 40 | write(b'***********************\n') |
|
41 | 41 | return list(map(ascii_bytes, data)) |
|
42 | 42 | |
|
43 | 43 | |
|
44 | 44 | BASE_ENVIRON = { |
|
45 | 45 | 'REQUEST_METHOD': 'GET', |
|
46 | 46 | 'SERVER_NAME': 'localhost', |
|
47 | 47 | 'SERVER_PORT': '80', |
|
48 | 48 | 'SCRIPT_NAME': '', |
|
49 | 49 | 'PATH_INFO': '/', |
|
50 | 50 | 'QUERY_STRING': '', |
|
51 | 51 | 'foo.var': 'bla', |
|
52 | 52 | } |
|
53 | 53 | |
|
54 | 54 | |
|
55 | 55 | def test_complete_environ(): |
|
56 | 56 | environ = dict(BASE_ENVIRON) |
|
57 | 57 | data = b"data" |
|
58 | 58 | wsgi_app_caller._complete_environ(environ, data) |
|
59 | 59 | wsgiref.validate.check_environ(environ) |
|
60 | 60 | |
|
61 | 61 | assert data == environ['wsgi.input'].read(1024) |
|
62 | 62 | |
|
63 | 63 | |
|
64 | 64 | def test_start_response(): |
|
65 | 65 | start_response = wsgi_app_caller._StartResponse() |
|
66 | 66 | status = '200 OK' |
|
67 | 67 | headers = [('Content-Type', 'text/plain')] |
|
68 | 68 | start_response(status, headers) |
|
69 | 69 | |
|
70 | 70 | assert status == start_response.status |
|
71 | 71 | assert headers == start_response.headers |
|
72 | 72 | |
|
73 | 73 | |
|
74 | 74 | def test_start_response_with_error(): |
|
75 | 75 | start_response = wsgi_app_caller._StartResponse() |
|
76 | 76 | status = '500 Internal Server Error' |
|
77 | 77 | headers = [('Content-Type', 'text/plain')] |
|
78 | 78 | start_response(status, headers, (None, None, None)) |
|
79 | 79 | |
|
80 | 80 | assert status == start_response.status |
|
81 | 81 | assert headers == start_response.headers |
|
82 | 82 | |
|
83 | 83 | |
|
84 | 84 | def test_wsgi_app_caller(): |
|
85 | 85 | environ = dict(BASE_ENVIRON) |
|
86 | 86 | input_data = 'some text' |
|
87 | 87 | |
|
88 | 88 | caller = wsgi_app_caller.WSGIAppCaller(demo_app) |
|
89 | 89 | responses, status, headers = caller.handle(environ, input_data) |
|
90 | 90 | response = b''.join(responses) |
|
91 | 91 | |
|
92 | 92 | assert status == '200 OK' |
|
93 | 93 | assert headers == [('Content-Type', 'text/plain')] |
|
94 | 94 | assert response.startswith(b'Old school write method\n***********************\n') |
|
95 | 95 | assert b'Hello World!\n' in response |
|
96 | 96 | assert b'foo.var=bla\n' in response |
|
97 | 97 | |
|
98 | 98 | assert ascii_bytes(f'input_data={input_data}\n') in response |
@@ -1,123 +1,123 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | import base64 |
|
18 | 18 | import logging |
|
19 | 19 | import time |
|
20 | 20 | |
|
21 | 21 | import msgpack |
|
22 | 22 | |
|
23 | 23 | import vcsserver |
|
24 | from vcsserver.str_utils import safe_str | |
|
24 | from vcsserver.lib.str_utils import safe_str | |
|
25 | 25 | |
|
26 | 26 | log = logging.getLogger(__name__) |
|
27 | 27 | |
|
28 | 28 | |
|
29 | 29 | def get_access_path(environ): |
|
30 | 30 | path = environ.get('PATH_INFO') |
|
31 | 31 | return path |
|
32 | 32 | |
|
33 | 33 | |
|
34 | 34 | def get_user_agent(environ): |
|
35 | 35 | return environ.get('HTTP_USER_AGENT') |
|
36 | 36 | |
|
37 | 37 | |
|
38 | 38 | def get_call_context(request) -> dict: |
|
39 | 39 | cc = {} |
|
40 | 40 | registry = request.registry |
|
41 | 41 | if hasattr(registry, 'vcs_call_context'): |
|
42 | 42 | cc.update({ |
|
43 | 43 | 'X-RC-Method': registry.vcs_call_context.get('method'), |
|
44 | 44 | 'X-RC-Repo-Name': registry.vcs_call_context.get('repo_name') |
|
45 | 45 | }) |
|
46 | 46 | |
|
47 | 47 | return cc |
|
48 | 48 | |
|
49 | 49 | |
|
50 | 50 | def get_headers_call_context(environ, strict=True): |
|
51 | 51 | if 'HTTP_X_RC_VCS_STREAM_CALL_CONTEXT' in environ: |
|
52 | 52 | packed_cc = base64.b64decode(environ['HTTP_X_RC_VCS_STREAM_CALL_CONTEXT']) |
|
53 | 53 | return msgpack.unpackb(packed_cc) |
|
54 | 54 | elif strict: |
|
55 | 55 | raise ValueError('Expected header HTTP_X_RC_VCS_STREAM_CALL_CONTEXT not found') |
|
56 | 56 | |
|
57 | 57 | |
|
58 | 58 | class RequestWrapperTween: |
|
59 | 59 | def __init__(self, handler, registry): |
|
60 | 60 | self.handler = handler |
|
61 | 61 | self.registry = registry |
|
62 | 62 | |
|
63 | 63 | # one-time configuration code goes here |
|
64 | 64 | |
|
65 | 65 | def __call__(self, request): |
|
66 | 66 | start = time.time() |
|
67 | 67 | log.debug('Starting request time measurement') |
|
68 | 68 | response = None |
|
69 | 69 | |
|
70 | 70 | try: |
|
71 | 71 | response = self.handler(request) |
|
72 | 72 | finally: |
|
73 | 73 | ua = get_user_agent(request.environ) |
|
74 | 74 | call_context = get_call_context(request) |
|
75 | 75 | vcs_method = call_context.get('X-RC-Method', '_NO_VCS_METHOD') |
|
76 | 76 | repo_name = call_context.get('X-RC-Repo-Name', '') |
|
77 | 77 | |
|
78 | 78 | count = request.request_count() |
|
79 | 79 | _ver_ = vcsserver.get_version() |
|
80 | 80 | _path = safe_str(get_access_path(request.environ)) |
|
81 | 81 | |
|
82 | 82 | ip = '127.0.0.1' |
|
83 | 83 | match_route = request.matched_route.name if request.matched_route else "NOT_FOUND" |
|
84 | 84 | resp_code = getattr(response, 'status_code', 'UNDEFINED') |
|
85 | 85 | |
|
86 | 86 | _view_path = f"{repo_name}@{_path}/{vcs_method}" |
|
87 | 87 | |
|
88 | 88 | total = time.time() - start |
|
89 | 89 | |
|
90 | 90 | log.info( |
|
91 | 91 | 'Req[%4s] IP: %s %s Request to %s time: %.4fs [%s], VCSServer %s', |
|
92 | 92 | count, ip, request.environ.get('REQUEST_METHOD'), |
|
93 | 93 | _view_path, total, ua, _ver_, |
|
94 | 94 | extra={"time": total, "ver": _ver_, "code": resp_code, |
|
95 | 95 | "path": _path, "view_name": match_route, "user_agent": ua, |
|
96 | 96 | "vcs_method": vcs_method, "repo_name": repo_name} |
|
97 | 97 | ) |
|
98 | 98 | |
|
99 | 99 | statsd = request.registry.statsd |
|
100 | 100 | if statsd: |
|
101 | 101 | match_route = request.matched_route.name if request.matched_route else _path |
|
102 | 102 | elapsed_time_ms = round(1000.0 * total) # use ms only |
|
103 | 103 | statsd.timing( |
|
104 | 104 | "vcsserver_req_timing.histogram", elapsed_time_ms, |
|
105 | 105 | tags=[ |
|
106 | 106 | f"view_name:{match_route}", |
|
107 | 107 | f"code:{resp_code}" |
|
108 | 108 | ], |
|
109 | 109 | use_decimals=False |
|
110 | 110 | ) |
|
111 | 111 | statsd.incr( |
|
112 | 112 | "vcsserver_req_total", tags=[ |
|
113 | 113 | f"view_name:{match_route}", |
|
114 | 114 | f"code:{resp_code}" |
|
115 | 115 | ]) |
|
116 | 116 | |
|
117 | 117 | return response |
|
118 | 118 | |
|
119 | 119 | |
|
120 | 120 | def includeme(config): |
|
121 | 121 | config.add_tween( |
|
122 | 122 | 'vcsserver.tweens.request_wrapper.RequestWrapperTween', |
|
123 | 123 | ) |
@@ -1,116 +1,116 b'' | |||
|
1 | 1 | # RhodeCode VCSServer provides access to different vcs backends via network. |
|
2 | 2 | # Copyright (C) 2014-2023 RhodeCode GmbH |
|
3 | 3 | # |
|
4 | 4 | # This program is free software; you can redistribute it and/or modify |
|
5 | 5 | # it under the terms of the GNU General Public License as published by |
|
6 | 6 | # the Free Software Foundation; either version 3 of the License, or |
|
7 | 7 | # (at your option) any later version. |
|
8 | 8 | # |
|
9 | 9 | # This program is distributed in the hope that it will be useful, |
|
10 | 10 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
11 | 11 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
12 | 12 | # GNU General Public License for more details. |
|
13 | 13 | # |
|
14 | 14 | # You should have received a copy of the GNU General Public License |
|
15 | 15 | # along with this program; if not, write to the Free Software Foundation, |
|
16 | 16 | # Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
|
17 | 17 | |
|
18 | 18 | """Extract the responses of a WSGI app.""" |
|
19 | 19 | |
|
20 | 20 | __all__ = ('WSGIAppCaller',) |
|
21 | 21 | |
|
22 | 22 | import io |
|
23 | 23 | import logging |
|
24 | 24 | import os |
|
25 | 25 | |
|
26 | from vcsserver.str_utils import ascii_bytes | |
|
26 | from vcsserver.lib.str_utils import ascii_bytes | |
|
27 | 27 | |
|
28 | 28 | log = logging.getLogger(__name__) |
|
29 | 29 | |
|
30 | 30 | DEV_NULL = open(os.devnull) |
|
31 | 31 | |
|
32 | 32 | |
|
33 | 33 | def _complete_environ(environ, input_data: bytes): |
|
34 | 34 | """Update the missing wsgi.* variables of a WSGI environment. |
|
35 | 35 | |
|
36 | 36 | :param environ: WSGI environment to update |
|
37 | 37 | :type environ: dict |
|
38 | 38 | :param input_data: data to be read by the app |
|
39 | 39 | :type input_data: bytes |
|
40 | 40 | """ |
|
41 | 41 | environ.update({ |
|
42 | 42 | 'wsgi.version': (1, 0), |
|
43 | 43 | 'wsgi.url_scheme': 'http', |
|
44 | 44 | 'wsgi.multithread': True, |
|
45 | 45 | 'wsgi.multiprocess': True, |
|
46 | 46 | 'wsgi.run_once': False, |
|
47 | 47 | 'wsgi.input': io.BytesIO(input_data), |
|
48 | 48 | 'wsgi.errors': DEV_NULL, |
|
49 | 49 | }) |
|
50 | 50 | |
|
51 | 51 | |
|
52 | 52 | # pylint: disable=too-few-public-methods |
|
53 | 53 | class _StartResponse: |
|
54 | 54 | """Save the arguments of a start_response call.""" |
|
55 | 55 | |
|
56 | 56 | __slots__ = ['status', 'headers', 'content'] |
|
57 | 57 | |
|
58 | 58 | def __init__(self): |
|
59 | 59 | self.status = None |
|
60 | 60 | self.headers = None |
|
61 | 61 | self.content = [] |
|
62 | 62 | |
|
63 | 63 | def __call__(self, status, headers, exc_info=None): |
|
64 | 64 | # TODO(skreft): do something meaningful with the exc_info |
|
65 | 65 | exc_info = None # avoid dangling circular reference |
|
66 | 66 | self.status = status |
|
67 | 67 | self.headers = headers |
|
68 | 68 | |
|
69 | 69 | return self.write |
|
70 | 70 | |
|
71 | 71 | def write(self, content): |
|
72 | 72 | """Write method returning when calling this object. |
|
73 | 73 | |
|
74 | 74 | All the data written is then available in content. |
|
75 | 75 | """ |
|
76 | 76 | self.content.append(content) |
|
77 | 77 | |
|
78 | 78 | |
|
79 | 79 | class WSGIAppCaller: |
|
80 | 80 | """Calls a WSGI app.""" |
|
81 | 81 | |
|
82 | 82 | def __init__(self, app): |
|
83 | 83 | """ |
|
84 | 84 | :param app: WSGI app to call |
|
85 | 85 | """ |
|
86 | 86 | self.app = app |
|
87 | 87 | |
|
88 | 88 | def handle(self, environ, input_data): |
|
89 | 89 | """Process a request with the WSGI app. |
|
90 | 90 | |
|
91 | 91 | The returned data of the app is fully consumed into a list. |
|
92 | 92 | |
|
93 | 93 | :param environ: WSGI environment to update |
|
94 | 94 | :type environ: dict |
|
95 | 95 | :param input_data: data to be read by the app |
|
96 | 96 | :type input_data: str/bytes |
|
97 | 97 | |
|
98 | 98 | :returns: a tuple with the contents, status and headers |
|
99 | 99 | :rtype: (list<str>, str, list<(str, str)>) |
|
100 | 100 | """ |
|
101 | 101 | _complete_environ(environ, ascii_bytes(input_data, allow_bytes=True)) |
|
102 | 102 | start_response = _StartResponse() |
|
103 | 103 | log.debug("Calling wrapped WSGI application") |
|
104 | 104 | responses = self.app(environ, start_response) |
|
105 | 105 | responses_list = list(responses) |
|
106 | 106 | existing_responses = start_response.content |
|
107 | 107 | if existing_responses: |
|
108 | 108 | log.debug("Adding returned response to response written via write()") |
|
109 | 109 | existing_responses.extend(responses_list) |
|
110 | 110 | responses_list = existing_responses |
|
111 | 111 | if hasattr(responses, 'close'): |
|
112 | 112 | log.debug("Closing iterator from WSGI application") |
|
113 | 113 | responses.close() |
|
114 | 114 | |
|
115 | 115 | log.debug("Handling of WSGI request done, returning response") |
|
116 | 116 | return responses_list, start_response.status, start_response.headers |
|
1 | NO CONTENT: file was removed |
General Comments 0
You need to be logged in to leave comments.
Login now