Show More
@@ -1,1053 +1,1053 b'' | |||
|
1 | 1 | # Copyright (C) 2014-2023 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 | """ |
|
20 | 20 | GIT repository module |
|
21 | 21 | """ |
|
22 | 22 | |
|
23 | 23 | import logging |
|
24 | 24 | import os |
|
25 | 25 | import re |
|
26 | 26 | |
|
27 | 27 | from zope.cachedescriptors.property import Lazy as LazyProperty |
|
28 | 28 | |
|
29 | 29 | from collections import OrderedDict |
|
30 | 30 | from rhodecode.lib.datelib import ( |
|
31 | 31 | utcdate_fromtimestamp, makedate, date_astimestamp) |
|
32 | 32 | from rhodecode.lib.hash_utils import safe_str |
|
33 | 33 | from rhodecode.lib.utils2 import CachedProperty |
|
34 | 34 | from rhodecode.lib.vcs import connection, path as vcspath |
|
35 | 35 | from rhodecode.lib.vcs.backends.base import ( |
|
36 | 36 | BaseRepository, CollectionGenerator, Config, MergeResponse, |
|
37 | 37 | MergeFailureReason, Reference) |
|
38 | 38 | from rhodecode.lib.vcs.backends.git.commit import GitCommit |
|
39 | 39 | from rhodecode.lib.vcs.backends.git.diff import GitDiff |
|
40 | 40 | from rhodecode.lib.vcs.backends.git.inmemory import GitInMemoryCommit |
|
41 | 41 | from rhodecode.lib.vcs.exceptions import ( |
|
42 | 42 | CommitDoesNotExistError, EmptyRepositoryError, |
|
43 | 43 | RepositoryError, TagAlreadyExistError, TagDoesNotExistError, VCSError, UnresolvedFilesInRepo) |
|
44 | 44 | |
|
45 | 45 | |
|
46 | 46 | SHA_PATTERN = re.compile(r'^([0-9a-fA-F]{12}|[0-9a-fA-F]{40})$') |
|
47 | 47 | |
|
48 | 48 | log = logging.getLogger(__name__) |
|
49 | 49 | |
|
50 | 50 | |
|
51 | 51 | class GitRepository(BaseRepository): |
|
52 | 52 | """ |
|
53 | 53 | Git repository backend. |
|
54 | 54 | """ |
|
55 | 55 | DEFAULT_BRANCH_NAME = os.environ.get('GIT_DEFAULT_BRANCH_NAME') or 'master' |
|
56 | 56 | DEFAULT_REF = f'branch:{DEFAULT_BRANCH_NAME}' |
|
57 | 57 | |
|
58 | 58 | contact = BaseRepository.DEFAULT_CONTACT |
|
59 | 59 | |
|
60 | 60 | def __init__(self, repo_path, config=None, create=False, src_url=None, |
|
61 | 61 | do_workspace_checkout=False, with_wire=None, bare=False): |
|
62 | 62 | |
|
63 | 63 | self.path = safe_str(os.path.abspath(repo_path)) |
|
64 | 64 | self.config = config if config else self.get_default_config() |
|
65 | 65 | self.with_wire = with_wire or {"cache": False} # default should not use cache |
|
66 | 66 | |
|
67 | 67 | self._init_repo(create, src_url, do_workspace_checkout, bare) |
|
68 | 68 | |
|
69 | 69 | # caches |
|
70 | 70 | self._commit_ids = {} |
|
71 | 71 | |
|
72 | 72 | @LazyProperty |
|
73 | 73 | def _remote(self): |
|
74 | 74 | repo_id = self.path |
|
75 | 75 | return connection.Git(self.path, repo_id, self.config, with_wire=self.with_wire) |
|
76 | 76 | |
|
77 | 77 | @LazyProperty |
|
78 | 78 | def bare(self): |
|
79 | 79 | return self._remote.bare() |
|
80 | 80 | |
|
81 | 81 | @LazyProperty |
|
82 | 82 | def head(self): |
|
83 | 83 | return self._remote.head() |
|
84 | 84 | |
|
85 | 85 | @CachedProperty |
|
86 | 86 | def commit_ids(self): |
|
87 | 87 | """ |
|
88 | 88 | Returns list of commit ids, in ascending order. Being lazy |
|
89 | 89 | attribute allows external tools to inject commit ids from cache. |
|
90 | 90 | """ |
|
91 | 91 | commit_ids = self._get_all_commit_ids() |
|
92 | 92 | self._rebuild_cache(commit_ids) |
|
93 | 93 | return commit_ids |
|
94 | 94 | |
|
95 | 95 | def _rebuild_cache(self, commit_ids): |
|
96 | 96 | self._commit_ids = {commit_id: index |
|
97 | 97 | for index, commit_id in enumerate(commit_ids)} |
|
98 | 98 | |
|
99 | 99 | def run_git_command(self, cmd, **opts): |
|
100 | 100 | """ |
|
101 | 101 | Runs given ``cmd`` as git command and returns tuple |
|
102 | 102 | (stdout, stderr). |
|
103 | 103 | |
|
104 | 104 | :param cmd: git command to be executed |
|
105 | 105 | :param opts: env options to pass into Subprocess command |
|
106 | 106 | """ |
|
107 | 107 | if not isinstance(cmd, list): |
|
108 | 108 | raise ValueError(f'cmd must be a list, got {type(cmd)} instead') |
|
109 | 109 | |
|
110 | 110 | skip_stderr_log = opts.pop('skip_stderr_log', False) |
|
111 | 111 | out, err = self._remote.run_git_command(cmd, **opts) |
|
112 | 112 | if err and not skip_stderr_log: |
|
113 | 113 | log.debug('Stderr output of git command "%s":\n%s', cmd, err) |
|
114 | 114 | return out, err |
|
115 | 115 | |
|
116 | 116 | @staticmethod |
|
117 | 117 | def check_url(url, config): |
|
118 | 118 | """ |
|
119 | 119 | Function will check given url and try to verify if it's a valid |
|
120 | 120 | link. Sometimes it may happened that git will issue basic |
|
121 | 121 | auth request that can cause whole API to hang when used from python |
|
122 | 122 | or other external calls. |
|
123 | 123 | |
|
124 | 124 | On failures it'll raise urllib2.HTTPError, exception is also thrown |
|
125 | 125 | when the return code is non 200 |
|
126 | 126 | """ |
|
127 | 127 | # check first if it's not an url |
|
128 | 128 | if os.path.isdir(url) or url.startswith('file:'): |
|
129 | 129 | return True |
|
130 | 130 | |
|
131 | 131 | if '+' in url.split('://', 1)[0]: |
|
132 | 132 | url = url.split('+', 1)[1] |
|
133 | 133 | |
|
134 | 134 | # Request the _remote to verify the url |
|
135 | 135 | return connection.Git.check_url(url, config.serialize()) |
|
136 | 136 | |
|
137 | 137 | @staticmethod |
|
138 | 138 | def is_valid_repository(path): |
|
139 | 139 | if os.path.isdir(os.path.join(path, '.git')): |
|
140 | 140 | return True |
|
141 | 141 | # check case of bare repository |
|
142 | 142 | try: |
|
143 | 143 | GitRepository(path) |
|
144 | 144 | return True |
|
145 | 145 | except VCSError: |
|
146 | 146 | pass |
|
147 | 147 | return False |
|
148 | 148 | |
|
149 | 149 | def _init_repo(self, create, src_url=None, do_workspace_checkout=False, |
|
150 | 150 | bare=False): |
|
151 | 151 | if create and os.path.exists(self.path): |
|
152 | 152 | raise RepositoryError( |
|
153 | 153 | f"Cannot create repository at {self.path}, location already exist") |
|
154 | 154 | |
|
155 | 155 | if bare and do_workspace_checkout: |
|
156 | 156 | raise RepositoryError("Cannot update a bare repository") |
|
157 | 157 | try: |
|
158 | 158 | |
|
159 | 159 | if src_url: |
|
160 | 160 | # check URL before any actions |
|
161 | 161 | GitRepository.check_url(src_url, self.config) |
|
162 | 162 | |
|
163 | 163 | if create: |
|
164 | 164 | if bare: |
|
165 | 165 | self._remote.init_bare() |
|
166 | 166 | else: |
|
167 | 167 | self._remote.init() |
|
168 | 168 | |
|
169 | 169 | if src_url and bare: |
|
170 | 170 | # bare repository only allows a fetch and checkout is not allowed |
|
171 | 171 | self.fetch(src_url, commit_ids=None) |
|
172 | 172 | elif src_url: |
|
173 | 173 | self.pull(src_url, commit_ids=None, |
|
174 | 174 | update_after=do_workspace_checkout) |
|
175 | 175 | |
|
176 | 176 | else: |
|
177 | 177 | if not self._remote.assert_correct_path(): |
|
178 | 178 | raise RepositoryError( |
|
179 | 179 | f'Path "{self.path}" does not contain a Git repository') |
|
180 | 180 | |
|
181 | 181 | # TODO: johbo: check if we have to translate the OSError here |
|
182 | 182 | except OSError as err: |
|
183 | 183 | raise RepositoryError(err) |
|
184 | 184 | |
|
185 | 185 | def _get_all_commit_ids(self): |
|
186 | 186 | return self._remote.get_all_commit_ids() |
|
187 | 187 | |
|
188 | 188 | def _get_commit_ids(self, filters=None): |
|
189 | 189 | # we must check if this repo is not empty, since later command |
|
190 | 190 | # fails if it is. And it's cheaper to ask than throw the subprocess |
|
191 | 191 | # errors |
|
192 | 192 | |
|
193 | 193 | head = self._remote.head(show_exc=False) |
|
194 | 194 | |
|
195 | 195 | if not head: |
|
196 | 196 | return [] |
|
197 | 197 | |
|
198 | 198 | rev_filter = ['--branches', '--tags'] |
|
199 | 199 | extra_filter = [] |
|
200 | 200 | |
|
201 | 201 | if filters: |
|
202 | 202 | if filters.get('since'): |
|
203 | 203 | extra_filter.append('--since=%s' % (filters['since'])) |
|
204 | 204 | if filters.get('until'): |
|
205 | 205 | extra_filter.append('--until=%s' % (filters['until'])) |
|
206 | 206 | if filters.get('branch_name'): |
|
207 | 207 | rev_filter = [] |
|
208 | 208 | extra_filter.append(filters['branch_name']) |
|
209 | 209 | rev_filter.extend(extra_filter) |
|
210 | 210 | |
|
211 | 211 | # if filters.get('start') or filters.get('end'): |
|
212 | 212 | # # skip is offset, max-count is limit |
|
213 | 213 | # if filters.get('start'): |
|
214 | 214 | # extra_filter += ' --skip=%s' % filters['start'] |
|
215 | 215 | # if filters.get('end'): |
|
216 | 216 | # extra_filter += ' --max-count=%s' % (filters['end'] - (filters['start'] or 0)) |
|
217 | 217 | |
|
218 | 218 | cmd = ['rev-list', '--reverse', '--date-order'] + rev_filter |
|
219 | 219 | try: |
|
220 | 220 | output, __ = self.run_git_command(cmd) |
|
221 | 221 | except RepositoryError: |
|
222 | 222 | # Can be raised for empty repositories |
|
223 | 223 | return [] |
|
224 | 224 | return output.splitlines() |
|
225 | 225 | |
|
226 | 226 | def _lookup_commit(self, commit_id_or_idx, translate_tag=True, maybe_unreachable=False, reference_obj=None): |
|
227 | 227 | |
|
228 | 228 | def is_null(value): |
|
229 | 229 | return len(value) == commit_id_or_idx.count('0') |
|
230 | 230 | |
|
231 | 231 | if commit_id_or_idx in (None, '', 'tip', 'HEAD', 'head', -1): |
|
232 | 232 | return self.commit_ids[-1] |
|
233 | 233 | |
|
234 | 234 | commit_missing_err = "Commit {} does not exist for `{}`".format( |
|
235 | 235 | *map(safe_str, [commit_id_or_idx, self.name])) |
|
236 | 236 | |
|
237 | 237 | is_bstr = isinstance(commit_id_or_idx, str) |
|
238 | 238 | is_branch = reference_obj and reference_obj.branch |
|
239 | 239 | |
|
240 | 240 | lookup_ok = False |
|
241 | 241 | if is_bstr: |
|
242 | 242 | # Need to call remote to translate id for tagging scenarios, |
|
243 | 243 | # or branch that are numeric |
|
244 | 244 | try: |
|
245 | 245 | remote_data = self._remote.get_object(commit_id_or_idx, |
|
246 | 246 | maybe_unreachable=maybe_unreachable) |
|
247 | 247 | commit_id_or_idx = remote_data["commit_id"] |
|
248 | 248 | lookup_ok = True |
|
249 | 249 | except (CommitDoesNotExistError,): |
|
250 | 250 | lookup_ok = False |
|
251 | 251 | |
|
252 | 252 | if lookup_ok is False: |
|
253 | 253 | is_numeric_idx = \ |
|
254 | 254 | (is_bstr and commit_id_or_idx.isdigit() and len(commit_id_or_idx) < 12) \ |
|
255 | 255 | or isinstance(commit_id_or_idx, int) |
|
256 | 256 | if not is_branch and (is_numeric_idx or is_null(commit_id_or_idx)): |
|
257 | 257 | try: |
|
258 | 258 | commit_id_or_idx = self.commit_ids[int(commit_id_or_idx)] |
|
259 | 259 | lookup_ok = True |
|
260 | 260 | except Exception: |
|
261 | 261 | raise CommitDoesNotExistError(commit_missing_err) |
|
262 | 262 | |
|
263 | 263 | # we failed regular lookup, and by integer number lookup |
|
264 | 264 | if lookup_ok is False: |
|
265 | 265 | raise CommitDoesNotExistError(commit_missing_err) |
|
266 | 266 | |
|
267 | 267 | # Ensure we return full id |
|
268 | 268 | if not SHA_PATTERN.match(str(commit_id_or_idx)): |
|
269 | 269 | raise CommitDoesNotExistError( |
|
270 | 270 | "Given commit id %s not recognized" % commit_id_or_idx) |
|
271 | 271 | return commit_id_or_idx |
|
272 | 272 | |
|
273 | 273 | def get_hook_location(self): |
|
274 | 274 | """ |
|
275 | 275 | returns absolute path to location where hooks are stored |
|
276 | 276 | """ |
|
277 | 277 | loc = os.path.join(self.path, 'hooks') |
|
278 | 278 | if not self.bare: |
|
279 | 279 | loc = os.path.join(self.path, '.git', 'hooks') |
|
280 | 280 | return loc |
|
281 | 281 | |
|
282 | 282 | @LazyProperty |
|
283 | 283 | def last_change(self): |
|
284 | 284 | """ |
|
285 | 285 | Returns last change made on this repository as |
|
286 | 286 | `datetime.datetime` object. |
|
287 | 287 | """ |
|
288 | 288 | try: |
|
289 | 289 | return self.get_commit().date |
|
290 | 290 | except RepositoryError: |
|
291 | 291 | tzoffset = makedate()[1] |
|
292 | 292 | return utcdate_fromtimestamp(self._get_fs_mtime(), tzoffset) |
|
293 | 293 | |
|
294 | 294 | def _get_fs_mtime(self): |
|
295 | 295 | idx_loc = '' if self.bare else '.git' |
|
296 | 296 | # fallback to filesystem |
|
297 | 297 | in_path = os.path.join(self.path, idx_loc, "index") |
|
298 | 298 | he_path = os.path.join(self.path, idx_loc, "HEAD") |
|
299 | 299 | if os.path.exists(in_path): |
|
300 | 300 | return os.stat(in_path).st_mtime |
|
301 | 301 | else: |
|
302 | 302 | return os.stat(he_path).st_mtime |
|
303 | 303 | |
|
304 | 304 | @LazyProperty |
|
305 | 305 | def description(self): |
|
306 | 306 | description = self._remote.get_description() |
|
307 | 307 | return safe_str(description or self.DEFAULT_DESCRIPTION) |
|
308 | 308 | |
|
309 | 309 | def _get_refs_entries(self, prefix='', reverse=False, strip_prefix=True): |
|
310 | 310 | if self.is_empty(): |
|
311 | 311 | return OrderedDict() |
|
312 | 312 | |
|
313 | 313 | result = [] |
|
314 | 314 | for ref, sha in self._refs.items(): |
|
315 | 315 | if ref.startswith(prefix): |
|
316 | 316 | ref_name = ref |
|
317 | 317 | if strip_prefix: |
|
318 | 318 | ref_name = ref[len(prefix):] |
|
319 | 319 | result.append((safe_str(ref_name), sha)) |
|
320 | 320 | |
|
321 | 321 | def get_name(entry): |
|
322 | 322 | return entry[0] |
|
323 | 323 | |
|
324 | 324 | return OrderedDict(sorted(result, key=get_name, reverse=reverse)) |
|
325 | 325 | |
|
326 | 326 | def _get_branches(self): |
|
327 | 327 | return self._get_refs_entries(prefix='refs/heads/', strip_prefix=True) |
|
328 | 328 | |
|
329 | 329 | @CachedProperty |
|
330 | 330 | def branches(self): |
|
331 | 331 | return self._get_branches() |
|
332 | 332 | |
|
333 | 333 | @CachedProperty |
|
334 | 334 | def branches_closed(self): |
|
335 | 335 | return {} |
|
336 | 336 | |
|
337 | 337 | @CachedProperty |
|
338 | 338 | def bookmarks(self): |
|
339 | 339 | return {} |
|
340 | 340 | |
|
341 | 341 | @CachedProperty |
|
342 | 342 | def branches_all(self): |
|
343 | 343 | all_branches = {} |
|
344 | 344 | all_branches.update(self.branches) |
|
345 | 345 | all_branches.update(self.branches_closed) |
|
346 | 346 | return all_branches |
|
347 | 347 | |
|
348 | 348 | @CachedProperty |
|
349 | 349 | def tags(self): |
|
350 | 350 | return self._get_tags() |
|
351 | 351 | |
|
352 | 352 | def _get_tags(self): |
|
353 | 353 | return self._get_refs_entries(prefix='refs/tags/', strip_prefix=True, reverse=True) |
|
354 | 354 | |
|
355 | 355 | def tag(self, name, user, commit_id=None, message=None, date=None, |
|
356 | 356 | **kwargs): |
|
357 | 357 | # TODO: fix this method to apply annotated tags correct with message |
|
358 | 358 | """ |
|
359 | 359 | Creates and returns a tag for the given ``commit_id``. |
|
360 | 360 | |
|
361 | 361 | :param name: name for new tag |
|
362 | 362 | :param user: full username, i.e.: "Joe Doe <joe.doe@example.com>" |
|
363 | 363 | :param commit_id: commit id for which new tag would be created |
|
364 | 364 | :param message: message of the tag's commit |
|
365 | 365 | :param date: date of tag's commit |
|
366 | 366 | |
|
367 | 367 | :raises TagAlreadyExistError: if tag with same name already exists |
|
368 | 368 | """ |
|
369 | 369 | if name in self.tags: |
|
370 | 370 | raise TagAlreadyExistError("Tag %s already exists" % name) |
|
371 | 371 | commit = self.get_commit(commit_id=commit_id) |
|
372 | 372 | message = message or f"Added tag {name} for commit {commit.raw_id}" |
|
373 | 373 | |
|
374 | 374 | self._remote.set_refs('refs/tags/%s' % name, commit.raw_id) |
|
375 | 375 | |
|
376 | 376 | self._invalidate_prop_cache('tags') |
|
377 | 377 | self._invalidate_prop_cache('_refs') |
|
378 | 378 | |
|
379 | 379 | return commit |
|
380 | 380 | |
|
381 | 381 | def remove_tag(self, name, user, message=None, date=None): |
|
382 | 382 | """ |
|
383 | 383 | Removes tag with the given ``name``. |
|
384 | 384 | |
|
385 | 385 | :param name: name of the tag to be removed |
|
386 | 386 | :param user: full username, i.e.: "Joe Doe <joe.doe@example.com>" |
|
387 | 387 | :param message: message of the tag's removal commit |
|
388 | 388 | :param date: date of tag's removal commit |
|
389 | 389 | |
|
390 | 390 | :raises TagDoesNotExistError: if tag with given name does not exists |
|
391 | 391 | """ |
|
392 | 392 | if name not in self.tags: |
|
393 | 393 | raise TagDoesNotExistError("Tag %s does not exist" % name) |
|
394 | 394 | |
|
395 | 395 | self._remote.tag_remove(name) |
|
396 | 396 | self._invalidate_prop_cache('tags') |
|
397 | 397 | self._invalidate_prop_cache('_refs') |
|
398 | 398 | |
|
399 | 399 | def _get_refs(self): |
|
400 | 400 | return self._remote.get_refs() |
|
401 | 401 | |
|
402 | 402 | @CachedProperty |
|
403 | 403 | def _refs(self): |
|
404 | 404 | return self._get_refs() |
|
405 | 405 | |
|
406 | 406 | @property |
|
407 | 407 | def _ref_tree(self): |
|
408 | 408 | node = tree = {} |
|
409 | 409 | for ref, sha in self._refs.items(): |
|
410 | 410 | path = ref.split('/') |
|
411 | 411 | for bit in path[:-1]: |
|
412 | 412 | node = node.setdefault(bit, {}) |
|
413 | 413 | node[path[-1]] = sha |
|
414 | 414 | node = tree |
|
415 | 415 | return tree |
|
416 | 416 | |
|
417 | 417 | def get_remote_ref(self, ref_name): |
|
418 | 418 | ref_key = f'refs/remotes/origin/{safe_str(ref_name)}' |
|
419 | 419 | try: |
|
420 | 420 | return self._refs[ref_key] |
|
421 | 421 | except Exception: |
|
422 | 422 | return |
|
423 | 423 | |
|
424 | 424 | def get_commit(self, commit_id=None, commit_idx=None, pre_load=None, |
|
425 | 425 | translate_tag=True, maybe_unreachable=False, reference_obj=None): |
|
426 | 426 | """ |
|
427 | 427 | Returns `GitCommit` object representing commit from git repository |
|
428 | 428 | at the given `commit_id` or head (most recent commit) if None given. |
|
429 | 429 | """ |
|
430 | 430 | |
|
431 | 431 | if self.is_empty(): |
|
432 | 432 | raise EmptyRepositoryError("There are no commits yet") |
|
433 | 433 | |
|
434 | 434 | if commit_id is not None: |
|
435 | 435 | self._validate_commit_id(commit_id) |
|
436 | 436 | try: |
|
437 | 437 | # we have cached idx, use it without contacting the remote |
|
438 | 438 | idx = self._commit_ids[commit_id] |
|
439 | 439 | return GitCommit(self, commit_id, idx, pre_load=pre_load) |
|
440 | 440 | except KeyError: |
|
441 | 441 | pass |
|
442 | 442 | |
|
443 | 443 | elif commit_idx is not None: |
|
444 | 444 | self._validate_commit_idx(commit_idx) |
|
445 | 445 | try: |
|
446 | 446 | _commit_id = self.commit_ids[commit_idx] |
|
447 | 447 | if commit_idx < 0: |
|
448 | 448 | commit_idx = self.commit_ids.index(_commit_id) |
|
449 | 449 | return GitCommit(self, _commit_id, commit_idx, pre_load=pre_load) |
|
450 | 450 | except IndexError: |
|
451 | 451 | commit_id = commit_idx |
|
452 | 452 | else: |
|
453 | 453 | commit_id = "tip" |
|
454 | 454 | |
|
455 | 455 | if translate_tag: |
|
456 | 456 | commit_id = self._lookup_commit( |
|
457 | 457 | commit_id, maybe_unreachable=maybe_unreachable, |
|
458 | 458 | reference_obj=reference_obj) |
|
459 | 459 | |
|
460 | 460 | try: |
|
461 | 461 | idx = self._commit_ids[commit_id] |
|
462 | 462 | except KeyError: |
|
463 | 463 | idx = -1 |
|
464 | 464 | |
|
465 | 465 | return GitCommit(self, commit_id, idx, pre_load=pre_load) |
|
466 | 466 | |
|
467 | 467 | def get_commits( |
|
468 | 468 | self, start_id=None, end_id=None, start_date=None, end_date=None, |
|
469 | 469 | branch_name=None, show_hidden=False, pre_load=None, translate_tags=True): |
|
470 | 470 | """ |
|
471 | 471 | Returns generator of `GitCommit` objects from start to end (both |
|
472 | 472 | are inclusive), in ascending date order. |
|
473 | 473 | |
|
474 | 474 | :param start_id: None, str(commit_id) |
|
475 | 475 | :param end_id: None, str(commit_id) |
|
476 | 476 | :param start_date: if specified, commits with commit date less than |
|
477 | 477 | ``start_date`` would be filtered out from returned set |
|
478 | 478 | :param end_date: if specified, commits with commit date greater than |
|
479 | 479 | ``end_date`` would be filtered out from returned set |
|
480 | 480 | :param branch_name: if specified, commits not reachable from given |
|
481 | 481 | branch would be filtered out from returned set |
|
482 | 482 | :param show_hidden: Show hidden commits such as obsolete or hidden from |
|
483 | 483 | Mercurial evolve |
|
484 | 484 | :raise BranchDoesNotExistError: If given `branch_name` does not |
|
485 | 485 | exist. |
|
486 | 486 | :raise CommitDoesNotExistError: If commits for given `start` or |
|
487 | 487 | `end` could not be found. |
|
488 | 488 | |
|
489 | 489 | """ |
|
490 | 490 | if self.is_empty(): |
|
491 | 491 | raise EmptyRepositoryError("There are no commits yet") |
|
492 | 492 | |
|
493 | 493 | self._validate_branch_name(branch_name) |
|
494 | 494 | |
|
495 | 495 | if start_id is not None: |
|
496 | 496 | self._validate_commit_id(start_id) |
|
497 | 497 | if end_id is not None: |
|
498 | 498 | self._validate_commit_id(end_id) |
|
499 | 499 | |
|
500 | 500 | start_raw_id = self._lookup_commit(start_id) |
|
501 | 501 | start_pos = self._commit_ids[start_raw_id] if start_id else None |
|
502 | 502 | end_raw_id = self._lookup_commit(end_id) |
|
503 | 503 | end_pos = max(0, self._commit_ids[end_raw_id]) if end_id else None |
|
504 | 504 | |
|
505 | 505 | if None not in [start_id, end_id] and start_pos > end_pos: |
|
506 | 506 | raise RepositoryError( |
|
507 | 507 | "Start commit '%s' cannot be after end commit '%s'" % |
|
508 | 508 | (start_id, end_id)) |
|
509 | 509 | |
|
510 | 510 | if end_pos is not None: |
|
511 | 511 | end_pos += 1 |
|
512 | 512 | |
|
513 | 513 | filter_ = [] |
|
514 | 514 | if branch_name: |
|
515 | 515 | filter_.append({'branch_name': branch_name}) |
|
516 | 516 | if start_date and not end_date: |
|
517 | 517 | filter_.append({'since': start_date}) |
|
518 | 518 | if end_date and not start_date: |
|
519 | 519 | filter_.append({'until': end_date}) |
|
520 | 520 | if start_date and end_date: |
|
521 | 521 | filter_.append({'since': start_date}) |
|
522 | 522 | filter_.append({'until': end_date}) |
|
523 | 523 | |
|
524 | 524 | # if start_pos or end_pos: |
|
525 | 525 | # filter_.append({'start': start_pos}) |
|
526 | 526 | # filter_.append({'end': end_pos}) |
|
527 | 527 | |
|
528 | 528 | if filter_: |
|
529 | 529 | revfilters = { |
|
530 | 530 | 'branch_name': branch_name, |
|
531 | 531 | 'since': start_date.strftime('%m/%d/%y %H:%M:%S') if start_date else None, |
|
532 | 532 | 'until': end_date.strftime('%m/%d/%y %H:%M:%S') if end_date else None, |
|
533 | 533 | 'start': start_pos, |
|
534 | 534 | 'end': end_pos, |
|
535 | 535 | } |
|
536 | 536 | commit_ids = self._get_commit_ids(filters=revfilters) |
|
537 | 537 | |
|
538 | 538 | else: |
|
539 | 539 | commit_ids = self.commit_ids |
|
540 | 540 | |
|
541 | 541 | if start_pos or end_pos: |
|
542 | 542 | commit_ids = commit_ids[start_pos: end_pos] |
|
543 | 543 | |
|
544 | 544 | return CollectionGenerator(self, commit_ids, pre_load=pre_load, |
|
545 | 545 | translate_tag=translate_tags) |
|
546 | 546 | |
|
547 | 547 | def get_diff( |
|
548 | 548 | self, commit1, commit2, path='', ignore_whitespace=False, |
|
549 | 549 | context=3, path1=None): |
|
550 | 550 | """ |
|
551 | 551 | Returns (git like) *diff*, as plain text. Shows changes introduced by |
|
552 | 552 | ``commit2`` since ``commit1``. |
|
553 | 553 | |
|
554 | 554 | :param commit1: Entry point from which diff is shown. Can be |
|
555 | 555 | ``self.EMPTY_COMMIT`` - in this case, patch showing all |
|
556 | 556 | the changes since empty state of the repository until ``commit2`` |
|
557 | 557 | :param commit2: Until which commits changes should be shown. |
|
558 | 558 | :param path: |
|
559 | 559 | :param ignore_whitespace: If set to ``True``, would not show whitespace |
|
560 | 560 | changes. Defaults to ``False``. |
|
561 | 561 | :param context: How many lines before/after changed lines should be |
|
562 | 562 | shown. Defaults to ``3``. |
|
563 | 563 | :param path1: |
|
564 | 564 | """ |
|
565 | 565 | self._validate_diff_commits(commit1, commit2) |
|
566 | 566 | if path1 is not None and path1 != path: |
|
567 | 567 | raise ValueError("Diff of two different paths not supported.") |
|
568 | 568 | |
|
569 | 569 | if path: |
|
570 | 570 | file_filter = path |
|
571 | 571 | else: |
|
572 | 572 | file_filter = None |
|
573 | 573 | |
|
574 | 574 | diff = self._remote.diff( |
|
575 | 575 | commit1.raw_id, commit2.raw_id, file_filter=file_filter, |
|
576 | 576 | opt_ignorews=ignore_whitespace, |
|
577 | 577 | context=context) |
|
578 | 578 | |
|
579 | 579 | return GitDiff(diff) |
|
580 | 580 | |
|
581 | 581 | def strip(self, commit_id, branch_name): |
|
582 | 582 | commit = self.get_commit(commit_id=commit_id) |
|
583 | 583 | if commit.merge: |
|
584 | 584 | raise Exception('Cannot reset to merge commit') |
|
585 | 585 | |
|
586 | 586 | if not branch_name: |
|
587 | 587 | raise ValueError(f'git strip requires a valid branch name, got {branch_name} instead') |
|
588 | 588 | |
|
589 | 589 | # parent is going to be the new head now |
|
590 | 590 | commit = commit.parents[0] |
|
591 | 591 | self._remote.update_refs(f'refs/heads/{branch_name}', commit.raw_id) |
|
592 | 592 | |
|
593 | 593 | # clear cached properties |
|
594 | 594 | self._invalidate_prop_cache('commit_ids') |
|
595 | 595 | self._invalidate_prop_cache('_refs') |
|
596 | 596 | self._invalidate_prop_cache('branches') |
|
597 | 597 | |
|
598 | 598 | return len(self.commit_ids) |
|
599 | 599 | |
|
600 | 600 | def get_common_ancestor(self, commit_id1, commit_id2, repo2): |
|
601 | 601 | log.debug('Calculating common ancestor between %sc1:%s and %sc2:%s', |
|
602 | 602 | self, commit_id1, repo2, commit_id2) |
|
603 | 603 | |
|
604 | 604 | if commit_id1 == commit_id2: |
|
605 | 605 | return commit_id1 |
|
606 | 606 | |
|
607 | 607 | if self != repo2: |
|
608 | 608 | commits = self._remote.get_missing_revs( |
|
609 | 609 | commit_id1, commit_id2, repo2.path) |
|
610 | 610 | if commits: |
|
611 | 611 | commit = repo2.get_commit(commits[-1]) |
|
612 | 612 | if commit.parents: |
|
613 | 613 | ancestor_id = commit.parents[0].raw_id |
|
614 | 614 | else: |
|
615 | 615 | ancestor_id = None |
|
616 | 616 | else: |
|
617 | 617 | # no commits from other repo, ancestor_id is the commit_id2 |
|
618 | 618 | ancestor_id = commit_id2 |
|
619 | 619 | else: |
|
620 | 620 | output, __ = self.run_git_command( |
|
621 | 621 | ['merge-base', commit_id1, commit_id2]) |
|
622 | 622 | ancestor_id = self.COMMIT_ID_PAT.findall(output)[0] |
|
623 | 623 | |
|
624 | 624 | log.debug('Found common ancestor with sha: %s', ancestor_id) |
|
625 | 625 | |
|
626 | 626 | return ancestor_id |
|
627 | 627 | |
|
628 | 628 | def compare(self, commit_id1, commit_id2, repo2, merge, pre_load=None): |
|
629 | 629 | repo1 = self |
|
630 | 630 | ancestor_id = None |
|
631 | 631 | |
|
632 | 632 | if commit_id1 == commit_id2: |
|
633 | 633 | commits = [] |
|
634 | 634 | elif repo1 != repo2: |
|
635 | 635 | missing_ids = self._remote.get_missing_revs(commit_id1, commit_id2, |
|
636 | 636 | repo2.path) |
|
637 | 637 | commits = [ |
|
638 | 638 | repo2.get_commit(commit_id=commit_id, pre_load=pre_load) |
|
639 | 639 | for commit_id in reversed(missing_ids)] |
|
640 | 640 | else: |
|
641 | 641 | output, __ = repo1.run_git_command( |
|
642 | 642 | ['log', '--reverse', '--pretty=format: %H', '-s', |
|
643 | 643 | f'{commit_id1}..{commit_id2}']) |
|
644 | 644 | commits = [ |
|
645 | 645 | repo1.get_commit(commit_id=commit_id, pre_load=pre_load) |
|
646 | 646 | for commit_id in self.COMMIT_ID_PAT.findall(output)] |
|
647 | 647 | |
|
648 | 648 | return commits |
|
649 | 649 | |
|
650 | 650 | @LazyProperty |
|
651 | 651 | def in_memory_commit(self): |
|
652 | 652 | """ |
|
653 | 653 | Returns ``GitInMemoryCommit`` object for this repository. |
|
654 | 654 | """ |
|
655 | 655 | return GitInMemoryCommit(self) |
|
656 | 656 | |
|
657 | 657 | def pull(self, url, commit_ids=None, update_after=False): |
|
658 | 658 | """ |
|
659 | 659 | Pull changes from external location. Pull is different in GIT |
|
660 | 660 | that fetch since it's doing a checkout |
|
661 | 661 | |
|
662 | 662 | :param commit_ids: Optional. Can be set to a list of commit ids |
|
663 | 663 | which shall be pulled from the other repository. |
|
664 | 664 | """ |
|
665 | 665 | refs = None |
|
666 | 666 | if commit_ids is not None: |
|
667 | 667 | remote_refs = self._remote.get_remote_refs(url) |
|
668 | 668 | refs = [ref for ref in remote_refs if remote_refs[ref] in commit_ids] |
|
669 | 669 | self._remote.pull(url, refs=refs, update_after=update_after) |
|
670 | 670 | self._remote.invalidate_vcs_cache() |
|
671 | 671 | |
|
672 | 672 | def fetch(self, url, commit_ids=None, **kwargs): |
|
673 | 673 | """ |
|
674 | 674 | Fetch all git objects from external location. |
|
675 | 675 | """ |
|
676 | 676 | self._remote.sync_fetch(url, refs=commit_ids, **kwargs) |
|
677 | 677 | self._remote.invalidate_vcs_cache() |
|
678 | 678 | |
|
679 | def push(self, url): | |
|
679 | def push(self, url, **kwargs): | |
|
680 | 680 | refs = None |
|
681 | self._remote.sync_push(url, refs=refs) | |
|
681 | self._remote.sync_push(url, refs=refs, **kwargs) | |
|
682 | 682 | |
|
683 | 683 | def set_refs(self, ref_name, commit_id): |
|
684 | 684 | self._remote.set_refs(ref_name, commit_id) |
|
685 | 685 | self._invalidate_prop_cache('_refs') |
|
686 | 686 | |
|
687 | 687 | def remove_ref(self, ref_name): |
|
688 | 688 | self._remote.remove_ref(ref_name) |
|
689 | 689 | self._invalidate_prop_cache('_refs') |
|
690 | 690 | |
|
691 | 691 | def run_gc(self, prune=True): |
|
692 | 692 | cmd = ['gc', '--aggressive'] |
|
693 | 693 | if prune: |
|
694 | 694 | cmd += ['--prune=now'] |
|
695 | 695 | _stdout, stderr = self.run_git_command(cmd, fail_on_stderr=False) |
|
696 | 696 | return stderr |
|
697 | 697 | |
|
698 | 698 | def _update_server_info(self): |
|
699 | 699 | """ |
|
700 | 700 | runs gits update-server-info command in this repo instance |
|
701 | 701 | """ |
|
702 | 702 | self._remote.update_server_info() |
|
703 | 703 | |
|
704 | 704 | def _current_branch(self): |
|
705 | 705 | """ |
|
706 | 706 | Return the name of the current branch. |
|
707 | 707 | |
|
708 | 708 | It only works for non bare repositories (i.e. repositories with a |
|
709 | 709 | working copy) |
|
710 | 710 | """ |
|
711 | 711 | if self.bare: |
|
712 | 712 | raise RepositoryError('Bare git repos do not have active branches') |
|
713 | 713 | |
|
714 | 714 | if self.is_empty(): |
|
715 | 715 | return None |
|
716 | 716 | |
|
717 | 717 | stdout, _ = self.run_git_command(['rev-parse', '--abbrev-ref', 'HEAD']) |
|
718 | 718 | return stdout.strip() |
|
719 | 719 | |
|
720 | 720 | def _checkout(self, branch_name, create=False, force=False): |
|
721 | 721 | """ |
|
722 | 722 | Checkout a branch in the working directory. |
|
723 | 723 | |
|
724 | 724 | It tries to create the branch if create is True, failing if the branch |
|
725 | 725 | already exists. |
|
726 | 726 | |
|
727 | 727 | It only works for non bare repositories (i.e. repositories with a |
|
728 | 728 | working copy) |
|
729 | 729 | """ |
|
730 | 730 | if self.bare: |
|
731 | 731 | raise RepositoryError('Cannot checkout branches in a bare git repo') |
|
732 | 732 | |
|
733 | 733 | cmd = ['checkout'] |
|
734 | 734 | if force: |
|
735 | 735 | cmd.append('-f') |
|
736 | 736 | if create: |
|
737 | 737 | cmd.append('-b') |
|
738 | 738 | cmd.append(branch_name) |
|
739 | 739 | self.run_git_command(cmd, fail_on_stderr=False) |
|
740 | 740 | |
|
741 | 741 | def _create_branch(self, branch_name, commit_id): |
|
742 | 742 | """ |
|
743 | 743 | creates a branch in a GIT repo |
|
744 | 744 | """ |
|
745 | 745 | self._remote.create_branch(branch_name, commit_id) |
|
746 | 746 | |
|
747 | 747 | def _identify(self): |
|
748 | 748 | """ |
|
749 | 749 | Return the current state of the working directory. |
|
750 | 750 | """ |
|
751 | 751 | if self.bare: |
|
752 | 752 | raise RepositoryError('Bare git repos do not have active branches') |
|
753 | 753 | |
|
754 | 754 | if self.is_empty(): |
|
755 | 755 | return None |
|
756 | 756 | |
|
757 | 757 | stdout, _ = self.run_git_command(['rev-parse', 'HEAD']) |
|
758 | 758 | return stdout.strip() |
|
759 | 759 | |
|
760 | 760 | def _local_clone(self, clone_path, branch_name, source_branch=None): |
|
761 | 761 | """ |
|
762 | 762 | Create a local clone of the current repo. |
|
763 | 763 | """ |
|
764 | 764 | # N.B.(skreft): the --branch option is required as otherwise the shallow |
|
765 | 765 | # clone will only fetch the active branch. |
|
766 | 766 | cmd = ['clone', '--branch', branch_name, |
|
767 | 767 | self.path, os.path.abspath(clone_path)] |
|
768 | 768 | |
|
769 | 769 | self.run_git_command(cmd, fail_on_stderr=False) |
|
770 | 770 | |
|
771 | 771 | # if we get the different source branch, make sure we also fetch it for |
|
772 | 772 | # merge conditions |
|
773 | 773 | if source_branch and source_branch != branch_name: |
|
774 | 774 | # check if the ref exists. |
|
775 | 775 | shadow_repo = GitRepository(os.path.abspath(clone_path)) |
|
776 | 776 | if shadow_repo.get_remote_ref(source_branch): |
|
777 | 777 | cmd = ['fetch', self.path, source_branch] |
|
778 | 778 | self.run_git_command(cmd, fail_on_stderr=False) |
|
779 | 779 | |
|
780 | 780 | def _local_fetch(self, repository_path, branch_name, use_origin=False): |
|
781 | 781 | """ |
|
782 | 782 | Fetch a branch from a local repository. |
|
783 | 783 | """ |
|
784 | 784 | repository_path = os.path.abspath(repository_path) |
|
785 | 785 | if repository_path == self.path: |
|
786 | 786 | raise ValueError('Cannot fetch from the same repository') |
|
787 | 787 | |
|
788 | 788 | if use_origin: |
|
789 | 789 | branch_name = '+{branch}:refs/heads/{branch}'.format( |
|
790 | 790 | branch=branch_name) |
|
791 | 791 | |
|
792 | 792 | cmd = ['fetch', '--no-tags', '--update-head-ok', |
|
793 | 793 | repository_path, branch_name] |
|
794 | 794 | self.run_git_command(cmd, fail_on_stderr=False) |
|
795 | 795 | |
|
796 | 796 | def _local_reset(self, branch_name): |
|
797 | 797 | branch_name = f'{branch_name}' |
|
798 | 798 | cmd = ['reset', '--hard', branch_name, '--'] |
|
799 | 799 | self.run_git_command(cmd, fail_on_stderr=False) |
|
800 | 800 | |
|
801 | 801 | def _last_fetch_heads(self): |
|
802 | 802 | """ |
|
803 | 803 | Return the last fetched heads that need merging. |
|
804 | 804 | |
|
805 | 805 | The algorithm is defined at |
|
806 | 806 | https://github.com/git/git/blob/v2.1.3/git-pull.sh#L283 |
|
807 | 807 | """ |
|
808 | 808 | if not self.bare: |
|
809 | 809 | fetch_heads_path = os.path.join(self.path, '.git', 'FETCH_HEAD') |
|
810 | 810 | else: |
|
811 | 811 | fetch_heads_path = os.path.join(self.path, 'FETCH_HEAD') |
|
812 | 812 | |
|
813 | 813 | heads = [] |
|
814 | 814 | with open(fetch_heads_path) as f: |
|
815 | 815 | for line in f: |
|
816 | 816 | if ' not-for-merge ' in line: |
|
817 | 817 | continue |
|
818 | 818 | line = re.sub('\t.*', '', line, flags=re.DOTALL) |
|
819 | 819 | heads.append(line) |
|
820 | 820 | |
|
821 | 821 | return heads |
|
822 | 822 | |
|
823 | 823 | def get_shadow_instance(self, shadow_repository_path, enable_hooks=False, cache=False): |
|
824 | 824 | return GitRepository(shadow_repository_path, with_wire={"cache": cache}) |
|
825 | 825 | |
|
826 | 826 | def _local_pull(self, repository_path, branch_name, ff_only=True): |
|
827 | 827 | """ |
|
828 | 828 | Pull a branch from a local repository. |
|
829 | 829 | """ |
|
830 | 830 | if self.bare: |
|
831 | 831 | raise RepositoryError('Cannot pull into a bare git repository') |
|
832 | 832 | # N.B.(skreft): The --ff-only option is to make sure this is a |
|
833 | 833 | # fast-forward (i.e., we are only pulling new changes and there are no |
|
834 | 834 | # conflicts with our current branch) |
|
835 | 835 | # Additionally, that option needs to go before --no-tags, otherwise git |
|
836 | 836 | # pull complains about it being an unknown flag. |
|
837 | 837 | cmd = ['pull'] |
|
838 | 838 | if ff_only: |
|
839 | 839 | cmd.append('--ff-only') |
|
840 | 840 | cmd.extend(['--no-tags', repository_path, branch_name]) |
|
841 | 841 | self.run_git_command(cmd, fail_on_stderr=False) |
|
842 | 842 | |
|
843 | 843 | def _local_merge(self, merge_message, user_name, user_email, heads): |
|
844 | 844 | """ |
|
845 | 845 | Merge the given head into the checked out branch. |
|
846 | 846 | |
|
847 | 847 | It will force a merge commit. |
|
848 | 848 | |
|
849 | 849 | Currently it raises an error if the repo is empty, as it is not possible |
|
850 | 850 | to create a merge commit in an empty repo. |
|
851 | 851 | |
|
852 | 852 | :param merge_message: The message to use for the merge commit. |
|
853 | 853 | :param heads: the heads to merge. |
|
854 | 854 | """ |
|
855 | 855 | if self.bare: |
|
856 | 856 | raise RepositoryError('Cannot merge into a bare git repository') |
|
857 | 857 | |
|
858 | 858 | if not heads: |
|
859 | 859 | return |
|
860 | 860 | |
|
861 | 861 | if self.is_empty(): |
|
862 | 862 | # TODO(skreft): do something more robust in this case. |
|
863 | 863 | raise RepositoryError('Do not know how to merge into empty repositories yet') |
|
864 | 864 | unresolved = None |
|
865 | 865 | |
|
866 | 866 | # N.B.(skreft): the --no-ff option is used to enforce the creation of a |
|
867 | 867 | # commit message. We also specify the user who is doing the merge. |
|
868 | 868 | cmd = ['-c', f'user.name="{user_name}"', |
|
869 | 869 | '-c', f'user.email={user_email}', |
|
870 | 870 | 'merge', '--no-ff', '-m', safe_str(merge_message)] |
|
871 | 871 | |
|
872 | 872 | merge_cmd = cmd + heads |
|
873 | 873 | |
|
874 | 874 | try: |
|
875 | 875 | self.run_git_command(merge_cmd, fail_on_stderr=False) |
|
876 | 876 | except RepositoryError: |
|
877 | 877 | files = self.run_git_command(['diff', '--name-only', '--diff-filter', 'U'], |
|
878 | 878 | fail_on_stderr=False)[0].splitlines() |
|
879 | 879 | # NOTE(marcink): we add U notation for consistent with HG backend output |
|
880 | 880 | unresolved = [f'U {f}' for f in files] |
|
881 | 881 | |
|
882 | 882 | # Cleanup any merge leftovers |
|
883 | 883 | self._remote.invalidate_vcs_cache() |
|
884 | 884 | self.run_git_command(['merge', '--abort'], fail_on_stderr=False) |
|
885 | 885 | |
|
886 | 886 | if unresolved: |
|
887 | 887 | raise UnresolvedFilesInRepo(unresolved) |
|
888 | 888 | else: |
|
889 | 889 | raise |
|
890 | 890 | |
|
891 | 891 | def _local_push( |
|
892 | 892 | self, source_branch, repository_path, target_branch, |
|
893 | 893 | enable_hooks=False, rc_scm_data=None): |
|
894 | 894 | """ |
|
895 | 895 | Push the source_branch to the given repository and target_branch. |
|
896 | 896 | |
|
897 | 897 | Currently it if the target_branch is not master and the target repo is |
|
898 | 898 | empty, the push will work, but then GitRepository won't be able to find |
|
899 | 899 | the pushed branch or the commits. As the HEAD will be corrupted (i.e., |
|
900 | 900 | pointing to master, which does not exist). |
|
901 | 901 | |
|
902 | 902 | It does not run the hooks in the target repo. |
|
903 | 903 | """ |
|
904 | 904 | # TODO(skreft): deal with the case in which the target repo is empty, |
|
905 | 905 | # and the target_branch is not master. |
|
906 | 906 | target_repo = GitRepository(repository_path) |
|
907 | 907 | if (not target_repo.bare and |
|
908 | 908 | target_repo._current_branch() == target_branch): |
|
909 | 909 | # Git prevents pushing to the checked out branch, so simulate it by |
|
910 | 910 | # pulling into the target repository. |
|
911 | 911 | target_repo._local_pull(self.path, source_branch) |
|
912 | 912 | else: |
|
913 | 913 | cmd = ['push', os.path.abspath(repository_path), |
|
914 | 914 | f'{source_branch}:{target_branch}'] |
|
915 | 915 | gitenv = {} |
|
916 | 916 | if rc_scm_data: |
|
917 | 917 | gitenv.update({'RC_SCM_DATA': rc_scm_data}) |
|
918 | 918 | |
|
919 | 919 | if not enable_hooks: |
|
920 | 920 | gitenv['RC_SKIP_HOOKS'] = '1' |
|
921 | 921 | self.run_git_command(cmd, fail_on_stderr=False, extra_env=gitenv) |
|
922 | 922 | |
|
923 | 923 | def _get_new_pr_branch(self, source_branch, target_branch): |
|
924 | 924 | prefix = f'pr_{source_branch}-{target_branch}_' |
|
925 | 925 | pr_branches = [] |
|
926 | 926 | for branch in self.branches: |
|
927 | 927 | if branch.startswith(prefix): |
|
928 | 928 | pr_branches.append(int(branch[len(prefix):])) |
|
929 | 929 | |
|
930 | 930 | if not pr_branches: |
|
931 | 931 | branch_id = 0 |
|
932 | 932 | else: |
|
933 | 933 | branch_id = max(pr_branches) + 1 |
|
934 | 934 | |
|
935 | 935 | return '%s%d' % (prefix, branch_id) |
|
936 | 936 | |
|
937 | 937 | def _maybe_prepare_merge_workspace( |
|
938 | 938 | self, repo_id, workspace_id, target_ref, source_ref): |
|
939 | 939 | shadow_repository_path = self._get_shadow_repository_path( |
|
940 | 940 | self.path, repo_id, workspace_id) |
|
941 | 941 | if not os.path.exists(shadow_repository_path): |
|
942 | 942 | self._local_clone( |
|
943 | 943 | shadow_repository_path, target_ref.name, source_ref.name) |
|
944 | 944 | log.debug('Prepared %s shadow repository in %s', |
|
945 | 945 | self.alias, shadow_repository_path) |
|
946 | 946 | |
|
947 | 947 | return shadow_repository_path |
|
948 | 948 | |
|
949 | 949 | def _merge_repo(self, repo_id, workspace_id, target_ref, |
|
950 | 950 | source_repo, source_ref, merge_message, |
|
951 | 951 | merger_name, merger_email, dry_run=False, |
|
952 | 952 | use_rebase=False, close_branch=False): |
|
953 | 953 | |
|
954 | 954 | log.debug('Executing merge_repo with %s strategy, dry_run mode:%s', |
|
955 | 955 | 'rebase' if use_rebase else 'merge', dry_run) |
|
956 | 956 | |
|
957 | 957 | if target_ref.commit_id != self.branches[target_ref.name]: |
|
958 | 958 | log.warning('Target ref %s commit mismatch %s vs %s', target_ref, |
|
959 | 959 | target_ref.commit_id, self.branches[target_ref.name]) |
|
960 | 960 | return MergeResponse( |
|
961 | 961 | False, False, None, MergeFailureReason.TARGET_IS_NOT_HEAD, |
|
962 | 962 | metadata={'target_ref': target_ref}) |
|
963 | 963 | |
|
964 | 964 | shadow_repository_path = self._maybe_prepare_merge_workspace( |
|
965 | 965 | repo_id, workspace_id, target_ref, source_ref) |
|
966 | 966 | shadow_repo = self.get_shadow_instance(shadow_repository_path) |
|
967 | 967 | |
|
968 | 968 | # checkout source, if it's different. Otherwise we could not |
|
969 | 969 | # fetch proper commits for merge testing |
|
970 | 970 | if source_ref.name != target_ref.name: |
|
971 | 971 | if shadow_repo.get_remote_ref(source_ref.name): |
|
972 | 972 | shadow_repo._checkout(source_ref.name, force=True) |
|
973 | 973 | |
|
974 | 974 | # checkout target, and fetch changes |
|
975 | 975 | shadow_repo._checkout(target_ref.name, force=True) |
|
976 | 976 | |
|
977 | 977 | # fetch/reset pull the target, in case it is changed |
|
978 | 978 | # this handles even force changes |
|
979 | 979 | shadow_repo._local_fetch(self.path, target_ref.name, use_origin=True) |
|
980 | 980 | shadow_repo._local_reset(target_ref.name) |
|
981 | 981 | |
|
982 | 982 | # Need to reload repo to invalidate the cache, or otherwise we cannot |
|
983 | 983 | # retrieve the last target commit. |
|
984 | 984 | shadow_repo = self.get_shadow_instance(shadow_repository_path) |
|
985 | 985 | if target_ref.commit_id != shadow_repo.branches[target_ref.name]: |
|
986 | 986 | log.warning('Shadow Target ref %s commit mismatch %s vs %s', |
|
987 | 987 | target_ref, target_ref.commit_id, |
|
988 | 988 | shadow_repo.branches[target_ref.name]) |
|
989 | 989 | return MergeResponse( |
|
990 | 990 | False, False, None, MergeFailureReason.TARGET_IS_NOT_HEAD, |
|
991 | 991 | metadata={'target_ref': target_ref}) |
|
992 | 992 | |
|
993 | 993 | # calculate new branch |
|
994 | 994 | pr_branch = shadow_repo._get_new_pr_branch( |
|
995 | 995 | source_ref.name, target_ref.name) |
|
996 | 996 | log.debug('using pull-request merge branch: `%s`', pr_branch) |
|
997 | 997 | # checkout to temp branch, and fetch changes |
|
998 | 998 | shadow_repo._checkout(pr_branch, create=True) |
|
999 | 999 | try: |
|
1000 | 1000 | shadow_repo._local_fetch(source_repo.path, source_ref.name) |
|
1001 | 1001 | except RepositoryError: |
|
1002 | 1002 | log.exception('Failure when doing local fetch on ' |
|
1003 | 1003 | 'shadow repo: %s', shadow_repo) |
|
1004 | 1004 | return MergeResponse( |
|
1005 | 1005 | False, False, None, MergeFailureReason.MISSING_SOURCE_REF, |
|
1006 | 1006 | metadata={'source_ref': source_ref}) |
|
1007 | 1007 | |
|
1008 | 1008 | merge_ref = None |
|
1009 | 1009 | merge_failure_reason = MergeFailureReason.NONE |
|
1010 | 1010 | metadata = {} |
|
1011 | 1011 | try: |
|
1012 | 1012 | shadow_repo._local_merge(merge_message, merger_name, merger_email, |
|
1013 | 1013 | [source_ref.commit_id]) |
|
1014 | 1014 | merge_possible = True |
|
1015 | 1015 | |
|
1016 | 1016 | # Need to invalidate the cache, or otherwise we |
|
1017 | 1017 | # cannot retrieve the merge commit. |
|
1018 | 1018 | shadow_repo = shadow_repo.get_shadow_instance(shadow_repository_path) |
|
1019 | 1019 | merge_commit_id = shadow_repo.branches[pr_branch] |
|
1020 | 1020 | |
|
1021 | 1021 | # Set a reference pointing to the merge commit. This reference may |
|
1022 | 1022 | # be used to easily identify the last successful merge commit in |
|
1023 | 1023 | # the shadow repository. |
|
1024 | 1024 | shadow_repo.set_refs('refs/heads/pr-merge', merge_commit_id) |
|
1025 | 1025 | merge_ref = Reference('branch', 'pr-merge', merge_commit_id) |
|
1026 | 1026 | except RepositoryError as e: |
|
1027 | 1027 | log.exception('Failure when doing local merge on git shadow repo') |
|
1028 | 1028 | if isinstance(e, UnresolvedFilesInRepo): |
|
1029 | 1029 | metadata['unresolved_files'] = '\n* conflict: ' + ('\n * conflict: '.join(e.args[0])) |
|
1030 | 1030 | |
|
1031 | 1031 | merge_possible = False |
|
1032 | 1032 | merge_failure_reason = MergeFailureReason.MERGE_FAILED |
|
1033 | 1033 | |
|
1034 | 1034 | if merge_possible and not dry_run: |
|
1035 | 1035 | try: |
|
1036 | 1036 | shadow_repo._local_push( |
|
1037 | 1037 | pr_branch, self.path, target_ref.name, enable_hooks=True, |
|
1038 | 1038 | rc_scm_data=self.config.get('rhodecode', 'RC_SCM_DATA')) |
|
1039 | 1039 | merge_succeeded = True |
|
1040 | 1040 | except RepositoryError: |
|
1041 | 1041 | log.exception( |
|
1042 | 1042 | 'Failure when doing local push from the shadow ' |
|
1043 | 1043 | 'repository to the target repository at %s.', self.path) |
|
1044 | 1044 | merge_succeeded = False |
|
1045 | 1045 | merge_failure_reason = MergeFailureReason.PUSH_FAILED |
|
1046 | 1046 | metadata['target'] = 'git shadow repo' |
|
1047 | 1047 | metadata['merge_commit'] = pr_branch |
|
1048 | 1048 | else: |
|
1049 | 1049 | merge_succeeded = False |
|
1050 | 1050 | |
|
1051 | 1051 | return MergeResponse( |
|
1052 | 1052 | merge_possible, merge_succeeded, merge_ref, merge_failure_reason, |
|
1053 | 1053 | metadata=metadata) |
@@ -1,1024 +1,1024 b'' | |||
|
1 | 1 | # Copyright (C) 2014-2023 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 | """ |
|
20 | 20 | HG repository module |
|
21 | 21 | """ |
|
22 | 22 | import os |
|
23 | 23 | import logging |
|
24 | 24 | import binascii |
|
25 | 25 | import configparser |
|
26 | 26 | import urllib.request |
|
27 | 27 | import urllib.parse |
|
28 | 28 | import urllib.error |
|
29 | 29 | |
|
30 | 30 | from zope.cachedescriptors.property import Lazy as LazyProperty |
|
31 | 31 | |
|
32 | 32 | from collections import OrderedDict |
|
33 | 33 | from rhodecode.lib.datelib import ( |
|
34 | 34 | date_to_timestamp_plus_offset, utcdate_fromtimestamp, makedate) |
|
35 | 35 | from rhodecode.lib.str_utils import safe_str |
|
36 | 36 | from rhodecode.lib.utils2 import CachedProperty |
|
37 | 37 | from rhodecode.lib.vcs import connection, exceptions |
|
38 | 38 | from rhodecode.lib.vcs.backends.base import ( |
|
39 | 39 | BaseRepository, CollectionGenerator, Config, MergeResponse, |
|
40 | 40 | MergeFailureReason, Reference, BasePathPermissionChecker) |
|
41 | 41 | from rhodecode.lib.vcs.backends.hg.commit import MercurialCommit |
|
42 | 42 | from rhodecode.lib.vcs.backends.hg.diff import MercurialDiff |
|
43 | 43 | from rhodecode.lib.vcs.backends.hg.inmemory import MercurialInMemoryCommit |
|
44 | 44 | from rhodecode.lib.vcs.exceptions import ( |
|
45 | 45 | EmptyRepositoryError, RepositoryError, TagAlreadyExistError, |
|
46 | 46 | TagDoesNotExistError, CommitDoesNotExistError, SubrepoMergeError, UnresolvedFilesInRepo) |
|
47 | 47 | |
|
48 | 48 | hexlify = binascii.hexlify |
|
49 | 49 | nullid = "\0" * 20 |
|
50 | 50 | |
|
51 | 51 | log = logging.getLogger(__name__) |
|
52 | 52 | |
|
53 | 53 | |
|
54 | 54 | class MercurialRepository(BaseRepository): |
|
55 | 55 | """ |
|
56 | 56 | Mercurial repository backend |
|
57 | 57 | """ |
|
58 | 58 | DEFAULT_BRANCH_NAME = 'default' |
|
59 | 59 | |
|
60 | 60 | def __init__(self, repo_path, config=None, create=False, src_url=None, |
|
61 | 61 | do_workspace_checkout=False, with_wire=None, bare=False): |
|
62 | 62 | """ |
|
63 | 63 | Raises RepositoryError if repository could not be find at the given |
|
64 | 64 | ``repo_path``. |
|
65 | 65 | |
|
66 | 66 | :param repo_path: local path of the repository |
|
67 | 67 | :param config: config object containing the repo configuration |
|
68 | 68 | :param create=False: if set to True, would try to create repository if |
|
69 | 69 | it does not exist rather than raising exception |
|
70 | 70 | :param src_url=None: would try to clone repository from given location |
|
71 | 71 | :param do_workspace_checkout=False: sets update of working copy after |
|
72 | 72 | making a clone |
|
73 | 73 | :param bare: not used, compatible with other VCS |
|
74 | 74 | """ |
|
75 | 75 | |
|
76 | 76 | self.path = safe_str(os.path.abspath(repo_path)) |
|
77 | 77 | # mercurial since 4.4.X requires certain configuration to be present |
|
78 | 78 | # because sometimes we init the repos with config we need to meet |
|
79 | 79 | # special requirements |
|
80 | 80 | self.config = config if config else self.get_default_config( |
|
81 | 81 | default=[('extensions', 'largefiles', '')]) |
|
82 | 82 | |
|
83 | 83 | # NOTE(marcink): since python3 hgsubversion is deprecated. |
|
84 | 84 | # From old installations we might still have this set enabled |
|
85 | 85 | # we explicitly remove this now here to make sure it wont propagate further |
|
86 | 86 | if config and config.get('extensions', 'hgsubversion') is not None: |
|
87 | 87 | config.drop_option('extensions', 'hgsubversion') |
|
88 | 88 | |
|
89 | 89 | self.with_wire = with_wire or {"cache": False} # default should not use cache |
|
90 | 90 | |
|
91 | 91 | self._init_repo(create, src_url, do_workspace_checkout) |
|
92 | 92 | |
|
93 | 93 | # caches |
|
94 | 94 | self._commit_ids = {} |
|
95 | 95 | |
|
96 | 96 | @LazyProperty |
|
97 | 97 | def _remote(self): |
|
98 | 98 | repo_id = self.path |
|
99 | 99 | return connection.Hg(self.path, repo_id, self.config, with_wire=self.with_wire) |
|
100 | 100 | |
|
101 | 101 | @CachedProperty |
|
102 | 102 | def commit_ids(self): |
|
103 | 103 | """ |
|
104 | 104 | Returns list of commit ids, in ascending order. Being lazy |
|
105 | 105 | attribute allows external tools to inject shas from cache. |
|
106 | 106 | """ |
|
107 | 107 | commit_ids = self._get_all_commit_ids() |
|
108 | 108 | self._rebuild_cache(commit_ids) |
|
109 | 109 | return commit_ids |
|
110 | 110 | |
|
111 | 111 | def _rebuild_cache(self, commit_ids): |
|
112 | 112 | self._commit_ids = {commit_id: index |
|
113 | 113 | for index, commit_id in enumerate(commit_ids)} |
|
114 | 114 | |
|
115 | 115 | @CachedProperty |
|
116 | 116 | def branches(self): |
|
117 | 117 | return self._get_branches() |
|
118 | 118 | |
|
119 | 119 | @CachedProperty |
|
120 | 120 | def branches_closed(self): |
|
121 | 121 | return self._get_branches(active=False, closed=True) |
|
122 | 122 | |
|
123 | 123 | @CachedProperty |
|
124 | 124 | def branches_all(self): |
|
125 | 125 | all_branches = {} |
|
126 | 126 | all_branches.update(self.branches) |
|
127 | 127 | all_branches.update(self.branches_closed) |
|
128 | 128 | return all_branches |
|
129 | 129 | |
|
130 | 130 | def _get_branches(self, active=True, closed=False): |
|
131 | 131 | """ |
|
132 | 132 | Gets branches for this repository |
|
133 | 133 | Returns only not closed active branches by default |
|
134 | 134 | |
|
135 | 135 | :param active: return also active branches |
|
136 | 136 | :param closed: return also closed branches |
|
137 | 137 | |
|
138 | 138 | """ |
|
139 | 139 | if self.is_empty(): |
|
140 | 140 | return {} |
|
141 | 141 | |
|
142 | 142 | def get_name(ctx): |
|
143 | 143 | return ctx[0] |
|
144 | 144 | |
|
145 | 145 | _branches = [(n, h,) for n, h in |
|
146 | 146 | self._remote.branches(active, closed).items()] |
|
147 | 147 | |
|
148 | 148 | return OrderedDict(sorted(_branches, key=get_name, reverse=False)) |
|
149 | 149 | |
|
150 | 150 | @CachedProperty |
|
151 | 151 | def tags(self): |
|
152 | 152 | """ |
|
153 | 153 | Gets tags for this repository |
|
154 | 154 | """ |
|
155 | 155 | return self._get_tags() |
|
156 | 156 | |
|
157 | 157 | def _get_tags(self): |
|
158 | 158 | if self.is_empty(): |
|
159 | 159 | return {} |
|
160 | 160 | |
|
161 | 161 | def get_name(ctx): |
|
162 | 162 | return ctx[0] |
|
163 | 163 | |
|
164 | 164 | _tags = [(n, h,) for n, h in |
|
165 | 165 | self._remote.tags().items()] |
|
166 | 166 | |
|
167 | 167 | return OrderedDict(sorted(_tags, key=get_name, reverse=True)) |
|
168 | 168 | |
|
169 | 169 | def tag(self, name, user, commit_id=None, message=None, date=None, **kwargs): |
|
170 | 170 | """ |
|
171 | 171 | Creates and returns a tag for the given ``commit_id``. |
|
172 | 172 | |
|
173 | 173 | :param name: name for new tag |
|
174 | 174 | :param user: full username, i.e.: "Joe Doe <joe.doe@example.com>" |
|
175 | 175 | :param commit_id: commit id for which new tag would be created |
|
176 | 176 | :param message: message of the tag's commit |
|
177 | 177 | :param date: date of tag's commit |
|
178 | 178 | |
|
179 | 179 | :raises TagAlreadyExistError: if tag with same name already exists |
|
180 | 180 | """ |
|
181 | 181 | if name in self.tags: |
|
182 | 182 | raise TagAlreadyExistError("Tag %s already exists" % name) |
|
183 | 183 | |
|
184 | 184 | commit = self.get_commit(commit_id=commit_id) |
|
185 | 185 | local = kwargs.setdefault('local', False) |
|
186 | 186 | |
|
187 | 187 | if message is None: |
|
188 | 188 | message = f"Added tag {name} for commit {commit.short_id}" |
|
189 | 189 | |
|
190 | 190 | date, tz = date_to_timestamp_plus_offset(date) |
|
191 | 191 | |
|
192 | 192 | self._remote.tag(name, commit.raw_id, message, local, user, date, tz) |
|
193 | 193 | self._remote.invalidate_vcs_cache() |
|
194 | 194 | |
|
195 | 195 | # Reinitialize tags |
|
196 | 196 | self._invalidate_prop_cache('tags') |
|
197 | 197 | tag_id = self.tags[name] |
|
198 | 198 | |
|
199 | 199 | return self.get_commit(commit_id=tag_id) |
|
200 | 200 | |
|
201 | 201 | def remove_tag(self, name, user, message=None, date=None): |
|
202 | 202 | """ |
|
203 | 203 | Removes tag with the given `name`. |
|
204 | 204 | |
|
205 | 205 | :param name: name of the tag to be removed |
|
206 | 206 | :param user: full username, i.e.: "Joe Doe <joe.doe@example.com>" |
|
207 | 207 | :param message: message of the tag's removal commit |
|
208 | 208 | :param date: date of tag's removal commit |
|
209 | 209 | |
|
210 | 210 | :raises TagDoesNotExistError: if tag with given name does not exists |
|
211 | 211 | """ |
|
212 | 212 | if name not in self.tags: |
|
213 | 213 | raise TagDoesNotExistError("Tag %s does not exist" % name) |
|
214 | 214 | |
|
215 | 215 | if message is None: |
|
216 | 216 | message = "Removed tag %s" % name |
|
217 | 217 | local = False |
|
218 | 218 | |
|
219 | 219 | date, tz = date_to_timestamp_plus_offset(date) |
|
220 | 220 | |
|
221 | 221 | self._remote.tag(name, nullid, message, local, user, date, tz) |
|
222 | 222 | self._remote.invalidate_vcs_cache() |
|
223 | 223 | self._invalidate_prop_cache('tags') |
|
224 | 224 | |
|
225 | 225 | @LazyProperty |
|
226 | 226 | def bookmarks(self): |
|
227 | 227 | """ |
|
228 | 228 | Gets bookmarks for this repository |
|
229 | 229 | """ |
|
230 | 230 | return self._get_bookmarks() |
|
231 | 231 | |
|
232 | 232 | def _get_bookmarks(self): |
|
233 | 233 | if self.is_empty(): |
|
234 | 234 | return {} |
|
235 | 235 | |
|
236 | 236 | def get_name(ctx): |
|
237 | 237 | return ctx[0] |
|
238 | 238 | |
|
239 | 239 | _bookmarks = [ |
|
240 | 240 | (n, h) for n, h in |
|
241 | 241 | self._remote.bookmarks().items()] |
|
242 | 242 | |
|
243 | 243 | return OrderedDict(sorted(_bookmarks, key=get_name)) |
|
244 | 244 | |
|
245 | 245 | def _get_all_commit_ids(self): |
|
246 | 246 | return self._remote.get_all_commit_ids('visible') |
|
247 | 247 | |
|
248 | 248 | def get_diff( |
|
249 | 249 | self, commit1, commit2, path='', ignore_whitespace=False, |
|
250 | 250 | context=3, path1=None): |
|
251 | 251 | """ |
|
252 | 252 | Returns (git like) *diff*, as plain text. Shows changes introduced by |
|
253 | 253 | `commit2` since `commit1`. |
|
254 | 254 | |
|
255 | 255 | :param commit1: Entry point from which diff is shown. Can be |
|
256 | 256 | ``self.EMPTY_COMMIT`` - in this case, patch showing all |
|
257 | 257 | the changes since empty state of the repository until `commit2` |
|
258 | 258 | :param commit2: Until which commit changes should be shown. |
|
259 | 259 | :param ignore_whitespace: If set to ``True``, would not show whitespace |
|
260 | 260 | changes. Defaults to ``False``. |
|
261 | 261 | :param context: How many lines before/after changed lines should be |
|
262 | 262 | shown. Defaults to ``3``. |
|
263 | 263 | """ |
|
264 | 264 | self._validate_diff_commits(commit1, commit2) |
|
265 | 265 | if path1 is not None and path1 != path: |
|
266 | 266 | raise ValueError("Diff of two different paths not supported.") |
|
267 | 267 | |
|
268 | 268 | if path: |
|
269 | 269 | file_filter = [self.path, path] |
|
270 | 270 | else: |
|
271 | 271 | file_filter = None |
|
272 | 272 | |
|
273 | 273 | diff = self._remote.diff( |
|
274 | 274 | commit1.raw_id, commit2.raw_id, file_filter=file_filter, |
|
275 | 275 | opt_git=True, opt_ignorews=ignore_whitespace, |
|
276 | 276 | context=context) |
|
277 | 277 | return MercurialDiff(diff) |
|
278 | 278 | |
|
279 | 279 | def strip(self, commit_id, branch=None): |
|
280 | 280 | self._remote.strip(commit_id, update=False, backup=False) |
|
281 | 281 | |
|
282 | 282 | self._remote.invalidate_vcs_cache() |
|
283 | 283 | # clear cache |
|
284 | 284 | self._invalidate_prop_cache('commit_ids') |
|
285 | 285 | |
|
286 | 286 | return len(self.commit_ids) |
|
287 | 287 | |
|
288 | 288 | def verify(self): |
|
289 | 289 | verify = self._remote.verify() |
|
290 | 290 | |
|
291 | 291 | self._remote.invalidate_vcs_cache() |
|
292 | 292 | return verify |
|
293 | 293 | |
|
294 | 294 | def hg_update_cache(self): |
|
295 | 295 | update_cache = self._remote.hg_update_cache() |
|
296 | 296 | |
|
297 | 297 | self._remote.invalidate_vcs_cache() |
|
298 | 298 | return update_cache |
|
299 | 299 | |
|
300 | 300 | def hg_rebuild_fn_cache(self): |
|
301 | 301 | update_cache = self._remote.hg_rebuild_fn_cache() |
|
302 | 302 | |
|
303 | 303 | self._remote.invalidate_vcs_cache() |
|
304 | 304 | return update_cache |
|
305 | 305 | |
|
306 | 306 | def get_common_ancestor(self, commit_id1, commit_id2, repo2): |
|
307 | 307 | log.debug('Calculating common ancestor between %sc1:%s and %sc2:%s', |
|
308 | 308 | self, commit_id1, repo2, commit_id2) |
|
309 | 309 | |
|
310 | 310 | if commit_id1 == commit_id2: |
|
311 | 311 | return commit_id1 |
|
312 | 312 | |
|
313 | 313 | ancestors = self._remote.revs_from_revspec( |
|
314 | 314 | "ancestor(id(%s), id(%s))", commit_id1, commit_id2, |
|
315 | 315 | other_path=repo2.path) |
|
316 | 316 | |
|
317 | 317 | ancestor_id = repo2[ancestors[0]].raw_id if ancestors else None |
|
318 | 318 | |
|
319 | 319 | log.debug('Found common ancestor with sha: %s', ancestor_id) |
|
320 | 320 | return ancestor_id |
|
321 | 321 | |
|
322 | 322 | def compare(self, commit_id1, commit_id2, repo2, merge, pre_load=None): |
|
323 | 323 | if commit_id1 == commit_id2: |
|
324 | 324 | commits = [] |
|
325 | 325 | else: |
|
326 | 326 | if merge: |
|
327 | 327 | indexes = self._remote.revs_from_revspec( |
|
328 | 328 | "ancestors(id(%s)) - ancestors(id(%s)) - id(%s)", |
|
329 | 329 | commit_id2, commit_id1, commit_id1, other_path=repo2.path) |
|
330 | 330 | else: |
|
331 | 331 | indexes = self._remote.revs_from_revspec( |
|
332 | 332 | "id(%s)..id(%s) - id(%s)", commit_id1, commit_id2, |
|
333 | 333 | commit_id1, other_path=repo2.path) |
|
334 | 334 | |
|
335 | 335 | commits = [repo2.get_commit(commit_idx=idx, pre_load=pre_load) |
|
336 | 336 | for idx in indexes] |
|
337 | 337 | |
|
338 | 338 | return commits |
|
339 | 339 | |
|
340 | 340 | @staticmethod |
|
341 | 341 | def check_url(url, config): |
|
342 | 342 | """ |
|
343 | 343 | Function will check given url and try to verify if it's a valid |
|
344 | 344 | link. Sometimes it may happened that mercurial will issue basic |
|
345 | 345 | auth request that can cause whole API to hang when used from python |
|
346 | 346 | or other external calls. |
|
347 | 347 | |
|
348 | 348 | On failures it'll raise urllib2.HTTPError, exception is also thrown |
|
349 | 349 | when the return code is non 200 |
|
350 | 350 | """ |
|
351 | 351 | # check first if it's not an local url |
|
352 | 352 | if os.path.isdir(url) or url.startswith('file:'): |
|
353 | 353 | return True |
|
354 | 354 | |
|
355 | 355 | # Request the _remote to verify the url |
|
356 | 356 | return connection.Hg.check_url(url, config.serialize()) |
|
357 | 357 | |
|
358 | 358 | @staticmethod |
|
359 | 359 | def is_valid_repository(path): |
|
360 | 360 | return os.path.isdir(os.path.join(path, '.hg')) |
|
361 | 361 | |
|
362 | 362 | def _init_repo(self, create, src_url=None, do_workspace_checkout=False): |
|
363 | 363 | """ |
|
364 | 364 | Function will check for mercurial repository in given path. If there |
|
365 | 365 | is no repository in that path it will raise an exception unless |
|
366 | 366 | `create` parameter is set to True - in that case repository would |
|
367 | 367 | be created. |
|
368 | 368 | |
|
369 | 369 | If `src_url` is given, would try to clone repository from the |
|
370 | 370 | location at given clone_point. Additionally it'll make update to |
|
371 | 371 | working copy accordingly to `do_workspace_checkout` flag. |
|
372 | 372 | """ |
|
373 | 373 | if create and os.path.exists(self.path): |
|
374 | 374 | raise RepositoryError( |
|
375 | 375 | f"Cannot create repository at {self.path}, location already exist") |
|
376 | 376 | |
|
377 | 377 | if src_url: |
|
378 | 378 | url = str(self._get_url(src_url)) |
|
379 | 379 | MercurialRepository.check_url(url, self.config) |
|
380 | 380 | |
|
381 | 381 | self._remote.clone(url, self.path, do_workspace_checkout) |
|
382 | 382 | |
|
383 | 383 | # Don't try to create if we've already cloned repo |
|
384 | 384 | create = False |
|
385 | 385 | |
|
386 | 386 | if create: |
|
387 | 387 | os.makedirs(self.path, mode=0o755) |
|
388 | 388 | |
|
389 | 389 | self._remote.localrepository(create) |
|
390 | 390 | |
|
391 | 391 | @LazyProperty |
|
392 | 392 | def in_memory_commit(self): |
|
393 | 393 | return MercurialInMemoryCommit(self) |
|
394 | 394 | |
|
395 | 395 | @LazyProperty |
|
396 | 396 | def description(self): |
|
397 | 397 | description = self._remote.get_config_value( |
|
398 | 398 | 'web', 'description', untrusted=True) |
|
399 | 399 | return safe_str(description or self.DEFAULT_DESCRIPTION) |
|
400 | 400 | |
|
401 | 401 | @LazyProperty |
|
402 | 402 | def contact(self): |
|
403 | 403 | contact = ( |
|
404 | 404 | self._remote.get_config_value("web", "contact") or |
|
405 | 405 | self._remote.get_config_value("ui", "username")) |
|
406 | 406 | return safe_str(contact or self.DEFAULT_CONTACT) |
|
407 | 407 | |
|
408 | 408 | @LazyProperty |
|
409 | 409 | def last_change(self): |
|
410 | 410 | """ |
|
411 | 411 | Returns last change made on this repository as |
|
412 | 412 | `datetime.datetime` object. |
|
413 | 413 | """ |
|
414 | 414 | try: |
|
415 | 415 | return self.get_commit().date |
|
416 | 416 | except RepositoryError: |
|
417 | 417 | tzoffset = makedate()[1] |
|
418 | 418 | return utcdate_fromtimestamp(self._get_fs_mtime(), tzoffset) |
|
419 | 419 | |
|
420 | 420 | def _get_fs_mtime(self): |
|
421 | 421 | # fallback to filesystem |
|
422 | 422 | cl_path = os.path.join(self.path, '.hg', "00changelog.i") |
|
423 | 423 | st_path = os.path.join(self.path, '.hg', "store") |
|
424 | 424 | if os.path.exists(cl_path): |
|
425 | 425 | return os.stat(cl_path).st_mtime |
|
426 | 426 | else: |
|
427 | 427 | return os.stat(st_path).st_mtime |
|
428 | 428 | |
|
429 | 429 | def _get_url(self, url): |
|
430 | 430 | """ |
|
431 | 431 | Returns normalized url. If schema is not given, would fall |
|
432 | 432 | to filesystem |
|
433 | 433 | (``file:///``) schema. |
|
434 | 434 | """ |
|
435 | 435 | if url != 'default' and '://' not in url: |
|
436 | 436 | url = "file:" + urllib.request.pathname2url(url) |
|
437 | 437 | return url |
|
438 | 438 | |
|
439 | 439 | def get_hook_location(self): |
|
440 | 440 | """ |
|
441 | 441 | returns absolute path to location where hooks are stored |
|
442 | 442 | """ |
|
443 | 443 | return os.path.join(self.path, '.hg', '.hgrc') |
|
444 | 444 | |
|
445 | 445 | def get_commit(self, commit_id=None, commit_idx=None, pre_load=None, |
|
446 | 446 | translate_tag=None, maybe_unreachable=False, reference_obj=None): |
|
447 | 447 | """ |
|
448 | 448 | Returns ``MercurialCommit`` object representing repository's |
|
449 | 449 | commit at the given `commit_id` or `commit_idx`. |
|
450 | 450 | """ |
|
451 | 451 | if self.is_empty(): |
|
452 | 452 | raise EmptyRepositoryError("There are no commits yet") |
|
453 | 453 | |
|
454 | 454 | if commit_id is not None: |
|
455 | 455 | self._validate_commit_id(commit_id) |
|
456 | 456 | try: |
|
457 | 457 | # we have cached idx, use it without contacting the remote |
|
458 | 458 | idx = self._commit_ids[commit_id] |
|
459 | 459 | return MercurialCommit(self, commit_id, idx, pre_load=pre_load) |
|
460 | 460 | except KeyError: |
|
461 | 461 | pass |
|
462 | 462 | |
|
463 | 463 | elif commit_idx is not None: |
|
464 | 464 | self._validate_commit_idx(commit_idx) |
|
465 | 465 | try: |
|
466 | 466 | _commit_id = self.commit_ids[commit_idx] |
|
467 | 467 | if commit_idx < 0: |
|
468 | 468 | commit_idx = self.commit_ids.index(_commit_id) |
|
469 | 469 | |
|
470 | 470 | return MercurialCommit(self, _commit_id, commit_idx, pre_load=pre_load) |
|
471 | 471 | except IndexError: |
|
472 | 472 | commit_id = commit_idx |
|
473 | 473 | else: |
|
474 | 474 | commit_id = "tip" |
|
475 | 475 | |
|
476 | 476 | # case here is no cached version, do an actual lookup instead |
|
477 | 477 | try: |
|
478 | 478 | raw_id, idx = self._remote.lookup(commit_id, both=True) |
|
479 | 479 | except CommitDoesNotExistError: |
|
480 | 480 | msg = "Commit {} does not exist for `{}`".format( |
|
481 | 481 | *map(safe_str, [commit_id, self.name])) |
|
482 | 482 | raise CommitDoesNotExistError(msg) |
|
483 | 483 | |
|
484 | 484 | return MercurialCommit(self, raw_id, idx, pre_load=pre_load) |
|
485 | 485 | |
|
486 | 486 | def get_commits( |
|
487 | 487 | self, start_id=None, end_id=None, start_date=None, end_date=None, |
|
488 | 488 | branch_name=None, show_hidden=False, pre_load=None, translate_tags=None): |
|
489 | 489 | """ |
|
490 | 490 | Returns generator of ``MercurialCommit`` objects from start to end |
|
491 | 491 | (both are inclusive) |
|
492 | 492 | |
|
493 | 493 | :param start_id: None, str(commit_id) |
|
494 | 494 | :param end_id: None, str(commit_id) |
|
495 | 495 | :param start_date: if specified, commits with commit date less than |
|
496 | 496 | ``start_date`` would be filtered out from returned set |
|
497 | 497 | :param end_date: if specified, commits with commit date greater than |
|
498 | 498 | ``end_date`` would be filtered out from returned set |
|
499 | 499 | :param branch_name: if specified, commits not reachable from given |
|
500 | 500 | branch would be filtered out from returned set |
|
501 | 501 | :param show_hidden: Show hidden commits such as obsolete or hidden from |
|
502 | 502 | Mercurial evolve |
|
503 | 503 | :raise BranchDoesNotExistError: If given ``branch_name`` does not |
|
504 | 504 | exist. |
|
505 | 505 | :raise CommitDoesNotExistError: If commit for given ``start`` or |
|
506 | 506 | ``end`` could not be found. |
|
507 | 507 | """ |
|
508 | 508 | # actually we should check now if it's not an empty repo |
|
509 | 509 | if self.is_empty(): |
|
510 | 510 | raise EmptyRepositoryError("There are no commits yet") |
|
511 | 511 | self._validate_branch_name(branch_name) |
|
512 | 512 | |
|
513 | 513 | branch_ancestors = False |
|
514 | 514 | if start_id is not None: |
|
515 | 515 | self._validate_commit_id(start_id) |
|
516 | 516 | c_start = self.get_commit(commit_id=start_id) |
|
517 | 517 | start_pos = self._commit_ids[c_start.raw_id] |
|
518 | 518 | else: |
|
519 | 519 | start_pos = None |
|
520 | 520 | |
|
521 | 521 | if end_id is not None: |
|
522 | 522 | self._validate_commit_id(end_id) |
|
523 | 523 | c_end = self.get_commit(commit_id=end_id) |
|
524 | 524 | end_pos = max(0, self._commit_ids[c_end.raw_id]) |
|
525 | 525 | else: |
|
526 | 526 | end_pos = None |
|
527 | 527 | |
|
528 | 528 | if None not in [start_id, end_id] and start_pos > end_pos: |
|
529 | 529 | raise RepositoryError( |
|
530 | 530 | "Start commit '%s' cannot be after end commit '%s'" % |
|
531 | 531 | (start_id, end_id)) |
|
532 | 532 | |
|
533 | 533 | if end_pos is not None: |
|
534 | 534 | end_pos += 1 |
|
535 | 535 | |
|
536 | 536 | commit_filter = [] |
|
537 | 537 | |
|
538 | 538 | if branch_name and not branch_ancestors: |
|
539 | 539 | commit_filter.append(f'branch("{branch_name}")') |
|
540 | 540 | elif branch_name and branch_ancestors: |
|
541 | 541 | commit_filter.append(f'ancestors(branch("{branch_name}"))') |
|
542 | 542 | |
|
543 | 543 | if start_date and not end_date: |
|
544 | 544 | commit_filter.append(f'date(">{start_date}")') |
|
545 | 545 | if end_date and not start_date: |
|
546 | 546 | commit_filter.append(f'date("<{end_date}")') |
|
547 | 547 | if start_date and end_date: |
|
548 | 548 | commit_filter.append( |
|
549 | 549 | f'date(">{start_date}") and date("<{end_date}")') |
|
550 | 550 | |
|
551 | 551 | if not show_hidden: |
|
552 | 552 | commit_filter.append('not obsolete()') |
|
553 | 553 | commit_filter.append('not hidden()') |
|
554 | 554 | |
|
555 | 555 | # TODO: johbo: Figure out a simpler way for this solution |
|
556 | 556 | collection_generator = CollectionGenerator |
|
557 | 557 | if commit_filter: |
|
558 | 558 | commit_filter = ' and '.join(map(safe_str, commit_filter)) |
|
559 | 559 | revisions = self._remote.rev_range([commit_filter]) |
|
560 | 560 | collection_generator = MercurialIndexBasedCollectionGenerator |
|
561 | 561 | else: |
|
562 | 562 | revisions = self.commit_ids |
|
563 | 563 | |
|
564 | 564 | if start_pos or end_pos: |
|
565 | 565 | revisions = revisions[start_pos:end_pos] |
|
566 | 566 | |
|
567 | 567 | return collection_generator(self, revisions, pre_load=pre_load) |
|
568 | 568 | |
|
569 | 569 | def pull(self, url, commit_ids=None): |
|
570 | 570 | """ |
|
571 | 571 | Pull changes from external location. |
|
572 | 572 | |
|
573 | 573 | :param commit_ids: Optional. Can be set to a list of commit ids |
|
574 | 574 | which shall be pulled from the other repository. |
|
575 | 575 | """ |
|
576 | 576 | url = self._get_url(url) |
|
577 | 577 | self._remote.pull(url, commit_ids=commit_ids) |
|
578 | 578 | self._remote.invalidate_vcs_cache() |
|
579 | 579 | |
|
580 | 580 | def fetch(self, url, commit_ids=None, **kwargs): |
|
581 | 581 | """ |
|
582 | 582 | Backward compatibility with GIT fetch==pull |
|
583 | 583 | """ |
|
584 | 584 | return self.pull(url, commit_ids=commit_ids) |
|
585 | 585 | |
|
586 | def push(self, url): | |
|
586 | def push(self, url, **kwargs): | |
|
587 | 587 | url = self._get_url(url) |
|
588 | 588 | self._remote.sync_push(url) |
|
589 | 589 | |
|
590 | 590 | def _local_clone(self, clone_path): |
|
591 | 591 | """ |
|
592 | 592 | Create a local clone of the current repo. |
|
593 | 593 | """ |
|
594 | 594 | self._remote.clone(self.path, clone_path, update_after_clone=True, |
|
595 | 595 | hooks=False) |
|
596 | 596 | |
|
597 | 597 | def _update(self, revision, clean=False): |
|
598 | 598 | """ |
|
599 | 599 | Update the working copy to the specified revision. |
|
600 | 600 | """ |
|
601 | 601 | log.debug('Doing checkout to commit: `%s` for %s', revision, self) |
|
602 | 602 | self._remote.update(revision, clean=clean) |
|
603 | 603 | |
|
604 | 604 | def _identify(self): |
|
605 | 605 | """ |
|
606 | 606 | Return the current state of the working directory. |
|
607 | 607 | """ |
|
608 | 608 | return self._remote.identify().strip().rstrip('+') |
|
609 | 609 | |
|
610 | 610 | def _heads(self, branch=None): |
|
611 | 611 | """ |
|
612 | 612 | Return the commit ids of the repository heads. |
|
613 | 613 | """ |
|
614 | 614 | return self._remote.heads(branch=branch).strip().split(' ') |
|
615 | 615 | |
|
616 | 616 | def _ancestor(self, revision1, revision2): |
|
617 | 617 | """ |
|
618 | 618 | Return the common ancestor of the two revisions. |
|
619 | 619 | """ |
|
620 | 620 | return self._remote.ancestor(revision1, revision2) |
|
621 | 621 | |
|
622 | 622 | def _local_push( |
|
623 | 623 | self, revision, repository_path, push_branches=False, |
|
624 | 624 | enable_hooks=False): |
|
625 | 625 | """ |
|
626 | 626 | Push the given revision to the specified repository. |
|
627 | 627 | |
|
628 | 628 | :param push_branches: allow to create branches in the target repo. |
|
629 | 629 | """ |
|
630 | 630 | self._remote.push( |
|
631 | 631 | [revision], repository_path, hooks=enable_hooks, |
|
632 | 632 | push_branches=push_branches) |
|
633 | 633 | |
|
634 | 634 | def _local_merge(self, target_ref, merge_message, user_name, user_email, |
|
635 | 635 | source_ref, use_rebase=False, close_commit_id=None, dry_run=False): |
|
636 | 636 | """ |
|
637 | 637 | Merge the given source_revision into the checked out revision. |
|
638 | 638 | |
|
639 | 639 | Returns the commit id of the merge and a boolean indicating if the |
|
640 | 640 | commit needs to be pushed. |
|
641 | 641 | """ |
|
642 | 642 | |
|
643 | 643 | source_ref_commit_id = source_ref.commit_id |
|
644 | 644 | target_ref_commit_id = target_ref.commit_id |
|
645 | 645 | |
|
646 | 646 | # update our workdir to target ref, for proper merge |
|
647 | 647 | self._update(target_ref_commit_id, clean=True) |
|
648 | 648 | |
|
649 | 649 | ancestor = self._ancestor(target_ref_commit_id, source_ref_commit_id) |
|
650 | 650 | is_the_same_branch = self._is_the_same_branch(target_ref, source_ref) |
|
651 | 651 | |
|
652 | 652 | if close_commit_id: |
|
653 | 653 | # NOTE(marcink): if we get the close commit, this is our new source |
|
654 | 654 | # which will include the close commit itself. |
|
655 | 655 | source_ref_commit_id = close_commit_id |
|
656 | 656 | |
|
657 | 657 | if ancestor == source_ref_commit_id: |
|
658 | 658 | # Nothing to do, the changes were already integrated |
|
659 | 659 | return target_ref_commit_id, False |
|
660 | 660 | |
|
661 | 661 | elif ancestor == target_ref_commit_id and is_the_same_branch: |
|
662 | 662 | # In this case we should force a commit message |
|
663 | 663 | return source_ref_commit_id, True |
|
664 | 664 | |
|
665 | 665 | unresolved = None |
|
666 | 666 | if use_rebase: |
|
667 | 667 | try: |
|
668 | 668 | bookmark_name = f'rcbook{source_ref_commit_id}{target_ref_commit_id}' |
|
669 | 669 | self.bookmark(bookmark_name, revision=source_ref.commit_id) |
|
670 | 670 | self._remote.rebase( |
|
671 | 671 | source=source_ref_commit_id, dest=target_ref_commit_id) |
|
672 | 672 | self._remote.invalidate_vcs_cache() |
|
673 | 673 | self._update(bookmark_name, clean=True) |
|
674 | 674 | return self._identify(), True |
|
675 | 675 | except RepositoryError as e: |
|
676 | 676 | # The rebase-abort may raise another exception which 'hides' |
|
677 | 677 | # the original one, therefore we log it here. |
|
678 | 678 | log.exception('Error while rebasing shadow repo during merge.') |
|
679 | 679 | if 'unresolved conflicts' in safe_str(e): |
|
680 | 680 | unresolved = self._remote.get_unresolved_files() |
|
681 | 681 | log.debug('unresolved files: %s', unresolved) |
|
682 | 682 | |
|
683 | 683 | # Cleanup any rebase leftovers |
|
684 | 684 | self._remote.invalidate_vcs_cache() |
|
685 | 685 | self._remote.rebase(abort=True) |
|
686 | 686 | self._remote.invalidate_vcs_cache() |
|
687 | 687 | self._remote.update(clean=True) |
|
688 | 688 | if unresolved: |
|
689 | 689 | raise UnresolvedFilesInRepo(unresolved) |
|
690 | 690 | else: |
|
691 | 691 | raise |
|
692 | 692 | else: |
|
693 | 693 | try: |
|
694 | 694 | self._remote.merge(source_ref_commit_id) |
|
695 | 695 | self._remote.invalidate_vcs_cache() |
|
696 | 696 | self._remote.commit( |
|
697 | 697 | message=safe_str(merge_message), |
|
698 | 698 | username=safe_str(f'{user_name} <{user_email}>')) |
|
699 | 699 | self._remote.invalidate_vcs_cache() |
|
700 | 700 | return self._identify(), True |
|
701 | 701 | except RepositoryError as e: |
|
702 | 702 | # The merge-abort may raise another exception which 'hides' |
|
703 | 703 | # the original one, therefore we log it here. |
|
704 | 704 | log.exception('Error while merging shadow repo during merge.') |
|
705 | 705 | if 'unresolved merge conflicts' in safe_str(e): |
|
706 | 706 | unresolved = self._remote.get_unresolved_files() |
|
707 | 707 | log.debug('unresolved files: %s', unresolved) |
|
708 | 708 | |
|
709 | 709 | # Cleanup any merge leftovers |
|
710 | 710 | self._remote.update(clean=True) |
|
711 | 711 | if unresolved: |
|
712 | 712 | raise UnresolvedFilesInRepo(unresolved) |
|
713 | 713 | else: |
|
714 | 714 | raise |
|
715 | 715 | |
|
716 | 716 | def _local_close(self, target_ref, user_name, user_email, |
|
717 | 717 | source_ref, close_message=''): |
|
718 | 718 | """ |
|
719 | 719 | Close the branch of the given source_revision |
|
720 | 720 | |
|
721 | 721 | Returns the commit id of the close and a boolean indicating if the |
|
722 | 722 | commit needs to be pushed. |
|
723 | 723 | """ |
|
724 | 724 | self._update(source_ref.commit_id) |
|
725 | 725 | message = close_message or f"Closing branch: `{source_ref.name}`" |
|
726 | 726 | try: |
|
727 | 727 | self._remote.commit( |
|
728 | 728 | message=safe_str(message), |
|
729 | 729 | username=safe_str(f'{user_name} <{user_email}>'), |
|
730 | 730 | close_branch=True) |
|
731 | 731 | self._remote.invalidate_vcs_cache() |
|
732 | 732 | return self._identify(), True |
|
733 | 733 | except RepositoryError: |
|
734 | 734 | # Cleanup any commit leftovers |
|
735 | 735 | self._remote.update(clean=True) |
|
736 | 736 | raise |
|
737 | 737 | |
|
738 | 738 | def _is_the_same_branch(self, target_ref, source_ref): |
|
739 | 739 | return ( |
|
740 | 740 | self._get_branch_name(target_ref) == |
|
741 | 741 | self._get_branch_name(source_ref)) |
|
742 | 742 | |
|
743 | 743 | def _get_branch_name(self, ref): |
|
744 | 744 | if ref.type == 'branch': |
|
745 | 745 | return ref.name |
|
746 | 746 | return self._remote.ctx_branch(ref.commit_id) |
|
747 | 747 | |
|
748 | 748 | def _maybe_prepare_merge_workspace( |
|
749 | 749 | self, repo_id, workspace_id, unused_target_ref, unused_source_ref): |
|
750 | 750 | shadow_repository_path = self._get_shadow_repository_path( |
|
751 | 751 | self.path, repo_id, workspace_id) |
|
752 | 752 | if not os.path.exists(shadow_repository_path): |
|
753 | 753 | self._local_clone(shadow_repository_path) |
|
754 | 754 | log.debug( |
|
755 | 755 | 'Prepared shadow repository in %s', shadow_repository_path) |
|
756 | 756 | |
|
757 | 757 | return shadow_repository_path |
|
758 | 758 | |
|
759 | 759 | def _merge_repo(self, repo_id, workspace_id, target_ref, |
|
760 | 760 | source_repo, source_ref, merge_message, |
|
761 | 761 | merger_name, merger_email, dry_run=False, |
|
762 | 762 | use_rebase=False, close_branch=False): |
|
763 | 763 | |
|
764 | 764 | log.debug('Executing merge_repo with %s strategy, dry_run mode:%s', |
|
765 | 765 | 'rebase' if use_rebase else 'merge', dry_run) |
|
766 | 766 | |
|
767 | 767 | if target_ref.commit_id not in self._heads(): |
|
768 | 768 | return MergeResponse( |
|
769 | 769 | False, False, None, MergeFailureReason.TARGET_IS_NOT_HEAD, |
|
770 | 770 | metadata={'target_ref': target_ref}) |
|
771 | 771 | |
|
772 | 772 | try: |
|
773 | 773 | if target_ref.type == 'branch' and len(self._heads(target_ref.name)) != 1: |
|
774 | 774 | heads_all = self._heads(target_ref.name) |
|
775 | 775 | max_heads = 10 |
|
776 | 776 | if len(heads_all) > max_heads: |
|
777 | 777 | heads = '\n,'.join( |
|
778 | 778 | heads_all[:max_heads] + |
|
779 | 779 | [f'and {len(heads_all)-max_heads} more.']) |
|
780 | 780 | else: |
|
781 | 781 | heads = '\n,'.join(heads_all) |
|
782 | 782 | metadata = { |
|
783 | 783 | 'target_ref': target_ref, |
|
784 | 784 | 'source_ref': source_ref, |
|
785 | 785 | 'heads': heads |
|
786 | 786 | } |
|
787 | 787 | return MergeResponse( |
|
788 | 788 | False, False, None, |
|
789 | 789 | MergeFailureReason.HG_TARGET_HAS_MULTIPLE_HEADS, |
|
790 | 790 | metadata=metadata) |
|
791 | 791 | except CommitDoesNotExistError: |
|
792 | 792 | log.exception('Failure when looking up branch heads on hg target') |
|
793 | 793 | return MergeResponse( |
|
794 | 794 | False, False, None, MergeFailureReason.MISSING_TARGET_REF, |
|
795 | 795 | metadata={'target_ref': target_ref}) |
|
796 | 796 | |
|
797 | 797 | shadow_repository_path = self._maybe_prepare_merge_workspace( |
|
798 | 798 | repo_id, workspace_id, target_ref, source_ref) |
|
799 | 799 | shadow_repo = self.get_shadow_instance(shadow_repository_path) |
|
800 | 800 | |
|
801 | 801 | log.debug('Pulling in target reference %s', target_ref) |
|
802 | 802 | self._validate_pull_reference(target_ref) |
|
803 | 803 | shadow_repo._local_pull(self.path, target_ref) |
|
804 | 804 | |
|
805 | 805 | try: |
|
806 | 806 | log.debug('Pulling in source reference %s', source_ref) |
|
807 | 807 | source_repo._validate_pull_reference(source_ref) |
|
808 | 808 | shadow_repo._local_pull(source_repo.path, source_ref) |
|
809 | 809 | except CommitDoesNotExistError: |
|
810 | 810 | log.exception('Failure when doing local pull on hg shadow repo') |
|
811 | 811 | return MergeResponse( |
|
812 | 812 | False, False, None, MergeFailureReason.MISSING_SOURCE_REF, |
|
813 | 813 | metadata={'source_ref': source_ref}) |
|
814 | 814 | |
|
815 | 815 | merge_ref = None |
|
816 | 816 | merge_commit_id = None |
|
817 | 817 | close_commit_id = None |
|
818 | 818 | merge_failure_reason = MergeFailureReason.NONE |
|
819 | 819 | metadata = {} |
|
820 | 820 | |
|
821 | 821 | # enforce that close branch should be used only in case we source from |
|
822 | 822 | # an actual Branch |
|
823 | 823 | close_branch = close_branch and source_ref.type == 'branch' |
|
824 | 824 | |
|
825 | 825 | # don't allow to close branch if source and target are the same |
|
826 | 826 | close_branch = close_branch and source_ref.name != target_ref.name |
|
827 | 827 | |
|
828 | 828 | needs_push_on_close = False |
|
829 | 829 | if close_branch and not use_rebase and not dry_run: |
|
830 | 830 | try: |
|
831 | 831 | close_commit_id, needs_push_on_close = shadow_repo._local_close( |
|
832 | 832 | target_ref, merger_name, merger_email, source_ref) |
|
833 | 833 | merge_possible = True |
|
834 | 834 | except RepositoryError: |
|
835 | 835 | log.exception('Failure when doing close branch on ' |
|
836 | 836 | 'shadow repo: %s', shadow_repo) |
|
837 | 837 | merge_possible = False |
|
838 | 838 | merge_failure_reason = MergeFailureReason.MERGE_FAILED |
|
839 | 839 | else: |
|
840 | 840 | merge_possible = True |
|
841 | 841 | |
|
842 | 842 | needs_push = False |
|
843 | 843 | if merge_possible: |
|
844 | 844 | |
|
845 | 845 | try: |
|
846 | 846 | merge_commit_id, needs_push = shadow_repo._local_merge( |
|
847 | 847 | target_ref, merge_message, merger_name, merger_email, |
|
848 | 848 | source_ref, use_rebase=use_rebase, |
|
849 | 849 | close_commit_id=close_commit_id, dry_run=dry_run) |
|
850 | 850 | merge_possible = True |
|
851 | 851 | |
|
852 | 852 | # read the state of the close action, if it |
|
853 | 853 | # maybe required a push |
|
854 | 854 | needs_push = needs_push or needs_push_on_close |
|
855 | 855 | |
|
856 | 856 | # Set a bookmark pointing to the merge commit. This bookmark |
|
857 | 857 | # may be used to easily identify the last successful merge |
|
858 | 858 | # commit in the shadow repository. |
|
859 | 859 | shadow_repo.bookmark('pr-merge', revision=merge_commit_id) |
|
860 | 860 | merge_ref = Reference('book', 'pr-merge', merge_commit_id) |
|
861 | 861 | except SubrepoMergeError: |
|
862 | 862 | log.exception( |
|
863 | 863 | 'Subrepo merge error during local merge on hg shadow repo.') |
|
864 | 864 | merge_possible = False |
|
865 | 865 | merge_failure_reason = MergeFailureReason.SUBREPO_MERGE_FAILED |
|
866 | 866 | needs_push = False |
|
867 | 867 | except RepositoryError as e: |
|
868 | 868 | log.exception('Failure when doing local merge on hg shadow repo') |
|
869 | 869 | metadata['unresolved_files'] = 'no unresolved files found' |
|
870 | 870 | |
|
871 | 871 | if isinstance(e, UnresolvedFilesInRepo): |
|
872 | 872 | all_conflicts = list(e.args[0]) |
|
873 | 873 | max_conflicts = 20 |
|
874 | 874 | if len(all_conflicts) > max_conflicts: |
|
875 | 875 | conflicts = all_conflicts[:max_conflicts] \ |
|
876 | 876 | + [f'and {len(all_conflicts)-max_conflicts} more.'] |
|
877 | 877 | else: |
|
878 | 878 | conflicts = all_conflicts |
|
879 | 879 | metadata['unresolved_files'] = \ |
|
880 | 880 | '\n* conflict: ' + \ |
|
881 | 881 | ('\n * conflict: '.join(conflicts)) |
|
882 | 882 | |
|
883 | 883 | merge_possible = False |
|
884 | 884 | merge_failure_reason = MergeFailureReason.MERGE_FAILED |
|
885 | 885 | needs_push = False |
|
886 | 886 | |
|
887 | 887 | if merge_possible and not dry_run: |
|
888 | 888 | if needs_push: |
|
889 | 889 | # In case the target is a bookmark, update it, so after pushing |
|
890 | 890 | # the bookmarks is also updated in the target. |
|
891 | 891 | if target_ref.type == 'book': |
|
892 | 892 | shadow_repo.bookmark( |
|
893 | 893 | target_ref.name, revision=merge_commit_id) |
|
894 | 894 | try: |
|
895 | 895 | shadow_repo_with_hooks = self.get_shadow_instance( |
|
896 | 896 | shadow_repository_path, |
|
897 | 897 | enable_hooks=True) |
|
898 | 898 | # This is the actual merge action, we push from shadow |
|
899 | 899 | # into origin. |
|
900 | 900 | # Note: the push_branches option will push any new branch |
|
901 | 901 | # defined in the source repository to the target. This may |
|
902 | 902 | # be dangerous as branches are permanent in Mercurial. |
|
903 | 903 | # This feature was requested in issue #441. |
|
904 | 904 | shadow_repo_with_hooks._local_push( |
|
905 | 905 | merge_commit_id, self.path, push_branches=True, |
|
906 | 906 | enable_hooks=True) |
|
907 | 907 | |
|
908 | 908 | # maybe we also need to push the close_commit_id |
|
909 | 909 | if close_commit_id: |
|
910 | 910 | shadow_repo_with_hooks._local_push( |
|
911 | 911 | close_commit_id, self.path, push_branches=True, |
|
912 | 912 | enable_hooks=True) |
|
913 | 913 | merge_succeeded = True |
|
914 | 914 | except RepositoryError: |
|
915 | 915 | log.exception( |
|
916 | 916 | 'Failure when doing local push from the shadow ' |
|
917 | 917 | 'repository to the target repository at %s.', self.path) |
|
918 | 918 | merge_succeeded = False |
|
919 | 919 | merge_failure_reason = MergeFailureReason.PUSH_FAILED |
|
920 | 920 | metadata['target'] = 'hg shadow repo' |
|
921 | 921 | metadata['merge_commit'] = merge_commit_id |
|
922 | 922 | else: |
|
923 | 923 | merge_succeeded = True |
|
924 | 924 | else: |
|
925 | 925 | merge_succeeded = False |
|
926 | 926 | |
|
927 | 927 | return MergeResponse( |
|
928 | 928 | merge_possible, merge_succeeded, merge_ref, merge_failure_reason, |
|
929 | 929 | metadata=metadata) |
|
930 | 930 | |
|
931 | 931 | def get_shadow_instance(self, shadow_repository_path, enable_hooks=False, cache=False): |
|
932 | 932 | config = self.config.copy() |
|
933 | 933 | if not enable_hooks: |
|
934 | 934 | config.clear_section('hooks') |
|
935 | 935 | return MercurialRepository(shadow_repository_path, config, with_wire={"cache": cache}) |
|
936 | 936 | |
|
937 | 937 | def _validate_pull_reference(self, reference): |
|
938 | 938 | if not (reference.name in self.bookmarks or |
|
939 | 939 | reference.name in self.branches or |
|
940 | 940 | self.get_commit(reference.commit_id)): |
|
941 | 941 | raise CommitDoesNotExistError( |
|
942 | 942 | 'Unknown branch, bookmark or commit id') |
|
943 | 943 | |
|
944 | 944 | def _local_pull(self, repository_path, reference): |
|
945 | 945 | """ |
|
946 | 946 | Fetch a branch, bookmark or commit from a local repository. |
|
947 | 947 | """ |
|
948 | 948 | repository_path = os.path.abspath(repository_path) |
|
949 | 949 | if repository_path == self.path: |
|
950 | 950 | raise ValueError('Cannot pull from the same repository') |
|
951 | 951 | |
|
952 | 952 | reference_type_to_option_name = { |
|
953 | 953 | 'book': 'bookmark', |
|
954 | 954 | 'branch': 'branch', |
|
955 | 955 | } |
|
956 | 956 | option_name = reference_type_to_option_name.get( |
|
957 | 957 | reference.type, 'revision') |
|
958 | 958 | |
|
959 | 959 | if option_name == 'revision': |
|
960 | 960 | ref = reference.commit_id |
|
961 | 961 | else: |
|
962 | 962 | ref = reference.name |
|
963 | 963 | |
|
964 | 964 | options = {option_name: [ref]} |
|
965 | 965 | self._remote.pull_cmd(repository_path, hooks=False, **options) |
|
966 | 966 | self._remote.invalidate_vcs_cache() |
|
967 | 967 | |
|
968 | 968 | def bookmark(self, bookmark, revision=None): |
|
969 | 969 | if isinstance(bookmark, str): |
|
970 | 970 | bookmark = safe_str(bookmark) |
|
971 | 971 | self._remote.bookmark(bookmark, revision=revision) |
|
972 | 972 | self._remote.invalidate_vcs_cache() |
|
973 | 973 | |
|
974 | 974 | def get_path_permissions(self, username): |
|
975 | 975 | hgacl_file = os.path.join(self.path, '.hg/hgacl') |
|
976 | 976 | |
|
977 | 977 | def read_patterns(suffix): |
|
978 | 978 | svalue = None |
|
979 | 979 | for section, option in [ |
|
980 | 980 | ('narrowacl', username + suffix), |
|
981 | 981 | ('narrowacl', 'default' + suffix), |
|
982 | 982 | ('narrowhgacl', username + suffix), |
|
983 | 983 | ('narrowhgacl', 'default' + suffix) |
|
984 | 984 | ]: |
|
985 | 985 | try: |
|
986 | 986 | svalue = hgacl.get(section, option) |
|
987 | 987 | break # stop at the first value we find |
|
988 | 988 | except configparser.NoOptionError: |
|
989 | 989 | pass |
|
990 | 990 | if not svalue: |
|
991 | 991 | return None |
|
992 | 992 | result = ['/'] |
|
993 | 993 | for pattern in svalue.split(): |
|
994 | 994 | result.append(pattern) |
|
995 | 995 | if '*' not in pattern and '?' not in pattern: |
|
996 | 996 | result.append(pattern + '/*') |
|
997 | 997 | return result |
|
998 | 998 | |
|
999 | 999 | if os.path.exists(hgacl_file): |
|
1000 | 1000 | try: |
|
1001 | 1001 | hgacl = configparser.RawConfigParser() |
|
1002 | 1002 | hgacl.read(hgacl_file) |
|
1003 | 1003 | |
|
1004 | 1004 | includes = read_patterns('.includes') |
|
1005 | 1005 | excludes = read_patterns('.excludes') |
|
1006 | 1006 | return BasePathPermissionChecker.create_from_patterns( |
|
1007 | 1007 | includes, excludes) |
|
1008 | 1008 | except BaseException as e: |
|
1009 | 1009 | msg = 'Cannot read ACL settings from {} on {}: {}'.format( |
|
1010 | 1010 | hgacl_file, self.name, e) |
|
1011 | 1011 | raise exceptions.RepositoryRequirementError(msg) |
|
1012 | 1012 | else: |
|
1013 | 1013 | return None |
|
1014 | 1014 | |
|
1015 | 1015 | |
|
1016 | 1016 | class MercurialIndexBasedCollectionGenerator(CollectionGenerator): |
|
1017 | 1017 | |
|
1018 | 1018 | def _commit_factory(self, commit_id): |
|
1019 | 1019 | if isinstance(commit_id, int): |
|
1020 | 1020 | return self.repo.get_commit( |
|
1021 | 1021 | commit_idx=commit_id, pre_load=self.pre_load) |
|
1022 | 1022 | else: |
|
1023 | 1023 | return self.repo.get_commit( |
|
1024 | 1024 | commit_id=commit_id, pre_load=self.pre_load) |
@@ -1,1044 +1,1044 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 | """ |
|
20 | 20 | Scm model for RhodeCode |
|
21 | 21 | """ |
|
22 | 22 | |
|
23 | 23 | import os.path |
|
24 | 24 | import traceback |
|
25 | 25 | import logging |
|
26 | 26 | import io |
|
27 | 27 | |
|
28 | 28 | from sqlalchemy import func |
|
29 | 29 | from zope.cachedescriptors.property import Lazy as LazyProperty |
|
30 | 30 | |
|
31 | 31 | import rhodecode |
|
32 | 32 | from rhodecode.lib.str_utils import safe_bytes |
|
33 | 33 | from rhodecode.lib.vcs import get_backend |
|
34 | 34 | from rhodecode.lib.vcs.exceptions import RepositoryError, NodeNotChangedError |
|
35 | 35 | from rhodecode.lib.vcs.nodes import FileNode |
|
36 | 36 | from rhodecode.lib.vcs.backends.base import EmptyCommit |
|
37 | 37 | from rhodecode.lib import helpers as h, rc_cache |
|
38 | 38 | from rhodecode.lib.auth import ( |
|
39 | 39 | HasRepoPermissionAny, HasRepoGroupPermissionAny, |
|
40 | 40 | HasUserGroupPermissionAny) |
|
41 | 41 | from rhodecode.lib.exceptions import NonRelativePathError, IMCCommitError |
|
42 | 42 | from rhodecode.lib import hooks_utils |
|
43 | 43 | from rhodecode.lib.utils import ( |
|
44 | 44 | get_filesystem_repos, make_db_config) |
|
45 | 45 | from rhodecode.lib.str_utils import safe_str |
|
46 | 46 | from rhodecode.lib.system_info import get_system_info |
|
47 | 47 | from rhodecode.model import BaseModel |
|
48 | 48 | from rhodecode.model.db import ( |
|
49 | 49 | or_, false, null, |
|
50 | 50 | Repository, CacheKey, UserFollowing, UserLog, User, RepoGroup, |
|
51 | 51 | PullRequest, FileStore) |
|
52 | 52 | from rhodecode.model.settings import VcsSettingsModel |
|
53 | 53 | from rhodecode.model.validation_schema.validators import url_validator, InvalidCloneUrl |
|
54 | 54 | |
|
55 | 55 | log = logging.getLogger(__name__) |
|
56 | 56 | |
|
57 | 57 | |
|
58 | 58 | class UserTemp(object): |
|
59 | 59 | def __init__(self, user_id): |
|
60 | 60 | self.user_id = user_id |
|
61 | 61 | |
|
62 | 62 | def __repr__(self): |
|
63 | 63 | return "<{}('id:{}')>".format(self.__class__.__name__, self.user_id) |
|
64 | 64 | |
|
65 | 65 | |
|
66 | 66 | class RepoTemp(object): |
|
67 | 67 | def __init__(self, repo_id): |
|
68 | 68 | self.repo_id = repo_id |
|
69 | 69 | |
|
70 | 70 | def __repr__(self): |
|
71 | 71 | return "<{}('id:{}')>".format(self.__class__.__name__, self.repo_id) |
|
72 | 72 | |
|
73 | 73 | |
|
74 | 74 | class SimpleCachedRepoList(object): |
|
75 | 75 | """ |
|
76 | 76 | Lighter version of of iteration of repos without the scm initialisation, |
|
77 | 77 | and with cache usage |
|
78 | 78 | """ |
|
79 | 79 | def __init__(self, db_repo_list, repos_path, order_by=None, perm_set=None): |
|
80 | 80 | self.db_repo_list = db_repo_list |
|
81 | 81 | self.repos_path = repos_path |
|
82 | 82 | self.order_by = order_by |
|
83 | 83 | self.reversed = (order_by or '').startswith('-') |
|
84 | 84 | if not perm_set: |
|
85 | 85 | perm_set = ['repository.read', 'repository.write', |
|
86 | 86 | 'repository.admin'] |
|
87 | 87 | self.perm_set = perm_set |
|
88 | 88 | |
|
89 | 89 | def __len__(self): |
|
90 | 90 | return len(self.db_repo_list) |
|
91 | 91 | |
|
92 | 92 | def __repr__(self): |
|
93 | 93 | return '<{} ({})>'.format(self.__class__.__name__, self.__len__()) |
|
94 | 94 | |
|
95 | 95 | def __iter__(self): |
|
96 | 96 | for dbr in self.db_repo_list: |
|
97 | 97 | # check permission at this level |
|
98 | 98 | has_perm = HasRepoPermissionAny(*self.perm_set)( |
|
99 | 99 | dbr.repo_name, 'SimpleCachedRepoList check') |
|
100 | 100 | if not has_perm: |
|
101 | 101 | continue |
|
102 | 102 | |
|
103 | 103 | tmp_d = { |
|
104 | 104 | 'name': dbr.repo_name, |
|
105 | 105 | 'dbrepo': dbr.get_dict(), |
|
106 | 106 | 'dbrepo_fork': dbr.fork.get_dict() if dbr.fork else {} |
|
107 | 107 | } |
|
108 | 108 | yield tmp_d |
|
109 | 109 | |
|
110 | 110 | |
|
111 | 111 | class _PermCheckIterator(object): |
|
112 | 112 | |
|
113 | 113 | def __init__( |
|
114 | 114 | self, obj_list, obj_attr, perm_set, perm_checker, |
|
115 | 115 | extra_kwargs=None): |
|
116 | 116 | """ |
|
117 | 117 | Creates iterator from given list of objects, additionally |
|
118 | 118 | checking permission for them from perm_set var |
|
119 | 119 | |
|
120 | 120 | :param obj_list: list of db objects |
|
121 | 121 | :param obj_attr: attribute of object to pass into perm_checker |
|
122 | 122 | :param perm_set: list of permissions to check |
|
123 | 123 | :param perm_checker: callable to check permissions against |
|
124 | 124 | """ |
|
125 | 125 | self.obj_list = obj_list |
|
126 | 126 | self.obj_attr = obj_attr |
|
127 | 127 | self.perm_set = perm_set |
|
128 | 128 | self.perm_checker = perm_checker(*self.perm_set) |
|
129 | 129 | self.extra_kwargs = extra_kwargs or {} |
|
130 | 130 | |
|
131 | 131 | def __len__(self): |
|
132 | 132 | return len(self.obj_list) |
|
133 | 133 | |
|
134 | 134 | def __repr__(self): |
|
135 | 135 | return '<{} ({})>'.format(self.__class__.__name__, self.__len__()) |
|
136 | 136 | |
|
137 | 137 | def __iter__(self): |
|
138 | 138 | for db_obj in self.obj_list: |
|
139 | 139 | # check permission at this level |
|
140 | 140 | # NOTE(marcink): the __dict__.get() is ~4x faster then getattr() |
|
141 | 141 | name = db_obj.__dict__.get(self.obj_attr, None) |
|
142 | 142 | if not self.perm_checker(name, self.__class__.__name__, **self.extra_kwargs): |
|
143 | 143 | continue |
|
144 | 144 | |
|
145 | 145 | yield db_obj |
|
146 | 146 | |
|
147 | 147 | |
|
148 | 148 | class RepoList(_PermCheckIterator): |
|
149 | 149 | |
|
150 | 150 | def __init__(self, db_repo_list, perm_set=None, extra_kwargs=None): |
|
151 | 151 | if not perm_set: |
|
152 | 152 | perm_set = ['repository.read', 'repository.write', 'repository.admin'] |
|
153 | 153 | |
|
154 | 154 | super().__init__( |
|
155 | 155 | obj_list=db_repo_list, |
|
156 | 156 | obj_attr='_repo_name', perm_set=perm_set, |
|
157 | 157 | perm_checker=HasRepoPermissionAny, |
|
158 | 158 | extra_kwargs=extra_kwargs) |
|
159 | 159 | |
|
160 | 160 | |
|
161 | 161 | class RepoGroupList(_PermCheckIterator): |
|
162 | 162 | |
|
163 | 163 | def __init__(self, db_repo_group_list, perm_set=None, extra_kwargs=None): |
|
164 | 164 | if not perm_set: |
|
165 | 165 | perm_set = ['group.read', 'group.write', 'group.admin'] |
|
166 | 166 | |
|
167 | 167 | super().__init__( |
|
168 | 168 | obj_list=db_repo_group_list, |
|
169 | 169 | obj_attr='_group_name', perm_set=perm_set, |
|
170 | 170 | perm_checker=HasRepoGroupPermissionAny, |
|
171 | 171 | extra_kwargs=extra_kwargs) |
|
172 | 172 | |
|
173 | 173 | |
|
174 | 174 | class UserGroupList(_PermCheckIterator): |
|
175 | 175 | |
|
176 | 176 | def __init__(self, db_user_group_list, perm_set=None, extra_kwargs=None): |
|
177 | 177 | if not perm_set: |
|
178 | 178 | perm_set = ['usergroup.read', 'usergroup.write', 'usergroup.admin'] |
|
179 | 179 | |
|
180 | 180 | super().__init__( |
|
181 | 181 | obj_list=db_user_group_list, |
|
182 | 182 | obj_attr='users_group_name', perm_set=perm_set, |
|
183 | 183 | perm_checker=HasUserGroupPermissionAny, |
|
184 | 184 | extra_kwargs=extra_kwargs) |
|
185 | 185 | |
|
186 | 186 | |
|
187 | 187 | class ScmModel(BaseModel): |
|
188 | 188 | """ |
|
189 | 189 | Generic Scm Model |
|
190 | 190 | """ |
|
191 | 191 | |
|
192 | 192 | @LazyProperty |
|
193 | 193 | def repos_path(self): |
|
194 | 194 | """ |
|
195 | 195 | Gets the repositories root path from database |
|
196 | 196 | """ |
|
197 | 197 | |
|
198 | 198 | settings_model = VcsSettingsModel(sa=self.sa) |
|
199 | 199 | return settings_model.get_repos_location() |
|
200 | 200 | |
|
201 | 201 | def repo_scan(self, repos_path=None): |
|
202 | 202 | """ |
|
203 | 203 | Listing of repositories in given path. This path should not be a |
|
204 | 204 | repository itself. Return a dictionary of repository objects |
|
205 | 205 | |
|
206 | 206 | :param repos_path: path to directory containing repositories |
|
207 | 207 | """ |
|
208 | 208 | |
|
209 | 209 | if repos_path is None: |
|
210 | 210 | repos_path = self.repos_path |
|
211 | 211 | |
|
212 | 212 | log.info('scanning for repositories in %s', repos_path) |
|
213 | 213 | |
|
214 | 214 | config = make_db_config() |
|
215 | 215 | config.set('extensions', 'largefiles', '') |
|
216 | 216 | repos = {} |
|
217 | 217 | |
|
218 | 218 | for name, path in get_filesystem_repos(repos_path, recursive=True): |
|
219 | 219 | # name need to be decomposed and put back together using the / |
|
220 | 220 | # since this is internal storage separator for rhodecode |
|
221 | 221 | name = Repository.normalize_repo_name(name) |
|
222 | 222 | |
|
223 | 223 | try: |
|
224 | 224 | if name in repos: |
|
225 | 225 | raise RepositoryError('Duplicate repository name %s ' |
|
226 | 226 | 'found in %s' % (name, path)) |
|
227 | 227 | elif path[0] in rhodecode.BACKENDS: |
|
228 | 228 | backend = get_backend(path[0]) |
|
229 | 229 | repos[name] = backend(path[1], config=config, |
|
230 | 230 | with_wire={"cache": False}) |
|
231 | 231 | except OSError: |
|
232 | 232 | continue |
|
233 | 233 | except RepositoryError: |
|
234 | 234 | log.exception('Failed to create a repo') |
|
235 | 235 | continue |
|
236 | 236 | |
|
237 | 237 | log.debug('found %s paths with repositories', len(repos)) |
|
238 | 238 | return repos |
|
239 | 239 | |
|
240 | 240 | def get_repos(self, all_repos=None, sort_key=None): |
|
241 | 241 | """ |
|
242 | 242 | Get all repositories from db and for each repo create it's |
|
243 | 243 | backend instance and fill that backed with information from database |
|
244 | 244 | |
|
245 | 245 | :param all_repos: list of repository names as strings |
|
246 | 246 | give specific repositories list, good for filtering |
|
247 | 247 | |
|
248 | 248 | :param sort_key: initial sorting of repositories |
|
249 | 249 | """ |
|
250 | 250 | if all_repos is None: |
|
251 | 251 | all_repos = self.sa.query(Repository)\ |
|
252 | 252 | .filter(Repository.group_id == null())\ |
|
253 | 253 | .order_by(func.lower(Repository.repo_name)).all() |
|
254 | 254 | repo_iter = SimpleCachedRepoList( |
|
255 | 255 | all_repos, repos_path=self.repos_path, order_by=sort_key) |
|
256 | 256 | return repo_iter |
|
257 | 257 | |
|
258 | 258 | def get_repo_groups(self, all_groups=None): |
|
259 | 259 | if all_groups is None: |
|
260 | 260 | all_groups = RepoGroup.query()\ |
|
261 | 261 | .filter(RepoGroup.group_parent_id == null()).all() |
|
262 | 262 | return [x for x in RepoGroupList(all_groups)] |
|
263 | 263 | |
|
264 | 264 | def mark_for_invalidation(self, repo_name, delete=False): |
|
265 | 265 | """ |
|
266 | 266 | Mark caches of this repo invalid in the database. `delete` flag |
|
267 | 267 | removes the cache entries |
|
268 | 268 | |
|
269 | 269 | :param repo_name: the repo_name for which caches should be marked |
|
270 | 270 | invalid, or deleted |
|
271 | 271 | :param delete: delete the entry keys instead of setting bool |
|
272 | 272 | flag on them, and also purge caches used by the dogpile |
|
273 | 273 | """ |
|
274 | 274 | repo = Repository.get_by_repo_name(repo_name) |
|
275 | 275 | |
|
276 | 276 | if repo: |
|
277 | 277 | invalidation_namespace = CacheKey.REPO_INVALIDATION_NAMESPACE.format( |
|
278 | 278 | repo_id=repo.repo_id) |
|
279 | 279 | CacheKey.set_invalidate(invalidation_namespace, delete=delete) |
|
280 | 280 | |
|
281 | 281 | repo_id = repo.repo_id |
|
282 | 282 | config = repo._config |
|
283 | 283 | config.set('extensions', 'largefiles', '') |
|
284 | 284 | repo.update_commit_cache(config=config, cs_cache=None) |
|
285 | 285 | if delete: |
|
286 | 286 | cache_namespace_uid = f'cache_repo.{repo_id}' |
|
287 | 287 | rc_cache.clear_cache_namespace('cache_repo', cache_namespace_uid, method=rc_cache.CLEAR_INVALIDATE) |
|
288 | 288 | |
|
289 | 289 | def toggle_following_repo(self, follow_repo_id, user_id): |
|
290 | 290 | |
|
291 | 291 | f = self.sa.query(UserFollowing)\ |
|
292 | 292 | .filter(UserFollowing.follows_repo_id == follow_repo_id)\ |
|
293 | 293 | .filter(UserFollowing.user_id == user_id).scalar() |
|
294 | 294 | |
|
295 | 295 | if f is not None: |
|
296 | 296 | try: |
|
297 | 297 | self.sa.delete(f) |
|
298 | 298 | return |
|
299 | 299 | except Exception: |
|
300 | 300 | log.error(traceback.format_exc()) |
|
301 | 301 | raise |
|
302 | 302 | |
|
303 | 303 | try: |
|
304 | 304 | f = UserFollowing() |
|
305 | 305 | f.user_id = user_id |
|
306 | 306 | f.follows_repo_id = follow_repo_id |
|
307 | 307 | self.sa.add(f) |
|
308 | 308 | except Exception: |
|
309 | 309 | log.error(traceback.format_exc()) |
|
310 | 310 | raise |
|
311 | 311 | |
|
312 | 312 | def toggle_following_user(self, follow_user_id, user_id): |
|
313 | 313 | f = self.sa.query(UserFollowing)\ |
|
314 | 314 | .filter(UserFollowing.follows_user_id == follow_user_id)\ |
|
315 | 315 | .filter(UserFollowing.user_id == user_id).scalar() |
|
316 | 316 | |
|
317 | 317 | if f is not None: |
|
318 | 318 | try: |
|
319 | 319 | self.sa.delete(f) |
|
320 | 320 | return |
|
321 | 321 | except Exception: |
|
322 | 322 | log.error(traceback.format_exc()) |
|
323 | 323 | raise |
|
324 | 324 | |
|
325 | 325 | try: |
|
326 | 326 | f = UserFollowing() |
|
327 | 327 | f.user_id = user_id |
|
328 | 328 | f.follows_user_id = follow_user_id |
|
329 | 329 | self.sa.add(f) |
|
330 | 330 | except Exception: |
|
331 | 331 | log.error(traceback.format_exc()) |
|
332 | 332 | raise |
|
333 | 333 | |
|
334 | 334 | def is_following_repo(self, repo_name, user_id, cache=False): |
|
335 | 335 | r = self.sa.query(Repository)\ |
|
336 | 336 | .filter(Repository.repo_name == repo_name).scalar() |
|
337 | 337 | |
|
338 | 338 | f = self.sa.query(UserFollowing)\ |
|
339 | 339 | .filter(UserFollowing.follows_repository == r)\ |
|
340 | 340 | .filter(UserFollowing.user_id == user_id).scalar() |
|
341 | 341 | |
|
342 | 342 | return f is not None |
|
343 | 343 | |
|
344 | 344 | def is_following_user(self, username, user_id, cache=False): |
|
345 | 345 | u = User.get_by_username(username) |
|
346 | 346 | |
|
347 | 347 | f = self.sa.query(UserFollowing)\ |
|
348 | 348 | .filter(UserFollowing.follows_user == u)\ |
|
349 | 349 | .filter(UserFollowing.user_id == user_id).scalar() |
|
350 | 350 | |
|
351 | 351 | return f is not None |
|
352 | 352 | |
|
353 | 353 | def get_followers(self, repo): |
|
354 | 354 | repo = self._get_repo(repo) |
|
355 | 355 | |
|
356 | 356 | return self.sa.query(UserFollowing)\ |
|
357 | 357 | .filter(UserFollowing.follows_repository == repo).count() |
|
358 | 358 | |
|
359 | 359 | def get_forks(self, repo): |
|
360 | 360 | repo = self._get_repo(repo) |
|
361 | 361 | return self.sa.query(Repository)\ |
|
362 | 362 | .filter(Repository.fork == repo).count() |
|
363 | 363 | |
|
364 | 364 | def get_pull_requests(self, repo): |
|
365 | 365 | repo = self._get_repo(repo) |
|
366 | 366 | return self.sa.query(PullRequest)\ |
|
367 | 367 | .filter(PullRequest.target_repo == repo)\ |
|
368 | 368 | .filter(PullRequest.status != PullRequest.STATUS_CLOSED).count() |
|
369 | 369 | |
|
370 | 370 | def get_artifacts(self, repo): |
|
371 | 371 | repo = self._get_repo(repo) |
|
372 | 372 | return self.sa.query(FileStore)\ |
|
373 | 373 | .filter(FileStore.repo == repo)\ |
|
374 | 374 | .filter(or_(FileStore.hidden == null(), FileStore.hidden == false())).count() |
|
375 | 375 | |
|
376 | 376 | def mark_as_fork(self, repo, fork, user): |
|
377 | 377 | repo = self._get_repo(repo) |
|
378 | 378 | fork = self._get_repo(fork) |
|
379 | 379 | if fork and repo.repo_id == fork.repo_id: |
|
380 | 380 | raise Exception("Cannot set repository as fork of itself") |
|
381 | 381 | |
|
382 | 382 | if fork and repo.repo_type != fork.repo_type: |
|
383 | 383 | raise RepositoryError( |
|
384 | 384 | "Cannot set repository as fork of repository with other type") |
|
385 | 385 | |
|
386 | 386 | repo.fork = fork |
|
387 | 387 | self.sa.add(repo) |
|
388 | 388 | return repo |
|
389 | 389 | |
|
390 | 390 | def pull_changes(self, repo, username, remote_uri=None, validate_uri=True, **kwargs): |
|
391 | 391 | dbrepo = self._get_repo(repo) |
|
392 | 392 | remote_uri = remote_uri or dbrepo.clone_uri |
|
393 | 393 | if not remote_uri: |
|
394 | 394 | raise Exception("This repository doesn't have a clone uri") |
|
395 | 395 | |
|
396 | 396 | repo = dbrepo.scm_instance(cache=False) |
|
397 | 397 | repo.config.clear_section('hooks') |
|
398 | 398 | |
|
399 | 399 | try: |
|
400 | 400 | # NOTE(marcink): add extra validation so we skip invalid urls |
|
401 | 401 | # this is due this tasks can be executed via scheduler without |
|
402 | 402 | # proper validation of remote_uri |
|
403 | 403 | if validate_uri: |
|
404 | 404 | config = make_db_config(clear_session=False) |
|
405 | 405 | url_validator(remote_uri, dbrepo.repo_type, config) |
|
406 | 406 | except InvalidCloneUrl: |
|
407 | 407 | raise |
|
408 | 408 | |
|
409 | 409 | repo_name = dbrepo.repo_name |
|
410 | 410 | try: |
|
411 | 411 | # TODO: we need to make sure those operations call proper hooks ! |
|
412 | 412 | repo.fetch(remote_uri, **kwargs) |
|
413 | 413 | |
|
414 | 414 | self.mark_for_invalidation(repo_name) |
|
415 | 415 | except Exception: |
|
416 | 416 | log.error(traceback.format_exc()) |
|
417 | 417 | raise |
|
418 | 418 | |
|
419 | def push_changes(self, repo, username, remote_uri=None, validate_uri=True): | |
|
419 | def push_changes(self, repo, username, remote_uri=None, validate_uri=True, **kwargs): | |
|
420 | 420 | dbrepo = self._get_repo(repo) |
|
421 | 421 | remote_uri = remote_uri or dbrepo.push_uri |
|
422 | 422 | if not remote_uri: |
|
423 | 423 | raise Exception("This repository doesn't have a clone uri") |
|
424 | 424 | |
|
425 | 425 | repo = dbrepo.scm_instance(cache=False) |
|
426 | 426 | repo.config.clear_section('hooks') |
|
427 | 427 | |
|
428 | 428 | try: |
|
429 | 429 | # NOTE(marcink): add extra validation so we skip invalid urls |
|
430 | 430 | # this is due this tasks can be executed via scheduler without |
|
431 | 431 | # proper validation of remote_uri |
|
432 | 432 | if validate_uri: |
|
433 | 433 | config = make_db_config(clear_session=False) |
|
434 | 434 | url_validator(remote_uri, dbrepo.repo_type, config) |
|
435 | 435 | except InvalidCloneUrl: |
|
436 | 436 | raise |
|
437 | 437 | |
|
438 | 438 | try: |
|
439 | repo.push(remote_uri) | |
|
439 | repo.push(remote_uri, **kwargs) | |
|
440 | 440 | except Exception: |
|
441 | 441 | log.error(traceback.format_exc()) |
|
442 | 442 | raise |
|
443 | 443 | |
|
444 | 444 | def commit_change(self, repo, repo_name, commit, user, author, message, |
|
445 | 445 | content: bytes, f_path: bytes, branch: str = None): |
|
446 | 446 | """ |
|
447 | 447 | Commits changes |
|
448 | 448 | """ |
|
449 | 449 | user = self._get_user(user) |
|
450 | 450 | |
|
451 | 451 | # message and author needs to be unicode |
|
452 | 452 | # proper backend should then translate that into required type |
|
453 | 453 | message = safe_str(message) |
|
454 | 454 | author = safe_str(author) |
|
455 | 455 | imc = repo.in_memory_commit |
|
456 | 456 | imc.change(FileNode(f_path, content, mode=commit.get_file_mode(f_path))) |
|
457 | 457 | try: |
|
458 | 458 | # TODO: handle pre-push action ! |
|
459 | 459 | tip = imc.commit( |
|
460 | 460 | message=message, author=author, parents=[commit], |
|
461 | 461 | branch=branch or commit.branch) |
|
462 | 462 | except Exception as e: |
|
463 | 463 | log.error(traceback.format_exc()) |
|
464 | 464 | raise IMCCommitError(str(e)) |
|
465 | 465 | finally: |
|
466 | 466 | # always clear caches, if commit fails we want fresh object also |
|
467 | 467 | self.mark_for_invalidation(repo_name) |
|
468 | 468 | |
|
469 | 469 | # We trigger the post-push action |
|
470 | 470 | hooks_utils.trigger_post_push_hook( |
|
471 | 471 | username=user.username, action='push_local', hook_type='post_push', |
|
472 | 472 | repo_name=repo_name, repo_type=repo.alias, commit_ids=[tip.raw_id]) |
|
473 | 473 | return tip |
|
474 | 474 | |
|
475 | 475 | def _sanitize_path(self, f_path: bytes): |
|
476 | 476 | if f_path.startswith(b'/') or f_path.startswith(b'./') or b'../' in f_path: |
|
477 | 477 | raise NonRelativePathError(b'%b is not an relative path' % f_path) |
|
478 | 478 | if f_path: |
|
479 | 479 | f_path = os.path.normpath(f_path) |
|
480 | 480 | return f_path |
|
481 | 481 | |
|
482 | 482 | def get_dirnode_metadata(self, request, commit, dir_node): |
|
483 | 483 | if not dir_node.is_dir(): |
|
484 | 484 | return [] |
|
485 | 485 | |
|
486 | 486 | data = [] |
|
487 | 487 | for node in dir_node: |
|
488 | 488 | if not node.is_file(): |
|
489 | 489 | # we skip file-nodes |
|
490 | 490 | continue |
|
491 | 491 | |
|
492 | 492 | last_commit = node.last_commit |
|
493 | 493 | last_commit_date = last_commit.date |
|
494 | 494 | data.append({ |
|
495 | 495 | 'name': node.name, |
|
496 | 496 | 'size': h.format_byte_size_binary(node.size), |
|
497 | 497 | 'modified_at': h.format_date(last_commit_date), |
|
498 | 498 | 'modified_ts': last_commit_date.isoformat(), |
|
499 | 499 | 'revision': last_commit.revision, |
|
500 | 500 | 'short_id': last_commit.short_id, |
|
501 | 501 | 'message': h.escape(last_commit.message), |
|
502 | 502 | 'author': h.escape(last_commit.author), |
|
503 | 503 | 'user_profile': h.gravatar_with_user( |
|
504 | 504 | request, last_commit.author), |
|
505 | 505 | }) |
|
506 | 506 | |
|
507 | 507 | return data |
|
508 | 508 | |
|
509 | 509 | def get_nodes(self, repo_name, commit_id, root_path='/', flat=True, |
|
510 | 510 | extended_info=False, content=False, max_file_bytes=None): |
|
511 | 511 | """ |
|
512 | 512 | recursive walk in root dir and return a set of all path in that dir |
|
513 | 513 | based on repository walk function |
|
514 | 514 | |
|
515 | 515 | :param repo_name: name of repository |
|
516 | 516 | :param commit_id: commit id for which to list nodes |
|
517 | 517 | :param root_path: root path to list |
|
518 | 518 | :param flat: return as a list, if False returns a dict with description |
|
519 | 519 | :param extended_info: show additional info such as md5, binary, size etc |
|
520 | 520 | :param content: add nodes content to the return data |
|
521 | 521 | :param max_file_bytes: will not return file contents over this limit |
|
522 | 522 | |
|
523 | 523 | """ |
|
524 | 524 | _files = list() |
|
525 | 525 | _dirs = list() |
|
526 | 526 | |
|
527 | 527 | try: |
|
528 | 528 | _repo = self._get_repo(repo_name) |
|
529 | 529 | commit = _repo.scm_instance().get_commit(commit_id=commit_id) |
|
530 | 530 | root_path = root_path.lstrip('/') |
|
531 | 531 | |
|
532 | 532 | # get RootNode, inject pre-load options before walking |
|
533 | 533 | top_node = commit.get_node(root_path) |
|
534 | 534 | extended_info_pre_load = [] |
|
535 | 535 | if extended_info: |
|
536 | 536 | extended_info_pre_load += ['md5'] |
|
537 | 537 | top_node.default_pre_load = ['is_binary', 'size'] + extended_info_pre_load |
|
538 | 538 | |
|
539 | 539 | for __, dirs, files in commit.walk(top_node): |
|
540 | 540 | |
|
541 | 541 | for f in files: |
|
542 | 542 | _content = None |
|
543 | 543 | _data = f_name = f.str_path |
|
544 | 544 | |
|
545 | 545 | if not flat: |
|
546 | 546 | _data = { |
|
547 | 547 | "name": h.escape(f_name), |
|
548 | 548 | "type": "file", |
|
549 | 549 | } |
|
550 | 550 | if extended_info: |
|
551 | 551 | _data.update({ |
|
552 | 552 | "md5": f.md5, |
|
553 | 553 | "binary": f.is_binary, |
|
554 | 554 | "size": f.size, |
|
555 | 555 | "extension": f.extension, |
|
556 | 556 | "mimetype": f.mimetype, |
|
557 | 557 | "lines": f.lines()[0] |
|
558 | 558 | }) |
|
559 | 559 | |
|
560 | 560 | if content: |
|
561 | 561 | over_size_limit = (max_file_bytes is not None |
|
562 | 562 | and f.size > max_file_bytes) |
|
563 | 563 | full_content = None |
|
564 | 564 | if not f.is_binary and not over_size_limit: |
|
565 | 565 | full_content = f.str_content |
|
566 | 566 | |
|
567 | 567 | _data.update({ |
|
568 | 568 | "content": full_content, |
|
569 | 569 | }) |
|
570 | 570 | _files.append(_data) |
|
571 | 571 | |
|
572 | 572 | for d in dirs: |
|
573 | 573 | _data = d_name = d.str_path |
|
574 | 574 | if not flat: |
|
575 | 575 | _data = { |
|
576 | 576 | "name": h.escape(d_name), |
|
577 | 577 | "type": "dir", |
|
578 | 578 | } |
|
579 | 579 | if extended_info: |
|
580 | 580 | _data.update({ |
|
581 | 581 | "md5": "", |
|
582 | 582 | "binary": False, |
|
583 | 583 | "size": 0, |
|
584 | 584 | "extension": "", |
|
585 | 585 | }) |
|
586 | 586 | if content: |
|
587 | 587 | _data.update({ |
|
588 | 588 | "content": None |
|
589 | 589 | }) |
|
590 | 590 | _dirs.append(_data) |
|
591 | 591 | except RepositoryError: |
|
592 | 592 | log.exception("Exception in get_nodes") |
|
593 | 593 | raise |
|
594 | 594 | |
|
595 | 595 | return _dirs, _files |
|
596 | 596 | |
|
597 | 597 | def get_quick_filter_nodes(self, repo_name, commit_id, root_path='/'): |
|
598 | 598 | """ |
|
599 | 599 | Generate files for quick filter in files view |
|
600 | 600 | """ |
|
601 | 601 | |
|
602 | 602 | _files = list() |
|
603 | 603 | _dirs = list() |
|
604 | 604 | try: |
|
605 | 605 | _repo = self._get_repo(repo_name) |
|
606 | 606 | commit = _repo.scm_instance().get_commit(commit_id=commit_id) |
|
607 | 607 | root_path = root_path.lstrip('/') |
|
608 | 608 | |
|
609 | 609 | top_node = commit.get_node(root_path) |
|
610 | 610 | top_node.default_pre_load = [] |
|
611 | 611 | |
|
612 | 612 | for __, dirs, files in commit.walk(top_node): |
|
613 | 613 | for f in files: |
|
614 | 614 | |
|
615 | 615 | _data = { |
|
616 | 616 | "name": h.escape(f.str_path), |
|
617 | 617 | "type": "file", |
|
618 | 618 | } |
|
619 | 619 | |
|
620 | 620 | _files.append(_data) |
|
621 | 621 | |
|
622 | 622 | for d in dirs: |
|
623 | 623 | |
|
624 | 624 | _data = { |
|
625 | 625 | "name": h.escape(d.str_path), |
|
626 | 626 | "type": "dir", |
|
627 | 627 | } |
|
628 | 628 | |
|
629 | 629 | _dirs.append(_data) |
|
630 | 630 | except RepositoryError: |
|
631 | 631 | log.exception("Exception in get_quick_filter_nodes") |
|
632 | 632 | raise |
|
633 | 633 | |
|
634 | 634 | return _dirs, _files |
|
635 | 635 | |
|
636 | 636 | def get_node(self, repo_name, commit_id, file_path, |
|
637 | 637 | extended_info=False, content=False, max_file_bytes=None, cache=True): |
|
638 | 638 | """ |
|
639 | 639 | retrieve single node from commit |
|
640 | 640 | """ |
|
641 | 641 | |
|
642 | 642 | try: |
|
643 | 643 | |
|
644 | 644 | _repo = self._get_repo(repo_name) |
|
645 | 645 | commit = _repo.scm_instance().get_commit(commit_id=commit_id) |
|
646 | 646 | |
|
647 | 647 | file_node = commit.get_node(file_path) |
|
648 | 648 | if file_node.is_dir(): |
|
649 | 649 | raise RepositoryError('The given path is a directory') |
|
650 | 650 | |
|
651 | 651 | _content = None |
|
652 | 652 | f_name = file_node.str_path |
|
653 | 653 | |
|
654 | 654 | file_data = { |
|
655 | 655 | "name": h.escape(f_name), |
|
656 | 656 | "type": "file", |
|
657 | 657 | } |
|
658 | 658 | |
|
659 | 659 | if extended_info: |
|
660 | 660 | file_data.update({ |
|
661 | 661 | "extension": file_node.extension, |
|
662 | 662 | "mimetype": file_node.mimetype, |
|
663 | 663 | }) |
|
664 | 664 | |
|
665 | 665 | if cache: |
|
666 | 666 | md5 = file_node.md5 |
|
667 | 667 | is_binary = file_node.is_binary |
|
668 | 668 | size = file_node.size |
|
669 | 669 | else: |
|
670 | 670 | is_binary, md5, size, _content = file_node.metadata_uncached() |
|
671 | 671 | |
|
672 | 672 | file_data.update({ |
|
673 | 673 | "md5": md5, |
|
674 | 674 | "binary": is_binary, |
|
675 | 675 | "size": size, |
|
676 | 676 | }) |
|
677 | 677 | |
|
678 | 678 | if content and cache: |
|
679 | 679 | # get content + cache |
|
680 | 680 | size = file_node.size |
|
681 | 681 | over_size_limit = (max_file_bytes is not None and size > max_file_bytes) |
|
682 | 682 | full_content = None |
|
683 | 683 | all_lines = 0 |
|
684 | 684 | if not file_node.is_binary and not over_size_limit: |
|
685 | 685 | full_content = safe_str(file_node.content) |
|
686 | 686 | all_lines, empty_lines = file_node.count_lines(full_content) |
|
687 | 687 | |
|
688 | 688 | file_data.update({ |
|
689 | 689 | "content": full_content, |
|
690 | 690 | "lines": all_lines |
|
691 | 691 | }) |
|
692 | 692 | elif content: |
|
693 | 693 | # get content *without* cache |
|
694 | 694 | if _content is None: |
|
695 | 695 | is_binary, md5, size, _content = file_node.metadata_uncached() |
|
696 | 696 | |
|
697 | 697 | over_size_limit = (max_file_bytes is not None and size > max_file_bytes) |
|
698 | 698 | full_content = None |
|
699 | 699 | all_lines = 0 |
|
700 | 700 | if not is_binary and not over_size_limit: |
|
701 | 701 | full_content = safe_str(_content) |
|
702 | 702 | all_lines, empty_lines = file_node.count_lines(full_content) |
|
703 | 703 | |
|
704 | 704 | file_data.update({ |
|
705 | 705 | "content": full_content, |
|
706 | 706 | "lines": all_lines |
|
707 | 707 | }) |
|
708 | 708 | |
|
709 | 709 | except RepositoryError: |
|
710 | 710 | log.exception("Exception in get_node") |
|
711 | 711 | raise |
|
712 | 712 | |
|
713 | 713 | return file_data |
|
714 | 714 | |
|
715 | 715 | def get_fts_data(self, repo_name, commit_id, root_path='/'): |
|
716 | 716 | """ |
|
717 | 717 | Fetch node tree for usage in full text search |
|
718 | 718 | """ |
|
719 | 719 | |
|
720 | 720 | tree_info = list() |
|
721 | 721 | |
|
722 | 722 | try: |
|
723 | 723 | _repo = self._get_repo(repo_name) |
|
724 | 724 | commit = _repo.scm_instance().get_commit(commit_id=commit_id) |
|
725 | 725 | root_path = root_path.lstrip('/') |
|
726 | 726 | top_node = commit.get_node(root_path) |
|
727 | 727 | top_node.default_pre_load = [] |
|
728 | 728 | |
|
729 | 729 | for __, dirs, files in commit.walk(top_node): |
|
730 | 730 | |
|
731 | 731 | for f in files: |
|
732 | 732 | is_binary, md5, size, _content = f.metadata_uncached() |
|
733 | 733 | _data = { |
|
734 | 734 | "name": f.str_path, |
|
735 | 735 | "md5": md5, |
|
736 | 736 | "extension": f.extension, |
|
737 | 737 | "binary": is_binary, |
|
738 | 738 | "size": size |
|
739 | 739 | } |
|
740 | 740 | |
|
741 | 741 | tree_info.append(_data) |
|
742 | 742 | |
|
743 | 743 | except RepositoryError: |
|
744 | 744 | log.exception("Exception in get_nodes") |
|
745 | 745 | raise |
|
746 | 746 | |
|
747 | 747 | return tree_info |
|
748 | 748 | |
|
749 | 749 | def create_nodes(self, user, repo, message, nodes, parent_commit=None, |
|
750 | 750 | author=None, trigger_push_hook=True): |
|
751 | 751 | """ |
|
752 | 752 | Commits given multiple nodes into repo |
|
753 | 753 | |
|
754 | 754 | :param user: RhodeCode User object or user_id, the commiter |
|
755 | 755 | :param repo: RhodeCode Repository object |
|
756 | 756 | :param message: commit message |
|
757 | 757 | :param nodes: mapping {filename:{'content':content},...} |
|
758 | 758 | :param parent_commit: parent commit, can be empty than it's |
|
759 | 759 | initial commit |
|
760 | 760 | :param author: author of commit, cna be different that commiter |
|
761 | 761 | only for git |
|
762 | 762 | :param trigger_push_hook: trigger push hooks |
|
763 | 763 | |
|
764 | 764 | :returns: new committed commit |
|
765 | 765 | """ |
|
766 | 766 | |
|
767 | 767 | user = self._get_user(user) |
|
768 | 768 | scm_instance = repo.scm_instance(cache=False) |
|
769 | 769 | |
|
770 | 770 | message = safe_str(message) |
|
771 | 771 | commiter = user.full_contact |
|
772 | 772 | author = safe_str(author) if author else commiter |
|
773 | 773 | |
|
774 | 774 | imc = scm_instance.in_memory_commit |
|
775 | 775 | |
|
776 | 776 | if not parent_commit: |
|
777 | 777 | parent_commit = EmptyCommit(alias=scm_instance.alias) |
|
778 | 778 | |
|
779 | 779 | if isinstance(parent_commit, EmptyCommit): |
|
780 | 780 | # EmptyCommit means we're editing empty repository |
|
781 | 781 | parents = None |
|
782 | 782 | else: |
|
783 | 783 | parents = [parent_commit] |
|
784 | 784 | |
|
785 | 785 | upload_file_types = (io.BytesIO, io.BufferedRandom) |
|
786 | 786 | processed_nodes = [] |
|
787 | 787 | for filename, content_dict in nodes.items(): |
|
788 | 788 | if not isinstance(filename, bytes): |
|
789 | 789 | raise ValueError(f'filename key in nodes needs to be bytes , or {upload_file_types}') |
|
790 | 790 | content = content_dict['content'] |
|
791 | 791 | if not isinstance(content, upload_file_types + (bytes,)): |
|
792 | 792 | raise ValueError('content key value in nodes needs to be bytes') |
|
793 | 793 | |
|
794 | 794 | for f_path in nodes: |
|
795 | 795 | f_path = self._sanitize_path(f_path) |
|
796 | 796 | content = nodes[f_path]['content'] |
|
797 | 797 | |
|
798 | 798 | # decoding here will force that we have proper encoded values |
|
799 | 799 | # in any other case this will throw exceptions and deny commit |
|
800 | 800 | |
|
801 | 801 | if isinstance(content, bytes): |
|
802 | 802 | pass |
|
803 | 803 | elif isinstance(content, upload_file_types): |
|
804 | 804 | content = content.read() |
|
805 | 805 | else: |
|
806 | 806 | raise Exception(f'Content is of unrecognized type {type(content)}, expected {upload_file_types}') |
|
807 | 807 | processed_nodes.append((f_path, content)) |
|
808 | 808 | |
|
809 | 809 | # add multiple nodes |
|
810 | 810 | for path, content in processed_nodes: |
|
811 | 811 | imc.add(FileNode(path, content=content)) |
|
812 | 812 | |
|
813 | 813 | # TODO: handle pre push scenario |
|
814 | 814 | tip = imc.commit(message=message, |
|
815 | 815 | author=author, |
|
816 | 816 | parents=parents, |
|
817 | 817 | branch=parent_commit.branch) |
|
818 | 818 | |
|
819 | 819 | self.mark_for_invalidation(repo.repo_name) |
|
820 | 820 | if trigger_push_hook: |
|
821 | 821 | hooks_utils.trigger_post_push_hook( |
|
822 | 822 | username=user.username, action='push_local', |
|
823 | 823 | repo_name=repo.repo_name, repo_type=scm_instance.alias, |
|
824 | 824 | hook_type='post_push', |
|
825 | 825 | commit_ids=[tip.raw_id]) |
|
826 | 826 | return tip |
|
827 | 827 | |
|
828 | 828 | def update_nodes(self, user, repo, message, nodes, parent_commit=None, |
|
829 | 829 | author=None, trigger_push_hook=True): |
|
830 | 830 | user = self._get_user(user) |
|
831 | 831 | scm_instance = repo.scm_instance(cache=False) |
|
832 | 832 | |
|
833 | 833 | message = safe_str(message) |
|
834 | 834 | commiter = user.full_contact |
|
835 | 835 | author = safe_str(author) if author else commiter |
|
836 | 836 | |
|
837 | 837 | imc = scm_instance.in_memory_commit |
|
838 | 838 | |
|
839 | 839 | if not parent_commit: |
|
840 | 840 | parent_commit = EmptyCommit(alias=scm_instance.alias) |
|
841 | 841 | |
|
842 | 842 | if isinstance(parent_commit, EmptyCommit): |
|
843 | 843 | # EmptyCommit means we we're editing empty repository |
|
844 | 844 | parents = None |
|
845 | 845 | else: |
|
846 | 846 | parents = [parent_commit] |
|
847 | 847 | |
|
848 | 848 | # add multiple nodes |
|
849 | 849 | for _filename, data in nodes.items(): |
|
850 | 850 | # new filename, can be renamed from the old one, also sanitaze |
|
851 | 851 | # the path for any hack around relative paths like ../../ etc. |
|
852 | 852 | filename = self._sanitize_path(data['filename']) |
|
853 | 853 | old_filename = self._sanitize_path(_filename) |
|
854 | 854 | content = data['content'] |
|
855 | 855 | file_mode = data.get('mode') |
|
856 | 856 | filenode = FileNode(old_filename, content=content, mode=file_mode) |
|
857 | 857 | op = data['op'] |
|
858 | 858 | if op == 'add': |
|
859 | 859 | imc.add(filenode) |
|
860 | 860 | elif op == 'del': |
|
861 | 861 | imc.remove(filenode) |
|
862 | 862 | elif op == 'mod': |
|
863 | 863 | if filename != old_filename: |
|
864 | 864 | # TODO: handle renames more efficient, needs vcs lib changes |
|
865 | 865 | imc.remove(filenode) |
|
866 | 866 | imc.add(FileNode(filename, content=content, mode=file_mode)) |
|
867 | 867 | else: |
|
868 | 868 | imc.change(filenode) |
|
869 | 869 | |
|
870 | 870 | try: |
|
871 | 871 | # TODO: handle pre push scenario commit changes |
|
872 | 872 | tip = imc.commit(message=message, |
|
873 | 873 | author=author, |
|
874 | 874 | parents=parents, |
|
875 | 875 | branch=parent_commit.branch) |
|
876 | 876 | except NodeNotChangedError: |
|
877 | 877 | raise |
|
878 | 878 | except Exception as e: |
|
879 | 879 | log.exception("Unexpected exception during call to imc.commit") |
|
880 | 880 | raise IMCCommitError(str(e)) |
|
881 | 881 | finally: |
|
882 | 882 | # always clear caches, if commit fails we want fresh object also |
|
883 | 883 | self.mark_for_invalidation(repo.repo_name) |
|
884 | 884 | |
|
885 | 885 | if trigger_push_hook: |
|
886 | 886 | hooks_utils.trigger_post_push_hook( |
|
887 | 887 | username=user.username, action='push_local', hook_type='post_push', |
|
888 | 888 | repo_name=repo.repo_name, repo_type=scm_instance.alias, |
|
889 | 889 | commit_ids=[tip.raw_id]) |
|
890 | 890 | |
|
891 | 891 | return tip |
|
892 | 892 | |
|
893 | 893 | def delete_nodes(self, user, repo, message, nodes, parent_commit=None, |
|
894 | 894 | author=None, trigger_push_hook=True): |
|
895 | 895 | """ |
|
896 | 896 | Deletes given multiple nodes into `repo` |
|
897 | 897 | |
|
898 | 898 | :param user: RhodeCode User object or user_id, the committer |
|
899 | 899 | :param repo: RhodeCode Repository object |
|
900 | 900 | :param message: commit message |
|
901 | 901 | :param nodes: mapping {filename:{'content':content},...} |
|
902 | 902 | :param parent_commit: parent commit, can be empty than it's initial |
|
903 | 903 | commit |
|
904 | 904 | :param author: author of commit, cna be different that commiter only |
|
905 | 905 | for git |
|
906 | 906 | :param trigger_push_hook: trigger push hooks |
|
907 | 907 | |
|
908 | 908 | :returns: new commit after deletion |
|
909 | 909 | """ |
|
910 | 910 | |
|
911 | 911 | user = self._get_user(user) |
|
912 | 912 | scm_instance = repo.scm_instance(cache=False) |
|
913 | 913 | |
|
914 | 914 | processed_nodes = [] |
|
915 | 915 | for f_path in nodes: |
|
916 | 916 | f_path = self._sanitize_path(f_path) |
|
917 | 917 | # content can be empty but for compatibility it allows same dicts |
|
918 | 918 | # structure as add_nodes |
|
919 | 919 | content = nodes[f_path].get('content') |
|
920 | 920 | processed_nodes.append((safe_bytes(f_path), content)) |
|
921 | 921 | |
|
922 | 922 | message = safe_str(message) |
|
923 | 923 | commiter = user.full_contact |
|
924 | 924 | author = safe_str(author) if author else commiter |
|
925 | 925 | |
|
926 | 926 | imc = scm_instance.in_memory_commit |
|
927 | 927 | |
|
928 | 928 | if not parent_commit: |
|
929 | 929 | parent_commit = EmptyCommit(alias=scm_instance.alias) |
|
930 | 930 | |
|
931 | 931 | if isinstance(parent_commit, EmptyCommit): |
|
932 | 932 | # EmptyCommit means we we're editing empty repository |
|
933 | 933 | parents = None |
|
934 | 934 | else: |
|
935 | 935 | parents = [parent_commit] |
|
936 | 936 | # add multiple nodes |
|
937 | 937 | for path, content in processed_nodes: |
|
938 | 938 | imc.remove(FileNode(path, content=content)) |
|
939 | 939 | |
|
940 | 940 | # TODO: handle pre push scenario |
|
941 | 941 | tip = imc.commit(message=message, |
|
942 | 942 | author=author, |
|
943 | 943 | parents=parents, |
|
944 | 944 | branch=parent_commit.branch) |
|
945 | 945 | |
|
946 | 946 | self.mark_for_invalidation(repo.repo_name) |
|
947 | 947 | if trigger_push_hook: |
|
948 | 948 | hooks_utils.trigger_post_push_hook( |
|
949 | 949 | username=user.username, action='push_local', hook_type='post_push', |
|
950 | 950 | repo_name=repo.repo_name, repo_type=scm_instance.alias, |
|
951 | 951 | commit_ids=[tip.raw_id]) |
|
952 | 952 | return tip |
|
953 | 953 | |
|
954 | 954 | def strip(self, repo, commit_id, branch): |
|
955 | 955 | scm_instance = repo.scm_instance(cache=False) |
|
956 | 956 | scm_instance.config.clear_section('hooks') |
|
957 | 957 | scm_instance.strip(commit_id, branch) |
|
958 | 958 | self.mark_for_invalidation(repo.repo_name) |
|
959 | 959 | |
|
960 | 960 | def get_unread_journal(self): |
|
961 | 961 | return self.sa.query(UserLog).count() |
|
962 | 962 | |
|
963 | 963 | @classmethod |
|
964 | 964 | def backend_landing_ref(cls, repo_type): |
|
965 | 965 | """ |
|
966 | 966 | Return a default landing ref based on a repository type. |
|
967 | 967 | """ |
|
968 | 968 | |
|
969 | 969 | landing_ref = { |
|
970 | 970 | 'hg': ('branch:default', 'default'), |
|
971 | 971 | 'git': ('branch:master', 'master'), |
|
972 | 972 | 'svn': ('rev:tip', 'latest tip'), |
|
973 | 973 | 'default': ('rev:tip', 'latest tip'), |
|
974 | 974 | } |
|
975 | 975 | |
|
976 | 976 | return landing_ref.get(repo_type) or landing_ref['default'] |
|
977 | 977 | |
|
978 | 978 | def get_repo_landing_revs(self, translator, repo=None): |
|
979 | 979 | """ |
|
980 | 980 | Generates select option with tags branches and bookmarks (for hg only) |
|
981 | 981 | grouped by type |
|
982 | 982 | |
|
983 | 983 | :param repo: |
|
984 | 984 | """ |
|
985 | 985 | from rhodecode.lib.vcs.backends.git import GitRepository |
|
986 | 986 | |
|
987 | 987 | _ = translator |
|
988 | 988 | repo = self._get_repo(repo) |
|
989 | 989 | |
|
990 | 990 | if repo: |
|
991 | 991 | repo_type = repo.repo_type |
|
992 | 992 | else: |
|
993 | 993 | repo_type = 'default' |
|
994 | 994 | |
|
995 | 995 | default_landing_ref, landing_ref_lbl = self.backend_landing_ref(repo_type) |
|
996 | 996 | |
|
997 | 997 | default_ref_options = [ |
|
998 | 998 | [default_landing_ref, landing_ref_lbl] |
|
999 | 999 | ] |
|
1000 | 1000 | default_choices = [ |
|
1001 | 1001 | default_landing_ref |
|
1002 | 1002 | ] |
|
1003 | 1003 | |
|
1004 | 1004 | if not repo: |
|
1005 | 1005 | # presented at NEW repo creation |
|
1006 | 1006 | return default_choices, default_ref_options |
|
1007 | 1007 | |
|
1008 | 1008 | repo = repo.scm_instance() |
|
1009 | 1009 | |
|
1010 | 1010 | ref_options = [(default_landing_ref, landing_ref_lbl)] |
|
1011 | 1011 | choices = [default_landing_ref] |
|
1012 | 1012 | |
|
1013 | 1013 | # branches |
|
1014 | 1014 | branch_group = [(f'branch:{safe_str(b)}', safe_str(b)) for b in repo.branches] |
|
1015 | 1015 | if not branch_group: |
|
1016 | 1016 | # new repo, or without maybe a branch? |
|
1017 | 1017 | branch_group = default_ref_options |
|
1018 | 1018 | |
|
1019 | 1019 | branches_group = (branch_group, _("Branches")) |
|
1020 | 1020 | ref_options.append(branches_group) |
|
1021 | 1021 | choices.extend([x[0] for x in branches_group[0]]) |
|
1022 | 1022 | |
|
1023 | 1023 | # bookmarks for HG |
|
1024 | 1024 | if repo.alias == 'hg': |
|
1025 | 1025 | bookmarks_group = ( |
|
1026 | 1026 | [(f'book:{safe_str(b)}', safe_str(b)) |
|
1027 | 1027 | for b in repo.bookmarks], |
|
1028 | 1028 | _("Bookmarks")) |
|
1029 | 1029 | ref_options.append(bookmarks_group) |
|
1030 | 1030 | choices.extend([x[0] for x in bookmarks_group[0]]) |
|
1031 | 1031 | |
|
1032 | 1032 | # tags |
|
1033 | 1033 | tags_group = ( |
|
1034 | 1034 | [(f'tag:{safe_str(t)}', safe_str(t)) |
|
1035 | 1035 | for t in repo.tags], |
|
1036 | 1036 | _("Tags")) |
|
1037 | 1037 | ref_options.append(tags_group) |
|
1038 | 1038 | choices.extend([x[0] for x in tags_group[0]]) |
|
1039 | 1039 | |
|
1040 | 1040 | return choices, ref_options |
|
1041 | 1041 | |
|
1042 | 1042 | def get_server_info(self, environ=None): |
|
1043 | 1043 | server_info = get_system_info(environ) |
|
1044 | 1044 | return server_info |
General Comments 0
You need to be logged in to leave comments.
Login now