feat(archive-cache): implemented s3 based backend for filecaches

rhodecode/lib/archive_cache/backends/__init__.py

@@ -0,0 +1,17 b''
+# Copyright (C) 2015-2024 RhodeCode GmbH
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License, version 3
+# (only), as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# This program is dual-licensed. If you wish to learn more about the
+# RhodeCode Enterprise Edition, including its added features, Support services,
+# and proprietary license terms, please see https://rhodecode.com/licenses/

rhodecode/lib/archive_cache/backends/base.py

@@ -0,0 +1,348 b''
+# Copyright (C) 2015-2024 RhodeCode GmbH
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License, version 3
+# (only), as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# This program is dual-licensed. If you wish to learn more about the
+# RhodeCode Enterprise Edition, including its added features, Support services,
+# and proprietary license terms, please see https://rhodecode.com/licenses/
+
+import os
+import functools
+import logging
+import typing
+import time
+import zlib
+
+from ...ext_json import json
+from ..utils import StatsDB, NOT_GIVEN, ShardFileReader, EVICTION_POLICY, format_size
+from ..lock import GenerationLock
+
+log = logging.getLogger(__name__)
+
+
+class BaseShard:
+    storage_type: str = ''
+    fs = None
+
+    @classmethod
+    def hash(cls, key):
+        """Compute portable hash for `key`.
+
+        :param key: key to hash
+        :return: hash value
+
+        """
+        mask = 0xFFFFFFFF
+        return zlib.adler32(key.encode('utf-8')) & mask  # noqa
+
+    def _write_file(self, full_path, read_iterator, mode):
+        raise NotImplementedError
+
+    def _get_keyfile(self, key):
+        raise NotImplementedError
+
+    def random_filename(self):
+        raise NotImplementedError
+
+    def _store(self, key, value_reader, metadata, mode):
+        (filename,  # hash-name
+         full_path  # full-path/hash-name
+         ) = self.random_filename()
+
+        key_file, key_file_path = self._get_keyfile(key)
+
+        # STORE METADATA
+        _metadata = {
+            "version": "v1",
+
+            "key_file": key_file,  # this is the .key.json file storing meta
+            "key_file_path": key_file_path,  # full path to key_file
+            "archive_key": key,  # original name we stored archive under, e.g my-archive.zip
+            "archive_filename": filename,  # the actual filename we stored that file under
+            "archive_full_path": full_path,
+
+            "store_time": time.time(),
+            "access_count": 0,
+            "access_time": 0,
+
+            "size": 0
+        }
+        if metadata:
+            _metadata.update(metadata)
+
+        read_iterator = iter(functools.partial(value_reader.read, 2**22), b'')
+        size, sha256 = self._write_file(full_path, read_iterator, mode)
+        _metadata['size'] = size
+        _metadata['sha256'] = sha256
+
+        # after archive is finished, we create a key to save the presence of the binary file
+        with self.fs.open(key_file_path, 'wb') as f:
+            f.write(json.dumps(_metadata))
+
+        return key, filename, size, _metadata
+
+    def _fetch(self,  key, retry, retry_attempts, retry_backoff):
+        if retry is NOT_GIVEN:
+            retry = False
+        if retry_attempts is NOT_GIVEN:
+            retry_attempts = 0
+
+        if retry and retry_attempts > 0:
+            for attempt in range(1, retry_attempts + 1):
+                if key in self:
+                    break
+                # we didn't find the key, wait retry_backoff N seconds, and re-check
+                time.sleep(retry_backoff)
+
+        if key not in self:
+            log.exception(f'requested key={key} not found in {self} retry={retry}, attempts={retry_attempts}')
+            raise KeyError(key)
+
+        key_file, key_file_path = self._get_keyfile(key)
+        with self.fs.open(key_file_path, 'rb') as f:
+            metadata = json.loads(f.read())
+
+        archive_path = metadata['archive_full_path']
+
+        try:
+            return ShardFileReader(self.fs.open(archive_path, 'rb')), metadata
+        finally:
+            # update usage stats, count and accessed
+            metadata["access_count"] = metadata.get("access_count", 0) + 1
+            metadata["access_time"] = time.time()
+            log.debug('Updated %s with access snapshot, access_count=%s access_time=%s',
+                      key_file, metadata['access_count'], metadata['access_time'])
+            with self.fs.open(key_file_path, 'wb') as f:
+                f.write(json.dumps(metadata))
+
+    def _remove(self, key):
+        if key not in self:
+            log.exception(f'requested key={key} not found in {self}')
+            raise KeyError(key)
+
+        key_file, key_file_path = self._get_keyfile(key)
+        with self.fs.open(key_file_path, 'rb') as f:
+            metadata = json.loads(f.read())
+
+        archive_path = metadata['archive_full_path']
+        self.fs.rm(archive_path)
+        self.fs.rm(key_file_path)
+        return 1
+
+    @property
+    def storage_medium(self):
+        return getattr(self, self.storage_type)
+
+    @property
+    def key_suffix(self):
+        return 'key.json'
+
+    def __contains__(self, key):
+        """Return `True` if `key` matching item is found in cache.
+
+        :param key: key matching item
+        :return: True if key matching item
+
+        """
+        key_file, key_file_path = self._get_keyfile(key)
+        return self.fs.exists(key_file_path)
+
+
+class BaseCache:
+    _locking_url: str = ''
+    _storage_path: str = ''
+    _config = {}
+    retry = False
+    retry_attempts = 0
+    retry_backoff = 1
+    _shards = tuple()
+
+    def __contains__(self, key):
+        """Return `True` if `key` matching item is found in cache.
+
+        :param key: key matching item
+        :return: True if key matching item
+
+        """
+        return self.has_key(key)
+
+    def __repr__(self):
+        return f'<{self.__class__.__name__}(storage={self._storage_path})>'
+
+    @classmethod
+    def gb_to_bytes(cls, gb):
+        return gb * (1024 ** 3)
+
+    @property
+    def storage_path(self):
+        return self._storage_path
+
+    @classmethod
+    def get_stats_db(cls):
+        return StatsDB()
+
+    def get_conf(self, key, pop=False):
+        if key not in self._config:
+            raise ValueError(f"No configuration key '{key}', please make sure it exists in archive_cache config")
+        val = self._config[key]
+        if pop:
+            del self._config[key]
+        return val
+
+    def _get_shard(self, key):
+        raise NotImplementedError
+
+    def _get_size(self, shard, archive_path):
+        raise NotImplementedError
+
+    def store(self, key, value_reader, metadata=None):
+        shard = self._get_shard(key)
+        return shard.store(key, value_reader, metadata)
+
+    def fetch(self, key, retry=NOT_GIVEN, retry_attempts=NOT_GIVEN) -> tuple[typing.BinaryIO, dict]:
+        """
+        Return file handle corresponding to `key` from specific shard cache.
+        """
+        if retry is NOT_GIVEN:
+            retry = self.retry
+        if retry_attempts is NOT_GIVEN:
+            retry_attempts = self.retry_attempts
+        retry_backoff = self.retry_backoff
+
+        shard = self._get_shard(key)
+        return shard.fetch(key, retry=retry, retry_attempts=retry_attempts, retry_backoff=retry_backoff)
+
+    def remove(self, key):
+        shard = self._get_shard(key)
+        return shard.remove(key)
+
+    def has_key(self, archive_key):
+        """Return `True` if `key` matching item is found in cache.
+
+        :param archive_key: key for item, this is a unique archive name we want to store data under. e.g my-archive-svn.zip
+        :return: True if key is found
+
+        """
+        shard = self._get_shard(archive_key)
+        return archive_key in shard
+
+    def iter_keys(self):
+        for shard in self._shards:
+            if shard.fs.exists(shard.storage_medium):
+                for path, _dirs, _files in shard.fs.walk(shard.storage_medium):
+                    for key_file_path in _files:
+                        if key_file_path.endswith(shard.key_suffix):
+                            yield shard, key_file_path
+
+    def get_lock(self, lock_key):
+        return GenerationLock(lock_key, self._locking_url)
+
+    def evict(self, policy=None, size_limit=None) -> int:
+        """
+        Remove old items based on the conditions
+
+
+        explanation of this algo:
+        iterate over each shard, then for each shard iterate over the .key files
+        read the key files metadata stored. This gives us a full list of keys, cached_archived, their size and
+        access data, time creation, and access counts.
+
+        Store that into a memory DB so we can run different sorting strategies easily.
+        Summing the size is a sum sql query.
+
+        Then we run a sorting strategy based on eviction policy.
+        We iterate over sorted keys, and remove each checking if we hit the overall limit.
+        """
+
+        policy = policy or self._eviction_policy
+        size_limit = size_limit or self._cache_size_limit
+
+        select_policy = EVICTION_POLICY[policy]['evict']
+
+        log.debug('Running eviction policy \'%s\', and checking for size limit: %s',
+                  policy, format_size(size_limit))
+
+        if select_policy is None:
+            return 0
+
+        db = self.get_stats_db()
+
+        data = []
+        cnt = 1
+
+        for shard, key_file in self.iter_keys():
+            with shard.fs.open(os.path.join(shard.storage_medium, key_file), 'rb') as f:
+                metadata = json.loads(f.read())
+
+            key_file_path = os.path.join(shard.storage_medium, key_file)
+
+            archive_key = metadata['archive_key']
+            archive_path = metadata['archive_full_path']
+
+            size = metadata.get('size')
+            if not size:
+                # in case we don't have size re-calc it...
+                size = self._get_size(shard, archive_path)
+
+            data.append([
+                cnt,
+                key_file,
+                key_file_path,
+                archive_key,
+                archive_path,
+                metadata.get('store_time', 0),
+                metadata.get('access_time', 0),
+                metadata.get('access_count', 0),
+                size,
+            ])
+            cnt += 1
+
+        # Insert bulk data using executemany
+        db.bulk_insert(data)
+
+        total_size = db.get_total_size()
+        log.debug('Analyzed %s keys, occupying: %s, running eviction to match %s',
+                  len(data), format_size(total_size), format_size(size_limit))
+
+        removed_items = 0
+        removed_size = 0
+        for key_file, archive_key, size in db.get_sorted_keys(select_policy):
+            # simulate removal impact BEFORE removal
+            total_size -= size
+
+            if total_size <= size_limit:
+                # we obtained what we wanted...
+                break
+
+            self.remove(archive_key)
+            removed_items += 1
+            removed_size += size
+
+        log.debug('Removed %s cache archives, and reduced size by: %s',
+                  removed_items, format_size(removed_size))
+        return removed_items
+
+    def get_statistics(self):
+        total_files = 0
+        total_size = 0
+        meta = {}
+
+        for shard, key_file in self.iter_keys():
+            json_key = f"{shard.storage_medium}/{key_file}"
+            with shard.fs.open(json_key, 'rb') as f:
+                total_files += 1
+                metadata = json.loads(f.read())
+                total_size += metadata['size']
+
+        return total_files, total_size, meta
+

rhodecode/lib/archive_cache/backends/objectstore_cache.py

@@ -0,0 +1,150 b''
+# Copyright (C) 2015-2024 RhodeCode GmbH
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License, version 3
+# (only), as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# This program is dual-licensed. If you wish to learn more about the
+# RhodeCode Enterprise Edition, including its added features, Support services,
+# and proprietary license terms, please see https://rhodecode.com/licenses/
+
+import codecs
+import hashlib
+import logging
+import os
+
+import fsspec
+
+from .base import BaseCache, BaseShard
+from ..utils import ShardFileReader, NOT_GIVEN
+from ...type_utils import str2bool
+
+log = logging.getLogger(__name__)
+
+
+class S3Shard(BaseShard):
+
+    def __init__(self, index, bucket, **settings):
+        self._index = index
+        self._bucket = bucket
+        self.storage_type = 'bucket'
+
+        endpoint_url = settings.pop('archive_cache.objectstore.url')
+        key = settings.pop('archive_cache.objectstore.key')
+        secret = settings.pop('archive_cache.objectstore.secret')
+
+        self.fs = fsspec.filesystem('s3', anon=False, endpoint_url=endpoint_url, key=key, secret=secret)
+
+    @property
+    def bucket(self):
+        """Cache bucket."""
+        return self._bucket
+
+    def _get_keyfile(self, archive_key) -> tuple[str, str]:
+        key_file = f'{archive_key}-{self.key_suffix}'
+        return key_file, os.path.join(self.bucket, key_file)
+
+    def _get_writer(self, path, mode):
+        return self.fs.open(path, 'wb')
+
+    def _write_file(self, full_path, iterator, mode):
+        # ensure bucket exists
+        destination = self.bucket
+        if not self.fs.exists(destination):
+            self.fs.mkdir(destination, s3_additional_kwargs={})
+
+        writer = self._get_writer(full_path, mode)
+
+        digest = hashlib.sha256()
+        with writer:
+            size = 0
+            for chunk in iterator:
+                size += len(chunk)
+                digest.update(chunk)
+                writer.write(chunk)
+
+            sha256 = digest.hexdigest()
+            log.debug('written new archive cache under %s, sha256: %s', full_path, sha256)
+            return size, sha256
+
+    def store(self, key, value_reader, metadata: dict | None = None):
+        return self._store(key, value_reader, metadata, mode='wb')
+
+    def fetch(self, key, retry=NOT_GIVEN, retry_attempts=NOT_GIVEN, retry_backoff=1) -> tuple[ShardFileReader, dict]:
+        return self._fetch(key, retry, retry_attempts, retry_backoff)
+
+    def remove(self, key):
+        return self._remove(key)
+
+    def random_filename(self):
+        """Return filename and full-path tuple for file storage.
+
+        Filename will be a randomly generated 28 character hexadecimal string
+        with ".archive_cache" suffixed. Two levels of sub-directories will be used to
+        reduce the size of directories. On older filesystems, lookups in
+        directories with many files may be slow.
+        """
+
+        hex_name = codecs.encode(os.urandom(16), 'hex').decode('utf-8')
+
+        archive_name = hex_name[4:] + '.archive_cache'
+        filename = f"{hex_name[:2]}-{hex_name[2:4]}-{archive_name}"
+
+        full_path = os.path.join(self.bucket, filename)
+        return archive_name, full_path
+
+    def __repr__(self):
+        return f'{self.__class__.__name__}(index={self._index}, bucket={self.bucket})'
+
+
+class ObjectStoreCache(BaseCache):
+
+    def __init__(self, locking_url, **settings):
+        """
+        Initialize objectstore cache instance.
+
+        :param str locking_url: redis url for a lock
+        :param settings: settings dict
+
+        """
+        self._locking_url = locking_url
+        self._config = settings
+
+        objectstore_url = self.get_conf('archive_cache.objectstore.url')
+        self._storage_path = objectstore_url
+
+        self._count = int(self.get_conf('archive_cache.objectstore.bucket_shards', pop=True))
+
+        self._eviction_policy = self.get_conf('archive_cache.objectstore.eviction_policy', pop=True)
+        self._cache_size_limit = self.gb_to_bytes(int(self.get_conf('archive_cache.objectstore.cache_size_gb')))
+
+        self.retry = str2bool(self.get_conf('archive_cache.objectstore.retry', pop=True))
+        self.retry_attempts = int(self.get_conf('archive_cache.objectstore.retry_attempts', pop=True))
+        self.retry_backoff = int(self.get_conf('archive_cache.objectstore.retry_backoff', pop=True))
+
+        log.debug('Initializing archival cache instance under %s', objectstore_url)
+        self._shards = tuple(
+            S3Shard(
+                index=num,
+                bucket='rhodecode-archivecache-%03d' % num,
+                **settings,
+            )
+            for num in range(self._count)
+        )
+        self._hash = self._shards[0].hash
+
+    def _get_shard(self, key) -> S3Shard:
+        index = self._hash(key) % self._count
+        shard = self._shards[index]
+        return shard
+
+    def _get_size(self, shard, archive_path):
+        return shard.fs.info(archive_path)['size']

rhodecode/tests/lib/test_archive_caches.py

@@ -0,0 +1,105 b''
+# Copyright (C) 2016-2023 RhodeCode GmbH
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License, version 3
+# (only), as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# This program is dual-licensed. If you wish to learn more about the
+# RhodeCode Enterprise Edition, including its added features, Support services,
+# and proprietary license terms, please see https://rhodecode.com/licenses/
+
+import time
+import pytest
+import rhodecode
+import os
+import shutil
+from tempfile import mkdtemp
+
+from rhodecode.lib import archive_cache
+
+
+def file_reader(temp_store):
+    with open(temp_store, 'w') as f:
+        for cnt in range(10000):
+            f.write(str(cnt))
+    return open(temp_store, 'rb')
+
+
[e-mail]()
+def d_cache_instance(ini_settings):
+    config = ini_settings
+    d_cache = archive_cache.get_archival_cache_store(config=config, always_init=True)
+    return d_cache
+
+
[e-mail]('app')
+class TestArchiveCaches(object):
+
+    def test_archivecache_empty_stats(self, d_cache_instance):
+        d_cache = d_cache_instance
+        shutil.rmtree(d_cache._directory)
+
+        stats = d_cache.get_statistics()
+        assert (0, 0, {}) == stats
+
+    def test_archivecache_store_keys(self, d_cache_instance, tmp_path):
+        d_cache = d_cache_instance
+        shutil.rmtree(d_cache._directory)
+
+        for n in range(100):
+
+            archive_name = f'my-archive-abc-{n}.zip'
+            temp_archive_path = os.path.join(tmp_path, archive_name)
+            d_cache.store(archive_name, file_reader(temp_archive_path ), {'foo': 'bar'})
+            reader, meta = d_cache.fetch(archive_name)
+            content = reader.read()
+            assert content == open(temp_archive_path, 'rb').read()
+
+        stats = d_cache.get_statistics()
+        assert (100, 3889000, {}) == stats
+
+    def test_archivecache_remove_keys(self, d_cache_instance, tmp_path):
+        d_cache = d_cache_instance
+        shutil.rmtree(d_cache._directory)
+
+        n = 1
+        archive_name = f'my-archive-abc-{n}.zip'
+        temp_archive_path = os.path.join(tmp_path, archive_name)
+
+        d_cache.store(archive_name, file_reader(temp_archive_path ), {'foo': 'bar'})
+        stats = d_cache.get_statistics()
+        assert (1, 38890, {}) == stats
+
+        assert 1 == d_cache.remove(archive_name)
+
+        stats = d_cache.get_statistics()
+        assert (0, 0, {}) == stats
+
+    def test_archivecache_evict_keys(self, d_cache_instance, tmp_path):
+        d_cache = d_cache_instance
+        shutil.rmtree(d_cache._directory)
+        tries = 500
+        for n in range(tries):
+
+            archive_name = f'my-archive-abc-{n}.zip'
+            temp_archive_path = os.path.join(tmp_path, archive_name)
+            d_cache.store(archive_name, file_reader(temp_archive_path ), {'foo': 'bar'})
+
+        stats = d_cache.get_statistics()
+        assert (tries, 19445000, {}) == stats
+        evict_to = 0.005  # around (5mb)
+        evicted_items = d_cache.evict(size_limit=d_cache.gb_to_bytes(evict_to))
+        evicted = 361
+        assert evicted == evicted_items
+
+        stats = d_cache.get_statistics()
+        assert (tries - evicted, 5405710, {}) == stats
+

configs/development.ini

@@ -1,813 +1,845 b''
 
 ; #########################################
 ; RHODECODE COMMUNITY EDITION CONFIGURATION
 ; #########################################
 
 [DEFAULT]
 ; Debug flag sets all loggers to debug, and enables request tracking
 debug = true
 
 ; ########################################################################
 ; EMAIL CONFIGURATION
 ; These settings will be used by the RhodeCode mailing system
 ; ########################################################################
 
 ; prefix all emails subjects with given prefix, helps filtering out emails
 #email_prefix = [RhodeCode]
 
 ; email FROM address all mails will be sent
 #app_email_from = rhodecode-noreply@localhost
 
 #smtp_server = mail.server.com
 #smtp_username =
 #smtp_password =
 #smtp_port =
 #smtp_use_tls = false
 #smtp_use_ssl = true
 
 [server:main]
 ; COMMON HOST/IP CONFIG, This applies mostly to develop setup,
 ; Host port for gunicorn are controlled by gunicorn_conf.py
 host = 127.0.0.1
 port = 10020
 
 
 ; ###########################
 ; GUNICORN APPLICATION SERVER
 ; ###########################
 
 ; run with gunicorn --paste rhodecode.ini --config gunicorn_conf.py
 
 ; Module to use, this setting shouldn't be changed
 use = egg:gunicorn#main
 
 ; Prefix middleware for RhodeCode.
 ; recommended when using proxy setup.
 ; allows to set RhodeCode under a prefix in server.
 ; eg https://server.com/custom_prefix. Enable `filter-with =` option below as well.
 ; And set your prefix like: `prefix = /custom_prefix`
 ; be sure to also set beaker.session.cookie_path = /custom_prefix if you need
 ; to make your cookies only work on prefix url
 [filter:proxy-prefix]
 use = egg:PasteDeploy#prefix
 prefix = /
 
 [app:main]
 ; The %(here)s variable will be replaced with the absolute path of parent directory
 ; of this file
 ; Each option in the app:main can be override by an environmental variable
 ;
 ;To override an option:
 ;
 ;RC_<KeyName>
 ;Everything should be uppercase, . and - should be replaced by _.
 ;For example, if you have these configuration settings:
 ;rc_cache.repo_object.backend = foo
 ;can be overridden by
 ;export RC_CACHE_REPO_OBJECT_BACKEND=foo
 
 use = egg:rhodecode-enterprise-ce
 
 ; enable proxy prefix middleware, defined above
 #filter-with = proxy-prefix
 
 ; #############
 ; DEBUG OPTIONS
 ; #############
 
 pyramid.reload_templates = true
 
 # During development the we want to have the debug toolbar enabled
 pyramid.includes =
     pyramid_debugtoolbar
 
 debugtoolbar.hosts = 0.0.0.0/0
 debugtoolbar.exclude_prefixes =
     /css
     /fonts
     /images
     /js
 
 ## RHODECODE PLUGINS ##
 rhodecode.includes =
     rhodecode.api
 
 
 # api prefix url
 rhodecode.api.url = /_admin/api
 
 ; enable debug style page
 debug_style = true
 
 ; #################
 ; END DEBUG OPTIONS
 ; #################
 
 ; encryption key used to encrypt social plugin tokens,
 ; remote_urls with credentials etc, if not set it defaults to
 ; `beaker.session.secret`
 #rhodecode.encrypted_values.secret =
 
 ; decryption strict mode (enabled by default). It controls if decryption raises
 ; `SignatureVerificationError` in case of wrong key, or damaged encryption data.
 #rhodecode.encrypted_values.strict = false
 
 ; Pick algorithm for encryption. Either fernet (more secure) or aes (default)
 ; fernet is safer, and we strongly recommend switching to it.
 ; Due to backward compatibility aes is used as default.
 #rhodecode.encrypted_values.algorithm = fernet
 
 ; Return gzipped responses from RhodeCode (static files/application)
 gzip_responses = false
 
 ; Auto-generate javascript routes file on startup
 generate_js_files = false
 
 ; System global default language.
 ; All available languages: en (default), be, de, es, fr, it, ja, pl, pt, ru, zh
 lang = en
 
 ; Perform a full repository scan and import on each server start.
 ; Settings this to true could lead to very long startup time.
 startup.import_repos = false
 
 ; URL at which the application is running. This is used for Bootstrapping
 ; requests in context when no web request is available. Used in ishell, or
 ; SSH calls. Set this for events to receive proper url for SSH calls.
 app.base_url = http://rhodecode.local
 
 ; Host at which the Service API is running.
 app.service_api.host = http://rhodecode.local:10020
 
 ; Secret for Service API authentication.
 app.service_api.token =
 
 ; Unique application ID. Should be a random unique string for security.
 app_instance_uuid = rc-production
 
 ; Cut off limit for large diffs (size in bytes). If overall diff size on
 ; commit, or pull request exceeds this limit this diff will be displayed
 ; partially. E.g 512000 == 512Kb
 cut_off_limit_diff = 512000
 
 ; Cut off limit for large files inside diffs (size in bytes). Each individual
 ; file inside diff which exceeds this limit will be displayed partially.
 ;  E.g 128000 == 128Kb
 cut_off_limit_file = 128000
 
 ; Use cached version of vcs repositories everywhere. Recommended to be `true`
 vcs_full_cache = true
 
 ; Force https in RhodeCode, fixes https redirects, assumes it's always https.
 ; Normally this is controlled by proper flags sent from http server such as Nginx or Apache
 force_https = false
 
 ; use Strict-Transport-Security headers
 use_htsts = false
 
 ; Set to true if your repos are exposed using the dumb protocol
 git_update_server_info = false
 
 ; RSS/ATOM feed options
 rss_cut_off_limit = 256000
 rss_items_per_page = 10
 rss_include_diff = false
 
 ; gist URL alias, used to create nicer urls for gist. This should be an
 ; url that does rewrites to _admin/gists/{gistid}.
 ; example: http://gist.rhodecode.org/{gistid}. Empty means use the internal
 ; RhodeCode url, ie. http[s]://rhodecode.server/_admin/gists/{gistid}
 gist_alias_url =
 
 ; List of views (using glob pattern syntax) that AUTH TOKENS could be
 ; used for access.
 ; Adding ?auth_token=TOKEN_HASH to the url authenticates this request as if it
 ; came from the the logged in user who own this authentication token.
 ; Additionally @TOKEN syntax can be used to bound the view to specific
 ; authentication token. Such view would be only accessible when used together
 ; with this authentication token
 ; list of all views can be found under `/_admin/permissions/auth_token_access`
 ; The list should be "," separated and on a single line.
 ; Most common views to enable:
 
 #    RepoCommitsView:repo_commit_download
 #    RepoCommitsView:repo_commit_patch
 #    RepoCommitsView:repo_commit_raw
 #    RepoCommitsView:repo_commit_raw@TOKEN
 #    RepoFilesView:repo_files_diff
 #    RepoFilesView:repo_archivefile
 #    RepoFilesView:repo_file_raw
 #    GistView:*
 api_access_controllers_whitelist =
 
 ; Default encoding used to convert from and to unicode
 ; can be also a comma separated list of encoding in case of mixed encodings
 default_encoding = UTF-8
 
 ; instance-id prefix
 ; a prefix key for this instance used for cache invalidation when running
 ; multiple instances of RhodeCode, make sure it's globally unique for
 ; all running RhodeCode instances. Leave empty if you don't use it
 instance_id =
 
 ; Fallback authentication plugin. Set this to a plugin ID to force the usage
 ; of an authentication plugin also if it is disabled by it's settings.
 ; This could be useful if you are unable to log in to the system due to broken
 ; authentication settings. Then you can enable e.g. the internal RhodeCode auth
 ; module to log in again and fix the settings.
 ; Available builtin plugin IDs (hash is part of the ID):
 ; egg:rhodecode-enterprise-ce#rhodecode
 ; egg:rhodecode-enterprise-ce#pam
 ; egg:rhodecode-enterprise-ce#ldap
 ; egg:rhodecode-enterprise-ce#jasig_cas
 ; egg:rhodecode-enterprise-ce#headers
 ; egg:rhodecode-enterprise-ce#crowd
 
 #rhodecode.auth_plugin_fallback = egg:rhodecode-enterprise-ce#rhodecode
 
 ; Flag to control loading of legacy plugins in py:/path format
 auth_plugin.import_legacy_plugins = true
 
 ; alternative return HTTP header for failed authentication. Default HTTP
 ; response is 401 HTTPUnauthorized. Currently HG clients have troubles with
 ; handling that causing a series of failed authentication calls.
 ; Set this variable to 403 to return HTTPForbidden, or any other HTTP code
 ; This will be served instead of default 401 on bad authentication
 auth_ret_code =
 
 ; use special detection method when serving auth_ret_code, instead of serving
 ; ret_code directly, use 401 initially (Which triggers credentials prompt)
 ; and then serve auth_ret_code to clients
 auth_ret_code_detection = false
 
 ; locking return code. When repository is locked return this HTTP code. 2XX
 ; codes don't break the transactions while 4XX codes do
 lock_ret_code = 423
 
 ; Filesystem location were repositories should be stored
 repo_store.path = /var/opt/rhodecode_repo_store
 
 ; allows to setup custom hooks in settings page
 allow_custom_hooks_settings = true
 
 ; Generated license token required for EE edition license.
 ; New generated token value can be found in Admin > settings > license page.
 license_token =
 
 ; This flag hides sensitive information on the license page such as token, and license data
 license.hide_license_info = false
 
 ; supervisor connection uri, for managing supervisor and logs.
 supervisor.uri =
 
 ; supervisord group name/id we only want this RC instance to handle
 supervisor.group_id = dev
 
 ; Display extended labs settings
 labs_settings_active = true
 
 ; Custom exception store path, defaults to TMPDIR
 ; This is used to store exception from RhodeCode in shared directory
 #exception_tracker.store_path =
 
 ; Send email with exception details when it happens
 #exception_tracker.send_email = false
 
 ; Comma separated list of recipients for exception emails,
 ; e.g [e-mail],[e-mail]
 ; Can be left empty, then emails will be sent to ALL super-admins
 #exception_tracker.send_email_recipients =
 
 ; optional prefix to Add to email Subject
 #exception_tracker.email_prefix = [RHODECODE ERROR]
 
 ; File store configuration. This is used to store and serve uploaded files
 file_store.enabled = true
 
 ; Storage backend, available options are: local
 file_store.backend = local
 
 ; path to store the uploaded binaries and artifacts
 file_store.storage_path = /var/opt/rhodecode_data/file_store
 
-; Uncomment and set this path to control settings for archive download cache.
+
+; Redis url to acquire/check generation of archives locks
+archive_cache.locking.url = redis://redis:6379/1
+
+; Storage backend, only 'filesystem' and 'objectstore' are available now
+archive_cache.backend.type = filesystem
+
+; url for s3 compatible storage that allows to upload artifacts
+; e.g http://minio:9000
+archive_cache.objectstore.url = http://s3-minio:9000
+
+; key for s3 auth
+archive_cache.objectstore.key = key
+
+; secret for s3 auth
+archive_cache.objectstore.secret = secret
+
+; number of sharded buckets to create to distribute archives across
+; default is 8 shards
+archive_cache.objectstore.bucket_shards = 8
+
+; if true, this cache will try to retry with retry_attempts=N times waiting retry_backoff time
+archive_cache.objectstore.retry = false
+
+; number of seconds to wait for next try using retry
+archive_cache.objectstore.retry_backoff = 1
+
+; how many tries do do a retry fetch from this backend
+archive_cache.objectstore.retry_attempts = 10
+
+; Default is $cache_dir/archive_cache if not set
 ; Generated repo archives will be cached at this location
 ; and served from the cache during subsequent requests for the same archive of
 ; the repository. This path is important to be shared across filesystems and with
 ; RhodeCode and vcsserver
-
-; Redis url to acquire/check generation of archives locks
-archive_cache.locking.url = redis://redis:6379/1
-
-; Storage backend, only 'filesystem' is available now
-archive_cache.backend.type = filesystem
-
-; Default is $cache_dir/archive_cache if not set
 archive_cache.filesystem.store_dir = /var/opt/rhodecode_data/archive_cache
 
 ; The limit in GB sets how much data we cache before recycling last used, defaults to 10 gb
 archive_cache.filesystem.cache_size_gb = 1
 
 ; Eviction policy used to clear out after cache_size_gb limit is reached
 archive_cache.filesystem.eviction_policy = least-recently-stored
 
 ; By default cache uses sharding technique, this specifies how many shards are there
+; default is 8 shards
 archive_cache.filesystem.cache_shards = 8
 
+; if true, this cache will try to retry with retry_attempts=N times waiting retry_backoff time
+archive_cache.filesystem.retry = false
+
+; number of seconds to wait for next try using retry
+archive_cache.filesystem.retry_backoff = 1
+
+; how many tries do do a retry fetch from this backend
+archive_cache.filesystem.retry_attempts = 10
+
 
 ; #############
 ; CELERY CONFIG
 ; #############
 
 ; manually run celery: /path/to/celery worker --task-events --beat --app rhodecode.lib.celerylib.loader --scheduler rhodecode.lib.celerylib.scheduler.RcScheduler --loglevel DEBUG --ini /path/to/rhodecode.ini
 
 use_celery = true
 
 ; path to store schedule database
 #celerybeat-schedule.path =
 
 ; connection url to the message broker (default redis)
 celery.broker_url = redis://redis:6379/8
 
 ; results backend to get results for (default redis)
 celery.result_backend = redis://redis:6379/8
 
 ; rabbitmq example
 #celery.broker_url = amqp://rabbitmq:qweqwe@localhost:5672/rabbitmqhost
 
 ; maximum tasks to execute before worker restart
 celery.max_tasks_per_child = 20
 
 ; tasks will never be sent to the queue, but executed locally instead.
 celery.task_always_eager = false
 
 ; #############
 ; DOGPILE CACHE
 ; #############
 
 ; Default cache dir for caches. Putting this into a ramdisk can boost performance.
 ; eg. /tmpfs/data_ramdisk, however this directory might require large amount of space
 cache_dir = /var/opt/rhodecode_data
 
 ; *********************************************
 ; `sql_cache_short` cache for heavy SQL queries
 ; Only supported backend is `memory_lru`
 ; *********************************************
 rc_cache.sql_cache_short.backend = dogpile.cache.rc.memory_lru
 rc_cache.sql_cache_short.expiration_time = 30
 
 
 ; *****************************************************
 ; `cache_repo_longterm` cache for repo object instances
 ; Only supported backend is `memory_lru`
 ; *****************************************************
 rc_cache.cache_repo_longterm.backend = dogpile.cache.rc.memory_lru
 ; by default we use 30 Days, cache is still invalidated on push
 rc_cache.cache_repo_longterm.expiration_time = 2592000
 ; max items in LRU cache, set to smaller number to save memory, and expire last used caches
 rc_cache.cache_repo_longterm.max_size = 10000
 
 
 ; *********************************************
 ; `cache_general` cache for general purpose use
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *********************************************
 rc_cache.cache_general.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_general.expiration_time = 43200
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_general.arguments.filename = /tmp/cache_general_db
 
 ; alternative `cache_general` redis backend with distributed lock
 #rc_cache.cache_general.backend = dogpile.cache.rc.redis
 #rc_cache.cache_general.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_general.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_general.arguments.host = localhost
 #rc_cache.cache_general.arguments.port = 6379
 #rc_cache.cache_general.arguments.db = 0
 #rc_cache.cache_general.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_general.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_general.arguments.lock_auto_renewal = true
 
 ; *************************************************
 ; `cache_perms` cache for permission tree, auth TTL
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *************************************************
 rc_cache.cache_perms.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_perms.expiration_time = 3600
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_perms.arguments.filename = /tmp/cache_perms_db
 
 ; alternative `cache_perms` redis backend with distributed lock
 #rc_cache.cache_perms.backend = dogpile.cache.rc.redis
 #rc_cache.cache_perms.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_perms.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_perms.arguments.host = localhost
 #rc_cache.cache_perms.arguments.port = 6379
 #rc_cache.cache_perms.arguments.db = 0
 #rc_cache.cache_perms.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_perms.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_perms.arguments.lock_auto_renewal = true
 
 ; ***************************************************
 ; `cache_repo` cache for file tree, Readme, RSS FEEDS
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; ***************************************************
 rc_cache.cache_repo.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_repo.expiration_time = 2592000
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_repo.arguments.filename = /tmp/cache_repo_db
 
 ; alternative `cache_repo` redis backend with distributed lock
 #rc_cache.cache_repo.backend = dogpile.cache.rc.redis
 #rc_cache.cache_repo.expiration_time = 2592000
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_repo.arguments.redis_expiration_time = 2678400
 
 #rc_cache.cache_repo.arguments.host = localhost
 #rc_cache.cache_repo.arguments.port = 6379
 #rc_cache.cache_repo.arguments.db = 1
 #rc_cache.cache_repo.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_repo.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_repo.arguments.lock_auto_renewal = true
 
 ; ##############
 ; BEAKER SESSION
 ; ##############
 
 ; beaker.session.type is type of storage options for the logged users sessions. Current allowed
 ; types are file, ext:redis, ext:database, ext:memcached
 ; Fastest ones are ext:redis and ext:database, DO NOT use memory type for session
 #beaker.session.type = file
 #beaker.session.data_dir = %(here)s/data/sessions
 
 ; Redis based sessions
 beaker.session.type = ext:redis
 beaker.session.url = redis://redis:6379/2
 
 ; DB based session, fast, and allows easy management over logged in users
 #beaker.session.type = ext:database
 #beaker.session.table_name = db_session
 #beaker.session.sa.url = postgresql://postgres:secret@localhost/rhodecode
 #beaker.session.sa.url = mysql://root:secret@127.0.0.1/rhodecode
 #beaker.session.sa.pool_recycle = 3600
 #beaker.session.sa.echo = false
 
 beaker.session.key = rhodecode
 beaker.session.secret = develop-rc-uytcxaz
 beaker.session.lock_dir = /data_ramdisk/lock
 
 ; Secure encrypted cookie. Requires AES and AES python libraries
 ; you must disable beaker.session.secret to use this
 #beaker.session.encrypt_key = key_for_encryption
 #beaker.session.validate_key = validation_key
 
 ; Sets session as invalid (also logging out user) if it haven not been
 ; accessed for given amount of time in seconds
 beaker.session.timeout = 2592000
 beaker.session.httponly = true
 
 ; Path to use for the cookie. Set to prefix if you use prefix middleware
 #beaker.session.cookie_path = /custom_prefix
 
 ; Set https secure cookie
 beaker.session.secure = false
 
 ; default cookie expiration time in seconds, set to `true` to set expire
 ; at browser close
 #beaker.session.cookie_expires = 3600
 
 ; #############################
 ; SEARCH INDEXING CONFIGURATION
 ; #############################
 
 ; Full text search indexer is available in rhodecode-tools under
 ; `rhodecode-tools index` command
 
 ; WHOOSH Backend, doesn't require additional services to run
 ; it works good with few dozen repos
 search.module = rhodecode.lib.index.whoosh
 search.location = %(here)s/data/index
 
 ; ####################
 ; CHANNELSTREAM CONFIG
 ; ####################
 
 ; channelstream enables persistent connections and live notification
 ; in the system. It's also used by the chat system
 
 channelstream.enabled = true
 
 ; server address for channelstream server on the backend
 channelstream.server = channelstream:9800
 
 ; location of the channelstream server from outside world
 ; use ws:// for http or wss:// for https. This address needs to be handled
 ; by external HTTP server such as Nginx or Apache
 ; see Nginx/Apache configuration examples in our docs
 channelstream.ws_url = ws://rhodecode.yourserver.com/_channelstream
 channelstream.secret = ENV_GENERATED
 channelstream.history.location = /var/opt/rhodecode_data/channelstream_history
 
 ; Internal application path that Javascript uses to connect into.
 ; If you use proxy-prefix the prefix should be added before /_channelstream
 channelstream.proxy_path = /_channelstream
 
 
 ; ##############################
 ; MAIN RHODECODE DATABASE CONFIG
 ; ##############################
 
 #sqlalchemy.db1.url = sqlite:///%(here)s/rhodecode.db?timeout=30
 #sqlalchemy.db1.url = postgresql://postgres:qweqwe@localhost/rhodecode
 #sqlalchemy.db1.url = mysql://root:qweqwe@localhost/rhodecode?charset=utf8
 ; pymysql is an alternative driver for MySQL, use in case of problems with default one
 #sqlalchemy.db1.url = mysql+pymysql://root:qweqwe@localhost/rhodecode
 
 sqlalchemy.db1.url = sqlite:///%(here)s/rhodecode.db?timeout=30
 
 ; see sqlalchemy docs for other advanced settings
 ; print the sql statements to output
 sqlalchemy.db1.echo = false
 
 ; recycle the connections after this amount of seconds
 sqlalchemy.db1.pool_recycle = 3600
 
 ; the number of connections to keep open inside the connection pool.
 ; 0 indicates no limit
 ; the general calculus with gevent is:
 ; if your system allows 500 concurrent greenlets (max_connections) that all do database access,
 ; then increase pool size + max overflow so that they add up to 500.
 #sqlalchemy.db1.pool_size = 5
 
 ; The number of connections to allow in connection pool "overflow", that is
 ; connections that can be opened above and beyond the pool_size setting,
 ; which defaults to five.
 #sqlalchemy.db1.max_overflow = 10
 
 ; Connection check ping, used to detect broken database connections
 ; could be enabled to better handle cases if MySQL has gone away errors
 #sqlalchemy.db1.ping_connection = true
 
 ; ##########
 ; VCS CONFIG
 ; ##########
 vcs.server.enable = true
 vcs.server = vcsserver:10010
 
 ; Web server connectivity protocol, responsible for web based VCS operations
 ; Available protocols are:
 ; `http` - use http-rpc backend (default)
 vcs.server.protocol = http
 
 ; Push/Pull operations protocol, available options are:
 ; `http` - use http-rpc backend (default)
 vcs.scm_app_implementation = http
 
 ; Push/Pull operations hooks protocol, available options are:
 ; `http` - use http-rpc backend (default)
 ; `celery` - use celery based hooks
 vcs.hooks.protocol = http
 
 ; Host on which this instance is listening for hooks. vcsserver will call this host to pull/push hooks so it should be
 ; accessible via network.
 ; Use vcs.hooks.host = "*" to bind to current hostname (for Docker)
 vcs.hooks.host = *
 
 ; Start VCSServer with this instance as a subprocess, useful for development
 vcs.start_server = false
 
 ; List of enabled VCS backends, available options are:
 ; `hg`  - mercurial
 ; `git` - git
 ; `svn` - subversion
 vcs.backends = hg, git, svn
 
 ; Wait this number of seconds before killing connection to the vcsserver
 vcs.connection_timeout = 3600
 
 ; Cache flag to cache vcsserver remote calls locally
 ; It uses cache_region `cache_repo`
 vcs.methods.cache = true
 
 ; ####################################################
 ; Subversion proxy support (mod_dav_svn)
 ; Maps RhodeCode repo groups into SVN paths for Apache
 ; ####################################################
 
 ; Compatibility version when creating SVN repositories. Defaults to newest version when commented out.
 ; Set a numeric version for your current SVN e.g 1.8, or 1.12
 ; Legacy available options are: pre-1.4-compatible, pre-1.5-compatible, pre-1.6-compatible, pre-1.8-compatible, pre-1.9-compatible
 #vcs.svn.compatible_version = 1.8
 
 ; Enable SVN proxy of requests over HTTP
 vcs.svn.proxy.enabled = true
 
 ; host to connect to running SVN subsystem
 vcs.svn.proxy.host = http://svn:8090
 
 ; Enable or disable the config file generation.
 svn.proxy.generate_config = true
 
 ; Generate config file with `SVNListParentPath` set to `On`.
 svn.proxy.list_parent_path = true
 
 ; Set location and file name of generated config file.
 svn.proxy.config_file_path = /etc/rhodecode/conf/svn/mod_dav_svn.conf
 
 ; alternative mod_dav config template. This needs to be a valid mako template
 ; Example template can be found in the source code:
 ; rhodecode/apps/svn_support/templates/mod-dav-svn.conf.mako
 #svn.proxy.config_template = ~/.rccontrol/enterprise-1/custom_svn_conf.mako
 
 ; Used as a prefix to the `Location` block in the generated config file.
 ; In most cases it should be set to `/`.
 svn.proxy.location_root = /
 
 ; Command to reload the mod dav svn configuration on change.
 ; Example: `/etc/init.d/apache2 reload` or /home/USER/apache_reload.sh
 ; Make sure user who runs RhodeCode process is allowed to reload Apache
 #svn.proxy.reload_cmd = /etc/init.d/apache2 reload
 
 ; If the timeout expires before the reload command finishes, the command will
 ; be killed. Setting it to zero means no timeout. Defaults to 10 seconds.
 #svn.proxy.reload_timeout = 10
 
 ; ####################
 ; SSH Support Settings
 ; ####################
 
 ; Defines if a custom authorized_keys file should be created and written on
 ; any change user ssh keys. Setting this to false also disables possibility
 ; of adding SSH keys by users from web interface. Super admins can still
 ; manage SSH Keys.
 ssh.generate_authorized_keyfile = true
 
 ; Options for ssh, default is `no-pty,no-port-forwarding,no-X11-forwarding,no-agent-forwarding`
 # ssh.authorized_keys_ssh_opts =
 
 ; Path to the authorized_keys file where the generate entries are placed.
 ; It is possible to have multiple key files specified in `sshd_config` e.g.
 ; AuthorizedKeysFile %h/.ssh/authorized_keys %h/.ssh/authorized_keys_rhodecode
 ssh.authorized_keys_file_path = /etc/rhodecode/conf/ssh/authorized_keys_rhodecode
 
 ; Command to execute the SSH wrapper. The binary is available in the
 ; RhodeCode installation directory.
 ; legacy: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 ; new rewrite: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper-v2
 ssh.wrapper_cmd = /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 
 ; Allow shell when executing the ssh-wrapper command
 ssh.wrapper_cmd_allow_shell = false
 
 ; Enables logging, and detailed output send back to the client during SSH
 ; operations. Useful for debugging, shouldn't be used in production.
 ssh.enable_debug_logging = true
 
 ; Paths to binary executable, by default they are the names, but we can
 ; override them if we want to use a custom one
 ssh.executable.hg = /usr/local/bin/rhodecode_bin/vcs_bin/hg
 ssh.executable.git = /usr/local/bin/rhodecode_bin/vcs_bin/git
 ssh.executable.svn = /usr/local/bin/rhodecode_bin/vcs_bin/svnserve
 
 ; Enables SSH key generator web interface. Disabling this still allows users
 ; to add their own keys.
 ssh.enable_ui_key_generator = true
 
 ; Statsd client config, this is used to send metrics to statsd
 ; We recommend setting statsd_exported and scrape them using Prometheus
 #statsd.enabled = false
 #statsd.statsd_host = 0.0.0.0
 #statsd.statsd_port = 8125
 #statsd.statsd_prefix =
 #statsd.statsd_ipv6 = false
 
 ; configure logging automatically at server startup set to false
 ; to use the below custom logging config.
 ; RC_LOGGING_FORMATTER
 ; RC_LOGGING_LEVEL
 ; env variables can control the settings for logging in case of autoconfigure
 
 #logging.autoconfigure = true
 
 ; specify your own custom logging config file to configure logging
 #logging.logging_conf_file = /path/to/custom_logging.ini
 
 ; Dummy marker to add new entries after.
 ; Add any custom entries below. Please don't remove this marker.
 custom.conf = 1
 
 
 ; #####################
 ; LOGGING CONFIGURATION
 ; #####################
 
 [loggers]
 keys = root, sqlalchemy, beaker, celery, rhodecode, ssh_wrapper
 
 [handlers]
 keys = console, console_sql
 
 [formatters]
 keys = generic, json, color_formatter, color_formatter_sql
 
 ; #######
 ; LOGGERS
 ; #######
 [logger_root]
 level = NOTSET
 handlers = console
 
 [logger_sqlalchemy]
 level = INFO
 handlers = console_sql
 qualname = sqlalchemy.engine
 propagate = 0
 
 [logger_beaker]
 level = DEBUG
 handlers =
 qualname = beaker.container
 propagate = 1
 
 [logger_rhodecode]
 level = DEBUG
 handlers =
 qualname = rhodecode
 propagate = 1
 
 [logger_ssh_wrapper]
 level = DEBUG
 handlers =
 qualname = ssh_wrapper
 propagate = 1
 
 [logger_celery]
 level = DEBUG
 handlers =
 qualname = celery
 
 
 ; ########
 ; HANDLERS
 ; ########
 
 [handler_console]
 class = StreamHandler
 args = (sys.stderr, )
 level = DEBUG
 ; To enable JSON formatted logs replace 'generic/color_formatter' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = color_formatter
 
 [handler_console_sql]
 ; "level = DEBUG" logs SQL queries and results.
 ; "level = INFO" logs SQL queries.
 ; "level = WARN" logs neither.  (Recommended for production systems.)
 class = StreamHandler
 args = (sys.stderr, )
 level = WARN
 ; To enable JSON formatted logs replace 'generic/color_formatter_sql' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = color_formatter_sql
 
 ; ##########
 ; FORMATTERS
 ; ##########
 
 [formatter_generic]
 class = rhodecode.lib.logging_formatter.ExceptionAwareFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter]
 class = rhodecode.lib.logging_formatter.ColorFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter_sql]
 class = rhodecode.lib.logging_formatter.ColorFormatterSql
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_json]
 format = %(timestamp)s %(levelname)s %(name)s %(message)s %(req_id)s
 class = rhodecode.lib._vendor.jsonlogger.JsonFormatter

configs/production.ini

@@ -1,781 +1,813 b''
 
 ; #########################################
 ; RHODECODE COMMUNITY EDITION CONFIGURATION
 ; #########################################
 
 [DEFAULT]
 ; Debug flag sets all loggers to debug, and enables request tracking
 debug = false
 
 ; ########################################################################
 ; EMAIL CONFIGURATION
 ; These settings will be used by the RhodeCode mailing system
 ; ########################################################################
 
 ; prefix all emails subjects with given prefix, helps filtering out emails
 #email_prefix = [RhodeCode]
 
 ; email FROM address all mails will be sent
 #app_email_from = rhodecode-noreply@localhost
 
 #smtp_server = mail.server.com
 #smtp_username =
 #smtp_password =
 #smtp_port =
 #smtp_use_tls = false
 #smtp_use_ssl = true
 
 [server:main]
 ; COMMON HOST/IP CONFIG, This applies mostly to develop setup,
 ; Host port for gunicorn are controlled by gunicorn_conf.py
 host = 127.0.0.1
 port = 10020
 
 
 ; ###########################
 ; GUNICORN APPLICATION SERVER
 ; ###########################
 
 ; run with gunicorn --paste rhodecode.ini --config gunicorn_conf.py
 
 ; Module to use, this setting shouldn't be changed
 use = egg:gunicorn#main
 
 ; Prefix middleware for RhodeCode.
 ; recommended when using proxy setup.
 ; allows to set RhodeCode under a prefix in server.
 ; eg https://server.com/custom_prefix. Enable `filter-with =` option below as well.
 ; And set your prefix like: `prefix = /custom_prefix`
 ; be sure to also set beaker.session.cookie_path = /custom_prefix if you need
 ; to make your cookies only work on prefix url
 [filter:proxy-prefix]
 use = egg:PasteDeploy#prefix
 prefix = /
 
 [app:main]
 ; The %(here)s variable will be replaced with the absolute path of parent directory
 ; of this file
 ; Each option in the app:main can be override by an environmental variable
 ;
 ;To override an option:
 ;
 ;RC_<KeyName>
 ;Everything should be uppercase, . and - should be replaced by _.
 ;For example, if you have these configuration settings:
 ;rc_cache.repo_object.backend = foo
 ;can be overridden by
 ;export RC_CACHE_REPO_OBJECT_BACKEND=foo
 
 use = egg:rhodecode-enterprise-ce
 
 ; enable proxy prefix middleware, defined above
 #filter-with = proxy-prefix
 
 ; encryption key used to encrypt social plugin tokens,
 ; remote_urls with credentials etc, if not set it defaults to
 ; `beaker.session.secret`
 #rhodecode.encrypted_values.secret =
 
 ; decryption strict mode (enabled by default). It controls if decryption raises
 ; `SignatureVerificationError` in case of wrong key, or damaged encryption data.
 #rhodecode.encrypted_values.strict = false
 
 ; Pick algorithm for encryption. Either fernet (more secure) or aes (default)
 ; fernet is safer, and we strongly recommend switching to it.
 ; Due to backward compatibility aes is used as default.
 #rhodecode.encrypted_values.algorithm = fernet
 
 ; Return gzipped responses from RhodeCode (static files/application)
 gzip_responses = false
 
 ; Auto-generate javascript routes file on startup
 generate_js_files = false
 
 ; System global default language.
 ; All available languages: en (default), be, de, es, fr, it, ja, pl, pt, ru, zh
 lang = en
 
 ; Perform a full repository scan and import on each server start.
 ; Settings this to true could lead to very long startup time.
 startup.import_repos = false
 
 ; URL at which the application is running. This is used for Bootstrapping
 ; requests in context when no web request is available. Used in ishell, or
 ; SSH calls. Set this for events to receive proper url for SSH calls.
 app.base_url = http://rhodecode.local
 
 ; Host at which the Service API is running.
 app.service_api.host = http://rhodecode.local:10020
 
 ; Secret for Service API authentication.
 app.service_api.token =
 
 ; Unique application ID. Should be a random unique string for security.
 app_instance_uuid = rc-production
 
 ; Cut off limit for large diffs (size in bytes). If overall diff size on
 ; commit, or pull request exceeds this limit this diff will be displayed
 ; partially. E.g 512000 == 512Kb
 cut_off_limit_diff = 512000
 
 ; Cut off limit for large files inside diffs (size in bytes). Each individual
 ; file inside diff which exceeds this limit will be displayed partially.
 ;  E.g 128000 == 128Kb
 cut_off_limit_file = 128000
 
 ; Use cached version of vcs repositories everywhere. Recommended to be `true`
 vcs_full_cache = true
 
 ; Force https in RhodeCode, fixes https redirects, assumes it's always https.
 ; Normally this is controlled by proper flags sent from http server such as Nginx or Apache
 force_https = false
 
 ; use Strict-Transport-Security headers
 use_htsts = false
 
 ; Set to true if your repos are exposed using the dumb protocol
 git_update_server_info = false
 
 ; RSS/ATOM feed options
 rss_cut_off_limit = 256000
 rss_items_per_page = 10
 rss_include_diff = false
 
 ; gist URL alias, used to create nicer urls for gist. This should be an
 ; url that does rewrites to _admin/gists/{gistid}.
 ; example: http://gist.rhodecode.org/{gistid}. Empty means use the internal
 ; RhodeCode url, ie. http[s]://rhodecode.server/_admin/gists/{gistid}
 gist_alias_url =
 
 ; List of views (using glob pattern syntax) that AUTH TOKENS could be
 ; used for access.
 ; Adding ?auth_token=TOKEN_HASH to the url authenticates this request as if it
 ; came from the the logged in user who own this authentication token.
 ; Additionally @TOKEN syntax can be used to bound the view to specific
 ; authentication token. Such view would be only accessible when used together
 ; with this authentication token
 ; list of all views can be found under `/_admin/permissions/auth_token_access`
 ; The list should be "," separated and on a single line.
 ; Most common views to enable:
 
 #    RepoCommitsView:repo_commit_download
 #    RepoCommitsView:repo_commit_patch
 #    RepoCommitsView:repo_commit_raw
 #    RepoCommitsView:repo_commit_raw@TOKEN
 #    RepoFilesView:repo_files_diff
 #    RepoFilesView:repo_archivefile
 #    RepoFilesView:repo_file_raw
 #    GistView:*
 api_access_controllers_whitelist =
 
 ; Default encoding used to convert from and to unicode
 ; can be also a comma separated list of encoding in case of mixed encodings
 default_encoding = UTF-8
 
 ; instance-id prefix
 ; a prefix key for this instance used for cache invalidation when running
 ; multiple instances of RhodeCode, make sure it's globally unique for
 ; all running RhodeCode instances. Leave empty if you don't use it
 instance_id =
 
 ; Fallback authentication plugin. Set this to a plugin ID to force the usage
 ; of an authentication plugin also if it is disabled by it's settings.
 ; This could be useful if you are unable to log in to the system due to broken
 ; authentication settings. Then you can enable e.g. the internal RhodeCode auth
 ; module to log in again and fix the settings.
 ; Available builtin plugin IDs (hash is part of the ID):
 ; egg:rhodecode-enterprise-ce#rhodecode
 ; egg:rhodecode-enterprise-ce#pam
 ; egg:rhodecode-enterprise-ce#ldap
 ; egg:rhodecode-enterprise-ce#jasig_cas
 ; egg:rhodecode-enterprise-ce#headers
 ; egg:rhodecode-enterprise-ce#crowd
 
 #rhodecode.auth_plugin_fallback = egg:rhodecode-enterprise-ce#rhodecode
 
 ; Flag to control loading of legacy plugins in py:/path format
 auth_plugin.import_legacy_plugins = true
 
 ; alternative return HTTP header for failed authentication. Default HTTP
 ; response is 401 HTTPUnauthorized. Currently HG clients have troubles with
 ; handling that causing a series of failed authentication calls.
 ; Set this variable to 403 to return HTTPForbidden, or any other HTTP code
 ; This will be served instead of default 401 on bad authentication
 auth_ret_code =
 
 ; use special detection method when serving auth_ret_code, instead of serving
 ; ret_code directly, use 401 initially (Which triggers credentials prompt)
 ; and then serve auth_ret_code to clients
 auth_ret_code_detection = false
 
 ; locking return code. When repository is locked return this HTTP code. 2XX
 ; codes don't break the transactions while 4XX codes do
 lock_ret_code = 423
 
 ; Filesystem location were repositories should be stored
 repo_store.path = /var/opt/rhodecode_repo_store
 
 ; allows to setup custom hooks in settings page
 allow_custom_hooks_settings = true
 
 ; Generated license token required for EE edition license.
 ; New generated token value can be found in Admin > settings > license page.
 license_token =
 
 ; This flag hides sensitive information on the license page such as token, and license data
 license.hide_license_info = false
 
 ; supervisor connection uri, for managing supervisor and logs.
 supervisor.uri =
 
 ; supervisord group name/id we only want this RC instance to handle
 supervisor.group_id = prod
 
 ; Display extended labs settings
 labs_settings_active = true
 
 ; Custom exception store path, defaults to TMPDIR
 ; This is used to store exception from RhodeCode in shared directory
 #exception_tracker.store_path =
 
 ; Send email with exception details when it happens
 #exception_tracker.send_email = false
 
 ; Comma separated list of recipients for exception emails,
 ; e.g [e-mail],[e-mail]
 ; Can be left empty, then emails will be sent to ALL super-admins
 #exception_tracker.send_email_recipients =
 
 ; optional prefix to Add to email Subject
 #exception_tracker.email_prefix = [RHODECODE ERROR]
 
 ; File store configuration. This is used to store and serve uploaded files
 file_store.enabled = true
 
 ; Storage backend, available options are: local
 file_store.backend = local
 
 ; path to store the uploaded binaries and artifacts
 file_store.storage_path = /var/opt/rhodecode_data/file_store
 
-; Uncomment and set this path to control settings for archive download cache.
+
+; Redis url to acquire/check generation of archives locks
+archive_cache.locking.url = redis://redis:6379/1
+
+; Storage backend, only 'filesystem' and 'objectstore' are available now
+archive_cache.backend.type = filesystem
+
+; url for s3 compatible storage that allows to upload artifacts
+; e.g http://minio:9000
+archive_cache.objectstore.url = http://s3-minio:9000
+
+; key for s3 auth
+archive_cache.objectstore.key = key
+
+; secret for s3 auth
+archive_cache.objectstore.secret = secret
+
+; number of sharded buckets to create to distribute archives across
+; default is 8 shards
+archive_cache.objectstore.bucket_shards = 8
+
+; if true, this cache will try to retry with retry_attempts=N times waiting retry_backoff time
+archive_cache.objectstore.retry = false
+
+; number of seconds to wait for next try using retry
+archive_cache.objectstore.retry_backoff = 1
+
+; how many tries do do a retry fetch from this backend
+archive_cache.objectstore.retry_attempts = 10
+
+; Default is $cache_dir/archive_cache if not set
 ; Generated repo archives will be cached at this location
 ; and served from the cache during subsequent requests for the same archive of
 ; the repository. This path is important to be shared across filesystems and with
 ; RhodeCode and vcsserver
-
-; Redis url to acquire/check generation of archives locks
-archive_cache.locking.url = redis://redis:6379/1
-
-; Storage backend, only 'filesystem' is available now
-archive_cache.backend.type = filesystem
-
-; Default is $cache_dir/archive_cache if not set
 archive_cache.filesystem.store_dir = /var/opt/rhodecode_data/archive_cache
 
 ; The limit in GB sets how much data we cache before recycling last used, defaults to 10 gb
 archive_cache.filesystem.cache_size_gb = 40
 
 ; Eviction policy used to clear out after cache_size_gb limit is reached
 archive_cache.filesystem.eviction_policy = least-recently-stored
 
 ; By default cache uses sharding technique, this specifies how many shards are there
+; default is 8 shards
 archive_cache.filesystem.cache_shards = 8
 
+; if true, this cache will try to retry with retry_attempts=N times waiting retry_backoff time
+archive_cache.filesystem.retry = false
+
+; number of seconds to wait for next try using retry
+archive_cache.filesystem.retry_backoff = 1
+
+; how many tries do do a retry fetch from this backend
+archive_cache.filesystem.retry_attempts = 10
+
 
 ; #############
 ; CELERY CONFIG
 ; #############
 
 ; manually run celery: /path/to/celery worker --task-events --beat --app rhodecode.lib.celerylib.loader --scheduler rhodecode.lib.celerylib.scheduler.RcScheduler --loglevel DEBUG --ini /path/to/rhodecode.ini
 
 use_celery = true
 
 ; path to store schedule database
 #celerybeat-schedule.path =
 
 ; connection url to the message broker (default redis)
 celery.broker_url = redis://redis:6379/8
 
 ; results backend to get results for (default redis)
 celery.result_backend = redis://redis:6379/8
 
 ; rabbitmq example
 #celery.broker_url = amqp://rabbitmq:qweqwe@localhost:5672/rabbitmqhost
 
 ; maximum tasks to execute before worker restart
 celery.max_tasks_per_child = 20
 
 ; tasks will never be sent to the queue, but executed locally instead.
 celery.task_always_eager = false
 
 ; #############
 ; DOGPILE CACHE
 ; #############
 
 ; Default cache dir for caches. Putting this into a ramdisk can boost performance.
 ; eg. /tmpfs/data_ramdisk, however this directory might require large amount of space
 cache_dir = /var/opt/rhodecode_data
 
 ; *********************************************
 ; `sql_cache_short` cache for heavy SQL queries
 ; Only supported backend is `memory_lru`
 ; *********************************************
 rc_cache.sql_cache_short.backend = dogpile.cache.rc.memory_lru
 rc_cache.sql_cache_short.expiration_time = 30
 
 
 ; *****************************************************
 ; `cache_repo_longterm` cache for repo object instances
 ; Only supported backend is `memory_lru`
 ; *****************************************************
 rc_cache.cache_repo_longterm.backend = dogpile.cache.rc.memory_lru
 ; by default we use 30 Days, cache is still invalidated on push
 rc_cache.cache_repo_longterm.expiration_time = 2592000
 ; max items in LRU cache, set to smaller number to save memory, and expire last used caches
 rc_cache.cache_repo_longterm.max_size = 10000
 
 
 ; *********************************************
 ; `cache_general` cache for general purpose use
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *********************************************
 rc_cache.cache_general.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_general.expiration_time = 43200
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_general.arguments.filename = /tmp/cache_general_db
 
 ; alternative `cache_general` redis backend with distributed lock
 #rc_cache.cache_general.backend = dogpile.cache.rc.redis
 #rc_cache.cache_general.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_general.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_general.arguments.host = localhost
 #rc_cache.cache_general.arguments.port = 6379
 #rc_cache.cache_general.arguments.db = 0
 #rc_cache.cache_general.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_general.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_general.arguments.lock_auto_renewal = true
 
 ; *************************************************
 ; `cache_perms` cache for permission tree, auth TTL
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *************************************************
 rc_cache.cache_perms.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_perms.expiration_time = 3600
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_perms.arguments.filename = /tmp/cache_perms_db
 
 ; alternative `cache_perms` redis backend with distributed lock
 #rc_cache.cache_perms.backend = dogpile.cache.rc.redis
 #rc_cache.cache_perms.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_perms.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_perms.arguments.host = localhost
 #rc_cache.cache_perms.arguments.port = 6379
 #rc_cache.cache_perms.arguments.db = 0
 #rc_cache.cache_perms.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_perms.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_perms.arguments.lock_auto_renewal = true
 
 ; ***************************************************
 ; `cache_repo` cache for file tree, Readme, RSS FEEDS
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; ***************************************************
 rc_cache.cache_repo.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_repo.expiration_time = 2592000
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_repo.arguments.filename = /tmp/cache_repo_db
 
 ; alternative `cache_repo` redis backend with distributed lock
 #rc_cache.cache_repo.backend = dogpile.cache.rc.redis
 #rc_cache.cache_repo.expiration_time = 2592000
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_repo.arguments.redis_expiration_time = 2678400
 
 #rc_cache.cache_repo.arguments.host = localhost
 #rc_cache.cache_repo.arguments.port = 6379
 #rc_cache.cache_repo.arguments.db = 1
 #rc_cache.cache_repo.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_repo.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_repo.arguments.lock_auto_renewal = true
 
 ; ##############
 ; BEAKER SESSION
 ; ##############
 
 ; beaker.session.type is type of storage options for the logged users sessions. Current allowed
 ; types are file, ext:redis, ext:database, ext:memcached
 ; Fastest ones are ext:redis and ext:database, DO NOT use memory type for session
 #beaker.session.type = file
 #beaker.session.data_dir = %(here)s/data/sessions
 
 ; Redis based sessions
 beaker.session.type = ext:redis
 beaker.session.url = redis://redis:6379/2
 
 ; DB based session, fast, and allows easy management over logged in users
 #beaker.session.type = ext:database
 #beaker.session.table_name = db_session
 #beaker.session.sa.url = postgresql://postgres:secret@localhost/rhodecode
 #beaker.session.sa.url = mysql://root:secret@127.0.0.1/rhodecode
 #beaker.session.sa.pool_recycle = 3600
 #beaker.session.sa.echo = false
 
 beaker.session.key = rhodecode
 beaker.session.secret = production-rc-uytcxaz
 beaker.session.lock_dir = /data_ramdisk/lock
 
 ; Secure encrypted cookie. Requires AES and AES python libraries
 ; you must disable beaker.session.secret to use this
 #beaker.session.encrypt_key = key_for_encryption
 #beaker.session.validate_key = validation_key
 
 ; Sets session as invalid (also logging out user) if it haven not been
 ; accessed for given amount of time in seconds
 beaker.session.timeout = 2592000
 beaker.session.httponly = true
 
 ; Path to use for the cookie. Set to prefix if you use prefix middleware
 #beaker.session.cookie_path = /custom_prefix
 
 ; Set https secure cookie
 beaker.session.secure = false
 
 ; default cookie expiration time in seconds, set to `true` to set expire
 ; at browser close
 #beaker.session.cookie_expires = 3600
 
 ; #############################
 ; SEARCH INDEXING CONFIGURATION
 ; #############################
 
 ; Full text search indexer is available in rhodecode-tools under
 ; `rhodecode-tools index` command
 
 ; WHOOSH Backend, doesn't require additional services to run
 ; it works good with few dozen repos
 search.module = rhodecode.lib.index.whoosh
 search.location = %(here)s/data/index
 
 ; ####################
 ; CHANNELSTREAM CONFIG
 ; ####################
 
 ; channelstream enables persistent connections and live notification
 ; in the system. It's also used by the chat system
 
 channelstream.enabled = true
 
 ; server address for channelstream server on the backend
 channelstream.server = channelstream:9800
 
 ; location of the channelstream server from outside world
 ; use ws:// for http or wss:// for https. This address needs to be handled
 ; by external HTTP server such as Nginx or Apache
 ; see Nginx/Apache configuration examples in our docs
 channelstream.ws_url = ws://rhodecode.yourserver.com/_channelstream
 channelstream.secret = ENV_GENERATED
 channelstream.history.location = /var/opt/rhodecode_data/channelstream_history
 
 ; Internal application path that Javascript uses to connect into.
 ; If you use proxy-prefix the prefix should be added before /_channelstream
 channelstream.proxy_path = /_channelstream
 
 
 ; ##############################
 ; MAIN RHODECODE DATABASE CONFIG
 ; ##############################
 
 #sqlalchemy.db1.url = sqlite:///%(here)s/rhodecode.db?timeout=30
 #sqlalchemy.db1.url = postgresql://postgres:qweqwe@localhost/rhodecode
 #sqlalchemy.db1.url = mysql://root:qweqwe@localhost/rhodecode?charset=utf8
 ; pymysql is an alternative driver for MySQL, use in case of problems with default one
 #sqlalchemy.db1.url = mysql+pymysql://root:qweqwe@localhost/rhodecode
 
 sqlalchemy.db1.url = postgresql://postgres:qweqwe@localhost/rhodecode
 
 ; see sqlalchemy docs for other advanced settings
 ; print the sql statements to output
 sqlalchemy.db1.echo = false
 
 ; recycle the connections after this amount of seconds
 sqlalchemy.db1.pool_recycle = 3600
 
 ; the number of connections to keep open inside the connection pool.
 ; 0 indicates no limit
 ; the general calculus with gevent is:
 ; if your system allows 500 concurrent greenlets (max_connections) that all do database access,
 ; then increase pool size + max overflow so that they add up to 500.
 #sqlalchemy.db1.pool_size = 5
 
 ; The number of connections to allow in connection pool "overflow", that is
 ; connections that can be opened above and beyond the pool_size setting,
 ; which defaults to five.
 #sqlalchemy.db1.max_overflow = 10
 
 ; Connection check ping, used to detect broken database connections
 ; could be enabled to better handle cases if MySQL has gone away errors
 #sqlalchemy.db1.ping_connection = true
 
 ; ##########
 ; VCS CONFIG
 ; ##########
 vcs.server.enable = true
 vcs.server = vcsserver:10010
 
 ; Web server connectivity protocol, responsible for web based VCS operations
 ; Available protocols are:
 ; `http` - use http-rpc backend (default)
 vcs.server.protocol = http
 
 ; Push/Pull operations protocol, available options are:
 ; `http` - use http-rpc backend (default)
 vcs.scm_app_implementation = http
 
 ; Push/Pull operations hooks protocol, available options are:
 ; `http` - use http-rpc backend (default)
 ; `celery` - use celery based hooks
 vcs.hooks.protocol = http
 
 ; Host on which this instance is listening for hooks. vcsserver will call this host to pull/push hooks so it should be
 ; accessible via network.
 ; Use vcs.hooks.host = "*" to bind to current hostname (for Docker)
 vcs.hooks.host = *
 
 ; Start VCSServer with this instance as a subprocess, useful for development
 vcs.start_server = false
 
 ; List of enabled VCS backends, available options are:
 ; `hg`  - mercurial
 ; `git` - git
 ; `svn` - subversion
 vcs.backends = hg, git, svn
 
 ; Wait this number of seconds before killing connection to the vcsserver
 vcs.connection_timeout = 3600
 
 ; Cache flag to cache vcsserver remote calls locally
 ; It uses cache_region `cache_repo`
 vcs.methods.cache = true
 
 ; ####################################################
 ; Subversion proxy support (mod_dav_svn)
 ; Maps RhodeCode repo groups into SVN paths for Apache
 ; ####################################################
 
 ; Compatibility version when creating SVN repositories. Defaults to newest version when commented out.
 ; Set a numeric version for your current SVN e.g 1.8, or 1.12
 ; Legacy available options are: pre-1.4-compatible, pre-1.5-compatible, pre-1.6-compatible, pre-1.8-compatible, pre-1.9-compatible
 #vcs.svn.compatible_version = 1.8
 
 ; Enable SVN proxy of requests over HTTP
 vcs.svn.proxy.enabled = true
 
 ; host to connect to running SVN subsystem
 vcs.svn.proxy.host = http://svn:8090
 
 ; Enable or disable the config file generation.
 svn.proxy.generate_config = true
 
 ; Generate config file with `SVNListParentPath` set to `On`.
 svn.proxy.list_parent_path = true
 
 ; Set location and file name of generated config file.
 svn.proxy.config_file_path = /etc/rhodecode/conf/svn/mod_dav_svn.conf
 
 ; alternative mod_dav config template. This needs to be a valid mako template
 ; Example template can be found in the source code:
 ; rhodecode/apps/svn_support/templates/mod-dav-svn.conf.mako
 #svn.proxy.config_template = ~/.rccontrol/enterprise-1/custom_svn_conf.mako
 
 ; Used as a prefix to the `Location` block in the generated config file.
 ; In most cases it should be set to `/`.
 svn.proxy.location_root = /
 
 ; Command to reload the mod dav svn configuration on change.
 ; Example: `/etc/init.d/apache2 reload` or /home/USER/apache_reload.sh
 ; Make sure user who runs RhodeCode process is allowed to reload Apache
 #svn.proxy.reload_cmd = /etc/init.d/apache2 reload
 
 ; If the timeout expires before the reload command finishes, the command will
 ; be killed. Setting it to zero means no timeout. Defaults to 10 seconds.
 #svn.proxy.reload_timeout = 10
 
 ; ####################
 ; SSH Support Settings
 ; ####################
 
 ; Defines if a custom authorized_keys file should be created and written on
 ; any change user ssh keys. Setting this to false also disables possibility
 ; of adding SSH keys by users from web interface. Super admins can still
 ; manage SSH Keys.
 ssh.generate_authorized_keyfile = true
 
 ; Options for ssh, default is `no-pty,no-port-forwarding,no-X11-forwarding,no-agent-forwarding`
 # ssh.authorized_keys_ssh_opts =
 
 ; Path to the authorized_keys file where the generate entries are placed.
 ; It is possible to have multiple key files specified in `sshd_config` e.g.
 ; AuthorizedKeysFile %h/.ssh/authorized_keys %h/.ssh/authorized_keys_rhodecode
 ssh.authorized_keys_file_path = /etc/rhodecode/conf/ssh/authorized_keys_rhodecode
 
 ; Command to execute the SSH wrapper. The binary is available in the
 ; RhodeCode installation directory.
 ; legacy: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 ; new rewrite: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper-v2
 ssh.wrapper_cmd = /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 
 ; Allow shell when executing the ssh-wrapper command
 ssh.wrapper_cmd_allow_shell = false
 
 ; Enables logging, and detailed output send back to the client during SSH
 ; operations. Useful for debugging, shouldn't be used in production.
 ssh.enable_debug_logging = false
 
 ; Paths to binary executable, by default they are the names, but we can
 ; override them if we want to use a custom one
 ssh.executable.hg = /usr/local/bin/rhodecode_bin/vcs_bin/hg
 ssh.executable.git = /usr/local/bin/rhodecode_bin/vcs_bin/git
 ssh.executable.svn = /usr/local/bin/rhodecode_bin/vcs_bin/svnserve
 
 ; Enables SSH key generator web interface. Disabling this still allows users
 ; to add their own keys.
 ssh.enable_ui_key_generator = true
 
 ; Statsd client config, this is used to send metrics to statsd
 ; We recommend setting statsd_exported and scrape them using Prometheus
 #statsd.enabled = false
 #statsd.statsd_host = 0.0.0.0
 #statsd.statsd_port = 8125
 #statsd.statsd_prefix =
 #statsd.statsd_ipv6 = false
 
 ; configure logging automatically at server startup set to false
 ; to use the below custom logging config.
 ; RC_LOGGING_FORMATTER
 ; RC_LOGGING_LEVEL
 ; env variables can control the settings for logging in case of autoconfigure
 
 #logging.autoconfigure = true
 
 ; specify your own custom logging config file to configure logging
 #logging.logging_conf_file = /path/to/custom_logging.ini
 
 ; Dummy marker to add new entries after.
 ; Add any custom entries below. Please don't remove this marker.
 custom.conf = 1
 
 
 ; #####################
 ; LOGGING CONFIGURATION
 ; #####################
 
 [loggers]
 keys = root, sqlalchemy, beaker, celery, rhodecode, ssh_wrapper
 
 [handlers]
 keys = console, console_sql
 
 [formatters]
 keys = generic, json, color_formatter, color_formatter_sql
 
 ; #######
 ; LOGGERS
 ; #######
 [logger_root]
 level = NOTSET
 handlers = console
 
 [logger_sqlalchemy]
 level = INFO
 handlers = console_sql
 qualname = sqlalchemy.engine
 propagate = 0
 
 [logger_beaker]
 level = DEBUG
 handlers =
 qualname = beaker.container
 propagate = 1
 
 [logger_rhodecode]
 level = DEBUG
 handlers =
 qualname = rhodecode
 propagate = 1
 
 [logger_ssh_wrapper]
 level = DEBUG
 handlers =
 qualname = ssh_wrapper
 propagate = 1
 
 [logger_celery]
 level = DEBUG
 handlers =
 qualname = celery
 
 
 ; ########
 ; HANDLERS
 ; ########
 
 [handler_console]
 class = StreamHandler
 args = (sys.stderr, )
 level = INFO
 ; To enable JSON formatted logs replace 'generic/color_formatter' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = generic
 
 [handler_console_sql]
 ; "level = DEBUG" logs SQL queries and results.
 ; "level = INFO" logs SQL queries.
 ; "level = WARN" logs neither.  (Recommended for production systems.)
 class = StreamHandler
 args = (sys.stderr, )
 level = WARN
 ; To enable JSON formatted logs replace 'generic/color_formatter_sql' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = generic
 
 ; ##########
 ; FORMATTERS
 ; ##########
 
 [formatter_generic]
 class = rhodecode.lib.logging_formatter.ExceptionAwareFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter]
 class = rhodecode.lib.logging_formatter.ColorFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter_sql]
 class = rhodecode.lib.logging_formatter.ColorFormatterSql
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_json]
 format = %(timestamp)s %(levelname)s %(name)s %(message)s %(req_id)s
 class = rhodecode.lib._vendor.jsonlogger.JsonFormatter

requirements.txt

@@ -1,286 +1,315 b''
 # deps, generated via pipdeptree --exclude setuptools,wheel,pipdeptree,pip -f | tr '[:upper:]' '[:lower:]'
 
 alembic==1.13.1
   mako==1.2.4
     markupsafe==2.1.2
   sqlalchemy==1.4.52
     greenlet==3.0.3
   typing_extensions==4.9.0
 async-timeout==4.0.3
 babel==2.12.1
 beaker==1.12.1
 celery==5.3.6
   billiard==4.2.0
   click==8.1.3
   click-didyoumean==0.3.0
     click==8.1.3
   click-plugins==1.1.1
     click==8.1.3
   click-repl==0.2.0
     click==8.1.3
     prompt-toolkit==3.0.38
       wcwidth==0.2.6
     six==1.16.0
   kombu==5.3.5
     amqp==5.2.0
       vine==5.1.0
     vine==5.1.0
   python-dateutil==2.8.2
     six==1.16.0
   tzdata==2024.1
   vine==5.1.0
 channelstream==0.7.1
   gevent==24.2.1
     greenlet==3.0.3
     zope.event==5.0.0
     zope.interface==6.3.0
   itsdangerous==1.1.0
   marshmallow==2.18.0
   pyramid==2.0.2
     hupper==1.12
     plaster==1.1.2
     plaster-pastedeploy==1.0.1
       pastedeploy==3.1.0
       plaster==1.1.2
     translationstring==1.4
     venusian==3.0.0
     webob==1.8.7
     zope.deprecation==5.0.0
     zope.interface==6.3.0
   pyramid-apispec==0.3.3
     apispec==1.3.3
   pyramid-jinja2==2.10
     jinja2==3.1.2
       markupsafe==2.1.2
     markupsafe==2.1.2
     pyramid==2.0.2
       hupper==1.12
       plaster==1.1.2
       plaster-pastedeploy==1.0.1
         pastedeploy==3.1.0
         plaster==1.1.2
       translationstring==1.4
       venusian==3.0.0
       webob==1.8.7
       zope.deprecation==5.0.0
       zope.interface==6.3.0
     zope.deprecation==5.0.0
   python-dateutil==2.8.2
     six==1.16.0
   requests==2.28.2
     certifi==2022.12.7
     charset-normalizer==3.1.0
     idna==3.4
     urllib3==1.26.14
   ws4py==0.5.1
 deform==2.0.15
   chameleon==3.10.2
   colander==2.0
     iso8601==1.1.0
     translationstring==1.4
   iso8601==1.1.0
   peppercorn==0.6
   translationstring==1.4
   zope.deprecation==5.0.0
 docutils==0.19
 dogpile.cache==1.3.3
   decorator==5.1.1
   stevedore==5.1.0
     pbr==5.11.1
 formencode==2.1.0
   six==1.16.0
+fsspec==2024.6.0
 gunicorn==21.2.0
   packaging==24.0
 gevent==24.2.1
   greenlet==3.0.3
   zope.event==5.0.0
   zope.interface==6.3.0
 ipython==8.14.0
   backcall==0.2.0
   decorator==5.1.1
   jedi==0.19.0
     parso==0.8.3
   matplotlib-inline==0.1.6
     traitlets==5.9.0
   pexpect==4.8.0
     ptyprocess==0.7.0
   pickleshare==0.7.5
   prompt-toolkit==3.0.38
     wcwidth==0.2.6
   pygments==2.15.1
   stack-data==0.6.2
     asttokens==2.2.1
       six==1.16.0
     executing==1.2.0
     pure-eval==0.2.2
   traitlets==5.9.0
 markdown==3.4.3
 msgpack==1.0.8
 mysqlclient==2.1.1
 nbconvert==7.7.3
   beautifulsoup4==4.12.3
     soupsieve==2.5
   bleach==6.1.0
     six==1.16.0
     webencodings==0.5.1
   defusedxml==0.7.1
   jinja2==3.1.2
     markupsafe==2.1.2
   jupyter_core==5.3.1
     platformdirs==3.10.0
     traitlets==5.9.0
   jupyterlab-pygments==0.2.2
   markupsafe==2.1.2
   mistune==2.0.5
   nbclient==0.8.0
     jupyter_client==8.3.0
       jupyter_core==5.3.1
         platformdirs==3.10.0
         traitlets==5.9.0
       python-dateutil==2.8.2
         six==1.16.0
       pyzmq==25.0.0
       tornado==6.2
       traitlets==5.9.0
     jupyter_core==5.3.1
       platformdirs==3.10.0
       traitlets==5.9.0
     nbformat==5.9.2
       fastjsonschema==2.18.0
       jsonschema==4.18.6
         attrs==22.2.0
         pyrsistent==0.19.3
       jupyter_core==5.3.1
         platformdirs==3.10.0
         traitlets==5.9.0
       traitlets==5.9.0
     traitlets==5.9.0
   nbformat==5.9.2
     fastjsonschema==2.18.0
     jsonschema==4.18.6
       attrs==22.2.0
       pyrsistent==0.19.3
     jupyter_core==5.3.1
       platformdirs==3.10.0
       traitlets==5.9.0
     traitlets==5.9.0
   pandocfilters==1.5.0
   pygments==2.15.1
   tinycss2==1.2.1
     webencodings==0.5.1
   traitlets==5.9.0
 orjson==3.10.3
 paste==3.10.1
 premailer==3.10.0
   cachetools==5.3.3
   cssselect==1.2.0
   cssutils==2.6.0
   lxml==4.9.3
   requests==2.28.2
     certifi==2022.12.7
     charset-normalizer==3.1.0
     idna==3.4
     urllib3==1.26.14
 psutil==5.9.8
 psycopg2==2.9.9
 py-bcrypt==0.4
 pycmarkgfm==1.2.0
   cffi==1.16.0
     pycparser==2.21
 pycryptodome==3.17
 pycurl==7.45.3
 pymysql==1.0.3
 pyotp==2.8.0
 pyparsing==3.1.1
 pyramid-debugtoolbar==4.12.1
   pygments==2.15.1
   pyramid==2.0.2
     hupper==1.12
     plaster==1.1.2
     plaster-pastedeploy==1.0.1
       pastedeploy==3.1.0
       plaster==1.1.2
     translationstring==1.4
     venusian==3.0.0
     webob==1.8.7
     zope.deprecation==5.0.0
     zope.interface==6.3.0
   pyramid-mako==1.1.0
     mako==1.2.4
       markupsafe==2.1.2
     pyramid==2.0.2
       hupper==1.12
       plaster==1.1.2
       plaster-pastedeploy==1.0.1
         pastedeploy==3.1.0
         plaster==1.1.2
       translationstring==1.4
       venusian==3.0.0
       webob==1.8.7
       zope.deprecation==5.0.0
       zope.interface==6.3.0
 pyramid-mailer==0.15.1
   pyramid==2.0.2
     hupper==1.12
     plaster==1.1.2
     plaster-pastedeploy==1.0.1
       pastedeploy==3.1.0
       plaster==1.1.2
     translationstring==1.4
     venusian==3.0.0
     webob==1.8.7
     zope.deprecation==5.0.0
     zope.interface==6.3.0
   repoze.sendmail==4.4.1
     transaction==3.1.0
       zope.interface==6.3.0
     zope.interface==6.3.0
   transaction==3.1.0
     zope.interface==6.3.0
 python-ldap==3.4.3
   pyasn1==0.4.8
   pyasn1-modules==0.2.8
     pyasn1==0.4.8
 python-memcached==1.59
   six==1.16.0
 python-pam==2.0.2
 python3-saml==1.15.0
   isodate==0.6.1
     six==1.16.0
   lxml==4.9.3
   xmlsec==1.3.13
     lxml==4.9.3
 pyyaml==6.0.1
 redis==5.0.4
   async-timeout==4.0.3
 regex==2022.10.31
 routes==2.5.1
   repoze.lru==0.7
   six==1.16.0
+s3fs==2024.6.0
+  aiobotocore==2.13.0
+    aiohttp==3.9.5
+      aiosignal==1.3.1
+        frozenlist==1.4.1
+      attrs==22.2.0
+      frozenlist==1.4.1
+      multidict==6.0.5
+      yarl==1.9.4
+        idna==3.4
+        multidict==6.0.5
+    aioitertools==0.11.0
+    botocore==1.34.106
+      jmespath==1.0.1
+      python-dateutil==2.8.2
+        six==1.16.0
+      urllib3==1.26.14
+    wrapt==1.16.0
+  aiohttp==3.9.5
+    aiosignal==1.3.1
+      frozenlist==1.4.1
+    attrs==22.2.0
+    frozenlist==1.4.1
+    multidict==6.0.5
+    yarl==1.9.4
+      idna==3.4
+      multidict==6.0.5
+  fsspec==2024.6.0
 simplejson==3.19.2
 sshpubkeys==3.3.1
   cryptography==40.0.2
     cffi==1.16.0
       pycparser==2.21
   ecdsa==0.18.0
     six==1.16.0
 sqlalchemy==1.4.52
   greenlet==3.0.3
   typing_extensions==4.9.0
 supervisor==4.2.5
 tzlocal==4.3
   pytz-deprecation-shim==0.1.0.post0
     tzdata==2024.1
 tempita==0.5.2
 unidecode==1.3.6
 urlobject==2.4.3
 waitress==3.0.0
 webhelpers2==2.1
   markupsafe==2.1.2
   six==1.16.0
 whoosh==2.7.4
 zope.cachedescriptors==5.0.0
 qrcode==7.4.2
 
 ## uncomment to add the debug libraries
 #-r requirements_debug.txt

rhodecode/apps/repository/views/repo_files.py

@@ -1,1716 +1,1716 b''
 # Copyright (C) 2011-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import itertools
 import logging
 import os
 import collections
 import urllib.request
 import urllib.parse
 import urllib.error
 import pathlib
 import time
 import random
 
 from pyramid.httpexceptions import HTTPNotFound, HTTPBadRequest, HTTPFound
 
 from pyramid.renderers import render
 from pyramid.response import Response
 
 import rhodecode
 from rhodecode.apps._base import RepoAppView
 
 
 from rhodecode.lib import diffs, helpers as h, rc_cache
 from rhodecode.lib import audit_logger
 from rhodecode.lib.hash_utils import sha1_safe
-from rhodecode.lib.rc_cache.archive_cache import (
+from rhodecode.lib.archive_cache import (
     get_archival_cache_store, get_archival_config, ArchiveCacheGenerationLock, archive_iterator)
 from rhodecode.lib.str_utils import safe_bytes, convert_special_chars
 from rhodecode.lib.view_utils import parse_path_ref
 from rhodecode.lib.exceptions import NonRelativePathError
 from rhodecode.lib.codeblocks import (
     filenode_as_lines_tokens, filenode_as_annotated_lines_tokens)
 from rhodecode.lib.utils2 import convert_line_endings, detect_mode
 from rhodecode.lib.type_utils import str2bool
 from rhodecode.lib.str_utils import safe_str, safe_int
 from rhodecode.lib.auth import (
     LoginRequired, HasRepoPermissionAnyDecorator, CSRFRequired)
 from rhodecode.lib.vcs import path as vcspath
 from rhodecode.lib.vcs.backends.base import EmptyCommit
 from rhodecode.lib.vcs.conf import settings
 from rhodecode.lib.vcs.nodes import FileNode
 from rhodecode.lib.vcs.exceptions import (
     RepositoryError, CommitDoesNotExistError, EmptyRepositoryError,
     ImproperArchiveTypeError, VCSError, NodeAlreadyExistsError,
     NodeDoesNotExistError, CommitError, NodeError)
 
 from rhodecode.model.scm import ScmModel
 from rhodecode.model.db import Repository
 
 log = logging.getLogger(__name__)
 
 
 def get_archive_name(db_repo_id, db_repo_name, commit_sha, ext, subrepos=False, path_sha='', with_hash=True):
     # original backward compat name of archive
     clean_name = safe_str(convert_special_chars(db_repo_name).replace('/', '_'))
 
     # e.g vcsserver-id-abcd-sub-1-abcfdef-archive-all.zip
     # vcsserver-id-abcd-sub-0-abcfdef-COMMIT_SHA-PATH_SHA.zip
     id_sha = sha1_safe(str(db_repo_id))[:4]
     sub_repo = 'sub-1' if subrepos else 'sub-0'
     commit = commit_sha if with_hash else 'archive'
     path_marker = (path_sha if with_hash else '') or 'all'
     archive_name = f'{clean_name}-id-{id_sha}-{sub_repo}-{commit}-{path_marker}{ext}'
 
     return archive_name
 
 
 def get_path_sha(at_path):
     return safe_str(sha1_safe(at_path)[:8])
 
 
 def _get_archive_spec(fname):
     log.debug('Detecting archive spec for: `%s`', fname)
 
     fileformat = None
     ext = None
     content_type = None
     for a_type, content_type, extension in settings.ARCHIVE_SPECS:
 
         if fname.endswith(extension):
             fileformat = a_type
             log.debug('archive is of type: %s', fileformat)
             ext = extension
             break
 
     if not fileformat:
         raise ValueError()
 
     # left over part of whole fname is the commit
     commit_id = fname[:-len(ext)]
 
     return commit_id, ext, fileformat, content_type
 
 
 class RepoFilesView(RepoAppView):
 
     @staticmethod
     def adjust_file_path_for_svn(f_path, repo):
         """
         Computes the relative path of `f_path`.
 
         This is mainly based on prefix matching of the recognized tags and
         branches in the underlying repository.
         """
         tags_and_branches = itertools.chain(
             repo.branches.keys(),
             repo.tags.keys())
         tags_and_branches = sorted(tags_and_branches, key=len, reverse=True)
 
         for name in tags_and_branches:
             if f_path.startswith(f'{name}/'):
                 f_path = vcspath.relpath(f_path, name)
                 break
         return f_path
 
     def load_default_context(self):
         c = self._get_local_tmpl_context(include_app_defaults=True)
         c.rhodecode_repo = self.rhodecode_vcs_repo
         c.enable_downloads = self.db_repo.enable_downloads
         return c
 
     def _ensure_not_locked(self, commit_id='tip'):
         _ = self.request.translate
 
         repo = self.db_repo
         if repo.enable_locking and repo.locked[0]:
             h.flash(_('This repository has been locked by %s on %s')
                     % (h.person_by_id(repo.locked[0]),
                     h.format_date(h.time_to_datetime(repo.locked[1]))),
                     'warning')
             files_url = h.route_path(
                 'repo_files:default_path',
                 repo_name=self.db_repo_name, commit_id=commit_id)
             raise HTTPFound(files_url)
 
     def forbid_non_head(self, is_head, f_path, commit_id='tip', json_mode=False):
         _ = self.request.translate
 
         if not is_head:
             message = _('Cannot modify file. '
                         'Given commit `{}` is not head of a branch.').format(commit_id)
             h.flash(message, category='warning')
 
             if json_mode:
                 return message
 
             files_url = h.route_path(
                 'repo_files', repo_name=self.db_repo_name, commit_id=commit_id,
                 f_path=f_path)
             raise HTTPFound(files_url)
 
     def check_branch_permission(self, branch_name, commit_id='tip', json_mode=False):
         _ = self.request.translate
 
         rule, branch_perm = self._rhodecode_user.get_rule_and_branch_permission(
             self.db_repo_name, branch_name)
         if branch_perm and branch_perm not in ['branch.push', 'branch.push_force']:
             message = _('Branch `{}` changes forbidden by rule {}.').format(
                 h.escape(branch_name), h.escape(rule))
             h.flash(message, 'warning')
 
             if json_mode:
                 return message
 
             files_url = h.route_path(
                 'repo_files:default_path', repo_name=self.db_repo_name, commit_id=commit_id)
 
             raise HTTPFound(files_url)
 
     def _get_commit_and_path(self):
         default_commit_id = self.db_repo.landing_ref_name
         default_f_path = '/'
 
         commit_id = self.request.matchdict.get(
             'commit_id', default_commit_id)
         f_path = self._get_f_path(self.request.matchdict, default_f_path)
         return commit_id, f_path
 
     def _get_default_encoding(self, c):
         enc_list = getattr(c, 'default_encodings', [])
         return enc_list[0] if enc_list else 'UTF-8'
 
     def _get_commit_or_redirect(self, commit_id, redirect_after=True):
         """
         This is a safe way to get commit. If an error occurs it redirects to
         tip with proper message
 
         :param commit_id: id of commit to fetch
         :param redirect_after: toggle redirection
         """
         _ = self.request.translate
 
         try:
             return self.rhodecode_vcs_repo.get_commit(commit_id)
         except EmptyRepositoryError:
             if not redirect_after:
                 return None
 
             add_new = upload_new = ""
             if h.HasRepoPermissionAny(
                     'repository.write', 'repository.admin')(self.db_repo_name):
                 _url = h.route_path(
                     'repo_files_add_file',
                     repo_name=self.db_repo_name, commit_id=0, f_path='')
                 add_new = h.link_to(
                     _('add a new file'), _url, class_="alert-link")
 
                 _url_upld = h.route_path(
                     'repo_files_upload_file',
                     repo_name=self.db_repo_name, commit_id=0, f_path='')
                 upload_new = h.link_to(
                     _('upload a new file'), _url_upld, class_="alert-link")
 
             h.flash(h.literal(
                 _('There are no files yet. Click here to %s or %s.') % (add_new, upload_new)), category='warning')
             raise HTTPFound(
                 h.route_path('repo_summary', repo_name=self.db_repo_name))
 
         except (CommitDoesNotExistError, LookupError) as e:
             msg = _('No such commit exists for this repository. Commit: {}').format(commit_id)
             h.flash(msg, category='error')
             raise HTTPNotFound()
         except RepositoryError as e:
             h.flash(h.escape(safe_str(e)), category='error')
             raise HTTPNotFound()
 
     def _get_filenode_or_redirect(self, commit_obj, path, pre_load=None):
         """
         Returns file_node, if error occurs or given path is directory,
         it'll redirect to top level path
         """
         _ = self.request.translate
 
         try:
             file_node = commit_obj.get_node(path, pre_load=pre_load)
             if file_node.is_dir():
                 raise RepositoryError('The given path is a directory')
         except CommitDoesNotExistError:
             log.exception('No such commit exists for this repository')
             h.flash(_('No such commit exists for this repository'), category='error')
             raise HTTPNotFound()
         except RepositoryError as e:
             log.warning('Repository error while fetching filenode `%s`. Err:%s', path, e)
             h.flash(h.escape(safe_str(e)), category='error')
             raise HTTPNotFound()
 
         return file_node
 
     def _is_valid_head(self, commit_id, repo, landing_ref):
         branch_name = sha_commit_id = ''
         is_head = False
         log.debug('Checking if commit_id `%s` is a head for %s.', commit_id, repo)
 
         for _branch_name, branch_commit_id in repo.branches.items():
             # simple case we pass in branch name, it's a HEAD
             if commit_id == _branch_name:
                 is_head = True
                 branch_name = _branch_name
                 sha_commit_id = branch_commit_id
                 break
             # case when we pass in full sha commit_id, which is a head
             elif commit_id == branch_commit_id:
                 is_head = True
                 branch_name = _branch_name
                 sha_commit_id = branch_commit_id
                 break
 
         if h.is_svn(repo) and not repo.is_empty():
             # Note: Subversion only has one head.
             if commit_id == repo.get_commit(commit_idx=-1).raw_id:
                 is_head = True
                 return branch_name, sha_commit_id, is_head
 
         # checked branches, means we only need to try to get the branch/commit_sha
         if repo.is_empty():
             is_head = True
             branch_name = landing_ref
             sha_commit_id = EmptyCommit().raw_id
         else:
             commit = repo.get_commit(commit_id=commit_id)
             if commit:
                 branch_name = commit.branch
                 sha_commit_id = commit.raw_id
 
         return branch_name, sha_commit_id, is_head
 
     def _get_tree_at_commit(self, c, commit_id, f_path, full_load=False, at_rev=None):
 
         repo_id = self.db_repo.repo_id
         force_recache = self.get_recache_flag()
 
         cache_seconds = safe_int(
             rhodecode.CONFIG.get('rc_cache.cache_repo.expiration_time'))
         cache_on = not force_recache and cache_seconds > 0
         log.debug(
             'Computing FILE TREE for repo_id %s commit_id `%s` and path `%s`'
             'with caching: %s[TTL: %ss]' % (
                 repo_id, commit_id, f_path, cache_on, cache_seconds or 0))
 
         cache_namespace_uid = f'repo.{rc_cache.FILE_TREE_CACHE_VER}.{repo_id}'
         region = rc_cache.get_or_create_region('cache_repo', cache_namespace_uid)
 
         @region.conditional_cache_on_arguments(namespace=cache_namespace_uid, condition=cache_on)
         def compute_file_tree(_name_hash, _repo_id, _commit_id, _f_path, _full_load, _at_rev):
             log.debug('Generating cached file tree at for repo_id: %s, %s, %s',
                       _repo_id, _commit_id, _f_path)
 
             c.full_load = _full_load
             return render(
                 'rhodecode:templates/files/files_browser_tree.mako',
                 self._get_template_context(c), self.request, _at_rev)
 
         return compute_file_tree(
             self.db_repo.repo_name_hash, self.db_repo.repo_id, commit_id, f_path, full_load, at_rev)
 
     def create_pure_path(self, *parts):
         # Split paths and sanitize them, removing any ../ etc
         sanitized_path = [
             x for x in pathlib.PurePath(*parts).parts
             if x not in ['.', '..']]
 
         pure_path = pathlib.PurePath(*sanitized_path)
         return pure_path
 
     def _is_lf_enabled(self, target_repo):
         lf_enabled = False
 
         lf_key_for_vcs_map = {
             'hg': 'extensions_largefiles',
             'git': 'vcs_git_lfs_enabled'
         }
 
         lf_key_for_vcs = lf_key_for_vcs_map.get(target_repo.repo_type)
 
         if lf_key_for_vcs:
             lf_enabled = self._get_repo_setting(target_repo, lf_key_for_vcs)
 
         return lf_enabled
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_archivefile(self):
         # archive cache config
         from rhodecode import CONFIG
         _ = self.request.translate
         self.load_default_context()
         default_at_path = '/'
         fname = self.request.matchdict['fname']
         subrepos = self.request.GET.get('subrepos') == 'true'
         with_hash = str2bool(self.request.GET.get('with_hash', '1'))
         at_path = self.request.GET.get('at_path') or default_at_path
 
         if not self.db_repo.enable_downloads:
             return Response(_('Downloads disabled'))
 
         try:
             commit_id, ext, fileformat, content_type = \
                 _get_archive_spec(fname)
         except ValueError:
             return Response(_('Unknown archive type for: `{}`').format(
                 h.escape(fname)))
 
         try:
             commit = self.rhodecode_vcs_repo.get_commit(commit_id)
         except CommitDoesNotExistError:
             return Response(_('Unknown commit_id {}').format(
                 h.escape(commit_id)))
         except EmptyRepositoryError:
             return Response(_('Empty repository'))
 
         # we used a ref, or a shorter version, lets redirect client ot use explicit hash
         if commit_id != commit.raw_id:
             fname=f'{commit.raw_id}{ext}'
             raise HTTPFound(self.request.current_route_path(fname=fname))
 
         try:
             at_path = commit.get_node(at_path).path or default_at_path
         except Exception:
             return Response(_('No node at path {} for this repository').format(h.escape(at_path)))
 
         path_sha = get_path_sha(at_path)
 
         # used for cache etc, consistent unique archive name
         archive_name_key = get_archive_name(
             self.db_repo.repo_id, self.db_repo_name, commit_sha=commit.short_id, ext=ext, subrepos=subrepos,
             path_sha=path_sha, with_hash=True)
 
         if not with_hash:
             path_sha = ''
 
         # what end client gets served
         response_archive_name = get_archive_name(
             self.db_repo.repo_id, self.db_repo_name, commit_sha=commit.short_id, ext=ext, subrepos=subrepos,
             path_sha=path_sha, with_hash=with_hash)
 
         # remove extension from our archive directory name
         archive_dir_name = response_archive_name[:-len(ext)]
 
         archive_cache_disable = self.request.GET.get('no_cache')
 
         d_cache = get_archival_cache_store(config=CONFIG)
 
         # NOTE: we get the config to pass to a call to lazy-init the SAME type of cache on vcsserver
         d_cache_conf = get_archival_config(config=CONFIG)
 
         # This is also a cache key, and lock key
         reentrant_lock_key = archive_name_key + '.lock'
 
         use_cached_archive = False
         if not archive_cache_disable and archive_name_key in d_cache:
             reader, metadata = d_cache.fetch(archive_name_key)
 
             use_cached_archive = True
             log.debug('Found cached archive as key=%s tag=%s, serving archive from cache reader=%s',
                       archive_name_key, metadata, reader.name)
         else:
             reader = None
             log.debug('Archive with key=%s is not yet cached, creating one now...', archive_name_key)
 
         if not reader:
             # generate new archive, as previous was not found in the cache
             try:
                 with d_cache.get_lock(reentrant_lock_key):
                     try:
                         commit.archive_repo(archive_name_key, archive_dir_name=archive_dir_name,
                                             kind=fileformat, subrepos=subrepos,
                                             archive_at_path=at_path, cache_config=d_cache_conf)
                     except ImproperArchiveTypeError:
                         return _('Unknown archive type')
 
             except ArchiveCacheGenerationLock:
                 retry_after = round(random.uniform(0.3, 3.0), 1)
                 time.sleep(retry_after)
 
                 location = self.request.url
                 response = Response(
                     f"archive {archive_name_key} generation in progress, Retry-After={retry_after}, Location={location}"
                 )
                 response.headers["Retry-After"] = str(retry_after)
                 response.status_code = 307  # temporary redirect
 
                 response.location = location
                 return response
 
             reader, metadata = d_cache.fetch(archive_name_key, retry=True, retry_attempts=30)
 
         response = Response(app_iter=archive_iterator(reader))
         response.content_disposition = f'attachment; filename={response_archive_name}'
         response.content_type = str(content_type)
 
         try:
             return response
         finally:
             # store download action
             audit_logger.store_web(
                 'repo.archive.download', action_data={
                     'user_agent': self.request.user_agent,
                     'archive_name': archive_name_key,
                     'archive_spec': fname,
                     'archive_cached': use_cached_archive},
                 user=self._rhodecode_user,
                 repo=self.db_repo,
                 commit=True
             )
 
     def _get_file_node(self, commit_id, f_path):
         if commit_id not in ['', None, 'None', '0' * 12, '0' * 40]:
             commit = self.rhodecode_vcs_repo.get_commit(commit_id=commit_id)
             try:
                 node = commit.get_node(f_path)
                 if node.is_dir():
                     raise NodeError(f'{node} path is a {type(node)} not a file')
             except NodeDoesNotExistError:
                 commit = EmptyCommit(
                     commit_id=commit_id,
                     idx=commit.idx,
                     repo=commit.repository,
                     alias=commit.repository.alias,
                     message=commit.message,
                     author=commit.author,
                     date=commit.date)
                 node = FileNode(safe_bytes(f_path), b'', commit=commit)
         else:
             commit = EmptyCommit(
                 repo=self.rhodecode_vcs_repo,
                 alias=self.rhodecode_vcs_repo.alias)
             node = FileNode(safe_bytes(f_path), b'', commit=commit)
         return node
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_files_diff(self):
         c = self.load_default_context()
         f_path = self._get_f_path(self.request.matchdict)
         diff1 = self.request.GET.get('diff1', '')
         diff2 = self.request.GET.get('diff2', '')
 
         path1, diff1 = parse_path_ref(diff1, default_path=f_path)
 
         ignore_whitespace = str2bool(self.request.GET.get('ignorews'))
         line_context = self.request.GET.get('context', 3)
 
         if not any((diff1, diff2)):
             h.flash(
                 'Need query parameter "diff1" or "diff2" to generate a diff.',
                 category='error')
             raise HTTPBadRequest()
 
         c.action = self.request.GET.get('diff')
         if c.action not in ['download', 'raw']:
             compare_url = h.route_path(
                 'repo_compare',
                 repo_name=self.db_repo_name,
                 source_ref_type='rev',
                 source_ref=diff1,
                 target_repo=self.db_repo_name,
                 target_ref_type='rev',
                 target_ref=diff2,
                 _query=dict(f_path=f_path))
             # redirect to new view if we render diff
             raise HTTPFound(compare_url)
 
         try:
             node1 = self._get_file_node(diff1, path1)
             node2 = self._get_file_node(diff2, f_path)
         except (RepositoryError, NodeError):
             log.exception("Exception while trying to get node from repository")
             raise HTTPFound(
                 h.route_path('repo_files', repo_name=self.db_repo_name,
                              commit_id='tip', f_path=f_path))
 
         if all(isinstance(node.commit, EmptyCommit)
                for node in (node1, node2)):
             raise HTTPNotFound()
 
         c.commit_1 = node1.commit
         c.commit_2 = node2.commit
 
         if c.action == 'download':
             _diff = diffs.get_gitdiff(node1, node2,
                                       ignore_whitespace=ignore_whitespace,
                                       context=line_context)
             # NOTE: this was using diff_format='gitdiff'
             diff = diffs.DiffProcessor(_diff, diff_format='newdiff')
 
             response = Response(self.path_filter.get_raw_patch(diff))
             response.content_type = 'text/plain'
             response.content_disposition = (
                 f'attachment; filename={f_path}_{diff1}_vs_{diff2}.diff'
             )
             charset = self._get_default_encoding(c)
             if charset:
                 response.charset = charset
             return response
 
         elif c.action == 'raw':
             _diff = diffs.get_gitdiff(node1, node2,
                                       ignore_whitespace=ignore_whitespace,
                                       context=line_context)
             # NOTE: this was using diff_format='gitdiff'
             diff = diffs.DiffProcessor(_diff, diff_format='newdiff')
 
             response = Response(self.path_filter.get_raw_patch(diff))
             response.content_type = 'text/plain'
             charset = self._get_default_encoding(c)
             if charset:
                 response.charset = charset
             return response
 
         # in case we ever end up here
         raise HTTPNotFound()
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_files_diff_2way_redirect(self):
         """
         Kept only to make OLD links work
         """
         f_path = self._get_f_path_unchecked(self.request.matchdict)
         diff1 = self.request.GET.get('diff1', '')
         diff2 = self.request.GET.get('diff2', '')
 
         if not any((diff1, diff2)):
             h.flash(
                 'Need query parameter "diff1" or "diff2" to generate a diff.',
                 category='error')
             raise HTTPBadRequest()
 
         compare_url = h.route_path(
             'repo_compare',
             repo_name=self.db_repo_name,
             source_ref_type='rev',
             source_ref=diff1,
             target_ref_type='rev',
             target_ref=diff2,
             _query=dict(f_path=f_path, diffmode='sideside',
                         target_repo=self.db_repo_name,))
         raise HTTPFound(compare_url)
 
     @LoginRequired()
     def repo_files_default_commit_redirect(self):
         """
         Special page that redirects to the landing page of files based on the default
         commit for repository
         """
         c = self.load_default_context()
         ref_name = c.rhodecode_db_repo.landing_ref_name
         landing_url = h.repo_files_by_ref_url(
             c.rhodecode_db_repo.repo_name,
             c.rhodecode_db_repo.repo_type,
             f_path='',
             ref_name=ref_name,
             commit_id='tip',
             query=dict(at=ref_name)
         )
 
         raise HTTPFound(landing_url)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_files(self):
         c = self.load_default_context()
 
         view_name = getattr(self.request.matched_route, 'name', None)
 
         c.annotate = view_name == 'repo_files:annotated'
         # default is false, but .rst/.md files later are auto rendered, we can
         # overwrite auto rendering by setting this GET flag
         c.renderer = view_name == 'repo_files:rendered' or not self.request.GET.get('no-render', False)
 
         commit_id, f_path = self._get_commit_and_path()
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.branch = self.request.GET.get('branch', None)
         c.f_path = f_path
         at_rev = self.request.GET.get('at')
 
         # files or dirs
         try:
             c.file = c.commit.get_node(f_path, pre_load=['is_binary', 'size', 'data'])
 
             c.file_author = True
             c.file_tree = ''
 
             # prev link
             try:
                 prev_commit = c.commit.prev(c.branch)
                 c.prev_commit = prev_commit
                 c.url_prev = h.route_path(
                     'repo_files', repo_name=self.db_repo_name,
                     commit_id=prev_commit.raw_id, f_path=f_path)
                 if c.branch:
                     c.url_prev += '?branch=%s' % c.branch
             except (CommitDoesNotExistError, VCSError):
                 c.url_prev = '#'
                 c.prev_commit = EmptyCommit()
 
             # next link
             try:
                 next_commit = c.commit.next(c.branch)
                 c.next_commit = next_commit
                 c.url_next = h.route_path(
                     'repo_files', repo_name=self.db_repo_name,
                     commit_id=next_commit.raw_id, f_path=f_path)
                 if c.branch:
                     c.url_next += '?branch=%s' % c.branch
             except (CommitDoesNotExistError, VCSError):
                 c.url_next = '#'
                 c.next_commit = EmptyCommit()
 
             # load file content
             if c.file.is_file():
 
                 c.lf_node = {}
 
                 has_lf_enabled = self._is_lf_enabled(self.db_repo)
                 if has_lf_enabled:
                     c.lf_node = c.file.get_largefile_node()
 
                 c.file_source_page = 'true'
                 c.file_last_commit = c.file.last_commit
 
                 c.file_size_too_big = c.file.size > c.visual.cut_off_limit_file
 
                 if not (c.file_size_too_big or c.file.is_binary):
                     if c.annotate:  # annotation has precedence over renderer
                         c.annotated_lines = filenode_as_annotated_lines_tokens(
                             c.file
                         )
                     else:
                         c.renderer = (
                             c.renderer and h.renderer_from_filename(c.file.path)
                         )
                         if not c.renderer:
                             c.lines = filenode_as_lines_tokens(c.file)
 
                 _branch_name, _sha_commit_id, is_head = \
                     self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                         landing_ref=self.db_repo.landing_ref_name)
                 c.on_branch_head = is_head
 
                 branch = c.commit.branch if (
                     c.commit.branch and '/' not in c.commit.branch) else None
                 c.branch_or_raw_id = branch or c.commit.raw_id
                 c.branch_name = c.commit.branch or h.short_id(c.commit.raw_id)
 
                 author = c.file_last_commit.author
                 c.authors = [[
                     h.email(author),
                     h.person(author, 'username_or_name_or_email'),
                     1
                 ]]
 
             else:  # load tree content at path
                 c.file_source_page = 'false'
                 c.authors = []
                 # this loads a simple tree without metadata to speed things up
                 # later via ajax we call repo_nodetree_full and fetch whole
                 c.file_tree = self._get_tree_at_commit(c, c.commit.raw_id, f_path, at_rev=at_rev)
 
                 c.readme_data, c.readme_file = \
                     self._get_readme_data(self.db_repo, c.visual.default_renderer,
                                           c.commit.raw_id, f_path)
 
         except RepositoryError as e:
             h.flash(h.escape(safe_str(e)), category='error')
             raise HTTPNotFound()
 
         if self.request.environ.get('HTTP_X_PJAX'):
             html = render('rhodecode:templates/files/files_pjax.mako',
                           self._get_template_context(c), self.request)
         else:
             html = render('rhodecode:templates/files/files.mako',
                           self._get_template_context(c), self.request)
         return Response(html)
 
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_files_annotated_previous(self):
         self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         prev_commit_id = commit.raw_id
         line_anchor = self.request.GET.get('line_anchor')
         is_file = False
         try:
             _file = commit.get_node(f_path)
             is_file = _file.is_file()
         except (NodeDoesNotExistError, CommitDoesNotExistError, VCSError):
             pass
 
         if is_file:
             history = commit.get_path_history(f_path)
             prev_commit_id = history[1].raw_id \
                 if len(history) > 1 else prev_commit_id
         prev_url = h.route_path(
             'repo_files:annotated', repo_name=self.db_repo_name,
             commit_id=prev_commit_id, f_path=f_path,
             _anchor=f'L{line_anchor}')
 
         raise HTTPFound(prev_url)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_nodetree_full(self):
         """
         Returns rendered html of file tree that contains commit date,
         author, commit_id for the specified combination of
         repo, commit_id and file path
         """
         c = self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         try:
             dir_node = commit.get_node(f_path)
         except RepositoryError as e:
             return Response(f'error: {h.escape(safe_str(e))}')
 
         if dir_node.is_file():
             return Response('')
 
         c.file = dir_node
         c.commit = commit
         at_rev = self.request.GET.get('at')
 
         html = self._get_tree_at_commit(
             c, commit.raw_id, dir_node.path, full_load=True, at_rev=at_rev)
 
         return Response(html)
 
     def _get_attachement_headers(self, f_path):
         f_name = safe_str(f_path.split(Repository.NAME_SEP)[-1])
         safe_path = f_name.replace('"', '\\"')
         encoded_path = urllib.parse.quote(f_name)
 
         headers = "attachment; " \
                   "filename=\"{}\"; " \
                   "filename*=UTF-8\'\'{}".format(safe_path, encoded_path)
 
         return safe_bytes(headers).decode('latin-1', errors='replace')
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_file_raw(self):
         """
         Action for show as raw, some mimetypes are "rendered",
         those include images, icons.
         """
         c = self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         file_node = self._get_filenode_or_redirect(commit, f_path)
 
         raw_mimetype_mapping = {
             # map original mimetype to a mimetype used for "show as raw"
             # you can also provide a content-disposition to override the
             # default "attachment" disposition.
             # orig_type: (new_type, new_dispo)
 
             # show images inline:
             # Do not re-add SVG: it is unsafe and permits XSS attacks. One can
             # for example render an SVG with javascript inside or even render
             # HTML.
             'image/x-icon': ('image/x-icon', 'inline'),
             'image/png': ('image/png', 'inline'),
             'image/gif': ('image/gif', 'inline'),
             'image/jpeg': ('image/jpeg', 'inline'),
             'application/pdf': ('application/pdf', 'inline'),
         }
 
         mimetype = file_node.mimetype
         try:
             mimetype, disposition = raw_mimetype_mapping[mimetype]
         except KeyError:
             # we don't know anything special about this, handle it safely
             if file_node.is_binary:
                 # do same as download raw for binary files
                 mimetype, disposition = 'application/octet-stream', 'attachment'
             else:
                 # do not just use the original mimetype, but force text/plain,
                 # otherwise it would serve text/html and that might be unsafe.
                 # Note: underlying vcs library fakes text/plain mimetype if the
                 # mimetype can not be determined and it thinks it is not
                 # binary.This might lead to erroneous text display in some
                 # cases, but helps in other cases, like with text files
                 # without extension.
                 mimetype, disposition = 'text/plain', 'inline'
 
         if disposition == 'attachment':
             disposition = self._get_attachement_headers(f_path)
 
         stream_content = file_node.stream_bytes()
 
         response = Response(app_iter=stream_content)
         response.content_disposition = disposition
         response.content_type = mimetype
 
         charset = self._get_default_encoding(c)
         if charset:
             response.charset = charset
 
         return response
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_file_download(self):
         c = self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         file_node = self._get_filenode_or_redirect(commit, f_path)
 
         if self.request.GET.get('lf'):
             # only if lf get flag is passed, we download this file
             # as LFS/Largefile
             lf_node = file_node.get_largefile_node()
             if lf_node:
                 # overwrite our pointer with the REAL large-file
                 file_node = lf_node
 
         disposition = self._get_attachement_headers(f_path)
 
         stream_content = file_node.stream_bytes()
 
         response = Response(app_iter=stream_content)
         response.content_disposition = disposition
         response.content_type = file_node.mimetype
 
         charset = self._get_default_encoding(c)
         if charset:
             response.charset = charset
 
         return response
 
     def _get_nodelist_at_commit(self, repo_name, repo_id, commit_id, f_path):
 
         cache_seconds = safe_int(
             rhodecode.CONFIG.get('rc_cache.cache_repo.expiration_time'))
         cache_on = cache_seconds > 0
         log.debug(
             'Computing FILE SEARCH for repo_id %s commit_id `%s` and path `%s`'
             'with caching: %s[TTL: %ss]' % (
                 repo_id, commit_id, f_path, cache_on, cache_seconds or 0))
 
         cache_namespace_uid = f'repo.{repo_id}'
         region = rc_cache.get_or_create_region('cache_repo', cache_namespace_uid)
 
         @region.conditional_cache_on_arguments(namespace=cache_namespace_uid, condition=cache_on)
         def compute_file_search(_name_hash, _repo_id, _commit_id, _f_path):
             log.debug('Generating cached nodelist for repo_id:%s, %s, %s',
                       _repo_id, commit_id, f_path)
             try:
                 _d, _f = ScmModel().get_quick_filter_nodes(repo_name, _commit_id, _f_path)
             except (RepositoryError, CommitDoesNotExistError, Exception) as e:
                 log.exception(safe_str(e))
                 h.flash(h.escape(safe_str(e)), category='error')
                 raise HTTPFound(h.route_path(
                     'repo_files', repo_name=self.db_repo_name,
                     commit_id='tip', f_path='/'))
 
             return _d + _f
 
         result = compute_file_search(self.db_repo.repo_name_hash, self.db_repo.repo_id,
                                      commit_id, f_path)
         return filter(lambda n: self.path_filter.path_access_allowed(n['name']), result)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_nodelist(self):
         self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
 
         metadata = self._get_nodelist_at_commit(
             self.db_repo_name, self.db_repo.repo_id, commit.raw_id, f_path)
         return {'nodes': [x for x in metadata]}
 
     def _create_references(self, branches_or_tags, symbolic_reference, f_path, ref_type):
         items = []
         for name, commit_id in branches_or_tags.items():
             sym_ref = symbolic_reference(commit_id, name, f_path, ref_type)
             items.append((sym_ref, name, ref_type))
         return items
 
     def _symbolic_reference(self, commit_id, name, f_path, ref_type):
         return commit_id
 
     def _symbolic_reference_svn(self, commit_id, name, f_path, ref_type):
         return commit_id
 
         # NOTE(dan): old code we used in "diff" mode compare
         new_f_path = vcspath.join(name, f_path)
         return f'{new_f_path}@{commit_id}'
 
     def _get_node_history(self, commit_obj, f_path, commits=None):
         """
         get commit history for given node
 
         :param commit_obj: commit to calculate history
         :param f_path: path for node to calculate history for
         :param commits: if passed don't calculate history and take
             commits defined in this list
         """
         _ = self.request.translate
 
         # calculate history based on tip
         tip = self.rhodecode_vcs_repo.get_commit()
         if commits is None:
             pre_load = ["author", "branch"]
             try:
                 commits = tip.get_path_history(f_path, pre_load=pre_load)
             except (NodeDoesNotExistError, CommitError):
                 # this node is not present at tip!
                 commits = commit_obj.get_path_history(f_path, pre_load=pre_load)
 
         history = []
         commits_group = ([], _("Changesets"))
         for commit in commits:
             branch = ' (%s)' % commit.branch if commit.branch else ''
             n_desc = f'r{commit.idx}:{commit.short_id}{branch}'
             commits_group[0].append((commit.raw_id, n_desc, 'sha'))
         history.append(commits_group)
 
         symbolic_reference = self._symbolic_reference
 
         if self.rhodecode_vcs_repo.alias == 'svn':
             adjusted_f_path = RepoFilesView.adjust_file_path_for_svn(
                 f_path, self.rhodecode_vcs_repo)
             if adjusted_f_path != f_path:
                 log.debug(
                     'Recognized svn tag or branch in file "%s", using svn '
                     'specific symbolic references', f_path)
                 f_path = adjusted_f_path
                 symbolic_reference = self._symbolic_reference_svn
 
         branches = self._create_references(
             self.rhodecode_vcs_repo.branches, symbolic_reference, f_path, 'branch')
         branches_group = (branches, _("Branches"))
 
         tags = self._create_references(
             self.rhodecode_vcs_repo.tags, symbolic_reference, f_path, 'tag')
         tags_group = (tags, _("Tags"))
 
         history.append(branches_group)
         history.append(tags_group)
 
         return history, commits
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_file_history(self):
         self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         file_node = self._get_filenode_or_redirect(commit, f_path)
 
         if file_node.is_file():
             file_history, _hist = self._get_node_history(commit, f_path)
 
             res = []
             for section_items, section in file_history:
                 items = []
                 for obj_id, obj_text, obj_type in section_items:
                     at_rev = ''
                     if obj_type in ['branch', 'bookmark', 'tag']:
                         at_rev = obj_text
                     entry = {
                         'id': obj_id,
                         'text': obj_text,
                         'type': obj_type,
                         'at_rev': at_rev
                     }
 
                     items.append(entry)
 
                 res.append({
                     'text': section,
                     'children': items
                 })
 
             data = {
                 'more': False,
                 'results': res
             }
             return data
 
         log.warning('Cannot fetch history for directory')
         raise HTTPBadRequest()
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_file_authors(self):
         c = self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         file_node = self._get_filenode_or_redirect(commit, f_path)
 
         if not file_node.is_file():
             raise HTTPBadRequest()
 
         c.file_last_commit = file_node.last_commit
         if self.request.GET.get('annotate') == '1':
             # use _hist from annotation if annotation mode is on
             commit_ids = {x[1] for x in file_node.annotate}
             _hist = (
                 self.rhodecode_vcs_repo.get_commit(commit_id)
                 for commit_id in commit_ids)
         else:
             _f_history, _hist = self._get_node_history(commit, f_path)
         c.file_author = False
 
         unique = collections.OrderedDict()
         for commit in _hist:
             author = commit.author
             if author not in unique:
                 unique[commit.author] = [
                     h.email(author),
                     h.person(author, 'username_or_name_or_email'),
                     1  # counter
                 ]
 
             else:
                 # increase counter
                 unique[commit.author][2] += 1
 
         c.authors = [val for val in unique.values()]
 
         return self._get_template_context(c)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     def repo_files_check_head(self):
         self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         new_path = self.request.POST.get('path')
         operation = self.request.POST.get('operation')
         path_exist = ''
 
         if new_path and operation in ['create', 'upload']:
             new_f_path = os.path.join(f_path.lstrip('/'), new_path)
             try:
                 commit_obj = self.rhodecode_vcs_repo.get_commit(commit_id)
                 # NOTE(dan): construct whole path without leading /
                 file_node = commit_obj.get_node(new_f_path)
                 if file_node is not None:
                     path_exist = new_f_path
             except EmptyRepositoryError:
                 pass
             except Exception:
                 pass
 
         return {
             'branch': _branch_name,
             'sha': _sha_commit_id,
             'is_head': is_head,
             'path_exists': path_exist
         }
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     def repo_files_remove_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path)
         self.check_branch_permission(_branch_name)
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.file = self._get_filenode_or_redirect(c.commit, f_path)
 
         c.default_message = _(
             'Deleted file {} via RhodeCode Enterprise').format(f_path)
         c.f_path = f_path
 
         return self._get_template_context(c)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_delete_file(self):
         _ = self.request.translate
 
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path)
         self.check_branch_permission(_branch_name)
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.file = self._get_filenode_or_redirect(c.commit, f_path)
 
         c.default_message = _(
             'Deleted file {} via RhodeCode Enterprise').format(f_path)
         c.f_path = f_path
         node_path = f_path
         author = self._rhodecode_db_user.full_contact
         message = self.request.POST.get('message') or c.default_message
         try:
             nodes = {
                 safe_bytes(node_path): {
                     'content': b''
                 }
             }
             ScmModel().delete_nodes(
                 user=self._rhodecode_db_user.user_id, repo=self.db_repo,
                 message=message,
                 nodes=nodes,
                 parent_commit=c.commit,
                 author=author,
             )
 
             h.flash(
                 _('Successfully deleted file `{}`').format(
                     h.escape(f_path)), category='success')
         except Exception:
             log.exception('Error during commit operation')
             h.flash(_('Error occurred during commit'), category='error')
         raise HTTPFound(
             h.route_path('repo_commit', repo_name=self.db_repo_name,
                          commit_id='tip'))
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     def repo_files_edit_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path, commit_id=commit_id)
         self.check_branch_permission(_branch_name, commit_id=commit_id)
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.file = self._get_filenode_or_redirect(c.commit, f_path)
 
         if c.file.is_binary:
             files_url = h.route_path(
                 'repo_files',
                 repo_name=self.db_repo_name,
                 commit_id=c.commit.raw_id, f_path=f_path)
             raise HTTPFound(files_url)
 
         c.default_message = _('Edited file {} via RhodeCode Enterprise').format(f_path)
         c.f_path = f_path
 
         return self._get_template_context(c)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_update_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.file = self._get_filenode_or_redirect(c.commit, f_path)
 
         if c.file.is_binary:
             raise HTTPFound(h.route_path('repo_files', repo_name=self.db_repo_name,
                                          commit_id=c.commit.raw_id, f_path=f_path))
 
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path, commit_id=commit_id)
         self.check_branch_permission(_branch_name, commit_id=commit_id)
 
         c.default_message = _('Edited file {} via RhodeCode Enterprise').format(f_path)
         c.f_path = f_path
 
         old_content = c.file.str_content
         sl = old_content.splitlines(1)
         first_line = sl[0] if sl else ''
 
         r_post = self.request.POST
         # line endings:  0 - Unix, 1 - Mac, 2 - DOS
         line_ending_mode = detect_mode(first_line, 0)
         content = convert_line_endings(r_post.get('content', ''), line_ending_mode)
 
         message = r_post.get('message') or c.default_message
 
         org_node_path = c.file.str_path
         filename = r_post['filename']
 
         root_path = c.file.dir_path
         pure_path = self.create_pure_path(root_path, filename)
         node_path = pure_path.as_posix()
 
         default_redirect_url = h.route_path('repo_commit', repo_name=self.db_repo_name,
                                             commit_id=commit_id)
         if content == old_content and node_path == org_node_path:
             h.flash(_('No changes detected on {}').format(h.escape(org_node_path)),
                     category='warning')
             raise HTTPFound(default_redirect_url)
 
         try:
             mapping = {
                 c.file.bytes_path: {
                     'org_filename': org_node_path,
                     'filename': safe_bytes(node_path),
                     'content': safe_bytes(content),
                     'lexer': '',
                     'op': 'mod',
                     'mode': c.file.mode
                 }
             }
 
             commit = ScmModel().update_nodes(
                 user=self._rhodecode_db_user.user_id,
                 repo=self.db_repo,
                 message=message,
                 nodes=mapping,
                 parent_commit=c.commit,
             )
 
             h.flash(_('Successfully committed changes to file `{}`').format(
                     h.escape(f_path)), category='success')
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id=commit.raw_id)
 
         except Exception:
             log.exception('Error occurred during commit')
             h.flash(_('Error occurred during commit'), category='error')
 
         raise HTTPFound(default_redirect_url)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     def repo_files_add_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         # Check if we need to use this page to upload binary
         upload_binary = str2bool(self.request.params.get('upload_binary', False))
 
         c.commit = self._get_commit_or_redirect(commit_id, redirect_after=False)
         if c.commit is None:
             c.commit = EmptyCommit(alias=self.rhodecode_vcs_repo.alias)
 
         if self.rhodecode_vcs_repo.is_empty():
             # for empty repository we cannot check for current branch, we rely on
             # c.commit.branch instead
             _branch_name, _sha_commit_id, is_head = c.commit.branch, '', True
         else:
             _branch_name, _sha_commit_id, is_head = \
                 self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                     landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path, commit_id=commit_id)
         self.check_branch_permission(_branch_name, commit_id=commit_id)
 
         c.default_message = (_('Added file via RhodeCode Enterprise')) \
             if not upload_binary else (_('Edited file {} via RhodeCode Enterprise').format(f_path))
         c.f_path = f_path.lstrip('/')  # ensure not relative path
         c.replace_binary = upload_binary
 
         return self._get_template_context(c)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_create_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         c.commit = self._get_commit_or_redirect(commit_id, redirect_after=False)
         if c.commit is None:
             c.commit = EmptyCommit(alias=self.rhodecode_vcs_repo.alias)
 
         # calculate redirect URL
         if self.rhodecode_vcs_repo.is_empty():
             default_redirect_url = h.route_path(
                 'repo_summary', repo_name=self.db_repo_name)
         else:
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id='tip')
 
         if self.rhodecode_vcs_repo.is_empty():
             # for empty repository we cannot check for current branch, we rely on
             # c.commit.branch instead
             _branch_name, _sha_commit_id, is_head = c.commit.branch, '', True
         else:
             _branch_name, _sha_commit_id, is_head = \
                 self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                     landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path, commit_id=commit_id)
         self.check_branch_permission(_branch_name, commit_id=commit_id)
 
         c.default_message = (_('Added file via RhodeCode Enterprise'))
         c.f_path = f_path
 
         r_post = self.request.POST
         message = r_post.get('message') or c.default_message
         filename = r_post.get('filename')
         unix_mode = 0
 
         if not filename:
             # If there's no commit, redirect to repo summary
             if type(c.commit) is EmptyCommit:
                 redirect_url = h.route_path(
                     'repo_summary', repo_name=self.db_repo_name)
             else:
                 redirect_url = default_redirect_url
             h.flash(_('No filename specified'), category='warning')
             raise HTTPFound(redirect_url)
 
         root_path = f_path
         pure_path = self.create_pure_path(root_path, filename)
         node_path = pure_path.as_posix().lstrip('/')
 
         author = self._rhodecode_db_user.full_contact
         content = convert_line_endings(r_post.get('content', ''), unix_mode)
         nodes = {
             safe_bytes(node_path): {
                 'content': safe_bytes(content)
             }
         }
 
         try:
 
             commit = ScmModel().create_nodes(
                 user=self._rhodecode_db_user.user_id,
                 repo=self.db_repo,
                 message=message,
                 nodes=nodes,
                 parent_commit=c.commit,
                 author=author,
             )
 
             h.flash(_('Successfully committed new file `{}`').format(
                     h.escape(node_path)), category='success')
 
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id=commit.raw_id)
 
         except NonRelativePathError:
             log.exception('Non Relative path found')
             h.flash(_('The location specified must be a relative path and must not '
                       'contain .. in the path'), category='warning')
             raise HTTPFound(default_redirect_url)
         except (NodeError, NodeAlreadyExistsError) as e:
             h.flash(h.escape(safe_str(e)), category='error')
         except Exception:
             log.exception('Error occurred during commit')
             h.flash(_('Error occurred during commit'), category='error')
 
         raise HTTPFound(default_redirect_url)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_upload_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         c.commit = self._get_commit_or_redirect(commit_id, redirect_after=False)
         if c.commit is None:
             c.commit = EmptyCommit(alias=self.rhodecode_vcs_repo.alias)
 
         # calculate redirect URL
         if self.rhodecode_vcs_repo.is_empty():
             default_redirect_url = h.route_path(
                 'repo_summary', repo_name=self.db_repo_name)
         else:
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id='tip')
 
         if self.rhodecode_vcs_repo.is_empty():
             # for empty repository we cannot check for current branch, we rely on
             # c.commit.branch instead
             _branch_name, _sha_commit_id, is_head = c.commit.branch, '', True
         else:
             _branch_name, _sha_commit_id, is_head = \
                 self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                     landing_ref=self.db_repo.landing_ref_name)
 
         error = self.forbid_non_head(is_head, f_path, json_mode=True)
         if error:
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         error = self.check_branch_permission(_branch_name, json_mode=True)
         if error:
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         c.default_message = (_('Added file via RhodeCode Enterprise'))
         c.f_path = f_path
 
         r_post = self.request.POST
 
         message = c.default_message
         user_message = r_post.getall('message')
         if isinstance(user_message, list) and user_message:
             # we take the first from duplicated results if it's not empty
             message = user_message[0] if user_message[0] else message
 
         nodes = {}
 
         for file_obj in r_post.getall('files_upload') or []:
             content = file_obj.file
             filename = file_obj.filename
 
             root_path = f_path
             pure_path = self.create_pure_path(root_path, filename)
             node_path = pure_path.as_posix().lstrip('/')
 
             nodes[safe_bytes(node_path)] = {
                 'content': content
             }
 
         if not nodes:
             error = 'missing files'
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         author = self._rhodecode_db_user.full_contact
 
         try:
             commit = ScmModel().create_nodes(
                 user=self._rhodecode_db_user.user_id,
                 repo=self.db_repo,
                 message=message,
                 nodes=nodes,
                 parent_commit=c.commit,
                 author=author,
             )
             if len(nodes) == 1:
                 flash_message = _('Successfully committed {} new files').format(len(nodes))
             else:
                 flash_message = _('Successfully committed 1 new file')
 
             h.flash(flash_message, category='success')
 
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id=commit.raw_id)
 
         except NonRelativePathError:
             log.exception('Non Relative path found')
             error = _('The location specified must be a relative path and must not '
                       'contain .. in the path')
             h.flash(error, category='warning')
 
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         except (NodeError, NodeAlreadyExistsError) as e:
             error = h.escape(e)
             h.flash(error, category='error')
 
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         except Exception:
             log.exception('Error occurred during commit')
             error = _('Error occurred during commit')
             h.flash(error, category='error')
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         return {
             'error': None,
             'redirect_url': default_redirect_url
         }
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_replace_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         c.commit = self._get_commit_or_redirect(commit_id, redirect_after=False)
         if c.commit is None:
             c.commit = EmptyCommit(alias=self.rhodecode_vcs_repo.alias)
 
         if self.rhodecode_vcs_repo.is_empty():
             default_redirect_url = h.route_path(
                 'repo_summary', repo_name=self.db_repo_name)
         else:
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id='tip')
 
         if self.rhodecode_vcs_repo.is_empty():
             # for empty repository we cannot check for current branch, we rely on
             # c.commit.branch instead
             _branch_name, _sha_commit_id, is_head = c.commit.branch, '', True
         else:
             _branch_name, _sha_commit_id, is_head = \
                 self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                     landing_ref=self.db_repo.landing_ref_name)
 
         error = self.forbid_non_head(is_head, f_path, json_mode=True)
         if error:
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         error = self.check_branch_permission(_branch_name, json_mode=True)
         if error:
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         c.default_message = (_('Edited file {} via RhodeCode Enterprise').format(f_path))
         c.f_path = f_path
 
         r_post = self.request.POST
 
         message = c.default_message
         user_message = r_post.getall('message')
         if isinstance(user_message, list) and user_message:
             # we take the first from duplicated results if it's not empty
             message = user_message[0] if user_message[0] else message
 
         data_for_replacement = r_post.getall('files_upload') or []
         if (objects_count := len(data_for_replacement)) > 1:
             return {
                 'error': 'too many files for replacement',
                 'redirect_url': default_redirect_url
             }
         elif not objects_count:
             return {
                 'error': 'missing files',
                 'redirect_url': default_redirect_url
             }
 
         content = data_for_replacement[0].file
         retrieved_filename = data_for_replacement[0].filename
 
         if retrieved_filename.split('.')[-1] != f_path.split('.')[-1]:
             return {
                 'error': 'file extension of uploaded file doesn\'t match an original file\'s extension',
                 'redirect_url': default_redirect_url
             }
 
         author = self._rhodecode_db_user.full_contact
 
         try:
             commit = ScmModel().update_binary_node(
                 user=self._rhodecode_db_user.user_id,
                 repo=self.db_repo,
                 message=message,
                 node={
                     'content': content,
                     'file_path': f_path.encode(),
                 },
                 parent_commit=c.commit,
                 author=author,
             )
 
             h.flash(_('Successfully committed 1 new file'), category='success')
 
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id=commit.raw_id)
 
         except (NodeError, NodeAlreadyExistsError) as e:
             error = h.escape(e)
             h.flash(error, category='error')
 
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         except Exception:
             log.exception('Error occurred during commit')
             error = _('Error occurred during commit')
             h.flash(error, category='error')
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         return {
             'error': None,
             'redirect_url': default_redirect_url
         }

rhodecode/config/config_maker.py

@@ -1,205 +1,221 b''
 # Copyright (C) 2010-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import os
 import tempfile
 import logging
 
 from pyramid.settings import asbool
 
 from rhodecode.config.settings_maker import SettingsMaker
 from rhodecode.config import utils as config_utils
 
 log = logging.getLogger(__name__)
 
 
 def sanitize_settings_and_apply_defaults(global_config, settings):
     """
     Applies settings defaults and does all type conversion.
 
     We would move all settings parsing and preparation into this place, so that
     we have only one place left which deals with this part. The remaining parts
     of the application would start to rely fully on well-prepared settings.
 
     This piece would later be split up per topic to avoid a big fat monster
     function.
     """
     jn = os.path.join
 
     global_settings_maker = SettingsMaker(global_config)
     global_settings_maker.make_setting('debug', default=False, parser='bool')
     debug_enabled = asbool(global_config.get('debug'))
 
     settings_maker = SettingsMaker(settings)
 
     settings_maker.make_setting(
         'logging.autoconfigure',
         default=False,
         parser='bool')
 
     logging_conf = jn(os.path.dirname(global_config.get('__file__')), 'logging.ini')
     settings_maker.enable_logging(logging_conf, level='INFO' if debug_enabled else 'DEBUG')
 
     # Default includes, possible to change as a user
     pyramid_includes = settings_maker.make_setting('pyramid.includes', [], parser='list:newline')
     log.debug(
         "Using the following pyramid.includes: %s",
         pyramid_includes)
 
     settings_maker.make_setting('rhodecode.edition', 'Community Edition')
     settings_maker.make_setting('rhodecode.edition_id', 'CE')
 
     if 'mako.default_filters' not in settings:
         # set custom default filters if we don't have it defined
         settings['mako.imports'] = 'from rhodecode.lib.base import h_filter'
         settings['mako.default_filters'] = 'h_filter'
 
     if 'mako.directories' not in settings:
         mako_directories = settings.setdefault('mako.directories', [
             # Base templates of the original application
             'rhodecode:templates',
         ])
         log.debug(
             "Using the following Mako template directories: %s",
             mako_directories)
 
     # NOTE(marcink): fix redis requirement for schema of connection since 3.X
     if 'beaker.session.type' in settings and settings['beaker.session.type'] == 'ext:redis':
         raw_url = settings['beaker.session.url']
         if not raw_url.startswith(('redis://', 'rediss://', 'unix://')):
             settings['beaker.session.url'] = 'redis://' + raw_url
 
     settings_maker.make_setting('__file__', global_config.get('__file__'))
 
     # TODO: johbo: Re-think this, usually the call to config.include
     # should allow to pass in a prefix.
     settings_maker.make_setting('rhodecode.api.url', '/_admin/api')
 
     # Sanitize generic settings.
     settings_maker.make_setting('default_encoding', 'UTF-8', parser='list')
     settings_maker.make_setting('gzip_responses', False, parser='bool')
     settings_maker.make_setting('startup.import_repos', 'false', parser='bool')
 
     # statsd
     settings_maker.make_setting('statsd.enabled', False, parser='bool')
     settings_maker.make_setting('statsd.statsd_host', 'statsd-exporter', parser='string')
     settings_maker.make_setting('statsd.statsd_port', 9125, parser='int')
     settings_maker.make_setting('statsd.statsd_prefix', '')
     settings_maker.make_setting('statsd.statsd_ipv6', False, parser='bool')
 
     settings_maker.make_setting('vcs.svn.compatible_version', '')
     settings_maker.make_setting('vcs.svn.proxy.enabled', True, parser='bool')
     settings_maker.make_setting('vcs.svn.proxy.host', 'http://svn:8090', parser='string')
     settings_maker.make_setting('vcs.hooks.protocol', 'http')
     settings_maker.make_setting('vcs.hooks.host', '*')
     settings_maker.make_setting('vcs.scm_app_implementation', 'http')
     settings_maker.make_setting('vcs.server', '')
     settings_maker.make_setting('vcs.server.protocol', 'http')
     settings_maker.make_setting('vcs.server.enable', 'true', parser='bool')
     settings_maker.make_setting('vcs.hooks.direct_calls', 'false', parser='bool')
     settings_maker.make_setting('vcs.start_server', 'false', parser='bool')
     settings_maker.make_setting('vcs.backends', 'hg, git, svn', parser='list')
     settings_maker.make_setting('vcs.connection_timeout', 3600, parser='int')
 
     settings_maker.make_setting('vcs.methods.cache', True, parser='bool')
 
     # repo_store path
     settings_maker.make_setting('repo_store.path', '/var/opt/rhodecode_repo_store')
     # Support legacy values of vcs.scm_app_implementation. Legacy
     # configurations may use 'rhodecode.lib.middleware.utils.scm_app_http', or
     # disabled since 4.13 'vcsserver.scm_app' which is now mapped to 'http'.
     scm_app_impl = settings['vcs.scm_app_implementation']
     if scm_app_impl in ['rhodecode.lib.middleware.utils.scm_app_http', 'vcsserver.scm_app']:
         settings['vcs.scm_app_implementation'] = 'http'
 
     settings_maker.make_setting('appenlight', False, parser='bool')
 
     temp_store = tempfile.gettempdir()
     tmp_cache_dir = jn(temp_store, 'rc_cache')
 
     # save default, cache dir, and use it for all backends later.
     default_cache_dir = settings_maker.make_setting(
         'cache_dir',
         default=tmp_cache_dir, default_when_empty=True,
         parser='dir:ensured')
 
     # exception store cache
     settings_maker.make_setting(
         'exception_tracker.store_path',
         default=jn(default_cache_dir, 'exc_store'), default_when_empty=True,
         parser='dir:ensured'
     )
 
     settings_maker.make_setting(
         'celerybeat-schedule.path',
         default=jn(default_cache_dir, 'celerybeat_schedule', 'celerybeat-schedule.db'), default_when_empty=True,
         parser='file:ensured'
     )
 
     settings_maker.make_setting('exception_tracker.send_email', False, parser='bool')
     settings_maker.make_setting('exception_tracker.email_prefix', '[RHODECODE ERROR]', default_when_empty=True)
 
     # sessions, ensure file since no-value is memory
     settings_maker.make_setting('beaker.session.type', 'file')
     settings_maker.make_setting('beaker.session.data_dir',  jn(default_cache_dir, 'session_data'))
 
     # cache_general
     settings_maker.make_setting('rc_cache.cache_general.backend', 'dogpile.cache.rc.file_namespace')
     settings_maker.make_setting('rc_cache.cache_general.expiration_time', 60 * 60 * 12, parser='int')
     settings_maker.make_setting('rc_cache.cache_general.arguments.filename', jn(default_cache_dir, 'rhodecode_cache_general.db'))
 
     # cache_perms
     settings_maker.make_setting('rc_cache.cache_perms.backend', 'dogpile.cache.rc.file_namespace')
     settings_maker.make_setting('rc_cache.cache_perms.expiration_time', 60 * 60, parser='int')
     settings_maker.make_setting('rc_cache.cache_perms.arguments.filename', jn(default_cache_dir, 'rhodecode_cache_perms_db'))
 
     # cache_repo
     settings_maker.make_setting('rc_cache.cache_repo.backend', 'dogpile.cache.rc.file_namespace')
     settings_maker.make_setting('rc_cache.cache_repo.expiration_time', 60 * 60 * 24 * 30, parser='int')
     settings_maker.make_setting('rc_cache.cache_repo.arguments.filename', jn(default_cache_dir, 'rhodecode_cache_repo_db'))
 
     # cache_license
     settings_maker.make_setting('rc_cache.cache_license.backend', 'dogpile.cache.rc.file_namespace')
     settings_maker.make_setting('rc_cache.cache_license.expiration_time', 60 * 5, parser='int')
     settings_maker.make_setting('rc_cache.cache_license.arguments.filename', jn(default_cache_dir, 'rhodecode_cache_license_db'))
 
     # cache_repo_longterm memory, 96H
     settings_maker.make_setting('rc_cache.cache_repo_longterm.backend', 'dogpile.cache.rc.memory_lru')
     settings_maker.make_setting('rc_cache.cache_repo_longterm.expiration_time', 345600, parser='int')
     settings_maker.make_setting('rc_cache.cache_repo_longterm.max_size', 10000, parser='int')
 
     # sql_cache_short
     settings_maker.make_setting('rc_cache.sql_cache_short.backend', 'dogpile.cache.rc.memory_lru')
     settings_maker.make_setting('rc_cache.sql_cache_short.expiration_time', 30, parser='int')
     settings_maker.make_setting('rc_cache.sql_cache_short.max_size', 10000, parser='int')
 
     # archive_cache
     settings_maker.make_setting('archive_cache.locking.url', 'redis://redis:6379/1')
     settings_maker.make_setting('archive_cache.backend.type', 'filesystem')
 
     settings_maker.make_setting('archive_cache.filesystem.store_dir', jn(default_cache_dir, 'archive_cache'), default_when_empty=True,)
+    settings_maker.make_setting('archive_cache.filesystem.cache_shards', 8, parser='int')
     settings_maker.make_setting('archive_cache.filesystem.cache_size_gb', 10, parser='float')
-    settings_maker.make_setting('archive_cache.filesystem.cache_shards', 8, parser='int')
     settings_maker.make_setting('archive_cache.filesystem.eviction_policy', 'least-recently-stored')
 
+    settings_maker.make_setting('archive_cache.filesystem.retry', False, parser='bool')
+    settings_maker.make_setting('archive_cache.filesystem.retry_backoff', 1, parser='int')
+    settings_maker.make_setting('archive_cache.filesystem.retry_attempts', 10, parser='int')
+
+    settings_maker.make_setting('archive_cache.objectstore.url', jn(default_cache_dir, 'archive_cache'), default_when_empty=True,)
+    settings_maker.make_setting('archive_cache.objectstore.key', '')
+    settings_maker.make_setting('archive_cache.objectstore.secret', '')
+    settings_maker.make_setting('archive_cache.objectstore.bucket_shards', 8, parser='int')
+
+    settings_maker.make_setting('archive_cache.objectstore.cache_size_gb', 10, parser='float')
+    settings_maker.make_setting('archive_cache.objectstore.eviction_policy', 'least-recently-stored')
+
+    settings_maker.make_setting('archive_cache.objectstore.retry', False, parser='bool')
+    settings_maker.make_setting('archive_cache.objectstore.retry_backoff', 1, parser='int')
+    settings_maker.make_setting('archive_cache.objectstore.retry_attempts', 10, parser='int')
+
     settings_maker.env_expand()
 
     # configure instance id
     config_utils.set_instance_id(settings)
 
     return settings

rhodecode/config/middleware.py

@@ -1,466 +1,466 b''
 # Copyright (C) 2010-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import os
 import sys
 import collections
 
 import time
 import logging.config
 
 from paste.gzipper import make_gzip_middleware
 import pyramid.events
 from pyramid.wsgi import wsgiapp
 from pyramid.config import Configurator
 from pyramid.settings import asbool, aslist
 from pyramid.httpexceptions import (
     HTTPException, HTTPError, HTTPInternalServerError, HTTPFound, HTTPNotFound)
 from pyramid.renderers import render_to_response
 
 from rhodecode.model import meta
 from rhodecode.config import patches
 
 from rhodecode.config.environment import load_pyramid_environment
 
 import rhodecode.events
 from rhodecode.config.config_maker import sanitize_settings_and_apply_defaults
 from rhodecode.lib.middleware.vcs import VCSMiddleware
 from rhodecode.lib.request import Request
 from rhodecode.lib.vcs import VCSCommunicationError
 from rhodecode.lib.exceptions import VCSServerUnavailable
 from rhodecode.lib.middleware.appenlight import wrap_in_appenlight_if_enabled
 from rhodecode.lib.middleware.https_fixup import HttpsFixup
 from rhodecode.lib.plugins.utils import register_rhodecode_plugin
 from rhodecode.lib.utils2 import AttributeDict
 from rhodecode.lib.exc_tracking import store_exception, format_exc
 from rhodecode.subscribers import (
     scan_repositories_if_enabled, write_js_routes_if_enabled,
     write_metadata_if_needed, write_usage_data)
 from rhodecode.lib.statsd_client import StatsdClient
 
 log = logging.getLogger(__name__)
 
 
 def is_http_error(response):
     # error which should have traceback
     return response.status_code > 499
 
 
 def should_load_all():
     """
     Returns if all application components should be loaded. In some cases it's
     desired to skip apps loading for faster shell script execution
     """
     ssh_cmd = os.environ.get('RC_CMD_SSH_WRAPPER')
     if ssh_cmd:
         return False
 
     return True
 
 
 def make_pyramid_app(global_config, **settings):
     """
     Constructs the WSGI application based on Pyramid.
 
     Specials:
 
     * The application can also be integrated like a plugin via the call to
       `includeme`. This is accompanied with the other utility functions which
       are called. Changing this should be done with great care to not break
       cases when these fragments are assembled from another place.
 
     """
     start_time = time.time()
     log.info('Pyramid app config starting')
 
     sanitize_settings_and_apply_defaults(global_config, settings)
 
     # init and bootstrap StatsdClient
     StatsdClient.setup(settings)
 
     config = Configurator(settings=settings)
     # Init our statsd at very start
     config.registry.statsd = StatsdClient.statsd
 
     # Apply compatibility patches
     patches.inspect_getargspec()
 
     load_pyramid_environment(global_config, settings)
 
     # Static file view comes first
     includeme_first(config)
 
     includeme(config)
 
     pyramid_app = config.make_wsgi_app()
     pyramid_app = wrap_app_in_wsgi_middlewares(pyramid_app, config)
     pyramid_app.config = config
 
     celery_settings = get_celery_config(settings)
     config.configure_celery(celery_settings)
 
     # creating the app uses a connection - return it after we are done
     meta.Session.remove()
 
     total_time = time.time() - start_time
     log.info('Pyramid app created and configured in %.2fs', total_time)
     return pyramid_app
 
 
 def get_celery_config(settings):
     """
     Converts basic ini configuration into celery 4.X options
     """
 
     def key_converter(key_name):
         pref = 'celery.'
         if key_name.startswith(pref):
             return key_name[len(pref):].replace('.', '_').lower()
 
     def type_converter(parsed_key, value):
         # cast to int
         if value.isdigit():
             return int(value)
 
         # cast to bool
         if value.lower() in ['true', 'false', 'True', 'False']:
             return value.lower() == 'true'
         return value
 
     celery_config = {}
     for k, v in settings.items():
         pref = 'celery.'
         if k.startswith(pref):
             celery_config[key_converter(k)] = type_converter(key_converter(k), v)
 
     # TODO:rethink if we want to support celerybeat based file config, probably NOT
     # beat_config = {}
     # for section in parser.sections():
     #     if section.startswith('celerybeat:'):
     #         name = section.split(':', 1)[1]
     #         beat_config[name] = get_beat_config(parser, section)
 
     # final compose of settings
     celery_settings = {}
 
     if celery_config:
         celery_settings.update(celery_config)
     # if beat_config:
     #     celery_settings.update({'beat_schedule': beat_config})
 
     return celery_settings
 
 
 def not_found_view(request):
     """
     This creates the view which should be registered as not-found-view to
     pyramid.
     """
 
     if not getattr(request, 'vcs_call', None):
         # handle like regular case with our error_handler
         return error_handler(HTTPNotFound(), request)
 
     # handle not found view as a vcs call
     settings = request.registry.settings
     ae_client = getattr(request, 'ae_client', None)
     vcs_app = VCSMiddleware(
         HTTPNotFound(), request.registry, settings,
         appenlight_client=ae_client)
 
     return wsgiapp(vcs_app)(None, request)
 
 
 def error_handler(exception, request):
     import rhodecode
     from rhodecode.lib import helpers
 
     rhodecode_title = rhodecode.CONFIG.get('rhodecode_title') or 'RhodeCode'
 
     base_response = HTTPInternalServerError()
     # prefer original exception for the response since it may have headers set
     if isinstance(exception, HTTPException):
         base_response = exception
     elif isinstance(exception, VCSCommunicationError):
         base_response = VCSServerUnavailable()
 
     if is_http_error(base_response):
         traceback_info = format_exc(request.exc_info)
         log.error(
             'error occurred handling this request for path: %s, \n%s',
             request.path, traceback_info)
 
     error_explanation = base_response.explanation or str(base_response)
     if base_response.status_code == 404:
         error_explanation += " Optionally you don't have permission to access this page."
     c = AttributeDict()
     c.error_message = base_response.status
     c.error_explanation = error_explanation
     c.visual = AttributeDict()
 
     c.visual.rhodecode_support_url = (
         request.registry.settings.get('rhodecode_support_url') or
         request.route_url('rhodecode_support')
     )
     c.redirect_time = 0
     c.rhodecode_name = rhodecode_title
     if not c.rhodecode_name:
         c.rhodecode_name = 'Rhodecode'
 
     c.causes = []
     if is_http_error(base_response):
         c.causes.append('Server is overloaded.')
         c.causes.append('Server database connection is lost.')
         c.causes.append('Server expected unhandled error.')
 
     if hasattr(base_response, 'causes'):
         c.causes = base_response.causes
 
     c.messages = helpers.flash.pop_messages(request=request)
     exc_info = sys.exc_info()
     c.exception_id = id(exc_info)
     c.show_exception_id = isinstance(base_response, VCSServerUnavailable) \
                           or base_response.status_code > 499
     c.exception_id_url = request.route_url(
         'admin_settings_exception_tracker_show', exception_id=c.exception_id)
 
     debug_mode = rhodecode.ConfigGet().get_bool('debug')
     if c.show_exception_id:
         store_exception(c.exception_id, exc_info)
     c.exception_debug = debug_mode
     c.exception_config_ini = rhodecode.CONFIG.get('__file__')
 
     if debug_mode:
         try:
             from rich.traceback import install
             install(show_locals=True)
             log.debug('Installing rich tracebacks...')
         except ImportError:
             pass
 
     response = render_to_response(
         '/errors/error_document.mako', {'c': c, 'h': helpers}, request=request,
         response=base_response)
 
     response.headers["X-RC-Exception-Id"] = str(c.exception_id)
 
     statsd = request.registry.statsd
     if statsd and base_response.status_code > 499:
         exc_type = f"{exception.__class__.__module__}.{exception.__class__.__name__}"
         statsd.incr('rhodecode_exception_total',
                     tags=["exc_source:web",
                           f"http_code:{base_response.status_code}",
                           f"type:{exc_type}"])
 
     return response
 
 
 def includeme_first(config):
     # redirect automatic browser favicon.ico requests to correct place
     def favicon_redirect(context, request):
         return HTTPFound(
             request.static_path('rhodecode:public/images/favicon.ico'))
 
     config.add_view(favicon_redirect, route_name='favicon')
     config.add_route('favicon', '/favicon.ico')
 
     def robots_redirect(context, request):
         return HTTPFound(
             request.static_path('rhodecode:public/robots.txt'))
 
     config.add_view(robots_redirect, route_name='robots')
     config.add_route('robots', '/robots.txt')
 
     config.add_static_view(
         '_static/deform', 'deform:static')
     config.add_static_view(
         '_static/rhodecode', path='rhodecode:public', cache_max_age=3600 * 24)
 
 
 ce_auth_resources = [
     'rhodecode.authentication.plugins.auth_crowd',
     'rhodecode.authentication.plugins.auth_headers',
     'rhodecode.authentication.plugins.auth_jasig_cas',
     'rhodecode.authentication.plugins.auth_ldap',
     'rhodecode.authentication.plugins.auth_pam',
     'rhodecode.authentication.plugins.auth_rhodecode',
     'rhodecode.authentication.plugins.auth_token',
 ]
 
 
 def includeme(config, auth_resources=None):
     from rhodecode.lib.celerylib.loader import configure_celery
     log.debug('Initializing main includeme from %s', os.path.basename(__file__))
     settings = config.registry.settings
     config.set_request_factory(Request)
 
     # plugin information
     config.registry.rhodecode_plugins = collections.OrderedDict()
 
     config.add_directive(
         'register_rhodecode_plugin', register_rhodecode_plugin)
 
     config.add_directive('configure_celery', configure_celery)
 
     if settings.get('appenlight', False):
         config.include('appenlight_client.ext.pyramid_tween')
 
     load_all = should_load_all()
 
     # Includes which are required. The application would fail without them.
     config.include('pyramid_mako')
     config.include('rhodecode.lib.rc_beaker')
     config.include('rhodecode.lib.rc_cache')
-    config.include('rhodecode.lib.rc_cache.archive_cache')
+    config.include('rhodecode.lib.archive_cache')
 
     config.include('rhodecode.apps._base.navigation')
     config.include('rhodecode.apps._base.subscribers')
     config.include('rhodecode.tweens')
     config.include('rhodecode.authentication')
 
     if load_all:
 
         # load CE authentication plugins
 
         if auth_resources:
             ce_auth_resources.extend(auth_resources)
 
         for resource in ce_auth_resources:
             config.include(resource)
 
         # Auto discover authentication plugins and include their configuration.
         if asbool(settings.get('auth_plugin.import_legacy_plugins', 'true')):
             from rhodecode.authentication import discover_legacy_plugins
             discover_legacy_plugins(config)
 
     # apps
     if load_all:
         log.debug('Starting config.include() calls')
         config.include('rhodecode.api.includeme')
         config.include('rhodecode.apps._base.includeme')
         config.include('rhodecode.apps._base.navigation.includeme')
         config.include('rhodecode.apps._base.subscribers.includeme')
         config.include('rhodecode.apps.hovercards.includeme')
         config.include('rhodecode.apps.ops.includeme')
         config.include('rhodecode.apps.channelstream.includeme')
         config.include('rhodecode.apps.file_store.includeme')
         config.include('rhodecode.apps.admin.includeme')
         config.include('rhodecode.apps.login.includeme')
         config.include('rhodecode.apps.home.includeme')
         config.include('rhodecode.apps.journal.includeme')
 
         config.include('rhodecode.apps.repository.includeme')
         config.include('rhodecode.apps.repo_group.includeme')
         config.include('rhodecode.apps.user_group.includeme')
         config.include('rhodecode.apps.search.includeme')
         config.include('rhodecode.apps.user_profile.includeme')
         config.include('rhodecode.apps.user_group_profile.includeme')
         config.include('rhodecode.apps.my_account.includeme')
         config.include('rhodecode.apps.gist.includeme')
 
         config.include('rhodecode.apps.svn_support.includeme')
         config.include('rhodecode.apps.ssh_support.includeme')
         config.include('rhodecode.apps.debug_style')
 
     if load_all:
         config.include('rhodecode.integrations.includeme')
         config.include('rhodecode.integrations.routes.includeme')
 
     config.add_route('rhodecode_support', 'https://rhodecode.com/help/', static=True)
     settings['default_locale_name'] = settings.get('lang', 'en')
     config.add_translation_dirs('rhodecode:i18n/')
 
     # Add subscribers.
     if load_all:
         log.debug('Adding subscribers...')
         config.add_subscriber(scan_repositories_if_enabled,
                               pyramid.events.ApplicationCreated)
         config.add_subscriber(write_metadata_if_needed,
                               pyramid.events.ApplicationCreated)
         config.add_subscriber(write_usage_data,
                               pyramid.events.ApplicationCreated)
         config.add_subscriber(write_js_routes_if_enabled,
                               pyramid.events.ApplicationCreated)
 
 
     # Set the default renderer for HTML templates to mako.
     config.add_mako_renderer('.html')
 
     config.add_renderer(
         name='json_ext',
         factory='rhodecode.lib.ext_json_renderer.pyramid_ext_json')
 
     config.add_renderer(
         name='string_html',
         factory='rhodecode.lib.string_renderer.html')
 
     # include RhodeCode plugins
     includes = aslist(settings.get('rhodecode.includes', []))
     log.debug('processing rhodecode.includes data...')
     for inc in includes:
         config.include(inc)
 
     # custom not found view, if our pyramid app doesn't know how to handle
     # the request pass it to potential VCS handling ap
     config.add_notfound_view(not_found_view)
     if not settings.get('debugtoolbar.enabled', False):
         # disabled debugtoolbar handle all exceptions via the error_handlers
         config.add_view(error_handler, context=Exception)
 
     # all errors including 403/404/50X
     config.add_view(error_handler, context=HTTPError)
 
 
 def wrap_app_in_wsgi_middlewares(pyramid_app, config):
     """
     Apply outer WSGI middlewares around the application.
     """
     registry = config.registry
     settings = registry.settings
 
     # enable https redirects based on HTTP_X_URL_SCHEME set by proxy
     pyramid_app = HttpsFixup(pyramid_app, settings)
 
     pyramid_app, _ae_client = wrap_in_appenlight_if_enabled(
         pyramid_app, settings)
     registry.ae_client = _ae_client
 
     if settings['gzip_responses']:
         pyramid_app = make_gzip_middleware(
             pyramid_app, settings, compress_level=1)
 
     # this should be the outer most middleware in the wsgi stack since
     # middleware like Routes make database calls
     def pyramid_app_with_cleanup(environ, start_response):
         start = time.time()
         try:
             return pyramid_app(environ, start_response)
         finally:
             # Dispose current database session and rollback uncommitted
             # transactions.
             meta.Session.remove()
 
             # In a single threaded mode server, on non sqlite db we should have
             # '0 Current Checked out connections' at the end of a request,
             # if not, then something, somewhere is leaving a connection open
             pool = meta.get_engine().pool
             log.debug('sa pool status: %s', pool.status())
             total = time.time() - start
             log.debug('Request processing finalized: %.4fs', total)
 
     return pyramid_app_with_cleanup

rhodecode/lib/archive_cache/__init__.py

@@ -1,29 +1,78 b''
 # Copyright (C) 2015-2024 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
-from .fanout_cache import get_archival_cache_store
-from .fanout_cache import get_archival_config
+import logging
+
+from .backends.fanout_cache import FileSystemFanoutCache
+from .backends.objectstore_cache import ObjectStoreCache
 
-from .utils import archive_iterator
-from .utils import ArchiveCacheGenerationLock
+from .utils import archive_iterator  # noqa
+from .lock import ArchiveCacheGenerationLock # noqa
+
+log = logging.getLogger(__name__)
+
+
+cache_meta = None
 
 
 def includeme(config):
     # init our cache at start
     settings = config.get_settings()
     get_archival_cache_store(settings)
+
+
+def get_archival_config(config):
+
+    final_config = {
+
+    }
+
+    for k, v in config.items():
+        if k.startswith('archive_cache'):
+            final_config[k] = v
+
+    return final_config
+
+
+def get_archival_cache_store(config, always_init=False):
+
+    global cache_meta
+    if cache_meta is not None and not always_init:
+        return cache_meta
+
+    config = get_archival_config(config)
+    backend = config['archive_cache.backend.type']
+
+    archive_cache_locking_url = config['archive_cache.locking.url']
+
+    match backend:
+        case 'filesystem':
+            d_cache = FileSystemFanoutCache(
+                locking_url=archive_cache_locking_url,
+                **config
+            )
+        case 'objectstore':
+            d_cache = ObjectStoreCache(
+                locking_url=archive_cache_locking_url,
+                **config
+            )
+        case _:
+            raise ValueError(f'archive_cache.backend.type only supports "filesystem" or "objectstore" got {backend} ')
+
+    cache_meta = d_cache
+    return cache_meta

rhodecode/lib/archive_cache/backends/fanout_cache.py

@@ -1,456 +1,166 b''
 # Copyright (C) 2015-2024 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import codecs
-import contextlib
-import functools
-import os
+import hashlib
 import logging
-import time
-import typing
-import zlib
-import sqlite3
+import os
+
+import fsspec
 
-from ...ext_json import json
-from .lock import GenerationLock
-from .utils import format_size
+from .base import BaseCache, BaseShard
+from ..utils import ShardFileReader, NOT_GIVEN
+from ...type_utils import str2bool
 
 log = logging.getLogger(__name__)
 
-cache_meta = None
 
-UNKNOWN = -241
-NO_VAL = -917
-
-MODE_BINARY = 'BINARY'
-
-
-EVICTION_POLICY = {
-    'none': {
-        'evict': None,
-    },
-    'least-recently-stored': {
-        'evict': 'SELECT {fields} FROM archive_cache ORDER BY store_time',
-    },
-    'least-recently-used': {
-        'evict': 'SELECT {fields} FROM archive_cache ORDER BY access_time',
-    },
-    'least-frequently-used': {
-        'evict': 'SELECT {fields} FROM archive_cache ORDER BY access_count',
-    },
-}
-
-
-class DB:
-
-    def __init__(self):
-        self.connection = sqlite3.connect(':memory:')
-        self._init_db()
-
-    def _init_db(self):
-        qry = '''
-            CREATE TABLE IF NOT EXISTS archive_cache (
-             rowid INTEGER PRIMARY KEY,
-             key_file TEXT,
-             key_file_path TEXT,
-             filename TEXT,
-             full_path TEXT,
-             store_time REAL,
-             access_time REAL,
-             access_count INTEGER DEFAULT 0,
-             size INTEGER DEFAULT 0
-             )
-        '''
-
-        self.sql(qry)
-        self.connection.commit()
-
-    @property
-    def sql(self):
-        return self.connection.execute
-
-    def bulk_insert(self, rows):
-        qry = '''
-            INSERT INTO archive_cache (
-             rowid,
-             key_file,
-             key_file_path,
-             filename,
-             full_path,
-             store_time,
-             access_time,
-             access_count,
-             size
-            )
-            VALUES (
-            ?, ?, ?, ?, ?, ?, ?, ?, ?
-            )
-        '''
-        cursor = self.connection.cursor()
-        cursor.executemany(qry, rows)
-        self.connection.commit()
-
-
-class FileSystemCache:
+class FileSystemShard(BaseShard):
 
     def __init__(self, index, directory, **settings):
         self._index = index
         self._directory = directory
+        self.storage_type = 'directory'
+        self.fs = fsspec.filesystem('file')
 
     @property
     def directory(self):
         """Cache directory."""
         return self._directory
 
-    def _write_file(self, full_path, iterator, mode, encoding=None):
-        full_dir, _ = os.path.split(full_path)
+    def _get_keyfile(self, archive_key) -> tuple[str, str]:
+        key_file = f'{archive_key}.{self.key_suffix}'
+        return key_file, os.path.join(self.directory, key_file)
 
+    def _get_writer(self, path, mode):
         for count in range(1, 11):
-            with contextlib.suppress(OSError):
-                os.makedirs(full_dir)
-
             try:
                 # Another cache may have deleted the directory before
                 # the file could be opened.
-                writer = open(full_path, mode, encoding=encoding)
+                return self.fs.open(path, mode)
             except OSError:
                 if count == 10:
                     # Give up after 10 tries to open the file.
                     raise
                 continue
 
+    def _write_file(self, full_path, iterator, mode):
+        # ensure dir exists
+        destination, _ = os.path.split(full_path)
+        if not self.fs.exists(destination):
+            self.fs.makedirs(destination)
+
+        writer = self._get_writer(full_path, mode)
+
+        digest = hashlib.sha256()
             with writer:
                 size = 0
                 for chunk in iterator:
                     size += len(chunk)
+                digest.update(chunk)
                     writer.write(chunk)
                 writer.flush()
                 # Get the file descriptor
                 fd = writer.fileno()
 
                 # Sync the file descriptor to disk, helps with NFS cases...
                 os.fsync(fd)
-                log.debug('written new archive cache under %s', full_path)
-                return size
-
-    def _get_keyfile(self, key):
-        return os.path.join(self._directory, f'{key}.key')
-
-    def store(self, key, value_reader, metadata):
-        filename, full_path = self.random_filename()
-        key_file = self._get_keyfile(key)
-
-        # STORE METADATA
-        _metadata = {
-            "version": "v1",
-            "filename": filename,
-            "full_path": full_path,
-            "key_file": key_file,
-            "store_time": time.time(),
-            "access_count": 1,
-            "access_time": 0,
-            "size": 0
-        }
-        if metadata:
-            _metadata.update(metadata)
-
-        reader = functools.partial(value_reader.read, 2**22)
-
-        iterator = iter(reader, b'')
-        size = self._write_file(full_path, iterator, 'xb')
-        metadata['size'] = size
+            sha256 = digest.hexdigest()
+            log.debug('written new archive cache under %s, sha256: %s', full_path, sha256)
+            return size, sha256
 
-        # after archive is finished, we create a key to save the presence of the binary file
-        with open(key_file, 'wb') as f:
-            f.write(json.dumps(_metadata))
-
-        return key, size, MODE_BINARY, filename, _metadata
-
-    def fetch(self, key, retry=False, retry_attempts=10) -> tuple[typing.BinaryIO, dict]:
-
-        if retry:
-            for attempt in range(retry_attempts):
-                if key in self:
-                    break
-                # we dind't find the key, wait 1s, and re-check
-                time.sleep(1)
+    def store(self, key, value_reader, metadata: dict | None = None):
+        return self._store(key, value_reader, metadata, mode='xb')
 
-        if key not in self:
-            log.exception('requested {key} not found in {self}', key, self)
-            raise KeyError(key)
-
-        key_file = self._get_keyfile(key)
-        with open(key_file, 'rb') as f:
-            metadata = json.loads(f.read())
-
-        filename = metadata['filename']
+    def fetch(self, key, retry=NOT_GIVEN, retry_attempts=NOT_GIVEN, retry_backoff=1) -> tuple[ShardFileReader, dict]:
+        return self._fetch(key, retry, retry_attempts, retry_backoff)
 
-        try:
-            return open(os.path.join(self.directory, filename), 'rb'), metadata
-        finally:
-            # update usage stats, count and accessed
-            metadata["access_count"] = metadata.get("access_count", 0) + 1
-            metadata["access_time"] = time.time()
-
-            with open(key_file, 'wb') as f:
-                f.write(json.dumps(metadata))
+    def remove(self, key):
+        return self._remove(key)
 
     def random_filename(self):
         """Return filename and full-path tuple for file storage.
 
         Filename will be a randomly generated 28 character hexadecimal string
         with ".archive_cache" suffixed. Two levels of sub-directories will be used to
         reduce the size of directories. On older filesystems, lookups in
         directories with many files may be slow.
         """
 
         hex_name = codecs.encode(os.urandom(16), 'hex').decode('utf-8')
-        sub_dir = os.path.join(hex_name[:2], hex_name[2:4])
-        name = hex_name[4:] + '.archive_cache'
-        filename = os.path.join(sub_dir, name)
-        full_path = os.path.join(self.directory, filename)
-        return filename, full_path
-
-    def hash(self, key):
-        """Compute portable hash for `key`.
-
-        :param key: key to hash
-        :return: hash value
 
-        """
-        mask = 0xFFFFFFFF
-        return zlib.adler32(key.encode('utf-8')) & mask  # noqa
-
-    def __contains__(self, key):
-        """Return `True` if `key` matching item is found in cache.
+        archive_name = hex_name[4:] + '.archive_cache'
+        filename = f"{hex_name[:2]}/{hex_name[2:4]}/{archive_name}"
 
-        :param key: key matching item
-        :return: True if key matching item
-
-        """
-        key_file = self._get_keyfile(key)
-        return os.path.exists(key_file)
+        full_path = os.path.join(self.directory, filename)
+        return archive_name, full_path
 
     def __repr__(self):
-        return f'FileSystemCache(index={self._index}, dir={self.directory})'
+        return f'{self.__class__.__name__}(index={self._index}, dir={self.directory})'
 
 
-class FanoutCache:
-    """Cache that shards keys and values."""
+class FileSystemFanoutCache(BaseCache):
 
-    def __init__(
-        self, directory=None, **settings
-    ):
-        """Initialize cache instance.
+    def __init__(self, locking_url, **settings):
+        """
+        Initialize file system cache instance.
 
-        :param str directory: cache directory
+        :param str locking_url: redis url for a lock
         :param settings: settings dict
 
         """
-        if directory is None:
-            raise ValueError('directory cannot be None')
-
-        directory = str(directory)
+        self._locking_url = locking_url
+        self._config = settings
+        cache_dir = self.get_conf('archive_cache.filesystem.store_dir')
+        directory = str(cache_dir)
         directory = os.path.expanduser(directory)
         directory = os.path.expandvars(directory)
         self._directory = directory
+        self._storage_path = directory
 
-        self._count = settings.pop('cache_shards')
-        self._locking_url = settings.pop('locking_url')
+        # check if it's ok to write, and re-create the archive cache
+        if not os.path.isdir(self._directory):
+            os.makedirs(self._directory, exist_ok=True)
+
+        self._count = int(self.get_conf('archive_cache.filesystem.cache_shards', pop=True))
 
-        self._eviction_policy = settings['cache_eviction_policy']
-        self._cache_size_limit = settings['cache_size_limit']
+        self._eviction_policy = self.get_conf('archive_cache.filesystem.eviction_policy', pop=True)
+        self._cache_size_limit = self.gb_to_bytes(int(self.get_conf('archive_cache.filesystem.cache_size_gb')))
 
+        self.retry = str2bool(self.get_conf('archive_cache.filesystem.retry', pop=True))
+        self.retry_attempts = int(self.get_conf('archive_cache.filesystem.retry_attempts', pop=True))
+        self.retry_backoff = int(self.get_conf('archive_cache.filesystem.retry_backoff', pop=True))
+
+        log.debug('Initializing archival cache instance under %s', self._directory)
         self._shards = tuple(
-            FileSystemCache(
+            FileSystemShard(
                 index=num,
                 directory=os.path.join(directory, 'shard_%03d' % num),
                 **settings,
             )
             for num in range(self._count)
         )
         self._hash = self._shards[0].hash
 
-    @property
-    def directory(self):
-        """Cache directory."""
-        return self._directory
-
-    def get_lock(self, lock_key):
-        return GenerationLock(lock_key, self._locking_url)
-
-    def _get_shard(self, key) -> FileSystemCache:
+    def _get_shard(self, key) -> FileSystemShard:
         index = self._hash(key) % self._count
         shard = self._shards[index]
         return shard
 
-    def store(self, key, value_reader, metadata=None):
-        shard = self._get_shard(key)
-        return shard.store(key, value_reader, metadata)
-
-    def fetch(self, key, retry=False, retry_attempts=10):
-        """Return file handle corresponding to `key` from cache.
-        """
-        shard = self._get_shard(key)
-        return shard.fetch(key, retry=retry, retry_attempts=retry_attempts)
-
-    def has_key(self, key):
-        """Return `True` if `key` matching item is found in cache.
-
-        :param key: key for item
-        :return: True if key is found
-
-        """
-        shard = self._get_shard(key)
-        return key in shard
-
-    def __contains__(self, item):
-        return self.has_key(item)
-
-    def evict(self, policy=None, size_limit=None):
-        """
-        Remove old items based on the conditions
-
-
-        explanation of this algo:
-        iterate over each shard, then for each shard iterate over the .key files
-        read the key files metadata stored. This gives us a full list of keys, cached_archived, their size and
-        access data, time creation, and access counts.
-
-        Store that into a memory DB so we can run different sorting strategies easily.
-        Summing the size is a sum sql query.
-
-        Then we run a sorting strategy based on eviction policy.
-        We iterate over sorted keys, and remove each checking if we hit the overall limit.
-        """
-
-        policy = policy or self._eviction_policy
-        size_limit = size_limit or self._cache_size_limit
-
-        select_policy = EVICTION_POLICY[policy]['evict']
-
-        log.debug('Running eviction policy \'%s\', and checking for size limit: %s',
-                  policy, format_size(size_limit))
-
-        if select_policy is None:
-            return 0
-
-        db = DB()
-
-        data = []
-        cnt = 1
-        for shard in self._shards:
-            for key_file in os.listdir(shard.directory):
-                if key_file.endswith('.key'):
-                    key_file_path = os.path.join(shard.directory, key_file)
-                    with open(key_file_path, 'rb') as f:
-                        metadata = json.loads(f.read())
-
-                        size = metadata.get('size')
-                        filename = metadata.get('filename')
-                        full_path = metadata.get('full_path')
-
-                        if not size:
-                            # in case we don't have size re-calc it...
-                            size = os.stat(full_path).st_size
-
-                        data.append([
-                          cnt,
-                          key_file,
-                          key_file_path,
-                          filename,
-                          full_path,
-                          metadata.get('store_time', 0),
-                          metadata.get('access_time', 0),
-                          metadata.get('access_count', 0),
-                          size,
-                        ])
-                        cnt += 1
-
-        # Insert bulk data using executemany
-        db.bulk_insert(data)
-
-        ((total_size,),) = db.sql('SELECT COALESCE(SUM(size), 0) FROM archive_cache').fetchall()
-        log.debug('Analyzed %s keys, occupied: %s', len(data), format_size(total_size))
-        select_policy_qry = select_policy.format(fields='key_file_path, full_path, size')
-        sorted_keys = db.sql(select_policy_qry).fetchall()
-
-        removed_items = 0
-        removed_size = 0
-        for key, cached_file, size in sorted_keys:
-            # simulate removal impact BEFORE removal
-            total_size -= size
-
-            if total_size <= size_limit:
-                # we obtained what we wanted...
-                break
-
-            os.remove(cached_file)
-            os.remove(key)
-            removed_items += 1
-            removed_size += size
-
-        log.debug('Removed %s cache archives, and reduced size: %s', removed_items, format_size(removed_size))
-        return removed_items
-
-
-def get_archival_config(config):
-
-    final_config = {
-
-    }
-
-    for k, v in config.items():
-        if k.startswith('archive_cache'):
-            final_config[k] = v
-
-    return final_config
-
-
-def get_archival_cache_store(config):
-
-    global cache_meta
-    if cache_meta is not None:
-        return cache_meta
-
-    config = get_archival_config(config)
-    backend = config['archive_cache.backend.type']
-    if backend != 'filesystem':
-        raise ValueError('archive_cache.backend.type only supports "filesystem"')
-
-    archive_cache_locking_url = config['archive_cache.locking.url']
-    archive_cache_dir = config['archive_cache.filesystem.store_dir']
-    archive_cache_size_gb = config['archive_cache.filesystem.cache_size_gb']
-    archive_cache_shards = config['archive_cache.filesystem.cache_shards']
-    archive_cache_eviction_policy = config['archive_cache.filesystem.eviction_policy']
-
-    log.debug('Initializing archival cache instance under %s', archive_cache_dir)
-
-    # check if it's ok to write, and re-create the archive cache
-    if not os.path.isdir(archive_cache_dir):
-        os.makedirs(archive_cache_dir, exist_ok=True)
-
-    d_cache = FanoutCache(
-        archive_cache_dir,
-        locking_url=archive_cache_locking_url,
-        cache_shards=archive_cache_shards,
-        cache_size_limit=archive_cache_size_gb * 1024 * 1024 * 1024,
-        cache_eviction_policy=archive_cache_eviction_policy
-    )
-    cache_meta = d_cache
-    return cache_meta
+    def _get_size(self, shard, archive_path):
+        return os.stat(archive_path).st_size

rhodecode/lib/archive_cache/lock.py

@@ -1,59 +1,62 b''
 # Copyright (C) 2015-2024 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import redis
-from ..._vendor import redis_lock
-from .utils import ArchiveCacheGenerationLock
+from .._vendor import redis_lock
+
+
+class ArchiveCacheGenerationLock(Exception):
+    pass
 
 
 class GenerationLock:
     """
     Locking mechanism that detects if a lock is acquired
 
     with GenerationLock(lock_key):
         compute_archive()
     """
     lock_timeout = 7200
 
     def __init__(self, lock_key, url):
         self.lock_key = lock_key
         self._create_client(url)
         self.lock = self.get_lock()
 
     def _create_client(self, url):
         connection_pool = redis.ConnectionPool.from_url(url)
         self.writer_client = redis.StrictRedis(
             connection_pool=connection_pool
         )
         self.reader_client = self.writer_client
 
     def get_lock(self):
         return redis_lock.Lock(
             redis_client=self.writer_client,
             name=self.lock_key,
             expire=self.lock_timeout,
             strict=True
         )
 
     def __enter__(self):
         acquired = self.lock.acquire(blocking=False)
         if not acquired:
             raise ArchiveCacheGenerationLock('Failed to create a lock')
 
     def __exit__(self, exc_type, exc_val, exc_tb):
         self.lock.release()

rhodecode/lib/archive_cache/utils.py

@@ -1,72 +1,134 b''
 # Copyright (C) 2015-2024 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
-import os
+import sqlite3
+import s3fs.core
+
+NOT_GIVEN = -917
 
 
-class ArchiveCacheGenerationLock(Exception):
-    pass
+EVICTION_POLICY = {
+    'none': {
+        'evict': None,
+    },
+    'least-recently-stored': {
+        'evict': 'SELECT {fields} FROM archive_cache ORDER BY store_time',
+    },
+    'least-recently-used': {
+        'evict': 'SELECT {fields} FROM archive_cache ORDER BY access_time',
+    },
+    'least-frequently-used': {
+        'evict': 'SELECT {fields} FROM archive_cache ORDER BY access_count',
+    },
+}
 
 
 def archive_iterator(_reader, block_size: int = 4096 * 512):
     # 4096 * 64 = 64KB
     while 1:
         data = _reader.read(block_size)
         if not data:
             break
         yield data
 
 
-def get_directory_statistics(start_path):
-    """
-    total_files, total_size, directory_stats = get_directory_statistics(start_path)
-
-    print(f"Directory statistics for: {start_path}\n")
-    print(f"Total files: {total_files}")
-    print(f"Total size: {format_size(total_size)}\n")
-
-    :param start_path:
-    :return:
-    """
-
-    total_files = 0
-    total_size = 0
-    directory_stats = {}
-
-    for dir_path, dir_names, file_names in os.walk(start_path):
-        dir_size = 0
-        file_count = len(file_names)
-
-        for file in file_names:
-            filepath = os.path.join(dir_path, file)
-            file_size = os.path.getsize(filepath)
-            dir_size += file_size
-
-        directory_stats[dir_path] = {'file_count': file_count, 'size': dir_size}
-        total_files += file_count
-        total_size += dir_size
-
-    return total_files, total_size, directory_stats
-
-
 def format_size(size):
     # Convert size in bytes to a human-readable format (e.g., KB, MB, GB)
     for unit in ['B', 'KB', 'MB', 'GB', 'TB']:
         if size < 1024:
             return f"{size:.2f} {unit}"
         size /= 1024
+
+
+class StatsDB:
+
+    def __init__(self):
+        self.connection = sqlite3.connect(':memory:')
+        self._init_db()
+
+    def _init_db(self):
+        qry = '''
+            CREATE TABLE IF NOT EXISTS archive_cache (
+             rowid INTEGER PRIMARY KEY,
+             key_file TEXT,
+             key_file_path TEXT,
+             archive_key TEXT,
+             archive_path TEXT,
+             store_time REAL,
+             access_time REAL,
+             access_count INTEGER DEFAULT 0,
+             size INTEGER DEFAULT 0
+             )
+        '''
+
+        self.sql(qry)
+        self.connection.commit()
+
+    @property
+    def sql(self):
+        return self.connection.execute
+
+    def bulk_insert(self, rows):
+        qry = '''
+            INSERT INTO archive_cache (
+             rowid,
+             key_file,
+             key_file_path,
+             archive_key,
+             archive_path,
+             store_time,
+             access_time,
+             access_count,
+             size
+            )
+            VALUES (
+            ?, ?, ?, ?, ?, ?, ?, ?, ?
+            )
+        '''
+        cursor = self.connection.cursor()
+        cursor.executemany(qry, rows)
+        self.connection.commit()
+
+    def get_total_size(self):
+        qry = 'SELECT COALESCE(SUM(size), 0) FROM archive_cache'
+        ((total_size,),) = self.sql(qry).fetchall()
+        return total_size
+
+    def get_sorted_keys(self, select_policy):
+        select_policy_qry = select_policy.format(fields='key_file, archive_key, size')
+        return self.sql(select_policy_qry).fetchall()
+
+
+class ShardFileReader:
+
+    def __init__(self, file_like_reader):
+        self._file_like_reader = file_like_reader
+
+    def __getattr__(self, item):
+        if isinstance(self._file_like_reader, s3fs.core.S3File):
+            match item:
+                case 'name':
+                    # S3 FileWrapper doesn't support name attribute, and we use it
+                    return self._file_like_reader.full_name
+                case _:
+                    return getattr(self._file_like_reader, item)
+        else:
+            return getattr(self._file_like_reader, item)
+
+    def __repr__(self):
+        return f'<{self.__class__.__name__}={self._file_like_reader}>'

rhodecode/lib/base.py

@@ -1,607 +1,607 b''
 # Copyright (C) 2010-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 """
 The base Controller API
 Provides the BaseController class for subclassing. And usage in different
 controllers
 """
 
 import logging
 import socket
 import base64
 
 import markupsafe
 import ipaddress
 
 import paste.httpheaders
 from paste.auth.basic import AuthBasicAuthenticator
 from paste.httpexceptions import HTTPUnauthorized, HTTPForbidden, get_exception
 
 import rhodecode
 from rhodecode.authentication.base import VCS_TYPE
 from rhodecode.lib import auth, utils2
 from rhodecode.lib import helpers as h
 from rhodecode.lib.auth import AuthUser, CookieStoreWrapper
 from rhodecode.lib.exceptions import UserCreationError
 from rhodecode.lib.utils import (password_changed, get_enabled_hook_classes)
 from rhodecode.lib.utils2 import AttributeDict
 from rhodecode.lib.str_utils import ascii_bytes, safe_int, safe_str
 from rhodecode.lib.type_utils import aslist, str2bool
 from rhodecode.lib.hash_utils import sha1
 from rhodecode.model.db import Repository, User, ChangesetComment, UserBookmark
 from rhodecode.model.notification import NotificationModel
 from rhodecode.model.settings import VcsSettingsModel, SettingsModel
 
 log = logging.getLogger(__name__)
 
 
 def _filter_proxy(ip):
     """
     Passed in IP addresses in HEADERS can be in a special format of multiple
     ips. Those comma separated IPs are passed from various proxies in the
     chain of request processing. The left-most being the original client.
     We only care about the first IP which came from the org. client.
 
     :param ip: ip string from headers
     """
     if ',' in ip:
         _ips = ip.split(',')
         _first_ip = _ips[0].strip()
         log.debug('Got multiple IPs %s, using %s', ','.join(_ips), _first_ip)
         return _first_ip
     return ip
 
 
 def _filter_port(ip):
     """
     Removes a port from ip, there are 4 main cases to handle here.
     - ipv4 eg. 127.0.0.1
     - ipv6 eg. ::1
     - ipv4+port eg. 127.0.0.1:8080
     - ipv6+port eg. [::1]:8080
 
     :param ip:
     """
     def is_ipv6(ip_addr):
         if hasattr(socket, 'inet_pton'):
             try:
                 socket.inet_pton(socket.AF_INET6, ip_addr)
             except socket.error:
                 return False
         else:
             # fallback to ipaddress
             try:
                 ipaddress.IPv6Address(safe_str(ip_addr))
             except Exception:
                 return False
         return True
 
     if ':' not in ip:  # must be ipv4 pure ip
         return ip
 
     if '[' in ip and ']' in ip:  # ipv6 with port
         return ip.split(']')[0][1:].lower()
 
     # must be ipv6 or ipv4 with port
     if is_ipv6(ip):
         return ip
     else:
         ip, _port = ip.split(':')[:2]  # means ipv4+port
         return ip
 
 
 def get_ip_addr(environ):
     proxy_key = 'HTTP_X_REAL_IP'
     proxy_key2 = 'HTTP_X_FORWARDED_FOR'
     def_key = 'REMOTE_ADDR'
 
     def ip_filters(ip_):
         return _filter_port(_filter_proxy(ip_))
 
     ip = environ.get(proxy_key)
     if ip:
         return ip_filters(ip)
 
     ip = environ.get(proxy_key2)
     if ip:
         return ip_filters(ip)
 
     ip = environ.get(def_key, '0.0.0.0')
     return ip_filters(ip)
 
 
 def get_server_ip_addr(environ, log_errors=True):
     hostname = environ.get('SERVER_NAME')
     try:
         return socket.gethostbyname(hostname)
     except Exception as e:
         if log_errors:
             # in some cases this lookup is not possible, and we don't want to
             # make it an exception in logs
             log.exception('Could not retrieve server ip address: %s', e)
         return hostname
 
 
 def get_server_port(environ):
     return environ.get('SERVER_PORT')
 
 
 
 def get_user_agent(environ):
     return environ.get('HTTP_USER_AGENT')
 
 
 def vcs_operation_context(
         environ, repo_name, username, action, scm, check_locking=True,
         is_shadow_repo=False, check_branch_perms=False, detect_force_push=False):
     """
     Generate the context for a vcs operation, e.g. push or pull.
 
     This context is passed over the layers so that hooks triggered by the
     vcs operation know details like the user, the user's IP address etc.
 
     :param check_locking: Allows to switch of the computation of the locking
         data. This serves mainly the need of the simplevcs middleware to be
         able to disable this for certain operations.
 
     """
     # Tri-state value: False: unlock, None: nothing, True: lock
     make_lock = None
     locked_by = [None, None, None]
     is_anonymous = username == User.DEFAULT_USER
     user = User.get_by_username(username)
     if not is_anonymous and check_locking:
         log.debug('Checking locking on repository "%s"', repo_name)
         repo = Repository.get_by_repo_name(repo_name)
         make_lock, __, locked_by = repo.get_locking_state(
             action, user.user_id)
     user_id = user.user_id
     settings_model = VcsSettingsModel(repo=repo_name)
     ui_settings = settings_model.get_ui_settings()
 
     # NOTE(marcink): This should be also in sync with
     # rhodecode/apps/ssh_support/lib/backends/base.py:update_environment scm_data
     store = [x for x in ui_settings if x.key == '/']
     repo_store = ''
     if store:
         repo_store = store[0].value
 
     scm_data = {
         'ip': get_ip_addr(environ),
         'username': username,
         'user_id': user_id,
         'action': action,
         'repository': repo_name,
         'scm': scm,
         'config': rhodecode.CONFIG['__file__'],
         'repo_store': repo_store,
         'make_lock': make_lock,
         'locked_by': locked_by,
         'server_url': utils2.get_server_url(environ),
         'user_agent': get_user_agent(environ),
         'hooks': get_enabled_hook_classes(ui_settings),
         'is_shadow_repo': is_shadow_repo,
         'detect_force_push': detect_force_push,
         'check_branch_perms': check_branch_perms,
     }
     return scm_data
 
 
 class BasicAuth(AuthBasicAuthenticator):
 
     def __init__(self, realm, authfunc, registry, auth_http_code=None,
                  initial_call_detection=False, acl_repo_name=None, rc_realm=''):
         super().__init__(realm=realm, authfunc=authfunc)
         self.realm = realm
         self.rc_realm = rc_realm
         self.initial_call = initial_call_detection
         self.authfunc = authfunc
         self.registry = registry
         self.acl_repo_name = acl_repo_name
         self._rc_auth_http_code = auth_http_code
 
     def _get_response_from_code(self, http_code, fallback):
         try:
             return get_exception(safe_int(http_code))
         except Exception:
             log.exception('Failed to fetch response class for code %s, using fallback: %s', http_code, fallback)
             return fallback
 
     def get_rc_realm(self):
         return safe_str(self.rc_realm)
 
     def build_authentication(self):
         header = [('WWW-Authenticate', f'Basic realm="{self.realm}"')]
 
         # NOTE: the initial_Call detection seems to be not working/not needed witg latest Mercurial
         # investigate if we still need it.
         if self._rc_auth_http_code and not self.initial_call:
             # return alternative HTTP code if alternative http return code
             # is specified in RhodeCode config, but ONLY if it's not the
             # FIRST call
             custom_response_klass = self._get_response_from_code(self._rc_auth_http_code, fallback=HTTPUnauthorized)
             log.debug('Using custom response class: %s', custom_response_klass)
             return custom_response_klass(headers=header)
         return HTTPUnauthorized(headers=header)
 
     def authenticate(self, environ):
         authorization = paste.httpheaders.AUTHORIZATION(environ)
         if not authorization:
             return self.build_authentication()
         (auth_meth, auth_creds_b64) = authorization.split(' ', 1)
         if 'basic' != auth_meth.lower():
             return self.build_authentication()
 
         credentials = safe_str(base64.b64decode(auth_creds_b64.strip()))
         _parts = credentials.split(':', 1)
         if len(_parts) == 2:
             username, password = _parts
             auth_data = self.authfunc(
                     username, password, environ, VCS_TYPE,
                     registry=self.registry, acl_repo_name=self.acl_repo_name)
             if auth_data:
                 return {'username': username, 'auth_data': auth_data}
             if username and password:
                 # we mark that we actually executed authentication once, at
                 # that point we can use the alternative auth code
                 self.initial_call = False
 
         return self.build_authentication()
 
     __call__ = authenticate
 
 
 def calculate_version_hash(config):
     return sha1(
         config.get(b'beaker.session.secret', b'') + ascii_bytes(rhodecode.__version__)
     )[:8]
 
 
 def get_current_lang(request):
     return getattr(request, '_LOCALE_', request.locale_name)
 
 
 def attach_context_attributes(context, request, user_id=None, is_api=None):
     """
     Attach variables into template context called `c`.
     """
     config = request.registry.settings
 
     rc_config = SettingsModel().get_all_settings(cache=True, from_request=False)
     context.rc_config = rc_config
     context.rhodecode_version = rhodecode.__version__
     context.rhodecode_edition = config.get('rhodecode.edition')
     context.rhodecode_edition_id = config.get('rhodecode.edition_id')
     # unique secret + version does not leak the version but keep consistency
     context.rhodecode_version_hash = calculate_version_hash(config)
 
     # Default language set for the incoming request
     context.language = get_current_lang(request)
 
     # Visual options
     context.visual = AttributeDict({})
 
     # DB stored Visual Items
     context.visual.show_public_icon = str2bool(
         rc_config.get('rhodecode_show_public_icon'))
     context.visual.show_private_icon = str2bool(
         rc_config.get('rhodecode_show_private_icon'))
     context.visual.stylify_metatags = str2bool(
         rc_config.get('rhodecode_stylify_metatags'))
     context.visual.dashboard_items = safe_int(
         rc_config.get('rhodecode_dashboard_items', 100))
     context.visual.admin_grid_items = safe_int(
         rc_config.get('rhodecode_admin_grid_items', 100))
     context.visual.show_revision_number = str2bool(
         rc_config.get('rhodecode_show_revision_number', True))
     context.visual.show_sha_length = safe_int(
         rc_config.get('rhodecode_show_sha_length', 100))
     context.visual.repository_fields = str2bool(
         rc_config.get('rhodecode_repository_fields'))
     context.visual.show_version = str2bool(
         rc_config.get('rhodecode_show_version'))
     context.visual.use_gravatar = str2bool(
         rc_config.get('rhodecode_use_gravatar'))
     context.visual.gravatar_url = rc_config.get('rhodecode_gravatar_url')
     context.visual.default_renderer = rc_config.get(
         'rhodecode_markup_renderer', 'rst')
     context.visual.comment_types = ChangesetComment.COMMENT_TYPES
     context.visual.rhodecode_support_url = \
         rc_config.get('rhodecode_support_url') or h.route_url('rhodecode_support')
 
     context.visual.affected_files_cut_off = 60
 
     context.pre_code = rc_config.get('rhodecode_pre_code')
     context.post_code = rc_config.get('rhodecode_post_code')
     context.rhodecode_name = rc_config.get('rhodecode_title')
     context.default_encodings = aslist(config.get('default_encoding'), sep=',')
     # if we have specified default_encoding in the request, it has more
     # priority
     if request.GET.get('default_encoding'):
         context.default_encodings.insert(0, request.GET.get('default_encoding'))
     context.clone_uri_tmpl = rc_config.get('rhodecode_clone_uri_tmpl')
     context.clone_uri_id_tmpl = rc_config.get('rhodecode_clone_uri_id_tmpl')
     context.clone_uri_ssh_tmpl = rc_config.get('rhodecode_clone_uri_ssh_tmpl')
 
     # INI stored
     context.labs_active = str2bool(
         config.get('labs_settings_active', 'false'))
     context.ssh_enabled = str2bool(
         config.get('ssh.generate_authorized_keyfile', 'false'))
     context.ssh_key_generator_enabled = str2bool(
         config.get('ssh.enable_ui_key_generator', 'true'))
 
     context.visual.allow_custom_hooks_settings = str2bool(
         config.get('allow_custom_hooks_settings', True))
     context.debug_style = str2bool(config.get('debug_style', False))
 
     context.rhodecode_instanceid = config.get('instance_id')
 
     context.visual.cut_off_limit_diff = safe_int(
         config.get('cut_off_limit_diff'), default=0)
     context.visual.cut_off_limit_file = safe_int(
         config.get('cut_off_limit_file'), default=0)
 
     context.license = AttributeDict({})
     context.license.hide_license_info = str2bool(
         config.get('license.hide_license_info', False))
 
     # AppEnlight
     context.appenlight_enabled = config.get('appenlight', False)
     context.appenlight_api_public_key = config.get(
         'appenlight.api_public_key', '')
     context.appenlight_server_url = config.get('appenlight.server_url', '')
 
     diffmode = {
         "unified": "unified",
         "sideside": "sideside"
     }.get(request.GET.get('diffmode'))
 
     if is_api is not None:
         is_api = hasattr(request, 'rpc_user')
     session_attrs = {
         # defaults
         "clone_url_format": "http",
         "diffmode": "sideside",
         "license_fingerprint": request.session.get('license_fingerprint')
     }
 
     if not is_api:
         # don't access pyramid session for API calls
         if diffmode and diffmode != request.session.get('rc_user_session_attr.diffmode'):
             request.session['rc_user_session_attr.diffmode'] = diffmode
 
         # session settings per user
 
         for k, v in list(request.session.items()):
             pref = 'rc_user_session_attr.'
             if k and k.startswith(pref):
                 k = k[len(pref):]
                 session_attrs[k] = v
 
     context.user_session_attrs = session_attrs
 
     # JS template context
     context.template_context = {
         'repo_name': None,
         'repo_type': None,
         'repo_landing_commit': None,
         'rhodecode_user': {
             'username': None,
             'email': None,
             'notification_status': False
         },
         'session_attrs': session_attrs,
         'visual': {
             'default_renderer': None
         },
         'commit_data': {
             'commit_id': None
         },
         'pull_request_data': {'pull_request_id': None},
         'timeago': {
             'refresh_time': 120 * 1000,
             'cutoff_limit': 1000 * 60 * 60 * 24 * 7
         },
         'pyramid_dispatch': {
 
         },
         'extra': {'plugins': {}}
     }
     # END CONFIG VARS
     if is_api:
         csrf_token = None
     else:
         csrf_token = auth.get_csrf_token(session=request.session)
 
     context.csrf_token = csrf_token
     context.backends = list(rhodecode.BACKENDS.keys())
 
     unread_count = 0
     user_bookmark_list = []
     if user_id:
         unread_count = NotificationModel().get_unread_cnt_for_user(user_id)
         user_bookmark_list = UserBookmark.get_bookmarks_for_user(user_id)
     context.unread_notifications = unread_count
     context.bookmark_items = user_bookmark_list
 
     # web case
     if hasattr(request, 'user'):
         context.auth_user = request.user
         context.rhodecode_user = request.user
 
     # api case
     if hasattr(request, 'rpc_user'):
         context.auth_user = request.rpc_user
         context.rhodecode_user = request.rpc_user
 
     # attach the whole call context to the request
     request.set_call_context(context)
 
 
 def get_auth_user(request):
     environ = request.environ
     session = request.session
 
     ip_addr = get_ip_addr(environ)
 
     # make sure that we update permissions each time we call controller
     _auth_token = (
             # ?auth_token=XXX
             request.GET.get('auth_token', '')
             # ?api_key=XXX !LEGACY
             or request.GET.get('api_key', '')
             # or headers....
             or request.headers.get('X-Rc-Auth-Token', '')
     )
     if not _auth_token and request.matchdict:
         url_auth_token = request.matchdict.get('_auth_token')
         _auth_token = url_auth_token
         if _auth_token:
             log.debug('Using URL extracted auth token `...%s`', _auth_token[-4:])
 
     if _auth_token:
         # when using API_KEY we assume user exists, and
         # doesn't need auth based on cookies.
         auth_user = AuthUser(api_key=_auth_token, ip_addr=ip_addr)
         authenticated = False
     else:
         cookie_store = CookieStoreWrapper(session.get('rhodecode_user'))
         try:
             auth_user = AuthUser(user_id=cookie_store.get('user_id', None),
                                  ip_addr=ip_addr)
         except UserCreationError as e:
             h.flash(e, 'error')
             # container auth or other auth functions that create users
             # on the fly can throw this exception signaling that there's
             # issue with user creation, explanation should be provided
             # in Exception itself. We then create a simple blank
             # AuthUser
             auth_user = AuthUser(ip_addr=ip_addr)
 
         # in case someone changes a password for user it triggers session
         # flush and forces a re-login
         if password_changed(auth_user, session):
             session.invalidate()
             cookie_store = CookieStoreWrapper(session.get('rhodecode_user'))
             auth_user = AuthUser(ip_addr=ip_addr)
 
         authenticated = cookie_store.get('is_authenticated')
 
     if not auth_user.is_authenticated and auth_user.is_user_object:
         # user is not authenticated and not empty
         auth_user.set_authenticated(authenticated)
 
     return auth_user, _auth_token
 
 
 def h_filter(s):
     """
     Custom filter for Mako templates. Mako by standard uses `markupsafe.escape`
     we wrap this with additional functionality that converts None to empty
     strings
     """
     if s is None:
         return markupsafe.Markup()
     return markupsafe.escape(s)
 
 
 def add_events_routes(config):
     """
     Adds routing that can be used in events. Because some events are triggered
     outside of pyramid context, we need to bootstrap request with some
     routing registered
     """
 
     from rhodecode.apps._base import ADMIN_PREFIX
 
     config.add_route(name='home', pattern='/')
     config.add_route(name='main_page_repos_data', pattern='/_home_repos')
     config.add_route(name='main_page_repo_groups_data', pattern='/_home_repo_groups')
 
     config.add_route(name='login', pattern=ADMIN_PREFIX + '/login')
     config.add_route(name='logout', pattern=ADMIN_PREFIX + '/logout')
     config.add_route(name='repo_summary', pattern='/{repo_name}')
     config.add_route(name='repo_summary_explicit', pattern='/{repo_name}/summary')
     config.add_route(name='repo_group_home', pattern='/{repo_group_name}')
 
     config.add_route(name='pullrequest_show',
                      pattern='/{repo_name}/pull-request/{pull_request_id}')
     config.add_route(name='pull_requests_global',
                      pattern='/pull-request/{pull_request_id}')
 
     config.add_route(name='repo_commit',
                      pattern='/{repo_name}/changeset/{commit_id}')
     config.add_route(name='repo_files',
                      pattern='/{repo_name}/files/{commit_id}/{f_path}')
 
     config.add_route(name='hovercard_user',
                      pattern='/_hovercard/user/{user_id}')
 
     config.add_route(name='hovercard_user_group',
                      pattern='/_hovercard/user_group/{user_group_id}')
 
     config.add_route(name='hovercard_pull_request',
                      pattern='/_hovercard/pull_request/{pull_request_id}')
 
     config.add_route(name='hovercard_repo_commit',
                      pattern='/_hovercard/commit/{repo_name}/{commit_id}')
 
 
 def bootstrap_config(request, registry_name='RcTestRegistry'):
     from rhodecode.config.config_maker import sanitize_settings_and_apply_defaults
     import pyramid.testing
     registry = pyramid.testing.Registry(registry_name)
 
     global_config = {'__file__': ''}
 
     config = pyramid.testing.setUp(registry=registry, request=request)
     sanitize_settings_and_apply_defaults(global_config, config.registry.settings)
 
     # allow pyramid lookup in testing
     config.include('pyramid_mako')
     config.include('rhodecode.lib.rc_beaker')
     config.include('rhodecode.lib.rc_cache')
-    config.include('rhodecode.lib.rc_cache.archive_cache')
+    config.include('rhodecode.lib.archive_cache')
     add_events_routes(config)
 
     return config
 
 
 def bootstrap_request(**kwargs):
     """
     Returns a thin version of Request Object that is used in non-web context like testing/celery
     """
 
     import pyramid.testing
     from rhodecode.lib.request import ThinRequest as _ThinRequest
 
     class ThinRequest(_ThinRequest):
         application_url = kwargs.pop('application_url', 'http://example.com')
         host = kwargs.pop('host', 'example.com:80')
         domain = kwargs.pop('domain', 'example.com')
 
     class ThinSession(pyramid.testing.DummySession):
         def save(*arg, **kw):
             pass
 
     request = ThinRequest(**kwargs)
     request.session = ThinSession()
 
     return request

rhodecode/lib/helpers.py

@@ -1,2198 +1,2230 b''
 # Copyright (C) 2010-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 """
 Helper functions
 
 Consists of functions to typically be used within templates, but also
 available to Controllers. This module is available to both as 'h'.
 """
 import base64
 import collections
 
 import os
 import random
 import hashlib
 import io
 import textwrap
 import urllib.request
 import urllib.parse
 import urllib.error
 import math
 import logging
 import re
 import time
 import string
 import regex
 from collections import OrderedDict
 
 import pygments
 import itertools
 import fnmatch
 
 from datetime import datetime
 from functools import partial
 from pygments.formatters.html import HtmlFormatter
 from pygments.lexers import (
     get_lexer_by_name, get_lexer_for_filename, get_lexer_for_mimetype)
 
 from pyramid.threadlocal import get_current_request
 from tempita import looper
 from webhelpers2.html import literal, HTML, escape
 from webhelpers2.html._autolink import _auto_link_urls
 from webhelpers2.html.tools import (
     button_to, highlight, js_obfuscate, strip_links, strip_tags)
 
 from webhelpers2.text import (
     chop_at, collapse, convert_accented_entities,
     convert_misc_entities, lchop, plural, rchop, remove_formatting,
     replace_whitespace, urlify, truncate, wrap_paragraphs)
 from webhelpers2.date import time_ago_in_words
 
 from webhelpers2.html.tags import (
     _input, NotGiven, _make_safe_id_component as safeid,
     form as insecure_form,
     auto_discovery_link, checkbox, end_form, file,
     hidden, image, javascript_link, link_to, link_to_if, link_to_unless, ol,
     stylesheet_link, submit, text, password, textarea,
     ul, radio, Options)
 
 from webhelpers2.number import format_byte_size
 # python3.11 backport fixes for webhelpers2
 from rhodecode import ConfigGet
 from rhodecode.lib._vendor.webhelpers_backports import raw_select
 
 from rhodecode.lib.action_parser import action_parser
 from rhodecode.lib.html_filters import sanitize_html
 from rhodecode.lib.pagination import Page, RepoPage, SqlPage
 from rhodecode.lib import ext_json
 from rhodecode.lib.ext_json import json
 from rhodecode.lib.str_utils import safe_bytes, convert_special_chars, base64_to_str
 from rhodecode.lib.utils import repo_name_slug, get_custom_lexer
 from rhodecode.lib.str_utils import safe_str
 from rhodecode.lib.utils2 import (
     str2bool,
     get_commit_safe, datetime_to_time, time_to_datetime, time_to_utcdatetime,
     AttributeDict, safe_int, md5, md5_safe, get_host_info)
 from rhodecode.lib.markup_renderer import MarkupRenderer, relative_links
 from rhodecode.lib.vcs.exceptions import CommitDoesNotExistError
 from rhodecode.lib.vcs.backends.base import BaseChangeset, EmptyCommit
 from rhodecode.lib.vcs.conf.settings import ARCHIVE_SPECS
 from rhodecode.lib.index.search_utils import get_matching_line_offsets
 from rhodecode.config.conf import DATE_FORMAT, DATETIME_FORMAT
 from rhodecode.model.changeset_status import ChangesetStatusModel
 from rhodecode.model.db import Permission, User, Repository, UserApiKeys, FileStore
 from rhodecode.model.repo_group import RepoGroupModel
 from rhodecode.model.settings import IssueTrackerSettingsModel
 
 
 log = logging.getLogger(__name__)
 
 
 DEFAULT_USER = User.DEFAULT_USER
 DEFAULT_USER_EMAIL = User.DEFAULT_USER_EMAIL
 
 
 def asset(path, ver=None, **kwargs):
     """
     Helper to generate a static asset file path for rhodecode assets
 
     eg. h.asset('images/image.png', ver='3923')
 
     :param path: path of asset
     :param ver: optional version query param to append as ?ver=
     """
     request = get_current_request()
     query = {}
     query.update(kwargs)
     if ver:
         query = {'ver': ver}
     return request.static_path(
         f'rhodecode:public/{path}', _query=query)
 
 
 default_html_escape_table = {
     ord('&'): '&amp;',
     ord('<'): '&lt;',
     ord('>'): '&gt;',
     ord('"'): '&quot;',
     ord("'"): '&#39;',
 }
 
 
 def html_escape(text, html_escape_table=default_html_escape_table):
     """Produce entities within text."""
     return text.translate(html_escape_table)
 
 
 def str_json(*args, **kwargs):
     return ext_json.str_json(*args, **kwargs)
 
 
 def formatted_str_json(*args, **kwargs):
     return ext_json.formatted_str_json(*args, **kwargs)
 
 
 def chop_at_smart(s, sub, inclusive=False, suffix_if_chopped=None):
     """
     Truncate string ``s`` at the first occurrence of ``sub``.
 
     If ``inclusive`` is true, truncate just after ``sub`` rather than at it.
     """
     suffix_if_chopped = suffix_if_chopped or ''
     pos = s.find(sub)
     if pos == -1:
         return s
 
     if inclusive:
         pos += len(sub)
 
     chopped = s[:pos]
     left = s[pos:].strip()
 
     if left and suffix_if_chopped:
         chopped += suffix_if_chopped
 
     return chopped
 
 
 def shorter(text, size=20, prefix=False):
     postfix = '...'
     if len(text) > size:
         if prefix:
             # shorten in front
             return postfix + text[-(size - len(postfix)):]
         else:
             return text[:size - len(postfix)] + postfix
     return text
 
 
 def reset(name, value=None, id=NotGiven, type="reset", **attrs):
     """
     Reset button
     """
     return _input(type, name, value, id, attrs)
 
 
 def select(name, selected_values, options, id=NotGiven, **attrs):
 
     if isinstance(options, (list, tuple)):
         options_iter = options
         # Handle old value,label lists ... where value also can be value,label lists
         options = Options()
         for opt in options_iter:
             if isinstance(opt, tuple) and len(opt) == 2:
                 value, label = opt
             elif isinstance(opt, str):
                 value = label = opt
             else:
                 raise ValueError('invalid select option type %r' % type(opt))
 
             if isinstance(value, (list, tuple)):
                 option_group = options.add_optgroup(label)
                 for opt2 in value:
                     if isinstance(opt2, tuple) and len(opt2) == 2:
                         group_value, group_label = opt2
                     elif isinstance(opt2, str):
                         group_value = group_label = opt2
                     else:
                         raise ValueError('invalid select option type %r' % type(opt2))
 
                     option_group.add_option(group_label, group_value)
             else:
                 options.add_option(label, value)
 
     return raw_select(name, selected_values, options, id=id, **attrs)
 
 
 def branding(name, length=40):
     return truncate(name, length, indicator="")
 
 
 def FID(raw_id, path):
     """
     Creates a unique ID for filenode based on it's hash of path and commit
     it's safe to use in urls
 
     :param raw_id:
     :param path:
     """
 
     return 'c-%s-%s' % (short_id(raw_id), md5_safe(path)[:12])
 
 
 class _GetError(object):
     """Get error from form_errors, and represent it as span wrapped error
     message
 
     :param field_name: field to fetch errors for
     :param form_errors: form errors dict
     """
 
     def __call__(self, field_name, form_errors):
         tmpl = """<span class="error_msg">%s</span>"""
         if form_errors and field_name in form_errors:
             return literal(tmpl % form_errors.get(field_name))
 
 
 get_error = _GetError()
 
 
 class _ToolTip(object):
 
     def __call__(self, tooltip_title, trim_at=50):
         """
         Special function just to wrap our text into nice formatted
         autowrapped text
 
         :param tooltip_title:
         """
         tooltip_title = escape(tooltip_title)
         tooltip_title = tooltip_title.replace('<', '&lt;').replace('>', '&gt;')
         return tooltip_title
 
 
 tooltip = _ToolTip()
 
 files_icon = '<i class="file-breadcrumb-copy tooltip icon-clipboard clipboard-action" data-clipboard-text="{}" title="Copy file path"></i>'
 
 
 def files_breadcrumbs(repo_name, repo_type, commit_id, file_path, landing_ref_name=None, at_ref=None,
                       limit_items=False, linkify_last_item=False, hide_last_item=False,
                       copy_path_icon=True):
 
     if at_ref:
         route_qry = {'at': at_ref}
         default_landing_ref = at_ref or landing_ref_name or commit_id
     else:
         route_qry = None
         default_landing_ref = commit_id
 
     # first segment is a `HOME` link to repo files root location
     root_name = literal('<i class="icon-home"></i>')
 
     url_segments = [
         link_to(
             root_name,
             repo_files_by_ref_url(
                 repo_name,
                 repo_type,
                 f_path=None,  # None here is a special case for SVN repos,
                               # that won't prefix with a ref
                 ref_name=default_landing_ref,
                 commit_id=commit_id,
                 query=route_qry
             )
         )]
 
     path_segments = file_path.split('/')
     last_cnt = len(path_segments) - 1
     for cnt, segment in enumerate(path_segments):
         if not segment:
             continue
         segment_html = escape(segment)
 
         last_item = cnt == last_cnt
 
         if last_item and hide_last_item:
             # iterate over and hide last element
             continue
 
         if last_item and linkify_last_item is False:
             # plain version
             url_segments.append(segment_html)
         else:
             url_segments.append(
                 link_to(
                     segment_html,
                     repo_files_by_ref_url(
                         repo_name,
                         repo_type,
                         f_path='/'.join(path_segments[:cnt + 1]),
                         ref_name=default_landing_ref,
                         commit_id=commit_id,
                         query=route_qry
                     ),
                 ))
 
     limited_url_segments = url_segments[:1] + ['...'] + url_segments[-5:]
     if limit_items and len(limited_url_segments) < len(url_segments):
         url_segments = limited_url_segments
 
     full_path = file_path
     if copy_path_icon:
         icon = files_icon.format(escape(full_path))
     else:
         icon = ''
 
     if file_path == '':
         return root_name
     else:
         return literal(' / '.join(url_segments) + icon)
 
 
 def files_url_data(request):
     matchdict = request.matchdict
 
     if 'f_path' not in matchdict:
         matchdict['f_path'] = ''
     else:
         matchdict['f_path'] = urllib.parse.quote(safe_str(matchdict['f_path']))
     if 'commit_id' not in matchdict:
         matchdict['commit_id'] = 'tip'
 
     return ext_json.str_json(matchdict)
 
 
 def repo_files_by_ref_url(db_repo_name, db_repo_type, f_path, ref_name, commit_id, query=None, ):
     _is_svn = is_svn(db_repo_type)
     final_f_path = f_path
 
     if _is_svn:
         """
         For SVN the ref_name cannot be used as a commit_id, it needs to be prefixed with
         actually commit_id followed by the ref_name. This should be done only in case
         This is a initial landing url, without additional paths.
 
         like: /1000/tags/1.0.0/?at=tags/1.0.0
         """
 
         if ref_name and ref_name != 'tip':
             # NOTE(marcink): for svn the ref_name is actually the stored path, so we prefix it
             # for SVN we only do this magic prefix if it's root, .eg landing revision
             # of files link. If we are in the tree we don't need this since we traverse the url
             # that has everything stored
             if f_path in ['', '/']:
                 final_f_path = '/'.join([ref_name, f_path])
 
         # SVN always needs a commit_id explicitly, without a named REF
         default_commit_id = commit_id
     else:
         """
         For git and mercurial we construct a new URL using the names instead of commit_id
         like: /master/some_path?at=master
         """
         # We currently do not support branches with slashes
         if '/' in ref_name:
             default_commit_id = commit_id
         else:
             default_commit_id = ref_name
 
     # sometimes we pass f_path as None, to indicate explicit no prefix,
     # we translate it to string to not have None
     final_f_path = final_f_path or ''
 
     files_url = route_path(
         'repo_files',
         repo_name=db_repo_name,
         commit_id=default_commit_id,
         f_path=final_f_path,
         _query=query
     )
     return files_url
 
 
 def code_highlight(code, lexer, formatter, use_hl_filter=False):
     """
     Lex ``code`` with ``lexer`` and format it with the formatter ``formatter``.
 
     If ``outfile`` is given and a valid file object (an object
     with a ``write`` method), the result will be written to it, otherwise
     it is returned as a string.
     """
     if use_hl_filter:
         # add HL filter
         from rhodecode.lib.index import search_utils
         lexer.add_filter(search_utils.ElasticSearchHLFilter())
     return pygments.format(pygments.lex(code, lexer), formatter)
 
 
 class CodeHtmlFormatter(HtmlFormatter):
     """
     My code Html Formatter for source codes
     """
 
     def wrap(self, source):
         return self._wrap_div(self._wrap_pre(self._wrap_code(source)))
 
     def _wrap_code(self, source):
         for cnt, it in enumerate(source):
             i, t = it
             t = f'<div id="L{cnt+1}">{t}</div>'
             yield i, t
 
     def _wrap_tablelinenos(self, inner):
         dummyoutfile = io.StringIO()
         lncount = 0
         for t, line in inner:
             if t:
                 lncount += 1
             dummyoutfile.write(line)
 
         fl = self.linenostart
         mw = len(str(lncount + fl - 1))
         sp = self.linenospecial
         st = self.linenostep
         la = self.lineanchors
         aln = self.anchorlinenos
         nocls = self.noclasses
         if sp:
             lines = []
 
             for i in range(fl, fl + lncount):
                 if i % st == 0:
                     if i % sp == 0:
                         if aln:
                             lines.append('<a href="#%s%d" class="special">%*d</a>' %
                                          (la, i, mw, i))
                         else:
                             lines.append('<span class="special">%*d</span>' % (mw, i))
                     else:
                         if aln:
                             lines.append('<a href="#%s%d">%*d</a>' % (la, i, mw, i))
                         else:
                             lines.append('%*d' % (mw, i))
                 else:
                     lines.append('')
             ls = '\n'.join(lines)
         else:
             lines = []
             for i in range(fl, fl + lncount):
                 if i % st == 0:
                     if aln:
                         lines.append('<a href="#%s%d">%*d</a>' % (la, i, mw, i))
                     else:
                         lines.append('%*d' % (mw, i))
                 else:
                     lines.append('')
             ls = '\n'.join(lines)
 
         # in case you wonder about the seemingly redundant <div> here: since the
         # content in the other cell also is wrapped in a div, some browsers in
         # some configurations seem to mess up the formatting...
         if nocls:
             yield 0, ('<table class="%stable">' % self.cssclass +
                       '<tr><td><div class="linenodiv" '
                       'style="background-color: #f0f0f0; padding-right: 10px">'
                       '<pre style="line-height: 125%">' +
                       ls + '</pre></div></td><td id="hlcode" class="code">')
         else:
             yield 0, ('<table class="%stable">' % self.cssclass +
                       '<tr><td class="linenos"><div class="linenodiv"><pre>' +
                       ls + '</pre></div></td><td id="hlcode" class="code">')
         yield 0, dummyoutfile.getvalue()
         yield 0, '</td></tr></table>'
 
 
 class SearchContentCodeHtmlFormatter(CodeHtmlFormatter):
     def __init__(self, **kw):
         # only show these line numbers if set
         self.only_lines = kw.pop('only_line_numbers', [])
         self.query_terms = kw.pop('query_terms', [])
         self.max_lines = kw.pop('max_lines', 5)
         self.line_context = kw.pop('line_context', 3)
         self.url = kw.pop('url', None)
 
         super(CodeHtmlFormatter, self).__init__(**kw)
 
     def _wrap_code(self, source):
         for cnt, it in enumerate(source):
             i, t = it
             t = '<pre>%s</pre>' % t
             yield i, t
 
     def _wrap_tablelinenos(self, inner):
         yield 0, '<table class="code-highlight %stable">' % self.cssclass
 
         last_shown_line_number = 0
         current_line_number = 1
 
         for t, line in inner:
             if not t:
                 yield t, line
                 continue
 
             if current_line_number in self.only_lines:
                 if last_shown_line_number + 1 != current_line_number:
                     yield 0, '<tr>'
                     yield 0, '<td class="line">...</td>'
                     yield 0, '<td id="hlcode" class="code"></td>'
                     yield 0, '</tr>'
 
                 yield 0, '<tr>'
                 if self.url:
                     yield 0, '<td class="line"><a href="%s#L%i">%i</a></td>' % (
                         self.url, current_line_number, current_line_number)
                 else:
                     yield 0, '<td class="line"><a href="">%i</a></td>' % (
                         current_line_number)
                 yield 0, '<td id="hlcode" class="code">' + line + '</td>'
                 yield 0, '</tr>'
 
                 last_shown_line_number = current_line_number
 
             current_line_number += 1
 
         yield 0, '</table>'
 
 
 def hsv_to_rgb(h, s, v):
     """ Convert hsv color values to rgb """
 
     if s == 0.0:
         return v, v, v
     i = int(h * 6.0)  # XXX assume int() truncates!
     f = (h * 6.0) - i
     p = v * (1.0 - s)
     q = v * (1.0 - s * f)
     t = v * (1.0 - s * (1.0 - f))
     i = i % 6
     if i == 0:
         return v, t, p
     if i == 1:
         return q, v, p
     if i == 2:
         return p, v, t
     if i == 3:
         return p, q, v
     if i == 4:
         return t, p, v
     if i == 5:
         return v, p, q
 
 
 def unique_color_generator(n=10000, saturation=0.10, lightness=0.95):
     """
     Generator for getting n of evenly distributed colors using
     hsv color and golden ratio. It always return same order of colors
 
     :param n: number of colors to generate
     :param saturation: saturation of returned colors
     :param lightness: lightness of returned colors
     :returns: RGB tuple
     """
 
     golden_ratio = 0.618033988749895
     h = 0.22717784590367374
 
     for _ in range(n):
         h += golden_ratio
         h %= 1
         HSV_tuple = [h, saturation, lightness]
         RGB_tuple = hsv_to_rgb(*HSV_tuple)
         yield [str(int(x * 256)) for x in RGB_tuple]
 
 
 def color_hasher(n=10000, saturation=0.10, lightness=0.95):
     """
     Returns a function which when called with an argument returns a unique
     color for that argument, eg.
 
     :param n: number of colors to generate
     :param saturation: saturation of returned colors
     :param lightness: lightness of returned colors
     :returns: css RGB string
 
     >>> color_hash = color_hasher()
     >>> color_hash('hello')
     'rgb(34, 12, 59)'
     >>> color_hash('hello')
     'rgb(34, 12, 59)'
     >>> color_hash('other')
     'rgb(90, 224, 159)'
     """
 
     color_dict = {}
     cgenerator = unique_color_generator(
         saturation=saturation, lightness=lightness)
 
     def get_color_string(thing):
         if thing in color_dict:
             col = color_dict[thing]
         else:
             col = color_dict[thing] = next(cgenerator)
         return "rgb(%s)" % (', '.join(col))
 
     return get_color_string
 
 
 def get_lexer_safe(mimetype=None, filepath=None):
     """
     Tries to return a relevant pygments lexer using mimetype/filepath name,
     defaulting to plain text if none could be found
     """
     lexer = None
     try:
         if mimetype:
             lexer = get_lexer_for_mimetype(mimetype)
         if not lexer:
             lexer = get_lexer_for_filename(filepath)
     except pygments.util.ClassNotFound:
         pass
 
     if not lexer:
         lexer = get_lexer_by_name('text')
 
     return lexer
 
 
 def get_lexer_for_filenode(filenode):
     lexer = get_custom_lexer(filenode.extension) or filenode.lexer
     return lexer
 
 
 def pygmentize(filenode, **kwargs):
     """
     pygmentize function using pygments
 
     :param filenode:
     """
     lexer = get_lexer_for_filenode(filenode)
     return literal(code_highlight(filenode.content, lexer,
                                   CodeHtmlFormatter(**kwargs)))
 
 
 def is_following_repo(repo_name, user_id):
     from rhodecode.model.scm import ScmModel
     return ScmModel().is_following_repo(repo_name, user_id)
 
 
 class _Message(object):
     """A message returned by ``Flash.pop_messages()``.
 
     Converting the message to a string returns the message text. Instances
     also have the following attributes:
 
     * ``message``: the message text.
     * ``category``: the category specified when the message was created.
     """
 
     def __init__(self, category, message, sub_data=None):
         self.category = category
         self.message = message
         self.sub_data = sub_data or {}
 
     def __str__(self):
         return self.message
 
     __unicode__ = __str__
 
     def __html__(self):
         return escape(safe_str(self.message))
 
 
 class Flash(object):
     # List of allowed categories.  If None, allow any category.
     categories = ["warning", "notice", "error", "success"]
 
     # Default category if none is specified.
     default_category = "notice"
 
     def __init__(self, session_key="flash", categories=None,
                  default_category=None):
         """
         Instantiate a ``Flash`` object.
 
         ``session_key`` is the key to save the messages under in the user's
         session.
 
         ``categories`` is an optional list which overrides the default list
         of categories.
 
         ``default_category`` overrides the default category used for messages
         when none is specified.
         """
         self.session_key = session_key
         if categories is not None:
             self.categories = categories
         if default_category is not None:
             self.default_category = default_category
         if self.categories and self.default_category not in self.categories:
             raise ValueError(
                 "unrecognized default category %r" % (self.default_category,))
 
     def pop_messages(self, session=None, request=None):
         """
         Return all accumulated messages and delete them from the session.
 
         The return value is a list of ``Message`` objects.
         """
         messages = []
 
         if not session:
             if not request:
                 request = get_current_request()
             session = request.session
 
         # Pop the 'old' pylons flash messages. They are tuples of the form
         # (category, message)
         for cat, msg in session.pop(self.session_key, []):
             messages.append(_Message(cat, msg))
 
         # Pop the 'new' pyramid flash messages for each category as list
         # of strings.
         for cat in self.categories:
             for msg in session.pop_flash(queue=cat):
                 sub_data = {}
                 if hasattr(msg, 'rsplit'):
                     flash_data = msg.rsplit('|DELIM|', 1)
                     org_message = flash_data[0]
                     if len(flash_data) > 1:
                         sub_data = json.loads(flash_data[1])
                 else:
                     org_message = msg
 
                 messages.append(_Message(cat, org_message, sub_data=sub_data))
 
         # Map messages from the default queue to the 'notice' category.
         for msg in session.pop_flash():
             messages.append(_Message('notice', msg))
 
         session.save()
         return messages
 
     def json_alerts(self, session=None, request=None):
         payloads = []
         messages = flash.pop_messages(session=session, request=request) or []
         for message in messages:
             payloads.append({
                 'message': {
                     'message': '{}'.format(message.message),
                     'level': message.category,
                     'force': True,
                     'subdata': message.sub_data
                 }
             })
         return safe_str(json.dumps(payloads))
 
     def __call__(self, message, category=None, ignore_duplicate=True,
                  session=None, request=None):
 
         if not session:
             if not request:
                 request = get_current_request()
             session = request.session
 
         session.flash(
             message, queue=category, allow_duplicate=not ignore_duplicate)
 
 
 flash = Flash()
 
 #==============================================================================
 # SCM FILTERS available via h.
 #==============================================================================
 from rhodecode.lib.vcs.utils import author_name, author_email
 from rhodecode.lib.utils2 import age, age_from_seconds
 from rhodecode.model.db import User, ChangesetStatus
 
 
 email = author_email
 
 
 def capitalize(raw_text):
     return raw_text.capitalize()
 
 
 def short_id(long_id):
     return long_id[:12]
 
 
 def hide_credentials(url):
     from rhodecode.lib.utils2 import credentials_filter
     return credentials_filter(url)
 
 import zoneinfo
 import tzlocal
 local_timezone = tzlocal.get_localzone()
 
 
 def get_timezone(datetime_iso, time_is_local=False):
     tzinfo = '+00:00'
 
     # detect if we have a timezone info, otherwise, add it
     if time_is_local and isinstance(datetime_iso, datetime) and not datetime_iso.tzinfo:
         force_timezone = os.environ.get('RC_TIMEZONE', '')
         if force_timezone:
             force_timezone = zoneinfo.ZoneInfo(force_timezone)
         timezone = force_timezone or local_timezone
 
         offset = datetime_iso.replace(tzinfo=timezone).strftime('%z')
         tzinfo = '{}:{}'.format(offset[:-2], offset[-2:])
     return tzinfo
 
 
 def age_component(datetime_iso, value=None, time_is_local=False, tooltip=True):
     title = value or format_date(datetime_iso)
     tzinfo = get_timezone(datetime_iso, time_is_local=time_is_local)
 
     return literal(
         '<time class="timeago {cls}" title="{tt_title}" datetime="{dt}{tzinfo}">{title}</time>'.format(
             cls='tooltip' if tooltip else '',
             tt_title=('{title}{tzinfo}'.format(title=title, tzinfo=tzinfo)) if tooltip else '',
             title=title, dt=datetime_iso, tzinfo=tzinfo
         ))
 
 
 def _shorten_commit_id(commit_id, commit_len=None):
     if commit_len is None:
         request = get_current_request()
         commit_len = request.call_context.visual.show_sha_length
     return commit_id[:commit_len]
 
 
 def show_id(commit, show_idx=None, commit_len=None):
     """
     Configurable function that shows ID
     by default it's r123:fffeeefffeee
 
     :param commit: commit instance
     """
     if show_idx is None:
         request = get_current_request()
         show_idx = request.call_context.visual.show_revision_number
 
     raw_id = _shorten_commit_id(commit.raw_id, commit_len=commit_len)
     if show_idx:
         return 'r%s:%s' % (commit.idx, raw_id)
     else:
         return '%s' % (raw_id, )
 
 
 def format_date(date):
     """
     use a standardized formatting for dates used in RhodeCode
 
     :param date: date/datetime object
     :return: formatted date
     """
 
     if date:
         _fmt = "%a, %d %b %Y %H:%M:%S"
         return safe_str(date.strftime(_fmt))
 
     return ""
 
 
 class _RepoChecker(object):
 
     def __init__(self, backend_alias):
         self._backend_alias = backend_alias
 
     def __call__(self, repository):
         if hasattr(repository, 'alias'):
             _type = repository.alias
         elif hasattr(repository, 'repo_type'):
             _type = repository.repo_type
         else:
             _type = repository
         return _type == self._backend_alias
 
 
 is_git = _RepoChecker('git')
 is_hg = _RepoChecker('hg')
 is_svn = _RepoChecker('svn')
 
 
 def get_repo_type_by_name(repo_name):
     repo = Repository.get_by_repo_name(repo_name)
     if repo:
         return repo.repo_type
 
 
 def is_svn_without_proxy(repository):
     if is_svn(repository):
         return not ConfigGet().get_bool('vcs.svn.proxy.enabled')
     return False
 
 
 def discover_user(author):
     """
     Tries to discover RhodeCode User based on the author string. Author string
     is typically `FirstName LastName <[e-mail]>`
     """
 
     # if author is already an instance use it for extraction
     if isinstance(author, User):
         return author
 
     # Valid email in the attribute passed, see if they're in the system
     _email = author_email(author)
     if _email != '':
         user = User.get_by_email(_email, case_insensitive=True, cache=True)
         if user is not None:
             return user
 
     # Maybe it's a username, we try to extract it and fetch by username ?
     _author = author_name(author)
     user = User.get_by_username(_author, case_insensitive=True, cache=True)
     if user is not None:
         return user
 
     return None
 
 
 def email_or_none(author):
     # extract email from the commit string
     _email = author_email(author)
 
     # If we have an email, use it, otherwise
     # see if it contains a username we can get an email from
     if _email != '':
         return _email
     else:
         user = User.get_by_username(
             author_name(author), case_insensitive=True, cache=True)
 
     if user is not None:
             return user.email
 
     # No valid email, not a valid user in the system, none!
     return None
 
 
 def link_to_user(author, length=0, **kwargs):
     user = discover_user(author)
     # user can be None, but if we have it already it means we can re-use it
     # in the person() function, so we save 1 intensive-query
     if user:
         author = user
 
     display_person = person(author, 'username_or_name_or_email')
     if length:
         display_person = shorter(display_person, length)
 
     if user and user.username != user.DEFAULT_USER:
         return link_to(
             escape(display_person),
             route_path('user_profile', username=user.username),
             **kwargs)
     else:
         return escape(display_person)
 
 
 def link_to_group(users_group_name, **kwargs):
     return link_to(
         escape(users_group_name),
         route_path('user_group_profile', user_group_name=users_group_name),
         **kwargs)
 
 
 def person(author, show_attr="username_and_name"):
     user = discover_user(author)
     if user:
         return getattr(user, show_attr)
     else:
         _author = author_name(author)
         _email = email(author)
         return _author or _email
 
 
 def author_string(email):
     if email:
         user = User.get_by_email(email, case_insensitive=True, cache=True)
         if user:
             if user.first_name or user.last_name:
                 return '%s %s &lt;%s&gt;' % (
                     user.first_name, user.last_name, email)
             else:
                 return email
         else:
             return email
     else:
         return None
 
 
 def person_by_id(id_, show_attr="username_and_name"):
     # attr to return from fetched user
     def person_getter(usr):
         return getattr(usr, show_attr)
 
     #maybe it's an ID ?
     if str(id_).isdigit() or isinstance(id_, int):
         id_ = int(id_)
         user = User.get(id_)
         if user is not None:
             return person_getter(user)
     return id_
 
 
 def gravatar_with_user(request, author, show_disabled=False, tooltip=False):
     _render = request.get_partial_renderer('rhodecode:templates/base/base.mako')
     return _render('gravatar_with_user', author, show_disabled=show_disabled, tooltip=tooltip)
 
 
 tags_patterns = OrderedDict(
     (
         (
             "lang",
             (
                 re.compile(r"\[(lang|language)\ \=\&gt;\ *([a-zA-Z\-\/\#\+\.]*)\]"),
                 '<div class="metatag" tag="lang">\\2</div>',
             ),
         ),
         (
             "see",
             (
                 re.compile(r"\[see\ \=\&gt;\ *([a-zA-Z0-9\/\=\?\&amp;\ \:\/\.\-]*)\]"),
                 '<div class="metatag" tag="see">see: \\1 </div>',
             ),
         ),
         (
             "url",
             (
                 re.compile(
                     r"\[url\ \=\&gt;\ \[([a-zA-Z0-9\ \.\-\_]+)\]\((http://|https://|/)(.*?)\)\]"
                 ),
                 '<div class="metatag" tag="url"> <a href="\\2\\3">\\1</a> </div>',
             ),
         ),
         (
             "license",
             (
                 re.compile(
                     r"\[license\ \=\&gt;\ *([a-zA-Z0-9\/\=\?\&amp;\ \:\/\.\-]*)\]"
                 ),
                 # don't make it a raw string here...
                 '<div class="metatag" tag="license"><a href="http:\/\/www.opensource.org/licenses/\\1">\\1</a></div>',
             ),
         ),
         (
             "ref",
             (
                 re.compile(
                     r"\[(requires|recommends|conflicts|base)\ \=\&gt;\ *([a-zA-Z0-9\-\/]*)\]"
                 ),
                 '<div class="metatag" tag="ref \\1">\\1: <a href="/\\2">\\2</a></div>',
             ),
         ),
         (
             "state",
             (
                 re.compile(r"\[(stable|featured|stale|dead|dev|deprecated)\]"),
                 '<div class="metatag" tag="state \\1">\\1</div>',
             ),
         ),
         # label in grey
         (
             "label",
             (re.compile(r"\[([a-z]+)\]"), '<div class="metatag" tag="label">\\1</div>'),
         ),
         # generic catch all in grey
         (
             "generic",
             (
                 re.compile(r"\[([a-zA-Z0-9\.\-\_]+)\]"),
                 '<div class="metatag" tag="generic">\\1</div>',
             ),
         ),
     )
 )
 
 
 def extract_metatags(value):
     """
     Extract supported meta-tags from given text value
     """
     tags = []
     if not value:
         return tags, ''
 
     for key, val in list(tags_patterns.items()):
         pat, replace_html = val
         tags.extend([(key, x.group()) for x in pat.finditer(value)])
         value = pat.sub('', value)
 
     return tags, value
 
 
 def style_metatag(tag_type, value):
     """
     converts tags from value into html equivalent
     """
     if not value:
         return ''
 
     html_value = value
     tag_data = tags_patterns.get(tag_type)
     if tag_data:
         pat, replace_html = tag_data
         # convert to plain `str` instead of a markup tag to be used in
         # regex expressions. safe_str doesn't work here
         html_value = pat.sub(replace_html, value)
 
     return html_value
 
 
 def bool2icon(value, show_at_false=True):
     """
     Returns boolean value of a given value, represented as html element with
     classes that will represent icons
 
     :param value: given value to convert to html node
     """
 
     if value:  # does bool conversion
         return HTML.tag('i', class_="icon-true", title='True')
     else:  # not true as bool
         if show_at_false:
             return HTML.tag('i', class_="icon-false", title='False')
         return HTML.tag('i')
 
 
 def b64(inp):
     return base64.b64encode(safe_bytes(inp))
 
 #==============================================================================
 # PERMS
 #==============================================================================
 from rhodecode.lib.auth import (
     HasPermissionAny, HasPermissionAll,
     HasRepoPermissionAny, HasRepoPermissionAll, HasRepoGroupPermissionAll,
     HasRepoGroupPermissionAny, HasRepoPermissionAnyApi, get_csrf_token,
     csrf_token_key, AuthUser)
 
 
 #==============================================================================
 # GRAVATAR URL
 #==============================================================================
 class InitialsGravatar(object):
     def __init__(self, email_address, first_name, last_name, size=30,
                  background=None, text_color='#fff'):
         self.size = size
         self.first_name = first_name
         self.last_name = last_name
         self.email_address = email_address
         self.background = background or self.str2color(email_address)
         self.text_color = text_color
 
     def get_color_bank(self):
         """
         returns a predefined list of colors that gravatars can use.
         Those are randomized distinct colors that guarantee readability and
         uniqueness.
 
         generated with: http://phrogz.net/css/distinct-colors.html
         """
         return [
             '#bf3030', '#a67f53', '#00ff00', '#5989b3', '#392040', '#d90000',
             '#402910', '#204020', '#79baf2', '#a700b3', '#bf6060', '#7f5320',
             '#008000', '#003059', '#ee00ff', '#ff0000', '#8c4b00', '#007300',
             '#005fb3', '#de73e6', '#ff4040', '#ffaa00', '#3df255', '#203140',
             '#47004d', '#591616', '#664400', '#59b365', '#0d2133', '#83008c',
             '#592d2d', '#bf9f60', '#73e682', '#1d3f73', '#73006b', '#402020',
             '#b2862d', '#397341', '#597db3', '#e600d6', '#a60000', '#736039',
             '#00b318', '#79aaf2', '#330d30', '#ff8080', '#403010', '#16591f',
             '#002459', '#8c4688', '#e50000', '#ffbf40', '#00732e', '#102340',
             '#bf60ac', '#8c4646', '#cc8800', '#00a642', '#1d3473', '#b32d98',
             '#660e00', '#ffd580', '#80ffb2', '#7391e6', '#733967', '#d97b6c',
             '#8c5e00', '#59b389', '#3967e6', '#590047', '#73281d', '#665200',
             '#00e67a', '#2d50b3', '#8c2377', '#734139', '#b2982d', '#16593a',
             '#001859', '#ff00aa', '#a65e53', '#ffcc00', '#0d3321', '#2d3959',
             '#731d56', '#401610', '#4c3d00', '#468c6c', '#002ca6', '#d936a3',
             '#d94c36', '#403920', '#36d9a3', '#0d1733', '#592d4a', '#993626',
             '#cca300', '#00734d', '#46598c', '#8c005e', '#7f1100', '#8c7000',
             '#00a66f', '#7382e6', '#b32d74', '#d9896c', '#ffe680', '#1d7362',
             '#364cd9', '#73003d', '#d93a00', '#998a4d', '#59b3a1', '#5965b3',
             '#e5007a', '#73341d', '#665f00', '#00b38f', '#0018b3', '#59163a',
             '#b2502d', '#bfb960', '#00ffcc', '#23318c', '#a6537f', '#734939',
             '#b2a700', '#104036', '#3d3df2', '#402031', '#e56739', '#736f39',
             '#79f2ea', '#000059', '#401029', '#4c1400', '#ffee00', '#005953',
             '#101040', '#990052', '#402820', '#403d10', '#00ffee', '#0000d9',
             '#ff80c4', '#a66953', '#eeff00', '#00ccbe', '#8080ff', '#e673a1',
             '#a62c00', '#474d00', '#1a3331', '#46468c', '#733950', '#662900',
             '#858c23', '#238c85', '#0f0073', '#b20047', '#d9986c', '#becc00',
             '#396f73', '#281d73', '#ff0066', '#ff6600', '#dee673', '#59adb3',
             '#6559b3', '#590024', '#b2622d', '#98b32d', '#36ced9', '#332d59',
             '#40001a', '#733f1d', '#526600', '#005359', '#242040', '#bf6079',
             '#735039', '#cef23d', '#007780', '#5630bf', '#66001b', '#b24700',
             '#acbf60', '#1d6273', '#25008c', '#731d34', '#a67453', '#50592d',
             '#00ccff', '#6600ff', '#ff0044', '#4c1f00', '#8a994d', '#79daf2',
             '#a173e6', '#d93662', '#402310', '#aaff00', '#2d98b3', '#8c40ff',
             '#592d39', '#ff8c40', '#354020', '#103640', '#1a0040', '#331a20',
             '#331400', '#334d00', '#1d5673', '#583973', '#7f0022', '#4c3626',
             '#88cc00', '#36a3d9', '#3d0073', '#d9364c', '#33241a', '#698c23',
             '#5995b3', '#300059', '#e57382', '#7f3300', '#366600', '#00aaff',
             '#3a1659', '#733941', '#663600', '#74b32d', '#003c59', '#7f53a6',
             '#73000f', '#ff8800', '#baf279', '#79caf2', '#291040', '#a6293a',
             '#b2742d', '#587339', '#0077b3', '#632699', '#400009', '#d9a66c',
             '#294010', '#2d4a59', '#aa00ff', '#4c131b', '#b25f00', '#5ce600',
             '#267399', '#a336d9', '#990014', '#664e33', '#86bf60', '#0088ff',
             '#7700b3', '#593a16', '#073300', '#1d4b73', '#ac60bf', '#e59539',
             '#4f8c46', '#368dd9', '#5c0073'
         ]
 
     def rgb_to_hex_color(self, rgb_tuple):
         """
         Converts an rgb_tuple passed to an hex color.
 
         :param rgb_tuple: tuple with 3 ints represents rgb color space
         """
         return '#' + ("".join(map(chr, rgb_tuple)).encode('hex'))
 
     def email_to_int_list(self, email_str):
         """
         Get every byte of the hex digest value of email and turn it to integer.
         It's going to be always between 0-255
         """
         digest = md5_safe(email_str.lower())
         return [int(digest[i * 2:i * 2 + 2], 16) for i in range(16)]
 
     def pick_color_bank_index(self, email_str, color_bank):
         return self.email_to_int_list(email_str)[0] % len(color_bank)
 
     def str2color(self, email_str):
         """
         Tries to map in a stable algorithm an email to color
 
         :param email_str:
         """
         color_bank = self.get_color_bank()
         # pick position (module it's length so we always find it in the
         # bank even if it's smaller than 256 values
         pos = self.pick_color_bank_index(email_str, color_bank)
         return color_bank[pos]
 
     def normalize_email(self, email_address):
         # default host used to fill in the fake/missing email
         default_host = 'localhost'
 
         if not email_address:
             email_address = f'{User.DEFAULT_USER}@{default_host}'
 
         email_address = safe_str(email_address)
 
         if '@' not in email_address:
             email_address = f'{email_address}@{default_host}'
 
         if email_address.endswith('@'):
             email_address = f'{email_address}{default_host}'
 
         email_address = convert_special_chars(email_address)
 
         return email_address
 
     def get_initials(self):
         """
         Returns 2 letter initials calculated based on the input.
         The algorithm picks first given email address, and takes first letter
         of part before @, and then the first letter of server name. In case
         the part before @ is in a format of `somestring.somestring2` it replaces
         the server letter with first letter of somestring2
 
         In case function was initialized with both first and lastname, this
         overrides the extraction from email by first letter of the first and
         last name. We add special logic to that functionality, In case Full name
         is compound, like Guido Von Rossum, we use last part of the last name
         (Von Rossum) picking `R`.
 
         Function also normalizes the non-ascii characters to they ascii
         representation, eg Ą => A
         """
         # replace non-ascii to ascii
         first_name = convert_special_chars(self.first_name)
         last_name = convert_special_chars(self.last_name)
         # multi word last names, Guido Von Rossum, we take the last part only
         last_name = last_name.split(' ', 1)[-1]
 
         # do NFKD encoding, and also make sure email has proper format
         email_address = self.normalize_email(self.email_address)
 
         # first push the email initials
         prefix, server = email_address.split('@', 1)
 
         # check if prefix is maybe a 'first_name.last_name' syntax
         _dot_split = prefix.rsplit('.', 1)
         if len(_dot_split) == 2 and _dot_split[1]:
             initials = [_dot_split[0][0], _dot_split[1][0]]
         else:
             initials = [prefix[0], server[0]]
 
         # get first letter of first and last names to create initials
         fn_letter = (first_name or " ")[0].strip()
         ln_letter = (last_name or " ")[0].strip()
 
         if fn_letter:
             initials[0] = fn_letter
 
         if ln_letter:
             initials[1] = ln_letter
 
         return ''.join(initials).upper()
 
     def get_img_data_by_type(self, font_family, img_type):
         default_user = """
         <svg xmlns="http://www.w3.org/2000/svg"
         version="1.1" x="0px" y="0px" width="{size}" height="{size}"
         viewBox="-15 -10 439.165 429.164"
 
         xml:space="preserve"
         font-family="{font_family}"
         style="background:{background};" >
 
         <path d="M204.583,216.671c50.664,0,91.74-48.075,
                  91.74-107.378c0-82.237-41.074-107.377-91.74-107.377
                  c-50.668,0-91.74,25.14-91.74,107.377C112.844,
                  168.596,153.916,216.671,
                  204.583,216.671z" fill="{text_color}"/>
         <path d="M407.164,374.717L360.88,
                  270.454c-2.117-4.771-5.836-8.728-10.465-11.138l-71.83-37.392
                  c-1.584-0.823-3.502-0.663-4.926,0.415c-20.316,
                  15.366-44.203,23.488-69.076,23.488c-24.877,
                  0-48.762-8.122-69.078-23.488
                  c-1.428-1.078-3.346-1.238-4.93-0.415L58.75,
                  259.316c-4.631,2.41-8.346,6.365-10.465,11.138L2.001,374.717
                  c-3.191,7.188-2.537,15.412,1.75,22.005c4.285,
                  6.592,11.537,10.526,19.4,10.526h362.861c7.863,0,15.117-3.936,
                  19.402-10.527 C409.699,390.129,
                  410.355,381.902,407.164,374.717z" fill="{text_color}"/>
         </svg>""".format(
             size=self.size,
             background='#979797',  # @grey4
             text_color=self.text_color,
             font_family=font_family)
 
         return {
             "default_user": default_user
         }[img_type]
 
     def get_img_data(self, svg_type=None):
         """
         generates the svg metadata for image
         """
         fonts = [
             '-apple-system',
              'BlinkMacSystemFont',
              'Segoe UI',
              'Roboto',
              'Oxygen-Sans',
              'Ubuntu',
              'Cantarell',
              'Helvetica Neue',
              'sans-serif'
         ]
         font_family = ','.join(fonts)
         if svg_type:
             return self.get_img_data_by_type(font_family, svg_type)
 
         initials = self.get_initials()
         img_data = """
         <svg xmlns="http://www.w3.org/2000/svg" pointer-events="none"
              width="{size}" height="{size}"
              style="width: 100%; height: 100%; background-color: {background}"
              viewBox="0 0 {size} {size}">
             <text text-anchor="middle" y="50%" x="50%" dy="0.35em"
                   pointer-events="auto" fill="{text_color}"
                   font-family="{font_family}"
                   style="font-weight: 400; font-size: {f_size}px;">{text}
             </text>
         </svg>""".format(
             size=self.size,
             f_size=self.size/2.05,  # scale the text inside the box nicely
             background=self.background,
             text_color=self.text_color,
             text=initials.upper(),
             font_family=font_family)
 
         return img_data
 
     def generate_svg(self, svg_type=None):
         img_data = base64_to_str(self.get_img_data(svg_type))
         return "data:image/svg+xml;base64,{}".format(img_data)
 
 
 def initials_gravatar(request, email_address, first_name, last_name, size=30, store_on_disk=False):
 
     svg_type = None
     if email_address == User.DEFAULT_USER_EMAIL:
         svg_type = 'default_user'
 
     klass = InitialsGravatar(email_address, first_name, last_name, size)
 
     if store_on_disk:
         from rhodecode.apps.file_store import utils as store_utils
         from rhodecode.apps.file_store.exceptions import FileNotAllowedException, \
             FileOverSizeException
         from rhodecode.model.db import Session
 
         image_key = md5_safe(email_address.lower()
                              + first_name.lower() + last_name.lower())
 
         storage = store_utils.get_file_storage(request.registry.settings)
         filename = '{}.svg'.format(image_key)
         subdir = 'gravatars'
         # since final name has a counter, we apply the 0
         uid = storage.apply_counter(0, store_utils.uid_filename(filename, randomized=False))
         store_uid = os.path.join(subdir, uid)
 
         db_entry = FileStore.get_by_store_uid(store_uid)
         if db_entry:
             return request.route_path('download_file', fid=store_uid)
 
         img_data = klass.get_img_data(svg_type=svg_type)
         img_file = store_utils.bytes_to_file_obj(img_data)
 
         try:
             store_uid, metadata = storage.save_file(
                 img_file, filename, directory=subdir,
                 extensions=['.svg'], randomized_name=False)
         except (FileNotAllowedException, FileOverSizeException):
             raise
 
         try:
             entry = FileStore.create(
                 file_uid=store_uid, filename=metadata["filename"],
                 file_hash=metadata["sha256"], file_size=metadata["size"],
                 file_display_name=filename,
                 file_description=f'user gravatar `{safe_str(filename)}`',
                 hidden=True, check_acl=False, user_id=1
             )
             Session().add(entry)
             Session().commit()
             log.debug('Stored upload in DB as %s', entry)
         except Exception:
             raise
 
         return request.route_path('download_file', fid=store_uid)
 
     else:
         return klass.generate_svg(svg_type=svg_type)
 
 
 def gravatar_external(request, gravatar_url_tmpl, email_address, size=30):
     return safe_str(gravatar_url_tmpl)\
         .replace('{email}', email_address) \
         .replace('{md5email}', md5_safe(email_address.lower())) \
         .replace('{netloc}', request.host) \
         .replace('{scheme}', request.scheme) \
         .replace('{size}', safe_str(size))
 
 
 def gravatar_url(email_address, size=30, request=None):
     request = request or get_current_request()
     _use_gravatar = request.call_context.visual.use_gravatar
 
     email_address = email_address or User.DEFAULT_USER_EMAIL
     if isinstance(email_address, str):
         # hashlib crashes on unicode items
         email_address = safe_str(email_address)
 
     # empty email or default user
     if not email_address or email_address == User.DEFAULT_USER_EMAIL:
         return initials_gravatar(request, User.DEFAULT_USER_EMAIL, '', '', size=size)
 
     if _use_gravatar:
         gravatar_url_tmpl = request.call_context.visual.gravatar_url \
                             or User.DEFAULT_GRAVATAR_URL
         return gravatar_external(request, gravatar_url_tmpl, email_address, size=size)
 
     else:
         return initials_gravatar(request, email_address, '', '', size=size)
 
 
 def breadcrumb_repo_link(repo):
     """
     Makes a breadcrumbs path link to repo
 
     ex::
         group >> subgroup >> repo
 
     :param repo: a Repository instance
     """
 
     path = [
         link_to(group.name, route_path('repo_group_home', repo_group_name=group.group_name),
                 title='last change:{}'.format(format_date(group.last_commit_change)))
         for group in repo.groups_with_parents
     ] + [
         link_to(repo.just_name, route_path('repo_summary', repo_name=repo.repo_name),
                 title='last change:{}'.format(format_date(repo.last_commit_change)))
     ]
 
     return literal(' &raquo; '.join(path))
 
 
 def breadcrumb_repo_group_link(repo_group):
     """
     Makes a breadcrumbs path link to repo
 
     ex::
         group >> subgroup
 
     :param repo_group: a Repository Group instance
     """
 
     path = [
         link_to(group.name,
                 route_path('repo_group_home', repo_group_name=group.group_name),
                 title='last change:{}'.format(format_date(group.last_commit_change)))
         for group in repo_group.parents
     ] + [
         link_to(repo_group.name,
                 route_path('repo_group_home', repo_group_name=repo_group.group_name),
                 title='last change:{}'.format(format_date(repo_group.last_commit_change)))
     ]
 
     return literal(' &raquo; '.join(path))
 
 
 def format_byte_size_binary(file_size):
     """
     Formats file/folder sizes to standard.
     """
     if file_size is None:
         file_size = 0
 
     formatted_size = format_byte_size(file_size, binary=True)
     return formatted_size
 
 
 def urlify_text(text_, safe=True, **href_attrs):
     """
     Extract urls from text and make html links out of them
     """
 
     url_pat = re.compile(r'''(http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@#.&+]'''
                          r'''|[!*\(\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+)''')
 
     def url_func(match_obj):
         url_full = match_obj.groups()[0]
         a_options = dict(href_attrs)
         a_options['href'] = url_full
         a_text = url_full
         return HTML.tag("a", a_text, **a_options)
 
     _new_text = url_pat.sub(url_func, text_)
 
     if safe:
         return literal(_new_text)
     return _new_text
 
 
 def urlify_commits(text_, repo_name):
     """
     Extract commit ids from text and make link from them
 
     :param text_:
     :param repo_name: repo name to build the URL with
     """
 
     url_pat = re.compile(r'(^|\s)([0-9a-fA-F]{12,40})($|\s)')
 
     def url_func(match_obj):
         commit_id = match_obj.groups()[1]
         pref = match_obj.groups()[0]
         suf = match_obj.groups()[2]
 
         tmpl = (
             '%(pref)s<a class="tooltip-hovercard %(cls)s" href="%(url)s" data-hovercard-alt="%(hovercard_alt)s" data-hovercard-url="%(hovercard_url)s">'
             '%(commit_id)s</a>%(suf)s'
         )
         return tmpl % {
             'pref': pref,
             'cls': 'revision-link',
             'url': route_url(
                 'repo_commit', repo_name=repo_name, commit_id=commit_id),
             'commit_id': commit_id,
             'suf': suf,
             'hovercard_alt': 'Commit: {}'.format(commit_id),
             'hovercard_url': route_url(
                 'hovercard_repo_commit', repo_name=repo_name, commit_id=commit_id)
         }
 
     new_text = url_pat.sub(url_func, text_)
 
     return new_text
 
 
 def _process_url_func(match_obj, repo_name, uid, entry,
                       return_raw_data=False, link_format='html'):
     pref = ''
     if match_obj.group().startswith(' '):
         pref = ' '
 
     issue_id = ''.join(match_obj.groups())
 
     if link_format == 'html':
         tmpl = (
             '%(pref)s<a class="tooltip %(cls)s" href="%(url)s" title="%(title)s">'
             '%(issue-prefix)s%(id-repr)s'
             '</a>')
     elif link_format == 'html+hovercard':
         tmpl = (
             '%(pref)s<a class="tooltip-hovercard %(cls)s" href="%(url)s" data-hovercard-url="%(hovercard_url)s">'
             '%(issue-prefix)s%(id-repr)s'
             '</a>')
     elif link_format in ['rst', 'rst+hovercard']:
         tmpl = '`%(issue-prefix)s%(id-repr)s <%(url)s>`_'
     elif link_format in ['markdown', 'markdown+hovercard']:
         tmpl = '[%(pref)s%(issue-prefix)s%(id-repr)s](%(url)s)'
     else:
         raise ValueError('Bad link_format:{}'.format(link_format))
 
     (repo_name_cleaned,
      parent_group_name) = RepoGroupModel()._get_group_name_and_parent(repo_name)
 
     # variables replacement
     named_vars = {
         'id': issue_id,
         'repo': repo_name,
         'repo_name': repo_name_cleaned,
         'group_name': parent_group_name,
         # set dummy keys so we always have them
         'hostname': '',
         'netloc': '',
         'scheme': ''
     }
 
     request = get_current_request()
     if request:
         # exposes, hostname, netloc, scheme
         host_data = get_host_info(request)
         named_vars.update(host_data)
 
     # named regex variables
     named_vars.update(match_obj.groupdict())
     _url = string.Template(entry['url']).safe_substitute(**named_vars)
     desc = string.Template(escape(entry['desc'])).safe_substitute(**named_vars)
     hovercard_url = string.Template(entry.get('hovercard_url', '')).safe_substitute(**named_vars)
 
     def quote_cleaner(input_str):
         """Remove quotes as it's HTML"""
         return input_str.replace('"', '')
 
     data = {
         'pref': pref,
         'cls': quote_cleaner('issue-tracker-link'),
         'url': quote_cleaner(_url),
         'id-repr': issue_id,
         'issue-prefix': entry['pref'],
         'serv': entry['url'],
         'title': sanitize_html(desc, strip=True),
         'hovercard_url': hovercard_url
     }
 
     if return_raw_data:
         return {
             'id': issue_id,
             'url': _url
         }
     return tmpl % data
 
 
 def get_active_pattern_entries(repo_name):
     repo = None
     if repo_name:
         # Retrieving repo_name to avoid invalid repo_name to explode on
         # IssueTrackerSettingsModel but still passing invalid name further down
         repo = Repository.get_by_repo_name(repo_name, cache=True)
 
     settings_model = IssueTrackerSettingsModel(repo=repo)
     active_entries = settings_model.get_settings(cache=True)
     return active_entries
 
 
 pr_pattern_re = regex.compile(r'(?:(?:^!)|(?: !))(\d+)')
 
 allowed_link_formats = [
     'html', 'rst', 'markdown', 'html+hovercard', 'rst+hovercard', 'markdown+hovercard']
 
 compile_cache = {
 
 }
 
 
 def process_patterns(text_string, repo_name, link_format='html', active_entries=None):
 
     if link_format not in allowed_link_formats:
         raise ValueError('Link format can be only one of:{} got {}'.format(
                          allowed_link_formats, link_format))
     issues_data = []
     errors = []
     new_text = text_string
 
     if active_entries is None:
         log.debug('Fetch active issue tracker patterns for repo: %s', repo_name)
         active_entries = get_active_pattern_entries(repo_name)
 
     log.debug('Got %s pattern entries to process', len(active_entries))
 
     for uid, entry in list(active_entries.items()):
 
         if not (entry['pat'] and entry['url']):
             log.debug('skipping due to missing data')
             continue
 
         log.debug('issue tracker entry: uid: `%s` PAT:%s URL:%s PREFIX:%s',
                   uid, entry['pat'], entry['url'], entry['pref'])
 
         if entry.get('pat_compiled'):
             pattern = entry['pat_compiled']
         elif entry['pat'] in compile_cache:
             pattern = compile_cache[entry['pat']]
         else:
             try:
                 pattern = regex.compile(r'%s' % entry['pat'])
             except regex.error as e:
                 regex_err = ValueError('{}:{}'.format(entry['pat'], e))
                 log.exception('issue tracker pattern: `%s` failed to compile', regex_err)
                 errors.append(regex_err)
                 continue
             compile_cache[entry['pat']] = pattern
 
         data_func = partial(
             _process_url_func, repo_name=repo_name, entry=entry, uid=uid,
             return_raw_data=True)
 
         for match_obj in pattern.finditer(text_string):
             issues_data.append(data_func(match_obj))
 
         url_func = partial(
             _process_url_func, repo_name=repo_name, entry=entry, uid=uid,
             link_format=link_format)
 
         new_text = pattern.sub(url_func, new_text)
         log.debug('processed prefix:uid `%s`', uid)
 
     # finally use global replace, eg !123 -> pr-link, those will not catch
     # if already similar pattern exists
     server_url = '${scheme}://${netloc}'
     pr_entry = {
         'pref': '!',
         'url': server_url + '/_admin/pull-requests/${id}',
         'desc': 'Pull Request !${id}',
         'hovercard_url': server_url + '/_hovercard/pull_request/${id}'
     }
     pr_url_func = partial(
         _process_url_func, repo_name=repo_name, entry=pr_entry, uid=None,
         link_format=link_format+'+hovercard')
     new_text = pr_pattern_re.sub(pr_url_func, new_text)
     log.debug('processed !pr pattern')
 
     return new_text, issues_data, errors
 
 
 def urlify_commit_message(commit_text, repository=None, active_pattern_entries=None,
                           issues_container_callback=None, error_container=None):
     """
     Parses given text message and makes proper links.
     issues are linked to given issue-server, and rest is a commit link
     """
 
     def escaper(_text):
         return _text.replace('<', '&lt;').replace('>', '&gt;')
 
     new_text = escaper(commit_text)
 
     # extract http/https links and make them real urls
     new_text = urlify_text(new_text, safe=False)
 
     # urlify commits - extract commit ids and make link out of them, if we have
     # the scope of repository present.
     if repository:
         new_text = urlify_commits(new_text, repository)
 
     # process issue tracker patterns
     new_text, issues, errors = process_patterns(
         new_text, repository or '', active_entries=active_pattern_entries)
 
     if issues_container_callback is not None:
         for issue in issues:
             issues_container_callback(issue)
 
     if error_container is not None:
         error_container.extend(errors)
 
     return literal(new_text)
 
 
 def render_binary(repo_name, file_obj):
     """
     Choose how to render a binary file
     """
 
     # unicode
     filename = file_obj.name
 
     # images
     for ext in ['*.png', '*.jpeg', '*.jpg', '*.ico', '*.gif']:
         if fnmatch.fnmatch(filename, pat=ext):
             src = route_path(
                 'repo_file_raw', repo_name=repo_name,
                 commit_id=file_obj.commit.raw_id,
                 f_path=file_obj.path)
 
             return literal(
                 '<img class="rendered-binary" alt="rendered-image" src="{}">'.format(src))
 
 
 def renderer_from_filename(filename, exclude=None):
     """
     choose a renderer based on filename, this works only for text based files
     """
 
     # ipython
     for ext in ['*.ipynb']:
         if fnmatch.fnmatch(filename, pat=ext):
             return 'jupyter'
 
     is_markup = MarkupRenderer.renderer_from_filename(filename, exclude=exclude)
     if is_markup:
         return is_markup
     return None
 
 
 def render(source, renderer='rst', mentions=False, relative_urls=None,
            repo_name=None, active_pattern_entries=None, issues_container_callback=None):
 
     def maybe_convert_relative_links(html_source):
         if relative_urls:
             return relative_links(html_source, relative_urls)
         return html_source
 
     if renderer == 'plain':
         return literal(
             MarkupRenderer.plain(source, leading_newline=False))
 
     elif renderer == 'rst':
         if repo_name:
             # process patterns on comments if we pass in repo name
             source, issues, errors = process_patterns(
                 source, repo_name, link_format='rst',
                 active_entries=active_pattern_entries)
             if issues_container_callback is not None:
                 for issue in issues:
                     issues_container_callback(issue)
 
         rendered_block = maybe_convert_relative_links(
                 MarkupRenderer.rst(source, mentions=mentions))
 
         return literal(f'<div class="rst-block">{rendered_block}</div>')
 
     elif renderer == 'markdown':
         if repo_name:
             # process patterns on comments if we pass in repo name
             source, issues, errors = process_patterns(
                 source, repo_name, link_format='markdown',
                 active_entries=active_pattern_entries)
             if issues_container_callback is not None:
                 for issue in issues:
                     issues_container_callback(issue)
 
         rendered_block = maybe_convert_relative_links(
             MarkupRenderer.markdown(source, flavored=True, mentions=mentions))
         return literal(f'<div class="markdown-block">{rendered_block}</div>')
 
     elif renderer == 'jupyter':
         rendered_block = maybe_convert_relative_links(
             MarkupRenderer.jupyter(source))
         return literal(f'<div class="ipynb">{rendered_block}</div>')
 
     # None means just show the file-source
     return None
 
 
 def commit_status(repo, commit_id):
     return ChangesetStatusModel().get_status(repo, commit_id)
 
 
 def commit_status_lbl(commit_status):
     return dict(ChangesetStatus.STATUSES).get(commit_status)
 
 
 def commit_time(repo_name, commit_id):
     repo = Repository.get_by_repo_name(repo_name)
     commit = repo.get_commit(commit_id=commit_id)
     return commit.date
 
 
 def get_permission_name(key):
     return dict(Permission.PERMS).get(key)
 
 
 def journal_filter_help(request):
     _ = request.translate
     from rhodecode.lib.audit_logger import ACTIONS
     actions = '\n'.join(textwrap.wrap(', '.join(sorted(ACTIONS.keys())), 80))
 
     return _(
         'Example filter terms:\n' +
         '     repository:vcs\n' +
         '     username:marcin\n' +
         '     username:(NOT marcin)\n' +
         '     action:*push*\n' +
         '     ip:127.0.0.1\n' +
         '     date:20120101\n' +
         '     date:[20120101100000 TO 20120102]\n' +
         '\n' +
         'Actions: {actions}\n' +
         '\n' +
         'Generate wildcards using \'*\' character:\n' +
         '     "repository:vcs*" - search everything starting with \'vcs\'\n' +
         '     "repository:*vcs*" - search for repository containing \'vcs\'\n' +
         '\n' +
         'Optional AND / OR operators in queries\n' +
         '     "repository:vcs OR repository:test"\n' +
         '     "username:test AND repository:test*"\n'
     ).format(actions=actions)
 
 
 def not_mapped_error(repo_name):
     from rhodecode.translation import _
     flash(_('%s repository is not mapped to db perhaps'
             ' it was created or renamed from the filesystem'
             ' please run the application again'
             ' in order to rescan repositories') % repo_name, category='error')
 
 
 def ip_range(ip_addr):
     from rhodecode.model.db import UserIpMap
     s, e = UserIpMap._get_ip_range(ip_addr)
     return '%s - %s' % (s, e)
 
 
 def form(url, method='post', needs_csrf_token=True, **attrs):
     """Wrapper around webhelpers.tags.form to prevent CSRF attacks."""
     if method.lower() != 'get' and needs_csrf_token:
         raise Exception(
             'Forms to POST/PUT/DELETE endpoints should have (in general) a ' +
             'CSRF token. If the endpoint does not require such token you can ' +
             'explicitly set the parameter needs_csrf_token to false.')
 
     return insecure_form(url, method=method, **attrs)
 
 
 def secure_form(form_url, method="POST", multipart=False, **attrs):
     """Start a form tag that points the action to an url. This
     form tag will also include the hidden field containing
     the auth token.
 
     The url options should be given either as a string, or as a
     ``url()`` function. The method for the form defaults to POST.
 
     Options:
 
     ``multipart``
         If set to True, the enctype is set to "multipart/form-data".
     ``method``
         The method to use when submitting the form, usually either
         "GET" or "POST". If "PUT", "DELETE", or another verb is used, a
         hidden input with name _method is added to simulate the verb
         over POST.
 
     """
 
     if 'request' in attrs:
         session = attrs['request'].session
         del attrs['request']
     else:
         raise ValueError(
             'Calling this form requires request= to be passed as argument')
 
     _form = insecure_form(form_url, method, multipart, **attrs)
     token = literal(
         '<input type="hidden" name="{}" value="{}">'.format(
             csrf_token_key, get_csrf_token(session)))
 
     return literal("%s\n%s" % (_form, token))
 
 
 def dropdownmenu(name, selected, options, enable_filter=False, **attrs):
     select_html = select(name, selected, options, **attrs)
 
     select2 = """
         <script>
             $(document).ready(function() {
                   $('#%s').select2({
                       containerCssClass: 'drop-menu %s',
                       dropdownCssClass: 'drop-menu-dropdown',
                       dropdownAutoWidth: true%s
                   });
             });
         </script>
     """
 
     filter_option = """,
                         minimumResultsForSearch: -1
     """
     input_id = attrs.get('id') or name
     extra_classes = ' '.join(attrs.pop('extra_classes', []))
     filter_enabled = "" if enable_filter else filter_option
     select_script = literal(select2 % (input_id, extra_classes, filter_enabled))
 
     return literal(select_html+select_script)
 
 
 def get_visual_attr(tmpl_context_var, attr_name):
     """
     A safe way to get a variable from visual variable of template context
 
     :param tmpl_context_var: instance of tmpl_context, usually present as `c`
     :param attr_name: name of the attribute we fetch from the c.visual
     """
     visual = getattr(tmpl_context_var, 'visual', None)
     if not visual:
         return
     else:
         return getattr(visual, attr_name, None)
 
 
 def get_last_path_part(file_node):
     if not file_node.path:
         return '/'
 
     path = safe_str(file_node.path.split('/')[-1])
     return '../' + path
 
 
 def route_url(*args, **kwargs):
     """
     Wrapper around pyramids `route_url` (fully qualified url) function.
     """
     req = get_current_request()
     return req.route_url(*args, **kwargs)
 
 
 def route_path(*args, **kwargs):
     """
     Wrapper around pyramids `route_path` function.
     """
     req = get_current_request()
     return req.route_path(*args, **kwargs)
 
 
 def route_path_or_none(*args, **kwargs):
     try:
         return route_path(*args, **kwargs)
     except KeyError:
         return None
 
 
 def current_route_path(request, **kw):
     new_args = request.GET.mixed()
     new_args.update(kw)
     return request.current_route_path(_query=new_args)
 
 
 def curl_api_example(method, args):
     args_json = json.dumps(OrderedDict([
         ('id', 1),
         ('auth_token', 'SECRET'),
         ('method', method),
         ('args', args)
     ]))
 
     return "curl {api_url} -X POST -H 'content-type:text/plain' --data-binary '{args_json}'".format(
         api_url=route_url('apiv2'),
         args_json=args_json
     )
 
 
 def api_call_example(method, args):
     """
     Generates an API call example via CURL
     """
     curl_call = curl_api_example(method, args)
 
     return literal(
         curl_call +
         "<br/><br/>SECRET can be found in <a href=\"{token_url}\">auth-tokens</a> page, "
         "and needs to be of `api calls` role."
         .format(token_url=route_url('my_account_auth_tokens')))
 
 
 def notification_description(notification, request):
     """
     Generate notification human readable description based on notification type
     """
     from rhodecode.model.notification import NotificationModel
     return NotificationModel().make_description(
         notification, translate=request.translate)
 
 
 def go_import_header(request, db_repo=None):
     """
     Creates a header for go-import functionality in Go Lang
     """
 
     if not db_repo:
         return
     if 'go-get' not in request.GET:
         return
 
     clone_url = db_repo.clone_url()
     prefix = re.split(r'^https?:\/\/', clone_url)[-1]
     # we have a repo and go-get flag,
     return literal('<meta name="go-import" content="{} {} {}">'.format(
         prefix, db_repo.repo_type, clone_url))
 
 
 def reviewer_as_json(*args, **kwargs):
     from rhodecode.apps.repository.utils import reviewer_as_json as _reviewer_as_json
     return _reviewer_as_json(*args, **kwargs)
 
 
 def get_repo_view_type(request):
     route_name = request.matched_route.name
     route_to_view_type = {
         'repo_changelog': 'commits',
         'repo_commits': 'commits',
         'repo_files': 'files',
         'repo_summary': 'summary',
         'repo_commit': 'commit'
     }
 
     return route_to_view_type.get(route_name)
 
 
 def is_active(menu_entry, selected):
     """
     Returns active class for selecting menus in templates
     <li class=${h.is_active('settings', current_active)}></li>
     """
     if not isinstance(menu_entry, list):
         menu_entry = [menu_entry]
 
     if selected in menu_entry:
         return "active"
 
 
 class IssuesRegistry(object):
     """
     issue_registry = IssuesRegistry()
     some_func(issues_callback=issues_registry(...))
     """
 
     def __init__(self):
         self.issues = []
         self.unique_issues = collections.defaultdict(lambda: [])
 
     def __call__(self, commit_dict=None):
         def callback(issue):
             if commit_dict and issue:
                 issue['commit'] = commit_dict
             self.issues.append(issue)
             self.unique_issues[issue['id']].append(issue)
         return callback
 
     def get_issues(self):
         return self.issues
 
     @property
     def issues_unique_count(self):
         return len(set(i['id'] for i in self.issues))
+
+
+def get_directory_statistics(start_path):
+    """
+    total_files, total_size, directory_stats = get_directory_statistics(start_path)
+
+    print(f"Directory statistics for: {start_path}\n")
+    print(f"Total files: {total_files}")
+    print(f"Total size: {format_size(total_size)}\n")
+
+    :param start_path:
+    :return:
+    """
+
+    total_files = 0
+    total_size = 0
+    directory_stats = {}
+
+    for dir_path, dir_names, file_names in os.walk(start_path):
+        dir_size = 0
+        file_count = len(file_names)
+
+        for fname in file_names:
+            filepath = os.path.join(dir_path, fname)
+            file_size = os.path.getsize(filepath)
+            dir_size += file_size
+
+        directory_stats[dir_path] = {'file_count': file_count, 'size': dir_size}
+        total_files += file_count
+        total_size += dir_size
+
+    return total_files, total_size, directory_stats

rhodecode/lib/system_info.py

@@ -1,834 +1,827 b''
 # Copyright (C) 2017-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 
 import os
 import sys
 import time
 import platform
 import collections
 import psutil
 from functools import wraps
 
 import pkg_resources
 import logging
 import resource
 
 import configparser
 
 from rc_license.models import LicenseModel
 from rhodecode.lib.str_utils import safe_str
 
 log = logging.getLogger(__name__)
 
 
 _NA = 'NOT AVAILABLE'
 _NA_FLOAT = 0.0
 
 STATE_OK = 'ok'
 STATE_ERR = 'error'
 STATE_WARN = 'warning'
 
 STATE_OK_DEFAULT = {'message': '', 'type': STATE_OK}
 
 
 registered_helpers = {}
 
 
 def register_sysinfo(func):
     """
     @register_helper
     def db_check():
         pass
 
     db_check == registered_helpers['db_check']
     """
     global registered_helpers
     registered_helpers[func.__name__] = func
 
     @wraps(func)
     def _wrapper(*args, **kwargs):
         return func(*args, **kwargs)
     return _wrapper
 
 
 # HELPERS
 def percentage(part: (int, float), whole: (int, float)):
     whole = float(whole)
     if whole > 0:
         return round(100 * float(part) / whole, 1)
     return 0.0
 
 
 def get_storage_size(storage_path):
     sizes = []
     for file_ in os.listdir(storage_path):
         storage_file = os.path.join(storage_path, file_)
         if os.path.isfile(storage_file):
             try:
                 sizes.append(os.path.getsize(storage_file))
             except OSError:
                 log.exception('Failed to get size of storage file %s', storage_file)
                 pass
 
     return sum(sizes)
 
 
 def get_resource(resource_type):
     try:
         return resource.getrlimit(resource_type)
     except Exception:
         return 'NOT_SUPPORTED'
 
 
 def get_cert_path(ini_path):
     default = '/etc/ssl/certs/ca-certificates.crt'
     control_ca_bundle = os.path.join(
         os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(ini_path)))),
         '/etc/ssl/certs/ca-certificates.crt')
     if os.path.isfile(control_ca_bundle):
         default = control_ca_bundle
 
     return default
 
 
 class SysInfoRes(object):
     def __init__(self, value, state=None, human_value=None):
         self.value = value
         self.state = state or STATE_OK_DEFAULT
         self.human_value = human_value or value
 
     def __json__(self):
         return {
             'value': self.value,
             'state': self.state,
             'human_value': self.human_value,
         }
 
     def get_value(self):
         return self.__json__()
 
     def __str__(self):
         return f'<SysInfoRes({self.__json__()})>'
 
 
 class SysInfo(object):
 
     def __init__(self, func_name, **kwargs):
         self.function_name = func_name
         self.value = _NA
         self.state = None
         self.kwargs = kwargs or {}
 
     def __call__(self):
         computed = self.compute(**self.kwargs)
         if not isinstance(computed, SysInfoRes):
             raise ValueError(
                 'computed value for {} is not instance of '
                 '{}, got {} instead'.format(
                     self.function_name, SysInfoRes, type(computed)))
         return computed.__json__()
 
     def __str__(self):
         return f'<SysInfo({self.function_name})>'
 
     def compute(self, **kwargs):
         return self.function_name(**kwargs)
 
 
 # SysInfo functions
 @register_sysinfo
 def python_info():
     value = dict(version=f'{platform.python_version()}:{platform.python_implementation()}',
                  executable=sys.executable)
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def py_modules():
     mods = dict([(p.project_name, {'version': p.version, 'location': p.location})
                  for p in pkg_resources.working_set])
 
     value = sorted(mods.items(), key=lambda k: k[0].lower())
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def platform_type():
     from rhodecode.lib.utils import generate_platform_uuid
 
     value = dict(
         name=safe_str(platform.platform()),
         uuid=generate_platform_uuid()
     )
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def locale_info():
     import locale
 
     def safe_get_locale(locale_name):
         try:
             locale.getlocale(locale_name)
         except TypeError:
             return f'FAILED_LOCALE_GET:{locale_name}'
 
     value = dict(
         locale_default=locale.getlocale(),
         locale_lc_all=safe_get_locale(locale.LC_ALL),
         locale_lc_ctype=safe_get_locale(locale.LC_CTYPE),
         lang_env=os.environ.get('LANG'),
         lc_all_env=os.environ.get('LC_ALL'),
         local_archive_env=os.environ.get('LOCALE_ARCHIVE'),
     )
     human_value = \
         f"LANG: {value['lang_env']}, \
           locale LC_ALL: {value['locale_lc_all']},  \
           locale LC_CTYPE: {value['locale_lc_ctype']},  \
           Default locales: {value['locale_default']}"
 
     return SysInfoRes(value=value, human_value=human_value)
 
 
 @register_sysinfo
 def ulimit_info():
     data = collections.OrderedDict([
         ('cpu time (seconds)', get_resource(resource.RLIMIT_CPU)),
         ('file size', get_resource(resource.RLIMIT_FSIZE)),
         ('stack size', get_resource(resource.RLIMIT_STACK)),
         ('core file size', get_resource(resource.RLIMIT_CORE)),
         ('address space size', get_resource(resource.RLIMIT_AS)),
         ('locked in mem size', get_resource(resource.RLIMIT_MEMLOCK)),
         ('heap size', get_resource(resource.RLIMIT_DATA)),
         ('rss size', get_resource(resource.RLIMIT_RSS)),
         ('number of processes', get_resource(resource.RLIMIT_NPROC)),
         ('open files', get_resource(resource.RLIMIT_NOFILE)),
     ])
 
     text = ', '.join(f'{k}:{v}' for k, v in data.items())
 
     value = {
         'limits': data,
         'text': text,
     }
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def uptime():
     from rhodecode.lib.helpers import age, time_to_datetime
     from rhodecode.translation import TranslationString
 
     value = dict(boot_time=0, uptime=0, text='')
     state = STATE_OK_DEFAULT
 
     boot_time = psutil.boot_time()
     value['boot_time'] = boot_time
     value['uptime'] = time.time() - boot_time
 
     date_or_age = age(time_to_datetime(boot_time))
     if isinstance(date_or_age, TranslationString):
         date_or_age = date_or_age.interpolate()
 
     human_value = value.copy()
     human_value['boot_time'] = time_to_datetime(boot_time)
     human_value['uptime'] = age(time_to_datetime(boot_time), show_suffix=False)
 
     human_value['text'] = f'Server started {date_or_age}'
     return SysInfoRes(value=value, human_value=human_value)
 
 
 @register_sysinfo
 def memory():
     from rhodecode.lib.helpers import format_byte_size_binary
     value = dict(available=0, used=0, used_real=0, cached=0, percent=0,
                  percent_used=0, free=0, inactive=0, active=0, shared=0,
                  total=0, buffers=0, text='')
 
     state = STATE_OK_DEFAULT
 
     value.update(dict(psutil.virtual_memory()._asdict()))
     value['used_real'] = value['total'] - value['available']
     value['percent_used'] = psutil._common.usage_percent(value['used_real'], value['total'], 1)
 
     human_value = value.copy()
     human_value['text'] = '{}/{}, {}% used'.format(
         format_byte_size_binary(value['used_real']),
         format_byte_size_binary(value['total']),
         value['percent_used'])
 
     keys = list(value.keys())[::]
     keys.pop(keys.index('percent'))
     keys.pop(keys.index('percent_used'))
     keys.pop(keys.index('text'))
     for k in keys:
         human_value[k] = format_byte_size_binary(value[k])
 
     if state['type'] == STATE_OK and value['percent_used'] > 90:
         msg = 'Critical: your available RAM memory is very low.'
         state = {'message': msg, 'type': STATE_ERR}
 
     elif state['type'] == STATE_OK and value['percent_used'] > 70:
         msg = 'Warning: your available RAM memory is running low.'
         state = {'message': msg, 'type': STATE_WARN}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def machine_load():
     value = {'1_min': _NA_FLOAT, '5_min': _NA_FLOAT, '15_min': _NA_FLOAT, 'text': ''}
     state = STATE_OK_DEFAULT
 
     # load averages
     if hasattr(psutil.os, 'getloadavg'):
         value.update(dict(
             list(zip(['1_min', '5_min', '15_min'], psutil.os.getloadavg()))
         ))
 
     human_value = value.copy()
     human_value['text'] = '1min: {}, 5min: {}, 15min: {}'.format(
         value['1_min'], value['5_min'], value['15_min'])
 
     if state['type'] == STATE_OK and value['15_min'] > 5.0:
         msg = 'Warning: your machine load is very high.'
         state = {'message': msg, 'type': STATE_WARN}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def cpu():
     value = {'cpu': 0, 'cpu_count': 0, 'cpu_usage': []}
     state = STATE_OK_DEFAULT
 
     value['cpu'] = psutil.cpu_percent(0.5)
     value['cpu_usage'] = psutil.cpu_percent(0.5, percpu=True)
     value['cpu_count'] = psutil.cpu_count()
 
     human_value = value.copy()
     human_value['text'] = f'{value["cpu_count"]} cores at {value["cpu"]} %'
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage():
     from rhodecode.lib.helpers import format_byte_size_binary
     from rhodecode.lib.utils import get_rhodecode_repo_store_path
     path = get_rhodecode_repo_store_path()
 
     value = dict(percent=0, used=0, total=0, path=path, text='')
     state = STATE_OK_DEFAULT
 
     try:
         value.update(dict(psutil.disk_usage(path)._asdict()))
     except Exception as e:
         log.exception('Failed to fetch disk info')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['used'] = format_byte_size_binary(value['used'])
     human_value['total'] = format_byte_size_binary(value['total'])
     human_value['text'] = "{}/{}, {}% used".format(
         format_byte_size_binary(value['used']),
         format_byte_size_binary(value['total']),
         value['percent'])
 
     if state['type'] == STATE_OK and value['percent'] > 90:
         msg = 'Critical: your disk space is very low.'
         state = {'message': msg, 'type': STATE_ERR}
 
     elif state['type'] == STATE_OK and value['percent'] > 70:
         msg = 'Warning: your disk space is running low.'
         state = {'message': msg, 'type': STATE_WARN}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage_inodes():
     from rhodecode.lib.utils import get_rhodecode_repo_store_path
     path = get_rhodecode_repo_store_path()
 
     value = dict(percent=0.0, free=0, used=0, total=0, path=path, text='')
     state = STATE_OK_DEFAULT
 
     try:
         i_stat = os.statvfs(path)
         value['free'] = i_stat.f_ffree
         value['used'] = i_stat.f_files-i_stat.f_favail
         value['total'] = i_stat.f_files
         value['percent'] = percentage(value['used'], value['total'])
     except Exception as e:
         log.exception('Failed to fetch disk inodes info')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['text'] = "{}/{}, {}% used".format(
         value['used'], value['total'], value['percent'])
 
     if state['type'] == STATE_OK and value['percent'] > 90:
         msg = 'Critical: your disk free inodes are very low.'
         state = {'message': msg, 'type': STATE_ERR}
 
     elif state['type'] == STATE_OK and value['percent'] > 70:
         msg = 'Warning: your disk free inodes are running low.'
         state = {'message': msg, 'type': STATE_WARN}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage_archives():
     import rhodecode
     from rhodecode.lib.helpers import format_byte_size_binary
-    from rhodecode.lib.rc_cache.archive_cache.utils import get_directory_statistics
+    from rhodecode.lib.archive_cache import get_archival_cache_store
 
     storage_type = rhodecode.ConfigGet().get_str('archive_cache.backend.type')
-    storage_key = 'archive_cache.filesystem.store_dir'
 
-    default_msg = 'Archive cache storage is controlled by '\
-                  f'{storage_key}=/path/to/cache option in the .ini file'
-    path = rhodecode.ConfigGet().get_str(storage_key, missing=default_msg)
-
-    value = dict(percent=0, used=0, total=0, items=0, path=path, text='', type=storage_type)
+    value = dict(percent=0, used=0, total=0, items=0, path='', text='', type=storage_type)
     state = STATE_OK_DEFAULT
     try:
-        if storage_type != 'filesystem':
-            # raise Exc to stop reporting on different type
-            raise ValueError('Storage type must be "filesystem"')
+        d_cache = get_archival_cache_store(config=rhodecode.CONFIG)
 
-        total_files, total_size, _directory_stats = get_directory_statistics(path)
+        total_files, total_size, _directory_stats = d_cache.get_statistics()
 
         value.update({
             'percent': 100,
             'used': total_size,
             'total': total_size,
-            'items': total_files
+            'items': total_files,
+            'path': d_cache.storage_path
         })
 
     except Exception as e:
         log.exception('failed to fetch archive cache storage')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['used'] = format_byte_size_binary(value['used'])
     human_value['total'] = format_byte_size_binary(value['total'])
     human_value['text'] = "{} ({} items)".format(
         human_value['used'], value['items'])
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage_gist():
     from rhodecode.model.gist import GIST_STORE_LOC
     from rhodecode.lib.utils import safe_str, get_rhodecode_repo_store_path
-    from rhodecode.lib.helpers import format_byte_size_binary
-    from rhodecode.lib.rc_cache.archive_cache.utils import get_directory_statistics
+    from rhodecode.lib.helpers import format_byte_size_binary, get_directory_statistics
 
     path = safe_str(os.path.join(
         get_rhodecode_repo_store_path(), GIST_STORE_LOC))
 
     # gist storage
     value = dict(percent=0, used=0, total=0, items=0, path=path, text='')
     state = STATE_OK_DEFAULT
 
     try:
         total_files, total_size, _directory_stats = get_directory_statistics(path)
         value.update({
             'percent': 100,
             'used': total_size,
             'total': total_size,
             'items': total_files
         })
     except Exception as e:
         log.exception('failed to fetch gist storage items')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['used'] = format_byte_size_binary(value['used'])
     human_value['total'] = format_byte_size_binary(value['total'])
     human_value['text'] = "{} ({} items)".format(
         human_value['used'], value['items'])
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage_temp():
     import tempfile
     from rhodecode.lib.helpers import format_byte_size_binary
 
     path = tempfile.gettempdir()
     value = dict(percent=0, used=0, total=0, items=0, path=path, text='')
     state = STATE_OK_DEFAULT
 
     if not psutil:
         return SysInfoRes(value=value, state=state)
 
     try:
         value.update(dict(psutil.disk_usage(path)._asdict()))
     except Exception as e:
         log.exception('Failed to fetch temp dir info')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['used'] = format_byte_size_binary(value['used'])
     human_value['total'] = format_byte_size_binary(value['total'])
     human_value['text'] = "{}/{}, {}% used".format(
         format_byte_size_binary(value['used']),
         format_byte_size_binary(value['total']),
         value['percent'])
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def search_info():
     import rhodecode
     from rhodecode.lib.index import searcher_from_config
 
     backend = rhodecode.CONFIG.get('search.module', '')
     location = rhodecode.CONFIG.get('search.location', '')
 
     try:
         searcher = searcher_from_config(rhodecode.CONFIG)
         searcher = searcher.__class__.__name__
     except Exception:
         searcher = None
 
     value = dict(
         backend=backend, searcher=searcher, location=location, text='')
     state = STATE_OK_DEFAULT
 
     human_value = value.copy()
     human_value['text'] = "backend:`{}`".format(human_value['backend'])
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def git_info():
     from rhodecode.lib.vcs.backends import git
     state = STATE_OK_DEFAULT
     value = human_value = ''
     try:
         value = git.discover_git_version(raise_on_exc=True)
         human_value = f'version reported from VCSServer: {value}'
     except Exception as e:
         state = {'message': str(e), 'type': STATE_ERR}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def hg_info():
     from rhodecode.lib.vcs.backends import hg
     state = STATE_OK_DEFAULT
     value = human_value = ''
     try:
         value = hg.discover_hg_version(raise_on_exc=True)
         human_value = f'version reported from VCSServer: {value}'
     except Exception as e:
         state = {'message': str(e), 'type': STATE_ERR}
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def svn_info():
     from rhodecode.lib.vcs.backends import svn
     state = STATE_OK_DEFAULT
     value = human_value = ''
     try:
         value = svn.discover_svn_version(raise_on_exc=True)
         human_value = f'version reported from VCSServer: {value}'
     except Exception as e:
         state = {'message': str(e), 'type': STATE_ERR}
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def vcs_backends():
     import rhodecode
     value = rhodecode.CONFIG.get('vcs.backends')
     human_value = 'Enabled backends in order: {}'.format(','.join(value))
     return SysInfoRes(value=value, human_value=human_value)
 
 
 @register_sysinfo
 def vcs_server():
     import rhodecode
     from rhodecode.lib.vcs.backends import get_vcsserver_service_data
 
     server_url = rhodecode.CONFIG.get('vcs.server')
     enabled = rhodecode.CONFIG.get('vcs.server.enable')
     protocol = rhodecode.CONFIG.get('vcs.server.protocol') or 'http'
     state = STATE_OK_DEFAULT
     version = None
     workers = 0
 
     try:
         data = get_vcsserver_service_data()
         if data and 'version' in data:
             version = data['version']
 
         if data and 'config' in data:
             conf = data['config']
             workers = conf.get('workers', 'NOT AVAILABLE')
 
         connection = 'connected'
     except Exception as e:
         connection = 'failed'
         state = {'message': str(e), 'type': STATE_ERR}
 
     value = dict(
         url=server_url,
         enabled=enabled,
         protocol=protocol,
         connection=connection,
         version=version,
         text='',
     )
 
     human_value = value.copy()
     human_value['text'] = \
         '{url}@ver:{ver} via {mode} mode[workers:{workers}], connection:{conn}'.format(
             url=server_url, ver=version, workers=workers, mode=protocol,
             conn=connection)
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def vcs_server_config():
     from rhodecode.lib.vcs.backends import get_vcsserver_service_data
     state = STATE_OK_DEFAULT
 
     value = {}
     try:
         data = get_vcsserver_service_data()
         value = data['app_config']
     except Exception as e:
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['text'] = 'VCS Server config'
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def rhodecode_app_info():
     import rhodecode
     edition = rhodecode.CONFIG.get('rhodecode.edition')
 
     value = dict(
         rhodecode_version=rhodecode.__version__,
         rhodecode_lib_path=os.path.abspath(rhodecode.__file__),
         text=''
     )
     human_value = value.copy()
     human_value['text'] = 'RhodeCode {edition}, version {ver}'.format(
         edition=edition, ver=value['rhodecode_version']
     )
     return SysInfoRes(value=value, human_value=human_value)
 
 
 @register_sysinfo
 def rhodecode_config():
     import rhodecode
     path = rhodecode.CONFIG.get('__file__')
     rhodecode_ini_safe = rhodecode.CONFIG.copy()
     cert_path = get_cert_path(path)
 
     try:
         config = configparser.ConfigParser()
         config.read(path)
         parsed_ini = config
         if parsed_ini.has_section('server:main'):
             parsed_ini = dict(parsed_ini.items('server:main'))
     except Exception:
         log.exception('Failed to read .ini file for display')
         parsed_ini = {}
 
     rhodecode_ini_safe['server:main'] = parsed_ini
 
     blacklist = [
         f'rhodecode_{LicenseModel.LICENSE_DB_KEY}',
         'routes.map',
         'sqlalchemy.db1.url',
         'channelstream.secret',
         'beaker.session.secret',
         'rhodecode.encrypted_values.secret',
         'rhodecode_auth_github_consumer_key',
         'rhodecode_auth_github_consumer_secret',
         'rhodecode_auth_google_consumer_key',
         'rhodecode_auth_google_consumer_secret',
         'rhodecode_auth_bitbucket_consumer_secret',
         'rhodecode_auth_bitbucket_consumer_key',
         'rhodecode_auth_twitter_consumer_secret',
         'rhodecode_auth_twitter_consumer_key',
 
         'rhodecode_auth_twitter_secret',
         'rhodecode_auth_github_secret',
         'rhodecode_auth_google_secret',
         'rhodecode_auth_bitbucket_secret',
 
         'appenlight.api_key',
         ('app_conf', 'sqlalchemy.db1.url')
     ]
     for k in blacklist:
         if isinstance(k, tuple):
             section, key = k
             if section in rhodecode_ini_safe:
                 rhodecode_ini_safe[section] = '**OBFUSCATED**'
         else:
             rhodecode_ini_safe.pop(k, None)
 
     # TODO: maybe put some CONFIG checks here ?
     return SysInfoRes(value={'config': rhodecode_ini_safe,
                              'path': path, 'cert_path': cert_path})
 
 
 @register_sysinfo
 def database_info():
     import rhodecode
     from sqlalchemy.engine import url as engine_url
     from rhodecode.model import meta
     from rhodecode.model.meta import Session
     from rhodecode.model.db import DbMigrateVersion
 
     state = STATE_OK_DEFAULT
 
     db_migrate = DbMigrateVersion.query().filter(
         DbMigrateVersion.repository_id == 'rhodecode_db_migrations').one()
 
     db_url_obj = engine_url.make_url(rhodecode.CONFIG['sqlalchemy.db1.url'])
 
     try:
         engine = meta.get_engine()
         db_server_info = engine.dialect._get_server_version_info(
             Session.connection(bind=engine))
         db_version = '.'.join(map(str, db_server_info))
     except Exception:
         log.exception('failed to fetch db version')
         db_version = 'UNKNOWN'
 
     db_info = dict(
         migrate_version=db_migrate.version,
         type=db_url_obj.get_backend_name(),
         version=db_version,
         url=repr(db_url_obj)
     )
     current_version = db_migrate.version
     expected_version = rhodecode.__dbversion__
     if state['type'] == STATE_OK and current_version != expected_version:
         msg = 'Critical: database schema mismatch, ' \
               'expected version {}, got {}. ' \
               'Please run migrations on your database.'.format(
             expected_version, current_version)
         state = {'message': msg, 'type': STATE_ERR}
 
     human_value = db_info.copy()
     human_value['url'] = "{} @ migration version: {}".format(
         db_info['url'], db_info['migrate_version'])
     human_value['version'] = "{} {}".format(db_info['type'], db_info['version'])
     return SysInfoRes(value=db_info, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def server_info(environ):
     import rhodecode
     from rhodecode.lib.base import get_server_ip_addr, get_server_port
 
     value = {
         'server_ip': '{}:{}'.format(
             get_server_ip_addr(environ, log_errors=False),
             get_server_port(environ)
         ),
         'server_id': rhodecode.CONFIG.get('instance_id'),
     }
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def usage_info():
     from rhodecode.model.db import User, Repository, true
     value = {
         'users': User.query().count(),
         'users_active': User.query().filter(User.active == true()).count(),
         'repositories': Repository.query().count(),
         'repository_types': {
             'hg': Repository.query().filter(
                 Repository.repo_type == 'hg').count(),
             'git': Repository.query().filter(
                 Repository.repo_type == 'git').count(),
             'svn': Repository.query().filter(
                 Repository.repo_type == 'svn').count(),
         },
     }
     return SysInfoRes(value=value)
 
 
 def get_system_info(environ):
     environ = environ or {}
     return {
         'rhodecode_app': SysInfo(rhodecode_app_info)(),
         'rhodecode_config': SysInfo(rhodecode_config)(),
         'rhodecode_usage': SysInfo(usage_info)(),
         'python': SysInfo(python_info)(),
         'py_modules': SysInfo(py_modules)(),
 
         'platform': SysInfo(platform_type)(),
         'locale': SysInfo(locale_info)(),
         'server': SysInfo(server_info, environ=environ)(),
         'database': SysInfo(database_info)(),
         'ulimit': SysInfo(ulimit_info)(),
         'storage': SysInfo(storage)(),
         'storage_inodes': SysInfo(storage_inodes)(),
         'storage_archive': SysInfo(storage_archives)(),
         'storage_gist': SysInfo(storage_gist)(),
         'storage_temp': SysInfo(storage_temp)(),
 
         'search': SysInfo(search_info)(),
 
         'uptime': SysInfo(uptime)(),
         'load': SysInfo(machine_load)(),
         'cpu': SysInfo(cpu)(),
         'memory': SysInfo(memory)(),
 
         'vcs_backends': SysInfo(vcs_backends)(),
         'vcs_server': SysInfo(vcs_server)(),
 
         'vcs_server_config': SysInfo(vcs_server_config)(),
 
         'git': SysInfo(git_info)(),
         'hg': SysInfo(hg_info)(),
         'svn': SysInfo(svn_info)(),
     }
 
 
 def load_system_info(key):
     """
     get_sys_info('vcs_server')
     get_sys_info('database')
     """
     return SysInfo(registered_helpers[key])()

rhodecode/tests/fixture_mods/fixture_pyramid.py

@@ -1,226 +1,226 b''
 
 # Copyright (C) 2010-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import pytest
 
 from rhodecode.lib.config_utils import get_app_config
 from rhodecode.tests.fixture import TestINI
 from rhodecode.tests import TESTS_TMP_PATH
 from rhodecode.tests.server_utils import RcVCSServer
 
 
 @pytest.fixture(scope='session')
 def vcsserver(request, vcsserver_port, vcsserver_factory):
     """
     Session scope VCSServer.
 
     Tests which need the VCSServer have to rely on this fixture in order
     to ensure it will be running.
 
     For specific needs, the fixture vcsserver_factory can be used. It allows to
     adjust the configuration file for the test run.
 
     Command line args:
 
     --without-vcsserver: Allows to switch this fixture off. You have to
     manually start the server.
 
     --vcsserver-port: Will expect the VCSServer to listen on this port.
     """
 
     if not request.config.getoption('with_vcsserver'):
         return None
 
     return vcsserver_factory(
         request, vcsserver_port=vcsserver_port)
 
 
 @pytest.fixture(scope='session')
 def vcsserver_factory(tmpdir_factory):
     """
     Use this if you need a running vcsserver with a special configuration.
     """
 
     def factory(request, overrides=(), vcsserver_port=None,
                 log_file=None, workers='2'):
 
         if vcsserver_port is None:
             vcsserver_port = get_available_port()
 
         overrides = list(overrides)
         overrides.append({'server:main': {'port': vcsserver_port}})
 
         option_name = 'vcsserver_config_http'
         override_option_name = 'vcsserver_config_override'
         config_file = get_config(
             request.config, option_name=option_name,
             override_option_name=override_option_name, overrides=overrides,
             basetemp=tmpdir_factory.getbasetemp().strpath,
             prefix='test_vcs_')
 
         server = RcVCSServer(config_file, log_file, workers)
         server.start()
 
         @request.addfinalizer
         def cleanup():
             server.shutdown()
 
         server.wait_until_ready()
         return server
 
     return factory
 
 
 def _use_log_level(config):
     level = config.getoption('test_loglevel') or 'critical'
     return level.upper()
 
 
 @pytest.fixture(scope='session')
 def ini_config(request, tmpdir_factory, rcserver_port, vcsserver_port):
     option_name = 'pyramid_config'
     log_level = _use_log_level(request.config)
 
     overrides = [
         {'server:main': {'port': rcserver_port}},
         {'app:main': {
-            'cache_dir': '%(here)s/rc_data',
+            'cache_dir': '%(here)s/rc-tests/rc_data',
             'vcs.server': f'localhost:{vcsserver_port}',
             # johbo: We will always start the VCSServer on our own based on the
             # fixtures of the test cases. For the test run it must always be
             # off in the INI file.
             'vcs.start_server': 'false',
 
             'vcs.server.protocol': 'http',
             'vcs.scm_app_implementation': 'http',
             'vcs.svn.proxy.enabled': 'true',
             'vcs.hooks.protocol': 'http',
             'vcs.hooks.host': '*',
             'repo_store.path': TESTS_TMP_PATH,
             'app.service_api.token': 'service_secret_token',
         }},
 
         {'handler_console': {
             'class': 'StreamHandler',
             'args': '(sys.stderr,)',
             'level': log_level,
         }},
 
     ]
 
     filename = get_config(
         request.config, option_name=option_name,
         override_option_name='{}_override'.format(option_name),
         overrides=overrides,
         basetemp=tmpdir_factory.getbasetemp().strpath,
         prefix='test_rce_')
     return filename
 
 
 @pytest.fixture(scope='session')
 def ini_settings(ini_config):
     ini_path = ini_config
     return get_app_config(ini_path)
 
 
 def get_available_port(min_port=40000, max_port=55555):
     from rhodecode.lib.utils2 import get_available_port as _get_port
     return _get_port(min_port, max_port)
 
 
 @pytest.fixture(scope='session')
 def rcserver_port(request):
     port = get_available_port()
     print(f'Using rhodecode port {port}')
     return port
 
 
 @pytest.fixture(scope='session')
 def vcsserver_port(request):
     port = request.config.getoption('--vcsserver-port')
     if port is None:
         port = get_available_port()
         print(f'Using vcsserver port {port}')
     return port
 
 
 @pytest.fixture(scope='session')
 def available_port_factory():
     """
     Returns a callable which returns free port numbers.
     """
     return get_available_port
 
 
 @pytest.fixture()
 def available_port(available_port_factory):
     """
     Gives you one free port for the current test.
 
     Uses "available_port_factory" to retrieve the port.
     """
     return available_port_factory()
 
 
 @pytest.fixture(scope='session')
 def testini_factory(tmpdir_factory, ini_config):
     """
     Factory to create an INI file based on TestINI.
 
     It will make sure to place the INI file in the correct directory.
     """
     basetemp = tmpdir_factory.getbasetemp().strpath
     return TestIniFactory(basetemp, ini_config)
 
 
 class TestIniFactory(object):
 
     def __init__(self, basetemp, template_ini):
         self._basetemp = basetemp
         self._template_ini = template_ini
 
     def __call__(self, ini_params, new_file_prefix='test'):
         ini_file = TestINI(
             self._template_ini, ini_params=ini_params,
             new_file_prefix=new_file_prefix, dir=self._basetemp)
         result = ini_file.create()
         return result
 
 
 def get_config(
         config, option_name, override_option_name, overrides=None,
         basetemp=None, prefix='test'):
     """
     Find a configuration file and apply overrides for the given `prefix`.
     """
     config_file = (
         config.getoption(option_name) or config.getini(option_name))
     if not config_file:
         pytest.exit(
             "Configuration error, could not extract {}.".format(option_name))
 
     overrides = overrides or []
     config_override = config.getoption(override_option_name)
     if config_override:
         overrides.append(config_override)
     temp_ini_file = TestINI(
         config_file, ini_params=overrides, new_file_prefix=prefix,
         dir=basetemp)
 
     return temp_ini_file.create()

rhodecode/tests/rhodecode.ini

@@ -1,785 +1,827 b''
 
 ; #########################################
 ; RHODECODE COMMUNITY EDITION CONFIGURATION
 ; #########################################
 
 [DEFAULT]
 ; Debug flag sets all loggers to debug, and enables request tracking
 debug = true
 
 ; ########################################################################
 ; EMAIL CONFIGURATION
 ; These settings will be used by the RhodeCode mailing system
 ; ########################################################################
 
 ; prefix all emails subjects with given prefix, helps filtering out emails
 #email_prefix = [RhodeCode]
 
 ; email FROM address all mails will be sent
 #app_email_from = rhodecode-noreply@localhost
 
 #smtp_server = mail.server.com
 #smtp_username =
 #smtp_password =
 #smtp_port =
 #smtp_use_tls = false
 #smtp_use_ssl = true
 
 [server:main]
 ; COMMON HOST/IP CONFIG, This applies mostly to develop setup,
 ; Host port for gunicorn are controlled by gunicorn_conf.py
 host = 127.0.0.1
 port = 10020
 
 
 ; ###########################
 ; GUNICORN APPLICATION SERVER
 ; ###########################
 
 ; run with gunicorn --paste rhodecode.ini --config gunicorn_conf.py
 
 ; Module to use, this setting shouldn't be changed
 use = egg:gunicorn#main
 
 ; Prefix middleware for RhodeCode.
 ; recommended when using proxy setup.
 ; allows to set RhodeCode under a prefix in server.
 ; eg https://server.com/custom_prefix. Enable `filter-with =` option below as well.
 ; And set your prefix like: `prefix = /custom_prefix`
 ; be sure to also set beaker.session.cookie_path = /custom_prefix if you need
 ; to make your cookies only work on prefix url
 [filter:proxy-prefix]
 use = egg:PasteDeploy#prefix
 prefix = /
 
 [app:main]
 ; The %(here)s variable will be replaced with the absolute path of parent directory
 ; of this file
 ; Each option in the app:main can be override by an environmental variable
 ;
 ;To override an option:
 ;
 ;RC_<KeyName>
 ;Everything should be uppercase, . and - should be replaced by _.
 ;For example, if you have these configuration settings:
 ;rc_cache.repo_object.backend = foo
 ;can be overridden by
 ;export RC_CACHE_REPO_OBJECT_BACKEND=foo
 
 use = egg:rhodecode-enterprise-ce
 
 ; enable proxy prefix middleware, defined above
 #filter-with = proxy-prefix
 
 ; encryption key used to encrypt social plugin tokens,
 ; remote_urls with credentials etc, if not set it defaults to
 ; `beaker.session.secret`
 #rhodecode.encrypted_values.secret =
 
 ; decryption strict mode (enabled by default). It controls if decryption raises
 ; `SignatureVerificationError` in case of wrong key, or damaged encryption data.
 #rhodecode.encrypted_values.strict = false
 
 ; Pick algorithm for encryption. Either fernet (more secure) or aes (default)
 ; fernet is safer, and we strongly recommend switching to it.
 ; Due to backward compatibility aes is used as default.
 #rhodecode.encrypted_values.algorithm = fernet
 
 ; Return gzipped responses from RhodeCode (static files/application)
 gzip_responses = false
 
 ; Auto-generate javascript routes file on startup
 generate_js_files = false
 
 ; System global default language.
 ; All available languages: en (default), be, de, es, fr, it, ja, pl, pt, ru, zh
 lang = en
 
 ; Perform a full repository scan and import on each server start.
 ; Settings this to true could lead to very long startup time.
 startup.import_repos = true
 
 ; URL at which the application is running. This is used for Bootstrapping
 ; requests in context when no web request is available. Used in ishell, or
 ; SSH calls. Set this for events to receive proper url for SSH calls.
 app.base_url = http://rhodecode.local
 
 ; Host at which the Service API is running.
 app.service_api.host = http://rhodecode.local:10020
 
 ; Secret for Service API authentication.
 app.service_api.token =
 
 ; Unique application ID. Should be a random unique string for security.
 app_instance_uuid = rc-production
 
 ; Cut off limit for large diffs (size in bytes). If overall diff size on
 ; commit, or pull request exceeds this limit this diff will be displayed
 ; partially. E.g 512000 == 512Kb
 cut_off_limit_diff = 1024000
 
 ; Cut off limit for large files inside diffs (size in bytes). Each individual
 ; file inside diff which exceeds this limit will be displayed partially.
 ;  E.g 128000 == 128Kb
 cut_off_limit_file = 256000
 
 ; Use cached version of vcs repositories everywhere. Recommended to be `true`
 vcs_full_cache = false
 
 ; Force https in RhodeCode, fixes https redirects, assumes it's always https.
 ; Normally this is controlled by proper flags sent from http server such as Nginx or Apache
 force_https = false
 
 ; use Strict-Transport-Security headers
 use_htsts = false
 
 ; Set to true if your repos are exposed using the dumb protocol
 git_update_server_info = false
 
 ; RSS/ATOM feed options
 rss_cut_off_limit = 256000
 rss_items_per_page = 10
 rss_include_diff = false
 
 ; gist URL alias, used to create nicer urls for gist. This should be an
 ; url that does rewrites to _admin/gists/{gistid}.
 ; example: http://gist.rhodecode.org/{gistid}. Empty means use the internal
 ; RhodeCode url, ie. http[s]://rhodecode.server/_admin/gists/{gistid}
 gist_alias_url =
 
 ; List of views (using glob pattern syntax) that AUTH TOKENS could be
 ; used for access.
 ; Adding ?auth_token=TOKEN_HASH to the url authenticates this request as if it
 ; came from the the logged in user who own this authentication token.
 ; Additionally @TOKEN syntax can be used to bound the view to specific
 ; authentication token. Such view would be only accessible when used together
 ; with this authentication token
 ; list of all views can be found under `/_admin/permissions/auth_token_access`
 ; The list should be "," separated and on a single line.
 ; Most common views to enable:
 
 #    RepoCommitsView:repo_commit_download
 #    RepoCommitsView:repo_commit_patch
 #    RepoCommitsView:repo_commit_raw
 #    RepoCommitsView:repo_commit_raw@TOKEN
 #    RepoFilesView:repo_files_diff
 #    RepoFilesView:repo_archivefile
 #    RepoFilesView:repo_file_raw
 #    GistView:*
 api_access_controllers_whitelist =
 
 ; Default encoding used to convert from and to unicode
 ; can be also a comma separated list of encoding in case of mixed encodings
 default_encoding = UTF-8
 
 ; instance-id prefix
 ; a prefix key for this instance used for cache invalidation when running
 ; multiple instances of RhodeCode, make sure it's globally unique for
 ; all running RhodeCode instances. Leave empty if you don't use it
 instance_id =
 
 ; Fallback authentication plugin. Set this to a plugin ID to force the usage
 ; of an authentication plugin also if it is disabled by it's settings.
 ; This could be useful if you are unable to log in to the system due to broken
 ; authentication settings. Then you can enable e.g. the internal RhodeCode auth
 ; module to log in again and fix the settings.
 ; Available builtin plugin IDs (hash is part of the ID):
 ; egg:rhodecode-enterprise-ce#rhodecode
 ; egg:rhodecode-enterprise-ce#pam
 ; egg:rhodecode-enterprise-ce#ldap
 ; egg:rhodecode-enterprise-ce#jasig_cas
 ; egg:rhodecode-enterprise-ce#headers
 ; egg:rhodecode-enterprise-ce#crowd
 
 #rhodecode.auth_plugin_fallback = egg:rhodecode-enterprise-ce#rhodecode
 
 ; Flag to control loading of legacy plugins in py:/path format
 auth_plugin.import_legacy_plugins = true
 
 ; alternative return HTTP header for failed authentication. Default HTTP
 ; response is 401 HTTPUnauthorized. Currently HG clients have troubles with
 ; handling that causing a series of failed authentication calls.
 ; Set this variable to 403 to return HTTPForbidden, or any other HTTP code
 ; This will be served instead of default 401 on bad authentication
 auth_ret_code =
 
 ; use special detection method when serving auth_ret_code, instead of serving
 ; ret_code directly, use 401 initially (Which triggers credentials prompt)
 ; and then serve auth_ret_code to clients
 auth_ret_code_detection = false
 
 ; locking return code. When repository is locked return this HTTP code. 2XX
 ; codes don't break the transactions while 4XX codes do
 lock_ret_code = 423
 
 ; Filesystem location were repositories should be stored
 repo_store.path = /var/opt/rhodecode_repo_store
 
 ; allows to setup custom hooks in settings page
 allow_custom_hooks_settings = true
 
 ; Generated license token required for EE edition license.
 ; New generated token value can be found in Admin > settings > license page.
 license_token = abra-cada-bra1-rce3
 
 ; This flag hides sensitive information on the license page such as token, and license data
 license.hide_license_info = false
 
 ; supervisor connection uri, for managing supervisor and logs.
 supervisor.uri =
 
 ; supervisord group name/id we only want this RC instance to handle
 supervisor.group_id = dev
 
 ; Display extended labs settings
 labs_settings_active = true
 
 ; Custom exception store path, defaults to TMPDIR
 ; This is used to store exception from RhodeCode in shared directory
 #exception_tracker.store_path =
 
 ; Send email with exception details when it happens
 #exception_tracker.send_email = false
 
 ; Comma separated list of recipients for exception emails,
 ; e.g [e-mail],[e-mail]
 ; Can be left empty, then emails will be sent to ALL super-admins
 #exception_tracker.send_email_recipients =
 
 ; optional prefix to Add to email Subject
 #exception_tracker.email_prefix = [RHODECODE ERROR]
 
 ; File store configuration. This is used to store and serve uploaded files
 file_store.enabled = true
 
 ; Storage backend, available options are: local
 file_store.backend = local
 
 ; path to store the uploaded binaries and artifacts
 file_store.storage_path = /var/opt/rhodecode_data/file_store
 
-; Uncomment and set this path to control settings for archive download cache.
+
+; Redis url to acquire/check generation of archives locks
+archive_cache.locking.url = redis://redis:6379/1
+
+; Storage backend, only 'filesystem' and 'objectstore' are available now
+archive_cache.backend.type = filesystem
+
+; url for s3 compatible storage that allows to upload artifacts
+; e.g http://minio:9000
+archive_cache.objectstore.url = http://s3-minio:9000
+
+; key for s3 auth
+archive_cache.objectstore.key = key
+
+; secret for s3 auth
+archive_cache.objectstore.secret = secret
+
+; number of sharded buckets to create to distribute archives across
+; default is 8 shards
+archive_cache.objectstore.bucket_shards = 8
+
+; if true, this cache will try to retry with retry_attempts=N times waiting retry_backoff time
+archive_cache.objectstore.retry = false
+
+; number of seconds to wait for next try using retry
+archive_cache.objectstore.retry_backoff = 1
+
+; how many tries do do a retry fetch from this backend
+archive_cache.objectstore.retry_attempts = 10
+
+; Default is $cache_dir/archive_cache if not set
 ; Generated repo archives will be cached at this location
 ; and served from the cache during subsequent requests for the same archive of
 ; the repository. This path is important to be shared across filesystems and with
 ; RhodeCode and vcsserver
-
-; Default is $cache_dir/archive_cache if not set
-archive_cache.store_dir = /var/opt/rhodecode_data/tarballcache
+archive_cache.filesystem.store_dir = %(here)s/rc-tests/archive_cache
 
 ; The limit in GB sets how much data we cache before recycling last used, defaults to 10 gb
-archive_cache.cache_size_gb = 10
+archive_cache.filesystem.cache_size_gb = 2
+
+; Eviction policy used to clear out after cache_size_gb limit is reached
+archive_cache.filesystem.eviction_policy = least-recently-stored
 
 ; By default cache uses sharding technique, this specifies how many shards are there
-archive_cache.cache_shards = 10
+; default is 8 shards
+archive_cache.filesystem.cache_shards = 8
+
+; if true, this cache will try to retry with retry_attempts=N times waiting retry_backoff time
+archive_cache.filesystem.retry = false
+
+; number of seconds to wait for next try using retry
+archive_cache.filesystem.retry_backoff = 1
+
+; how many tries do do a retry fetch from this backend
+archive_cache.filesystem.retry_attempts = 10
+
 
 ; #############
 ; CELERY CONFIG
 ; #############
 
 ; manually run celery: /path/to/celery worker --task-events --beat --app rhodecode.lib.celerylib.loader --scheduler rhodecode.lib.celerylib.scheduler.RcScheduler --loglevel DEBUG --ini /path/to/rhodecode.ini
 
 use_celery = false
 
 ; path to store schedule database
 #celerybeat-schedule.path =
 
 ; connection url to the message broker (default redis)
 celery.broker_url = redis://redis:6379/8
 
 ; results backend to get results for (default redis)
 celery.result_backend = redis://redis:6379/8
 
 ; rabbitmq example
 #celery.broker_url = amqp://rabbitmq:qweqwe@localhost:5672/rabbitmqhost
 
 ; maximum tasks to execute before worker restart
 celery.max_tasks_per_child = 20
 
 ; tasks will never be sent to the queue, but executed locally instead.
 celery.task_always_eager = true
 celery.task_store_eager_result = true
 
 ; #############
 ; DOGPILE CACHE
 ; #############
 
 ; Default cache dir for caches. Putting this into a ramdisk can boost performance.
 ; eg. /tmpfs/data_ramdisk, however this directory might require large amount of space
 cache_dir = /var/opt/rhodecode_data
 
 ; *********************************************
 ; `sql_cache_short` cache for heavy SQL queries
 ; Only supported backend is `memory_lru`
 ; *********************************************
 rc_cache.sql_cache_short.backend = dogpile.cache.rc.memory_lru
 rc_cache.sql_cache_short.expiration_time = 0
 
 
 ; *****************************************************
 ; `cache_repo_longterm` cache for repo object instances
 ; Only supported backend is `memory_lru`
 ; *****************************************************
 rc_cache.cache_repo_longterm.backend = dogpile.cache.rc.memory_lru
 ; by default we use 30 Days, cache is still invalidated on push
 rc_cache.cache_repo_longterm.expiration_time = 2592000
 ; max items in LRU cache, set to smaller number to save memory, and expire last used caches
 rc_cache.cache_repo_longterm.max_size = 10000
 
 
 ; *********************************************
 ; `cache_general` cache for general purpose use
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *********************************************
 rc_cache.cache_general.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_general.expiration_time = 43200
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
-rc_cache.cache_general.arguments.filename = %(here)s/cache-backend/cache_general_db
+rc_cache.cache_general.arguments.filename = %(here)s/rc-tests/cache-backend/cache_general_db
 
 ; alternative `cache_general` redis backend with distributed lock
 #rc_cache.cache_general.backend = dogpile.cache.rc.redis
 #rc_cache.cache_general.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_general.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_general.arguments.host = localhost
 #rc_cache.cache_general.arguments.port = 6379
 #rc_cache.cache_general.arguments.db = 0
 #rc_cache.cache_general.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_general.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_general.arguments.lock_auto_renewal = true
 
 ; *************************************************
 ; `cache_perms` cache for permission tree, auth TTL
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *************************************************
 rc_cache.cache_perms.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_perms.expiration_time = 0
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
-rc_cache.cache_perms.arguments.filename = %(here)s/cache-backend/cache_perms_db
+rc_cache.cache_perms.arguments.filename = %(here)s/rc-tests/cache-backend/cache_perms_db
 
 ; alternative `cache_perms` redis backend with distributed lock
 #rc_cache.cache_perms.backend = dogpile.cache.rc.redis
 #rc_cache.cache_perms.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_perms.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_perms.arguments.host = localhost
 #rc_cache.cache_perms.arguments.port = 6379
 #rc_cache.cache_perms.arguments.db = 0
 #rc_cache.cache_perms.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_perms.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_perms.arguments.lock_auto_renewal = true
 
 ; ***************************************************
 ; `cache_repo` cache for file tree, Readme, RSS FEEDS
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; ***************************************************
 rc_cache.cache_repo.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_repo.expiration_time = 2592000
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
-rc_cache.cache_repo.arguments.filename = %(here)s/cache-backend/cache_repo_db
+rc_cache.cache_repo.arguments.filename = %(here)s/rc-tests/cache-backend/cache_repo_db
 
 ; alternative `cache_repo` redis backend with distributed lock
 #rc_cache.cache_repo.backend = dogpile.cache.rc.redis
 #rc_cache.cache_repo.expiration_time = 2592000
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_repo.arguments.redis_expiration_time = 2678400
 
 #rc_cache.cache_repo.arguments.host = localhost
 #rc_cache.cache_repo.arguments.port = 6379
 #rc_cache.cache_repo.arguments.db = 1
 #rc_cache.cache_repo.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_repo.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_repo.arguments.lock_auto_renewal = true
 
 ; ##############
 ; BEAKER SESSION
 ; ##############
 
 ; beaker.session.type is type of storage options for the logged users sessions. Current allowed
 ; types are file, ext:redis, ext:database, ext:memcached
 ; Fastest ones are ext:redis and ext:database, DO NOT use memory type for session
 beaker.session.type = file
 beaker.session.data_dir = %(here)s/rc-tests/data/sessions
 
 ; Redis based sessions
 #beaker.session.type = ext:redis
 #beaker.session.url = redis://redis:6379/2
 
 ; DB based session, fast, and allows easy management over logged in users
 #beaker.session.type = ext:database
 #beaker.session.table_name = db_session
 #beaker.session.sa.url = postgresql://postgres:secret@localhost/rhodecode
 #beaker.session.sa.url = mysql://root:secret@127.0.0.1/rhodecode
 #beaker.session.sa.pool_recycle = 3600
 #beaker.session.sa.echo = false
 
 beaker.session.key = rhodecode
 beaker.session.secret = test-rc-uytcxaz
-beaker.session.lock_dir = %(here)s/data/sessions/lock
+beaker.session.lock_dir = %(here)s/rc-tests/data/sessions/lock
 
 ; Secure encrypted cookie. Requires AES and AES python libraries
 ; you must disable beaker.session.secret to use this
 #beaker.session.encrypt_key = key_for_encryption
 #beaker.session.validate_key = validation_key
 
 ; Sets session as invalid (also logging out user) if it haven not been
 ; accessed for given amount of time in seconds
 beaker.session.timeout = 2592000
 beaker.session.httponly = true
 
 ; Path to use for the cookie. Set to prefix if you use prefix middleware
 #beaker.session.cookie_path = /custom_prefix
 
 ; Set https secure cookie
 beaker.session.secure = false
 
 ; default cookie expiration time in seconds, set to `true` to set expire
 ; at browser close
 #beaker.session.cookie_expires = 3600
 
 ; #############################
 ; SEARCH INDEXING CONFIGURATION
 ; #############################
 
 ; Full text search indexer is available in rhodecode-tools under
 ; `rhodecode-tools index` command
 
 ; WHOOSH Backend, doesn't require additional services to run
 ; it works good with few dozen repos
 search.module = rhodecode.lib.index.whoosh
-search.location = %(here)s/data/index
+search.location = %(here)s/rc-tests/data/index
 
 ; ####################
 ; CHANNELSTREAM CONFIG
 ; ####################
 
 ; channelstream enables persistent connections and live notification
 ; in the system. It's also used by the chat system
 
 channelstream.enabled = false
 
 ; server address for channelstream server on the backend
 channelstream.server = channelstream:9800
 
 ; location of the channelstream server from outside world
 ; use ws:// for http or wss:// for https. This address needs to be handled
 ; by external HTTP server such as Nginx or Apache
 ; see Nginx/Apache configuration examples in our docs
 channelstream.ws_url = ws://rhodecode.yourserver.com/_channelstream
 channelstream.secret = ENV_GENERATED
-channelstream.history.location = %(here)s/channelstream_history
+channelstream.history.location = %(here)s/rc-tests/channelstream_history
 
 ; Internal application path that Javascript uses to connect into.
 ; If you use proxy-prefix the prefix should be added before /_channelstream
 channelstream.proxy_path = /_channelstream
 
 
 ; ##############################
 ; MAIN RHODECODE DATABASE CONFIG
 ; ##############################
 
 #sqlalchemy.db1.url = sqlite:///%(here)s/rhodecode.db?timeout=30
 #sqlalchemy.db1.url = postgresql://postgres:qweqwe@localhost/rhodecode
 #sqlalchemy.db1.url = mysql://root:qweqwe@localhost/rhodecode?charset=utf8
 ; pymysql is an alternative driver for MySQL, use in case of problems with default one
 #sqlalchemy.db1.url = mysql+pymysql://root:qweqwe@localhost/rhodecode
 
-sqlalchemy.db1.url = sqlite:///%(here)s/rhodecode_test.db?timeout=30
+sqlalchemy.db1.url = sqlite:///%(here)s/rc-tests/rhodecode_test.db?timeout=30
 
 ; see sqlalchemy docs for other advanced settings
 ; print the sql statements to output
 sqlalchemy.db1.echo = false
 
 ; recycle the connections after this amount of seconds
 sqlalchemy.db1.pool_recycle = 3600
 
 ; the number of connections to keep open inside the connection pool.
 ; 0 indicates no limit
 ; the general calculus with gevent is:
 ; if your system allows 500 concurrent greenlets (max_connections) that all do database access,
 ; then increase pool size + max overflow so that they add up to 500.
 #sqlalchemy.db1.pool_size = 5
 
 ; The number of connections to allow in connection pool "overflow", that is
 ; connections that can be opened above and beyond the pool_size setting,
 ; which defaults to five.
 #sqlalchemy.db1.max_overflow = 10
 
 ; Connection check ping, used to detect broken database connections
 ; could be enabled to better handle cases if MySQL has gone away errors
 #sqlalchemy.db1.ping_connection = true
 
 ; ##########
 ; VCS CONFIG
 ; ##########
 vcs.server.enable = true
 vcs.server = vcsserver:10010
 
 ; Web server connectivity protocol, responsible for web based VCS operations
 ; Available protocols are:
 ; `http` - use http-rpc backend (default)
 vcs.server.protocol = http
 
 ; Push/Pull operations protocol, available options are:
 ; `http` - use http-rpc backend (default)
 vcs.scm_app_implementation = http
 
 ; Push/Pull operations hooks protocol, available options are:
 ; `http` - use http-rpc backend (default)
 ; `celery` - use celery based hooks
 vcs.hooks.protocol = http
 
 ; Host on which this instance is listening for hooks. vcsserver will call this host to pull/push hooks so it should be
 ; accessible via network.
 ; Use vcs.hooks.host = "*" to bind to current hostname (for Docker)
 vcs.hooks.host = *
 
 ; Start VCSServer with this instance as a subprocess, useful for development
 vcs.start_server = false
 
 ; List of enabled VCS backends, available options are:
 ; `hg`  - mercurial
 ; `git` - git
 ; `svn` - subversion
 vcs.backends = hg, git, svn
 
 ; Wait this number of seconds before killing connection to the vcsserver
 vcs.connection_timeout = 3600
 
 ; Cache flag to cache vcsserver remote calls locally
 ; It uses cache_region `cache_repo`
 vcs.methods.cache = false
 
 ; ####################################################
 ; Subversion proxy support (mod_dav_svn)
 ; Maps RhodeCode repo groups into SVN paths for Apache
 ; ####################################################
 
 ; Compatibility version when creating SVN repositories. Defaults to newest version when commented out.
 ; Set a numeric version for your current SVN e.g 1.8, or 1.12
 ; Legacy available options are: pre-1.4-compatible, pre-1.5-compatible, pre-1.6-compatible, pre-1.8-compatible, pre-1.9-compatible
 #vcs.svn.compatible_version = 1.8
 
 ; Enable SVN proxy of requests over HTTP
 vcs.svn.proxy.enabled = true
 
 ; host to connect to running SVN subsystem
 vcs.svn.proxy.host = http://svn:8090
 
 ; Enable or disable the config file generation.
 svn.proxy.generate_config = false
 
 ; Generate config file with `SVNListParentPath` set to `On`.
 svn.proxy.list_parent_path = true
 
 ; Set location and file name of generated config file.
-svn.proxy.config_file_path = %(here)s/mod_dav_svn.conf
+svn.proxy.config_file_path = %(here)s/rc-tests/mod_dav_svn.conf
 
 ; alternative mod_dav config template. This needs to be a valid mako template
 ; Example template can be found in the source code:
 ; rhodecode/apps/svn_support/templates/mod-dav-svn.conf.mako
 #svn.proxy.config_template = ~/.rccontrol/enterprise-1/custom_svn_conf.mako
 
 ; Used as a prefix to the `Location` block in the generated config file.
 ; In most cases it should be set to `/`.
 svn.proxy.location_root = /
 
 ; Command to reload the mod dav svn configuration on change.
 ; Example: `/etc/init.d/apache2 reload` or /home/USER/apache_reload.sh
 ; Make sure user who runs RhodeCode process is allowed to reload Apache
 #svn.proxy.reload_cmd = /etc/init.d/apache2 reload
 
 ; If the timeout expires before the reload command finishes, the command will
 ; be killed. Setting it to zero means no timeout. Defaults to 10 seconds.
 #svn.proxy.reload_timeout = 10
 
 ; ####################
 ; SSH Support Settings
 ; ####################
 
 ; Defines if a custom authorized_keys file should be created and written on
 ; any change user ssh keys. Setting this to false also disables possibility
 ; of adding SSH keys by users from web interface. Super admins can still
 ; manage SSH Keys.
 ssh.generate_authorized_keyfile = true
 
 ; Options for ssh, default is `no-pty,no-port-forwarding,no-X11-forwarding,no-agent-forwarding`
 # ssh.authorized_keys_ssh_opts =
 
 ; Path to the authorized_keys file where the generate entries are placed.
 ; It is possible to have multiple key files specified in `sshd_config` e.g.
 ; AuthorizedKeysFile %h/.ssh/authorized_keys %h/.ssh/authorized_keys_rhodecode
-ssh.authorized_keys_file_path = %(here)s/rc/authorized_keys_rhodecode
+ssh.authorized_keys_file_path = %(here)s/rc-tests/authorized_keys_rhodecode
 
 ; Command to execute the SSH wrapper. The binary is available in the
 ; RhodeCode installation directory.
 ; legacy: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 ; new rewrite: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper-v2
 ssh.wrapper_cmd = /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 
 ; Allow shell when executing the ssh-wrapper command
 ssh.wrapper_cmd_allow_shell = false
 
 ; Enables logging, and detailed output send back to the client during SSH
 ; operations. Useful for debugging, shouldn't be used in production.
 ssh.enable_debug_logging = false
 
 ; Paths to binary executable, by default they are the names, but we can
 ; override them if we want to use a custom one
 ssh.executable.hg = /usr/local/bin/rhodecode_bin/vcs_bin/hg
 ssh.executable.git = /usr/local/bin/rhodecode_bin/vcs_bin/git
 ssh.executable.svn = /usr/local/bin/rhodecode_bin/vcs_bin/svnserve
 
 ; Enables SSH key generator web interface. Disabling this still allows users
 ; to add their own keys.
 ssh.enable_ui_key_generator = true
 
 ; Statsd client config, this is used to send metrics to statsd
 ; We recommend setting statsd_exported and scrape them using Prometheus
 #statsd.enabled = false
 #statsd.statsd_host = 0.0.0.0
 #statsd.statsd_port = 8125
 #statsd.statsd_prefix =
 #statsd.statsd_ipv6 = false
 
 ; configure logging automatically at server startup set to false
 ; to use the below custom logging config.
 ; RC_LOGGING_FORMATTER
 ; RC_LOGGING_LEVEL
 ; env variables can control the settings for logging in case of autoconfigure
 
 logging.autoconfigure = false
 
 ; specify your own custom logging config file to configure logging
 #logging.logging_conf_file = /path/to/custom_logging.ini
 
 ; Dummy marker to add new entries after.
 ; Add any custom entries below. Please don't remove this marker.
 custom.conf = 1
 
 
 ; #####################
 ; LOGGING CONFIGURATION
 ; #####################
 
 [loggers]
 keys = root, sqlalchemy, beaker, celery, rhodecode, ssh_wrapper, dogpile
 
 [handlers]
 keys = console, console_sql
 
 [formatters]
 keys = generic, json, color_formatter, color_formatter_sql
 
 ; #######
 ; LOGGERS
 ; #######
 [logger_root]
 level = NOTSET
 handlers = console
 
 [logger_routes]
 level = DEBUG
 handlers =
 qualname = routes.middleware
 ## "level = DEBUG" logs the route matched and routing variables.
 propagate = 1
 
 [logger_sqlalchemy]
 level = INFO
 handlers = console_sql
 qualname = sqlalchemy.engine
 propagate = 0
 
 [logger_beaker]
 level = DEBUG
 handlers =
 qualname = beaker.container
 propagate = 1
 
 [logger_dogpile]
 level = INFO
 handlers = console
 qualname = dogpile
 propagate = 1
 
 [logger_rhodecode]
 level = DEBUG
 handlers =
 qualname = rhodecode
 propagate = 1
 
 [logger_ssh_wrapper]
 level = DEBUG
 handlers =
 qualname = ssh_wrapper
 propagate = 1
 
 [logger_celery]
 level = DEBUG
 handlers =
 qualname = celery
 
 
 ; ########
 ; HANDLERS
 ; ########
 
 [handler_console]
 class = StreamHandler
 args = (sys.stderr, )
 level = DEBUG
 ; To enable JSON formatted logs replace 'generic/color_formatter' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = generic
 
 [handler_console_sql]
 ; "level = DEBUG" logs SQL queries and results.
 ; "level = INFO" logs SQL queries.
 ; "level = WARN" logs neither.  (Recommended for production systems.)
 class = StreamHandler
 args = (sys.stderr, )
 level = WARN
 ; To enable JSON formatted logs replace 'generic/color_formatter_sql' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = generic
 
 ; ##########
 ; FORMATTERS
 ; ##########
 
 [formatter_generic]
 class = rhodecode.lib.logging_formatter.ExceptionAwareFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter]
 class = rhodecode.lib.logging_formatter.ColorFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter_sql]
 class = rhodecode.lib.logging_formatter.ColorFormatterSql
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_json]
 format = %(timestamp)s %(levelname)s %(name)s %(message)s %(req_id)s
 class = rhodecode.lib._vendor.jsonlogger.JsonFormatter

rhodecode/tests/vcs/test_archives.py

@@ -1,198 +1,198 b''
 
 # Copyright (C) 2010-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import datetime
 import os
 import shutil
 import tarfile
 import zipfile
 import io
 
 import mock
 import pytest
 
 import rhodecode
-from rhodecode.lib.rc_cache.archive_cache import get_archival_config
+from rhodecode.lib.archive_cache import get_archival_config
 from rhodecode.lib.str_utils import ascii_bytes
 from rhodecode.lib.vcs.backends import base
 from rhodecode.lib.vcs.exceptions import ImproperArchiveTypeError, VCSError
 from rhodecode.lib.vcs.nodes import FileNode
 from rhodecode.tests.vcs.conftest import BackendTestMixin
 
 
 @pytest.fixture()
 def d_cache_config():
     return get_archival_config(config=rhodecode.CONFIG)
 
 
 @pytest.mark.usefixtures("vcs_repository_support")
 class TestArchives(BackendTestMixin):
 
     @classmethod
     def _get_commits(cls):
         start_date = datetime.datetime(2010, 1, 1, 20)
         yield {
             'message': 'Initial Commit',
             'author': 'Joe Doe <[e-mail]>',
             'date': start_date + datetime.timedelta(hours=12),
             'added': [
                 FileNode(b'executable_0o100755', b'mode_755', mode=0o100755),
                 FileNode(b'executable_0o100500', b'mode_500', mode=0o100500),
                 FileNode(b'not_executable', b'mode_644', mode=0o100644),
             ],
         }
         for x in range(5):
             yield {
                 'message': 'Commit %d' % x,
                 'author': 'Joe Doe <[e-mail]>',
                 'date': start_date + datetime.timedelta(hours=12 * x),
                 'added': [
                     FileNode(b'%d/file_%d.txt' % (x, x), content=b'Foobar %d' % x),
                 ],
             }
 
     @pytest.mark.parametrize('compressor', ['gz', 'bz2'])
     def test_archive_tar(self, compressor, tmpdir, tmp_path, d_cache_config):
 
         archive_node = tmp_path / 'archive-node'
         archive_node.touch()
 
         archive_lnk = self.tip.archive_repo(
             str(archive_node), kind=f't{compressor}', archive_dir_name='repo', cache_config=d_cache_config)
 
         out_dir = tmpdir
         out_file = tarfile.open(str(archive_lnk), f'r|{compressor}')
         out_file.extractall(out_dir)
         out_file.close()
 
         for x in range(5):
             node_path = '%d/file_%d.txt' % (x, x)
             with open(os.path.join(out_dir, 'repo/' + node_path), 'rb') as f:
                 file_content = f.read()
             assert file_content == self.tip.get_node(node_path).content
 
         shutil.rmtree(out_dir)
 
     @pytest.mark.parametrize('compressor', ['gz', 'bz2'])
     def test_archive_tar_symlink(self, compressor):
         pytest.skip('Not supported')
 
     @pytest.mark.parametrize('compressor', ['gz', 'bz2'])
     def test_archive_tar_file_modes(self, compressor, tmpdir, tmp_path, d_cache_config):
         archive_node = tmp_path / 'archive-node'
         archive_node.touch()
 
         archive_lnk = self.tip.archive_repo(
             str(archive_node), kind='t{}'.format(compressor), archive_dir_name='repo', cache_config=d_cache_config)
 
         out_dir = tmpdir
         out_file = tarfile.open(str(archive_lnk), 'r|{}'.format(compressor))
         out_file.extractall(out_dir)
         out_file.close()
 
         def dest(inp):
             return os.path.join(out_dir, "repo/" + inp)
 
         assert oct(os.stat(dest('not_executable')).st_mode) == '0o100644'
 
     def test_archive_zip(self, tmp_path, d_cache_config):
         archive_node = tmp_path / 'archive-node'
         archive_node.touch()
 
         archive_lnk = self.tip.archive_repo(str(archive_node), kind='zip',
                                             archive_dir_name='repo', cache_config=d_cache_config)
         zip_file = zipfile.ZipFile(str(archive_lnk))
 
         for x in range(5):
             node_path = '%d/file_%d.txt' % (x, x)
             data = zip_file.read(f'repo/{node_path}')
 
             decompressed = io.BytesIO()
             decompressed.write(data)
             assert decompressed.getvalue() == \
                 self.tip.get_node(node_path).content
             decompressed.close()
 
     def test_archive_zip_with_metadata(self, tmp_path, d_cache_config):
         archive_node = tmp_path / 'archive-node'
         archive_node.touch()
 
         archive_lnk = self.tip.archive_repo(
             str(archive_node), kind='zip',
             archive_dir_name='repo', write_metadata=True, cache_config=d_cache_config)
 
         zip_file = zipfile.ZipFile(str(archive_lnk))
         metafile = zip_file.read('repo/.archival.txt')
 
         raw_id = ascii_bytes(self.tip.raw_id)
         assert b'commit_id:%b' % raw_id in metafile
 
         for x in range(5):
             node_path = '%d/file_%d.txt' % (x, x)
             data = zip_file.read(f'repo/{node_path}')
             decompressed = io.BytesIO()
             decompressed.write(data)
             assert decompressed.getvalue() == \
                 self.tip.get_node(node_path).content
             decompressed.close()
 
     def test_archive_wrong_kind(self, tmp_path, d_cache_config):
         archive_node = tmp_path / 'archive-node'
         archive_node.touch()
 
         with pytest.raises(ImproperArchiveTypeError):
             self.tip.archive_repo(str(archive_node), kind='wrong kind', cache_config=d_cache_config)
 
 
 @pytest.fixture()
 def base_commit():
     """
     Prepare a `base.BaseCommit` just enough for `_validate_archive_prefix`.
     """
     commit = base.BaseCommit()
     commit.repository = mock.Mock()
     commit.repository.name = 'fake_repo'
     commit.short_id = 'fake_id'
     return commit
 
 
 def test_validate_archive_prefix_enforces_non_ascii_as_prefix(base_commit):
     with pytest.raises(VCSError):
         base_commit._validate_archive_prefix("Ünïcödë")
 
 
 def test_validate_archive_prefix_empty_prefix(base_commit):
     # TODO: johbo: Should raise a ValueError here.
     with pytest.raises(VCSError):
         base_commit._validate_archive_prefix('')
 
 
 def test_validate_archive_prefix_with_leading_slash(base_commit):
     # TODO: johbo: Should raise a ValueError here.
     with pytest.raises(VCSError):
         base_commit._validate_archive_prefix('/any')
 
 
 def test_validate_archive_prefix_falls_back_to_repository_name(base_commit):
     prefix = base_commit._validate_archive_prefix(None)
     expected_prefix = base_commit._ARCHIVE_PREFIX_TEMPLATE.format(
         repo_name='fake_repo',
         short_id='fake_id')
     assert isinstance(prefix, str)
     assert prefix == expected_prefix