vcs: change way refs are retrieved for git so same name branch/tags

and remotes can be supported, fixes  #298

rhodecode/lib/vcs/backends/git/inmemory.py

@@ -1,105 +1,105 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2014-2016  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/
 
 """
 GIT inmemory module
 """
 
 from rhodecode.lib.datelib import date_to_timestamp_plus_offset
 from rhodecode.lib.utils import safe_str
 from rhodecode.lib.vcs.backends import base
 
 
 class GitInMemoryCommit(base.BaseInMemoryCommit):
 
     def commit(self, message, author, parents=None, branch=None, date=None,
                **kwargs):
         """
         Performs in-memory commit (doesn't check workdir in any way) and
         returns newly created `GitCommit`. Updates repository's
         `commit_ids`.
 
         :param message: message of the commit
         :param author: full username, i.e. "Joe Doe <[e-mail]>"
         :param parents: single parent or sequence of parents from which commit
           would be derived
         :param date: `datetime.datetime` instance. Defaults to
           ``datetime.datetime.now()``.
         :param branch: branch name, as string. If none given, default backend's
           branch would be used.
 
         :raises `CommitError`: if any error occurs while committing
         """
         self.check_integrity(parents)
         if branch is None:
             branch = self.repository.DEFAULT_BRANCH_NAME
 
         ENCODING = "UTF-8"
 
         commit_tree = None
         if self.parents[0]:
             commit_tree = self.parents[0]._commit['tree']
 
         updated = []
         for node in self.added + self.changed:
             if not node.is_binary:
                 content = node.content.encode(ENCODING)
             else:
                 content = node.content
             updated.append({
                 'path': node.path,
                 'node_path': node.name.encode(ENCODING),
                 'content': content,
                 'mode': node.mode,
             })
 
         removed = [node.path for node in self.removed]
 
         date, tz = date_to_timestamp_plus_offset(date)
 
         # TODO: johbo: Make kwargs explicit and check if this is needed.
         author_time = kwargs.pop('author_time', date)
         author_tz = kwargs.pop('author_timezone', tz)
 
         commit_data = {
             'parents': [p._commit['id'] for p in self.parents if p],
             'author': safe_str(author),
             'committer': safe_str(author),
             'encoding': ENCODING,
             'message': safe_str(message),
             'commit_time': int(date),
             'author_time': int(author_time),
             'commit_timezone': tz,
             'author_timezone': author_tz,
         }
 
         commit_id = self.repository._remote.commit(
             commit_data, branch, commit_tree, updated, removed)
 
         # Update vcs repository object
         self.repository.commit_ids.append(commit_id)
         self.repository._rebuild_cache(self.repository.commit_ids)
 
         # invalidate parsed refs after commit
-        self.repository._parsed_refs = self.repository._get_parsed_refs()
+        self.repository._refs = self.repository._get_refs()
         self.repository.branches = self.repository._get_branches()
         tip = self.repository.get_commit()
         self.reset()
         return tip

rhodecode/lib/vcs/backends/git/repository.py

@@ -1,911 +1,922 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2014-2016  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/
 
 """
 GIT repository module
 """
 
 import logging
 import os
 import re
 import shutil
 import time
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 from rhodecode.lib.compat import OrderedDict
 from rhodecode.lib.datelib import makedate, utcdate_fromtimestamp
 from rhodecode.lib.utils import safe_unicode, safe_str
 from rhodecode.lib.vcs import connection, path as vcspath
 from rhodecode.lib.vcs.backends.base import (
     BaseRepository, CollectionGenerator, Config, MergeResponse,
     MergeFailureReason)
 from rhodecode.lib.vcs.backends.git.commit import GitCommit
 from rhodecode.lib.vcs.backends.git.diff import GitDiff
 from rhodecode.lib.vcs.backends.git.inmemory import GitInMemoryCommit
 from rhodecode.lib.vcs.conf import settings
 from rhodecode.lib.vcs.exceptions import (
     CommitDoesNotExistError, EmptyRepositoryError,
     RepositoryError, TagAlreadyExistError, TagDoesNotExistError, VCSError)
 
 
 SHA_PATTERN = re.compile(r'^[[0-9a-fA-F]{12}|[0-9a-fA-F]{40}]$')
 
 log = logging.getLogger(__name__)
 
 
 class GitRepository(BaseRepository):
     """
     Git repository backend.
     """
     DEFAULT_BRANCH_NAME = 'master'
 
     contact = BaseRepository.DEFAULT_CONTACT
 
     def __init__(self, repo_path, config=None, create=False, src_url=None,
                  update_after_clone=False, with_wire=None, bare=False):
 
         self.path = safe_str(os.path.abspath(repo_path))
         self.config = config if config else Config()
         self._remote = connection.Git(
             self.path, self.config, with_wire=with_wire)
 
         self._init_repo(create, src_url, update_after_clone, bare)
 
         # caches
         self._commit_ids = {}
 
         self.bookmarks = {}
 
     @LazyProperty
     def bare(self):
         return self._remote.bare()
 
     @LazyProperty
     def head(self):
         return self._remote.head()
 
     @LazyProperty
     def commit_ids(self):
         """
         Returns list of commit ids, in ascending order.  Being lazy
         attribute allows external tools to inject commit ids from cache.
         """
         commit_ids = self._get_all_commit_ids()
         self._rebuild_cache(commit_ids)
         return commit_ids
 
     def _rebuild_cache(self, commit_ids):
         self._commit_ids = dict((commit_id, index)
                                 for index, commit_id in enumerate(commit_ids))
 
     def run_git_command(self, cmd, **opts):
         """
         Runs given ``cmd`` as git command and returns tuple
         (stdout, stderr).
 
         :param cmd: git command to be executed
         :param opts: env options to pass into Subprocess command
         """
         if not isinstance(cmd, list):
             raise ValueError('cmd must be a list, got %s instead' % type(cmd))
 
         out, err = self._remote.run_git_command(cmd, **opts)
         log.debug('Stderr output of git command "%s":\n%s', cmd, err)
         return out, err
 
     @staticmethod
     def check_url(url, config):
         """
         Function will check given url and try to verify if it's a valid
         link. Sometimes it may happened that git will issue basic
         auth request that can cause whole API to hang when used from python
         or other external calls.
 
         On failures it'll raise urllib2.HTTPError, exception is also thrown
         when the return code is non 200
         """
         # check first if it's not an url
         if os.path.isdir(url) or url.startswith('file:'):
             return True
 
         if '+' in url.split('://', 1)[0]:
             url = url.split('+', 1)[1]
 
         # Request the _remote to verify the url
         return connection.Git.check_url(url, config.serialize())
 
     @staticmethod
     def is_valid_repository(path):
         if os.path.isdir(os.path.join(path, '.git')):
             return True
         # check case of bare repository
         try:
             GitRepository(path)
             return True
         except VCSError:
             pass
         return False
 
     def _init_repo(self, create, src_url=None, update_after_clone=False,
                    bare=False):
         if create and os.path.exists(self.path):
             raise RepositoryError(
                 "Cannot create repository at %s, location already exist"
                 % self.path)
 
         try:
             if create and src_url:
                 GitRepository.check_url(src_url, self.config)
                 self.clone(src_url, update_after_clone, bare)
             elif create:
                 os.makedirs(self.path, mode=0755)
 
                 if bare:
                     self._remote.init_bare()
                 else:
                     self._remote.init()
             else:
                 self._remote.assert_correct_path()
         # TODO: johbo: check if we have to translate the OSError here
         except OSError as err:
             raise RepositoryError(err)
 
     def _get_all_commit_ids(self, filters=None):
         # we must check if this repo is not empty, since later command
         # fails if it is. And it's cheaper to ask than throw the subprocess
         # errors
         try:
             self._remote.head()
         except KeyError:
             return []
 
         rev_filter = ['--branches', '--tags']
         extra_filter = []
 
         if filters:
             if filters.get('since'):
                 extra_filter.append('--since=%s' % (filters['since']))
             if filters.get('until'):
                 extra_filter.append('--until=%s' % (filters['until']))
             if filters.get('branch_name'):
                 rev_filter = ['--tags']
                 extra_filter.append(filters['branch_name'])
         rev_filter.extend(extra_filter)
 
             # if filters.get('start') or filters.get('end'):
             #     # skip is offset, max-count is limit
             #     if filters.get('start'):
             #         extra_filter += ' --skip=%s' % filters['start']
             #     if filters.get('end'):
             #         extra_filter += ' --max-count=%s' % (filters['end'] - (filters['start'] or 0))
 
         cmd = ['rev-list', '--reverse', '--date-order'] + rev_filter
         try:
             output, __ = self.run_git_command(cmd)
         except RepositoryError:
             # Can be raised for empty repositories
             return []
         return output.splitlines()
 
-    def _get_all_commit_ids2(self):
-        # alternate implementation
-        includes = [x[1][0] for x in self._parsed_refs.iteritems()
-                    if x[1][1] != 'T']
-        return [c.commit.id for c in self._remote.get_walker(include=includes)]
-
     def _get_commit_id(self, commit_id_or_idx):
         def is_null(value):
             return len(value) == commit_id_or_idx.count('0')
 
         if self.is_empty():
             raise EmptyRepositoryError("There are no commits yet")
 
         if commit_id_or_idx in (None, '', 'tip', 'HEAD', 'head', -1):
             return self.commit_ids[-1]
 
         is_bstr = isinstance(commit_id_or_idx, (str, unicode))
         if ((is_bstr and commit_id_or_idx.isdigit() and len(commit_id_or_idx) < 12)
             or isinstance(commit_id_or_idx, int) or is_null(commit_id_or_idx)):
             try:
                 commit_id_or_idx = self.commit_ids[int(commit_id_or_idx)]
             except Exception:
                 msg = "Commit %s does not exist for %s" % (
                     commit_id_or_idx, self)
                 raise CommitDoesNotExistError(msg)
 
         elif is_bstr:
-            # get by branch/tag name
-            ref_id = self._parsed_refs.get(commit_id_or_idx)
-            if ref_id:  # and ref_id[1] in ['H', 'RH', 'T']:
-                return ref_id[0]
+            # check full path ref, eg. refs/heads/master
+            ref_id = self._refs.get(commit_id_or_idx)
+            if ref_id:
+                return ref_id
 
-            tag_ids = self.tags.values()
-            # maybe it's a tag ? we don't have them in self.commit_ids
-            if commit_id_or_idx in tag_ids:
-                return commit_id_or_idx
+            # check branch name
+            branch_ids = self.branches.values()
+            ref_id = self._refs.get('refs/heads/%s' % commit_id_or_idx)
+            if ref_id:
+                return ref_id
 
-            elif (not SHA_PATTERN.match(commit_id_or_idx) or
+            # check tag name
+            ref_id = self._refs.get('refs/tags/%s' % commit_id_or_idx)
+            if ref_id:
+                return ref_id
+
+            if (not SHA_PATTERN.match(commit_id_or_idx) or
                     commit_id_or_idx not in self.commit_ids):
                 msg = "Commit %s does not exist for %s" % (
                     commit_id_or_idx, self)
                 raise CommitDoesNotExistError(msg)
 
         # Ensure we return full id
         if not SHA_PATTERN.match(str(commit_id_or_idx)):
             raise CommitDoesNotExistError(
                 "Given commit id %s not recognized" % commit_id_or_idx)
         return commit_id_or_idx
 
     def get_hook_location(self):
         """
         returns absolute path to location where hooks are stored
         """
         loc = os.path.join(self.path, 'hooks')
         if not self.bare:
             loc = os.path.join(self.path, '.git', 'hooks')
         return loc
 
     @LazyProperty
     def last_change(self):
         """
         Returns last change made on this repository as
         `datetime.datetime` object.
         """
         return utcdate_fromtimestamp(self._get_mtime(), makedate()[1])
 
     def _get_mtime(self):
         try:
             return time.mktime(self.get_commit().date.timetuple())
         except RepositoryError:
             idx_loc = '' if self.bare else '.git'
             # fallback to filesystem
             in_path = os.path.join(self.path, idx_loc, "index")
             he_path = os.path.join(self.path, idx_loc, "HEAD")
             if os.path.exists(in_path):
                 return os.stat(in_path).st_mtime
             else:
                 return os.stat(he_path).st_mtime
 
     @LazyProperty
     def description(self):
         description = self._remote.get_description()
         return safe_unicode(description or self.DEFAULT_DESCRIPTION)
 
-    def _get_refs_entry(self, value, reverse):
+    def _get_refs_entries(self, prefix='', reverse=False, strip_prefix=True):
         if self.is_empty():
-            return {}
+            return OrderedDict()
 
-        def get_name(ctx):
-            return ctx[0]
+        result = []
+        for ref, sha in self._refs.iteritems():
+            if ref.startswith(prefix):
+                ref_name = ref
+                if strip_prefix:
+                    ref_name = ref[len(prefix):]
+                result.append((safe_unicode(ref_name), sha))
 
-        _branches = [
-            (safe_unicode(x[0]), x[1][0])
-            for x in self._parsed_refs.iteritems() if x[1][1] == value]
-        return OrderedDict(sorted(_branches, key=get_name, reverse=reverse))
+        def get_name(entry):
+            return entry[0]
+
+        return OrderedDict(sorted(result, key=get_name, reverse=reverse))
 
     def _get_branches(self):
-        return self._get_refs_entry('H', False)
+        return self._get_refs_entries(prefix='refs/heads/', strip_prefix=True)
 
     @LazyProperty
     def branches(self):
         return self._get_branches()
 
     @LazyProperty
     def branches_closed(self):
         return {}
 
     @LazyProperty
     def branches_all(self):
         all_branches = {}
         all_branches.update(self.branches)
         all_branches.update(self.branches_closed)
         return all_branches
 
     @LazyProperty
     def tags(self):
         return self._get_tags()
 
     def _get_tags(self):
-        return self._get_refs_entry('T', True)
+        return self._get_refs_entries(
+            prefix='refs/tags/', strip_prefix=True, reverse=True)
 
     def tag(self, name, user, commit_id=None, message=None, date=None,
             **kwargs):
+        # TODO: fix this method to apply annotated tags correct with message
         """
         Creates and returns a tag for the given ``commit_id``.
 
         :param name: name for new tag
         :param user: full username, i.e.: "Joe Doe <[e-mail]>"
         :param commit_id: commit id for which new tag would be created
         :param message: message of the tag's commit
         :param date: date of tag's commit
 
         :raises TagAlreadyExistError: if tag with same name already exists
         """
         if name in self.tags:
             raise TagAlreadyExistError("Tag %s already exists" % name)
         commit = self.get_commit(commit_id=commit_id)
         message = message or "Added tag %s for commit %s" % (
             name, commit.raw_id)
         self._remote.set_refs('refs/tags/%s' % name, commit._commit['id'])
 
-        self._parsed_refs = self._get_parsed_refs()
+        self._refs = self._get_refs()
         self.tags = self._get_tags()
         return commit
 
     def remove_tag(self, name, user, message=None, date=None):
         """
         Removes tag with the given ``name``.
 
         :param name: name of the tag to be removed
         :param user: full username, i.e.: "Joe Doe <[e-mail]>"
         :param message: message of the tag's removal commit
         :param date: date of tag's removal commit
 
         :raises TagDoesNotExistError: if tag with given name does not exists
         """
         if name not in self.tags:
             raise TagDoesNotExistError("Tag %s does not exist" % name)
         tagpath = vcspath.join(
             self._remote.get_refs_path(), 'refs', 'tags', name)
         try:
             os.remove(tagpath)
-            self._parsed_refs = self._get_parsed_refs()
+            self._refs = self._get_refs()
             self.tags = self._get_tags()
         except OSError as e:
             raise RepositoryError(e.strerror)
 
+    def _get_refs(self):
+        return self._remote.get_refs()
+
     @LazyProperty
-    def _parsed_refs(self):
-        return self._get_parsed_refs()
+    def _refs(self):
+        return self._get_refs()
 
-    def _get_parsed_refs(self):
-        # TODO: (oliver) who needs RH; branches?
-        # Remote Heads were commented out, as they may overwrite local branches
-        # See the TODO note in rhodecode.lib.vcs.remote.git:get_refs for more
-        # details.
-        keys = [('refs/heads/', 'H'),
-                #('refs/remotes/origin/', 'RH'),
-                ('refs/tags/', 'T')]
-        return self._remote.get_refs(keys=keys)
+    @property
+    def _ref_tree(self):
+        node = tree = {}
+        for ref, sha in self._refs.iteritems():
+            path = ref.split('/')
+            for bit in path[:-1]:
+                node = node.setdefault(bit, {})
+            node[path[-1]] = sha
+            node = tree
+        return tree
 
     def get_commit(self, commit_id=None, commit_idx=None, pre_load=None):
         """
         Returns `GitCommit` object representing commit from git repository
         at the given `commit_id` or head (most recent commit) if None given.
         """
         if commit_id is not None:
             self._validate_commit_id(commit_id)
         elif commit_idx is not None:
             self._validate_commit_idx(commit_idx)
             commit_id = commit_idx
         commit_id = self._get_commit_id(commit_id)
         try:
             # Need to call remote to translate id for tagging scenario
             commit_id = self._remote.get_object(commit_id)["commit_id"]
             idx = self._commit_ids[commit_id]
         except KeyError:
             raise RepositoryError("Cannot get object with id %s" % commit_id)
 
         return GitCommit(self, commit_id, idx, pre_load=pre_load)
 
     def get_commits(
             self, start_id=None, end_id=None, start_date=None, end_date=None,
             branch_name=None, pre_load=None):
         """
         Returns generator of `GitCommit` objects from start to end (both
         are inclusive), in ascending date order.
 
         :param start_id: None, str(commit_id)
         :param end_id: None, str(commit_id)
         :param start_date: if specified, commits with commit date less than
           ``start_date`` would be filtered out from returned set
         :param end_date: if specified, commits with commit date greater than
           ``end_date`` would be filtered out from returned set
         :param branch_name: if specified, commits not reachable from given
           branch would be filtered out from returned set
 
         :raise BranchDoesNotExistError: If given `branch_name` does not
             exist.
         :raise CommitDoesNotExistError: If commits for given `start` or
           `end` could not be found.
 
         """
         if self.is_empty():
             raise EmptyRepositoryError("There are no commits yet")
         self._validate_branch_name(branch_name)
 
         if start_id is not None:
             self._validate_commit_id(start_id)
         if end_id is not None:
             self._validate_commit_id(end_id)
 
         start_raw_id = self._get_commit_id(start_id)
         start_pos = self._commit_ids[start_raw_id] if start_id else None
         end_raw_id = self._get_commit_id(end_id)
         end_pos = max(0, self._commit_ids[end_raw_id]) if end_id else None
 
         if None not in [start_id, end_id] and start_pos > end_pos:
             raise RepositoryError(
                 "Start commit '%s' cannot be after end commit '%s'" %
                 (start_id, end_id))
 
         if end_pos is not None:
             end_pos += 1
 
         filter_ = []
         if branch_name:
             filter_.append({'branch_name': branch_name})
         if start_date and not end_date:
             filter_.append({'since': start_date})
         if end_date and not start_date:
             filter_.append({'until': end_date})
         if start_date and end_date:
             filter_.append({'since': start_date})
             filter_.append({'until': end_date})
 
         # if start_pos or end_pos:
         #     filter_.append({'start': start_pos})
         #     filter_.append({'end': end_pos})
 
         if filter_:
             revfilters = {
                 'branch_name': branch_name,
                 'since': start_date.strftime('%m/%d/%y %H:%M:%S') if start_date else None,
                 'until': end_date.strftime('%m/%d/%y %H:%M:%S') if end_date else None,
                 'start': start_pos,
                 'end': end_pos,
             }
             commit_ids = self._get_all_commit_ids(filters=revfilters)
 
             # pure python stuff, it's slow due to walker walking whole repo
             # def get_revs(walker):
             #     for walker_entry in walker:
             #         yield walker_entry.commit.id
             # revfilters = {}
             # commit_ids = list(reversed(list(get_revs(self._repo.get_walker(**revfilters)))))
         else:
             commit_ids = self.commit_ids
 
         if start_pos or end_pos:
             commit_ids = commit_ids[start_pos: end_pos]
 
         return CollectionGenerator(self, commit_ids, pre_load=pre_load)
 
     def get_diff(
             self, commit1, commit2, path='', ignore_whitespace=False,
             context=3, path1=None):
         """
         Returns (git like) *diff*, as plain text. Shows changes introduced by
         ``commit2`` since ``commit1``.
 
         :param commit1: Entry point from which diff is shown. Can be
           ``self.EMPTY_COMMIT`` - in this case, patch showing all
           the changes since empty state of the repository until ``commit2``
         :param commit2: Until which commits changes should be shown.
         :param ignore_whitespace: If set to ``True``, would not show whitespace
           changes. Defaults to ``False``.
         :param context: How many lines before/after changed lines should be
           shown. Defaults to ``3``.
         """
         self._validate_diff_commits(commit1, commit2)
         if path1 is not None and path1 != path:
             raise ValueError("Diff of two different paths not supported.")
 
         flags = [
             '-U%s' % context, '--full-index', '--binary', '-p',
             '-M', '--abbrev=40']
         if ignore_whitespace:
             flags.append('-w')
 
         if commit1 == self.EMPTY_COMMIT:
             cmd = ['show'] + flags + [commit2.raw_id]
         else:
             cmd = ['diff'] + flags + [commit1.raw_id, commit2.raw_id]
 
         if path:
             cmd.extend(['--', path])
 
         stdout, __ = self.run_git_command(cmd)
         # If we used 'show' command, strip first few lines (until actual diff
         # starts)
         if commit1 == self.EMPTY_COMMIT:
             lines = stdout.splitlines()
             x = 0
             for line in lines:
                 if line.startswith('diff'):
                     break
                 x += 1
             # Append new line just like 'diff' command do
             stdout = '\n'.join(lines[x:]) + '\n'
         return GitDiff(stdout)
 
     def strip(self, commit_id, branch_name):
         commit = self.get_commit(commit_id=commit_id)
         if commit.merge:
             raise Exception('Cannot reset to merge commit')
 
         # parent is going to be the new head now
         commit = commit.parents[0]
         self._remote.set_refs('refs/heads/%s' % branch_name, commit.raw_id)
 
         self.commit_ids = self._get_all_commit_ids()
         self._rebuild_cache(self.commit_ids)
 
     def get_common_ancestor(self, commit_id1, commit_id2, repo2):
         if commit_id1 == commit_id2:
             return commit_id1
 
         if self != repo2:
             commits = self._remote.get_missing_revs(
                 commit_id1, commit_id2, repo2.path)
             if commits:
                 commit = repo2.get_commit(commits[-1])
                 if commit.parents:
                     ancestor_id = commit.parents[0].raw_id
                 else:
                     ancestor_id = None
             else:
                 # no commits from other repo, ancestor_id is the commit_id2
                 ancestor_id = commit_id2
         else:
             output, __ = self.run_git_command(
                 ['merge-base', commit_id1, commit_id2])
             ancestor_id = re.findall(r'[0-9a-fA-F]{40}', output)[0]
 
         return ancestor_id
 
     def compare(self, commit_id1, commit_id2, repo2, merge, pre_load=None):
         repo1 = self
         ancestor_id = None
 
         if commit_id1 == commit_id2:
             commits = []
         elif repo1 != repo2:
             missing_ids = self._remote.get_missing_revs(commit_id1, commit_id2,
                                                         repo2.path)
             commits = [
                 repo2.get_commit(commit_id=commit_id, pre_load=pre_load)
                 for commit_id in reversed(missing_ids)]
         else:
             output, __ = repo1.run_git_command(
                 ['log', '--reverse', '--pretty=format: %H', '-s',
                  '%s..%s' % (commit_id1, commit_id2)])
             commits = [
                 repo1.get_commit(commit_id=commit_id, pre_load=pre_load)
                 for commit_id in re.findall(r'[0-9a-fA-F]{40}', output)]
 
         return commits
 
     @LazyProperty
     def in_memory_commit(self):
         """
         Returns ``GitInMemoryCommit`` object for this repository.
         """
         return GitInMemoryCommit(self)
 
     def clone(self, url, update_after_clone=True, bare=False):
         """
         Tries to clone commits from external location.
 
         :param update_after_clone: If set to ``False``, git won't checkout
           working directory
         :param bare: If set to ``True``, repository would be cloned into
           *bare* git repository (no working directory at all).
         """
         # init_bare and init expect empty dir created to proceed
         if not os.path.exists(self.path):
             os.mkdir(self.path)
 
         if bare:
             self._remote.init_bare()
         else:
             self._remote.init()
 
         deferred = '^{}'
         valid_refs = ('refs/heads', 'refs/tags', 'HEAD')
 
         return self._remote.clone(
             url, deferred, valid_refs, update_after_clone)
 
     def pull(self, url, commit_ids=None):
         """
         Tries to pull changes from external location. We use fetch here since
         pull in get does merges and we want to be compatible with hg backend so
         pull == fetch in this case
         """
         self.fetch(url, commit_ids=commit_ids)
 
     def fetch(self, url, commit_ids=None):
         """
         Tries to fetch changes from external location.
         """
         refs = None
 
         if commit_ids is not None:
             remote_refs = self._remote.get_remote_refs(url)
             refs = [
                 ref for ref in remote_refs if remote_refs[ref] in commit_ids]
         self._remote.fetch(url, refs=refs)
 
     def set_refs(self, ref_name, commit_id):
         self._remote.set_refs(ref_name, commit_id)
 
     def remove_ref(self, ref_name):
         self._remote.remove_ref(ref_name)
 
     def _update_server_info(self):
         """
         runs gits update-server-info command in this repo instance
         """
         self._remote.update_server_info()
 
     def _current_branch(self):
         """
         Return the name of the current branch.
 
         It only works for non bare repositories (i.e. repositories with a
         working copy)
         """
         if self.bare:
             raise RepositoryError('Bare git repos do not have active branches')
 
         if self.is_empty():
             return None
 
         stdout, _ = self.run_git_command(['rev-parse', '--abbrev-ref', 'HEAD'])
         return stdout.strip()
 
     def _checkout(self, branch_name, create=False):
         """
         Checkout a branch in the working directory.
 
         It tries to create the branch if create is True, failing if the branch
         already exists.
 
         It only works for non bare repositories (i.e. repositories with a
         working copy)
         """
         if self.bare:
             raise RepositoryError('Cannot checkout branches in a bare git repo')
 
         cmd = ['checkout']
         if create:
             cmd.append('-b')
         cmd.append(branch_name)
         self.run_git_command(cmd, fail_on_stderr=False)
 
     def _local_clone(self, clone_path, branch_name):
         """
         Create a local clone of the current repo.
         """
         # N.B.(skreft): the --branch option is required as otherwise the shallow
         # clone will only fetch the active branch.
         cmd = ['clone', '--branch', branch_name, '--single-branch',
                self.path, os.path.abspath(clone_path)]
         self.run_git_command(cmd, fail_on_stderr=False)
 
     def _local_fetch(self, repository_path, branch_name):
         """
         Fetch a branch from a local repository.
         """
         repository_path = os.path.abspath(repository_path)
         if repository_path == self.path:
             raise ValueError('Cannot fetch from the same repository')
 
         cmd = ['fetch', '--no-tags', repository_path, branch_name]
         self.run_git_command(cmd, fail_on_stderr=False)
 
     def _last_fetch_heads(self):
         """
         Return the last fetched heads that need merging.
 
         The algorithm is defined at
         https://github.com/git/git/blob/v2.1.3/git-pull.sh#L283
         """
         if not self.bare:
             fetch_heads_path = os.path.join(self.path, '.git', 'FETCH_HEAD')
         else:
             fetch_heads_path = os.path.join(self.path, 'FETCH_HEAD')
 
         heads = []
         with open(fetch_heads_path) as f:
             for line in f:
                 if '    not-for-merge   ' in line:
                     continue
                 line = re.sub('\t.*', '', line, flags=re.DOTALL)
                 heads.append(line)
 
         return heads
 
     def _local_pull(self, repository_path, branch_name):
         """
         Pull a branch from a local repository.
         """
         if self.bare:
             raise RepositoryError('Cannot pull into a bare git repository')
         # N.B.(skreft): The --ff-only option is to make sure this is a
         # fast-forward (i.e., we are only pulling new changes and there are no
         # conflicts with our current branch)
         # Additionally, that option needs to go before --no-tags, otherwise git
         # pull complains about it being an unknown flag.
         cmd = ['pull', '--ff-only', '--no-tags', repository_path, branch_name]
         self.run_git_command(cmd, fail_on_stderr=False)
 
     def _local_merge(self, merge_message, user_name, user_email, heads):
         """
         Merge the given head into the checked out branch.
 
         It will force a merge commit.
 
         Currently it raises an error if the repo is empty, as it is not possible
         to create a merge commit in an empty repo.
 
         :param merge_message: The message to use for the merge commit.
         :param heads: the heads to merge.
         """
         if self.bare:
             raise RepositoryError('Cannot merge into a bare git repository')
 
         if not heads:
             return
 
         if self.is_empty():
             # TODO(skreft): do somehting more robust in this case.
             raise RepositoryError(
                 'Do not know how to merge into empty repositories yet')
 
         # N.B.(skreft): the --no-ff option is used to enforce the creation of a
         # commit message. We also specify the user who is doing the merge.
         cmd = ['-c', 'user.name=%s' % safe_str(user_name),
                '-c', 'user.email=%s' % safe_str(user_email),
                'merge', '--no-ff', '-m', safe_str(merge_message)]
         cmd.extend(heads)
         try:
             self.run_git_command(cmd, fail_on_stderr=False)
         except RepositoryError:
             # Cleanup any merge leftovers
             self.run_git_command(['merge', '--abort'], fail_on_stderr=False)
             raise
 
     def _local_push(
             self, source_branch, repository_path, target_branch,
             enable_hooks=False, rc_scm_data=None):
         """
         Push the source_branch to the given repository and target_branch.
 
         Currently it if the target_branch is not master and the target repo is
         empty, the push will work, but then GitRepository won't be able to find
         the pushed branch or the commits. As the HEAD will be corrupted (i.e.,
         pointing to master, which does not exist).
 
         It does not run the hooks in the target repo.
         """
         # TODO(skreft): deal with the case in which the target repo is empty,
         # and the target_branch is not master.
         target_repo = GitRepository(repository_path)
         if (not target_repo.bare and
                 target_repo._current_branch() == target_branch):
             # Git prevents pushing to the checked out branch, so simulate it by
             # pulling into the target repository.
             target_repo._local_pull(self.path, source_branch)
         else:
             cmd = ['push', os.path.abspath(repository_path),
                    '%s:%s' % (source_branch, target_branch)]
             gitenv = {}
             if rc_scm_data:
                 gitenv.update({'RC_SCM_DATA': rc_scm_data})
 
             if not enable_hooks:
                 gitenv['RC_SKIP_HOOKS'] = '1'
             self.run_git_command(cmd, fail_on_stderr=False, extra_env=gitenv)
 
     def _get_new_pr_branch(self, source_branch, target_branch):
         prefix = 'pr_%s-%s_' % (source_branch, target_branch)
         pr_branches = []
         for branch in self.branches:
             if branch.startswith(prefix):
                 pr_branches.append(int(branch[len(prefix):]))
 
         if not pr_branches:
             branch_id = 0
         else:
             branch_id = max(pr_branches) + 1
 
         return '%s%d' % (prefix, branch_id)
 
     def _merge_repo(self, shadow_repository_path, target_ref,
                     source_repo, source_ref, merge_message,
                     merger_name, merger_email, dry_run=False,
                     use_rebase=False):
         if target_ref.commit_id != self.branches[target_ref.name]:
             return MergeResponse(
                 False, False, None, MergeFailureReason.TARGET_IS_NOT_HEAD)
 
         shadow_repo = GitRepository(shadow_repository_path)
         shadow_repo._checkout(target_ref.name)
         shadow_repo._local_pull(self.path, target_ref.name)
         # Need to reload repo to invalidate the cache, or otherwise we cannot
         # retrieve the last target commit.
         shadow_repo = GitRepository(shadow_repository_path)
         if target_ref.commit_id != shadow_repo.branches[target_ref.name]:
             return MergeResponse(
                 False, False, None, MergeFailureReason.TARGET_IS_NOT_HEAD)
 
         pr_branch = shadow_repo._get_new_pr_branch(
             source_ref.name, target_ref.name)
         shadow_repo._checkout(pr_branch, create=True)
         try:
             shadow_repo._local_fetch(source_repo.path, source_ref.name)
         except RepositoryError as e:
             log.exception('Failure when doing local fetch on git shadow repo')
             return MergeResponse(
                 False, False, None, MergeFailureReason.MISSING_COMMIT)
 
         merge_commit_id = None
         merge_failure_reason = MergeFailureReason.NONE
         try:
             shadow_repo._local_merge(merge_message, merger_name, merger_email,
                                      [source_ref.commit_id])
             merge_possible = True
         except RepositoryError as e:
             log.exception('Failure when doing local merge on git shadow repo')
             merge_possible = False
             merge_failure_reason = MergeFailureReason.MERGE_FAILED
 
         if merge_possible and not dry_run:
             try:
                 shadow_repo._local_push(
                     pr_branch, self.path, target_ref.name, enable_hooks=True,
                     rc_scm_data=self.config.get('rhodecode', 'RC_SCM_DATA'))
                 merge_succeeded = True
                 # Need to reload repo to invalidate the cache, or otherwise we
                 # cannot retrieve the merge commit.
                 shadow_repo = GitRepository(shadow_repository_path)
                 merge_commit_id = shadow_repo.branches[pr_branch]
             except RepositoryError as e:
                 log.exception(
                     'Failure when doing local push on git shadow repo')
                 merge_succeeded = False
                 merge_failure_reason = MergeFailureReason.PUSH_FAILED
         else:
             merge_succeeded = False
 
         return MergeResponse(
             merge_possible, merge_succeeded, merge_commit_id,
             merge_failure_reason)
 
     def _get_shadow_repository_path(self, workspace_id):
         # The name of the shadow repository must start with '.', so it is
         # skipped by 'rhodecode.lib.utils.get_filesystem_repos'.
         return os.path.join(
             os.path.dirname(self.path),
             '.__shadow_%s_%s' % (os.path.basename(self.path), workspace_id))
 
     def _maybe_prepare_merge_workspace(self, workspace_id, target_ref):
         shadow_repository_path = self._get_shadow_repository_path(workspace_id)
         if not os.path.exists(shadow_repository_path):
             self._local_clone(shadow_repository_path, target_ref.name)
 
         return shadow_repository_path
 
     def cleanup_merge_workspace(self, workspace_id):
         shadow_repository_path = self._get_shadow_repository_path(workspace_id)
         shutil.rmtree(shadow_repository_path, ignore_errors=True)

rhodecode/tests/vcs/test_git.py

@@ -1,1228 +1,1242 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2010-2016  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 mock
 import os
 import sys
 
 import pytest
 
 from rhodecode.lib.vcs.backends.base import Reference
 from rhodecode.lib.vcs.backends.git import (
     GitRepository, GitCommit, discover_git_version)
 from rhodecode.lib.vcs.exceptions import (
     RepositoryError, VCSError, NodeDoesNotExistError
 )
 from rhodecode.lib.vcs.nodes import (
     NodeKind, FileNode, DirNode, NodeState, SubModuleNode)
 from rhodecode.tests import TEST_GIT_REPO, TEST_GIT_REPO_CLONE, get_new_dir
 from rhodecode.tests.vcs.base import BackendTestMixin
 
 
 pytestmark = pytest.mark.backends("git")
 
 
 def repo_path_generator():
     """
     Return a different path to be used for cloning repos.
     """
     i = 0
     while True:
         i += 1
         yield '%s-%d' % (TEST_GIT_REPO_CLONE, i)
 
 
 REPO_PATH_GENERATOR = repo_path_generator()
 
 
 class TestGitRepository:
 
     # pylint: disable=protected-access
 
     def __check_for_existing_repo(self):
         if os.path.exists(TEST_GIT_REPO_CLONE):
             self.fail('Cannot test git clone repo as location %s already '
                       'exists. You should manually remove it first.'
                       % TEST_GIT_REPO_CLONE)
 
     @pytest.fixture(autouse=True)
     def prepare(self, request, pylonsapp):
         self.repo = GitRepository(TEST_GIT_REPO, bare=True)
 
     def get_clone_repo(self):
         """
         Return a non bare clone of the base repo.
         """
         clone_path = next(REPO_PATH_GENERATOR)
         repo_clone = GitRepository(
             clone_path, create=True, src_url=self.repo.path, bare=False)
 
         return repo_clone
 
     def get_empty_repo(self, bare=False):
         """
         Return a non bare empty repo.
         """
         return GitRepository(next(REPO_PATH_GENERATOR), create=True, bare=bare)
 
     def test_wrong_repo_path(self):
         wrong_repo_path = '/tmp/errorrepo'
         with pytest.raises(RepositoryError):
             GitRepository(wrong_repo_path)
 
     def test_repo_clone(self):
         self.__check_for_existing_repo()
         repo = GitRepository(TEST_GIT_REPO)
         repo_clone = GitRepository(
             TEST_GIT_REPO_CLONE,
             src_url=TEST_GIT_REPO, create=True, update_after_clone=True)
         assert len(repo.commit_ids) == len(repo_clone.commit_ids)
         # Checking hashes of commits should be enough
         for commit in repo.get_commits():
             raw_id = commit.raw_id
             assert raw_id == repo_clone.get_commit(raw_id).raw_id
 
     def test_repo_clone_without_create(self):
         with pytest.raises(RepositoryError):
             GitRepository(
                 TEST_GIT_REPO_CLONE + '_wo_create', src_url=TEST_GIT_REPO)
 
     def test_repo_clone_with_update(self):
         repo = GitRepository(TEST_GIT_REPO)
         clone_path = TEST_GIT_REPO_CLONE + '_with_update'
         repo_clone = GitRepository(
             clone_path,
             create=True, src_url=TEST_GIT_REPO, update_after_clone=True)
         assert len(repo.commit_ids) == len(repo_clone.commit_ids)
 
         # check if current workdir was updated
         fpath = os.path.join(clone_path, 'MANIFEST.in')
         assert os.path.isfile(fpath)
 
     def test_repo_clone_without_update(self):
         repo = GitRepository(TEST_GIT_REPO)
         clone_path = TEST_GIT_REPO_CLONE + '_without_update'
         repo_clone = GitRepository(
             clone_path,
             create=True, src_url=TEST_GIT_REPO, update_after_clone=False)
         assert len(repo.commit_ids) == len(repo_clone.commit_ids)
         # check if current workdir was *NOT* updated
         fpath = os.path.join(clone_path, 'MANIFEST.in')
         # Make sure it's not bare repo
         assert not repo_clone.bare
         assert not os.path.isfile(fpath)
 
     def test_repo_clone_into_bare_repo(self):
         repo = GitRepository(TEST_GIT_REPO)
         clone_path = TEST_GIT_REPO_CLONE + '_bare.git'
         repo_clone = GitRepository(
             clone_path, create=True, src_url=repo.path, bare=True)
         assert repo_clone.bare
 
     def test_create_repo_is_not_bare_by_default(self):
         repo = GitRepository(get_new_dir('not-bare-by-default'), create=True)
         assert not repo.bare
 
     def test_create_bare_repo(self):
         repo = GitRepository(get_new_dir('bare-repo'), create=True, bare=True)
         assert repo.bare
 
     def test_update_server_info(self):
         self.repo._update_server_info()
 
     def test_fetch(self, vcsbackend_git):
         # Note: This is a git specific part of the API, it's only implemented
         # by the git backend.
         source_repo = vcsbackend_git.repo
         target_repo = vcsbackend_git.create_repo()
         target_repo.fetch(source_repo.path)
         # Note: Get a fresh instance, avoids caching trouble
         target_repo = vcsbackend_git.backend(target_repo.path)
         assert len(source_repo.commit_ids) == len(target_repo.commit_ids)
 
     def test_commit_ids(self):
         # there are 112 commits (by now)
         # so we can assume they would be available from now on
         subset = set([
             'c1214f7e79e02fc37156ff215cd71275450cffc3',
             '38b5fe81f109cb111f549bfe9bb6b267e10bc557',
             'fa6600f6848800641328adbf7811fd2372c02ab2',
             '102607b09cdd60e2793929c4f90478be29f85a17',
             '49d3fd156b6f7db46313fac355dca1a0b94a0017',
             '2d1028c054665b962fa3d307adfc923ddd528038',
             'd7e0d30fbcae12c90680eb095a4f5f02505ce501',
             'ff7ca51e58c505fec0dd2491de52c622bb7a806b',
             'dd80b0f6cf5052f17cc738c2951c4f2070200d7f',
             '8430a588b43b5d6da365400117c89400326e7992',
             'd955cd312c17b02143c04fa1099a352b04368118',
             'f67b87e5c629c2ee0ba58f85197e423ff28d735b',
             'add63e382e4aabc9e1afdc4bdc24506c269b7618',
             'f298fe1189f1b69779a4423f40b48edf92a703fc',
             'bd9b619eb41994cac43d67cf4ccc8399c1125808',
             '6e125e7c890379446e98980d8ed60fba87d0f6d1',
             'd4a54db9f745dfeba6933bf5b1e79e15d0af20bd',
             '0b05e4ed56c802098dfc813cbe779b2f49e92500',
             '191caa5b2c81ed17c0794bf7bb9958f4dcb0b87e',
             '45223f8f114c64bf4d6f853e3c35a369a6305520',
             'ca1eb7957a54bce53b12d1a51b13452f95bc7c7e',
             'f5ea29fc42ef67a2a5a7aecff10e1566699acd68',
             '27d48942240f5b91dfda77accd2caac94708cc7d',
             '622f0eb0bafd619d2560c26f80f09e3b0b0d78af',
             'e686b958768ee96af8029fe19c6050b1a8dd3b2b'])
         assert subset.issubset(set(self.repo.commit_ids))
 
     def test_slicing(self):
         # 4 1 5 10 95
         for sfrom, sto, size in [(0, 4, 4), (1, 2, 1), (10, 15, 5),
                                  (10, 20, 10), (5, 100, 95)]:
             commit_ids = list(self.repo[sfrom:sto])
             assert len(commit_ids) == size
             assert commit_ids[0] == self.repo.get_commit(commit_idx=sfrom)
             assert commit_ids[-1] == self.repo.get_commit(commit_idx=sto - 1)
 
     def test_branches(self):
         # TODO: Need more tests here
         # Removed (those are 'remotes' branches for cloned repo)
         # assert 'master' in self.repo.branches
         # assert 'gittree' in self.repo.branches
         # assert 'web-branch' in self.repo.branches
         for __, commit_id in self.repo.branches.items():
             assert isinstance(self.repo.get_commit(commit_id), GitCommit)
 
     def test_tags(self):
         # TODO: Need more tests here
         assert 'v0.1.1' in self.repo.tags
         assert 'v0.1.2' in self.repo.tags
         for __, commit_id in self.repo.tags.items():
             assert isinstance(self.repo.get_commit(commit_id), GitCommit)
 
     def _test_single_commit_cache(self, commit_id):
         commit = self.repo.get_commit(commit_id)
         assert commit_id in self.repo.commits
         assert commit is self.repo.commits[commit_id]
 
     def test_initial_commit(self):
         commit_id = self.repo.commit_ids[0]
         init_commit = self.repo.get_commit(commit_id)
         init_author = init_commit.author
 
         assert init_commit.message == 'initial import\n'
         assert init_author == 'Marcin Kuzminski <[e-mail]>'
         assert init_author == init_commit.committer
         for path in ('vcs/__init__.py',
                      'vcs/backends/BaseRepository.py',
                      'vcs/backends/__init__.py'):
             assert isinstance(init_commit.get_node(path), FileNode)
         for path in ('', 'vcs', 'vcs/backends'):
             assert isinstance(init_commit.get_node(path), DirNode)
 
         with pytest.raises(NodeDoesNotExistError):
             init_commit.get_node(path='foobar')
 
         node = init_commit.get_node('vcs/')
         assert hasattr(node, 'kind')
         assert node.kind == NodeKind.DIR
 
         node = init_commit.get_node('vcs')
         assert hasattr(node, 'kind')
         assert node.kind == NodeKind.DIR
 
         node = init_commit.get_node('vcs/__init__.py')
         assert hasattr(node, 'kind')
         assert node.kind == NodeKind.FILE
 
     def test_not_existing_commit(self):
         with pytest.raises(RepositoryError):
             self.repo.get_commit('f' * 40)
 
     def test_commit10(self):
 
         commit10 = self.repo.get_commit(self.repo.commit_ids[9])
         README = """===
 VCS
 ===
 
 Various Version Control System management abstraction layer for Python.
 
 Introduction
 ------------
 
 TODO: To be written...
 
 """
         node = commit10.get_node('README.rst')
         assert node.kind == NodeKind.FILE
         assert node.content == README
 
     def test_head(self):
         assert self.repo.head == self.repo.get_commit().raw_id
 
     def test_checkout_with_create(self):
         repo_clone = self.get_clone_repo()
 
         new_branch = 'new_branch'
         assert repo_clone._current_branch() == 'master'
         assert set(repo_clone.branches) == set(('master',))
         repo_clone._checkout(new_branch, create=True)
 
         # Branches is a lazy property so we need to recrete the Repo object.
         repo_clone = GitRepository(repo_clone.path)
         assert set(repo_clone.branches) == set(('master', new_branch))
         assert repo_clone._current_branch() == new_branch
 
     def test_checkout(self):
         repo_clone = self.get_clone_repo()
 
         repo_clone._checkout('new_branch', create=True)
         repo_clone._checkout('master')
 
         assert repo_clone._current_branch() == 'master'
 
     def test_checkout_same_branch(self):
         repo_clone = self.get_clone_repo()
 
         repo_clone._checkout('master')
         assert repo_clone._current_branch() == 'master'
 
     def test_checkout_branch_already_exists(self):
         repo_clone = self.get_clone_repo()
 
         with pytest.raises(RepositoryError):
             repo_clone._checkout('master', create=True)
 
     def test_checkout_bare_repo(self):
         with pytest.raises(RepositoryError):
             self.repo._checkout('master')
 
     def test_current_branch_bare_repo(self):
         with pytest.raises(RepositoryError):
             self.repo._current_branch()
 
     def test_current_branch_empty_repo(self):
         repo = self.get_empty_repo()
         assert repo._current_branch() is None
 
     def test_local_clone(self):
         clone_path = next(REPO_PATH_GENERATOR)
         self.repo._local_clone(clone_path, 'master')
         repo_clone = GitRepository(clone_path)
 
         assert self.repo.commit_ids == repo_clone.commit_ids
 
     def test_local_clone_with_specific_branch(self):
         source_repo = self.get_clone_repo()
 
         # Create a new branch in source repo
         new_branch_commit = source_repo.commit_ids[-3]
         source_repo._checkout(new_branch_commit)
         source_repo._checkout('new_branch', create=True)
 
         clone_path = next(REPO_PATH_GENERATOR)
         source_repo._local_clone(clone_path, 'new_branch')
         repo_clone = GitRepository(clone_path)
 
         assert source_repo.commit_ids[:-3 + 1] == repo_clone.commit_ids
 
         clone_path = next(REPO_PATH_GENERATOR)
         source_repo._local_clone(clone_path, 'master')
         repo_clone = GitRepository(clone_path)
 
         assert source_repo.commit_ids == repo_clone.commit_ids
 
     def test_local_clone_fails_if_target_exists(self):
         with pytest.raises(RepositoryError):
             self.repo._local_clone(self.repo.path, 'master')
 
     def test_local_fetch(self):
         target_repo = self.get_empty_repo()
         source_repo = self.get_clone_repo()
 
         # Create a new branch in source repo
         master_commit = source_repo.commit_ids[-1]
         new_branch_commit = source_repo.commit_ids[-3]
         source_repo._checkout(new_branch_commit)
         source_repo._checkout('new_branch', create=True)
 
         target_repo._local_fetch(source_repo.path, 'new_branch')
         assert target_repo._last_fetch_heads() == [new_branch_commit]
 
         target_repo._local_fetch(source_repo.path, 'master')
         assert target_repo._last_fetch_heads() == [master_commit]
 
     def test_local_fetch_from_bare_repo(self):
         target_repo = self.get_empty_repo()
         target_repo._local_fetch(self.repo.path, 'master')
 
         master_commit = self.repo.commit_ids[-1]
         assert target_repo._last_fetch_heads() == [master_commit]
 
     def test_local_fetch_from_same_repo(self):
         with pytest.raises(ValueError):
             self.repo._local_fetch(self.repo.path, 'master')
 
     def test_local_fetch_branch_does_not_exist(self):
         target_repo = self.get_empty_repo()
 
         with pytest.raises(RepositoryError):
             target_repo._local_fetch(self.repo.path, 'new_branch')
 
     def test_local_pull(self):
         target_repo = self.get_empty_repo()
         source_repo = self.get_clone_repo()
 
         # Create a new branch in source repo
         master_commit = source_repo.commit_ids[-1]
         new_branch_commit = source_repo.commit_ids[-3]
         source_repo._checkout(new_branch_commit)
         source_repo._checkout('new_branch', create=True)
 
         target_repo._local_pull(source_repo.path, 'new_branch')
         target_repo = GitRepository(target_repo.path)
         assert target_repo.head == new_branch_commit
 
         target_repo._local_pull(source_repo.path, 'master')
         target_repo = GitRepository(target_repo.path)
         assert target_repo.head == master_commit
 
     def test_local_pull_in_bare_repo(self):
         with pytest.raises(RepositoryError):
             self.repo._local_pull(self.repo.path, 'master')
 
     def test_local_merge(self):
         target_repo = self.get_empty_repo()
         source_repo = self.get_clone_repo()
 
         # Create a new branch in source repo
         master_commit = source_repo.commit_ids[-1]
         new_branch_commit = source_repo.commit_ids[-3]
         source_repo._checkout(new_branch_commit)
         source_repo._checkout('new_branch', create=True)
 
         # This is required as one cannot do a -ff-only merge in an empty repo.
         target_repo._local_pull(source_repo.path, 'new_branch')
 
         target_repo._local_fetch(source_repo.path, 'master')
         merge_message = 'Merge message\n\nDescription:...'
         user_name = 'Albert Einstein'
         user_email = '[e-mail]'
         target_repo._local_merge(merge_message, user_name, user_email,
                                  target_repo._last_fetch_heads())
 
         target_repo = GitRepository(target_repo.path)
         assert target_repo.commit_ids[-2] == master_commit
         last_commit = target_repo.get_commit(target_repo.head)
         assert last_commit.message.strip() == merge_message
         assert last_commit.author == '%s <%s>' % (user_name, user_email)
 
         assert not os.path.exists(
             os.path.join(target_repo.path, '.git', 'MERGE_HEAD'))
 
     def test_local_merge_raises_exception_on_conflict(self, vcsbackend_git):
         target_repo = vcsbackend_git.create_repo(number_of_commits=1)
         vcsbackend_git.ensure_file('README', 'I will conflict with you!!!')
 
         target_repo._local_fetch(self.repo.path, 'master')
         with pytest.raises(RepositoryError):
             target_repo._local_merge(
                 'merge_message', 'user name', '[e-mail]',
                 target_repo._last_fetch_heads())
 
         # Check we are not left in an intermediate merge state
         assert not os.path.exists(
             os.path.join(target_repo.path, '.git', 'MERGE_HEAD'))
 
     def test_local_merge_into_empty_repo(self):
         target_repo = self.get_empty_repo()
 
         # This is required as one cannot do a -ff-only merge in an empty repo.
         target_repo._local_fetch(self.repo.path, 'master')
         with pytest.raises(RepositoryError):
             target_repo._local_merge(
                 'merge_message', 'user name', '[e-mail]',
                 target_repo._last_fetch_heads())
 
     def test_local_merge_in_bare_repo(self):
         with pytest.raises(RepositoryError):
             self.repo._local_merge(
                 'merge_message', 'user name', '[e-mail]', None)
 
     def test_local_push_non_bare(self):
         target_repo = self.get_empty_repo()
 
         pushed_branch = 'pushed_branch'
         self.repo._local_push('master', target_repo.path, pushed_branch)
         # Fix the HEAD of the target repo, or otherwise GitRepository won't
         # report any branches.
         with open(os.path.join(target_repo.path, '.git', 'HEAD'), 'w') as f:
             f.write('ref: refs/heads/%s' % pushed_branch)
 
         target_repo = GitRepository(target_repo.path)
 
         assert (target_repo.branches[pushed_branch] ==
                 self.repo.branches['master'])
 
     def test_local_push_bare(self):
         target_repo = self.get_empty_repo(bare=True)
 
         pushed_branch = 'pushed_branch'
         self.repo._local_push('master', target_repo.path, pushed_branch)
         # Fix the HEAD of the target repo, or otherwise GitRepository won't
         # report any branches.
         with open(os.path.join(target_repo.path, 'HEAD'), 'w') as f:
             f.write('ref: refs/heads/%s' % pushed_branch)
 
         target_repo = GitRepository(target_repo.path)
 
         assert (target_repo.branches[pushed_branch] ==
                 self.repo.branches['master'])
 
     def test_local_push_non_bare_target_branch_is_checked_out(self):
         target_repo = self.get_clone_repo()
 
         pushed_branch = 'pushed_branch'
         # Create a new branch in source repo
         new_branch_commit = target_repo.commit_ids[-3]
         target_repo._checkout(new_branch_commit)
         target_repo._checkout(pushed_branch, create=True)
 
         self.repo._local_push('master', target_repo.path, pushed_branch)
 
         target_repo = GitRepository(target_repo.path)
 
         assert (target_repo.branches[pushed_branch] ==
                 self.repo.branches['master'])
 
     def test_local_push_raises_exception_on_conflict(self, vcsbackend_git):
         target_repo = vcsbackend_git.create_repo(number_of_commits=1)
         with pytest.raises(RepositoryError):
             self.repo._local_push('master', target_repo.path, 'master')
 
     def test_hooks_can_be_enabled_via_env_variable_for_local_push(self):
         target_repo = self.get_empty_repo(bare=True)
 
         with mock.patch.object(self.repo, 'run_git_command') as run_mock:
             self.repo._local_push(
                 'master', target_repo.path, 'master', enable_hooks=True)
         env = run_mock.call_args[1]['extra_env']
         assert 'RC_SKIP_HOOKS' not in env
 
     def _add_failing_hook(self, repo_path, hook_name, bare=False):
         path_components = (
             ['hooks', hook_name] if bare else ['.git', 'hooks', hook_name])
         hook_path = os.path.join(repo_path, *path_components)
         with open(hook_path, 'w') as f:
             script_lines = [
                 '#!%s' % sys.executable,
                 'import os',
                 'import sys',
                 'if os.environ.get("RC_SKIP_HOOKS"):',
                 '    sys.exit(0)',
                 'sys.exit(1)',
             ]
             f.write('\n'.join(script_lines))
         os.chmod(hook_path, 0755)
 
     def test_local_push_does_not_execute_hook(self):
         target_repo = self.get_empty_repo()
 
         pushed_branch = 'pushed_branch'
         self._add_failing_hook(target_repo.path, 'pre-receive')
         self.repo._local_push('master', target_repo.path, pushed_branch)
         # Fix the HEAD of the target repo, or otherwise GitRepository won't
         # report any branches.
         with open(os.path.join(target_repo.path, '.git', 'HEAD'), 'w') as f:
             f.write('ref: refs/heads/%s' % pushed_branch)
 
         target_repo = GitRepository(target_repo.path)
 
         assert (target_repo.branches[pushed_branch] ==
                 self.repo.branches['master'])
 
     def test_local_push_executes_hook(self):
         target_repo = self.get_empty_repo(bare=True)
         self._add_failing_hook(target_repo.path, 'pre-receive', bare=True)
         with pytest.raises(RepositoryError):
             self.repo._local_push(
                 'master', target_repo.path, 'master', enable_hooks=True)
 
     def test_maybe_prepare_merge_workspace(self):
         workspace = self.repo._maybe_prepare_merge_workspace(
             'pr2', Reference('branch', 'master', 'unused'))
 
         assert os.path.isdir(workspace)
         workspace_repo = GitRepository(workspace)
         assert workspace_repo.branches == self.repo.branches
 
         # Calling it a second time should also succeed
         workspace = self.repo._maybe_prepare_merge_workspace(
             'pr2', Reference('branch', 'master', 'unused'))
         assert os.path.isdir(workspace)
 
     def test_cleanup_merge_workspace(self):
         workspace = self.repo._maybe_prepare_merge_workspace(
             'pr3', Reference('branch', 'master', 'unused'))
         self.repo.cleanup_merge_workspace('pr3')
 
         assert not os.path.exists(workspace)
 
     def test_cleanup_merge_workspace_invalid_workspace_id(self):
         # No assert: because in case of an inexistent workspace this function
         # should still succeed.
         self.repo.cleanup_merge_workspace('pr4')
 
     def test_set_refs(self):
         test_ref = 'refs/test-refs/abcde'
         test_commit_id = 'ecb86e1f424f2608262b130db174a7dfd25a6623'
 
         self.repo.set_refs(test_ref, test_commit_id)
         stdout, _ = self.repo.run_git_command(['show-ref'])
         assert test_ref in stdout
         assert test_commit_id in stdout
 
     def test_remove_ref(self):
         test_ref = 'refs/test-refs/abcde'
         test_commit_id = 'ecb86e1f424f2608262b130db174a7dfd25a6623'
         self.repo.set_refs(test_ref, test_commit_id)
         stdout, _ = self.repo.run_git_command(['show-ref'])
         assert test_ref in stdout
         assert test_commit_id in stdout
 
         self.repo.remove_ref(test_ref)
         stdout, _ = self.repo.run_git_command(['show-ref'])
         assert test_ref not in stdout
         assert test_commit_id not in stdout
 
 
 class TestGitCommit(object):
 
     @pytest.fixture(autouse=True)
     def prepare(self):
         self.repo = GitRepository(TEST_GIT_REPO)
 
     def test_default_commit(self):
         tip = self.repo.get_commit()
         assert tip == self.repo.get_commit(None)
         assert tip == self.repo.get_commit('tip')
 
     def test_root_node(self):
         tip = self.repo.get_commit()
         assert tip.root is tip.get_node('')
 
     def test_lazy_fetch(self):
         """
         Test if commit's nodes expands and are cached as we walk through
         the commit. This test is somewhat hard to write as order of tests
         is a key here. Written by running command after command in a shell.
         """
         commit_id = '2a13f185e4525f9d4b59882791a2d397b90d5ddc'
         assert commit_id in self.repo.commit_ids
         commit = self.repo.get_commit(commit_id)
         assert len(commit.nodes) == 0
         root = commit.root
         assert len(commit.nodes) == 1
         assert len(root.nodes) == 8
         # accessing root.nodes updates commit.nodes
         assert len(commit.nodes) == 9
 
         docs = root.get_node('docs')
         # we haven't yet accessed anything new as docs dir was already cached
         assert len(commit.nodes) == 9
         assert len(docs.nodes) == 8
         # accessing docs.nodes updates commit.nodes
         assert len(commit.nodes) == 17
 
         assert docs is commit.get_node('docs')
         assert docs is root.nodes[0]
         assert docs is root.dirs[0]
         assert docs is commit.get_node('docs')
 
     def test_nodes_with_commit(self):
         commit_id = '2a13f185e4525f9d4b59882791a2d397b90d5ddc'
         commit = self.repo.get_commit(commit_id)
         root = commit.root
         docs = root.get_node('docs')
         assert docs is commit.get_node('docs')
         api = docs.get_node('api')
         assert api is commit.get_node('docs/api')
         index = api.get_node('index.rst')
         assert index is commit.get_node('docs/api/index.rst')
         assert index is commit.get_node('docs')\
             .get_node('api')\
             .get_node('index.rst')
 
     def test_branch_and_tags(self):
         """
         rev0 = self.repo.commit_ids[0]
         commit0 = self.repo.get_commit(rev0)
         assert commit0.branch == 'master'
         assert commit0.tags == []
 
         rev10 = self.repo.commit_ids[10]
         commit10 = self.repo.get_commit(rev10)
         assert commit10.branch == 'master'
         assert commit10.tags == []
 
         rev44 = self.repo.commit_ids[44]
         commit44 = self.repo.get_commit(rev44)
         assert commit44.branch == 'web-branch'
 
         tip = self.repo.get_commit('tip')
         assert 'tip' in tip.tags
         """
         # Those tests would fail - branches are now going
         # to be changed at main API in order to support git backend
         pass
 
     def test_file_size(self):
         to_check = (
             ('c1214f7e79e02fc37156ff215cd71275450cffc3',
                 'vcs/backends/BaseRepository.py', 502),
             ('d7e0d30fbcae12c90680eb095a4f5f02505ce501',
                 'vcs/backends/hg.py', 854),
             ('6e125e7c890379446e98980d8ed60fba87d0f6d1',
                 'setup.py', 1068),
 
             ('d955cd312c17b02143c04fa1099a352b04368118',
                 'vcs/backends/base.py', 2921),
             ('ca1eb7957a54bce53b12d1a51b13452f95bc7c7e',
                 'vcs/backends/base.py', 3936),
             ('f50f42baeed5af6518ef4b0cb2f1423f3851a941',
                 'vcs/backends/base.py', 6189),
         )
         for commit_id, path, size in to_check:
             node = self.repo.get_commit(commit_id).get_node(path)
             assert node.is_file()
             assert node.size == size
 
     def test_file_history_from_commits(self):
         node = self.repo[10].get_node('setup.py')
         commit_ids = [commit.raw_id for commit in node.history]
         assert ['ff7ca51e58c505fec0dd2491de52c622bb7a806b'] == commit_ids
 
         node = self.repo[20].get_node('setup.py')
         node_ids = [commit.raw_id for commit in node.history]
         assert ['191caa5b2c81ed17c0794bf7bb9958f4dcb0b87e',
                 'ff7ca51e58c505fec0dd2491de52c622bb7a806b'] == node_ids
 
         # special case we check history from commit that has this particular
         # file changed this means we check if it's included as well
         node = self.repo.get_commit('191caa5b2c81ed17c0794bf7bb9958f4dcb0b87e') \
             .get_node('setup.py')
         node_ids = [commit.raw_id for commit in node.history]
         assert ['191caa5b2c81ed17c0794bf7bb9958f4dcb0b87e',
                 'ff7ca51e58c505fec0dd2491de52c622bb7a806b'] == node_ids
 
     def test_file_history(self):
         # we can only check if those commits are present in the history
         # as we cannot update this test every time file is changed
         files = {
             'setup.py': [
                 '54386793436c938cff89326944d4c2702340037d',
                 '51d254f0ecf5df2ce50c0b115741f4cf13985dab',
                 '998ed409c795fec2012b1c0ca054d99888b22090',
                 '5e0eb4c47f56564395f76333f319d26c79e2fb09',
                 '0115510b70c7229dbc5dc49036b32e7d91d23acd',
                 '7cb3fd1b6d8c20ba89e2264f1c8baebc8a52d36e',
                 '2a13f185e4525f9d4b59882791a2d397b90d5ddc',
                 '191caa5b2c81ed17c0794bf7bb9958f4dcb0b87e',
                 'ff7ca51e58c505fec0dd2491de52c622bb7a806b',
             ],
             'vcs/nodes.py': [
                 '33fa3223355104431402a888fa77a4e9956feb3e',
                 'fa014c12c26d10ba682fadb78f2a11c24c8118e1',
                 'e686b958768ee96af8029fe19c6050b1a8dd3b2b',
                 'ab5721ca0a081f26bf43d9051e615af2cc99952f',
                 'c877b68d18e792a66b7f4c529ea02c8f80801542',
                 '4313566d2e417cb382948f8d9d7c765330356054',
                 '6c2303a793671e807d1cfc70134c9ca0767d98c2',
                 '54386793436c938cff89326944d4c2702340037d',
                 '54000345d2e78b03a99d561399e8e548de3f3203',
                 '1c6b3677b37ea064cb4b51714d8f7498f93f4b2b',
                 '2d03ca750a44440fb5ea8b751176d1f36f8e8f46',
                 '2a08b128c206db48c2f0b8f70df060e6db0ae4f8',
                 '30c26513ff1eb8e5ce0e1c6b477ee5dc50e2f34b',
                 'ac71e9503c2ca95542839af0ce7b64011b72ea7c',
                 '12669288fd13adba2a9b7dd5b870cc23ffab92d2',
                 '5a0c84f3e6fe3473e4c8427199d5a6fc71a9b382',
                 '12f2f5e2b38e6ff3fbdb5d722efed9aa72ecb0d5',
                 '5eab1222a7cd4bfcbabc218ca6d04276d4e27378',
                 'f50f42baeed5af6518ef4b0cb2f1423f3851a941',
                 'd7e390a45f6aa96f04f5e7f583ad4f867431aa25',
                 'f15c21f97864b4f071cddfbf2750ec2e23859414',
                 'e906ef056cf539a4e4e5fc8003eaf7cf14dd8ade',
                 'ea2b108b48aa8f8c9c4a941f66c1a03315ca1c3b',
                 '84dec09632a4458f79f50ddbbd155506c460b4f9',
                 '0115510b70c7229dbc5dc49036b32e7d91d23acd',
                 '2a13f185e4525f9d4b59882791a2d397b90d5ddc',
                 '3bf1c5868e570e39569d094f922d33ced2fa3b2b',
                 'b8d04012574729d2c29886e53b1a43ef16dd00a1',
                 '6970b057cffe4aab0a792aa634c89f4bebf01441',
                 'dd80b0f6cf5052f17cc738c2951c4f2070200d7f',
                 'ff7ca51e58c505fec0dd2491de52c622bb7a806b',
             ],
             'vcs/backends/git.py': [
                 '4cf116ad5a457530381135e2f4c453e68a1b0105',
                 '9a751d84d8e9408e736329767387f41b36935153',
                 'cb681fb539c3faaedbcdf5ca71ca413425c18f01',
                 '428f81bb652bcba8d631bce926e8834ff49bdcc6',
                 '180ab15aebf26f98f714d8c68715e0f05fa6e1c7',
                 '2b8e07312a2e89e92b90426ab97f349f4bce2a3a',
                 '50e08c506174d8645a4bb517dd122ac946a0f3bf',
                 '54000345d2e78b03a99d561399e8e548de3f3203',
             ],
         }
         for path, commit_ids in files.items():
             node = self.repo.get_commit(commit_ids[0]).get_node(path)
             node_ids = [commit.raw_id for commit in node.history]
             assert set(commit_ids).issubset(set(node_ids)), (
                 "We assumed that %s is subset of commit_ids for which file %s "
                 "has been changed, and history of that node returned: %s"
                 % (commit_ids, path, node_ids))
 
     def test_file_annotate(self):
         files = {
             'vcs/backends/__init__.py': {
                 'c1214f7e79e02fc37156ff215cd71275450cffc3': {
                     'lines_no': 1,
                     'commits': [
                         'c1214f7e79e02fc37156ff215cd71275450cffc3',
                     ],
                 },
                 '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647': {
                     'lines_no': 21,
                     'commits': [
                         '49d3fd156b6f7db46313fac355dca1a0b94a0017',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                     ],
                 },
                 'e29b67bd158580fc90fc5e9111240b90e6e86064': {
                     'lines_no': 32,
                     'commits': [
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '5eab1222a7cd4bfcbabc218ca6d04276d4e27378',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '54000345d2e78b03a99d561399e8e548de3f3203',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '78c3f0c23b7ee935ec276acb8b8212444c33c396',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '2a13f185e4525f9d4b59882791a2d397b90d5ddc',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '78c3f0c23b7ee935ec276acb8b8212444c33c396',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '992f38217b979d0b0987d0bae3cc26dac85d9b19',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                         '16fba1ae9334d79b66d7afed2c2dfbfa2ae53647',
                     ],
                 },
             },
         }
 
         for fname, commit_dict in files.items():
             for commit_id, __ in commit_dict.items():
                 commit = self.repo.get_commit(commit_id)
 
                 l1_1 = [x[1] for x in commit.get_file_annotate(fname)]
                 l1_2 = [x[2]().raw_id for x in commit.get_file_annotate(fname)]
                 assert l1_1 == l1_2
                 l1 = l1_1
                 l2 = files[fname][commit_id]['commits']
                 assert l1 == l2, (
                     "The lists of commit_ids for %s@commit_id %s"
                     "from annotation list should match each other, "
                     "got \n%s \nvs \n%s " % (fname, commit_id, l1, l2))
 
     def test_files_state(self):
         """
         Tests state of FileNodes.
         """
         node = self.repo\
             .get_commit('e6ea6d16e2f26250124a1f4b4fe37a912f9d86a0')\
             .get_node('vcs/utils/diffs.py')
         assert node.state, NodeState.ADDED
         assert node.added
         assert not node.changed
         assert not node.not_changed
         assert not node.removed
 
         node = self.repo\
             .get_commit('33fa3223355104431402a888fa77a4e9956feb3e')\
             .get_node('.hgignore')
         assert node.state, NodeState.CHANGED
         assert not node.added
         assert node.changed
         assert not node.not_changed
         assert not node.removed
 
         node = self.repo\
             .get_commit('e29b67bd158580fc90fc5e9111240b90e6e86064')\
             .get_node('setup.py')
         assert node.state, NodeState.NOT_CHANGED
         assert not node.added
         assert not node.changed
         assert node.not_changed
         assert not node.removed
 
         # If node has REMOVED state then trying to fetch it would raise
         # CommitError exception
         commit = self.repo.get_commit(
             'fa6600f6848800641328adbf7811fd2372c02ab2')
         path = 'vcs/backends/BaseRepository.py'
         with pytest.raises(NodeDoesNotExistError):
             commit.get_node(path)
         # but it would be one of ``removed`` (commit's attribute)
         assert path in [rf.path for rf in commit.removed]
 
         commit = self.repo.get_commit(
             '54386793436c938cff89326944d4c2702340037d')
         changed = [
             'setup.py', 'tests/test_nodes.py', 'vcs/backends/hg.py',
             'vcs/nodes.py']
         assert set(changed) == set([f.path for f in commit.changed])
 
-    def test_unicode_refs(self):
+    def test_unicode_branch_refs(self):
         unicode_branches = {
-            'unicode': ['6c0ce52b229aa978889e91b38777f800e85f330b', 'H'],
-            u'uniçö∂e': ['ürl', 'H']
+            'refs/heads/unicode': '6c0ce52b229aa978889e91b38777f800e85f330b',
+            u'refs/heads/uniçö∂e': 'ürl',
         }
         with mock.patch(
             ("rhodecode.lib.vcs.backends.git.repository"
-                ".GitRepository._parsed_refs"),
+                ".GitRepository._refs"),
                 unicode_branches):
             branches = self.repo.branches
 
         assert 'unicode' in branches
         assert u'uniçö∂e' in branches
 
+    def test_unicode_tag_refs(self):
+        unicode_tags = {
+            'refs/tags/unicode': '6c0ce52b229aa978889e91b38777f800e85f330b',
+            u'refs/tags/uniçö∂e': '6c0ce52b229aa978889e91b38777f800e85f330b',
+        }
+        with mock.patch(
+            ("rhodecode.lib.vcs.backends.git.repository"
+                ".GitRepository._refs"),
+                unicode_tags):
+            tags = self.repo.tags
+
+        assert 'unicode' in tags
+        assert u'uniçö∂e' in tags
+
     def test_commit_message_is_unicode(self):
         for commit in self.repo:
             assert type(commit.message) == unicode
 
     def test_commit_author_is_unicode(self):
         for commit in self.repo:
             assert type(commit.author) == unicode
 
     def test_repo_files_content_is_unicode(self):
         commit = self.repo.get_commit()
         for node in commit.get_node('/'):
             if node.is_file():
                 assert type(node.content) == unicode
 
     def test_wrong_path(self):
         # There is 'setup.py' in the root dir but not there:
         path = 'foo/bar/setup.py'
         tip = self.repo.get_commit()
         with pytest.raises(VCSError):
             tip.get_node(path)
 
     @pytest.mark.parametrize("author_email, commit_id", [
         ('[e-mail]', 'c1214f7e79e02fc37156ff215cd71275450cffc3'),
         ('[e-mail]',
          'ff7ca51e58c505fec0dd2491de52c622bb7a806b'),
         ('none@none', '8430a588b43b5d6da365400117c89400326e7992'),
     ])
     def test_author_email(self, author_email, commit_id):
         commit = self.repo.get_commit(commit_id)
         assert author_email == commit.author_email
 
     @pytest.mark.parametrize("author, commit_id", [
         ('Marcin Kuzminski', 'c1214f7e79e02fc37156ff215cd71275450cffc3'),
         ('Lukasz Balcerzak', 'ff7ca51e58c505fec0dd2491de52c622bb7a806b'),
         ('marcink', '8430a588b43b5d6da365400117c89400326e7992'),
     ])
     def test_author_username(self, author, commit_id):
         commit = self.repo.get_commit(commit_id)
         assert author == commit.author_name
 
 
 class TestGitSpecificWithRepo(BackendTestMixin):
 
     @classmethod
     def _get_commits(cls):
         return [
             {
                 'message': 'Initial',
                 'author': 'Joe Doe <[e-mail]>',
                 'date': datetime.datetime(2010, 1, 1, 20),
                 'added': [
                     FileNode('foobar/static/js/admin/base.js', content='base'),
                     FileNode(
                         'foobar/static/admin', content='admin',
                         mode=0120000),  # this is a link
                     FileNode('foo', content='foo'),
                 ],
             },
             {
                 'message': 'Second',
                 'author': 'Joe Doe <[e-mail]>',
                 'date': datetime.datetime(2010, 1, 1, 22),
                 'added': [
                     FileNode('foo2', content='foo2'),
                 ],
             },
         ]
 
     def test_paths_slow_traversing(self):
         commit = self.repo.get_commit()
         assert commit.get_node('foobar').get_node('static').get_node('js')\
             .get_node('admin').get_node('base.js').content == 'base'
 
     def test_paths_fast_traversing(self):
         commit = self.repo.get_commit()
         assert (
             commit.get_node('foobar/static/js/admin/base.js').content ==
             'base')
 
     def test_get_diff_runs_git_command_with_hashes(self):
         self.repo.run_git_command = mock.Mock(return_value=['', ''])
         self.repo.get_diff(self.repo[0], self.repo[1])
         self.repo.run_git_command.assert_called_once_with(
             ['diff', '-U3', '--full-index', '--binary', '-p', '-M',
              '--abbrev=40', self.repo._get_commit_id(0),
              self.repo._get_commit_id(1)])
 
     def test_get_diff_runs_git_command_with_str_hashes(self):
         self.repo.run_git_command = mock.Mock(return_value=['', ''])
         self.repo.get_diff(self.repo.EMPTY_COMMIT, self.repo[1])
         self.repo.run_git_command.assert_called_once_with(
             ['show', '-U3', '--full-index', '--binary', '-p', '-M',
              '--abbrev=40', self.repo._get_commit_id(1)])
 
     def test_get_diff_runs_git_command_with_path_if_its_given(self):
         self.repo.run_git_command = mock.Mock(return_value=['', ''])
         self.repo.get_diff(self.repo[0], self.repo[1], 'foo')
         self.repo.run_git_command.assert_called_once_with(
             ['diff', '-U3', '--full-index', '--binary', '-p', '-M',
              '--abbrev=40', self.repo._get_commit_id(0),
              self.repo._get_commit_id(1), '--', 'foo'])
 
 
 class TestGitRegression(BackendTestMixin):
 
     @classmethod
     def _get_commits(cls):
         return [
             {
                 'message': 'Initial',
                 'author': 'Joe Doe <[e-mail]>',
                 'date': datetime.datetime(2010, 1, 1, 20),
                 'added': [
                     FileNode('bot/__init__.py', content='base'),
                     FileNode('bot/templates/404.html', content='base'),
                     FileNode('bot/templates/500.html', content='base'),
                 ],
             },
             {
                 'message': 'Second',
                 'author': 'Joe Doe <[e-mail]>',
                 'date': datetime.datetime(2010, 1, 1, 22),
                 'added': [
                     FileNode('bot/build/migrations/1.py', content='foo2'),
                     FileNode('bot/build/migrations/2.py', content='foo2'),
                     FileNode(
                         'bot/build/static/templates/f.html', content='foo2'),
                     FileNode(
                         'bot/build/static/templates/f1.html', content='foo2'),
                     FileNode('bot/build/templates/err.html', content='foo2'),
                     FileNode('bot/build/templates/err2.html', content='foo2'),
                 ],
             },
         ]
 
     @pytest.mark.parametrize("path, expected_paths", [
         ('bot', [
             'bot/build',
             'bot/templates',
             'bot/__init__.py']),
         ('bot/build', [
             'bot/build/migrations',
             'bot/build/static',
             'bot/build/templates']),
         ('bot/build/static', [
             'bot/build/static/templates']),
         ('bot/build/static/templates', [
             'bot/build/static/templates/f.html',
             'bot/build/static/templates/f1.html']),
         ('bot/build/templates', [
             'bot/build/templates/err.html',
             'bot/build/templates/err2.html']),
         ('bot/templates/', [
             'bot/templates/404.html',
             'bot/templates/500.html']),
     ])
     def test_similar_paths(self, path, expected_paths):
         commit = self.repo.get_commit()
         paths = [n.path for n in commit.get_nodes(path)]
         assert paths == expected_paths
 
 
 class TestDiscoverGitVersion:
 
     def test_returns_git_version(self, pylonsapp):
         version = discover_git_version()
         assert version
 
     def test_returns_empty_string_without_vcsserver(self):
         mock_connection = mock.Mock()
         mock_connection.discover_git_version = mock.Mock(
             side_effect=Exception)
         with mock.patch('rhodecode.lib.vcs.connection.Git', mock_connection):
             version = discover_git_version()
         assert version == ''
 
 
 class TestGetSubmoduleUrl(object):
     def test_submodules_file_found(self):
         commit = GitCommit(repository=mock.Mock(), raw_id='abcdef12', idx=1)
         node = mock.Mock()
         with mock.patch.object(
                 commit, 'get_node', return_value=node) as get_node_mock:
             node.content = (
                 '[submodule "subrepo1"]\n'
                 '\tpath = subrepo1\n'
                 '\turl = https://code.rhodecode.com/dulwich\n'
             )
             result = commit._get_submodule_url('subrepo1')
         get_node_mock.assert_called_once_with('.gitmodules')
         assert result == 'https://code.rhodecode.com/dulwich'
 
     def test_complex_submodule_path(self):
         commit = GitCommit(repository=mock.Mock(), raw_id='abcdef12', idx=1)
         node = mock.Mock()
         with mock.patch.object(
                 commit, 'get_node', return_value=node) as get_node_mock:
             node.content = (
                 '[submodule "complex/subrepo/path"]\n'
                 '\tpath = complex/subrepo/path\n'
                 '\turl = https://code.rhodecode.com/dulwich\n'
             )
             result = commit._get_submodule_url('complex/subrepo/path')
         get_node_mock.assert_called_once_with('.gitmodules')
         assert result == 'https://code.rhodecode.com/dulwich'
 
     def test_submodules_file_not_found(self):
         commit = GitCommit(repository=mock.Mock(), raw_id='abcdef12', idx=1)
         with mock.patch.object(
                 commit, 'get_node', side_effect=NodeDoesNotExistError):
             result = commit._get_submodule_url('complex/subrepo/path')
         assert result is None
 
     def test_path_not_found(self):
         commit = GitCommit(repository=mock.Mock(), raw_id='abcdef12', idx=1)
         node = mock.Mock()
         with mock.patch.object(
                 commit, 'get_node', return_value=node) as get_node_mock:
             node.content = (
                 '[submodule "subrepo1"]\n'
                 '\tpath = subrepo1\n'
                 '\turl = https://code.rhodecode.com/dulwich\n'
             )
             result = commit._get_submodule_url('subrepo2')
         get_node_mock.assert_called_once_with('.gitmodules')
         assert result is None
 
     def test_returns_cached_values(self):
         commit = GitCommit(repository=mock.Mock(), raw_id='abcdef12', idx=1)
         node = mock.Mock()
         with mock.patch.object(
                 commit, 'get_node', return_value=node) as get_node_mock:
             node.content = (
                 '[submodule "subrepo1"]\n'
                 '\tpath = subrepo1\n'
                 '\turl = https://code.rhodecode.com/dulwich\n'
             )
             for _ in range(3):
                 commit._get_submodule_url('subrepo1')
         get_node_mock.assert_called_once_with('.gitmodules')
 
     def test_get_node_returns_a_link(self):
         repository = mock.Mock()
         repository.alias = 'git'
         commit = GitCommit(repository=repository, raw_id='abcdef12', idx=1)
         submodule_url = 'https://code.rhodecode.com/dulwich'
         get_id_patch = mock.patch.object(
             commit, '_get_id_for_path', return_value=(1, 'link'))
         get_submodule_patch = mock.patch.object(
             commit, '_get_submodule_url', return_value=submodule_url)
 
         with get_id_patch, get_submodule_patch as submodule_mock:
             node = commit.get_node('/abcde')
 
         submodule_mock.assert_called_once_with('/abcde')
         assert type(node) == SubModuleNode
         assert node.url == submodule_url
 
     def test_get_nodes_returns_links(self):
         repository = mock.MagicMock()
         repository.alias = 'git'
         repository._remote.tree_items.return_value = [
             ('subrepo', 'stat', 1, 'link')
         ]
         commit = GitCommit(repository=repository, raw_id='abcdef12', idx=1)
         submodule_url = 'https://code.rhodecode.com/dulwich'
         get_id_patch = mock.patch.object(
             commit, '_get_id_for_path', return_value=(1, 'tree'))
         get_submodule_patch = mock.patch.object(
             commit, '_get_submodule_url', return_value=submodule_url)
 
         with get_id_patch, get_submodule_patch as submodule_mock:
             nodes = commit.get_nodes('/abcde')
 
         submodule_mock.assert_called_once_with('/abcde/subrepo')
         assert len(nodes) == 1
         assert type(nodes[0]) == SubModuleNode
         assert nodes[0].url == submodule_url