git: fixed readme searcher for libgit2

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

@@ -1,493 +1,494 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2014-2020 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 commit module
 """
 
 import re
 import stat
 from itertools import chain
 from StringIO import StringIO
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 from rhodecode.lib.datelib import utcdate_fromtimestamp
 from rhodecode.lib.utils import safe_unicode, safe_str
 from rhodecode.lib.utils2 import safe_int
 from rhodecode.lib.vcs.conf import settings
 from rhodecode.lib.vcs.backends import base
 from rhodecode.lib.vcs.exceptions import CommitError, NodeDoesNotExistError
 from rhodecode.lib.vcs.nodes import (
     FileNode, DirNode, NodeKind, RootNode, SubModuleNode,
     ChangedFileNodesGenerator, AddedFileNodesGenerator,
     RemovedFileNodesGenerator, LargeFileNode)
 from rhodecode.lib.vcs.compat import configparser
 
 
 class GitCommit(base.BaseCommit):
     """
     Represents state of the repository at single commit id.
     """
 
     _filter_pre_load = [
         # done through a more complex tree walk on parents
         "affected_files",
         # done through subprocess not remote call
         "children",
         # done through a more complex tree walk on parents
         "status",
         # mercurial specific property not supported here
         "_file_paths",
         # mercurial specific property not supported here
         'obsolete',
         # mercurial specific property not supported here
         'phase',
         # mercurial specific property not supported here
         'hidden'
     ]
 
     def __init__(self, repository, raw_id, idx, pre_load=None):
         self.repository = repository
         self._remote = repository._remote
         # TODO: johbo: Tweak of raw_id should not be necessary
         self.raw_id = safe_str(raw_id)
         self.idx = idx
 
         self._set_bulk_properties(pre_load)
 
         # caches
         self._stat_modes = {}  # stat info for paths
         self._paths = {}  # path processed with parse_tree
         self.nodes = {}
         self._submodules = None
 
     def _set_bulk_properties(self, pre_load):
 
         if not pre_load:
             return
         pre_load = [entry for entry in pre_load
                     if entry not in self._filter_pre_load]
         if not pre_load:
             return
 
         result = self._remote.bulk_request(self.raw_id, pre_load)
         for attr, value in result.items():
             if attr in ["author", "message"]:
                 if value:
                     value = safe_unicode(value)
             elif attr == "date":
                 value = utcdate_fromtimestamp(*value)
             elif attr == "parents":
                 value = self._make_commits(value)
             elif attr == "branch":
                 value = value[0] if value else None
             self.__dict__[attr] = value
 
     @LazyProperty
     def _commit(self):
         return self._remote[self.raw_id]
 
     @LazyProperty
     def _tree_id(self):
         return self._remote[self._commit['tree']]['id']
 
     @LazyProperty
     def id(self):
         return self.raw_id
 
     @LazyProperty
     def short_id(self):
         return self.raw_id[:12]
 
     @LazyProperty
     def message(self):
         return safe_unicode(self._remote.message(self.id))
 
     @LazyProperty
     def committer(self):
         return safe_unicode(self._remote.author(self.id))
 
     @LazyProperty
     def author(self):
         return safe_unicode(self._remote.author(self.id))
 
     @LazyProperty
     def date(self):
         unix_ts, tz = self._remote.date(self.raw_id)
         return utcdate_fromtimestamp(unix_ts, tz)
 
     @LazyProperty
     def status(self):
         """
         Returns modified, added, removed, deleted files for current commit
         """
         return self.changed, self.added, self.removed
 
     @LazyProperty
     def tags(self):
         tags = [safe_unicode(name) for name,
                 commit_id in self.repository.tags.iteritems()
                 if commit_id == self.raw_id]
         return tags
 
     @LazyProperty
     def commit_branches(self):
         branches = []
         for name, commit_id in self.repository.branches.iteritems():
             if commit_id == self.raw_id:
                 branches.append(name)
         return branches
 
     @LazyProperty
     def branch(self):
         branches = self._remote.branch(self.raw_id)
 
         if branches:
             # actually commit can have multiple branches in git
             return safe_unicode(branches[0])
 
     def _get_tree_id_for_path(self, path):
         path = safe_str(path)
         if path in self._paths:
             return self._paths[path]
 
         tree_id = self._tree_id
 
         path = path.strip('/')
         if path == '':
             data = [tree_id, "tree"]
             self._paths[''] = data
             return data
 
         tree_id, tree_type, tree_mode = \
             self._remote.tree_and_type_for_path(self.raw_id, path)
         if tree_id is None:
             raise self.no_node_at_path(path)
 
         self._paths[path] = [tree_id, tree_type]
         self._stat_modes[path] = tree_mode
 
         if path not in self._paths:
             raise self.no_node_at_path(path)
 
         return self._paths[path]
 
     def _get_kind(self, path):
         tree_id, type_ = self._get_tree_id_for_path(path)
         if type_ == 'blob':
             return NodeKind.FILE
         elif type_ == 'tree':
             return NodeKind.DIR
         elif type_ == 'link':
             return NodeKind.SUBMODULE
         return None
 
     def _get_filectx(self, path):
         path = self._fix_path(path)
         if self._get_kind(path) != NodeKind.FILE:
             raise CommitError(
                 "File does not exist for commit %s at  '%s'" % (self.raw_id, path))
         return path
 
     def _get_file_nodes(self):
         return chain(*(t[2] for t in self.walk()))
 
     @LazyProperty
     def parents(self):
         """
         Returns list of parent commits.
         """
         parent_ids = self._remote.parents(self.id)
         return self._make_commits(parent_ids)
 
     @LazyProperty
     def children(self):
         """
         Returns list of child commits.
         """
 
         children = self._remote.children(self.raw_id)
         return self._make_commits(children)
 
     def _make_commits(self, commit_ids):
         def commit_maker(_commit_id):
             return self.repository.get_commit(commit_id=commit_id)
 
         return [commit_maker(commit_id) for commit_id in commit_ids]
 
     def get_file_mode(self, path):
         """
         Returns stat mode of the file at the given `path`.
         """
         path = safe_str(path)
         # ensure path is traversed
         self._get_tree_id_for_path(path)
         return self._stat_modes[path]
 
     def is_link(self, path):
         return stat.S_ISLNK(self.get_file_mode(path))
 
     def is_node_binary(self, path):
         tree_id, _ = self._get_tree_id_for_path(path)
         return self._remote.is_binary(tree_id)
 
     def get_file_content(self, path):
         """
         Returns content of the file at given `path`.
         """
         tree_id, _ = self._get_tree_id_for_path(path)
         return self._remote.blob_as_pretty_string(tree_id)
 
     def get_file_content_streamed(self, path):
         tree_id, _ = self._get_tree_id_for_path(path)
         stream_method = getattr(self._remote, 'stream:blob_as_pretty_string')
         return stream_method(tree_id)
 
     def get_file_size(self, path):
         """
         Returns size of the file at given `path`.
         """
         tree_id, _ = self._get_tree_id_for_path(path)
         return self._remote.blob_raw_length(tree_id)
 
     def get_path_history(self, path, limit=None, pre_load=None):
         """
         Returns history of file as reversed list of `GitCommit` objects for
         which file at given `path` has been modified.
         """
 
         path = self._get_filectx(path)
         hist = self._remote.node_history(self.raw_id, path, limit)
         return [
             self.repository.get_commit(commit_id=commit_id, pre_load=pre_load)
             for commit_id in hist]
 
     def get_file_annotate(self, path, pre_load=None):
         """
         Returns a generator of four element tuples with
             lineno, commit_id, commit lazy loader and line
         """
 
         result = self._remote.node_annotate(self.raw_id, path)
 
         for ln_no, commit_id, content in result:
             yield (
                 ln_no, commit_id,
                 lambda: self.repository.get_commit(commit_id=commit_id, pre_load=pre_load),
                 content)
 
     def get_nodes(self, path):
 
         if self._get_kind(path) != NodeKind.DIR:
             raise CommitError(
                 "Directory does not exist for commit %s at '%s'" % (self.raw_id, path))
         path = self._fix_path(path)
 
         tree_id, _ = self._get_tree_id_for_path(path)
 
         dirnodes = []
         filenodes = []
 
         # extracted tree ID gives us our files...
+        bytes_path = safe_str(path)  # libgit operates on bytes
         for name, stat_, id_, type_ in self._remote.tree_items(tree_id):
             if type_ == 'link':
-                url = self._get_submodule_url('/'.join((path, name)))
+                url = self._get_submodule_url('/'.join((bytes_path, name)))
                 dirnodes.append(SubModuleNode(
                     name, url=url, commit=id_, alias=self.repository.alias))
                 continue
 
-            if path != '':
-                obj_path = '/'.join((path, name))
+            if bytes_path != '':
+                obj_path = '/'.join((bytes_path, name))
             else:
                 obj_path = name
             if obj_path not in self._stat_modes:
                 self._stat_modes[obj_path] = stat_
 
             if type_ == 'tree':
                 dirnodes.append(DirNode(obj_path, commit=self))
             elif type_ == 'blob':
                 filenodes.append(FileNode(obj_path, commit=self, mode=stat_))
             else:
                 raise CommitError(
                     "Requested object should be Tree or Blob, is %s", type_)
 
         nodes = dirnodes + filenodes
         for node in nodes:
             if node.path not in self.nodes:
                 self.nodes[node.path] = node
         nodes.sort()
         return nodes
 
     def get_node(self, path, pre_load=None):
         if isinstance(path, unicode):
             path = path.encode('utf-8')
         path = self._fix_path(path)
         if path not in self.nodes:
             try:
                 tree_id, type_ = self._get_tree_id_for_path(path)
             except CommitError:
                 raise NodeDoesNotExistError(
                     "Cannot find one of parents' directories for a given "
                     "path: %s" % path)
 
             if type_ in ['link', 'commit']:
                 url = self._get_submodule_url(path)
                 node = SubModuleNode(path, url=url, commit=tree_id,
                                      alias=self.repository.alias)
             elif type_ == 'tree':
                 if path == '':
                     node = RootNode(commit=self)
                 else:
                     node = DirNode(path, commit=self)
             elif type_ == 'blob':
                 node = FileNode(path, commit=self, pre_load=pre_load)
                 self._stat_modes[path] = node.mode
             else:
                 raise self.no_node_at_path(path)
 
             # cache node
             self.nodes[path] = node
 
         return self.nodes[path]
 
     def get_largefile_node(self, path):
         tree_id, _ = self._get_tree_id_for_path(path)
         pointer_spec = self._remote.is_large_file(tree_id)
 
         if pointer_spec:
             # content of that file regular FileNode is the hash of largefile
             file_id = pointer_spec.get('oid_hash')
             if self._remote.in_largefiles_store(file_id):
                 lf_path = self._remote.store_path(file_id)
                 return LargeFileNode(lf_path, commit=self, org_path=path)
 
     @LazyProperty
     def affected_files(self):
         """
         Gets a fast accessible file changes for given commit
         """
         added, modified, deleted = self._changes_cache
         return list(added.union(modified).union(deleted))
 
     @LazyProperty
     def _changes_cache(self):
         added = set()
         modified = set()
         deleted = set()
         _r = self._remote
 
         parents = self.parents
         if not self.parents:
             parents = [base.EmptyCommit()]
         for parent in parents:
             if isinstance(parent, base.EmptyCommit):
                 oid = None
             else:
                 oid = parent.raw_id
             changes = _r.tree_changes(oid, self.raw_id)
             for (oldpath, newpath), (_, _), (_, _) in changes:
                 if newpath and oldpath:
                     modified.add(newpath)
                 elif newpath and not oldpath:
                     added.add(newpath)
                 elif not newpath and oldpath:
                     deleted.add(oldpath)
         return added, modified, deleted
 
     def _get_paths_for_status(self, status):
         """
         Returns sorted list of paths for given ``status``.
 
         :param status: one of: *added*, *modified* or *deleted*
         """
         added, modified, deleted = self._changes_cache
         return sorted({
             'added': list(added),
             'modified': list(modified),
             'deleted': list(deleted)}[status]
         )
 
     @LazyProperty
     def added(self):
         """
         Returns list of added ``FileNode`` objects.
         """
         if not self.parents:
             return list(self._get_file_nodes())
         return AddedFileNodesGenerator(self.added_paths, self)
 
     @LazyProperty
     def added_paths(self):
         return [n for n in self._get_paths_for_status('added')]
 
     @LazyProperty
     def changed(self):
         """
         Returns list of modified ``FileNode`` objects.
         """
         if not self.parents:
             return []
         return ChangedFileNodesGenerator(self.changed_paths, self)
 
     @LazyProperty
     def changed_paths(self):
         return [n for n in self._get_paths_for_status('modified')]
 
     @LazyProperty
     def removed(self):
         """
         Returns list of removed ``FileNode`` objects.
         """
         if not self.parents:
             return []
         return RemovedFileNodesGenerator(self.removed_paths, self)
 
     @LazyProperty
     def removed_paths(self):
         return [n for n in self._get_paths_for_status('deleted')]
 
     def _get_submodule_url(self, submodule_path):
         git_modules_path = '.gitmodules'
 
         if self._submodules is None:
             self._submodules = {}
 
             try:
                 submodules_node = self.get_node(git_modules_path)
             except NodeDoesNotExistError:
                 return None
 
             # ConfigParser fails if there are whitespaces, also it needs an iterable
             # file like content
             def iter_content(_content):
                 for line in _content.splitlines():
                     yield line
 
             parser = configparser.RawConfigParser()
             parser.read_file(iter_content(submodules_node.content))
 
             for section in parser.sections():
                 path = parser.get(section, 'path')
                 url = parser.get(section, 'url')
                 if path and url:
                     self._submodules[path.strip('/')] = url
 
         return self._submodules.get(submodule_path.strip('/'))

rhodecode/model/repo.py

@@ -1,1172 +1,1172 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2010-2020 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import os
 import re
 import shutil
 import time
 import logging
 import traceback
 import datetime
 
 from pyramid.threadlocal import get_current_request
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 from rhodecode import events
 from rhodecode.lib.auth import HasUserGroupPermissionAny
 from rhodecode.lib.caching_query import FromCache
 from rhodecode.lib.exceptions import AttachedForksError, AttachedPullRequestsError
 from rhodecode.lib.hooks_base import log_delete_repository
 from rhodecode.lib.user_log_filter import user_log_filter
 from rhodecode.lib.utils import make_db_config
 from rhodecode.lib.utils2 import (
     safe_str, safe_unicode, remove_prefix, obfuscate_url_pw,
     get_current_rhodecode_user, safe_int, action_logger_generic)
 from rhodecode.lib.vcs.backends import get_backend
 from rhodecode.model import BaseModel
 from rhodecode.model.db import (
     _hash_key, func, case, joinedload, or_, in_filter_generator,
     Session, Repository, UserRepoToPerm, UserGroupRepoToPerm,
     UserRepoGroupToPerm, UserGroupRepoGroupToPerm, User, Permission,
     Statistics, UserGroup, RepoGroup, RepositoryField, UserLog)
 from rhodecode.model.settings import VcsSettingsModel
 
 log = logging.getLogger(__name__)
 
 
 class RepoModel(BaseModel):
 
     cls = Repository
 
     def _get_user_group(self, users_group):
         return self._get_instance(UserGroup, users_group,
                                   callback=UserGroup.get_by_group_name)
 
     def _get_repo_group(self, repo_group):
         return self._get_instance(RepoGroup, repo_group,
                                   callback=RepoGroup.get_by_group_name)
 
     def _create_default_perms(self, repository, private):
         # create default permission
         default = 'repository.read'
         def_user = User.get_default_user()
         for p in def_user.user_perms:
             if p.permission.permission_name.startswith('repository.'):
                 default = p.permission.permission_name
                 break
 
         default_perm = 'repository.none' if private else default
 
         repo_to_perm = UserRepoToPerm()
         repo_to_perm.permission = Permission.get_by_key(default_perm)
 
         repo_to_perm.repository = repository
         repo_to_perm.user_id = def_user.user_id
 
         return repo_to_perm
 
     @LazyProperty
     def repos_path(self):
         """
         Gets the repositories root path from database
         """
         settings_model = VcsSettingsModel(sa=self.sa)
         return settings_model.get_repos_location()
 
     def get(self, repo_id):
         repo = self.sa.query(Repository) \
             .filter(Repository.repo_id == repo_id)
 
         return repo.scalar()
 
     def get_repo(self, repository):
         return self._get_repo(repository)
 
     def get_by_repo_name(self, repo_name, cache=False):
         repo = self.sa.query(Repository) \
             .filter(Repository.repo_name == repo_name)
 
         if cache:
             name_key = _hash_key(repo_name)
             repo = repo.options(
                 FromCache("sql_cache_short", "get_repo_%s" % name_key))
         return repo.scalar()
 
     def _extract_id_from_repo_name(self, repo_name):
         if repo_name.startswith('/'):
             repo_name = repo_name.lstrip('/')
         by_id_match = re.match(r'^_(\d{1,})', repo_name)
         if by_id_match:
             return by_id_match.groups()[0]
 
     def get_repo_by_id(self, repo_name):
         """
         Extracts repo_name by id from special urls.
         Example url is _11/repo_name
 
         :param repo_name:
         :return: repo object if matched else None
         """
 
         try:
             _repo_id = self._extract_id_from_repo_name(repo_name)
             if _repo_id:
                 return self.get(_repo_id)
         except Exception:
             log.exception('Failed to extract repo_name from URL')
 
         return None
 
     def get_repos_for_root(self, root, traverse=False):
         if traverse:
             like_expression = u'{}%'.format(safe_unicode(root))
             repos = Repository.query().filter(
                 Repository.repo_name.like(like_expression)).all()
         else:
             if root and not isinstance(root, RepoGroup):
                 raise ValueError(
                     'Root must be an instance '
                     'of RepoGroup, got:{} instead'.format(type(root)))
             repos = Repository.query().filter(Repository.group == root).all()
         return repos
 
     def get_url(self, repo, request=None, permalink=False):
         if not request:
             request = get_current_request()
 
         if not request:
             return
 
         if permalink:
             return request.route_url(
                 'repo_summary', repo_name='_{}'.format(safe_str(repo.repo_id)))
         else:
             return request.route_url(
                 'repo_summary', repo_name=safe_str(repo.repo_name))
 
     def get_commit_url(self, repo, commit_id, request=None, permalink=False):
         if not request:
             request = get_current_request()
 
         if not request:
             return
 
         if permalink:
             return request.route_url(
                 'repo_commit', repo_name=safe_str(repo.repo_id),
                 commit_id=commit_id)
 
         else:
             return request.route_url(
                 'repo_commit', repo_name=safe_str(repo.repo_name),
                 commit_id=commit_id)
 
     def get_repo_log(self, repo, filter_term):
         repo_log = UserLog.query()\
             .filter(or_(UserLog.repository_id == repo.repo_id,
                         UserLog.repository_name == repo.repo_name))\
             .options(joinedload(UserLog.user))\
             .options(joinedload(UserLog.repository))\
             .order_by(UserLog.action_date.desc())
 
         repo_log = user_log_filter(repo_log, filter_term)
         return repo_log
 
     @classmethod
     def update_commit_cache(cls, repositories=None):
         if not repositories:
             repositories = Repository.getAll()
         for repo in repositories:
             repo.update_commit_cache()
 
     def get_repos_as_dict(self, repo_list=None, admin=False,
                           super_user_actions=False, short_name=None):
 
         _render = get_current_request().get_partial_renderer(
             'rhodecode:templates/data_table/_dt_elements.mako')
         c = _render.get_call_context()
         h = _render.get_helpers()
 
         def quick_menu(repo_name):
             return _render('quick_menu', repo_name)
 
         def repo_lnk(name, rtype, rstate, private, archived, fork_of):
             if short_name is not None:
                 short_name_var = short_name
             else:
                 short_name_var = not admin
             return _render('repo_name', name, rtype, rstate, private, archived, fork_of,
                            short_name=short_name_var, admin=False)
 
         def last_change(last_change):
             if admin and isinstance(last_change, datetime.datetime) and not last_change.tzinfo:
                 ts = time.time()
                 utc_offset = (datetime.datetime.fromtimestamp(ts)
                               - datetime.datetime.utcfromtimestamp(ts)).total_seconds()
                 last_change = last_change + datetime.timedelta(seconds=utc_offset)
 
             return _render("last_change", last_change)
 
         def rss_lnk(repo_name):
             return _render("rss", repo_name)
 
         def atom_lnk(repo_name):
             return _render("atom", repo_name)
 
         def last_rev(repo_name, cs_cache):
             return _render('revision', repo_name, cs_cache.get('revision'),
                            cs_cache.get('raw_id'), cs_cache.get('author'),
                            cs_cache.get('message'), cs_cache.get('date'))
 
         def desc(desc):
             return _render('repo_desc', desc, c.visual.stylify_metatags)
 
         def state(repo_state):
             return _render("repo_state", repo_state)
 
         def repo_actions(repo_name):
             return _render('repo_actions', repo_name, super_user_actions)
 
         def user_profile(username):
             return _render('user_profile', username)
 
         repos_data = []
         for repo in repo_list:
             # NOTE(marcink): because we use only raw column we need to load it like that
             changeset_cache = Repository._load_changeset_cache(
                 repo.repo_id, repo._changeset_cache)
 
             row = {
                 "menu": quick_menu(repo.repo_name),
 
                 "name": repo_lnk(repo.repo_name, repo.repo_type, repo.repo_state,
                                  repo.private, repo.archived, repo.fork),
 
                 "desc": desc(h.escape(repo.description)),
 
                 "last_change": last_change(repo.updated_on),
 
                 "last_changeset": last_rev(repo.repo_name, changeset_cache),
                 "last_changeset_raw": changeset_cache.get('revision'),
 
                 "owner": user_profile(repo.User.username),
 
                 "state": state(repo.repo_state),
                 "rss": rss_lnk(repo.repo_name),
                 "atom": atom_lnk(repo.repo_name),
             }
             if admin:
                 row.update({
                     "action": repo_actions(repo.repo_name),
                 })
             repos_data.append(row)
 
         return repos_data
 
     def get_repos_data_table(
             self, draw, start, limit,
             search_q, order_by, order_dir,
             auth_user, repo_group_id):
         from rhodecode.model.scm import RepoList
 
         _perms = ['repository.read', 'repository.write', 'repository.admin']
 
         repos = Repository.query() \
             .filter(Repository.group_id == repo_group_id) \
             .all()
         auth_repo_list = RepoList(
             repos, perm_set=_perms,
             extra_kwargs=dict(user=auth_user))
 
         allowed_ids = [-1]
         for repo in auth_repo_list:
             allowed_ids.append(repo.repo_id)
 
         repos_data_total_count = Repository.query() \
             .filter(Repository.group_id == repo_group_id) \
             .filter(or_(
                 # generate multiple IN to fix limitation problems
                 *in_filter_generator(Repository.repo_id, allowed_ids))
             ) \
             .count()
 
         base_q = Session.query(
             Repository.repo_id,
             Repository.repo_name,
             Repository.description,
             Repository.repo_type,
             Repository.repo_state,
             Repository.private,
             Repository.archived,
             Repository.fork,
             Repository.updated_on,
             Repository._changeset_cache,
             User,
             ) \
             .filter(Repository.group_id == repo_group_id) \
             .filter(or_(
                 # generate multiple IN to fix limitation problems
                 *in_filter_generator(Repository.repo_id, allowed_ids))
             ) \
             .join(User, User.user_id == Repository.user_id) \
             .group_by(Repository, User)
 
         repos_data_total_filtered_count = base_q.count()
 
         sort_defined = False
         if order_by == 'repo_name':
             sort_col = func.lower(Repository.repo_name)
             sort_defined = True
         elif order_by == 'user_username':
             sort_col = User.username
         else:
             sort_col = getattr(Repository, order_by, None)
 
         if sort_defined or sort_col:
             if order_dir == 'asc':
                 sort_col = sort_col.asc()
             else:
                 sort_col = sort_col.desc()
 
         base_q = base_q.order_by(sort_col)
         base_q = base_q.offset(start).limit(limit)
 
         repos_list = base_q.all()
 
         repos_data = RepoModel().get_repos_as_dict(
             repo_list=repos_list, admin=False)
 
         data = ({
             'draw': draw,
             'data': repos_data,
             'recordsTotal': repos_data_total_count,
             'recordsFiltered': repos_data_total_filtered_count,
         })
         return data
 
     def _get_defaults(self, repo_name):
         """
         Gets information about repository, and returns a dict for
         usage in forms
 
         :param repo_name:
         """
 
         repo_info = Repository.get_by_repo_name(repo_name)
 
         if repo_info is None:
             return None
 
         defaults = repo_info.get_dict()
         defaults['repo_name'] = repo_info.just_name
 
         groups = repo_info.groups_with_parents
         parent_group = groups[-1] if groups else None
 
         # we use -1 as this is how in HTML, we mark an empty group
         defaults['repo_group'] = getattr(parent_group, 'group_id', -1)
 
         keys_to_process = (
             {'k': 'repo_type', 'strip': False},
             {'k': 'repo_enable_downloads', 'strip': True},
             {'k': 'repo_description', 'strip': True},
             {'k': 'repo_enable_locking', 'strip': True},
             {'k': 'repo_landing_rev', 'strip': True},
             {'k': 'clone_uri', 'strip': False},
             {'k': 'push_uri', 'strip': False},
             {'k': 'repo_private', 'strip': True},
             {'k': 'repo_enable_statistics', 'strip': True}
         )
 
         for item in keys_to_process:
             attr = item['k']
             if item['strip']:
                 attr = remove_prefix(item['k'], 'repo_')
 
             val = defaults[attr]
             if item['k'] == 'repo_landing_rev':
                 val = ':'.join(defaults[attr])
             defaults[item['k']] = val
             if item['k'] == 'clone_uri':
                 defaults['clone_uri_hidden'] = repo_info.clone_uri_hidden
             if item['k'] == 'push_uri':
                 defaults['push_uri_hidden'] = repo_info.push_uri_hidden
 
         # fill owner
         if repo_info.user:
             defaults.update({'user': repo_info.user.username})
         else:
             replacement_user = User.get_first_super_admin().username
             defaults.update({'user': replacement_user})
 
         return defaults
 
     def update(self, repo, **kwargs):
         try:
             cur_repo = self._get_repo(repo)
             source_repo_name = cur_repo.repo_name
             if 'user' in kwargs:
                 cur_repo.user = User.get_by_username(kwargs['user'])
 
             if 'repo_group' in kwargs:
                 cur_repo.group = RepoGroup.get(kwargs['repo_group'])
             log.debug('Updating repo %s with params:%s', cur_repo, kwargs)
 
             update_keys = [
                 (1, 'repo_description'),
                 (1, 'repo_landing_rev'),
                 (1, 'repo_private'),
                 (1, 'repo_enable_downloads'),
                 (1, 'repo_enable_locking'),
                 (1, 'repo_enable_statistics'),
                 (0, 'clone_uri'),
                 (0, 'push_uri'),
                 (0, 'fork_id')
             ]
             for strip, k in update_keys:
                 if k in kwargs:
                     val = kwargs[k]
                     if strip:
                         k = remove_prefix(k, 'repo_')
 
                     setattr(cur_repo, k, val)
 
             new_name = cur_repo.get_new_name(kwargs['repo_name'])
             cur_repo.repo_name = new_name
 
             # if private flag is set, reset default permission to NONE
             if kwargs.get('repo_private'):
                 EMPTY_PERM = 'repository.none'
                 RepoModel().grant_user_permission(
                     repo=cur_repo, user=User.DEFAULT_USER, perm=EMPTY_PERM
                 )
 
             # handle extra fields
             for field in filter(lambda k: k.startswith(RepositoryField.PREFIX), kwargs):
                 k = RepositoryField.un_prefix_key(field)
                 ex_field = RepositoryField.get_by_key_name(
                     key=k, repo=cur_repo)
                 if ex_field:
                     ex_field.field_value = kwargs[field]
                     self.sa.add(ex_field)
 
             self.sa.add(cur_repo)
 
             if source_repo_name != new_name:
                 # rename repository
                 self._rename_filesystem_repo(
                     old=source_repo_name, new=new_name)
 
             return cur_repo
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def _create_repo(self, repo_name, repo_type, description, owner,
                      private=False, clone_uri=None, repo_group=None,
                      landing_rev='rev:tip', fork_of=None,
                      copy_fork_permissions=False, enable_statistics=False,
                      enable_locking=False, enable_downloads=False,
                      copy_group_permissions=False,
                      state=Repository.STATE_PENDING):
         """
         Create repository inside database with PENDING state, this should be
         only executed by create() repo. With exception of importing existing
         repos
         """
         from rhodecode.model.scm import ScmModel
 
         owner = self._get_user(owner)
         fork_of = self._get_repo(fork_of)
         repo_group = self._get_repo_group(safe_int(repo_group))
 
         try:
             repo_name = safe_unicode(repo_name)
             description = safe_unicode(description)
             # repo name is just a name of repository
             # while repo_name_full is a full qualified name that is combined
             # with name and path of group
             repo_name_full = repo_name
             repo_name = repo_name.split(Repository.NAME_SEP)[-1]
 
             new_repo = Repository()
             new_repo.repo_state = state
             new_repo.enable_statistics = False
             new_repo.repo_name = repo_name_full
             new_repo.repo_type = repo_type
             new_repo.user = owner
             new_repo.group = repo_group
             new_repo.description = description or repo_name
             new_repo.private = private
             new_repo.archived = False
             new_repo.clone_uri = clone_uri
             new_repo.landing_rev = landing_rev
 
             new_repo.enable_statistics = enable_statistics
             new_repo.enable_locking = enable_locking
             new_repo.enable_downloads = enable_downloads
 
             if repo_group:
                 new_repo.enable_locking = repo_group.enable_locking
 
             if fork_of:
                 parent_repo = fork_of
                 new_repo.fork = parent_repo
 
             events.trigger(events.RepoPreCreateEvent(new_repo))
 
             self.sa.add(new_repo)
 
             EMPTY_PERM = 'repository.none'
             if fork_of and copy_fork_permissions:
                 repo = fork_of
                 user_perms = UserRepoToPerm.query() \
                     .filter(UserRepoToPerm.repository == repo).all()
                 group_perms = UserGroupRepoToPerm.query() \
                     .filter(UserGroupRepoToPerm.repository == repo).all()
 
                 for perm in user_perms:
                     UserRepoToPerm.create(
                         perm.user, new_repo, perm.permission)
 
                 for perm in group_perms:
                     UserGroupRepoToPerm.create(
                         perm.users_group, new_repo, perm.permission)
                 # in case we copy permissions and also set this repo to private
                 # override the default user permission to make it a private repo
                 if private:
                     RepoModel(self.sa).grant_user_permission(
                         repo=new_repo, user=User.DEFAULT_USER, perm=EMPTY_PERM)
 
             elif repo_group and copy_group_permissions:
                 user_perms = UserRepoGroupToPerm.query() \
                     .filter(UserRepoGroupToPerm.group == repo_group).all()
 
                 group_perms = UserGroupRepoGroupToPerm.query() \
                     .filter(UserGroupRepoGroupToPerm.group == repo_group).all()
 
                 for perm in user_perms:
                     perm_name = perm.permission.permission_name.replace(
                         'group.', 'repository.')
                     perm_obj = Permission.get_by_key(perm_name)
                     UserRepoToPerm.create(perm.user, new_repo, perm_obj)
 
                 for perm in group_perms:
                     perm_name = perm.permission.permission_name.replace(
                         'group.', 'repository.')
                     perm_obj = Permission.get_by_key(perm_name)
                     UserGroupRepoToPerm.create(perm.users_group, new_repo, perm_obj)
 
                 if private:
                     RepoModel(self.sa).grant_user_permission(
                         repo=new_repo, user=User.DEFAULT_USER, perm=EMPTY_PERM)
 
             else:
                 perm_obj = self._create_default_perms(new_repo, private)
                 self.sa.add(perm_obj)
 
             # now automatically start following this repository as owner
             ScmModel(self.sa).toggle_following_repo(new_repo.repo_id, owner.user_id)
 
             # we need to flush here, in order to check if database won't
             # throw any exceptions, create filesystem dirs at the very end
             self.sa.flush()
             events.trigger(events.RepoCreateEvent(new_repo))
             return new_repo
 
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def create(self, form_data, cur_user):
         """
         Create repository using celery tasks
 
         :param form_data:
         :param cur_user:
         """
         from rhodecode.lib.celerylib import tasks, run_task
         return run_task(tasks.create_repo, form_data, cur_user)
 
     def update_permissions(self, repo, perm_additions=None, perm_updates=None,
                            perm_deletions=None, check_perms=True,
                            cur_user=None):
         if not perm_additions:
             perm_additions = []
         if not perm_updates:
             perm_updates = []
         if not perm_deletions:
             perm_deletions = []
 
         req_perms = ('usergroup.read', 'usergroup.write', 'usergroup.admin')
 
         changes = {
             'added': [],
             'updated': [],
             'deleted': [],
             'default_user_changed': None
         }
 
         repo = self._get_repo(repo)
 
         # update permissions
         for member_id, perm, member_type in perm_updates:
             member_id = int(member_id)
             if member_type == 'user':
                 member_name = User.get(member_id).username
                 if member_name == User.DEFAULT_USER:
                     # NOTE(dan): detect if we changed permissions for default user
                     perm_obj = self.sa.query(UserRepoToPerm) \
                         .filter(UserRepoToPerm.user_id == member_id) \
                         .filter(UserRepoToPerm.repository == repo) \
                         .scalar()
                     if perm_obj and perm_obj.permission.permission_name != perm:
                         changes['default_user_changed'] = True
 
                 # this updates also current one if found
                 self.grant_user_permission(
                     repo=repo, user=member_id, perm=perm)
             elif member_type == 'user_group':
                 # check if we have permissions to alter this usergroup
                 member_name = UserGroup.get(member_id).users_group_name
                 if not check_perms or HasUserGroupPermissionAny(
                         *req_perms)(member_name, user=cur_user):
                     self.grant_user_group_permission(
                         repo=repo, group_name=member_id, perm=perm)
             else:
                 raise ValueError("member_type must be 'user' or 'user_group' "
                                  "got {} instead".format(member_type))
             changes['updated'].append({'type': member_type, 'id': member_id,
                                        'name': member_name, 'new_perm': perm})
 
         # set new permissions
         for member_id, perm, member_type in perm_additions:
             member_id = int(member_id)
             if member_type == 'user':
                 member_name = User.get(member_id).username
                 self.grant_user_permission(
                     repo=repo, user=member_id, perm=perm)
             elif member_type == 'user_group':
                 # check if we have permissions to alter this usergroup
                 member_name = UserGroup.get(member_id).users_group_name
                 if not check_perms or HasUserGroupPermissionAny(
                         *req_perms)(member_name, user=cur_user):
                     self.grant_user_group_permission(
                         repo=repo, group_name=member_id, perm=perm)
             else:
                 raise ValueError("member_type must be 'user' or 'user_group' "
                                  "got {} instead".format(member_type))
 
             changes['added'].append({'type': member_type, 'id': member_id,
                                      'name': member_name, 'new_perm': perm})
         # delete permissions
         for member_id, perm, member_type in perm_deletions:
             member_id = int(member_id)
             if member_type == 'user':
                 member_name = User.get(member_id).username
                 self.revoke_user_permission(repo=repo, user=member_id)
             elif member_type == 'user_group':
                 # check if we have permissions to alter this usergroup
                 member_name = UserGroup.get(member_id).users_group_name
                 if not check_perms or HasUserGroupPermissionAny(
                         *req_perms)(member_name, user=cur_user):
                     self.revoke_user_group_permission(
                         repo=repo, group_name=member_id)
             else:
                 raise ValueError("member_type must be 'user' or 'user_group' "
                                  "got {} instead".format(member_type))
 
             changes['deleted'].append({'type': member_type, 'id': member_id,
                                        'name': member_name, 'new_perm': perm})
         return changes
 
     def create_fork(self, form_data, cur_user):
         """
         Simple wrapper into executing celery task for fork creation
 
         :param form_data:
         :param cur_user:
         """
         from rhodecode.lib.celerylib import tasks, run_task
         return run_task(tasks.create_repo_fork, form_data, cur_user)
 
     def archive(self, repo):
         """
         Archive given repository. Set archive flag.
 
         :param repo:
         """
         repo = self._get_repo(repo)
         if repo:
 
             try:
                 repo.archived = True
                 self.sa.add(repo)
                 self.sa.commit()
             except Exception:
                 log.error(traceback.format_exc())
                 raise
 
     def delete(self, repo, forks=None, pull_requests=None, fs_remove=True, cur_user=None):
         """
         Delete given repository, forks parameter defines what do do with
         attached forks. Throws AttachedForksError if deleted repo has attached
         forks
 
         :param repo:
         :param forks: str 'delete' or 'detach'
         :param pull_requests: str 'delete' or None
         :param fs_remove: remove(archive) repo from filesystem
         """
         if not cur_user:
             cur_user = getattr(get_current_rhodecode_user(), 'username', None)
         repo = self._get_repo(repo)
         if repo:
             if forks == 'detach':
                 for r in repo.forks:
                     r.fork = None
                     self.sa.add(r)
             elif forks == 'delete':
                 for r in repo.forks:
                     self.delete(r, forks='delete')
             elif [f for f in repo.forks]:
                 raise AttachedForksError()
 
             # check for pull requests
             pr_sources = repo.pull_requests_source
             pr_targets = repo.pull_requests_target
             if pull_requests != 'delete' and (pr_sources or pr_targets):
                 raise AttachedPullRequestsError()
 
             old_repo_dict = repo.get_dict()
             events.trigger(events.RepoPreDeleteEvent(repo))
             try:
                 self.sa.delete(repo)
                 if fs_remove:
                     self._delete_filesystem_repo(repo)
                 else:
                     log.debug('skipping removal from filesystem')
                 old_repo_dict.update({
                     'deleted_by': cur_user,
                     'deleted_on': time.time(),
                 })
                 log_delete_repository(**old_repo_dict)
                 events.trigger(events.RepoDeleteEvent(repo))
             except Exception:
                 log.error(traceback.format_exc())
                 raise
 
     def grant_user_permission(self, repo, user, perm):
         """
         Grant permission for user on given repository, or update existing one
         if found
 
         :param repo: Instance of Repository, repository_id, or repository name
         :param user: Instance of User, user_id or username
         :param perm: Instance of Permission, or permission_name
         """
         user = self._get_user(user)
         repo = self._get_repo(repo)
         permission = self._get_perm(perm)
 
         # check if we have that permission already
         obj = self.sa.query(UserRepoToPerm) \
             .filter(UserRepoToPerm.user == user) \
             .filter(UserRepoToPerm.repository == repo) \
             .scalar()
         if obj is None:
             # create new !
             obj = UserRepoToPerm()
         obj.repository = repo
         obj.user = user
         obj.permission = permission
         self.sa.add(obj)
         log.debug('Granted perm %s to %s on %s', perm, user, repo)
         action_logger_generic(
             'granted permission: {} to user: {} on repo: {}'.format(
                 perm, user, repo), namespace='security.repo')
         return obj
 
     def revoke_user_permission(self, repo, user):
         """
         Revoke permission for user on given repository
 
         :param repo: Instance of Repository, repository_id, or repository name
         :param user: Instance of User, user_id or username
         """
 
         user = self._get_user(user)
         repo = self._get_repo(repo)
 
         obj = self.sa.query(UserRepoToPerm) \
             .filter(UserRepoToPerm.repository == repo) \
             .filter(UserRepoToPerm.user == user) \
             .scalar()
         if obj:
             self.sa.delete(obj)
             log.debug('Revoked perm on %s on %s', repo, user)
             action_logger_generic(
                 'revoked permission from user: {} on repo: {}'.format(
                     user, repo), namespace='security.repo')
 
     def grant_user_group_permission(self, repo, group_name, perm):
         """
         Grant permission for user group on given repository, or update
         existing one if found
 
         :param repo: Instance of Repository, repository_id, or repository name
         :param group_name: Instance of UserGroup, users_group_id,
             or user group name
         :param perm: Instance of Permission, or permission_name
         """
         repo = self._get_repo(repo)
         group_name = self._get_user_group(group_name)
         permission = self._get_perm(perm)
 
         # check if we have that permission already
         obj = self.sa.query(UserGroupRepoToPerm) \
             .filter(UserGroupRepoToPerm.users_group == group_name) \
             .filter(UserGroupRepoToPerm.repository == repo) \
             .scalar()
 
         if obj is None:
             # create new
             obj = UserGroupRepoToPerm()
 
         obj.repository = repo
         obj.users_group = group_name
         obj.permission = permission
         self.sa.add(obj)
         log.debug('Granted perm %s to %s on %s', perm, group_name, repo)
         action_logger_generic(
             'granted permission: {} to usergroup: {} on repo: {}'.format(
                 perm, group_name, repo), namespace='security.repo')
 
         return obj
 
     def revoke_user_group_permission(self, repo, group_name):
         """
         Revoke permission for user group on given repository
 
         :param repo: Instance of Repository, repository_id, or repository name
         :param group_name: Instance of UserGroup, users_group_id,
             or user group name
         """
         repo = self._get_repo(repo)
         group_name = self._get_user_group(group_name)
 
         obj = self.sa.query(UserGroupRepoToPerm) \
             .filter(UserGroupRepoToPerm.repository == repo) \
             .filter(UserGroupRepoToPerm.users_group == group_name) \
             .scalar()
         if obj:
             self.sa.delete(obj)
             log.debug('Revoked perm to %s on %s', repo, group_name)
             action_logger_generic(
                 'revoked permission from usergroup: {} on repo: {}'.format(
                     group_name, repo), namespace='security.repo')
 
     def delete_stats(self, repo_name):
         """
         removes stats for given repo
 
         :param repo_name:
         """
         repo = self._get_repo(repo_name)
         try:
             obj = self.sa.query(Statistics) \
                 .filter(Statistics.repository == repo).scalar()
             if obj:
                 self.sa.delete(obj)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def add_repo_field(self, repo_name, field_key, field_label, field_value='',
                        field_type='str', field_desc=''):
 
         repo = self._get_repo(repo_name)
 
         new_field = RepositoryField()
         new_field.repository = repo
         new_field.field_key = field_key
         new_field.field_type = field_type  # python type
         new_field.field_value = field_value
         new_field.field_desc = field_desc
         new_field.field_label = field_label
         self.sa.add(new_field)
         return new_field
 
     def delete_repo_field(self, repo_name, field_key):
         repo = self._get_repo(repo_name)
         field = RepositoryField.get_by_key_name(field_key, repo)
         if field:
             self.sa.delete(field)
 
     def _create_filesystem_repo(self, repo_name, repo_type, repo_group,
                                 clone_uri=None, repo_store_location=None,
                                 use_global_config=False, install_hooks=True):
         """
         makes repository on filesystem. It's group aware means it'll create
         a repository within a group, and alter the paths accordingly of
         group location
 
         :param repo_name:
         :param alias:
         :param parent:
         :param clone_uri:
         :param repo_store_location:
         """
         from rhodecode.lib.utils import is_valid_repo, is_valid_repo_group
         from rhodecode.model.scm import ScmModel
 
         if Repository.NAME_SEP in repo_name:
             raise ValueError(
                 'repo_name must not contain groups got `%s`' % repo_name)
 
         if isinstance(repo_group, RepoGroup):
             new_parent_path = os.sep.join(repo_group.full_path_splitted)
         else:
             new_parent_path = repo_group or ''
 
         if repo_store_location:
             _paths = [repo_store_location]
         else:
             _paths = [self.repos_path, new_parent_path, repo_name]
             # we need to make it str for mercurial
         repo_path = os.path.join(*map(lambda x: safe_str(x), _paths))
 
         # check if this path is not a repository
         if is_valid_repo(repo_path, self.repos_path):
             raise Exception('This path %s is a valid repository' % repo_path)
 
         # check if this path is a group
         if is_valid_repo_group(repo_path, self.repos_path):
             raise Exception('This path %s is a valid group' % repo_path)
 
         log.info('creating repo %s in %s from url: `%s`',
                  repo_name, safe_unicode(repo_path),
                  obfuscate_url_pw(clone_uri))
 
         backend = get_backend(repo_type)
 
         config_repo = None if use_global_config else repo_name
         if config_repo and new_parent_path:
             config_repo = Repository.NAME_SEP.join(
                 (new_parent_path, config_repo))
         config = make_db_config(clear_session=False, repo=config_repo)
         config.set('extensions', 'largefiles', '')
 
         # patch and reset hooks section of UI config to not run any
         # hooks on creating remote repo
         config.clear_section('hooks')
 
         # TODO: johbo: Unify this, hardcoded "bare=True" does not look nice
         if repo_type == 'git':
             repo = backend(
                 repo_path, config=config, create=True, src_url=clone_uri, bare=True,
                 with_wire={"cache": False})
         else:
             repo = backend(
                 repo_path, config=config, create=True, src_url=clone_uri,
                 with_wire={"cache": False})
 
         if install_hooks:
             repo.install_hooks()
 
         log.debug('Created repo %s with %s backend',
                   safe_unicode(repo_name), safe_unicode(repo_type))
         return repo
 
     def _rename_filesystem_repo(self, old, new):
         """
         renames repository on filesystem
 
         :param old: old name
         :param new: new name
         """
         log.info('renaming repo from %s to %s', old, new)
 
         old_path = os.path.join(self.repos_path, old)
         new_path = os.path.join(self.repos_path, new)
         if os.path.isdir(new_path):
             raise Exception(
                 'Was trying to rename to already existing dir %s' % new_path
             )
         shutil.move(old_path, new_path)
 
     def _delete_filesystem_repo(self, repo):
         """
         removes repo from filesystem, the removal is acctually made by
         added rm__ prefix into dir, and rename internat .hg/.git dirs so this
         repository is no longer valid for rhodecode, can be undeleted later on
         by reverting the renames on this repository
 
         :param repo: repo object
         """
         rm_path = os.path.join(self.repos_path, repo.repo_name)
         repo_group = repo.group
         log.info("Removing repository %s", rm_path)
         # disable hg/git internal that it doesn't get detected as repo
         alias = repo.repo_type
 
         config = make_db_config(clear_session=False)
         config.set('extensions', 'largefiles', '')
         bare = getattr(repo.scm_instance(config=config), 'bare', False)
 
         # skip this for bare git repos
         if not bare:
             # disable VCS repo
             vcs_path = os.path.join(rm_path, '.%s' % alias)
             if os.path.exists(vcs_path):
                 shutil.move(vcs_path, os.path.join(rm_path, 'rm__.%s' % alias))
 
         _now = datetime.datetime.now()
         _ms = str(_now.microsecond).rjust(6, '0')
         _d = 'rm__%s__%s' % (_now.strftime('%Y%m%d_%H%M%S_' + _ms),
                              repo.just_name)
         if repo_group:
             # if repository is in group, prefix the removal path with the group
             args = repo_group.full_path_splitted + [_d]
             _d = os.path.join(*args)
 
         if os.path.isdir(rm_path):
             shutil.move(rm_path, os.path.join(self.repos_path, _d))
 
         # finally cleanup diff-cache if it exists
         cached_diffs_dir = repo.cached_diffs_dir
         if os.path.isdir(cached_diffs_dir):
             shutil.rmtree(cached_diffs_dir)
 
 
 class ReadmeFinder:
     """
     Utility which knows how to find a readme for a specific commit.
 
     The main idea is that this is a configurable algorithm. When creating an
     instance you can define parameters, currently only the `default_renderer`.
     Based on this configuration the method :meth:`search` behaves slightly
     different.
     """
 
     readme_re = re.compile(r'^readme(\.[^\.]+)?$', re.IGNORECASE)
     path_re = re.compile(r'^docs?', re.IGNORECASE)
 
     default_priorities = {
         None: 0,
         '.text': 2,
         '.txt': 3,
         '.rst': 1,
         '.rest': 2,
         '.md': 1,
         '.mkdn': 2,
         '.mdown': 3,
         '.markdown': 4,
     }
 
     path_priority = {
         'doc': 0,
         'docs': 1,
     }
 
     FALLBACK_PRIORITY = 99
 
     RENDERER_TO_EXTENSION = {
         'rst': ['.rst', '.rest'],
         'markdown': ['.md', 'mkdn', '.mdown', '.markdown'],
     }
 
     def __init__(self, default_renderer=None):
         self._default_renderer = default_renderer
         self._renderer_extensions = self.RENDERER_TO_EXTENSION.get(
             default_renderer, [])
 
-    def search(self, commit, path='/'):
+    def search(self, commit, path=u'/'):
         """
         Find a readme in the given `commit`.
         """
         nodes = commit.get_nodes(path)
         matches = self._match_readmes(nodes)
         matches = self._sort_according_to_priority(matches)
         if matches:
             return matches[0].node
 
         paths = self._match_paths(nodes)
         paths = self._sort_paths_according_to_priority(paths)
         for path in paths:
             match = self.search(commit, path=path)
             if match:
                 return match
 
         return None
 
     def _match_readmes(self, nodes):
         for node in nodes:
             if not node.is_file():
                 continue
             path = node.path.rsplit('/', 1)[-1]
             match = self.readme_re.match(path)
             if match:
                 extension = match.group(1)
                 yield ReadmeMatch(node, match, self._priority(extension))
 
     def _match_paths(self, nodes):
         for node in nodes:
             if not node.is_dir():
                 continue
             match = self.path_re.match(node.path)
             if match:
                 yield node.path
 
     def _priority(self, extension):
         renderer_priority = (
             0 if extension in self._renderer_extensions else 1)
         extension_priority = self.default_priorities.get(
             extension, self.FALLBACK_PRIORITY)
         return (renderer_priority, extension_priority)
 
     def _sort_according_to_priority(self, matches):
 
         def priority_and_path(match):
             return (match.priority, match.path)
 
         return sorted(matches, key=priority_and_path)
 
     def _sort_paths_according_to_priority(self, paths):
 
         def priority_and_path(path):
             return (self.path_priority.get(path, self.FALLBACK_PRIORITY), path)
 
         return sorted(paths, key=priority_and_path)
 
 
 class ReadmeMatch:
 
     def __init__(self, node, match, priority):
         self.node = node
         self._match = match
         self.priority = priority
 
     @property
     def path(self):
         return self.node.path
 
     def __repr__(self):
         return '<ReadmeMatch {} priority={}'.format(self.path, self.priority)