python3: removed compat modules

rhodecode/apps/ops/views.py

@@ -1,96 +1,96 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2016-2020 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 time
 import logging
 
 
 from pyramid.httpexceptions import HTTPFound
 
 from rhodecode.apps._base import BaseAppView
 from rhodecode.lib import helpers as h
 from rhodecode.lib.auth import LoginRequired
-from rhodecode.lib.compat import OrderedDict
+from collections import OrderedDict
 from rhodecode.model.db import UserApiKeys
 
 log = logging.getLogger(__name__)
 
 
 class OpsView(BaseAppView):
 
     def load_default_context(self):
         c = self._get_local_tmpl_context()
         c.user = c.auth_user.get_instance()
 
         return c
 
     def ops_ping(self):
         data = OrderedDict()
         data['instance'] = self.request.registry.settings.get('instance_id')
 
         if getattr(self.request, 'user'):
             caller_name = 'anonymous'
             if self.request.user.user_id:
                 caller_name = self.request.user.username
 
             data['caller_ip'] = self.request.user.ip_addr
             data['caller_name'] = caller_name
 
         return {'ok': data}
 
     def ops_error_test(self):
         """
         Test exception handling and emails on errors
         """
 
         class TestException(Exception):
             pass
         # add timeout so we add some sort of rate limiter
         time.sleep(2)
         msg = ('RhodeCode Enterprise test exception. '
                'Client:{}. Generation time: {}.'.format(self.request.user, time.time()))
         raise TestException(msg)
 
     def ops_redirect_test(self):
         """
         Test redirect handling
         """
         redirect_to = self.request.GET.get('to') or h.route_path('home')
         raise HTTPFound(redirect_to)
 
     @LoginRequired(auth_token_access=[UserApiKeys.ROLE_HTTP])
     def ops_healthcheck(self):
         from rhodecode.lib.system_info import load_system_info
 
         vcsserver_info = load_system_info('vcs_server')
         if vcsserver_info:
             vcsserver_info = vcsserver_info['human_value']
 
         db_info = load_system_info('database_info')
         if db_info:
             db_info = db_info['human_value']
 
         health_spec = {
             'caller_ip': self.request.user.ip_addr,
             'vcsserver': vcsserver_info,
             'db': db_info,
         }
 
         return {'healthcheck': health_spec}

rhodecode/apps/repository/tests/test_repo_files.py

@@ -1,1092 +1,1092 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 mock
 import pytest
 
 from rhodecode.apps.repository.tests.test_repo_compare import ComparePage
 from rhodecode.apps.repository.views.repo_files import RepoFilesView
 from rhodecode.lib import helpers as h
-from rhodecode.lib.compat import OrderedDict
+from collections import OrderedDict
 from rhodecode.lib.ext_json import json
 from rhodecode.lib.vcs import nodes
 
 from rhodecode.lib.vcs.conf import settings
 from rhodecode.tests import assert_session_flash
 from rhodecode.tests.fixture import Fixture
 from rhodecode.model.db import Session
 
 fixture = Fixture()
 
 
 def get_node_history(backend_type):
     return {
         'hg': json.loads(fixture.load_resource('hg_node_history_response.json')),
         'git': json.loads(fixture.load_resource('git_node_history_response.json')),
         'svn': json.loads(fixture.load_resource('svn_node_history_response.json')),
     }[backend_type]
 
 
 def route_path(name, params=None, **kwargs):
     import urllib.request, urllib.parse, urllib.error
 
     base_url = {
         'repo_summary': '/{repo_name}',
         'repo_archivefile': '/{repo_name}/archive/{fname}',
         'repo_files_diff': '/{repo_name}/diff/{f_path}',
         'repo_files_diff_2way_redirect':  '/{repo_name}/diff-2way/{f_path}',
         'repo_files': '/{repo_name}/files/{commit_id}/{f_path}',
         'repo_files:default_path': '/{repo_name}/files/{commit_id}/',
         'repo_files:default_commit': '/{repo_name}/files',
         'repo_files:rendered': '/{repo_name}/render/{commit_id}/{f_path}',
         'repo_files:annotated': '/{repo_name}/annotate/{commit_id}/{f_path}',
         'repo_files:annotated_previous': '/{repo_name}/annotate-previous/{commit_id}/{f_path}',
         'repo_files_nodelist': '/{repo_name}/nodelist/{commit_id}/{f_path}',
         'repo_file_raw': '/{repo_name}/raw/{commit_id}/{f_path}',
         'repo_file_download': '/{repo_name}/download/{commit_id}/{f_path}',
         'repo_file_history': '/{repo_name}/history/{commit_id}/{f_path}',
         'repo_file_authors': '/{repo_name}/authors/{commit_id}/{f_path}',
         'repo_files_remove_file': '/{repo_name}/remove_file/{commit_id}/{f_path}',
         'repo_files_delete_file': '/{repo_name}/delete_file/{commit_id}/{f_path}',
         'repo_files_edit_file': '/{repo_name}/edit_file/{commit_id}/{f_path}',
         'repo_files_update_file': '/{repo_name}/update_file/{commit_id}/{f_path}',
         'repo_files_add_file': '/{repo_name}/add_file/{commit_id}/{f_path}',
         'repo_files_create_file': '/{repo_name}/create_file/{commit_id}/{f_path}',
         'repo_nodetree_full': '/{repo_name}/nodetree_full/{commit_id}/{f_path}',
         'repo_nodetree_full:default_path': '/{repo_name}/nodetree_full/{commit_id}/',
     }[name].format(**kwargs)
 
     if params:
         base_url = '{}?{}'.format(base_url, urllib.parse.urlencode(params))
     return base_url
 
 
 def assert_files_in_response(response, files, params):
     template = (
         'href="/%(repo_name)s/files/%(commit_id)s/%(name)s"')
     _assert_items_in_response(response, files, template, params)
 
 
 def assert_dirs_in_response(response, dirs, params):
     template = (
         'href="/%(repo_name)s/files/%(commit_id)s/%(name)s"')
     _assert_items_in_response(response, dirs, template, params)
 
 
 def _assert_items_in_response(response, items, template, params):
     for item in items:
         item_params = {'name': item}
         item_params.update(params)
         response.mustcontain(template % item_params)
 
 
 def assert_timeago_in_response(response, items, params):
     for item in items:
         response.mustcontain(h.age_component(params['date']))
 
 
 @pytest.mark.usefixtures("app")
 class TestFilesViews(object):
 
     def test_show_files(self, backend):
         response = self.app.get(
             route_path('repo_files',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='/'))
         commit = backend.repo.get_commit()
 
         params = {
             'repo_name': backend.repo_name,
             'commit_id': commit.raw_id,
             'date': commit.date
         }
         assert_dirs_in_response(response, ['docs', 'vcs'], params)
         files = [
             '.gitignore',
             '.hgignore',
             '.hgtags',
             # TODO: missing in Git
             # '.travis.yml',
             'MANIFEST.in',
             'README.rst',
             # TODO: File is missing in svn repository
             # 'run_test_and_report.sh',
             'setup.cfg',
             'setup.py',
             'test_and_report.sh',
             'tox.ini',
         ]
         assert_files_in_response(response, files, params)
         assert_timeago_in_response(response, files, params)
 
     def test_show_files_links_submodules_with_absolute_url(self, backend_hg):
         repo = backend_hg['subrepos']
         response = self.app.get(
             route_path('repo_files',
                        repo_name=repo.repo_name,
                        commit_id='tip', f_path='/'))
         assert_response = response.assert_response()
         assert_response.contains_one_link(
             'absolute-path @ 000000000000', 'http://example.com/absolute-path')
 
     def test_show_files_links_submodules_with_absolute_url_subpaths(
             self, backend_hg):
         repo = backend_hg['subrepos']
         response = self.app.get(
             route_path('repo_files',
                        repo_name=repo.repo_name,
                        commit_id='tip', f_path='/'))
         assert_response = response.assert_response()
         assert_response.contains_one_link(
             'subpaths-path @ 000000000000',
             'http://sub-base.example.com/subpaths-path')
 
     @pytest.mark.xfail_backends("svn", reason="Depends on branch support")
     def test_files_menu(self, backend):
         new_branch = "temp_branch_name"
         commits = [
             {'message': 'a'},
             {'message': 'b', 'branch': new_branch}
         ]
         backend.create_repo(commits)
         backend.repo.landing_rev = "branch:%s" % new_branch
         Session().commit()
 
         # get response based on tip and not new commit
         response = self.app.get(
             route_path('repo_files',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='/'))
 
         # make sure Files menu url is not tip but new commit
         landing_rev = backend.repo.landing_ref_name
         files_url = route_path('repo_files:default_path',
                                repo_name=backend.repo_name,
                                commit_id=landing_rev, params={'at': landing_rev})
 
         assert landing_rev != 'tip'
         response.mustcontain(
             '<li class="active"><a class="menulink" href="%s">' % files_url)
 
     def test_show_files_commit(self, backend):
         commit = backend.repo.get_commit(commit_idx=32)
 
         response = self.app.get(
             route_path('repo_files',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='/'))
 
         dirs = ['docs', 'tests']
         files = ['README.rst']
         params = {
             'repo_name': backend.repo_name,
             'commit_id': commit.raw_id,
         }
         assert_dirs_in_response(response, dirs, params)
         assert_files_in_response(response, files, params)
 
     def test_show_files_different_branch(self, backend):
         branches = dict(
             hg=(150, ['git']),
             # TODO: Git test repository does not contain other branches
             git=(633, ['master']),
             # TODO: Branch support in Subversion
             svn=(150, [])
         )
         idx, branches = branches[backend.alias]
         commit = backend.repo.get_commit(commit_idx=idx)
         response = self.app.get(
             route_path('repo_files',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='/'))
 
         assert_response = response.assert_response()
         for branch in branches:
             assert_response.element_contains('.tags .branchtag', branch)
 
     def test_show_files_paging(self, backend):
         repo = backend.repo
         indexes = [73, 92, 109, 1, 0]
         idx_map = [(rev, repo.get_commit(commit_idx=rev).raw_id)
                    for rev in indexes]
 
         for idx in idx_map:
             response = self.app.get(
             route_path('repo_files',
                        repo_name=backend.repo_name,
                        commit_id=idx[1], f_path='/'))
 
             response.mustcontain("""r%s:%s""" % (idx[0], idx[1][:8]))
 
     def test_file_source(self, backend):
         commit = backend.repo.get_commit(commit_idx=167)
         response = self.app.get(
             route_path('repo_files',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='vcs/nodes.py'))
 
         msgbox = """<div class="commit">%s</div>"""
         response.mustcontain(msgbox % (commit.message, ))
 
         assert_response = response.assert_response()
         if commit.branch:
             assert_response.element_contains(
                 '.tags.tags-main .branchtag', commit.branch)
         if commit.tags:
             for tag in commit.tags:
                 assert_response.element_contains('.tags.tags-main .tagtag', tag)
 
     def test_file_source_annotated(self, backend):
         response = self.app.get(
             route_path('repo_files:annotated',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='vcs/nodes.py'))
         expected_commits = {
             'hg': 'r356',
             'git': 'r345',
             'svn': 'r208',
         }
         response.mustcontain(expected_commits[backend.alias])
 
     def test_file_source_authors(self, backend):
         response = self.app.get(
             route_path('repo_file_authors',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='vcs/nodes.py'))
         expected_authors = {
             'hg': ('Marcin Kuzminski', 'Lukasz Balcerzak'),
             'git': ('Marcin Kuzminski', 'Lukasz Balcerzak'),
             'svn': ('marcin', 'lukasz'),
         }
 
         for author in expected_authors[backend.alias]:
             response.mustcontain(author)
 
     def test_file_source_authors_with_annotation(self, backend):
         response = self.app.get(
             route_path('repo_file_authors',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='vcs/nodes.py',
                        params=dict(annotate=1)))
         expected_authors = {
             'hg': ('Marcin Kuzminski', 'Lukasz Balcerzak'),
             'git': ('Marcin Kuzminski', 'Lukasz Balcerzak'),
             'svn': ('marcin', 'lukasz'),
         }
 
         for author in expected_authors[backend.alias]:
             response.mustcontain(author)
 
     def test_file_source_history(self, backend, xhr_header):
         response = self.app.get(
             route_path('repo_file_history',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='vcs/nodes.py'),
             extra_environ=xhr_header)
         assert get_node_history(backend.alias) == json.loads(response.body)
 
     def test_file_source_history_svn(self, backend_svn, xhr_header):
         simple_repo = backend_svn['svn-simple-layout']
         response = self.app.get(
             route_path('repo_file_history',
                        repo_name=simple_repo.repo_name,
                        commit_id='tip', f_path='trunk/example.py'),
             extra_environ=xhr_header)
 
         expected_data = json.loads(
             fixture.load_resource('svn_node_history_branches.json'))
 
         assert expected_data == response.json
 
     def test_file_source_history_with_annotation(self, backend, xhr_header):
         response = self.app.get(
             route_path('repo_file_history',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='vcs/nodes.py',
                        params=dict(annotate=1)),
 
             extra_environ=xhr_header)
         assert get_node_history(backend.alias) == json.loads(response.body)
 
     def test_tree_search_top_level(self, backend, xhr_header):
         commit = backend.repo.get_commit(commit_idx=173)
         response = self.app.get(
             route_path('repo_files_nodelist',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='/'),
             extra_environ=xhr_header)
         assert 'nodes' in response.json
         assert {'name': 'docs', 'type': 'dir'} in response.json['nodes']
 
     def test_tree_search_missing_xhr(self, backend):
         self.app.get(
             route_path('repo_files_nodelist',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='/'),
             status=404)
 
     def test_tree_search_at_path(self, backend, xhr_header):
         commit = backend.repo.get_commit(commit_idx=173)
         response = self.app.get(
             route_path('repo_files_nodelist',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='/docs'),
             extra_environ=xhr_header)
         assert 'nodes' in response.json
         nodes = response.json['nodes']
         assert {'name': 'docs/api', 'type': 'dir'} in nodes
         assert {'name': 'docs/index.rst', 'type': 'file'} in nodes
 
     def test_tree_search_at_path_2nd_level(self, backend, xhr_header):
         commit = backend.repo.get_commit(commit_idx=173)
         response = self.app.get(
             route_path('repo_files_nodelist',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='/docs/api'),
             extra_environ=xhr_header)
         assert 'nodes' in response.json
         nodes = response.json['nodes']
         assert {'name': 'docs/api/index.rst', 'type': 'file'} in nodes
 
     def test_tree_search_at_path_missing_xhr(self, backend):
         self.app.get(
             route_path('repo_files_nodelist',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='/docs'),
             status=404)
 
     def test_nodetree(self, backend, xhr_header):
         commit = backend.repo.get_commit(commit_idx=173)
         response = self.app.get(
             route_path('repo_nodetree_full',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='/'),
             extra_environ=xhr_header)
 
         assert_response = response.assert_response()
 
         for attr in ['data-commit-id', 'data-date', 'data-author']:
             elements = assert_response.get_elements('[{}]'.format(attr))
             assert len(elements) > 1
 
             for element in elements:
                 assert element.get(attr)
 
     def test_nodetree_if_file(self, backend, xhr_header):
         commit = backend.repo.get_commit(commit_idx=173)
         response = self.app.get(
             route_path('repo_nodetree_full',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='README.rst'),
             extra_environ=xhr_header)
         assert response.body == ''
 
     def test_nodetree_wrong_path(self, backend, xhr_header):
         commit = backend.repo.get_commit(commit_idx=173)
         response = self.app.get(
             route_path('repo_nodetree_full',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='/dont-exist'),
             extra_environ=xhr_header)
 
         err = 'error: There is no file nor ' \
               'directory at the given path'
         assert err in response.body
 
     def test_nodetree_missing_xhr(self, backend):
         self.app.get(
             route_path('repo_nodetree_full',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='/'),
             status=404)
 
 
 @pytest.mark.usefixtures("app", "autologin_user")
 class TestRawFileHandling(object):
 
     def test_download_file(self, backend):
         commit = backend.repo.get_commit(commit_idx=173)
         response = self.app.get(
             route_path('repo_file_download',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='vcs/nodes.py'),)
 
         assert response.content_disposition == 'attachment; filename="nodes.py"; filename*=UTF-8\'\'nodes.py'
         assert response.content_type == "text/x-python"
 
     def test_download_file_wrong_cs(self, backend):
         raw_id = u'ERRORce30c96924232dffcd24178a07ffeb5dfc'
 
         response = self.app.get(
             route_path('repo_file_download',
                        repo_name=backend.repo_name,
                        commit_id=raw_id, f_path='vcs/nodes.svg'),
             status=404)
 
         msg = """No such commit exists for this repository"""
         response.mustcontain(msg)
 
     def test_download_file_wrong_f_path(self, backend):
         commit = backend.repo.get_commit(commit_idx=173)
         f_path = 'vcs/ERRORnodes.py'
 
         response = self.app.get(
             route_path('repo_file_download',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path=f_path),
             status=404)
 
         msg = (
             "There is no file nor directory at the given path: "
             "`%s` at commit %s" % (f_path, commit.short_id))
         response.mustcontain(msg)
 
     def test_file_raw(self, backend):
         commit = backend.repo.get_commit(commit_idx=173)
         response = self.app.get(
             route_path('repo_file_raw',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path='vcs/nodes.py'),)
 
         assert response.content_type == "text/plain"
 
     def test_file_raw_binary(self, backend):
         commit = backend.repo.get_commit()
         response = self.app.get(
             route_path('repo_file_raw',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id,
                        f_path='docs/theme/ADC/static/breadcrumb_background.png'),)
 
         assert response.content_disposition == 'inline'
 
     def test_raw_file_wrong_cs(self, backend):
         raw_id = u'ERRORcce30c96924232dffcd24178a07ffeb5dfc'
 
         response = self.app.get(
             route_path('repo_file_raw',
                        repo_name=backend.repo_name,
                        commit_id=raw_id, f_path='vcs/nodes.svg'),
             status=404)
 
         msg = """No such commit exists for this repository"""
         response.mustcontain(msg)
 
     def test_raw_wrong_f_path(self, backend):
         commit = backend.repo.get_commit(commit_idx=173)
         f_path = 'vcs/ERRORnodes.py'
         response = self.app.get(
             route_path('repo_file_raw',
                        repo_name=backend.repo_name,
                        commit_id=commit.raw_id, f_path=f_path),
             status=404)
 
         msg = (
             "There is no file nor directory at the given path: "
             "`%s` at commit %s" % (f_path, commit.short_id))
         response.mustcontain(msg)
 
     def test_raw_svg_should_not_be_rendered(self, backend):
         backend.create_repo()
         backend.ensure_file("xss.svg")
         response = self.app.get(
             route_path('repo_file_raw',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='xss.svg'),)
         # If the content type is image/svg+xml then it allows to render HTML
         # and malicious SVG.
         assert response.content_type == "text/plain"
 
 
 @pytest.mark.usefixtures("app")
 class TestRepositoryArchival(object):
 
     def test_archival(self, backend):
         backend.enable_downloads()
         commit = backend.repo.get_commit(commit_idx=173)
         for a_type, content_type, extension in settings.ARCHIVE_SPECS:
 
             short = commit.short_id + extension
             fname = commit.raw_id + extension
             filename = '%s-%s' % (backend.repo_name, short)
             response = self.app.get(
                 route_path('repo_archivefile',
                            repo_name=backend.repo_name,
                            fname=fname))
 
             assert response.status == '200 OK'
             headers = [
                 ('Content-Disposition', 'attachment; filename=%s' % filename),
                 ('Content-Type', '%s' % content_type),
             ]
 
             for header in headers:
                 assert header in response.headers.items()
 
     def test_archival_no_hash(self, backend):
         backend.enable_downloads()
         commit = backend.repo.get_commit(commit_idx=173)
         for a_type, content_type, extension in settings.ARCHIVE_SPECS:
 
             short = 'plain' + extension
             fname = commit.raw_id + extension
             filename = '%s-%s' % (backend.repo_name, short)
             response = self.app.get(
                 route_path('repo_archivefile',
                            repo_name=backend.repo_name,
                            fname=fname, params={'with_hash': 0}))
 
             assert response.status == '200 OK'
             headers = [
                 ('Content-Disposition', 'attachment; filename=%s' % filename),
                 ('Content-Type', '%s' % content_type),
             ]
 
             for header in headers:
                 assert header in response.headers.items()
 
     @pytest.mark.parametrize('arch_ext',[
         'tar', 'rar', 'x', '..ax', '.zipz', 'tar.gz.tar'])
     def test_archival_wrong_ext(self, backend, arch_ext):
         backend.enable_downloads()
         commit = backend.repo.get_commit(commit_idx=173)
 
         fname = commit.raw_id + '.' + arch_ext
 
         response = self.app.get(
             route_path('repo_archivefile',
                        repo_name=backend.repo_name,
                        fname=fname))
         response.mustcontain(
             'Unknown archive type for: `{}`'.format(fname))
 
     @pytest.mark.parametrize('commit_id', [
         '00x000000', 'tar', 'wrong', '@$@$42413232', '232dffcd'])
     def test_archival_wrong_commit_id(self, backend, commit_id):
         backend.enable_downloads()
         fname = '%s.zip' % commit_id
 
         response = self.app.get(
             route_path('repo_archivefile',
                        repo_name=backend.repo_name,
                        fname=fname))
         response.mustcontain('Unknown commit_id')
 
 
 @pytest.mark.usefixtures("app")
 class TestFilesDiff(object):
 
     @pytest.mark.parametrize("diff", ['diff', 'download', 'raw'])
     def test_file_full_diff(self, backend, diff):
         commit1 = backend.repo.get_commit(commit_idx=-1)
         commit2 = backend.repo.get_commit(commit_idx=-2)
 
         response = self.app.get(
             route_path('repo_files_diff',
                        repo_name=backend.repo_name,
                        f_path='README'),
             params={
                 'diff1': commit2.raw_id,
                 'diff2': commit1.raw_id,
                 'fulldiff': '1',
                 'diff': diff,
             })
 
         if diff == 'diff':
             # use redirect since this is OLD view redirecting to compare page
             response = response.follow()
 
         # It's a symlink to README.rst
         response.mustcontain('README.rst')
         response.mustcontain('No newline at end of file')
 
     def test_file_binary_diff(self, backend):
         commits = [
             {'message': 'First commit'},
             {'message': 'Commit with binary',
              'added': [nodes.FileNode('file.bin', content='\0BINARY\0')]},
         ]
         repo = backend.create_repo(commits=commits)
 
         response = self.app.get(
             route_path('repo_files_diff',
                        repo_name=backend.repo_name,
                        f_path='file.bin'),
             params={
                 'diff1': repo.get_commit(commit_idx=0).raw_id,
                 'diff2': repo.get_commit(commit_idx=1).raw_id,
                 'fulldiff': '1',
                 'diff': 'diff',
             })
         # use redirect since this is OLD view redirecting to compare page
         response = response.follow()
         response.mustcontain('Collapse 1 commit')
         file_changes = (1, 0, 0)
 
         compare_page = ComparePage(response)
         compare_page.contains_change_summary(*file_changes)
 
         if backend.alias == 'svn':
             response.mustcontain('new file 10644')
             # TODO(marcink): SVN doesn't yet detect binary changes
         else:
             response.mustcontain('new file 100644')
             response.mustcontain('binary diff hidden')
 
     def test_diff_2way(self, backend):
         commit1 = backend.repo.get_commit(commit_idx=-1)
         commit2 = backend.repo.get_commit(commit_idx=-2)
         response = self.app.get(
             route_path('repo_files_diff_2way_redirect',
                        repo_name=backend.repo_name,
                        f_path='README'),
             params={
                 'diff1': commit2.raw_id,
                 'diff2': commit1.raw_id,
             })
         # use redirect since this is OLD view redirecting to compare page
         response = response.follow()
 
         # It's a symlink to README.rst
         response.mustcontain('README.rst')
         response.mustcontain('No newline at end of file')
 
     def test_requires_one_commit_id(self, backend, autologin_user):
         response = self.app.get(
             route_path('repo_files_diff',
                        repo_name=backend.repo_name,
                        f_path='README.rst'),
             status=400)
         response.mustcontain(
             'Need query parameter', 'diff1', 'diff2', 'to generate a diff.')
 
     def test_returns_no_files_if_file_does_not_exist(self, vcsbackend):
         repo = vcsbackend.repo
         response = self.app.get(
             route_path('repo_files_diff',
                        repo_name=repo.name,
                        f_path='does-not-exist-in-any-commit'),
             params={
                 'diff1': repo[0].raw_id,
                 'diff2': repo[1].raw_id
             })
 
         response = response.follow()
         response.mustcontain('No files')
 
     def test_returns_redirect_if_file_not_changed(self, backend):
         commit = backend.repo.get_commit(commit_idx=-1)
         response = self.app.get(
             route_path('repo_files_diff_2way_redirect',
                        repo_name=backend.repo_name,
                        f_path='README'),
             params={
                 'diff1': commit.raw_id,
                 'diff2': commit.raw_id,
             })
 
         response = response.follow()
         response.mustcontain('No files')
         response.mustcontain('No commits in this compare')
 
     def test_supports_diff_to_different_path_svn(self, backend_svn):
         #TODO: check this case
         return
 
         repo = backend_svn['svn-simple-layout'].scm_instance()
         commit_id_1 = '24'
         commit_id_2 = '26'
 
         response = self.app.get(
             route_path('repo_files_diff',
                        repo_name=backend_svn.repo_name,
                        f_path='trunk/example.py'),
             params={
                 'diff1': 'tags/v0.2/example.py@' + commit_id_1,
                 'diff2': commit_id_2,
             })
 
         response = response.follow()
         response.mustcontain(
             # diff contains this
             "Will print out a useful message on invocation.")
 
         # Note: Expecting that we indicate the user what's being compared
         response.mustcontain("trunk/example.py")
         response.mustcontain("tags/v0.2/example.py")
 
     def test_show_rev_redirects_to_svn_path(self, backend_svn):
         #TODO: check this case
         return
 
         repo = backend_svn['svn-simple-layout'].scm_instance()
         commit_id = repo[-1].raw_id
 
         response = self.app.get(
             route_path('repo_files_diff',
                        repo_name=backend_svn.repo_name,
                        f_path='trunk/example.py'),
             params={
                 'diff1': 'branches/argparse/example.py@' + commit_id,
                 'diff2': commit_id,
             },
             status=302)
         response = response.follow()
         assert response.headers['Location'].endswith(
             'svn-svn-simple-layout/files/26/branches/argparse/example.py')
 
     def test_show_rev_and_annotate_redirects_to_svn_path(self, backend_svn):
         #TODO: check this case
         return
 
         repo = backend_svn['svn-simple-layout'].scm_instance()
         commit_id = repo[-1].raw_id
         response = self.app.get(
             route_path('repo_files_diff',
                        repo_name=backend_svn.repo_name,
                        f_path='trunk/example.py'),
             params={
                 'diff1': 'branches/argparse/example.py@' + commit_id,
                 'diff2': commit_id,
                 'show_rev': 'Show at Revision',
                 'annotate': 'true',
             },
             status=302)
         response = response.follow()
         assert response.headers['Location'].endswith(
             'svn-svn-simple-layout/annotate/26/branches/argparse/example.py')
 
 
 @pytest.mark.usefixtures("app", "autologin_user")
 class TestModifyFilesWithWebInterface(object):
 
     def test_add_file_view(self, backend):
         self.app.get(
             route_path('repo_files_add_file',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='/')
             )
 
     @pytest.mark.xfail_backends("svn", reason="Depends on online editing")
     def test_add_file_into_repo_missing_content(self, backend, csrf_token):
         backend.create_repo()
         filename = 'init.py'
         response = self.app.post(
             route_path('repo_files_create_file',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path='/'),
             params={
                 'content': "",
                 'filename': filename,
                 'csrf_token': csrf_token,
             },
             status=302)
         expected_msg = 'Successfully committed new file `{}`'.format(os.path.join(filename))
         assert_session_flash(response, expected_msg)
 
     def test_add_file_into_repo_missing_filename(self, backend, csrf_token):
         commit_id = backend.repo.get_commit().raw_id
         response = self.app.post(
             route_path('repo_files_create_file',
                        repo_name=backend.repo_name,
                        commit_id=commit_id, f_path='/'),
             params={
                 'content': "foo",
                 'csrf_token': csrf_token,
             },
             status=302)
 
         assert_session_flash(response, 'No filename specified')
 
     def test_add_file_into_repo_errors_and_no_commits(
             self, backend, csrf_token):
         repo = backend.create_repo()
         # Create a file with no filename, it will display an error but
         # the repo has no commits yet
         response = self.app.post(
             route_path('repo_files_create_file',
                        repo_name=repo.repo_name,
                        commit_id='tip', f_path='/'),
             params={
                 'content': "foo",
                 'csrf_token': csrf_token,
             },
             status=302)
 
         assert_session_flash(response, 'No filename specified')
 
         # Not allowed, redirect to the summary
         redirected = response.follow()
         summary_url = h.route_path('repo_summary', repo_name=repo.repo_name)
 
         # As there are no commits, displays the summary page with the error of
         # creating a file with no filename
 
         assert redirected.request.path == summary_url
 
     @pytest.mark.parametrize("filename, clean_filename", [
         ('/abs/foo', 'abs/foo'),
         ('../rel/foo', 'rel/foo'),
         ('file/../foo/foo', 'file/foo/foo'),
     ])
     def test_add_file_into_repo_bad_filenames(self, filename, clean_filename, backend, csrf_token):
         repo = backend.create_repo()
         commit_id = repo.get_commit().raw_id
 
         response = self.app.post(
             route_path('repo_files_create_file',
                        repo_name=repo.repo_name,
                        commit_id=commit_id, f_path='/'),
             params={
                 'content': "foo",
                 'filename': filename,
                 'csrf_token': csrf_token,
             },
             status=302)
 
         expected_msg = 'Successfully committed new file `{}`'.format(clean_filename)
         assert_session_flash(response, expected_msg)
 
     @pytest.mark.parametrize("cnt, filename, content", [
         (1, 'foo.txt', "Content"),
         (2, 'dir/foo.rst', "Content"),
         (3, 'dir/foo-second.rst', "Content"),
         (4, 'rel/dir/foo.bar', "Content"),
     ])
     def test_add_file_into_empty_repo(self, cnt, filename, content, backend, csrf_token):
         repo = backend.create_repo()
         commit_id = repo.get_commit().raw_id
         response = self.app.post(
             route_path('repo_files_create_file',
                        repo_name=repo.repo_name,
                        commit_id=commit_id, f_path='/'),
             params={
                 'content': content,
                 'filename': filename,
                 'csrf_token': csrf_token,
             },
             status=302)
 
         expected_msg = 'Successfully committed new file `{}`'.format(filename)
         assert_session_flash(response, expected_msg)
 
     def test_edit_file_view(self, backend):
         response = self.app.get(
             route_path('repo_files_edit_file',
                        repo_name=backend.repo_name,
                        commit_id=backend.default_head_id,
                        f_path='vcs/nodes.py'),
             status=200)
         response.mustcontain("Module holding everything related to vcs nodes.")
 
     def test_edit_file_view_not_on_branch(self, backend):
         repo = backend.create_repo()
         backend.ensure_file("vcs/nodes.py")
 
         response = self.app.get(
             route_path('repo_files_edit_file',
                        repo_name=repo.repo_name,
                        commit_id='tip',
                        f_path='vcs/nodes.py'),
             status=302)
         assert_session_flash(
             response, 'Cannot modify file. Given commit `tip` is not head of a branch.')
 
     def test_edit_file_view_commit_changes(self, backend, csrf_token):
         repo = backend.create_repo()
         backend.ensure_file("vcs/nodes.py", content="print 'hello'")
 
         response = self.app.post(
             route_path('repo_files_update_file',
                        repo_name=repo.repo_name,
                        commit_id=backend.default_head_id,
                        f_path='vcs/nodes.py'),
             params={
                 'content': "print 'hello world'",
                 'message': 'I committed',
                 'filename': "vcs/nodes.py",
                 'csrf_token': csrf_token,
             },
             status=302)
         assert_session_flash(
             response, 'Successfully committed changes to file `vcs/nodes.py`')
         tip = repo.get_commit(commit_idx=-1)
         assert tip.message == 'I committed'
 
     def test_edit_file_view_commit_changes_default_message(self, backend,
                                                            csrf_token):
         repo = backend.create_repo()
         backend.ensure_file("vcs/nodes.py", content="print 'hello'")
 
         commit_id = (
             backend.default_branch_name or
             backend.repo.scm_instance().commit_ids[-1])
 
         response = self.app.post(
             route_path('repo_files_update_file',
                        repo_name=repo.repo_name,
                        commit_id=commit_id,
                        f_path='vcs/nodes.py'),
             params={
                 'content': "print 'hello world'",
                 'message': '',
                 'filename': "vcs/nodes.py",
                 'csrf_token': csrf_token,
             },
             status=302)
         assert_session_flash(
             response, 'Successfully committed changes to file `vcs/nodes.py`')
         tip = repo.get_commit(commit_idx=-1)
         assert tip.message == 'Edited file vcs/nodes.py via RhodeCode Enterprise'
 
     def test_delete_file_view(self, backend):
         self.app.get(
             route_path('repo_files_remove_file',
                        repo_name=backend.repo_name,
                        commit_id=backend.default_head_id,
                        f_path='vcs/nodes.py'),
             status=200)
 
     def test_delete_file_view_not_on_branch(self, backend):
         repo = backend.create_repo()
         backend.ensure_file('vcs/nodes.py')
 
         response = self.app.get(
             route_path('repo_files_remove_file',
                        repo_name=repo.repo_name,
                        commit_id='tip',
                        f_path='vcs/nodes.py'),
             status=302)
         assert_session_flash(
             response, 'Cannot modify file. Given commit `tip` is not head of a branch.')
 
     def test_delete_file_view_commit_changes(self, backend, csrf_token):
         repo = backend.create_repo()
         backend.ensure_file("vcs/nodes.py")
 
         response = self.app.post(
             route_path('repo_files_delete_file',
                        repo_name=repo.repo_name,
                        commit_id=backend.default_head_id,
                        f_path='vcs/nodes.py'),
             params={
                 'message': 'i committed',
                 'csrf_token': csrf_token,
             },
             status=302)
         assert_session_flash(
             response, 'Successfully deleted file `vcs/nodes.py`')
 
 
 @pytest.mark.usefixtures("app")
 class TestFilesViewOtherCases(object):
 
     def test_access_empty_repo_redirect_to_summary_with_alert_write_perms(
             self, backend_stub, autologin_regular_user, user_regular,
             user_util):
 
         repo = backend_stub.create_repo()
         user_util.grant_user_permission_to_repo(
             repo, user_regular, 'repository.write')
         response = self.app.get(
             route_path('repo_files',
                        repo_name=repo.repo_name,
                        commit_id='tip', f_path='/'))
 
         repo_file_add_url = route_path(
             'repo_files_add_file',
             repo_name=repo.repo_name,
             commit_id=0, f_path='')
 
         assert_session_flash(
             response,
             'There are no files yet. <a class="alert-link" '
             'href="{}">Click here to add a new file.</a>'
             .format(repo_file_add_url))
 
     def test_access_empty_repo_redirect_to_summary_with_alert_no_write_perms(
             self, backend_stub, autologin_regular_user):
         repo = backend_stub.create_repo()
         # init session for anon user
         route_path('repo_summary', repo_name=repo.repo_name)
 
         repo_file_add_url = route_path(
             'repo_files_add_file',
             repo_name=repo.repo_name,
             commit_id=0, f_path='')
 
         response = self.app.get(
             route_path('repo_files',
                        repo_name=repo.repo_name,
                        commit_id='tip', f_path='/'))
 
         assert_session_flash(response, no_=repo_file_add_url)
 
     @pytest.mark.parametrize('file_node', [
         'archive/file.zip',
         'diff/my-file.txt',
         'render.py',
         'render',
         'remove_file',
         'remove_file/to-delete.txt',
     ])
     def test_file_names_equal_to_routes_parts(self, backend, file_node):
         backend.create_repo()
         backend.ensure_file(file_node)
 
         self.app.get(
             route_path('repo_files',
                        repo_name=backend.repo_name,
                        commit_id='tip', f_path=file_node),
             status=200)
 
 
 class TestAdjustFilePathForSvn(object):
     """
     SVN specific adjustments of node history in RepoFilesView.
     """
 
     def test_returns_path_relative_to_matched_reference(self):
         repo = self._repo(branches=['trunk'])
         self.assert_file_adjustment('trunk/file', 'file', repo)
 
     def test_does_not_modify_file_if_no_reference_matches(self):
         repo = self._repo(branches=['trunk'])
         self.assert_file_adjustment('notes/file', 'notes/file', repo)
 
     def test_does_not_adjust_partial_directory_names(self):
         repo = self._repo(branches=['trun'])
         self.assert_file_adjustment('trunk/file', 'trunk/file', repo)
 
     def test_is_robust_to_patterns_which_prefix_other_patterns(self):
         repo = self._repo(branches=['trunk', 'trunk/new', 'trunk/old'])
         self.assert_file_adjustment('trunk/new/file', 'file', repo)
 
     def assert_file_adjustment(self, f_path, expected, repo):
         result = RepoFilesView.adjust_file_path_for_svn(f_path, repo)
         assert result == expected
 
     def _repo(self, branches=None):
         repo = mock.Mock()
         repo.branches = OrderedDict((name, '0') for name in branches or [])
         repo.tags = {}
         return repo

rhodecode/apps/repository/tests/test_repo_summary.py

@@ -1,524 +1,524 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 re
 
 import mock
 import pytest
 
 from rhodecode.apps.repository.views.repo_summary import RepoSummaryView
 from rhodecode.lib import helpers as h
-from rhodecode.lib.compat import OrderedDict
+from collections import OrderedDict
 from rhodecode.lib.utils2 import AttributeDict, safe_str
 from rhodecode.lib.vcs.exceptions import RepositoryRequirementError
 from rhodecode.model.db import Repository
 from rhodecode.model.meta import Session
 from rhodecode.model.repo import RepoModel
 from rhodecode.model.scm import ScmModel
 from rhodecode.tests import assert_session_flash
 from rhodecode.tests.fixture import Fixture
 from rhodecode.tests.utils import AssertResponse, repo_on_filesystem
 
 
 fixture = Fixture()
 
 
 def route_path(name, params=None, **kwargs):
     import urllib.request, urllib.parse, urllib.error
 
     base_url = {
         'repo_summary': '/{repo_name}',
         'repo_stats': '/{repo_name}/repo_stats/{commit_id}',
         'repo_refs_data': '/{repo_name}/refs-data',
         'repo_refs_changelog_data': '/{repo_name}/refs-data-changelog',
         'repo_creating_check': '/{repo_name}/repo_creating_check',
     }[name].format(**kwargs)
 
     if params:
         base_url = '{}?{}'.format(base_url, urllib.parse.urlencode(params))
     return base_url
 
 
 def assert_clone_url(response, server, repo, disabled=False):
 
     response.mustcontain(
         '<input type="text" class="input-monospace clone_url_input" '
         '{disabled}readonly="readonly" '
         'value="http://test_admin@{server}/{repo}"/>'.format(
             server=server, repo=repo, disabled='disabled ' if disabled else ' ')
     )
 
 
 @pytest.mark.usefixtures('app')
 class TestSummaryView(object):
     def test_index(self, autologin_user, backend, http_host_only_stub):
         repo_id = backend.repo.repo_id
         repo_name = backend.repo_name
         with mock.patch('rhodecode.lib.helpers.is_svn_without_proxy',
                         return_value=False):
             response = self.app.get(
                 route_path('repo_summary', repo_name=repo_name))
 
         # repo type
         response.mustcontain(
             '<i class="icon-%s">' % (backend.alias, )
         )
         # public/private
         response.mustcontain(
             """<i class="icon-unlock-alt">"""
         )
 
         # clone url...
         assert_clone_url(response, http_host_only_stub, repo_name)
         assert_clone_url(response, http_host_only_stub, '_{}'.format(repo_id))
 
     def test_index_svn_without_proxy(
             self, autologin_user, backend_svn, http_host_only_stub):
         repo_id = backend_svn.repo.repo_id
         repo_name = backend_svn.repo_name
         response = self.app.get(route_path('repo_summary', repo_name=repo_name))
         # clone url...
 
         assert_clone_url(response, http_host_only_stub, repo_name, disabled=True)
         assert_clone_url(response, http_host_only_stub, '_{}'.format(repo_id), disabled=True)
 
     def test_index_with_trailing_slash(
             self, autologin_user, backend, http_host_only_stub):
 
         repo_id = backend.repo.repo_id
         repo_name = backend.repo_name
         with mock.patch('rhodecode.lib.helpers.is_svn_without_proxy',
                         return_value=False):
             response = self.app.get(
                 route_path('repo_summary', repo_name=repo_name) + '/',
                 status=200)
 
         # clone url...
         assert_clone_url(response, http_host_only_stub, repo_name)
         assert_clone_url(response, http_host_only_stub, '_{}'.format(repo_id))
 
     def test_index_by_id(self, autologin_user, backend):
         repo_id = backend.repo.repo_id
         response = self.app.get(
             route_path('repo_summary', repo_name='_%s' % (repo_id,)))
 
         # repo type
         response.mustcontain(
             '<i class="icon-%s">' % (backend.alias, )
         )
         # public/private
         response.mustcontain(
             """<i class="icon-unlock-alt">"""
         )
 
     def test_index_by_repo_having_id_path_in_name_hg(self, autologin_user):
         fixture.create_repo(name='repo_1')
         response = self.app.get(route_path('repo_summary', repo_name='repo_1'))
 
         try:
             response.mustcontain("repo_1")
         finally:
             RepoModel().delete(Repository.get_by_repo_name('repo_1'))
             Session().commit()
 
     def test_index_with_anonymous_access_disabled(
             self, backend, disable_anonymous_user):
         response = self.app.get(
             route_path('repo_summary', repo_name=backend.repo_name), status=302)
         assert 'login' in response.location
 
     def _enable_stats(self, repo):
         r = Repository.get_by_repo_name(repo)
         r.enable_statistics = True
         Session().add(r)
         Session().commit()
 
     expected_trending = {
         'hg': {
             "py": {"count": 68, "desc": ["Python"]},
             "rst": {"count": 16, "desc": ["Rst"]},
             "css": {"count": 2, "desc": ["Css"]},
             "sh": {"count": 2, "desc": ["Bash"]},
             "bat": {"count": 1, "desc": ["Batch"]},
             "cfg": {"count": 1, "desc": ["Ini"]},
             "html": {"count": 1, "desc": ["EvoqueHtml", "Html"]},
             "ini": {"count": 1, "desc": ["Ini"]},
             "js": {"count": 1, "desc": ["Javascript"]},
             "makefile": {"count": 1, "desc": ["Makefile", "Makefile"]}
         },
         'git': {
             "py": {"count": 68, "desc": ["Python"]},
             "rst": {"count": 16, "desc": ["Rst"]},
             "css": {"count": 2, "desc": ["Css"]},
             "sh": {"count": 2, "desc": ["Bash"]},
             "bat": {"count": 1, "desc": ["Batch"]},
             "cfg": {"count": 1, "desc": ["Ini"]},
             "html": {"count": 1, "desc": ["EvoqueHtml", "Html"]},
             "ini": {"count": 1, "desc": ["Ini"]},
             "js": {"count": 1, "desc": ["Javascript"]},
             "makefile": {"count": 1, "desc": ["Makefile", "Makefile"]}
         },
         'svn': {
             "py": {"count": 75, "desc": ["Python"]},
             "rst": {"count": 16, "desc": ["Rst"]},
             "html": {"count": 11, "desc": ["EvoqueHtml", "Html"]},
             "css": {"count": 2, "desc": ["Css"]},
             "bat": {"count": 1, "desc": ["Batch"]},
             "cfg": {"count": 1, "desc": ["Ini"]},
             "ini": {"count": 1, "desc": ["Ini"]},
             "js": {"count": 1, "desc": ["Javascript"]},
             "makefile": {"count": 1, "desc": ["Makefile", "Makefile"]},
             "sh": {"count": 1, "desc": ["Bash"]}
         },
     }
 
     def test_repo_stats(self, autologin_user, backend, xhr_header):
         response = self.app.get(
             route_path(
                 'repo_stats', repo_name=backend.repo_name, commit_id='tip'),
             extra_environ=xhr_header,
             status=200)
         assert re.match(r'6[\d\.]+ KiB', response.json['size'])
 
     def test_repo_stats_code_stats_enabled(self, autologin_user, backend, xhr_header):
         repo_name = backend.repo_name
 
         # codes stats
         self._enable_stats(repo_name)
         ScmModel().mark_for_invalidation(repo_name)
 
         response = self.app.get(
             route_path(
                 'repo_stats', repo_name=backend.repo_name, commit_id='tip'),
             extra_environ=xhr_header,
             status=200)
 
         expected_data = self.expected_trending[backend.alias]
         returned_stats = response.json['code_stats']
         for k, v in expected_data.items():
             assert v == returned_stats[k]
 
     def test_repo_refs_data(self, backend):
         response = self.app.get(
             route_path('repo_refs_data', repo_name=backend.repo_name),
             status=200)
 
         # Ensure that there is the correct amount of items in the result
         repo = backend.repo.scm_instance()
         data = response.json['results']
         items = sum(len(section['children']) for section in data)
         repo_refs = len(repo.branches) + len(repo.tags) + len(repo.bookmarks)
         assert items == repo_refs
 
     def test_index_shows_missing_requirements_message(
             self, backend, autologin_user):
         repo_name = backend.repo_name
         scm_patcher = mock.patch.object(
             Repository, 'scm_instance', side_effect=RepositoryRequirementError)
 
         with scm_patcher:
             response = self.app.get(
                 route_path('repo_summary', repo_name=repo_name))
         assert_response = response.assert_response()
         assert_response.element_contains(
             '.main .alert-warning strong', 'Missing requirements')
         assert_response.element_contains(
             '.main .alert-warning',
             'Commits cannot be displayed, because this repository '
             'uses one or more extensions, which was not enabled.')
 
     def test_missing_requirements_page_does_not_contains_switch_to(
             self, autologin_user, backend):
         repo_name = backend.repo_name
         scm_patcher = mock.patch.object(
             Repository, 'scm_instance', side_effect=RepositoryRequirementError)
 
         with scm_patcher:
             response = self.app.get(route_path('repo_summary', repo_name=repo_name))
         response.mustcontain(no='Switch To')
 
 
 @pytest.mark.usefixtures('app')
 class TestRepoLocation(object):
 
     @pytest.mark.parametrize("suffix", [u'', u'ąęł'], ids=['', 'non-ascii'])
     def test_missing_filesystem_repo(
             self, autologin_user, backend, suffix, csrf_token):
         repo = backend.create_repo(name_suffix=suffix)
         repo_name = repo.repo_name
 
         # delete from file system
         RepoModel()._delete_filesystem_repo(repo)
 
         # test if the repo is still in the database
         new_repo = RepoModel().get_by_repo_name(repo_name)
         assert new_repo.repo_name == repo_name
 
         # check if repo is not in the filesystem
         assert not repo_on_filesystem(repo_name)
 
         response = self.app.get(
             route_path('repo_summary', repo_name=safe_str(repo_name)), status=302)
 
         msg = 'The repository `%s` cannot be loaded in filesystem. ' \
               'Please check if it exist, or is not damaged.' % repo_name
         assert_session_flash(response, msg)
 
     @pytest.mark.parametrize("suffix", [u'', u'ąęł'], ids=['', 'non-ascii'])
     def test_missing_filesystem_repo_on_repo_check(
             self, autologin_user, backend, suffix, csrf_token):
         repo = backend.create_repo(name_suffix=suffix)
         repo_name = repo.repo_name
 
         # delete from file system
         RepoModel()._delete_filesystem_repo(repo)
 
         # test if the repo is still in the database
         new_repo = RepoModel().get_by_repo_name(repo_name)
         assert new_repo.repo_name == repo_name
 
         # check if repo is not in the filesystem
         assert not repo_on_filesystem(repo_name)
 
         # flush the session
         self.app.get(
             route_path('repo_summary', repo_name=safe_str(repo_name)),
             status=302)
 
         response = self.app.get(
             route_path('repo_creating_check', repo_name=safe_str(repo_name)),
             status=200)
         msg = 'The repository `%s` cannot be loaded in filesystem. ' \
               'Please check if it exist, or is not damaged.' % repo_name
         assert_session_flash(response, msg )
 
 
 @pytest.fixture()
 def summary_view(context_stub, request_stub, user_util):
     """
     Bootstrap view to test the view functions
     """
     request_stub.matched_route = AttributeDict(name='test_view')
 
     request_stub.user = user_util.create_user().AuthUser()
     request_stub.db_repo = user_util.create_repo()
 
     view = RepoSummaryView(context=context_stub, request=request_stub)
     return view
 
 
 @pytest.mark.usefixtures('app')
 class TestCreateReferenceData(object):
 
     @pytest.fixture()
     def example_refs(self):
         section_1_refs = OrderedDict((('a', 'a_id'), ('b', 'b_id')))
         example_refs = [
             ('section_1', section_1_refs, 't1'),
             ('section_2', {'c': 'c_id'}, 't2'),
         ]
         return example_refs
 
     def test_generates_refs_based_on_commit_ids(self, example_refs, summary_view):
         repo = mock.Mock()
         repo.name = 'test-repo'
         repo.alias = 'git'
         full_repo_name = 'pytest-repo-group/' + repo.name
 
         result = summary_view._create_reference_data(
             repo, full_repo_name, example_refs)
 
         expected_files_url = '/{}/files/'.format(full_repo_name)
         expected_result = [
             {
                 'children': [
                     {
                         'id': 'a', 'idx': 0, 'raw_id': 'a_id', 'text': 'a', 'type': 't1',
                         'files_url': expected_files_url + 'a/?at=a',
                     },
                     {
                         'id': 'b', 'idx': 0, 'raw_id': 'b_id', 'text': 'b', 'type': 't1',
                         'files_url': expected_files_url + 'b/?at=b',
                     }
                 ],
                 'text': 'section_1'
             },
             {
                 'children': [
                     {
                         'id': 'c', 'idx': 0, 'raw_id': 'c_id', 'text': 'c', 'type': 't2',
                         'files_url': expected_files_url + 'c/?at=c',
                     }
                 ],
                 'text': 'section_2'
             }]
         assert result == expected_result
 
     def test_generates_refs_with_path_for_svn(self, example_refs, summary_view):
         repo = mock.Mock()
         repo.name = 'test-repo'
         repo.alias = 'svn'
         full_repo_name = 'pytest-repo-group/' + repo.name
 
         result = summary_view._create_reference_data(
             repo, full_repo_name, example_refs)
 
         expected_files_url = '/{}/files/'.format(full_repo_name)
         expected_result = [
             {
                 'children': [
                     {
                         'id': 'a@a_id', 'idx': 0, 'raw_id': 'a_id',
                         'text': 'a', 'type': 't1',
                         'files_url': expected_files_url + 'a_id/a?at=a',
                     },
                     {
                         'id': 'b@b_id', 'idx': 0, 'raw_id': 'b_id',
                         'text': 'b', 'type': 't1',
                         'files_url': expected_files_url + 'b_id/b?at=b',
                     }
                 ],
                 'text': 'section_1'
             },
             {
                 'children': [
                     {
                         'id': 'c@c_id', 'idx': 0, 'raw_id': 'c_id',
                         'text': 'c', 'type': 't2',
                         'files_url': expected_files_url + 'c_id/c?at=c',
                     }
                 ],
                 'text': 'section_2'
             }
         ]
         assert result == expected_result
 
 
 class TestCreateFilesUrl(object):
 
     def test_creates_non_svn_url(self, app, summary_view):
         repo = mock.Mock()
         repo.name = 'abcde'
         full_repo_name = 'test-repo-group/' + repo.name
         ref_name = 'branch1'
         raw_id = 'deadbeef0123456789'
         is_svn = False
 
         with mock.patch('rhodecode.lib.helpers.route_path') as url_mock:
             result = summary_view._create_files_url(
                 repo, full_repo_name, ref_name, raw_id, is_svn)
         url_mock.assert_called_once_with(
             'repo_files', repo_name=full_repo_name, commit_id=ref_name,
             f_path='', _query=dict(at=ref_name))
         assert result == url_mock.return_value
 
     def test_creates_svn_url(self, app, summary_view):
         repo = mock.Mock()
         repo.name = 'abcde'
         full_repo_name = 'test-repo-group/' + repo.name
         ref_name = 'branch1'
         raw_id = 'deadbeef0123456789'
         is_svn = True
 
         with mock.patch('rhodecode.lib.helpers.route_path') as url_mock:
             result = summary_view._create_files_url(
                 repo, full_repo_name, ref_name, raw_id, is_svn)
         url_mock.assert_called_once_with(
             'repo_files', repo_name=full_repo_name, f_path=ref_name,
             commit_id=raw_id, _query=dict(at=ref_name))
         assert result == url_mock.return_value
 
     def test_name_has_slashes(self, app, summary_view):
         repo = mock.Mock()
         repo.name = 'abcde'
         full_repo_name = 'test-repo-group/' + repo.name
         ref_name = 'branch1/branch2'
         raw_id = 'deadbeef0123456789'
         is_svn = False
 
         with mock.patch('rhodecode.lib.helpers.route_path') as url_mock:
             result = summary_view._create_files_url(
                 repo, full_repo_name, ref_name, raw_id, is_svn)
         url_mock.assert_called_once_with(
             'repo_files', repo_name=full_repo_name, commit_id=raw_id,
             f_path='', _query=dict(at=ref_name))
         assert result == url_mock.return_value
 
 
 class TestReferenceItems(object):
     repo = mock.Mock()
     repo.name = 'pytest-repo'
     repo_full_name = 'pytest-repo-group/' + repo.name
     ref_type = 'branch'
     fake_url = '/abcde/'
 
     @staticmethod
     def _format_function(name, id_):
         return 'format_function_{}_{}'.format(name, id_)
 
     def test_creates_required_amount_of_items(self, summary_view):
         amount = 100
         refs = {
             'ref{}'.format(i): '{0:040d}'.format(i)
             for i in range(amount)
         }
 
         url_patcher = mock.patch.object(summary_view, '_create_files_url')
         svn_patcher = mock.patch('rhodecode.lib.helpers.is_svn',
                                  return_value=False)
 
         with url_patcher as url_mock, svn_patcher:
             result = summary_view._create_reference_items(
                 self.repo, self.repo_full_name, refs, self.ref_type,
                 self._format_function)
         assert len(result) == amount
         assert url_mock.call_count == amount
 
     def test_single_item_details(self, summary_view):
         ref_name = 'ref1'
         ref_id = 'deadbeef'
         refs = {
             ref_name: ref_id
         }
 
         svn_patcher = mock.patch('rhodecode.lib.helpers.is_svn',
                                  return_value=False)
 
         url_patcher = mock.patch.object(
             summary_view, '_create_files_url', return_value=self.fake_url)
 
         with url_patcher as url_mock, svn_patcher:
             result = summary_view._create_reference_items(
                 self.repo, self.repo_full_name, refs, self.ref_type,
                 self._format_function)
 
         url_mock.assert_called_once_with(
             self.repo, self.repo_full_name, ref_name, ref_id, False)
         expected_result = [
             {
                 'text': ref_name,
                 'id': self._format_function(ref_name, ref_id),
                 'raw_id': ref_id,
                 'idx': 0,
                 'type': self.ref_type,
                 'files_url': self.fake_url
             }
         ]
         assert result == expected_result

rhodecode/apps/repository/views/repo_commits.py

@@ -1,819 +1,819 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 logging
 import collections
 
 from pyramid.httpexceptions import (
     HTTPNotFound, HTTPBadRequest, HTTPFound, HTTPForbidden, HTTPConflict)
 from pyramid.renderers import render
 from pyramid.response import Response
 
 from rhodecode.apps._base import RepoAppView
 from rhodecode.apps.file_store import utils as store_utils
 from rhodecode.apps.file_store.exceptions import FileNotAllowedException, FileOverSizeException
 
 from rhodecode.lib import diffs, codeblocks, channelstream
 from rhodecode.lib.auth import (
     LoginRequired, HasRepoPermissionAnyDecorator, NotAnonymous, CSRFRequired)
 from rhodecode.lib.ext_json import json
-from rhodecode.lib.compat import OrderedDict
+from collections import OrderedDict
 from rhodecode.lib.diffs import (
     cache_diff, load_cached_diff, diff_cache_exist, get_diff_context,
     get_diff_whitespace_flag)
 from rhodecode.lib.exceptions import StatusChangeOnClosedPullRequestError, CommentVersionMismatch
 import rhodecode.lib.helpers as h
 from rhodecode.lib.utils2 import safe_unicode, str2bool, StrictAttributeDict, safe_str
 from rhodecode.lib.vcs.backends.base import EmptyCommit
 from rhodecode.lib.vcs.exceptions import (
     RepositoryError, CommitDoesNotExistError)
 from rhodecode.model.db import ChangesetComment, ChangesetStatus, FileStore, \
     ChangesetCommentHistory
 from rhodecode.model.changeset_status import ChangesetStatusModel
 from rhodecode.model.comment import CommentsModel
 from rhodecode.model.meta import Session
 from rhodecode.model.settings import VcsSettingsModel
 
 log = logging.getLogger(__name__)
 
 
 def _update_with_GET(params, request):
     for k in ['diff1', 'diff2', 'diff']:
         params[k] += request.GET.getall(k)
 
 
 class RepoCommitsView(RepoAppView):
     def load_default_context(self):
         c = self._get_local_tmpl_context(include_app_defaults=True)
         c.rhodecode_repo = self.rhodecode_vcs_repo
 
         return c
 
     def _is_diff_cache_enabled(self, target_repo):
         caching_enabled = self._get_general_setting(
             target_repo, 'rhodecode_diff_cache')
         log.debug('Diff caching enabled: %s', caching_enabled)
         return caching_enabled
 
     def _commit(self, commit_id_range, method):
         _ = self.request.translate
         c = self.load_default_context()
         c.fulldiff = self.request.GET.get('fulldiff')
         redirect_to_combined = str2bool(self.request.GET.get('redirect_combined'))
 
         # fetch global flags of ignore ws or context lines
         diff_context = get_diff_context(self.request)
         hide_whitespace_changes = get_diff_whitespace_flag(self.request)
 
         # diff_limit will cut off the whole diff if the limit is applied
         # otherwise it will just hide the big files from the front-end
         diff_limit = c.visual.cut_off_limit_diff
         file_limit = c.visual.cut_off_limit_file
 
         # get ranges of commit ids if preset
         commit_range = commit_id_range.split('...')[:2]
 
         try:
             pre_load = ['affected_files', 'author', 'branch', 'date',
                         'message', 'parents']
             if self.rhodecode_vcs_repo.alias == 'hg':
                 pre_load += ['hidden', 'obsolete', 'phase']
 
             if len(commit_range) == 2:
                 commits = self.rhodecode_vcs_repo.get_commits(
                     start_id=commit_range[0], end_id=commit_range[1],
                     pre_load=pre_load, translate_tags=False)
                 commits = list(commits)
             else:
                 commits = [self.rhodecode_vcs_repo.get_commit(
                     commit_id=commit_id_range, pre_load=pre_load)]
 
             c.commit_ranges = commits
             if not c.commit_ranges:
                 raise RepositoryError('The commit range returned an empty result')
         except CommitDoesNotExistError as e:
             msg = _('No such commit exists. Org exception: `{}`').format(safe_str(e))
             h.flash(msg, category='error')
             raise HTTPNotFound()
         except Exception:
             log.exception("General failure")
             raise HTTPNotFound()
         single_commit = len(c.commit_ranges) == 1
 
         if redirect_to_combined and not single_commit:
             source_ref = getattr(c.commit_ranges[0].parents[0]
                                  if c.commit_ranges[0].parents else h.EmptyCommit(), 'raw_id')
             target_ref = c.commit_ranges[-1].raw_id
             next_url = h.route_path(
                 'repo_compare',
                 repo_name=c.repo_name,
                 source_ref_type='rev',
                 source_ref=source_ref,
                 target_ref_type='rev',
                 target_ref=target_ref)
             raise HTTPFound(next_url)
 
         c.changes = OrderedDict()
         c.lines_added = 0
         c.lines_deleted = 0
 
         # auto collapse if we have more than limit
         collapse_limit = diffs.DiffProcessor._collapse_commits_over
         c.collapse_all_commits = len(c.commit_ranges) > collapse_limit
 
         c.commit_statuses = ChangesetStatus.STATUSES
         c.inline_comments = []
         c.files = []
 
         c.comments = []
         c.unresolved_comments = []
         c.resolved_comments = []
 
         # Single commit
         if single_commit:
             commit = c.commit_ranges[0]
             c.comments = CommentsModel().get_comments(
                 self.db_repo.repo_id,
                 revision=commit.raw_id)
 
             # comments from PR
             statuses = ChangesetStatusModel().get_statuses(
                 self.db_repo.repo_id, commit.raw_id,
                 with_revisions=True)
 
             prs = set()
             reviewers = list()
             reviewers_duplicates = set()  # to not have duplicates from multiple votes
             for c_status in statuses:
 
                 # extract associated pull-requests from votes
                 if c_status.pull_request:
                     prs.add(c_status.pull_request)
 
                 # extract reviewers
                 _user_id = c_status.author.user_id
                 if _user_id not in reviewers_duplicates:
                     reviewers.append(
                         StrictAttributeDict({
                             'user': c_status.author,
 
                             # fake attributed for commit, page that we don't have
                             # but we share the display with PR page
                             'mandatory': False,
                             'reasons': [],
                             'rule_user_group_data': lambda: None
                         })
                     )
                     reviewers_duplicates.add(_user_id)
 
             c.reviewers_count = len(reviewers)
             c.observers_count = 0
 
             # from associated statuses, check the pull requests, and
             # show comments from them
             for pr in prs:
                 c.comments.extend(pr.comments)
 
             c.unresolved_comments = CommentsModel()\
                 .get_commit_unresolved_todos(commit.raw_id)
             c.resolved_comments = CommentsModel()\
                 .get_commit_resolved_todos(commit.raw_id)
 
             c.inline_comments_flat = CommentsModel()\
                 .get_commit_inline_comments(commit.raw_id)
 
             review_statuses = ChangesetStatusModel().aggregate_votes_by_user(
                 statuses, reviewers)
 
             c.commit_review_status = ChangesetStatus.STATUS_NOT_REVIEWED
 
             c.commit_set_reviewers_data_json = collections.OrderedDict({'reviewers': []})
 
             for review_obj, member, reasons, mandatory, status in review_statuses:
                 member_reviewer = h.reviewer_as_json(
                     member, reasons=reasons, mandatory=mandatory, role=None,
                     user_group=None
                 )
 
                 current_review_status = status[0][1].status if status else ChangesetStatus.STATUS_NOT_REVIEWED
                 member_reviewer['review_status'] = current_review_status
                 member_reviewer['review_status_label'] = h.commit_status_lbl(current_review_status)
                 member_reviewer['allowed_to_update'] = False
                 c.commit_set_reviewers_data_json['reviewers'].append(member_reviewer)
 
             c.commit_set_reviewers_data_json = json.dumps(c.commit_set_reviewers_data_json)
 
             # NOTE(marcink): this uses the same voting logic as in pull-requests
             c.commit_review_status = ChangesetStatusModel().calculate_status(review_statuses)
             c.commit_broadcast_channel = channelstream.comment_channel(c.repo_name, commit_obj=commit)
 
         diff = None
         # Iterate over ranges (default commit view is always one commit)
         for commit in c.commit_ranges:
             c.changes[commit.raw_id] = []
 
             commit2 = commit
             commit1 = commit.first_parent
 
             if method == 'show':
                 inline_comments = CommentsModel().get_inline_comments(
                     self.db_repo.repo_id, revision=commit.raw_id)
                 c.inline_cnt = len(CommentsModel().get_inline_comments_as_list(
                     inline_comments))
                 c.inline_comments = inline_comments
 
                 cache_path = self.rhodecode_vcs_repo.get_create_shadow_cache_pr_path(
                     self.db_repo)
                 cache_file_path = diff_cache_exist(
                     cache_path, 'diff', commit.raw_id,
                     hide_whitespace_changes, diff_context, c.fulldiff)
 
                 caching_enabled = self._is_diff_cache_enabled(self.db_repo)
                 force_recache = str2bool(self.request.GET.get('force_recache'))
 
                 cached_diff = None
                 if caching_enabled:
                     cached_diff = load_cached_diff(cache_file_path)
 
                 has_proper_diff_cache = cached_diff and cached_diff.get('diff')
                 if not force_recache and has_proper_diff_cache:
                     diffset = cached_diff['diff']
                 else:
                     vcs_diff = self.rhodecode_vcs_repo.get_diff(
                         commit1, commit2,
                         ignore_whitespace=hide_whitespace_changes,
                         context=diff_context)
 
                     diff_processor = diffs.DiffProcessor(
                         vcs_diff, format='newdiff', diff_limit=diff_limit,
                         file_limit=file_limit, show_full_diff=c.fulldiff)
 
                     _parsed = diff_processor.prepare()
 
                     diffset = codeblocks.DiffSet(
                         repo_name=self.db_repo_name,
                         source_node_getter=codeblocks.diffset_node_getter(commit1),
                         target_node_getter=codeblocks.diffset_node_getter(commit2))
 
                     diffset = self.path_filter.render_patchset_filtered(
                         diffset, _parsed, commit1.raw_id, commit2.raw_id)
 
                     # save cached diff
                     if caching_enabled:
                         cache_diff(cache_file_path, diffset, None)
 
                 c.limited_diff = diffset.limited_diff
                 c.changes[commit.raw_id] = diffset
             else:
                 # TODO(marcink): no cache usage here...
                 _diff = self.rhodecode_vcs_repo.get_diff(
                     commit1, commit2,
                     ignore_whitespace=hide_whitespace_changes, context=diff_context)
                 diff_processor = diffs.DiffProcessor(
                     _diff, format='newdiff', diff_limit=diff_limit,
                     file_limit=file_limit, show_full_diff=c.fulldiff)
                 # downloads/raw we only need RAW diff nothing else
                 diff = self.path_filter.get_raw_patch(diff_processor)
                 c.changes[commit.raw_id] = [None, None, None, None, diff, None, None]
 
         # sort comments by how they were generated
         c.comments = sorted(c.comments, key=lambda x: x.comment_id)
         c.at_version_num = None
 
         if len(c.commit_ranges) == 1:
             c.commit = c.commit_ranges[0]
             c.parent_tmpl = ''.join(
                 '# Parent  %s\n' % x.raw_id for x in c.commit.parents)
 
         if method == 'download':
             response = Response(diff)
             response.content_type = 'text/plain'
             response.content_disposition = (
                 'attachment; filename=%s.diff' % commit_id_range[:12])
             return response
         elif method == 'patch':
             c.diff = safe_unicode(diff)
             patch = render(
                 'rhodecode:templates/changeset/patch_changeset.mako',
                 self._get_template_context(c), self.request)
             response = Response(patch)
             response.content_type = 'text/plain'
             return response
         elif method == 'raw':
             response = Response(diff)
             response.content_type = 'text/plain'
             return response
         elif method == 'show':
             if len(c.commit_ranges) == 1:
                 html = render(
                     'rhodecode:templates/changeset/changeset.mako',
                     self._get_template_context(c), self.request)
                 return Response(html)
             else:
                 c.ancestor = None
                 c.target_repo = self.db_repo
                 html = render(
                     'rhodecode:templates/changeset/changeset_range.mako',
                     self._get_template_context(c), self.request)
                 return Response(html)
 
         raise HTTPBadRequest()
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_commit_show(self):
         commit_id = self.request.matchdict['commit_id']
         return self._commit(commit_id, method='show')
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_commit_raw(self):
         commit_id = self.request.matchdict['commit_id']
         return self._commit(commit_id, method='raw')
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_commit_patch(self):
         commit_id = self.request.matchdict['commit_id']
         return self._commit(commit_id, method='patch')
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_commit_download(self):
         commit_id = self.request.matchdict['commit_id']
         return self._commit(commit_id, method='download')
 
     def _commit_comments_create(self, commit_id, comments):
         _ = self.request.translate
         data = {}
         if not comments:
             return
 
         commit = self.db_repo.get_commit(commit_id)
 
         all_drafts = len([x for x in comments if str2bool(x['is_draft'])]) == len(comments)
         for entry in comments:
             c = self.load_default_context()
             comment_type = entry['comment_type']
             text = entry['text']
             status = entry['status']
             is_draft = str2bool(entry['is_draft'])
             resolves_comment_id = entry['resolves_comment_id']
             f_path = entry['f_path']
             line_no = entry['line']
             target_elem_id = 'file-{}'.format(h.safeid(h.safe_unicode(f_path)))
 
             if status:
                 text = text or (_('Status change %(transition_icon)s %(status)s')
                                 % {'transition_icon': '>',
                                    'status': ChangesetStatus.get_status_lbl(status)})
 
             comment = CommentsModel().create(
                 text=text,
                 repo=self.db_repo.repo_id,
                 user=self._rhodecode_db_user.user_id,
                 commit_id=commit_id,
                 f_path=f_path,
                 line_no=line_no,
                 status_change=(ChangesetStatus.get_status_lbl(status)
                                if status else None),
                 status_change_type=status,
                 comment_type=comment_type,
                 is_draft=is_draft,
                 resolves_comment_id=resolves_comment_id,
                 auth_user=self._rhodecode_user,
                 send_email=not is_draft,  # skip notification for draft comments
             )
             is_inline = comment.is_inline
 
             # get status if set !
             if status:
                 # `dont_allow_on_closed_pull_request = True` means
                 # if latest status was from pull request and it's closed
                 # disallow changing status !
 
                 try:
                     ChangesetStatusModel().set_status(
                         self.db_repo.repo_id,
                         status,
                         self._rhodecode_db_user.user_id,
                         comment,
                         revision=commit_id,
                         dont_allow_on_closed_pull_request=True
                     )
                 except StatusChangeOnClosedPullRequestError:
                     msg = _('Changing the status of a commit associated with '
                             'a closed pull request is not allowed')
                     log.exception(msg)
                     h.flash(msg, category='warning')
                     raise HTTPFound(h.route_path(
                         'repo_commit', repo_name=self.db_repo_name,
                         commit_id=commit_id))
 
             Session().flush()
             # this is somehow required to get access to some relationship
             # loaded on comment
             Session().refresh(comment)
 
             # skip notifications for drafts
             if not is_draft:
                 CommentsModel().trigger_commit_comment_hook(
                     self.db_repo, self._rhodecode_user, 'create',
                     data={'comment': comment, 'commit': commit})
 
             comment_id = comment.comment_id
             data[comment_id] = {
                 'target_id': target_elem_id
             }
             Session().flush()
 
             c.co = comment
             c.at_version_num = 0
             c.is_new = True
             rendered_comment = render(
                 'rhodecode:templates/changeset/changeset_comment_block.mako',
                 self._get_template_context(c), self.request)
 
             data[comment_id].update(comment.get_dict())
             data[comment_id].update({'rendered_text': rendered_comment})
 
         # finalize, commit and redirect
         Session().commit()
 
         # skip channelstream for draft comments
         if not all_drafts:
             comment_broadcast_channel = channelstream.comment_channel(
                 self.db_repo_name, commit_obj=commit)
 
             comment_data = data
             posted_comment_type = 'inline' if is_inline else 'general'
             if len(data) == 1:
                 msg = _('posted {} new {} comment').format(len(data), posted_comment_type)
             else:
                 msg = _('posted {} new {} comments').format(len(data), posted_comment_type)
 
             channelstream.comment_channelstream_push(
                 self.request, comment_broadcast_channel, self._rhodecode_user, msg,
                 comment_data=comment_data)
 
         return data
 
     @LoginRequired()
     @NotAnonymous()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_commit_comment_create(self):
         _ = self.request.translate
         commit_id = self.request.matchdict['commit_id']
 
         multi_commit_ids = []
         for _commit_id in self.request.POST.get('commit_ids', '').split(','):
             if _commit_id not in ['', None, EmptyCommit.raw_id]:
                 if _commit_id not in multi_commit_ids:
                     multi_commit_ids.append(_commit_id)
 
         commit_ids = multi_commit_ids or [commit_id]
 
         data = []
         # Multiple comments for each passed commit id
         for current_id in filter(None, commit_ids):
             comment_data = {
                 'comment_type': self.request.POST.get('comment_type'),
                 'text': self.request.POST.get('text'),
                 'status': self.request.POST.get('changeset_status', None),
                 'is_draft': self.request.POST.get('draft'),
                 'resolves_comment_id': self.request.POST.get('resolves_comment_id', None),
                 'close_pull_request': self.request.POST.get('close_pull_request'),
                 'f_path': self.request.POST.get('f_path'),
                 'line': self.request.POST.get('line'),
             }
             comment = self._commit_comments_create(commit_id=current_id, comments=[comment_data])
             data.append(comment)
 
         return data if len(data) > 1 else data[0]
 
     @LoginRequired()
     @NotAnonymous()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_commit_comment_preview(self):
         # Technically a CSRF token is not needed as no state changes with this
         # call. However, as this is a POST is better to have it, so automated
         # tools don't flag it as potential CSRF.
         # Post is required because the payload could be bigger than the maximum
         # allowed by GET.
 
         text = self.request.POST.get('text')
         renderer = self.request.POST.get('renderer') or 'rst'
         if text:
             return h.render(text, renderer=renderer, mentions=True,
                             repo_name=self.db_repo_name)
         return ''
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_commit_comment_history_view(self):
         c = self.load_default_context()
         comment_id = self.request.matchdict['comment_id']
         comment_history_id = self.request.matchdict['comment_history_id']
 
         comment = ChangesetComment.get_or_404(comment_id)
         comment_owner = (comment.author.user_id == self._rhodecode_db_user.user_id)
         if comment.draft and not comment_owner:
             # if we see draft comments history, we only allow this for owner
             raise HTTPNotFound()
 
         comment_history = ChangesetCommentHistory.get_or_404(comment_history_id)
         is_repo_comment = comment_history.comment.repo.repo_id == self.db_repo.repo_id
 
         if is_repo_comment:
             c.comment_history = comment_history
 
             rendered_comment = render(
                 'rhodecode:templates/changeset/comment_history.mako',
                 self._get_template_context(c), self.request)
             return rendered_comment
         else:
             log.warning('No permissions for user %s to show comment_history_id: %s',
                         self._rhodecode_db_user, comment_history_id)
             raise HTTPNotFound()
 
     @LoginRequired()
     @NotAnonymous()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_commit_comment_attachment_upload(self):
         c = self.load_default_context()
         upload_key = 'attachment'
 
         file_obj = self.request.POST.get(upload_key)
 
         if file_obj is None:
             self.request.response.status = 400
             return {'store_fid': None,
                     'access_path': None,
                     'error': '{} data field is missing'.format(upload_key)}
 
         if not hasattr(file_obj, 'filename'):
             self.request.response.status = 400
             return {'store_fid': None,
                     'access_path': None,
                     'error': 'filename cannot be read from the data field'}
 
         filename = file_obj.filename
         file_display_name = filename
 
         metadata = {
             'user_uploaded': {'username': self._rhodecode_user.username,
                               'user_id': self._rhodecode_user.user_id,
                               'ip': self._rhodecode_user.ip_addr}}
 
         # TODO(marcink): allow .ini configuration for allowed_extensions, and file-size
         allowed_extensions = [
             'gif', '.jpeg', '.jpg', '.png', '.docx', '.gz', '.log', '.pdf',
             '.pptx', '.txt', '.xlsx', '.zip']
         max_file_size = 10 * 1024 * 1024  # 10MB, also validated via dropzone.js
 
         try:
             storage = store_utils.get_file_storage(self.request.registry.settings)
             store_uid, metadata = storage.save_file(
                 file_obj.file, filename, extra_metadata=metadata,
                 extensions=allowed_extensions, max_filesize=max_file_size)
         except FileNotAllowedException:
             self.request.response.status = 400
             permitted_extensions = ', '.join(allowed_extensions)
             error_msg = 'File `{}` is not allowed. ' \
                         'Only following extensions are permitted: {}'.format(
                             filename, permitted_extensions)
             return {'store_fid': None,
                     'access_path': None,
                     'error': error_msg}
         except FileOverSizeException:
             self.request.response.status = 400
             limit_mb = h.format_byte_size_binary(max_file_size)
             return {'store_fid': None,
                     'access_path': None,
                     'error': 'File {} is exceeding allowed limit of {}.'.format(
                         filename, limit_mb)}
 
         try:
             entry = FileStore.create(
                 file_uid=store_uid, filename=metadata["filename"],
                 file_hash=metadata["sha256"], file_size=metadata["size"],
                 file_display_name=file_display_name,
                 file_description=u'comment attachment `{}`'.format(safe_unicode(filename)),
                 hidden=True, check_acl=True, user_id=self._rhodecode_user.user_id,
                 scope_repo_id=self.db_repo.repo_id
             )
             Session().add(entry)
             Session().commit()
             log.debug('Stored upload in DB as %s', entry)
         except Exception:
             log.exception('Failed to store file %s', filename)
             self.request.response.status = 400
             return {'store_fid': None,
                     'access_path': None,
                     'error': 'File {} failed to store in DB.'.format(filename)}
 
         Session().commit()
 
         return {
             'store_fid': store_uid,
             'access_path': h.route_path(
                 'download_file', fid=store_uid),
             'fqn_access_path': h.route_url(
                 'download_file', fid=store_uid),
             'repo_access_path': h.route_path(
                 'repo_artifacts_get', repo_name=self.db_repo_name, uid=store_uid),
             'repo_fqn_access_path': h.route_url(
                 'repo_artifacts_get', repo_name=self.db_repo_name, uid=store_uid),
         }
 
     @LoginRequired()
     @NotAnonymous()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_commit_comment_delete(self):
         commit_id = self.request.matchdict['commit_id']
         comment_id = self.request.matchdict['comment_id']
 
         comment = ChangesetComment.get_or_404(comment_id)
         if not comment:
             log.debug('Comment with id:%s not found, skipping', comment_id)
             # comment already deleted in another call probably
             return True
 
         if comment.immutable:
             # don't allow deleting comments that are immutable
             raise HTTPForbidden()
 
         is_repo_admin = h.HasRepoPermissionAny('repository.admin')(self.db_repo_name)
         super_admin = h.HasPermissionAny('hg.admin')()
         comment_owner = (comment.author.user_id == self._rhodecode_db_user.user_id)
         is_repo_comment = comment.repo.repo_id == self.db_repo.repo_id
         comment_repo_admin = is_repo_admin and is_repo_comment
 
         if comment.draft and not comment_owner:
             # We never allow to delete draft comments for other than owners
             raise HTTPNotFound()
 
         if super_admin or comment_owner or comment_repo_admin:
             CommentsModel().delete(comment=comment, auth_user=self._rhodecode_user)
             Session().commit()
             return True
         else:
             log.warning('No permissions for user %s to delete comment_id: %s',
                         self._rhodecode_db_user, comment_id)
             raise HTTPNotFound()
 
     @LoginRequired()
     @NotAnonymous()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_commit_comment_edit(self):
         self.load_default_context()
 
         commit_id = self.request.matchdict['commit_id']
         comment_id = self.request.matchdict['comment_id']
         comment = ChangesetComment.get_or_404(comment_id)
 
         if comment.immutable:
             # don't allow deleting comments that are immutable
             raise HTTPForbidden()
 
         is_repo_admin = h.HasRepoPermissionAny('repository.admin')(self.db_repo_name)
         super_admin = h.HasPermissionAny('hg.admin')()
         comment_owner = (comment.author.user_id == self._rhodecode_db_user.user_id)
         is_repo_comment = comment.repo.repo_id == self.db_repo.repo_id
         comment_repo_admin = is_repo_admin and is_repo_comment
 
         if super_admin or comment_owner or comment_repo_admin:
             text = self.request.POST.get('text')
             version = self.request.POST.get('version')
             if text == comment.text:
                 log.warning(
                     'Comment(repo): '
                     'Trying to create new version '
                     'with the same comment body {}'.format(
                         comment_id,
                     )
                 )
                 raise HTTPNotFound()
 
             if version.isdigit():
                 version = int(version)
             else:
                 log.warning(
                     'Comment(repo): Wrong version type {} {} '
                     'for comment {}'.format(
                         version,
                         type(version),
                         comment_id,
                     )
                 )
                 raise HTTPNotFound()
 
             try:
                 comment_history = CommentsModel().edit(
                     comment_id=comment_id,
                     text=text,
                     auth_user=self._rhodecode_user,
                     version=version,
                 )
             except CommentVersionMismatch:
                 raise HTTPConflict()
 
             if not comment_history:
                 raise HTTPNotFound()
 
             if not comment.draft:
                 commit = self.db_repo.get_commit(commit_id)
                 CommentsModel().trigger_commit_comment_hook(
                     self.db_repo, self._rhodecode_user, 'edit',
                     data={'comment': comment, 'commit': commit})
 
             Session().commit()
             return {
                 'comment_history_id': comment_history.comment_history_id,
                 'comment_id': comment.comment_id,
                 'comment_version': comment_history.version,
                 'comment_author_username': comment_history.author.username,
                 'comment_author_gravatar': h.gravatar_url(comment_history.author.email, 16),
                 'comment_created_on': h.age_component(comment_history.created_on,
                                                       time_is_local=True),
             }
         else:
             log.warning('No permissions for user %s to edit comment_id: %s',
                         self._rhodecode_db_user, comment_id)
             raise HTTPNotFound()
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_commit_data(self):
         commit_id = self.request.matchdict['commit_id']
         self.load_default_context()
 
         try:
             return self.rhodecode_vcs_repo.get_commit(commit_id=commit_id)
         except CommitDoesNotExistError as e:
             return EmptyCommit(message=str(e))
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_commit_children(self):
         commit_id = self.request.matchdict['commit_id']
         self.load_default_context()
 
         try:
             commit = self.rhodecode_vcs_repo.get_commit(commit_id=commit_id)
             children = commit.children
         except CommitDoesNotExistError:
             children = []
 
         result = {"results": children}
         return result
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_commit_parents(self):
         commit_id = self.request.matchdict['commit_id']
         self.load_default_context()
 
         try:
             commit = self.rhodecode_vcs_repo.get_commit(commit_id=commit_id)
             parents = commit.parents
         except CommitDoesNotExistError:
             parents = []
         result = {"results": parents}
         return result

rhodecode/lib/pidlock.py

@@ -1,149 +1,148 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 errno
 
 from multiprocessing.util import Finalize
 
-from rhodecode.lib.compat import kill
 
 
 class LockHeld(Exception):
     pass
 
 
 class DaemonLock(object):
     """daemon locking
     USAGE:
     try:
         l = DaemonLock(file_='/path/tolockfile',desc='test lock')
         main()
         l.release()
     except LockHeld:
         sys.exit(1)
     """
 
     def __init__(self, file_=None, callbackfn=None,
                  desc='daemon lock', debug=False):
 
         lock_name = os.path.join(os.path.dirname(__file__), 'running.lock')
         self.pidfile = file_ if file_ else lock_name
         self.callbackfn = callbackfn
         self.desc = desc
         self.debug = debug
         self.held = False
         #run the lock automatically !
         self.lock()
         self._finalize = Finalize(self, DaemonLock._on_finalize,
                                   args=(self, debug), exitpriority=10)
 
     @staticmethod
     def _on_finalize(lock, debug):
         if lock.held:
             if debug:
                 print('leck held finilazing and running lock.release()')
             lock.release()
 
     def lock(self):
         """
         locking function, if lock is present it
         will raise LockHeld exception
         """
         lockname = '%s' % (os.getpid())
         if self.debug:
             print('running lock')
         self.trylock()
         self.makelock(lockname, self.pidfile)
         return True
 
     def trylock(self):
         running_pid = False
         if self.debug:
             print('checking for already running process')
         try:
             with open(self.pidfile, 'r') as f:
                 try:
                     running_pid = int(f.readline())
                 except ValueError:
                     running_pid = -1
 
             if self.debug:
                 print('lock file present running_pid: %s, '
                       'checking for execution' % (running_pid,))
             # Now we check the PID from lock file matches to the current
             # process PID
             if running_pid:
                 try:
-                    kill(running_pid, 0)
+                    os.kill(running_pid, 0)
                 except OSError as exc:
                     if exc.errno in (errno.ESRCH, errno.EPERM):
                         print("Lock File is there but"
                               " the program is not running")
                         print("Removing lock file for the: %s" % running_pid)
                         self.release()
                     else:
                         raise
                 else:
                     print("You already have an instance of the program running")
                     print("It is running as process %s" % running_pid)
                     raise LockHeld()
 
         except IOError as e:
             if e.errno != 2:
                 raise
 
     def release(self):
         """releases the pid by removing the pidfile
         """
         if self.debug:
             print('trying to release the pidlock')
 
         if self.callbackfn:
             # execute callback function on release
             if self.debug:
                 print('executing callback function %s' % self.callbackfn)
             self.callbackfn()
         try:
             if self.debug:
                 print('removing pidfile %s' % self.pidfile)
             os.remove(self.pidfile)
             self.held = False
         except OSError as e:
             if self.debug:
                 print('removing pidfile failed %s' % e)
             pass
 
     def makelock(self, lockname, pidfile):
         """
         this function will make an actual lock
 
         :param lockname: acctual pid of file
         :param pidfile: the file to write the pid in
         """
         if self.debug:
             print('creating a file %s and pid: %s' % (pidfile, lockname))
 
         dir_, file_ = os.path.split(pidfile)
         if not os.path.isdir(dir_):
             os.makedirs(dir_)
         with open(self.pidfile, 'wb') as f:
             f.write(lockname)
         self.held = True

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

@@ -1,1052 +1,1052 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 repository module
 """
 
 import logging
 import os
 import re
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
-from rhodecode.lib.compat import OrderedDict
+from collections import OrderedDict
 from rhodecode.lib.datelib import (
     utcdate_fromtimestamp, makedate, date_astimestamp)
 from rhodecode.lib.utils import safe_unicode, safe_str
 from rhodecode.lib.utils2 import CachedProperty
 from rhodecode.lib.vcs import connection, path as vcspath
 from rhodecode.lib.vcs.backends.base import (
     BaseRepository, CollectionGenerator, Config, MergeResponse,
     MergeFailureReason, Reference)
 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.exceptions import (
     CommitDoesNotExistError, EmptyRepositoryError,
     RepositoryError, TagAlreadyExistError, TagDoesNotExistError, VCSError, UnresolvedFilesInRepo)
 
 
 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 = os.environ.get('GIT_DEFAULT_BRANCH_NAME') or 'master'
     DEFAULT_REF = 'branch:{}'.format(DEFAULT_BRANCH_NAME)
 
     contact = BaseRepository.DEFAULT_CONTACT
 
     def __init__(self, repo_path, config=None, create=False, src_url=None,
                  do_workspace_checkout=False, with_wire=None, bare=False):
 
         self.path = safe_str(os.path.abspath(repo_path))
         self.config = config if config else self.get_default_config()
         self.with_wire = with_wire or {"cache": False}  # default should not use cache
 
         self._init_repo(create, src_url, do_workspace_checkout, bare)
 
         # caches
         self._commit_ids = {}
 
     @LazyProperty
     def _remote(self):
         repo_id = self.path
         return connection.Git(self.path, repo_id, self.config, with_wire=self.with_wire)
 
     @LazyProperty
     def bare(self):
         return self._remote.bare()
 
     @LazyProperty
     def head(self):
         return self._remote.head()
 
     @CachedProperty
     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))
 
         skip_stderr_log = opts.pop('skip_stderr_log', False)
         out, err = self._remote.run_git_command(cmd, **opts)
         if err and not skip_stderr_log:
             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, do_workspace_checkout=False,
                    bare=False):
         if create and os.path.exists(self.path):
             raise RepositoryError(
                 "Cannot create repository at %s, location already exist"
                 % self.path)
 
         if bare and do_workspace_checkout:
             raise RepositoryError("Cannot update a bare repository")
         try:
 
             if src_url:
                 # check URL before any actions
                 GitRepository.check_url(src_url, self.config)
 
             if create:
                 os.makedirs(self.path, mode=0o755)
 
                 if bare:
                     self._remote.init_bare()
                 else:
                     self._remote.init()
 
                 if src_url and bare:
                     # bare repository only allows a fetch and checkout is not allowed
                     self.fetch(src_url, commit_ids=None)
                 elif src_url:
                     self.pull(src_url, commit_ids=None,
                               update_after=do_workspace_checkout)
 
             else:
                 if not self._remote.assert_correct_path():
                     raise RepositoryError(
                         'Path "%s" does not contain a Git repository' %
                         (self.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):
         return self._remote.get_all_commit_ids()
 
     def _get_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
 
         head = self._remote.head(show_exc=False)
 
         if not head:
             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 = []
                 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 _lookup_commit(self, commit_id_or_idx, translate_tag=True, maybe_unreachable=False, reference_obj=None):
 
         def is_null(value):
             return len(value) == commit_id_or_idx.count('0')
 
         if commit_id_or_idx in (None, '', 'tip', 'HEAD', 'head', -1):
             return self.commit_ids[-1]
 
         commit_missing_err = "Commit {} does not exist for `{}`".format(
             *map(safe_str, [commit_id_or_idx, self.name]))
 
         is_bstr = isinstance(commit_id_or_idx, (str, unicode))
         is_branch = reference_obj and reference_obj.branch
 
         lookup_ok = False
         if is_bstr:
             # Need to call remote to translate id for tagging scenarios,
             # or branch that are numeric
             try:
                 remote_data = self._remote.get_object(commit_id_or_idx,
                                                       maybe_unreachable=maybe_unreachable)
                 commit_id_or_idx = remote_data["commit_id"]
                 lookup_ok = True
             except (CommitDoesNotExistError,):
                 lookup_ok = False
 
         if lookup_ok is False:
             is_numeric_idx = \
                 (is_bstr and commit_id_or_idx.isdigit() and len(commit_id_or_idx) < 12) \
                 or isinstance(commit_id_or_idx, int)
             if not is_branch and (is_numeric_idx or is_null(commit_id_or_idx)):
                 try:
                     commit_id_or_idx = self.commit_ids[int(commit_id_or_idx)]
                     lookup_ok = True
                 except Exception:
                     raise CommitDoesNotExistError(commit_missing_err)
 
         # we failed regular lookup, and by integer number lookup
         if lookup_ok is False:
             raise CommitDoesNotExistError(commit_missing_err)
 
         # 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.
         """
         try:
             return self.get_commit().date
         except RepositoryError:
             tzoffset = makedate()[1]
             return utcdate_fromtimestamp(self._get_fs_mtime(), tzoffset)
 
     def _get_fs_mtime(self):
         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_entries(self, prefix='', reverse=False, strip_prefix=True):
         if self.is_empty():
             return OrderedDict()
 
         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))
 
         def get_name(entry):
             return entry[0]
 
         return OrderedDict(sorted(result, key=get_name, reverse=reverse))
 
     def _get_branches(self):
         return self._get_refs_entries(prefix='refs/heads/', strip_prefix=True)
 
     @CachedProperty
     def branches(self):
         return self._get_branches()
 
     @CachedProperty
     def branches_closed(self):
         return {}
 
     @CachedProperty
     def bookmarks(self):
         return {}
 
     @CachedProperty
     def branches_all(self):
         all_branches = {}
         all_branches.update(self.branches)
         all_branches.update(self.branches_closed)
         return all_branches
 
     @CachedProperty
     def tags(self):
         return self._get_tags()
 
     def _get_tags(self):
         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.raw_id)
 
         self._invalidate_prop_cache('tags')
         self._invalidate_prop_cache('_refs')
 
         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)
 
         self._remote.tag_remove(name)
         self._invalidate_prop_cache('tags')
         self._invalidate_prop_cache('_refs')
 
     def _get_refs(self):
         return self._remote.get_refs()
 
     @CachedProperty
     def _refs(self):
         return self._get_refs()
 
     @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_remote_ref(self, ref_name):
         ref_key = 'refs/remotes/origin/{}'.format(safe_str(ref_name))
         try:
             return self._refs[ref_key]
         except Exception:
             return
 
     def get_commit(self, commit_id=None, commit_idx=None, pre_load=None,
                    translate_tag=True, maybe_unreachable=False, reference_obj=None):
         """
         Returns `GitCommit` object representing commit from git repository
         at the given `commit_id` or head (most recent commit) if None given.
         """
 
         if self.is_empty():
             raise EmptyRepositoryError("There are no commits yet")
 
         if commit_id is not None:
             self._validate_commit_id(commit_id)
             try:
                 # we have cached idx, use it without contacting the remote
                 idx = self._commit_ids[commit_id]
                 return GitCommit(self, commit_id, idx, pre_load=pre_load)
             except KeyError:
                 pass
 
         elif commit_idx is not None:
             self._validate_commit_idx(commit_idx)
             try:
                 _commit_id = self.commit_ids[commit_idx]
                 if commit_idx < 0:
                     commit_idx = self.commit_ids.index(_commit_id)
                 return GitCommit(self, _commit_id, commit_idx, pre_load=pre_load)
             except IndexError:
                 commit_id = commit_idx
         else:
             commit_id = "tip"
 
         if translate_tag:
             commit_id = self._lookup_commit(
                 commit_id, maybe_unreachable=maybe_unreachable,
                 reference_obj=reference_obj)
 
         try:
             idx = self._commit_ids[commit_id]
         except KeyError:
             idx = -1
 
         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, show_hidden=False, pre_load=None, translate_tags=True):
         """
         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
         :param show_hidden: Show hidden commits such as obsolete or hidden from
             Mercurial evolve
         :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._lookup_commit(start_id)
         start_pos = self._commit_ids[start_raw_id] if start_id else None
         end_raw_id = self._lookup_commit(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_commit_ids(filters=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,
                                    translate_tag=translate_tags)
 
     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.")
 
         if path:
             file_filter = path
         else:
             file_filter = None
 
         diff = self._remote.diff(
             commit1.raw_id, commit2.raw_id, file_filter=file_filter,
             opt_ignorews=ignore_whitespace,
             context=context)
         return GitDiff(diff)
 
     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)
 
         # clear cached properties
         self._invalidate_prop_cache('commit_ids')
         self._invalidate_prop_cache('_refs')
         self._invalidate_prop_cache('branches')
 
         return len(self.commit_ids)
 
     def get_common_ancestor(self, commit_id1, commit_id2, repo2):
         log.debug('Calculating common ancestor between %sc1:%s and %sc2:%s',
                   self, commit_id1, repo2, commit_id2)
 
         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 = self.COMMIT_ID_PAT.findall(output)[0]
 
         log.debug('Found common ancestor with sha: %s', ancestor_id)
 
         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 self.COMMIT_ID_PAT.findall(output)]
 
         return commits
 
     @LazyProperty
     def in_memory_commit(self):
         """
         Returns ``GitInMemoryCommit`` object for this repository.
         """
         return GitInMemoryCommit(self)
 
     def pull(self, url, commit_ids=None, update_after=False):
         """
         Pull changes from external location. Pull is different in GIT
         that fetch since it's doing a checkout
 
         :param commit_ids: Optional. Can be set to a list of commit ids
            which shall be pulled from the other repository.
         """
         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.pull(url, refs=refs, update_after=update_after)
         self._remote.invalidate_vcs_cache()
 
     def fetch(self, url, commit_ids=None):
         """
         Fetch all git objects from external location.
         """
         self._remote.sync_fetch(url, refs=commit_ids)
         self._remote.invalidate_vcs_cache()
 
     def push(self, url):
         refs = None
         self._remote.sync_push(url, refs=refs)
 
     def set_refs(self, ref_name, commit_id):
         self._remote.set_refs(ref_name, commit_id)
         self._invalidate_prop_cache('_refs')
 
     def remove_ref(self, ref_name):
         self._remote.remove_ref(ref_name)
         self._invalidate_prop_cache('_refs')
 
     def run_gc(self, prune=True):
         cmd = ['gc', '--aggressive']
         if prune:
             cmd += ['--prune=now']
         _stdout, stderr = self.run_git_command(cmd, fail_on_stderr=False)
         return stderr
 
     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, force=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 force:
             cmd.append('-f')
         if create:
             cmd.append('-b')
         cmd.append(branch_name)
         self.run_git_command(cmd, fail_on_stderr=False)
 
     def _create_branch(self, branch_name, commit_id):
         """
         creates a branch in a GIT repo
         """
         self._remote.create_branch(branch_name, commit_id)
 
     def _identify(self):
         """
         Return the current state of the working directory.
         """
         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', 'HEAD'])
         return stdout.strip()
 
     def _local_clone(self, clone_path, branch_name, source_branch=None):
         """
         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,
                self.path, os.path.abspath(clone_path)]
 
         self.run_git_command(cmd, fail_on_stderr=False)
 
         # if we get the different source branch, make sure we also fetch it for
         # merge conditions
         if source_branch and source_branch != branch_name:
             # check if the ref exists.
             shadow_repo = GitRepository(os.path.abspath(clone_path))
             if shadow_repo.get_remote_ref(source_branch):
                 cmd = ['fetch', self.path, source_branch]
                 self.run_git_command(cmd, fail_on_stderr=False)
 
     def _local_fetch(self, repository_path, branch_name, use_origin=False):
         """
         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')
 
         if use_origin:
             branch_name = '+{branch}:refs/heads/{branch}'.format(
                 branch=branch_name)
 
         cmd = ['fetch', '--no-tags', '--update-head-ok',
                repository_path, branch_name]
         self.run_git_command(cmd, fail_on_stderr=False)
 
     def _local_reset(self, branch_name):
         branch_name = '{}'.format(branch_name)
         cmd = ['reset', '--hard', 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 get_shadow_instance(self, shadow_repository_path, enable_hooks=False, cache=False):
         return GitRepository(shadow_repository_path, with_wire={"cache": cache})
 
     def _local_pull(self, repository_path, branch_name, ff_only=True):
         """
         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']
         if ff_only:
             cmd.append('--ff-only')
         cmd.extend(['--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 something more robust in this case.
             raise RepositoryError('Do not know how to merge into empty repositories yet')
         unresolved = None
 
         # 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)]
 
         merge_cmd = cmd + heads
 
         try:
             self.run_git_command(merge_cmd, fail_on_stderr=False)
         except RepositoryError:
             files = self.run_git_command(['diff', '--name-only', '--diff-filter', 'U'],
                                          fail_on_stderr=False)[0].splitlines()
             # NOTE(marcink): we add U notation for consistent with HG backend output
             unresolved = ['U {}'.format(f) for f in files]
 
             # Cleanup any merge leftovers
             self._remote.invalidate_vcs_cache()
             self.run_git_command(['merge', '--abort'], fail_on_stderr=False)
 
             if unresolved:
                 raise UnresolvedFilesInRepo(unresolved)
             else:
                 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 _maybe_prepare_merge_workspace(
             self, repo_id, workspace_id, target_ref, source_ref):
         shadow_repository_path = self._get_shadow_repository_path(
             self.path, repo_id, workspace_id)
         if not os.path.exists(shadow_repository_path):
             self._local_clone(
                 shadow_repository_path, target_ref.name, source_ref.name)
             log.debug('Prepared %s shadow repository in %s',
                       self.alias, shadow_repository_path)
 
         return shadow_repository_path
 
     def _merge_repo(self, repo_id, workspace_id, target_ref,
                     source_repo, source_ref, merge_message,
                     merger_name, merger_email, dry_run=False,
                     use_rebase=False, close_branch=False):
 
         log.debug('Executing merge_repo with %s strategy, dry_run mode:%s',
                   'rebase' if use_rebase else 'merge', dry_run)
         if target_ref.commit_id != self.branches[target_ref.name]:
             log.warning('Target ref %s commit mismatch %s vs %s', target_ref,
                         target_ref.commit_id, self.branches[target_ref.name])
             return MergeResponse(
                 False, False, None, MergeFailureReason.TARGET_IS_NOT_HEAD,
                 metadata={'target_ref': target_ref})
 
         shadow_repository_path = self._maybe_prepare_merge_workspace(
             repo_id, workspace_id, target_ref, source_ref)
         shadow_repo = self.get_shadow_instance(shadow_repository_path)
 
         # checkout source, if it's different. Otherwise we could not
         # fetch proper commits for merge testing
         if source_ref.name != target_ref.name:
             if shadow_repo.get_remote_ref(source_ref.name):
                 shadow_repo._checkout(source_ref.name, force=True)
 
         # checkout target, and fetch changes
         shadow_repo._checkout(target_ref.name, force=True)
 
         # fetch/reset pull the target, in case it is changed
         # this handles even force changes
         shadow_repo._local_fetch(self.path, target_ref.name, use_origin=True)
         shadow_repo._local_reset(target_ref.name)
 
         # Need to reload repo to invalidate the cache, or otherwise we cannot
         # retrieve the last target commit.
         shadow_repo = self.get_shadow_instance(shadow_repository_path)
         if target_ref.commit_id != shadow_repo.branches[target_ref.name]:
             log.warning('Shadow Target ref %s commit mismatch %s vs %s',
                         target_ref, target_ref.commit_id,
                         shadow_repo.branches[target_ref.name])
             return MergeResponse(
                 False, False, None, MergeFailureReason.TARGET_IS_NOT_HEAD,
                 metadata={'target_ref': target_ref})
 
         # calculate new branch
         pr_branch = shadow_repo._get_new_pr_branch(
             source_ref.name, target_ref.name)
         log.debug('using pull-request merge branch: `%s`', pr_branch)
         # checkout to temp branch, and fetch changes
         shadow_repo._checkout(pr_branch, create=True)
         try:
             shadow_repo._local_fetch(source_repo.path, source_ref.name)
         except RepositoryError:
             log.exception('Failure when doing local fetch on '
                           'shadow repo: %s', shadow_repo)
             return MergeResponse(
                 False, False, None, MergeFailureReason.MISSING_SOURCE_REF,
                 metadata={'source_ref': source_ref})
 
         merge_ref = None
         merge_failure_reason = MergeFailureReason.NONE
         metadata = {}
         try:
             shadow_repo._local_merge(merge_message, merger_name, merger_email,
                                      [source_ref.commit_id])
             merge_possible = True
 
             # Need to invalidate the cache, or otherwise we
             # cannot retrieve the merge commit.
             shadow_repo = shadow_repo.get_shadow_instance(shadow_repository_path)
             merge_commit_id = shadow_repo.branches[pr_branch]
 
             # Set a reference pointing to the merge commit. This reference may
             # be used to easily identify the last successful merge commit in
             # the shadow repository.
             shadow_repo.set_refs('refs/heads/pr-merge', merge_commit_id)
             merge_ref = Reference('branch', 'pr-merge', merge_commit_id)
         except RepositoryError as e:
             log.exception('Failure when doing local merge on git shadow repo')
             if isinstance(e, UnresolvedFilesInRepo):
                 metadata['unresolved_files'] = '\n* conflict: ' + ('\n * conflict: '.join(e.args[0]))
 
             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
             except RepositoryError:
                 log.exception(
                     'Failure when doing local push from the shadow '
                     'repository to the target repository at %s.', self.path)
                 merge_succeeded = False
                 merge_failure_reason = MergeFailureReason.PUSH_FAILED
                 metadata['target'] = 'git shadow repo'
                 metadata['merge_commit'] = pr_branch
         else:
             merge_succeeded = False
 
         return MergeResponse(
             merge_possible, merge_succeeded, merge_ref, merge_failure_reason,
             metadata=metadata)

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

@@ -1,1012 +1,1012 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/
 
 """
 HG repository module
 """
 import os
 import logging
 import binascii
 import configparser
 import urllib.request, urllib.parse, urllib.error
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
-from rhodecode.lib.compat import OrderedDict
+from collections import OrderedDict
 from rhodecode.lib.datelib import (
     date_to_timestamp_plus_offset, utcdate_fromtimestamp, makedate)
 from rhodecode.lib.utils import safe_unicode, safe_str
 from rhodecode.lib.utils2 import CachedProperty
 from rhodecode.lib.vcs import connection, exceptions
 from rhodecode.lib.vcs.backends.base import (
     BaseRepository, CollectionGenerator, Config, MergeResponse,
     MergeFailureReason, Reference, BasePathPermissionChecker)
 from rhodecode.lib.vcs.backends.hg.commit import MercurialCommit
 from rhodecode.lib.vcs.backends.hg.diff import MercurialDiff
 from rhodecode.lib.vcs.backends.hg.inmemory import MercurialInMemoryCommit
 from rhodecode.lib.vcs.exceptions import (
     EmptyRepositoryError, RepositoryError, TagAlreadyExistError,
     TagDoesNotExistError, CommitDoesNotExistError, SubrepoMergeError, UnresolvedFilesInRepo)
 
 hexlify = binascii.hexlify
 nullid = "\0" * 20
 
 log = logging.getLogger(__name__)
 
 
 class MercurialRepository(BaseRepository):
     """
     Mercurial repository backend
     """
     DEFAULT_BRANCH_NAME = 'default'
 
     def __init__(self, repo_path, config=None, create=False, src_url=None,
                  do_workspace_checkout=False, with_wire=None, bare=False):
         """
         Raises RepositoryError if repository could not be find at the given
         ``repo_path``.
 
         :param repo_path: local path of the repository
         :param config: config object containing the repo configuration
         :param create=False: if set to True, would try to create repository if
            it does not exist rather than raising exception
         :param src_url=None: would try to clone repository from given location
         :param do_workspace_checkout=False: sets update of working copy after
            making a clone
         :param bare: not used, compatible with other VCS
         """
 
         self.path = safe_str(os.path.abspath(repo_path))
         # mercurial since 4.4.X requires certain configuration to be present
         # because sometimes we init the repos with config we need to meet
         # special requirements
         self.config = config if config else self.get_default_config(
             default=[('extensions', 'largefiles', '1')])
         self.with_wire = with_wire or {"cache": False}  # default should not use cache
 
         self._init_repo(create, src_url, do_workspace_checkout)
 
         # caches
         self._commit_ids = {}
 
     @LazyProperty
     def _remote(self):
         repo_id = self.path
         return connection.Hg(self.path, repo_id, self.config, with_wire=self.with_wire)
 
     @CachedProperty
     def commit_ids(self):
         """
         Returns list of commit ids, in ascending order.  Being lazy
         attribute allows external tools to inject shas 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))
 
     @CachedProperty
     def branches(self):
         return self._get_branches()
 
     @CachedProperty
     def branches_closed(self):
         return self._get_branches(active=False, closed=True)
 
     @CachedProperty
     def branches_all(self):
         all_branches = {}
         all_branches.update(self.branches)
         all_branches.update(self.branches_closed)
         return all_branches
 
     def _get_branches(self, active=True, closed=False):
         """
         Gets branches for this repository
         Returns only not closed active branches by default
 
         :param active: return also active branches
         :param closed: return also closed branches
 
         """
         if self.is_empty():
             return {}
 
         def get_name(ctx):
             return ctx[0]
 
         _branches = [(safe_unicode(n), hexlify(h),) for n, h in
                      self._remote.branches(active, closed).items()]
 
         return OrderedDict(sorted(_branches, key=get_name, reverse=False))
 
     @CachedProperty
     def tags(self):
         """
         Gets tags for this repository
         """
         return self._get_tags()
 
     def _get_tags(self):
         if self.is_empty():
             return {}
 
         def get_name(ctx):
             return ctx[0]
 
         _tags = [(safe_unicode(n), hexlify(h),) for n, h in
                  self._remote.tags().items()]
 
         return OrderedDict(sorted(_tags, key=get_name, reverse=True))
 
     def tag(self, name, user, commit_id=None, message=None, date=None, **kwargs):
         """
         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)
         local = kwargs.setdefault('local', False)
 
         if message is None:
             message = "Added tag %s for commit %s" % (name, commit.short_id)
 
         date, tz = date_to_timestamp_plus_offset(date)
 
         self._remote.tag(name, commit.raw_id, message, local, user, date, tz)
         self._remote.invalidate_vcs_cache()
 
         # Reinitialize tags
         self._invalidate_prop_cache('tags')
         tag_id = self.tags[name]
 
         return self.get_commit(commit_id=tag_id)
 
     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)
 
         if message is None:
             message = "Removed tag %s" % name
         local = False
 
         date, tz = date_to_timestamp_plus_offset(date)
 
         self._remote.tag(name, nullid, message, local, user, date, tz)
         self._remote.invalidate_vcs_cache()
         self._invalidate_prop_cache('tags')
 
     @LazyProperty
     def bookmarks(self):
         """
         Gets bookmarks for this repository
         """
         return self._get_bookmarks()
 
     def _get_bookmarks(self):
         if self.is_empty():
             return {}
 
         def get_name(ctx):
             return ctx[0]
 
         _bookmarks = [
             (safe_unicode(n), hexlify(h)) for n, h in
             self._remote.bookmarks().items()]
 
         return OrderedDict(sorted(_bookmarks, key=get_name))
 
     def _get_all_commit_ids(self):
         return self._remote.get_all_commit_ids('visible')
 
     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 commit 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.")
 
         if path:
             file_filter = [self.path, path]
         else:
             file_filter = None
 
         diff = self._remote.diff(
             commit1.raw_id, commit2.raw_id, file_filter=file_filter,
             opt_git=True, opt_ignorews=ignore_whitespace,
             context=context)
         return MercurialDiff(diff)
 
     def strip(self, commit_id, branch=None):
         self._remote.strip(commit_id, update=False, backup="none")
 
         self._remote.invalidate_vcs_cache()
         # clear cache
         self._invalidate_prop_cache('commit_ids')
 
         return len(self.commit_ids)
 
     def verify(self):
         verify = self._remote.verify()
 
         self._remote.invalidate_vcs_cache()
         return verify
 
     def hg_update_cache(self):
         update_cache = self._remote.hg_update_cache()
 
         self._remote.invalidate_vcs_cache()
         return update_cache
 
     def hg_rebuild_fn_cache(self):
         update_cache = self._remote.hg_rebuild_fn_cache()
 
         self._remote.invalidate_vcs_cache()
         return update_cache
 
     def get_common_ancestor(self, commit_id1, commit_id2, repo2):
         log.debug('Calculating common ancestor between %sc1:%s and %sc2:%s',
                   self, commit_id1, repo2, commit_id2)
 
         if commit_id1 == commit_id2:
             return commit_id1
 
         ancestors = self._remote.revs_from_revspec(
             "ancestor(id(%s), id(%s))", commit_id1, commit_id2,
             other_path=repo2.path)
 
         ancestor_id = repo2[ancestors[0]].raw_id if ancestors else None
 
         log.debug('Found common ancestor with sha: %s', ancestor_id)
         return ancestor_id
 
     def compare(self, commit_id1, commit_id2, repo2, merge, pre_load=None):
         if commit_id1 == commit_id2:
             commits = []
         else:
             if merge:
                 indexes = self._remote.revs_from_revspec(
                     "ancestors(id(%s)) - ancestors(id(%s)) - id(%s)",
                     commit_id2, commit_id1, commit_id1, other_path=repo2.path)
             else:
                 indexes = self._remote.revs_from_revspec(
                     "id(%s)..id(%s) - id(%s)", commit_id1, commit_id2,
                     commit_id1, other_path=repo2.path)
 
             commits = [repo2.get_commit(commit_idx=idx, pre_load=pre_load)
                        for idx in indexes]
 
         return commits
 
     @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 mercurial 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 local url
         if os.path.isdir(url) or url.startswith('file:'):
             return True
 
         # Request the _remote to verify the url
         return connection.Hg.check_url(url, config.serialize())
 
     @staticmethod
     def is_valid_repository(path):
         return os.path.isdir(os.path.join(path, '.hg'))
 
     def _init_repo(self, create, src_url=None, do_workspace_checkout=False):
         """
         Function will check for mercurial repository in given path. If there
         is no repository in that path it will raise an exception unless
         `create` parameter is set to True - in that case repository would
         be created.
 
         If `src_url` is given, would try to clone repository from the
         location at given clone_point. Additionally it'll make update to
         working copy accordingly to `do_workspace_checkout` flag.
         """
         if create and os.path.exists(self.path):
             raise RepositoryError(
                 "Cannot create repository at %s, location already exist"
                 % self.path)
 
         if src_url:
             url = str(self._get_url(src_url))
             MercurialRepository.check_url(url, self.config)
 
             self._remote.clone(url, self.path, do_workspace_checkout)
 
             # Don't try to create if we've already cloned repo
             create = False
 
         if create:
             os.makedirs(self.path, mode=0o755)
         self._remote.localrepository(create)
 
     @LazyProperty
     def in_memory_commit(self):
         return MercurialInMemoryCommit(self)
 
     @LazyProperty
     def description(self):
         description = self._remote.get_config_value(
             'web', 'description', untrusted=True)
         return safe_unicode(description or self.DEFAULT_DESCRIPTION)
 
     @LazyProperty
     def contact(self):
         contact = (
             self._remote.get_config_value("web", "contact") or
             self._remote.get_config_value("ui", "username"))
         return safe_unicode(contact or self.DEFAULT_CONTACT)
 
     @LazyProperty
     def last_change(self):
         """
         Returns last change made on this repository as
         `datetime.datetime` object.
         """
         try:
             return self.get_commit().date
         except RepositoryError:
             tzoffset = makedate()[1]
             return utcdate_fromtimestamp(self._get_fs_mtime(), tzoffset)
 
     def _get_fs_mtime(self):
         # fallback to filesystem
         cl_path = os.path.join(self.path, '.hg', "00changelog.i")
         st_path = os.path.join(self.path, '.hg', "store")
         if os.path.exists(cl_path):
             return os.stat(cl_path).st_mtime
         else:
             return os.stat(st_path).st_mtime
 
     def _get_url(self, url):
         """
         Returns normalized url. If schema is not given, would fall
         to filesystem
         (``file:///``) schema.
         """
         url = url.encode('utf8')
         if url != 'default' and '://' not in url:
             url = "file:" + urllib.request.pathname2url(url)
         return url
 
     def get_hook_location(self):
         """
         returns absolute path to location where hooks are stored
         """
         return os.path.join(self.path, '.hg', '.hgrc')
 
     def get_commit(self, commit_id=None, commit_idx=None, pre_load=None,
                    translate_tag=None, maybe_unreachable=False, reference_obj=None):
         """
         Returns ``MercurialCommit`` object representing repository's
         commit at the given `commit_id` or `commit_idx`.
         """
         if self.is_empty():
             raise EmptyRepositoryError("There are no commits yet")
 
         if commit_id is not None:
             self._validate_commit_id(commit_id)
             try:
                 # we have cached idx, use it without contacting the remote
                 idx = self._commit_ids[commit_id]
                 return MercurialCommit(self, commit_id, idx, pre_load=pre_load)
             except KeyError:
                 pass
 
         elif commit_idx is not None:
             self._validate_commit_idx(commit_idx)
             try:
                 _commit_id = self.commit_ids[commit_idx]
                 if commit_idx < 0:
                     commit_idx = self.commit_ids.index(_commit_id)
 
                 return MercurialCommit(self, _commit_id, commit_idx, pre_load=pre_load)
             except IndexError:
                 commit_id = commit_idx
         else:
             commit_id = "tip"
 
         if isinstance(commit_id, unicode):
             commit_id = safe_str(commit_id)
 
         try:
             raw_id, idx = self._remote.lookup(commit_id, both=True)
         except CommitDoesNotExistError:
             msg = "Commit {} does not exist for `{}`".format(
                 *map(safe_str, [commit_id, self.name]))
             raise CommitDoesNotExistError(msg)
 
         return MercurialCommit(self, raw_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, show_hidden=False, pre_load=None, translate_tags=None):
         """
         Returns generator of ``MercurialCommit`` objects from start to end
         (both are inclusive)
 
         :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
         :param show_hidden: Show hidden commits such as obsolete or hidden from
             Mercurial evolve
         :raise BranchDoesNotExistError: If given ``branch_name`` does not
             exist.
         :raise CommitDoesNotExistError: If commit for given ``start`` or
           ``end`` could not be found.
         """
         # actually we should check now if it's not an empty repo
         if self.is_empty():
             raise EmptyRepositoryError("There are no commits yet")
         self._validate_branch_name(branch_name)
 
         branch_ancestors = False
         if start_id is not None:
             self._validate_commit_id(start_id)
             c_start = self.get_commit(commit_id=start_id)
             start_pos = self._commit_ids[c_start.raw_id]
         else:
             start_pos = None
 
         if end_id is not None:
             self._validate_commit_id(end_id)
             c_end = self.get_commit(commit_id=end_id)
             end_pos = max(0, self._commit_ids[c_end.raw_id])
         else:
             end_pos = 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
 
         commit_filter = []
 
         if branch_name and not branch_ancestors:
             commit_filter.append('branch("%s")' % (branch_name,))
         elif branch_name and branch_ancestors:
             commit_filter.append('ancestors(branch("%s"))' % (branch_name,))
 
         if start_date and not end_date:
             commit_filter.append('date(">%s")' % (start_date,))
         if end_date and not start_date:
             commit_filter.append('date("<%s")' % (end_date,))
         if start_date and end_date:
             commit_filter.append(
                 'date(">%s") and date("<%s")' % (start_date, end_date))
 
         if not show_hidden:
             commit_filter.append('not obsolete()')
             commit_filter.append('not hidden()')
 
         # TODO: johbo: Figure out a simpler way for this solution
         collection_generator = CollectionGenerator
         if commit_filter:
             commit_filter = ' and '.join(map(safe_str, commit_filter))
             revisions = self._remote.rev_range([commit_filter])
             collection_generator = MercurialIndexBasedCollectionGenerator
         else:
             revisions = self.commit_ids
 
         if start_pos or end_pos:
             revisions = revisions[start_pos:end_pos]
 
         return collection_generator(self, revisions, pre_load=pre_load)
 
     def pull(self, url, commit_ids=None):
         """
         Pull changes from external location.
 
         :param commit_ids: Optional. Can be set to a list of commit ids
            which shall be pulled from the other repository.
         """
         url = self._get_url(url)
         self._remote.pull(url, commit_ids=commit_ids)
         self._remote.invalidate_vcs_cache()
 
     def fetch(self, url, commit_ids=None):
         """
         Backward compatibility with GIT fetch==pull
         """
         return self.pull(url, commit_ids=commit_ids)
 
     def push(self, url):
         url = self._get_url(url)
         self._remote.sync_push(url)
 
     def _local_clone(self, clone_path):
         """
         Create a local clone of the current repo.
         """
         self._remote.clone(self.path, clone_path, update_after_clone=True,
                            hooks=False)
 
     def _update(self, revision, clean=False):
         """
         Update the working copy to the specified revision.
         """
         log.debug('Doing checkout to commit: `%s` for %s', revision, self)
         self._remote.update(revision, clean=clean)
 
     def _identify(self):
         """
         Return the current state of the working directory.
         """
         return self._remote.identify().strip().rstrip('+')
 
     def _heads(self, branch=None):
         """
         Return the commit ids of the repository heads.
         """
         return self._remote.heads(branch=branch).strip().split(' ')
 
     def _ancestor(self, revision1, revision2):
         """
         Return the common ancestor of the two revisions.
         """
         return self._remote.ancestor(revision1, revision2)
 
     def _local_push(
             self, revision, repository_path, push_branches=False,
             enable_hooks=False):
         """
         Push the given revision to the specified repository.
 
         :param push_branches: allow to create branches in the target repo.
         """
         self._remote.push(
             [revision], repository_path, hooks=enable_hooks,
             push_branches=push_branches)
 
     def _local_merge(self, target_ref, merge_message, user_name, user_email,
                      source_ref, use_rebase=False, close_commit_id=None, dry_run=False):
         """
         Merge the given source_revision into the checked out revision.
 
         Returns the commit id of the merge and a boolean indicating if the
         commit needs to be pushed.
         """
         source_ref_commit_id = source_ref.commit_id
         target_ref_commit_id = target_ref.commit_id
 
         # update our workdir to target ref, for proper merge
         self._update(target_ref_commit_id, clean=True)
 
         ancestor = self._ancestor(target_ref_commit_id, source_ref_commit_id)
         is_the_same_branch = self._is_the_same_branch(target_ref, source_ref)
 
         if close_commit_id:
             # NOTE(marcink): if we get the close commit, this is our new source
             # which will include the close commit itself.
             source_ref_commit_id = close_commit_id
 
         if ancestor == source_ref_commit_id:
             # Nothing to do, the changes were already integrated
             return target_ref_commit_id, False
 
         elif ancestor == target_ref_commit_id and is_the_same_branch:
             # In this case we should force a commit message
             return source_ref_commit_id, True
 
         unresolved = None
         if use_rebase:
             try:
                 bookmark_name = 'rcbook%s%s' % (source_ref_commit_id, target_ref_commit_id)
                 self.bookmark(bookmark_name, revision=source_ref.commit_id)
                 self._remote.rebase(
                     source=source_ref_commit_id, dest=target_ref_commit_id)
                 self._remote.invalidate_vcs_cache()
                 self._update(bookmark_name, clean=True)
                 return self._identify(), True
             except RepositoryError as e:
                 # The rebase-abort may raise another exception which 'hides'
                 # the original one, therefore we log it here.
                 log.exception('Error while rebasing shadow repo during merge.')
                 if 'unresolved conflicts' in safe_str(e):
                     unresolved = self._remote.get_unresolved_files()
                     log.debug('unresolved files: %s', unresolved)
 
                 # Cleanup any rebase leftovers
                 self._remote.invalidate_vcs_cache()
                 self._remote.rebase(abort=True)
                 self._remote.invalidate_vcs_cache()
                 self._remote.update(clean=True)
                 if unresolved:
                     raise UnresolvedFilesInRepo(unresolved)
                 else:
                     raise
         else:
             try:
                 self._remote.merge(source_ref_commit_id)
                 self._remote.invalidate_vcs_cache()
                 self._remote.commit(
                     message=safe_str(merge_message),
                     username=safe_str('%s <%s>' % (user_name, user_email)))
                 self._remote.invalidate_vcs_cache()
                 return self._identify(), True
             except RepositoryError as e:
                 # The merge-abort may raise another exception which 'hides'
                 # the original one, therefore we log it here.
                 log.exception('Error while merging shadow repo during merge.')
                 if 'unresolved merge conflicts' in safe_str(e):
                     unresolved = self._remote.get_unresolved_files()
                     log.debug('unresolved files: %s', unresolved)
 
                 # Cleanup any merge leftovers
                 self._remote.update(clean=True)
                 if unresolved:
                     raise UnresolvedFilesInRepo(unresolved)
                 else:
                     raise
 
     def _local_close(self, target_ref, user_name, user_email,
                      source_ref, close_message=''):
         """
         Close the branch of the given source_revision
 
         Returns the commit id of the close and a boolean indicating if the
         commit needs to be pushed.
         """
         self._update(source_ref.commit_id)
         message = close_message or "Closing branch: `{}`".format(source_ref.name)
         try:
             self._remote.commit(
                 message=safe_str(message),
                 username=safe_str('%s <%s>' % (user_name, user_email)),
                 close_branch=True)
             self._remote.invalidate_vcs_cache()
             return self._identify(), True
         except RepositoryError:
             # Cleanup any commit leftovers
             self._remote.update(clean=True)
             raise
 
     def _is_the_same_branch(self, target_ref, source_ref):
         return (
             self._get_branch_name(target_ref) ==
             self._get_branch_name(source_ref))
 
     def _get_branch_name(self, ref):
         if ref.type == 'branch':
             return ref.name
         return self._remote.ctx_branch(ref.commit_id)
 
     def _maybe_prepare_merge_workspace(
             self, repo_id, workspace_id, unused_target_ref, unused_source_ref):
         shadow_repository_path = self._get_shadow_repository_path(
             self.path, repo_id, workspace_id)
         if not os.path.exists(shadow_repository_path):
             self._local_clone(shadow_repository_path)
             log.debug(
                 'Prepared shadow repository in %s', shadow_repository_path)
 
         return shadow_repository_path
 
     def _merge_repo(self, repo_id, workspace_id, target_ref,
                     source_repo, source_ref, merge_message,
                     merger_name, merger_email, dry_run=False,
                     use_rebase=False, close_branch=False):
 
         log.debug('Executing merge_repo with %s strategy, dry_run mode:%s',
                   'rebase' if use_rebase else 'merge', dry_run)
         if target_ref.commit_id not in self._heads():
             return MergeResponse(
                 False, False, None, MergeFailureReason.TARGET_IS_NOT_HEAD,
                 metadata={'target_ref': target_ref})
 
         try:
             if target_ref.type == 'branch' and len(self._heads(target_ref.name)) != 1:
                 heads_all = self._heads(target_ref.name)
                 max_heads = 10
                 if len(heads_all) > max_heads:
                     heads = '\n,'.join(
                         heads_all[:max_heads] +
                         ['and {} more.'.format(len(heads_all)-max_heads)])
                 else:
                     heads = '\n,'.join(heads_all)
                 metadata = {
                     'target_ref': target_ref,
                     'source_ref': source_ref,
                     'heads': heads
                 }
                 return MergeResponse(
                     False, False, None,
                     MergeFailureReason.HG_TARGET_HAS_MULTIPLE_HEADS,
                     metadata=metadata)
         except CommitDoesNotExistError:
             log.exception('Failure when looking up branch heads on hg target')
             return MergeResponse(
                 False, False, None, MergeFailureReason.MISSING_TARGET_REF,
                 metadata={'target_ref': target_ref})
 
         shadow_repository_path = self._maybe_prepare_merge_workspace(
             repo_id, workspace_id, target_ref, source_ref)
         shadow_repo = self.get_shadow_instance(shadow_repository_path)
 
         log.debug('Pulling in target reference %s', target_ref)
         self._validate_pull_reference(target_ref)
         shadow_repo._local_pull(self.path, target_ref)
 
         try:
             log.debug('Pulling in source reference %s', source_ref)
             source_repo._validate_pull_reference(source_ref)
             shadow_repo._local_pull(source_repo.path, source_ref)
         except CommitDoesNotExistError:
             log.exception('Failure when doing local pull on hg shadow repo')
             return MergeResponse(
                 False, False, None, MergeFailureReason.MISSING_SOURCE_REF,
                 metadata={'source_ref': source_ref})
 
         merge_ref = None
         merge_commit_id = None
         close_commit_id = None
         merge_failure_reason = MergeFailureReason.NONE
         metadata = {}
 
         # enforce that close branch should be used only in case we source from
         # an actual Branch
         close_branch = close_branch and source_ref.type == 'branch'
 
         # don't allow to close branch if source and target are the same
         close_branch = close_branch and source_ref.name != target_ref.name
 
         needs_push_on_close = False
         if close_branch and not use_rebase and not dry_run:
             try:
                 close_commit_id, needs_push_on_close = shadow_repo._local_close(
                     target_ref, merger_name, merger_email, source_ref)
                 merge_possible = True
             except RepositoryError:
                 log.exception('Failure when doing close branch on '
                               'shadow repo: %s', shadow_repo)
                 merge_possible = False
                 merge_failure_reason = MergeFailureReason.MERGE_FAILED
         else:
             merge_possible = True
 
         needs_push = False
         if merge_possible:
 
             try:
                 merge_commit_id, needs_push = shadow_repo._local_merge(
                     target_ref, merge_message, merger_name, merger_email,
                     source_ref, use_rebase=use_rebase,
                     close_commit_id=close_commit_id, dry_run=dry_run)
                 merge_possible = True
 
                 # read the state of the close action, if it
                 # maybe required a push
                 needs_push = needs_push or needs_push_on_close
 
                 # Set a bookmark pointing to the merge commit. This bookmark
                 # may be used to easily identify the last successful merge
                 # commit in the shadow repository.
                 shadow_repo.bookmark('pr-merge', revision=merge_commit_id)
                 merge_ref = Reference('book', 'pr-merge', merge_commit_id)
             except SubrepoMergeError:
                 log.exception(
                     'Subrepo merge error during local merge on hg shadow repo.')
                 merge_possible = False
                 merge_failure_reason = MergeFailureReason.SUBREPO_MERGE_FAILED
                 needs_push = False
             except RepositoryError as e:
                 log.exception('Failure when doing local merge on hg shadow repo')
                 if isinstance(e, UnresolvedFilesInRepo):
                     all_conflicts = list(e.args[0])
                     max_conflicts = 20
                     if len(all_conflicts) > max_conflicts:
                         conflicts = all_conflicts[:max_conflicts] \
                                     + ['and {} more.'.format(len(all_conflicts)-max_conflicts)]
                     else:
                         conflicts = all_conflicts
                     metadata['unresolved_files'] = \
                         '\n* conflict: ' + \
                         ('\n * conflict: '.join(conflicts))
 
                 merge_possible = False
                 merge_failure_reason = MergeFailureReason.MERGE_FAILED
                 needs_push = False
 
         if merge_possible and not dry_run:
             if needs_push:
                 # In case the target is a bookmark, update it, so after pushing
                 # the bookmarks is also updated in the target.
                 if target_ref.type == 'book':
                     shadow_repo.bookmark(
                         target_ref.name, revision=merge_commit_id)
                 try:
                     shadow_repo_with_hooks = self.get_shadow_instance(
                         shadow_repository_path,
                         enable_hooks=True)
                     # This is the actual merge action, we push from shadow
                     # into origin.
                     # Note: the push_branches option will push any new branch
                     # defined in the source repository to the target. This may
                     # be dangerous as branches are permanent in Mercurial.
                     # This feature was requested in issue #441.
                     shadow_repo_with_hooks._local_push(
                         merge_commit_id, self.path, push_branches=True,
                         enable_hooks=True)
 
                     # maybe we also need to push the close_commit_id
                     if close_commit_id:
                         shadow_repo_with_hooks._local_push(
                             close_commit_id, self.path, push_branches=True,
                             enable_hooks=True)
                     merge_succeeded = True
                 except RepositoryError:
                     log.exception(
                         'Failure when doing local push from the shadow '
                         'repository to the target repository at %s.', self.path)
                     merge_succeeded = False
                     merge_failure_reason = MergeFailureReason.PUSH_FAILED
                     metadata['target'] = 'hg shadow repo'
                     metadata['merge_commit'] = merge_commit_id
             else:
                 merge_succeeded = True
         else:
             merge_succeeded = False
 
         return MergeResponse(
             merge_possible, merge_succeeded, merge_ref, merge_failure_reason,
             metadata=metadata)
 
     def get_shadow_instance(self, shadow_repository_path, enable_hooks=False, cache=False):
         config = self.config.copy()
         if not enable_hooks:
             config.clear_section('hooks')
         return MercurialRepository(shadow_repository_path, config, with_wire={"cache": cache})
 
     def _validate_pull_reference(self, reference):
         if not (reference.name in self.bookmarks or
                 reference.name in self.branches or
                 self.get_commit(reference.commit_id)):
             raise CommitDoesNotExistError(
                 'Unknown branch, bookmark or commit id')
 
     def _local_pull(self, repository_path, reference):
         """
         Fetch a branch, bookmark or commit from a local repository.
         """
         repository_path = os.path.abspath(repository_path)
         if repository_path == self.path:
             raise ValueError('Cannot pull from the same repository')
 
         reference_type_to_option_name = {
             'book': 'bookmark',
             'branch': 'branch',
         }
         option_name = reference_type_to_option_name.get(
             reference.type, 'revision')
 
         if option_name == 'revision':
             ref = reference.commit_id
         else:
             ref = reference.name
 
         options = {option_name: [ref]}
         self._remote.pull_cmd(repository_path, hooks=False, **options)
         self._remote.invalidate_vcs_cache()
 
     def bookmark(self, bookmark, revision=None):
         if isinstance(bookmark, unicode):
             bookmark = safe_str(bookmark)
         self._remote.bookmark(bookmark, revision=revision)
         self._remote.invalidate_vcs_cache()
 
     def get_path_permissions(self, username):
         hgacl_file = os.path.join(self.path, '.hg/hgacl')
 
         def read_patterns(suffix):
             svalue = None
             for section, option in [
                 ('narrowacl', username + suffix),
                 ('narrowacl', 'default' + suffix),
                 ('narrowhgacl', username + suffix),
                 ('narrowhgacl', 'default' + suffix)
             ]:
                 try:
                     svalue = hgacl.get(section, option)
                     break  # stop at the first value we find
                 except configparser.NoOptionError:
                     pass
             if not svalue:
                 return None
             result = ['/']
             for pattern in svalue.split():
                 result.append(pattern)
                 if '*' not in pattern and '?' not in pattern:
                     result.append(pattern + '/*')
             return result
 
         if os.path.exists(hgacl_file):
             try:
                 hgacl = configparser.RawConfigParser()
                 hgacl.read(hgacl_file)
 
                 includes = read_patterns('.includes')
                 excludes = read_patterns('.excludes')
                 return BasePathPermissionChecker.create_from_patterns(
                     includes, excludes)
             except BaseException as e:
                 msg = 'Cannot read ACL settings from {} on {}: {}'.format(
                     hgacl_file, self.name, e)
                 raise exceptions.RepositoryRequirementError(msg)
         else:
             return None
 
 
 class MercurialIndexBasedCollectionGenerator(CollectionGenerator):
 
     def _commit_factory(self, commit_id):
         return self.repo.get_commit(
             commit_idx=commit_id, pre_load=self.pre_load)

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

@@ -1,370 +1,370 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/
 
 """
 SVN repository module
 """
 
 import logging
 import os
 import urllib.request, urllib.parse, urllib.error
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
-from rhodecode.lib.compat import OrderedDict
+from collections import OrderedDict
 from rhodecode.lib.datelib import date_astimestamp
 from rhodecode.lib.utils import safe_str, safe_unicode
 from rhodecode.lib.utils2 import CachedProperty
 from rhodecode.lib.vcs import connection, path as vcspath
 from rhodecode.lib.vcs.backends import base
 from rhodecode.lib.vcs.backends.svn.commit import (
     SubversionCommit, _date_from_svn_properties)
 from rhodecode.lib.vcs.backends.svn.diff import SubversionDiff
 from rhodecode.lib.vcs.backends.svn.inmemory import SubversionInMemoryCommit
 from rhodecode.lib.vcs.conf import settings
 from rhodecode.lib.vcs.exceptions import (
     CommitDoesNotExistError, EmptyRepositoryError, RepositoryError,
     VCSError, NodeDoesNotExistError)
 
 
 log = logging.getLogger(__name__)
 
 
 class SubversionRepository(base.BaseRepository):
     """
     Subversion backend implementation
 
     .. important::
 
        It is very important to distinguish the commit index and the commit id
        which is assigned by Subversion. The first one is always handled as an
        `int` by this implementation. The commit id assigned by Subversion on
        the other side will always be a `str`.
 
        There is a specific trap since the first commit will have the index
        ``0`` but the svn id will be ``"1"``.
 
     """
 
     # Note: Subversion does not really have a default branch name.
     DEFAULT_BRANCH_NAME = None
 
     contact = base.BaseRepository.DEFAULT_CONTACT
     description = base.BaseRepository.DEFAULT_DESCRIPTION
 
     def __init__(self, repo_path, config=None, create=False, src_url=None, with_wire=None,
                  bare=False, **kwargs):
         self.path = safe_str(os.path.abspath(repo_path))
         self.config = config if config else self.get_default_config()
         self.with_wire = with_wire or {"cache": False}  # default should not use cache
 
         self._init_repo(create, src_url)
 
         # caches
         self._commit_ids = {}
 
     @LazyProperty
     def _remote(self):
         repo_id = self.path
         return connection.Svn(self.path, repo_id, self.config, with_wire=self.with_wire)
 
     def _init_repo(self, create, src_url):
         if create and os.path.exists(self.path):
             raise RepositoryError(
                 "Cannot create repository at %s, location already exist"
                 % self.path)
 
         if create:
             self._remote.create_repository(settings.SVN_COMPATIBLE_VERSION)
             if src_url:
                 src_url = _sanitize_url(src_url)
                 self._remote.import_remote_repository(src_url)
         else:
             self._check_path()
 
     @CachedProperty
     def commit_ids(self):
         head = self._remote.lookup(None)
         return [str(r) for r in range(1, head + 1)]
 
     def _rebuild_cache(self, commit_ids):
         pass
 
     def run_svn_command(self, cmd, **opts):
         """
         Runs given ``cmd`` as svn command and returns tuple
         (stdout, stderr).
 
         :param cmd: full svn 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))
 
         skip_stderr_log = opts.pop('skip_stderr_log', False)
         out, err = self._remote.run_svn_command(cmd, **opts)
         if err and not skip_stderr_log:
             log.debug('Stderr output of svn command "%s":\n%s', cmd, err)
         return out, err
 
     @LazyProperty
     def branches(self):
         return self._tags_or_branches('vcs_svn_branch')
 
     @LazyProperty
     def branches_closed(self):
         return {}
 
     @LazyProperty
     def bookmarks(self):
         return {}
 
     @LazyProperty
     def branches_all(self):
         # TODO: johbo: Implement proper branch support
         all_branches = {}
         all_branches.update(self.branches)
         all_branches.update(self.branches_closed)
         return all_branches
 
     @LazyProperty
     def tags(self):
         return self._tags_or_branches('vcs_svn_tag')
 
     def _tags_or_branches(self, config_section):
         found_items = {}
 
         if self.is_empty():
             return {}
 
         for pattern in self._patterns_from_section(config_section):
             pattern = vcspath.sanitize(pattern)
             tip = self.get_commit()
             try:
                 if pattern.endswith('*'):
                     basedir = tip.get_node(vcspath.dirname(pattern))
                     directories = basedir.dirs
                 else:
                     directories = (tip.get_node(pattern), )
             except NodeDoesNotExistError:
                 continue
             found_items.update(
                 (safe_unicode(n.path),
                  self.commit_ids[-1])
                 for n in directories)
 
         def get_name(item):
             return item[0]
 
         return OrderedDict(sorted(found_items.items(), key=get_name))
 
     def _patterns_from_section(self, section):
         return (pattern for key, pattern in self.config.items(section))
 
     def get_common_ancestor(self, commit_id1, commit_id2, repo2):
         if self != repo2:
             raise ValueError(
                 "Subversion does not support getting common ancestor of"
                 " different repositories.")
 
         if int(commit_id1) < int(commit_id2):
             return commit_id1
         return commit_id2
 
     def verify(self):
         verify = self._remote.verify()
 
         self._remote.invalidate_vcs_cache()
         return verify
 
     def compare(self, commit_id1, commit_id2, repo2, merge, pre_load=None):
         # TODO: johbo: Implement better comparison, this is a very naive
         # version which does not allow to compare branches, tags or folders
         # at all.
         if repo2 != self:
             raise ValueError(
                 "Subversion does not support comparison of of different "
                 "repositories.")
 
         if commit_id1 == commit_id2:
             return []
 
         commit_idx1 = self._get_commit_idx(commit_id1)
         commit_idx2 = self._get_commit_idx(commit_id2)
 
         commits = [
             self.get_commit(commit_idx=idx)
             for idx in range(commit_idx1 + 1, commit_idx2 + 1)]
 
         return commits
 
     def _get_commit_idx(self, commit_id):
         try:
             svn_rev = int(commit_id)
         except:
             # TODO: johbo: this might be only one case, HEAD, check this
             svn_rev = self._remote.lookup(commit_id)
         commit_idx = svn_rev - 1
         if commit_idx >= len(self.commit_ids):
             raise CommitDoesNotExistError(
                 "Commit at index %s does not exist." % (commit_idx, ))
         return commit_idx
 
     @staticmethod
     def check_url(url, config):
         """
         Check if `url` is a valid source to import a Subversion repository.
         """
         # convert to URL if it's a local directory
         if os.path.isdir(url):
             url = 'file://' + urllib.request.pathname2url(url)
         return connection.Svn.check_url(url, config.serialize())
 
     @staticmethod
     def is_valid_repository(path):
         try:
             SubversionRepository(path)
             return True
         except VCSError:
             pass
         return False
 
     def _check_path(self):
         if not os.path.exists(self.path):
             raise VCSError('Path "%s" does not exist!' % (self.path, ))
         if not self._remote.is_path_valid_repository(self.path):
             raise VCSError(
                 'Path "%s" does not contain a Subversion repository' %
                 (self.path, ))
 
     @LazyProperty
     def last_change(self):
         """
         Returns last change made on this repository as
         `datetime.datetime` object.
         """
         # Subversion always has a first commit which has id "0" and contains
         # what we are looking for.
         last_id = len(self.commit_ids)
         properties = self._remote.revision_properties(last_id)
         return _date_from_svn_properties(properties)
 
     @LazyProperty
     def in_memory_commit(self):
         return SubversionInMemoryCommit(self)
 
     def get_hook_location(self):
         """
         returns absolute path to location where hooks are stored
         """
         return os.path.join(self.path, 'hooks')
 
     def get_commit(self, commit_id=None, commit_idx=None, pre_load=None,
                    translate_tag=None, maybe_unreachable=False, reference_obj=None):
         if self.is_empty():
             raise EmptyRepositoryError("There are no commits yet")
         if commit_id is not None:
             self._validate_commit_id(commit_id)
         elif commit_idx is not None:
             self._validate_commit_idx(commit_idx)
             try:
                 commit_id = self.commit_ids[commit_idx]
             except IndexError:
                 raise CommitDoesNotExistError('No commit with idx: {}'.format(commit_idx))
 
         commit_id = self._sanitize_commit_id(commit_id)
         commit = SubversionCommit(repository=self, commit_id=commit_id)
         return commit
 
     def get_commits(
             self, start_id=None, end_id=None, start_date=None, end_date=None,
             branch_name=None, show_hidden=False, pre_load=None, translate_tags=None):
         if self.is_empty():
             raise EmptyRepositoryError("There are no commit_ids 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._sanitize_commit_id(start_id)
         start_pos = self.commit_ids.index(start_raw_id) if start_id else None
         end_raw_id = self._sanitize_commit_id(end_id)
         end_pos = max(0, self.commit_ids.index(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
 
         # Date based filtering
         if start_date or end_date:
             start_raw_id, end_raw_id = self._remote.lookup_interval(
                 date_astimestamp(start_date) if start_date else None,
                 date_astimestamp(end_date) if end_date else None)
             start_pos = start_raw_id - 1
             end_pos = end_raw_id
 
         commit_ids = self.commit_ids
 
         # TODO: johbo: Reconsider impact of DEFAULT_BRANCH_NAME here
         if branch_name not in [None, self.DEFAULT_BRANCH_NAME]:
             svn_rev = long(self.commit_ids[-1])
             commit_ids = self._remote.node_history(
                 path=branch_name, revision=svn_rev, limit=None)
             commit_ids = [str(i) for i in reversed(commit_ids)]
 
         if start_pos or end_pos:
             commit_ids = commit_ids[start_pos:end_pos]
         return base.CollectionGenerator(self, commit_ids, pre_load=pre_load)
 
     def _sanitize_commit_id(self, commit_id):
         if commit_id and commit_id.isdigit():
             if int(commit_id) <= len(self.commit_ids):
                 return commit_id
             else:
                 raise CommitDoesNotExistError(
                     "Commit %s does not exist." % (commit_id, ))
         if commit_id not in [
                 None, 'HEAD', 'tip', self.DEFAULT_BRANCH_NAME]:
             raise CommitDoesNotExistError(
                 "Commit id %s not understood." % (commit_id, ))
         svn_rev = self._remote.lookup('HEAD')
         return str(svn_rev)
 
     def get_diff(
             self, commit1, commit2, path=None, ignore_whitespace=False,
             context=3, path1=None):
         self._validate_diff_commits(commit1, commit2)
         svn_rev1 = long(commit1.raw_id)
         svn_rev2 = long(commit2.raw_id)
         diff = self._remote.diff(
             svn_rev1, svn_rev2, path1=path1, path2=path,
             ignore_whitespace=ignore_whitespace, context=context)
         return SubversionDiff(diff)
 
 
 def _sanitize_url(url):
     if '://' not in url:
         url = 'file://' + urllib.request.pathname2url(url)
     return url

rhodecode/model/pull_request.py

@@ -1,2380 +1,2380 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2012-2020 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/
 
 
 """
 pull request model for RhodeCode
 """
 
 
 import json
 import logging
 import os
 
 import datetime
 import urllib.request, urllib.parse, urllib.error
 import collections
 
 from pyramid.threadlocal import get_current_request
 
 from rhodecode.lib.vcs.nodes import FileNode
 from rhodecode.translation import lazy_ugettext
 from rhodecode.lib import helpers as h, hooks_utils, diffs
 from rhodecode.lib import audit_logger
-from rhodecode.lib.compat import OrderedDict
+from collections import OrderedDict
 from rhodecode.lib.hooks_daemon import prepare_callback_daemon
 from rhodecode.lib.markup_renderer import (
     DEFAULT_COMMENTS_RENDERER, RstTemplateRenderer)
 from rhodecode.lib.utils2 import (
     safe_unicode, safe_str, md5_safe, AttributeDict, safe_int,
     get_current_rhodecode_user)
 from rhodecode.lib.vcs.backends.base import (
     Reference, MergeResponse, MergeFailureReason, UpdateFailureReason,
     TargetRefMissing, SourceRefMissing)
 from rhodecode.lib.vcs.conf import settings as vcs_settings
 from rhodecode.lib.vcs.exceptions import (
     CommitDoesNotExistError, EmptyRepositoryError)
 from rhodecode.model import BaseModel
 from rhodecode.model.changeset_status import ChangesetStatusModel
 from rhodecode.model.comment import CommentsModel
 from rhodecode.model.db import (
     aliased, null, lazyload, and_, or_, func, String, cast, PullRequest, PullRequestReviewers, ChangesetStatus,
     PullRequestVersion, ChangesetComment, Repository, RepoReviewRule, User)
 from rhodecode.model.meta import Session
 from rhodecode.model.notification import NotificationModel, \
     EmailNotificationModel
 from rhodecode.model.scm import ScmModel
 from rhodecode.model.settings import VcsSettingsModel
 
 
 log = logging.getLogger(__name__)
 
 
 # Data structure to hold the response data when updating commits during a pull
 # request update.
 class UpdateResponse(object):
 
     def __init__(self, executed, reason, new, old, common_ancestor_id,
                  commit_changes, source_changed, target_changed):
 
         self.executed = executed
         self.reason = reason
         self.new = new
         self.old = old
         self.common_ancestor_id = common_ancestor_id
         self.changes = commit_changes
         self.source_changed = source_changed
         self.target_changed = target_changed
 
 
 def get_diff_info(
         source_repo, source_ref, target_repo, target_ref, get_authors=False,
         get_commit_authors=True):
     """
     Calculates detailed diff information for usage in preview of creation of a pull-request.
     This is also used for default reviewers logic
     """
 
     source_scm = source_repo.scm_instance()
     target_scm = target_repo.scm_instance()
 
     ancestor_id = target_scm.get_common_ancestor(target_ref, source_ref, source_scm)
     if not ancestor_id:
         raise ValueError(
             'cannot calculate diff info without a common ancestor. '
             'Make sure both repositories are related, and have a common forking commit.')
 
     # case here is that want a simple diff without incoming commits,
     # previewing what will be merged based only on commits in the source.
     log.debug('Using ancestor %s as source_ref instead of %s',
               ancestor_id, source_ref)
 
     # source of changes now is the common ancestor
     source_commit = source_scm.get_commit(commit_id=ancestor_id)
     # target commit becomes the source ref as it is the last commit
     # for diff generation this logic gives proper diff
     target_commit = source_scm.get_commit(commit_id=source_ref)
 
     vcs_diff = \
         source_scm.get_diff(commit1=source_commit, commit2=target_commit,
                             ignore_whitespace=False, context=3)
 
     diff_processor = diffs.DiffProcessor(
         vcs_diff, format='newdiff', diff_limit=None,
         file_limit=None, show_full_diff=True)
 
     _parsed = diff_processor.prepare()
 
     all_files = []
     all_files_changes = []
     changed_lines = {}
     stats = [0, 0]
     for f in _parsed:
         all_files.append(f['filename'])
         all_files_changes.append({
             'filename': f['filename'],
             'stats': f['stats']
         })
         stats[0] += f['stats']['added']
         stats[1] += f['stats']['deleted']
 
         changed_lines[f['filename']] = []
         if len(f['chunks']) < 2:
             continue
         # first line is "context" information
         for chunks in f['chunks'][1:]:
             for chunk in chunks['lines']:
                 if chunk['action'] not in ('del', 'mod'):
                     continue
                 changed_lines[f['filename']].append(chunk['old_lineno'])
 
     commit_authors = []
     user_counts = {}
     email_counts = {}
     author_counts = {}
     _commit_cache = {}
 
     commits = []
     if get_commit_authors:
         log.debug('Obtaining commit authors from set of commits')
         _compare_data = target_scm.compare(
             target_ref, source_ref, source_scm, merge=True,
             pre_load=["author", "date", "message"]
         )
 
         for commit in _compare_data:
             # NOTE(marcink): we serialize here, so we don't produce more vcsserver calls on data returned
             # at this function which is later called via JSON serialization
             serialized_commit = dict(
                 author=commit.author,
                 date=commit.date,
                 message=commit.message,
                 commit_id=commit.raw_id,
                 raw_id=commit.raw_id
             )
             commits.append(serialized_commit)
             user = User.get_from_cs_author(serialized_commit['author'])
             if user and user not in commit_authors:
                 commit_authors.append(user)
 
     # lines
     if get_authors:
         log.debug('Calculating authors of changed files')
         target_commit = source_repo.get_commit(ancestor_id)
 
         for fname, lines in changed_lines.items():
 
             try:
                 node = target_commit.get_node(fname, pre_load=["is_binary"])
             except Exception:
                 log.exception("Failed to load node with path %s", fname)
                 continue
 
             if not isinstance(node, FileNode):
                 continue
 
             # NOTE(marcink): for binary node we don't do annotation, just use last author
             if node.is_binary:
                 author = node.last_commit.author
                 email = node.last_commit.author_email
 
                 user = User.get_from_cs_author(author)
                 if user:
                     user_counts[user.user_id] = user_counts.get(user.user_id, 0) + 1
                 author_counts[author] = author_counts.get(author, 0) + 1
                 email_counts[email] = email_counts.get(email, 0) + 1
 
                 continue
 
             for annotation in node.annotate:
                 line_no, commit_id, get_commit_func, line_text = annotation
                 if line_no in lines:
                     if commit_id not in _commit_cache:
                         _commit_cache[commit_id] = get_commit_func()
                     commit = _commit_cache[commit_id]
                     author = commit.author
                     email = commit.author_email
                     user = User.get_from_cs_author(author)
                     if user:
                         user_counts[user.user_id] = user_counts.get(user.user_id, 0) + 1
                     author_counts[author] = author_counts.get(author, 0) + 1
                     email_counts[email] = email_counts.get(email, 0) + 1
 
     log.debug('Default reviewers processing finished')
 
     return {
         'commits': commits,
         'files': all_files_changes,
         'stats': stats,
         'ancestor': ancestor_id,
         # original authors of modified files
         'original_authors': {
             'users': user_counts,
             'authors': author_counts,
             'emails': email_counts,
         },
         'commit_authors': commit_authors
     }
 
 
 class PullRequestModel(BaseModel):
 
     cls = PullRequest
 
     DIFF_CONTEXT = diffs.DEFAULT_CONTEXT
 
     UPDATE_STATUS_MESSAGES = {
         UpdateFailureReason.NONE: lazy_ugettext(
             'Pull request update successful.'),
         UpdateFailureReason.UNKNOWN: lazy_ugettext(
             'Pull request update failed because of an unknown error.'),
         UpdateFailureReason.NO_CHANGE: lazy_ugettext(
             'No update needed because the source and target have not changed.'),
         UpdateFailureReason.WRONG_REF_TYPE: lazy_ugettext(
             'Pull request cannot be updated because the reference type is '
             'not supported for an update. Only Branch, Tag or Bookmark is allowed.'),
         UpdateFailureReason.MISSING_TARGET_REF: lazy_ugettext(
             'This pull request cannot be updated because the target '
             'reference is missing.'),
         UpdateFailureReason.MISSING_SOURCE_REF: lazy_ugettext(
             'This pull request cannot be updated because the source '
             'reference is missing.'),
     }
     REF_TYPES = ['bookmark', 'book', 'tag', 'branch']
     UPDATABLE_REF_TYPES = ['bookmark', 'book', 'branch']
 
     def __get_pull_request(self, pull_request):
         return self._get_instance((
             PullRequest, PullRequestVersion), pull_request)
 
     def _check_perms(self, perms, pull_request, user, api=False):
         if not api:
             return h.HasRepoPermissionAny(*perms)(
                 user=user, repo_name=pull_request.target_repo.repo_name)
         else:
             return h.HasRepoPermissionAnyApi(*perms)(
                 user=user, repo_name=pull_request.target_repo.repo_name)
 
     def check_user_read(self, pull_request, user, api=False):
         _perms = ('repository.admin', 'repository.write', 'repository.read',)
         return self._check_perms(_perms, pull_request, user, api)
 
     def check_user_merge(self, pull_request, user, api=False):
         _perms = ('repository.admin', 'repository.write', 'hg.admin',)
         return self._check_perms(_perms, pull_request, user, api)
 
     def check_user_update(self, pull_request, user, api=False):
         owner = user.user_id == pull_request.user_id
         return self.check_user_merge(pull_request, user, api) or owner
 
     def check_user_delete(self, pull_request, user):
         owner = user.user_id == pull_request.user_id
         _perms = ('repository.admin',)
         return self._check_perms(_perms, pull_request, user) or owner
 
     def is_user_reviewer(self, pull_request, user):
         return user.user_id in [
             x.user_id for x in
             pull_request.get_pull_request_reviewers(PullRequestReviewers.ROLE_REVIEWER)
             if x.user
         ]
 
     def check_user_change_status(self, pull_request, user, api=False):
         return self.check_user_update(pull_request, user, api) \
                or self.is_user_reviewer(pull_request, user)
 
     def check_user_comment(self, pull_request, user):
         owner = user.user_id == pull_request.user_id
         return self.check_user_read(pull_request, user) or owner
 
     def get(self, pull_request):
         return self.__get_pull_request(pull_request)
 
     def _prepare_get_all_query(self, repo_name, search_q=None, source=False,
                                statuses=None, opened_by=None, order_by=None,
                                order_dir='desc', only_created=False):
         repo = None
         if repo_name:
             repo = self._get_repo(repo_name)
 
         q = PullRequest.query()
 
         if search_q:
             like_expression = u'%{}%'.format(safe_unicode(search_q))
             q = q.join(User, User.user_id == PullRequest.user_id)
             q = q.filter(or_(
                 cast(PullRequest.pull_request_id, String).ilike(like_expression),
                 User.username.ilike(like_expression),
                 PullRequest.title.ilike(like_expression),
                 PullRequest.description.ilike(like_expression),
             ))
 
         # source or target
         if repo and source:
             q = q.filter(PullRequest.source_repo == repo)
         elif repo:
             q = q.filter(PullRequest.target_repo == repo)
 
         # closed,opened
         if statuses:
             q = q.filter(PullRequest.status.in_(statuses))
 
         # opened by filter
         if opened_by:
             q = q.filter(PullRequest.user_id.in_(opened_by))
 
         # only get those that are in "created" state
         if only_created:
             q = q.filter(PullRequest.pull_request_state == PullRequest.STATE_CREATED)
 
         order_map = {
             'name_raw': PullRequest.pull_request_id,
             'id': PullRequest.pull_request_id,
             'title': PullRequest.title,
             'updated_on_raw': PullRequest.updated_on,
             'target_repo': PullRequest.target_repo_id
         }
         if order_by and order_by in order_map:
             if order_dir == 'asc':
                 q = q.order_by(order_map[order_by].asc())
             else:
                 q = q.order_by(order_map[order_by].desc())
 
         return q
 
     def count_all(self, repo_name, search_q=None, source=False, statuses=None,
                   opened_by=None):
         """
         Count the number of pull requests for a specific repository.
 
         :param repo_name: target or source repo
         :param search_q: filter by text
         :param source: boolean flag to specify if repo_name refers to source
         :param statuses: list of pull request statuses
         :param opened_by: author user of the pull request
         :returns: int number of pull requests
         """
         q = self._prepare_get_all_query(
             repo_name, search_q=search_q, source=source, statuses=statuses,
             opened_by=opened_by)
 
         return q.count()
 
     def get_all(self, repo_name, search_q=None, source=False, statuses=None,
                 opened_by=None, offset=0, length=None, order_by=None, order_dir='desc'):
         """
         Get all pull requests for a specific repository.
 
         :param repo_name: target or source repo
         :param search_q: filter by text
         :param source: boolean flag to specify if repo_name refers to source
         :param statuses: list of pull request statuses
         :param opened_by: author user of the pull request
         :param offset: pagination offset
         :param length: length of returned list
         :param order_by: order of the returned list
         :param order_dir: 'asc' or 'desc' ordering direction
         :returns: list of pull requests
         """
         q = self._prepare_get_all_query(
             repo_name, search_q=search_q, source=source, statuses=statuses,
             opened_by=opened_by, order_by=order_by, order_dir=order_dir)
 
         if length:
             pull_requests = q.limit(length).offset(offset).all()
         else:
             pull_requests = q.all()
 
         return pull_requests
 
     def count_awaiting_review(self, repo_name, search_q=None, statuses=None):
         """
         Count the number of pull requests for a specific repository that are
         awaiting review.
 
         :param repo_name: target or source repo
         :param search_q: filter by text
         :param statuses: list of pull request statuses
         :returns: int number of pull requests
         """
         pull_requests = self.get_awaiting_review(
             repo_name, search_q=search_q, statuses=statuses)
 
         return len(pull_requests)
 
     def get_awaiting_review(self, repo_name, search_q=None, statuses=None,
                             offset=0, length=None, order_by=None, order_dir='desc'):
         """
         Get all pull requests for a specific repository that are awaiting
         review.
 
         :param repo_name: target or source repo
         :param search_q: filter by text
         :param statuses: list of pull request statuses
         :param offset: pagination offset
         :param length: length of returned list
         :param order_by: order of the returned list
         :param order_dir: 'asc' or 'desc' ordering direction
         :returns: list of pull requests
         """
         pull_requests = self.get_all(
             repo_name, search_q=search_q, statuses=statuses,
             order_by=order_by, order_dir=order_dir)
 
         _filtered_pull_requests = []
         for pr in pull_requests:
             status = pr.calculated_review_status()
             if status in [ChangesetStatus.STATUS_NOT_REVIEWED,
                           ChangesetStatus.STATUS_UNDER_REVIEW]:
                 _filtered_pull_requests.append(pr)
         if length:
             return _filtered_pull_requests[offset:offset+length]
         else:
             return _filtered_pull_requests
 
     def _prepare_awaiting_my_review_review_query(
             self, repo_name, user_id, search_q=None, statuses=None,
             order_by=None, order_dir='desc'):
 
         for_review_statuses = [
             ChangesetStatus.STATUS_UNDER_REVIEW, ChangesetStatus.STATUS_NOT_REVIEWED
         ]
 
         pull_request_alias = aliased(PullRequest)
         status_alias = aliased(ChangesetStatus)
         reviewers_alias = aliased(PullRequestReviewers)
         repo_alias = aliased(Repository)
 
         last_ver_subq = Session()\
             .query(func.min(ChangesetStatus.version)) \
             .filter(ChangesetStatus.pull_request_id == reviewers_alias.pull_request_id)\
             .filter(ChangesetStatus.user_id == reviewers_alias.user_id) \
             .subquery()
 
         q = Session().query(pull_request_alias) \
             .options(lazyload(pull_request_alias.author)) \
             .join(reviewers_alias,
                   reviewers_alias.pull_request_id == pull_request_alias.pull_request_id) \
             .join(repo_alias,
                   repo_alias.repo_id == pull_request_alias.target_repo_id) \
             .outerjoin(status_alias,
                        and_(status_alias.user_id == reviewers_alias.user_id,
                             status_alias.pull_request_id == reviewers_alias.pull_request_id)) \
             .filter(or_(status_alias.version == null(),
                         status_alias.version == last_ver_subq)) \
             .filter(reviewers_alias.user_id == user_id) \
             .filter(repo_alias.repo_name == repo_name) \
             .filter(or_(status_alias.status == null(), status_alias.status.in_(for_review_statuses))) \
             .group_by(pull_request_alias)
 
         # closed,opened
         if statuses:
             q = q.filter(pull_request_alias.status.in_(statuses))
 
         if search_q:
             like_expression = u'%{}%'.format(safe_unicode(search_q))
             q = q.join(User, User.user_id == pull_request_alias.user_id)
             q = q.filter(or_(
                 cast(pull_request_alias.pull_request_id, String).ilike(like_expression),
                 User.username.ilike(like_expression),
                 pull_request_alias.title.ilike(like_expression),
                 pull_request_alias.description.ilike(like_expression),
             ))
 
         order_map = {
             'name_raw': pull_request_alias.pull_request_id,
             'title': pull_request_alias.title,
             'updated_on_raw': pull_request_alias.updated_on,
             'target_repo': pull_request_alias.target_repo_id
         }
         if order_by and order_by in order_map:
             if order_dir == 'asc':
                 q = q.order_by(order_map[order_by].asc())
             else:
                 q = q.order_by(order_map[order_by].desc())
 
         return q
 
     def count_awaiting_my_review(self, repo_name, user_id, search_q=None, statuses=None):
         """
         Count the number of pull requests for a specific repository that are
         awaiting review from a specific user.
 
         :param repo_name: target or source repo
         :param user_id: reviewer user of the pull request
         :param search_q: filter by text
         :param statuses: list of pull request statuses
         :returns: int number of pull requests
         """
         q = self._prepare_awaiting_my_review_review_query(
             repo_name, user_id, search_q=search_q, statuses=statuses)
         return q.count()
 
     def get_awaiting_my_review(self, repo_name, user_id, search_q=None, statuses=None,
                                offset=0, length=None, order_by=None, order_dir='desc'):
         """
         Get all pull requests for a specific repository that are awaiting
         review from a specific user.
 
         :param repo_name: target or source repo
         :param user_id: reviewer user of the pull request
         :param search_q: filter by text
         :param statuses: list of pull request statuses
         :param offset: pagination offset
         :param length: length of returned list
         :param order_by: order of the returned list
         :param order_dir: 'asc' or 'desc' ordering direction
         :returns: list of pull requests
         """
 
         q = self._prepare_awaiting_my_review_review_query(
             repo_name, user_id, search_q=search_q, statuses=statuses,
             order_by=order_by, order_dir=order_dir)
 
         if length:
             pull_requests = q.limit(length).offset(offset).all()
         else:
             pull_requests = q.all()
 
         return pull_requests
 
     def _prepare_im_participating_query(self, user_id=None, statuses=None, query='',
                                         order_by=None, order_dir='desc'):
         """
         return a query of pull-requests user is an creator, or he's added as a reviewer
         """
         q = PullRequest.query()
         if user_id:
             reviewers_subquery = Session().query(
                 PullRequestReviewers.pull_request_id).filter(
                 PullRequestReviewers.user_id == user_id).subquery()
             user_filter = or_(
                 PullRequest.user_id == user_id,
                 PullRequest.pull_request_id.in_(reviewers_subquery)
             )
             q = PullRequest.query().filter(user_filter)
 
         # closed,opened
         if statuses:
             q = q.filter(PullRequest.status.in_(statuses))
 
         if query:
             like_expression = u'%{}%'.format(safe_unicode(query))
             q = q.join(User, User.user_id == PullRequest.user_id)
             q = q.filter(or_(
                 cast(PullRequest.pull_request_id, String).ilike(like_expression),
                 User.username.ilike(like_expression),
                 PullRequest.title.ilike(like_expression),
                 PullRequest.description.ilike(like_expression),
             ))
 
         order_map = {
             'name_raw': PullRequest.pull_request_id,
             'title': PullRequest.title,
             'updated_on_raw': PullRequest.updated_on,
             'target_repo': PullRequest.target_repo_id
         }
         if order_by and order_by in order_map:
             if order_dir == 'asc':
                 q = q.order_by(order_map[order_by].asc())
             else:
                 q = q.order_by(order_map[order_by].desc())
 
         return q
 
     def count_im_participating_in(self, user_id=None, statuses=None, query=''):
         q = self._prepare_im_participating_query(user_id, statuses=statuses, query=query)
         return q.count()
 
     def get_im_participating_in(
             self, user_id=None, statuses=None, query='', offset=0,
             length=None, order_by=None, order_dir='desc'):
         """
         Get all Pull requests that i'm participating in as a reviewer, or i have opened
         """
 
         q = self._prepare_im_participating_query(
             user_id, statuses=statuses, query=query, order_by=order_by,
             order_dir=order_dir)
 
         if length:
             pull_requests = q.limit(length).offset(offset).all()
         else:
             pull_requests = q.all()
 
         return pull_requests
 
     def _prepare_participating_in_for_review_query(
             self, user_id, statuses=None, query='', order_by=None, order_dir='desc'):
 
         for_review_statuses = [
             ChangesetStatus.STATUS_UNDER_REVIEW, ChangesetStatus.STATUS_NOT_REVIEWED
         ]
 
         pull_request_alias = aliased(PullRequest)
         status_alias = aliased(ChangesetStatus)
         reviewers_alias = aliased(PullRequestReviewers)
 
         last_ver_subq = Session()\
             .query(func.min(ChangesetStatus.version)) \
             .filter(ChangesetStatus.pull_request_id == reviewers_alias.pull_request_id)\
             .filter(ChangesetStatus.user_id == reviewers_alias.user_id) \
             .subquery()
 
         q = Session().query(pull_request_alias) \
             .options(lazyload(pull_request_alias.author)) \
             .join(reviewers_alias,
                   reviewers_alias.pull_request_id == pull_request_alias.pull_request_id) \
             .outerjoin(status_alias,
                        and_(status_alias.user_id == reviewers_alias.user_id,
                             status_alias.pull_request_id == reviewers_alias.pull_request_id)) \
             .filter(or_(status_alias.version == null(),
                         status_alias.version == last_ver_subq)) \
             .filter(reviewers_alias.user_id == user_id) \
             .filter(or_(status_alias.status == null(), status_alias.status.in_(for_review_statuses))) \
             .group_by(pull_request_alias)
 
         # closed,opened
         if statuses:
             q = q.filter(pull_request_alias.status.in_(statuses))
 
         if query:
             like_expression = u'%{}%'.format(safe_unicode(query))
             q = q.join(User, User.user_id == pull_request_alias.user_id)
             q = q.filter(or_(
                 cast(pull_request_alias.pull_request_id, String).ilike(like_expression),
                 User.username.ilike(like_expression),
                 pull_request_alias.title.ilike(like_expression),
                 pull_request_alias.description.ilike(like_expression),
             ))
 
         order_map = {
             'name_raw': pull_request_alias.pull_request_id,
             'title': pull_request_alias.title,
             'updated_on_raw': pull_request_alias.updated_on,
             'target_repo': pull_request_alias.target_repo_id
         }
         if order_by and order_by in order_map:
             if order_dir == 'asc':
                 q = q.order_by(order_map[order_by].asc())
             else:
                 q = q.order_by(order_map[order_by].desc())
 
         return q
 
     def count_im_participating_in_for_review(self, user_id, statuses=None, query=''):
         q = self._prepare_participating_in_for_review_query(user_id, statuses=statuses, query=query)
         return q.count()
 
     def get_im_participating_in_for_review(
             self, user_id, statuses=None, query='', offset=0,
             length=None, order_by=None, order_dir='desc'):
         """
         Get all Pull requests that needs user approval or rejection
         """
 
         q = self._prepare_participating_in_for_review_query(
             user_id, statuses=statuses, query=query, order_by=order_by,
             order_dir=order_dir)
 
         if length:
             pull_requests = q.limit(length).offset(offset).all()
         else:
             pull_requests = q.all()
 
         return pull_requests
 
     def get_versions(self, pull_request):
         """
         returns version of pull request sorted by ID descending
         """
         return PullRequestVersion.query()\
             .filter(PullRequestVersion.pull_request == pull_request)\
             .order_by(PullRequestVersion.pull_request_version_id.asc())\
             .all()
 
     def get_pr_version(self, pull_request_id, version=None):
         at_version = None
 
         if version and version == 'latest':
             pull_request_ver = PullRequest.get(pull_request_id)
             pull_request_obj = pull_request_ver
             _org_pull_request_obj = pull_request_obj
             at_version = 'latest'
         elif version:
             pull_request_ver = PullRequestVersion.get_or_404(version)
             pull_request_obj = pull_request_ver
             _org_pull_request_obj = pull_request_ver.pull_request
             at_version = pull_request_ver.pull_request_version_id
         else:
             _org_pull_request_obj = pull_request_obj = PullRequest.get_or_404(
                 pull_request_id)
 
         pull_request_display_obj = PullRequest.get_pr_display_object(
             pull_request_obj, _org_pull_request_obj)
 
         return _org_pull_request_obj, pull_request_obj, \
                pull_request_display_obj, at_version
 
     def pr_commits_versions(self, versions):
         """
         Maps the pull-request commits into all known PR versions. This way we can obtain
         each pr version the commit was introduced in.
         """
         commit_versions = collections.defaultdict(list)
         num_versions = [x.pull_request_version_id for x in versions]
         for ver in versions:
             for commit_id in ver.revisions:
                 ver_idx = ChangesetComment.get_index_from_version(
                     ver.pull_request_version_id, num_versions=num_versions)
                 commit_versions[commit_id].append(ver_idx)
         return commit_versions
 
     def create(self, created_by, source_repo, source_ref, target_repo,
                target_ref, revisions, reviewers, observers, title, description=None,
                common_ancestor_id=None,
                description_renderer=None,
                reviewer_data=None, translator=None, auth_user=None):
         translator = translator or get_current_request().translate
 
         created_by_user = self._get_user(created_by)
         auth_user = auth_user or created_by_user.AuthUser()
         source_repo = self._get_repo(source_repo)
         target_repo = self._get_repo(target_repo)
 
         pull_request = PullRequest()
         pull_request.source_repo = source_repo
         pull_request.source_ref = source_ref
         pull_request.target_repo = target_repo
         pull_request.target_ref = target_ref
         pull_request.revisions = revisions
         pull_request.title = title
         pull_request.description = description
         pull_request.description_renderer = description_renderer
         pull_request.author = created_by_user
         pull_request.reviewer_data = reviewer_data
         pull_request.pull_request_state = pull_request.STATE_CREATING
         pull_request.common_ancestor_id = common_ancestor_id
 
         Session().add(pull_request)
         Session().flush()
 
         reviewer_ids = set()
         # members / reviewers
         for reviewer_object in reviewers:
             user_id, reasons, mandatory, role, rules = reviewer_object
             user = self._get_user(user_id)
 
             # skip duplicates
             if user.user_id in reviewer_ids:
                 continue
 
             reviewer_ids.add(user.user_id)
 
             reviewer = PullRequestReviewers()
             reviewer.user = user
             reviewer.pull_request = pull_request
             reviewer.reasons = reasons
             reviewer.mandatory = mandatory
             reviewer.role = role
 
             # NOTE(marcink): pick only first rule for now
             rule_id = list(rules)[0] if rules else None
             rule = RepoReviewRule.get(rule_id) if rule_id else None
             if rule:
                 review_group = rule.user_group_vote_rule(user_id)
                 # we check if this particular reviewer is member of a voting group
                 if review_group:
                     # NOTE(marcink):
                     # can be that user is member of more but we pick the first same,
                     # same as default reviewers algo
                     review_group = review_group[0]
 
                     rule_data = {
                         'rule_name':
                             rule.review_rule_name,
                         'rule_user_group_entry_id':
                             review_group.repo_review_rule_users_group_id,
                         'rule_user_group_name':
                             review_group.users_group.users_group_name,
                         'rule_user_group_members':
                             [x.user.username for x in review_group.users_group.members],
                         'rule_user_group_members_id':
                             [x.user.user_id for x in review_group.users_group.members],
                     }
                     # e.g {'vote_rule': -1, 'mandatory': True}
                     rule_data.update(review_group.rule_data())
 
                     reviewer.rule_data = rule_data
 
             Session().add(reviewer)
             Session().flush()
 
         for observer_object in observers:
             user_id, reasons, mandatory, role, rules = observer_object
             user = self._get_user(user_id)
 
             # skip duplicates from reviewers
             if user.user_id in reviewer_ids:
                 continue
 
             #reviewer_ids.add(user.user_id)
 
             observer = PullRequestReviewers()
             observer.user = user
             observer.pull_request = pull_request
             observer.reasons = reasons
             observer.mandatory = mandatory
             observer.role = role
 
             # NOTE(marcink): pick only first rule for now
             rule_id = list(rules)[0] if rules else None
             rule = RepoReviewRule.get(rule_id) if rule_id else None
             if rule:
                 # TODO(marcink): do we need this for observers ??
                 pass
 
             Session().add(observer)
             Session().flush()
 
         # Set approval status to "Under Review" for all commits which are
         # part of this pull request.
         ChangesetStatusModel().set_status(
             repo=target_repo,
             status=ChangesetStatus.STATUS_UNDER_REVIEW,
             user=created_by_user,
             pull_request=pull_request
         )
         # we commit early at this point. This has to do with a fact
         # that before queries do some row-locking. And because of that
         # we need to commit and finish transaction before below validate call
         # that for large repos could be long resulting in long row locks
         Session().commit()
 
         # prepare workspace, and run initial merge simulation. Set state during that
         # operation
         pull_request = PullRequest.get(pull_request.pull_request_id)
 
         # set as merging, for merge simulation, and if finished to created so we mark
         # simulation is working fine
         with pull_request.set_state(PullRequest.STATE_MERGING,
                                     final_state=PullRequest.STATE_CREATED) as state_obj:
             MergeCheck.validate(
                 pull_request, auth_user=auth_user, translator=translator)
 
         self.notify_reviewers(pull_request, reviewer_ids, created_by_user)
         self.trigger_pull_request_hook(pull_request, created_by_user, 'create')
 
         creation_data = pull_request.get_api_data(with_merge_state=False)
         self._log_audit_action(
             'repo.pull_request.create', {'data': creation_data},
             auth_user, pull_request)
 
         return pull_request
 
     def trigger_pull_request_hook(self, pull_request, user, action, data=None):
         pull_request = self.__get_pull_request(pull_request)
         target_scm = pull_request.target_repo.scm_instance()
         if action == 'create':
             trigger_hook = hooks_utils.trigger_create_pull_request_hook
         elif action == 'merge':
             trigger_hook = hooks_utils.trigger_merge_pull_request_hook
         elif action == 'close':
             trigger_hook = hooks_utils.trigger_close_pull_request_hook
         elif action == 'review_status_change':
             trigger_hook = hooks_utils.trigger_review_pull_request_hook
         elif action == 'update':
             trigger_hook = hooks_utils.trigger_update_pull_request_hook
         elif action == 'comment':
             trigger_hook = hooks_utils.trigger_comment_pull_request_hook
         elif action == 'comment_edit':
             trigger_hook = hooks_utils.trigger_comment_pull_request_edit_hook
         else:
             return
 
         log.debug('Handling pull_request %s trigger_pull_request_hook with action %s and hook: %s',
                   pull_request, action, trigger_hook)
         trigger_hook(
             username=user.username,
             repo_name=pull_request.target_repo.repo_name,
             repo_type=target_scm.alias,
             pull_request=pull_request,
             data=data)
 
     def _get_commit_ids(self, pull_request):
         """
         Return the commit ids of the merged pull request.
 
         This method is not dealing correctly yet with the lack of autoupdates
         nor with the implicit target updates.
         For example: if a commit in the source repo is already in the target it
         will be reported anyways.
         """
         merge_rev = pull_request.merge_rev
         if merge_rev is None:
             raise ValueError('This pull request was not merged yet')
 
         commit_ids = list(pull_request.revisions)
         if merge_rev not in commit_ids:
             commit_ids.append(merge_rev)
 
         return commit_ids
 
     def merge_repo(self, pull_request, user, extras):
         repo_type = pull_request.source_repo.repo_type
         log.debug("Merging pull request %s", pull_request.pull_request_id)
         extras['user_agent'] = '{}/internal-merge'.format(repo_type)
         merge_state = self._merge_pull_request(pull_request, user, extras)
         if merge_state.executed:
             log.debug("Merge was successful, updating the pull request comments.")
             self._comment_and_close_pr(pull_request, user, merge_state)
 
             self._log_audit_action(
                 'repo.pull_request.merge',
                 {'merge_state': merge_state.__dict__},
                 user, pull_request)
 
         else:
             log.warn("Merge failed, not updating the pull request.")
         return merge_state
 
     def _merge_pull_request(self, pull_request, user, extras, merge_msg=None):
         target_vcs = pull_request.target_repo.scm_instance()
         source_vcs = pull_request.source_repo.scm_instance()
 
         message = safe_unicode(merge_msg or vcs_settings.MERGE_MESSAGE_TMPL).format(
             pr_id=pull_request.pull_request_id,
             pr_title=pull_request.title,
             pr_desc=pull_request.description,
             source_repo=source_vcs.name,
             source_ref_name=pull_request.source_ref_parts.name,
             target_repo=target_vcs.name,
             target_ref_name=pull_request.target_ref_parts.name,
         )
 
         workspace_id = self._workspace_id(pull_request)
         repo_id = pull_request.target_repo.repo_id
         use_rebase = self._use_rebase_for_merging(pull_request)
         close_branch = self._close_branch_before_merging(pull_request)
         user_name = self._user_name_for_merging(pull_request, user)
 
         target_ref = self._refresh_reference(
             pull_request.target_ref_parts, target_vcs)
 
         callback_daemon, extras = prepare_callback_daemon(
             extras, protocol=vcs_settings.HOOKS_PROTOCOL,
             host=vcs_settings.HOOKS_HOST,
             use_direct_calls=vcs_settings.HOOKS_DIRECT_CALLS)
 
         with callback_daemon:
             # TODO: johbo: Implement a clean way to run a config_override
             # for a single call.
             target_vcs.config.set(
                 'rhodecode', 'RC_SCM_DATA', json.dumps(extras))
 
             merge_state = target_vcs.merge(
                 repo_id, workspace_id, target_ref, source_vcs,
                 pull_request.source_ref_parts,
                 user_name=user_name, user_email=user.email,
                 message=message, use_rebase=use_rebase,
                 close_branch=close_branch)
         return merge_state
 
     def _comment_and_close_pr(self, pull_request, user, merge_state, close_msg=None):
         pull_request.merge_rev = merge_state.merge_ref.commit_id
         pull_request.updated_on = datetime.datetime.now()
         close_msg = close_msg or 'Pull request merged and closed'
 
         CommentsModel().create(
             text=safe_unicode(close_msg),
             repo=pull_request.target_repo.repo_id,
             user=user.user_id,
             pull_request=pull_request.pull_request_id,
             f_path=None,
             line_no=None,
             closing_pr=True
         )
 
         Session().add(pull_request)
         Session().flush()
         # TODO: paris: replace invalidation with less radical solution
         ScmModel().mark_for_invalidation(
             pull_request.target_repo.repo_name)
         self.trigger_pull_request_hook(pull_request, user, 'merge')
 
     def has_valid_update_type(self, pull_request):
         source_ref_type = pull_request.source_ref_parts.type
         return source_ref_type in self.REF_TYPES
 
     def get_flow_commits(self, pull_request):
 
         # source repo
         source_ref_name = pull_request.source_ref_parts.name
         source_ref_type = pull_request.source_ref_parts.type
         source_ref_id = pull_request.source_ref_parts.commit_id
         source_repo = pull_request.source_repo.scm_instance()
 
         try:
             if source_ref_type in self.REF_TYPES:
                 source_commit = source_repo.get_commit(
                     source_ref_name, reference_obj=pull_request.source_ref_parts)
             else:
                 source_commit = source_repo.get_commit(source_ref_id)
         except CommitDoesNotExistError:
             raise SourceRefMissing()
 
         # target repo
         target_ref_name = pull_request.target_ref_parts.name
         target_ref_type = pull_request.target_ref_parts.type
         target_ref_id = pull_request.target_ref_parts.commit_id
         target_repo = pull_request.target_repo.scm_instance()
 
         try:
             if target_ref_type in self.REF_TYPES:
                 target_commit = target_repo.get_commit(
                     target_ref_name, reference_obj=pull_request.target_ref_parts)
             else:
                 target_commit = target_repo.get_commit(target_ref_id)
         except CommitDoesNotExistError:
             raise TargetRefMissing()
 
         return source_commit, target_commit
 
     def update_commits(self, pull_request, updating_user):
         """
         Get the updated list of commits for the pull request
         and return the new pull request version and the list
         of commits processed by this update action
 
         updating_user is the user_object who triggered the update
         """
         pull_request = self.__get_pull_request(pull_request)
         source_ref_type = pull_request.source_ref_parts.type
         source_ref_name = pull_request.source_ref_parts.name
         source_ref_id = pull_request.source_ref_parts.commit_id
 
         target_ref_type = pull_request.target_ref_parts.type
         target_ref_name = pull_request.target_ref_parts.name
         target_ref_id = pull_request.target_ref_parts.commit_id
 
         if not self.has_valid_update_type(pull_request):
             log.debug("Skipping update of pull request %s due to ref type: %s",
                       pull_request, source_ref_type)
             return UpdateResponse(
                 executed=False,
                 reason=UpdateFailureReason.WRONG_REF_TYPE,
                 old=pull_request, new=None, common_ancestor_id=None, commit_changes=None,
                 source_changed=False, target_changed=False)
 
         try:
             source_commit, target_commit = self.get_flow_commits(pull_request)
         except SourceRefMissing:
             return UpdateResponse(
                 executed=False,
                 reason=UpdateFailureReason.MISSING_SOURCE_REF,
                 old=pull_request, new=None, common_ancestor_id=None, commit_changes=None,
                 source_changed=False, target_changed=False)
         except TargetRefMissing:
             return UpdateResponse(
                 executed=False,
                 reason=UpdateFailureReason.MISSING_TARGET_REF,
                 old=pull_request, new=None, common_ancestor_id=None, commit_changes=None,
                 source_changed=False, target_changed=False)
 
         source_changed = source_ref_id != source_commit.raw_id
         target_changed = target_ref_id != target_commit.raw_id
 
         if not (source_changed or target_changed):
             log.debug("Nothing changed in pull request %s", pull_request)
             return UpdateResponse(
                 executed=False,
                 reason=UpdateFailureReason.NO_CHANGE,
                 old=pull_request, new=None, common_ancestor_id=None, commit_changes=None,
                 source_changed=target_changed, target_changed=source_changed)
 
         change_in_found = 'target repo' if target_changed else 'source repo'
         log.debug('Updating pull request because of change in %s detected',
                   change_in_found)
 
         # Finally there is a need for an update, in case of source change
         # we create a new version, else just an update
         if source_changed:
             pull_request_version = self._create_version_from_snapshot(pull_request)
             self._link_comments_to_version(pull_request_version)
         else:
             try:
                 ver = pull_request.versions[-1]
             except IndexError:
                 ver = None
 
             pull_request.pull_request_version_id = \
                 ver.pull_request_version_id if ver else None
             pull_request_version = pull_request
 
         source_repo = pull_request.source_repo.scm_instance()
         target_repo = pull_request.target_repo.scm_instance()
 
         # re-compute commit ids
         old_commit_ids = pull_request.revisions
         pre_load = ["author", "date", "message", "branch"]
         commit_ranges = target_repo.compare(
             target_commit.raw_id, source_commit.raw_id, source_repo, merge=True,
             pre_load=pre_load)
 
         target_ref = target_commit.raw_id
         source_ref = source_commit.raw_id
         ancestor_commit_id = target_repo.get_common_ancestor(
             target_ref, source_ref, source_repo)
 
         if not ancestor_commit_id:
             raise ValueError(
                 'cannot calculate diff info without a common ancestor. '
                 'Make sure both repositories are related, and have a common forking commit.')
 
         pull_request.common_ancestor_id = ancestor_commit_id
 
         pull_request.source_ref = '%s:%s:%s' % (
             source_ref_type, source_ref_name, source_commit.raw_id)
         pull_request.target_ref = '%s:%s:%s' % (
             target_ref_type, target_ref_name, ancestor_commit_id)
 
         pull_request.revisions = [
             commit.raw_id for commit in reversed(commit_ranges)]
         pull_request.updated_on = datetime.datetime.now()
         Session().add(pull_request)
         new_commit_ids = pull_request.revisions
 
         old_diff_data, new_diff_data = self._generate_update_diffs(
             pull_request, pull_request_version)
 
         # calculate commit and file changes
         commit_changes = self._calculate_commit_id_changes(
             old_commit_ids, new_commit_ids)
         file_changes = self._calculate_file_changes(
             old_diff_data, new_diff_data)
 
         # set comments as outdated if DIFFS changed
         CommentsModel().outdate_comments(
             pull_request, old_diff_data=old_diff_data,
             new_diff_data=new_diff_data)
 
         valid_commit_changes = (commit_changes.added or commit_changes.removed)
         file_node_changes = (
             file_changes.added or file_changes.modified or file_changes.removed)
         pr_has_changes = valid_commit_changes or file_node_changes
 
         # Add an automatic comment to the pull request, in case
         # anything has changed
         if pr_has_changes:
             update_comment = CommentsModel().create(
                 text=self._render_update_message(ancestor_commit_id, commit_changes, file_changes),
                 repo=pull_request.target_repo,
                 user=pull_request.author,
                 pull_request=pull_request,
                 send_email=False, renderer=DEFAULT_COMMENTS_RENDERER)
 
             # Update status to "Under Review" for added commits
             for commit_id in commit_changes.added:
                 ChangesetStatusModel().set_status(
                     repo=pull_request.source_repo,
                     status=ChangesetStatus.STATUS_UNDER_REVIEW,
                     comment=update_comment,
                     user=pull_request.author,
                     pull_request=pull_request,
                     revision=commit_id)
 
         # initial commit
         Session().commit()
 
         if pr_has_changes:
             # send update email to users
             try:
                 self.notify_users(pull_request=pull_request, updating_user=updating_user,
                                   ancestor_commit_id=ancestor_commit_id,
                                   commit_changes=commit_changes,
                                   file_changes=file_changes)
                 Session().commit()
             except Exception:
                 log.exception('Failed to send email notification to users')
                 Session().rollback()
 
         log.debug(
             'Updated pull request %s, added_ids: %s, common_ids: %s, '
             'removed_ids: %s', pull_request.pull_request_id,
             commit_changes.added, commit_changes.common, commit_changes.removed)
         log.debug(
             'Updated pull request with the following file changes: %s',
             file_changes)
 
         log.info(
             "Updated pull request %s from commit %s to commit %s, "
             "stored new version %s of this pull request.",
             pull_request.pull_request_id, source_ref_id,
             pull_request.source_ref_parts.commit_id,
             pull_request_version.pull_request_version_id)
 
         self.trigger_pull_request_hook(pull_request, pull_request.author, 'update')
 
         return UpdateResponse(
             executed=True, reason=UpdateFailureReason.NONE,
             old=pull_request, new=pull_request_version,
             common_ancestor_id=ancestor_commit_id, commit_changes=commit_changes,
             source_changed=source_changed, target_changed=target_changed)
 
     def _create_version_from_snapshot(self, pull_request):
         version = PullRequestVersion()
         version.title = pull_request.title
         version.description = pull_request.description
         version.status = pull_request.status
         version.pull_request_state = pull_request.pull_request_state
         version.created_on = datetime.datetime.now()
         version.updated_on = pull_request.updated_on
         version.user_id = pull_request.user_id
         version.source_repo = pull_request.source_repo
         version.source_ref = pull_request.source_ref
         version.target_repo = pull_request.target_repo
         version.target_ref = pull_request.target_ref
 
         version._last_merge_source_rev = pull_request._last_merge_source_rev
         version._last_merge_target_rev = pull_request._last_merge_target_rev
         version.last_merge_status = pull_request.last_merge_status
         version.last_merge_metadata = pull_request.last_merge_metadata
         version.shadow_merge_ref = pull_request.shadow_merge_ref
         version.merge_rev = pull_request.merge_rev
         version.reviewer_data = pull_request.reviewer_data
 
         version.revisions = pull_request.revisions
         version.common_ancestor_id = pull_request.common_ancestor_id
         version.pull_request = pull_request
         Session().add(version)
         Session().flush()
 
         return version
 
     def _generate_update_diffs(self, pull_request, pull_request_version):
 
         diff_context = (
             self.DIFF_CONTEXT +
             CommentsModel.needed_extra_diff_context())
         hide_whitespace_changes = False
         source_repo = pull_request_version.source_repo
         source_ref_id = pull_request_version.source_ref_parts.commit_id
         target_ref_id = pull_request_version.target_ref_parts.commit_id
         old_diff = self._get_diff_from_pr_or_version(
             source_repo, source_ref_id, target_ref_id,
             hide_whitespace_changes=hide_whitespace_changes, diff_context=diff_context)
 
         source_repo = pull_request.source_repo
         source_ref_id = pull_request.source_ref_parts.commit_id
         target_ref_id = pull_request.target_ref_parts.commit_id
 
         new_diff = self._get_diff_from_pr_or_version(
             source_repo, source_ref_id, target_ref_id,
             hide_whitespace_changes=hide_whitespace_changes, diff_context=diff_context)
 
         old_diff_data = diffs.DiffProcessor(old_diff)
         old_diff_data.prepare()
         new_diff_data = diffs.DiffProcessor(new_diff)
         new_diff_data.prepare()
 
         return old_diff_data, new_diff_data
 
     def _link_comments_to_version(self, pull_request_version):
         """
         Link all unlinked comments of this pull request to the given version.
 
         :param pull_request_version: The `PullRequestVersion` to which
             the comments shall be linked.
 
         """
         pull_request = pull_request_version.pull_request
         comments = ChangesetComment.query()\
             .filter(
                 # TODO: johbo: Should we query for the repo at all here?
                 # Pending decision on how comments of PRs are to be related
                 # to either the source repo, the target repo or no repo at all.
                 ChangesetComment.repo_id == pull_request.target_repo.repo_id,
                 ChangesetComment.pull_request == pull_request,
                 ChangesetComment.pull_request_version == None)\
             .order_by(ChangesetComment.comment_id.asc())
 
         # TODO: johbo: Find out why this breaks if it is done in a bulk
         # operation.
         for comment in comments:
             comment.pull_request_version_id = (
                 pull_request_version.pull_request_version_id)
             Session().add(comment)
 
     def _calculate_commit_id_changes(self, old_ids, new_ids):
         added = [x for x in new_ids if x not in old_ids]
         common = [x for x in new_ids if x in old_ids]
         removed = [x for x in old_ids if x not in new_ids]
         total = new_ids
         return ChangeTuple(added, common, removed, total)
 
     def _calculate_file_changes(self, old_diff_data, new_diff_data):
 
         old_files = OrderedDict()
         for diff_data in old_diff_data.parsed_diff:
             old_files[diff_data['filename']] = md5_safe(diff_data['raw_diff'])
 
         added_files = []
         modified_files = []
         removed_files = []
         for diff_data in new_diff_data.parsed_diff:
             new_filename = diff_data['filename']
             new_hash = md5_safe(diff_data['raw_diff'])
 
             old_hash = old_files.get(new_filename)
             if not old_hash:
                 # file is not present in old diff, we have to figure out from parsed diff
                 # operation ADD/REMOVE
                 operations_dict = diff_data['stats']['ops']
                 if diffs.DEL_FILENODE in operations_dict:
                     removed_files.append(new_filename)
                 else:
                     added_files.append(new_filename)
             else:
                 if new_hash != old_hash:
                     modified_files.append(new_filename)
                 # now remove a file from old, since we have seen it already
                 del old_files[new_filename]
 
         # removed files is when there are present in old, but not in NEW,
         # since we remove old files that are present in new diff, left-overs
         # if any should be the removed files
         removed_files.extend(old_files.keys())
 
         return FileChangeTuple(added_files, modified_files, removed_files)
 
     def _render_update_message(self, ancestor_commit_id, changes, file_changes):
         """
         render the message using DEFAULT_COMMENTS_RENDERER (RST renderer),
         so it's always looking the same disregarding on which default
         renderer system is using.
 
         :param ancestor_commit_id: ancestor raw_id
         :param changes: changes named tuple
         :param file_changes: file changes named tuple
 
         """
         new_status = ChangesetStatus.get_status_lbl(
             ChangesetStatus.STATUS_UNDER_REVIEW)
 
         changed_files = (
             file_changes.added + file_changes.modified + file_changes.removed)
 
         params = {
             'under_review_label': new_status,
             'added_commits': changes.added,
             'removed_commits': changes.removed,
             'changed_files': changed_files,
             'added_files': file_changes.added,
             'modified_files': file_changes.modified,
             'removed_files': file_changes.removed,
             'ancestor_commit_id': ancestor_commit_id
         }
         renderer = RstTemplateRenderer()
         return renderer.render('pull_request_update.mako', **params)
 
     def edit(self, pull_request, title, description, description_renderer, user):
         pull_request = self.__get_pull_request(pull_request)
         old_data = pull_request.get_api_data(with_merge_state=False)
         if pull_request.is_closed():
             raise ValueError('This pull request is closed')
         if title:
             pull_request.title = title
         pull_request.description = description
         pull_request.updated_on = datetime.datetime.now()
         pull_request.description_renderer = description_renderer
         Session().add(pull_request)
         self._log_audit_action(
             'repo.pull_request.edit', {'old_data': old_data},
             user, pull_request)
 
     def update_reviewers(self, pull_request, reviewer_data, user):
         """
         Update the reviewers in the pull request
 
         :param pull_request: the pr to update
         :param reviewer_data: list of tuples
             [(user, ['reason1', 'reason2'], mandatory_flag, role, [rules])]
         :param user: current use who triggers this action
         """
 
         pull_request = self.__get_pull_request(pull_request)
         if pull_request.is_closed():
             raise ValueError('This pull request is closed')
 
         reviewers = {}
         for user_id, reasons, mandatory, role, rules in reviewer_data:
             if isinstance(user_id, (int, str)):
                 user_id = self._get_user(user_id).user_id
             reviewers[user_id] = {
                 'reasons': reasons, 'mandatory': mandatory, 'role': role}
 
         reviewers_ids = set(reviewers.keys())
         current_reviewers = PullRequestReviewers.get_pull_request_reviewers(
             pull_request.pull_request_id, role=PullRequestReviewers.ROLE_REVIEWER)
 
         current_reviewers_ids = set([x.user.user_id for x in current_reviewers])
 
         ids_to_add = reviewers_ids.difference(current_reviewers_ids)
         ids_to_remove = current_reviewers_ids.difference(reviewers_ids)
 
         log.debug("Adding %s reviewers", ids_to_add)
         log.debug("Removing %s reviewers", ids_to_remove)
         changed = False
         added_audit_reviewers = []
         removed_audit_reviewers = []
 
         for uid in ids_to_add:
             changed = True
             _usr = self._get_user(uid)
             reviewer = PullRequestReviewers()
             reviewer.user = _usr
             reviewer.pull_request = pull_request
             reviewer.reasons = reviewers[uid]['reasons']
             # NOTE(marcink): mandatory shouldn't be changed now
             # reviewer.mandatory = reviewers[uid]['reasons']
             # NOTE(marcink): role should be hardcoded, so we won't edit it.
             reviewer.role = PullRequestReviewers.ROLE_REVIEWER
             Session().add(reviewer)
             added_audit_reviewers.append(reviewer.get_dict())
 
         for uid in ids_to_remove:
             changed = True
             # NOTE(marcink): we fetch "ALL" reviewers objects using .all().
             # This is an edge case that handles previous state of having the same reviewer twice.
             # this CAN happen due to the lack of DB checks
             reviewers = PullRequestReviewers.query()\
                 .filter(PullRequestReviewers.user_id == uid,
                         PullRequestReviewers.role == PullRequestReviewers.ROLE_REVIEWER,
                         PullRequestReviewers.pull_request == pull_request)\
                 .all()
 
             for obj in reviewers:
                 added_audit_reviewers.append(obj.get_dict())
                 Session().delete(obj)
 
         if changed:
             Session().expire_all()
             pull_request.updated_on = datetime.datetime.now()
             Session().add(pull_request)
 
         # finally store audit logs
         for user_data in added_audit_reviewers:
             self._log_audit_action(
                 'repo.pull_request.reviewer.add', {'data': user_data},
                 user, pull_request)
         for user_data in removed_audit_reviewers:
             self._log_audit_action(
                 'repo.pull_request.reviewer.delete', {'old_data': user_data},
                 user, pull_request)
 
         self.notify_reviewers(pull_request, ids_to_add, user)
         return ids_to_add, ids_to_remove
 
     def update_observers(self, pull_request, observer_data, user):
         """
         Update the observers in the pull request
 
         :param pull_request: the pr to update
         :param observer_data: list of tuples
             [(user, ['reason1', 'reason2'], mandatory_flag, role, [rules])]
         :param user: current use who triggers this action
         """
         pull_request = self.__get_pull_request(pull_request)
         if pull_request.is_closed():
             raise ValueError('This pull request is closed')
 
         observers = {}
         for user_id, reasons, mandatory, role, rules in observer_data:
             if isinstance(user_id, (int, str)):
                 user_id = self._get_user(user_id).user_id
             observers[user_id] = {
                 'reasons': reasons, 'observers': mandatory, 'role': role}
 
         observers_ids = set(observers.keys())
         current_observers = PullRequestReviewers.get_pull_request_reviewers(
             pull_request.pull_request_id, role=PullRequestReviewers.ROLE_OBSERVER)
 
         current_observers_ids = set([x.user.user_id for x in current_observers])
 
         ids_to_add = observers_ids.difference(current_observers_ids)
         ids_to_remove = current_observers_ids.difference(observers_ids)
 
         log.debug("Adding %s observer", ids_to_add)
         log.debug("Removing %s observer", ids_to_remove)
         changed = False
         added_audit_observers = []
         removed_audit_observers = []
 
         for uid in ids_to_add:
             changed = True
             _usr = self._get_user(uid)
             observer = PullRequestReviewers()
             observer.user = _usr
             observer.pull_request = pull_request
             observer.reasons = observers[uid]['reasons']
             # NOTE(marcink): mandatory shouldn't be changed now
             # observer.mandatory = observer[uid]['reasons']
 
             # NOTE(marcink): role should be hardcoded, so we won't edit it.
             observer.role = PullRequestReviewers.ROLE_OBSERVER
             Session().add(observer)
             added_audit_observers.append(observer.get_dict())
 
         for uid in ids_to_remove:
             changed = True
             # NOTE(marcink): we fetch "ALL" reviewers objects using .all().
             # This is an edge case that handles previous state of having the same reviewer twice.
             # this CAN happen due to the lack of DB checks
             observers = PullRequestReviewers.query()\
                 .filter(PullRequestReviewers.user_id == uid,
                         PullRequestReviewers.role == PullRequestReviewers.ROLE_OBSERVER,
                         PullRequestReviewers.pull_request == pull_request)\
                 .all()
 
             for obj in observers:
                 added_audit_observers.append(obj.get_dict())
                 Session().delete(obj)
 
         if changed:
             Session().expire_all()
             pull_request.updated_on = datetime.datetime.now()
             Session().add(pull_request)
 
         # finally store audit logs
         for user_data in added_audit_observers:
             self._log_audit_action(
                 'repo.pull_request.observer.add', {'data': user_data},
                 user, pull_request)
         for user_data in removed_audit_observers:
             self._log_audit_action(
                 'repo.pull_request.observer.delete', {'old_data': user_data},
                 user, pull_request)
 
         self.notify_observers(pull_request, ids_to_add, user)
         return ids_to_add, ids_to_remove
 
     def get_url(self, pull_request, request=None, permalink=False):
         if not request:
             request = get_current_request()
 
         if permalink:
             return request.route_url(
                 'pull_requests_global',
                 pull_request_id=pull_request.pull_request_id,)
         else:
             return request.route_url('pullrequest_show',
                 repo_name=safe_str(pull_request.target_repo.repo_name),
                 pull_request_id=pull_request.pull_request_id,)
 
     def get_shadow_clone_url(self, pull_request, request=None):
         """
         Returns qualified url pointing to the shadow repository. If this pull
         request is closed there is no shadow repository and ``None`` will be
         returned.
         """
         if pull_request.is_closed():
             return None
         else:
             pr_url = urllib.parse.unquote(self.get_url(pull_request, request=request))
             return safe_unicode('{pr_url}/repository'.format(pr_url=pr_url))
 
     def _notify_reviewers(self, pull_request, user_ids, role, user):
         # notification to reviewers/observers
         if not user_ids:
             return
 
         log.debug('Notify following %s users about pull-request %s', role, user_ids)
 
         pull_request_obj = pull_request
         # get the current participants of this pull request
         recipients = user_ids
         notification_type = EmailNotificationModel.TYPE_PULL_REQUEST
 
         pr_source_repo = pull_request_obj.source_repo
         pr_target_repo = pull_request_obj.target_repo
 
         pr_url = h.route_url('pullrequest_show',
                              repo_name=pr_target_repo.repo_name,
                              pull_request_id=pull_request_obj.pull_request_id,)
 
         # set some variables for email notification
         pr_target_repo_url = h.route_url(
             'repo_summary', repo_name=pr_target_repo.repo_name)
 
         pr_source_repo_url = h.route_url(
             'repo_summary', repo_name=pr_source_repo.repo_name)
 
         # pull request specifics
         pull_request_commits = [
             (x.raw_id, x.message)
             for x in map(pr_source_repo.get_commit, pull_request.revisions)]
 
         current_rhodecode_user = user
         kwargs = {
             'user': current_rhodecode_user,
             'pull_request_author': pull_request.author,
             'pull_request': pull_request_obj,
             'pull_request_commits': pull_request_commits,
 
             'pull_request_target_repo': pr_target_repo,
             'pull_request_target_repo_url': pr_target_repo_url,
 
             'pull_request_source_repo': pr_source_repo,
             'pull_request_source_repo_url': pr_source_repo_url,
 
             'pull_request_url': pr_url,
             'thread_ids': [pr_url],
             'user_role': role
         }
 
         # create notification objects, and emails
         NotificationModel().create(
             created_by=current_rhodecode_user,
             notification_subject='',  # Filled in based on the notification_type
             notification_body='',  # Filled in based on the notification_type
             notification_type=notification_type,
             recipients=recipients,
             email_kwargs=kwargs,
         )
 
     def notify_reviewers(self, pull_request, reviewers_ids, user):
         return self._notify_reviewers(pull_request, reviewers_ids,
                                       PullRequestReviewers.ROLE_REVIEWER, user)
 
     def notify_observers(self, pull_request, observers_ids, user):
         return self._notify_reviewers(pull_request, observers_ids,
                                       PullRequestReviewers.ROLE_OBSERVER, user)
 
     def notify_users(self, pull_request, updating_user, ancestor_commit_id,
                      commit_changes, file_changes):
 
         updating_user_id = updating_user.user_id
         reviewers = set([x.user.user_id for x in pull_request.get_pull_request_reviewers()])
         # NOTE(marcink): send notification to all other users except to
         # person who updated the PR
         recipients = reviewers.difference(set([updating_user_id]))
 
         log.debug('Notify following recipients about pull-request update %s', recipients)
 
         pull_request_obj = pull_request
 
         # send email about the update
         changed_files = (
                 file_changes.added + file_changes.modified + file_changes.removed)
 
         pr_source_repo = pull_request_obj.source_repo
         pr_target_repo = pull_request_obj.target_repo
 
         pr_url = h.route_url('pullrequest_show',
                              repo_name=pr_target_repo.repo_name,
                              pull_request_id=pull_request_obj.pull_request_id,)
 
         # set some variables for email notification
         pr_target_repo_url = h.route_url(
             'repo_summary', repo_name=pr_target_repo.repo_name)
 
         pr_source_repo_url = h.route_url(
             'repo_summary', repo_name=pr_source_repo.repo_name)
 
         email_kwargs = {
             'date': datetime.datetime.now(),
             'updating_user': updating_user,
 
             'pull_request': pull_request_obj,
 
             'pull_request_target_repo': pr_target_repo,
             'pull_request_target_repo_url': pr_target_repo_url,
 
             'pull_request_source_repo': pr_source_repo,
             'pull_request_source_repo_url': pr_source_repo_url,
 
             'pull_request_url': pr_url,
 
             'ancestor_commit_id': ancestor_commit_id,
             'added_commits': commit_changes.added,
             'removed_commits': commit_changes.removed,
             'changed_files': changed_files,
             'added_files': file_changes.added,
             'modified_files': file_changes.modified,
             'removed_files': file_changes.removed,
             'thread_ids': [pr_url],
         }
 
         # create notification objects, and emails
         NotificationModel().create(
             created_by=updating_user,
             notification_subject='',  # Filled in based on the notification_type
             notification_body='',  # Filled in based on the notification_type
             notification_type=EmailNotificationModel.TYPE_PULL_REQUEST_UPDATE,
             recipients=recipients,
             email_kwargs=email_kwargs,
         )
 
     def delete(self, pull_request, user=None):
         if not user:
             user = getattr(get_current_rhodecode_user(), 'username', None)
 
         pull_request = self.__get_pull_request(pull_request)
         old_data = pull_request.get_api_data(with_merge_state=False)
         self._cleanup_merge_workspace(pull_request)
         self._log_audit_action(
             'repo.pull_request.delete', {'old_data': old_data},
             user, pull_request)
         Session().delete(pull_request)
 
     def close_pull_request(self, pull_request, user):
         pull_request = self.__get_pull_request(pull_request)
         self._cleanup_merge_workspace(pull_request)
         pull_request.status = PullRequest.STATUS_CLOSED
         pull_request.updated_on = datetime.datetime.now()
         Session().add(pull_request)
         self.trigger_pull_request_hook(pull_request, pull_request.author, 'close')
 
         pr_data = pull_request.get_api_data(with_merge_state=False)
         self._log_audit_action(
             'repo.pull_request.close', {'data': pr_data}, user, pull_request)
 
     def close_pull_request_with_comment(
             self, pull_request, user, repo, message=None, auth_user=None):
 
         pull_request_review_status = pull_request.calculated_review_status()
 
         if pull_request_review_status == ChangesetStatus.STATUS_APPROVED:
             # approved only if we have voting consent
             status = ChangesetStatus.STATUS_APPROVED
         else:
             status = ChangesetStatus.STATUS_REJECTED
         status_lbl = ChangesetStatus.get_status_lbl(status)
 
         default_message = (
             'Closing with status change {transition_icon} {status}.'
         ).format(transition_icon='>', status=status_lbl)
         text = message or default_message
 
         # create a comment, and link it to new status
         comment = CommentsModel().create(
             text=text,
             repo=repo.repo_id,
             user=user.user_id,
             pull_request=pull_request.pull_request_id,
             status_change=status_lbl,
             status_change_type=status,
             closing_pr=True,
             auth_user=auth_user,
         )
 
         # calculate old status before we change it
         old_calculated_status = pull_request.calculated_review_status()
         ChangesetStatusModel().set_status(
             repo.repo_id,
             status,
             user.user_id,
             comment=comment,
             pull_request=pull_request.pull_request_id
         )
 
         Session().flush()
 
         self.trigger_pull_request_hook(pull_request, user, 'comment',
                                        data={'comment': comment})
 
         # we now calculate the status of pull request again, and based on that
         # calculation trigger status change. This might happen in cases
         # that non-reviewer admin closes a pr, which means his vote doesn't
         # change the status, while if he's a reviewer this might change it.
         calculated_status = pull_request.calculated_review_status()
         if old_calculated_status != calculated_status:
             self.trigger_pull_request_hook(pull_request, user, 'review_status_change',
                                            data={'status': calculated_status})
 
         # finally close the PR
         PullRequestModel().close_pull_request(pull_request.pull_request_id, user)
 
         return comment, status
 
     def merge_status(self, pull_request, translator=None, force_shadow_repo_refresh=False):
         _ = translator or get_current_request().translate
 
         if not self._is_merge_enabled(pull_request):
             return None, False, _('Server-side pull request merging is disabled.')
 
         if pull_request.is_closed():
             return None, False, _('This pull request is closed.')
 
         merge_possible, msg = self._check_repo_requirements(
             target=pull_request.target_repo, source=pull_request.source_repo,
             translator=_)
         if not merge_possible:
             return None, merge_possible, msg
 
         try:
             merge_response = self._try_merge(
                 pull_request, force_shadow_repo_refresh=force_shadow_repo_refresh)
             log.debug("Merge response: %s", merge_response)
             return merge_response, merge_response.possible, merge_response.merge_status_message
         except NotImplementedError:
             return None, False, _('Pull request merging is not supported.')
 
     def _check_repo_requirements(self, target, source, translator):
         """
         Check if `target` and `source` have compatible requirements.
 
         Currently this is just checking for largefiles.
         """
         _ = translator
         target_has_largefiles = self._has_largefiles(target)
         source_has_largefiles = self._has_largefiles(source)
         merge_possible = True
         message = u''
 
         if target_has_largefiles != source_has_largefiles:
             merge_possible = False
             if source_has_largefiles:
                 message = _(
                     'Target repository large files support is disabled.')
             else:
                 message = _(
                     'Source repository large files support is disabled.')
 
         return merge_possible, message
 
     def _has_largefiles(self, repo):
         largefiles_ui = VcsSettingsModel(repo=repo).get_ui_settings(
             'extensions', 'largefiles')
         return largefiles_ui and largefiles_ui[0].active
 
     def _try_merge(self, pull_request, force_shadow_repo_refresh=False):
         """
         Try to merge the pull request and return the merge status.
         """
         log.debug(
             "Trying out if the pull request %s can be merged. Force_refresh=%s",
             pull_request.pull_request_id, force_shadow_repo_refresh)
         target_vcs = pull_request.target_repo.scm_instance()
         # Refresh the target reference.
         try:
             target_ref = self._refresh_reference(
                 pull_request.target_ref_parts, target_vcs)
         except CommitDoesNotExistError:
             merge_state = MergeResponse(
                 False, False, None, MergeFailureReason.MISSING_TARGET_REF,
                 metadata={'target_ref': pull_request.target_ref_parts})
             return merge_state
 
         target_locked = pull_request.target_repo.locked
         if target_locked and target_locked[0]:
             locked_by = 'user:{}'.format(target_locked[0])
             log.debug("The target repository is locked by %s.", locked_by)
             merge_state = MergeResponse(
                 False, False, None, MergeFailureReason.TARGET_IS_LOCKED,
                 metadata={'locked_by': locked_by})
         elif force_shadow_repo_refresh or self._needs_merge_state_refresh(
                 pull_request, target_ref):
             log.debug("Refreshing the merge status of the repository.")
             merge_state = self._refresh_merge_state(
                 pull_request, target_vcs, target_ref)
         else:
             possible = pull_request.last_merge_status == MergeFailureReason.NONE
             metadata = {
                 'unresolved_files': '',
                 'target_ref': pull_request.target_ref_parts,
                 'source_ref': pull_request.source_ref_parts,
             }
             if pull_request.last_merge_metadata:
                 metadata.update(pull_request.last_merge_metadata_parsed)
 
             if not possible and target_ref.type == 'branch':
                 # NOTE(marcink): case for mercurial multiple heads on branch
                 heads = target_vcs._heads(target_ref.name)
                 if len(heads) != 1:
                     heads = '\n,'.join(target_vcs._heads(target_ref.name))
                     metadata.update({
                         'heads': heads
                     })
 
             merge_state = MergeResponse(
                 possible, False, None, pull_request.last_merge_status, metadata=metadata)
 
         return merge_state
 
     def _refresh_reference(self, reference, vcs_repository):
         if reference.type in self.UPDATABLE_REF_TYPES:
             name_or_id = reference.name
         else:
             name_or_id = reference.commit_id
 
         refreshed_commit = vcs_repository.get_commit(name_or_id)
         refreshed_reference = Reference(
             reference.type, reference.name, refreshed_commit.raw_id)
         return refreshed_reference
 
     def _needs_merge_state_refresh(self, pull_request, target_reference):
         return not(
             pull_request.revisions and
             pull_request.revisions[0] == pull_request._last_merge_source_rev and
             target_reference.commit_id == pull_request._last_merge_target_rev)
 
     def _refresh_merge_state(self, pull_request, target_vcs, target_reference):
         workspace_id = self._workspace_id(pull_request)
         source_vcs = pull_request.source_repo.scm_instance()
         repo_id = pull_request.target_repo.repo_id
         use_rebase = self._use_rebase_for_merging(pull_request)
         close_branch = self._close_branch_before_merging(pull_request)
         merge_state = target_vcs.merge(
             repo_id, workspace_id,
             target_reference, source_vcs, pull_request.source_ref_parts,
             dry_run=True, use_rebase=use_rebase,
             close_branch=close_branch)
 
         # Do not store the response if there was an unknown error.
         if merge_state.failure_reason != MergeFailureReason.UNKNOWN:
             pull_request._last_merge_source_rev = \
                 pull_request.source_ref_parts.commit_id
             pull_request._last_merge_target_rev = target_reference.commit_id
             pull_request.last_merge_status = merge_state.failure_reason
             pull_request.last_merge_metadata = merge_state.metadata
 
             pull_request.shadow_merge_ref = merge_state.merge_ref
             Session().add(pull_request)
             Session().commit()
 
         return merge_state
 
     def _workspace_id(self, pull_request):
         workspace_id = 'pr-%s' % pull_request.pull_request_id
         return workspace_id
 
     def generate_repo_data(self, repo, commit_id=None, branch=None,
                            bookmark=None, translator=None):
         from rhodecode.model.repo import RepoModel
 
         all_refs, selected_ref = \
             self._get_repo_pullrequest_sources(
                 repo.scm_instance(), commit_id=commit_id,
                 branch=branch, bookmark=bookmark, translator=translator)
 
         refs_select2 = []
         for element in all_refs:
             children = [{'id': x[0], 'text': x[1]} for x in element[0]]
             refs_select2.append({'text': element[1], 'children': children})
 
         return {
             'user': {
                 'user_id': repo.user.user_id,
                 'username': repo.user.username,
                 'firstname': repo.user.first_name,
                 'lastname': repo.user.last_name,
                 'gravatar_link': h.gravatar_url(repo.user.email, 14),
             },
             'name': repo.repo_name,
             'link': RepoModel().get_url(repo),
             'description': h.chop_at_smart(repo.description_safe, '\n'),
             'refs': {
                 'all_refs': all_refs,
                 'selected_ref': selected_ref,
                 'select2_refs': refs_select2
             }
         }
 
     def generate_pullrequest_title(self, source, source_ref, target):
         return u'{source}#{at_ref} to {target}'.format(
             source=source,
             at_ref=source_ref,
             target=target,
         )
 
     def _cleanup_merge_workspace(self, pull_request):
         # Merging related cleanup
         repo_id = pull_request.target_repo.repo_id
         target_scm = pull_request.target_repo.scm_instance()
         workspace_id = self._workspace_id(pull_request)
 
         try:
             target_scm.cleanup_merge_workspace(repo_id, workspace_id)
         except NotImplementedError:
             pass
 
     def _get_repo_pullrequest_sources(
             self, repo, commit_id=None, branch=None, bookmark=None,
             translator=None):
         """
         Return a structure with repo's interesting commits, suitable for
         the selectors in pullrequest controller
 
         :param commit_id: a commit that must be in the list somehow
             and selected by default
         :param branch: a branch that must be in the list and selected
             by default - even if closed
         :param bookmark: a bookmark that must be in the list and selected
         """
         _ = translator or get_current_request().translate
 
         commit_id = safe_str(commit_id) if commit_id else None
         branch = safe_unicode(branch) if branch else None
         bookmark = safe_unicode(bookmark) if bookmark else None
 
         selected = None
 
         # order matters: first source that has commit_id in it will be selected
         sources = []
         sources.append(('book', repo.bookmarks.items(), _('Bookmarks'), bookmark))
         sources.append(('branch', repo.branches.items(), _('Branches'), branch))
 
         if commit_id:
             ref_commit = (h.short_id(commit_id), commit_id)
             sources.append(('rev', [ref_commit], _('Commit IDs'), commit_id))
 
         sources.append(
             ('branch', repo.branches_closed.items(), _('Closed Branches'), branch),
         )
 
         groups = []
 
         for group_key, ref_list, group_name, match in sources:
             group_refs = []
             for ref_name, ref_id in ref_list:
                 ref_key = u'{}:{}:{}'.format(group_key, ref_name, ref_id)
                 group_refs.append((ref_key, ref_name))
 
                 if not selected:
                     if set([commit_id, match]) & set([ref_id, ref_name]):
                         selected = ref_key
 
             if group_refs:
                 groups.append((group_refs, group_name))
 
         if not selected:
             ref = commit_id or branch or bookmark
             if ref:
                 raise CommitDoesNotExistError(
                     u'No commit refs could be found matching: {}'.format(ref))
             elif repo.DEFAULT_BRANCH_NAME in repo.branches:
                 selected = u'branch:{}:{}'.format(
                     safe_unicode(repo.DEFAULT_BRANCH_NAME),
                     safe_unicode(repo.branches[repo.DEFAULT_BRANCH_NAME])
                 )
             elif repo.commit_ids:
                 # make the user select in this case
                 selected = None
             else:
                 raise EmptyRepositoryError()
         return groups, selected
 
     def get_diff(self, source_repo, source_ref_id, target_ref_id,
                  hide_whitespace_changes, diff_context):
 
         return self._get_diff_from_pr_or_version(
             source_repo, source_ref_id, target_ref_id,
             hide_whitespace_changes=hide_whitespace_changes, diff_context=diff_context)
 
     def _get_diff_from_pr_or_version(
             self, source_repo, source_ref_id, target_ref_id,
             hide_whitespace_changes, diff_context):
 
         target_commit = source_repo.get_commit(
             commit_id=safe_str(target_ref_id))
         source_commit = source_repo.get_commit(
             commit_id=safe_str(source_ref_id), maybe_unreachable=True)
         if isinstance(source_repo, Repository):
             vcs_repo = source_repo.scm_instance()
         else:
             vcs_repo = source_repo
 
         # TODO: johbo: In the context of an update, we cannot reach
         # the old commit anymore with our normal mechanisms. It needs
         # some sort of special support in the vcs layer to avoid this
         # workaround.
         if (source_commit.raw_id == vcs_repo.EMPTY_COMMIT_ID and
                 vcs_repo.alias == 'git'):
             source_commit.raw_id = safe_str(source_ref_id)
 
         log.debug('calculating diff between '
                   'source_ref:%s and target_ref:%s for repo `%s`',
                   target_ref_id, source_ref_id,
                   safe_unicode(vcs_repo.path))
 
         vcs_diff = vcs_repo.get_diff(
             commit1=target_commit, commit2=source_commit,
             ignore_whitespace=hide_whitespace_changes, context=diff_context)
         return vcs_diff
 
     def _is_merge_enabled(self, pull_request):
         return self._get_general_setting(
             pull_request, 'rhodecode_pr_merge_enabled')
 
     def _use_rebase_for_merging(self, pull_request):
         repo_type = pull_request.target_repo.repo_type
         if repo_type == 'hg':
             return self._get_general_setting(
                 pull_request, 'rhodecode_hg_use_rebase_for_merging')
         elif repo_type == 'git':
             return self._get_general_setting(
                 pull_request, 'rhodecode_git_use_rebase_for_merging')
 
         return False
 
     def _user_name_for_merging(self, pull_request, user):
         env_user_name_attr = os.environ.get('RC_MERGE_USER_NAME_ATTR', '')
         if env_user_name_attr and hasattr(user, env_user_name_attr):
             user_name_attr = env_user_name_attr
         else:
             user_name_attr = 'short_contact'
 
         user_name = getattr(user, user_name_attr)
         return user_name
 
     def _close_branch_before_merging(self, pull_request):
         repo_type = pull_request.target_repo.repo_type
         if repo_type == 'hg':
             return self._get_general_setting(
                 pull_request, 'rhodecode_hg_close_branch_before_merging')
         elif repo_type == 'git':
             return self._get_general_setting(
                 pull_request, 'rhodecode_git_close_branch_before_merging')
 
         return False
 
     def _get_general_setting(self, pull_request, settings_key, default=False):
         settings_model = VcsSettingsModel(repo=pull_request.target_repo)
         settings = settings_model.get_general_settings()
         return settings.get(settings_key, default)
 
     def _log_audit_action(self, action, action_data, user, pull_request):
         audit_logger.store(
             action=action,
             action_data=action_data,
             user=user,
             repo=pull_request.target_repo)
 
     def get_reviewer_functions(self):
         """
         Fetches functions for validation and fetching default reviewers.
         If available we use the EE package, else we fallback to CE
         package functions
         """
         try:
             from rc_reviewers.utils import get_default_reviewers_data
             from rc_reviewers.utils import validate_default_reviewers
             from rc_reviewers.utils import validate_observers
         except ImportError:
             from rhodecode.apps.repository.utils import get_default_reviewers_data
             from rhodecode.apps.repository.utils import validate_default_reviewers
             from rhodecode.apps.repository.utils import validate_observers
 
         return get_default_reviewers_data, validate_default_reviewers, validate_observers
 
 
 class MergeCheck(object):
     """
     Perform Merge Checks and returns a check object which stores information
     about merge errors, and merge conditions
     """
     TODO_CHECK = 'todo'
     PERM_CHECK = 'perm'
     REVIEW_CHECK = 'review'
     MERGE_CHECK = 'merge'
     WIP_CHECK = 'wip'
 
     def __init__(self):
         self.review_status = None
         self.merge_possible = None
         self.merge_msg = ''
         self.merge_response = None
         self.failed = None
         self.errors = []
         self.error_details = OrderedDict()
         self.source_commit = AttributeDict()
         self.target_commit = AttributeDict()
         self.reviewers_count = 0
         self.observers_count = 0
 
     def __repr__(self):
         return '<MergeCheck(possible:{}, failed:{}, errors:{})>'.format(
             self.merge_possible, self.failed, self.errors)
 
     def push_error(self, error_type, message, error_key, details):
         self.failed = True
         self.errors.append([error_type, message])
         self.error_details[error_key] = dict(
             details=details,
             error_type=error_type,
             message=message
         )
 
     @classmethod
     def validate(cls, pull_request, auth_user, translator, fail_early=False,
                  force_shadow_repo_refresh=False):
         _ = translator
         merge_check = cls()
 
         # title has WIP:
         if pull_request.work_in_progress:
             log.debug("MergeCheck: cannot merge, title has wip: marker.")
 
             msg = _('WIP marker in title prevents from accidental merge.')
             merge_check.push_error('error', msg, cls.WIP_CHECK, pull_request.title)
             if fail_early:
                 return merge_check
 
         # permissions to merge
         user_allowed_to_merge = PullRequestModel().check_user_merge(pull_request, auth_user)
         if not user_allowed_to_merge:
             log.debug("MergeCheck: cannot merge, approval is pending.")
 
             msg = _('User `{}` not allowed to perform merge.').format(auth_user.username)
             merge_check.push_error('error', msg, cls.PERM_CHECK, auth_user.username)
             if fail_early:
                 return merge_check
 
         # permission to merge into the target branch
         target_commit_id = pull_request.target_ref_parts.commit_id
         if pull_request.target_ref_parts.type == 'branch':
             branch_name = pull_request.target_ref_parts.name
         else:
             # for mercurial we can always figure out the branch from the commit
             # in case of bookmark
             target_commit = pull_request.target_repo.get_commit(target_commit_id)
             branch_name = target_commit.branch
 
         rule, branch_perm = auth_user.get_rule_and_branch_permission(
             pull_request.target_repo.repo_name, branch_name)
         if branch_perm and branch_perm == 'branch.none':
             msg = _('Target branch `{}` changes rejected by rule {}.').format(
                 branch_name, rule)
             merge_check.push_error('error', msg, cls.PERM_CHECK, auth_user.username)
             if fail_early:
                 return merge_check
 
         # review status, must be always present
         review_status = pull_request.calculated_review_status()
         merge_check.review_status = review_status
         merge_check.reviewers_count = pull_request.reviewers_count
         merge_check.observers_count = pull_request.observers_count
 
         status_approved = review_status == ChangesetStatus.STATUS_APPROVED
         if not status_approved and merge_check.reviewers_count:
             log.debug("MergeCheck: cannot merge, approval is pending.")
             msg = _('Pull request reviewer approval is pending.')
 
             merge_check.push_error('warning', msg, cls.REVIEW_CHECK, review_status)
 
             if fail_early:
                 return merge_check
 
         # left over TODOs
         todos = CommentsModel().get_pull_request_unresolved_todos(pull_request)
         if todos:
             log.debug("MergeCheck: cannot merge, {} "
                       "unresolved TODOs left.".format(len(todos)))
 
             if len(todos) == 1:
                 msg = _('Cannot merge, {} TODO still not resolved.').format(
                     len(todos))
             else:
                 msg = _('Cannot merge, {} TODOs still not resolved.').format(
                     len(todos))
 
             merge_check.push_error('warning', msg, cls.TODO_CHECK, todos)
 
             if fail_early:
                 return merge_check
 
         # merge possible, here is the filesystem simulation + shadow repo
         merge_response, merge_status, msg = PullRequestModel().merge_status(
             pull_request, translator=translator,
             force_shadow_repo_refresh=force_shadow_repo_refresh)
 
         merge_check.merge_possible = merge_status
         merge_check.merge_msg = msg
         merge_check.merge_response = merge_response
 
         source_ref_id = pull_request.source_ref_parts.commit_id
         target_ref_id = pull_request.target_ref_parts.commit_id
 
         try:
             source_commit, target_commit = PullRequestModel().get_flow_commits(pull_request)
             merge_check.source_commit.changed = source_ref_id != source_commit.raw_id
             merge_check.source_commit.ref_spec = pull_request.source_ref_parts
             merge_check.source_commit.current_raw_id = source_commit.raw_id
             merge_check.source_commit.previous_raw_id = source_ref_id
 
             merge_check.target_commit.changed = target_ref_id != target_commit.raw_id
             merge_check.target_commit.ref_spec = pull_request.target_ref_parts
             merge_check.target_commit.current_raw_id = target_commit.raw_id
             merge_check.target_commit.previous_raw_id = target_ref_id
         except (SourceRefMissing, TargetRefMissing):
             pass
 
         if not merge_status:
             log.debug("MergeCheck: cannot merge, pull request merge not possible.")
             merge_check.push_error('warning', msg, cls.MERGE_CHECK, None)
 
             if fail_early:
                 return merge_check
 
         log.debug('MergeCheck: is failed: %s', merge_check.failed)
         return merge_check
 
     @classmethod
     def get_merge_conditions(cls, pull_request, translator):
         _ = translator
         merge_details = {}
 
         model = PullRequestModel()
         use_rebase = model._use_rebase_for_merging(pull_request)
 
         if use_rebase:
             merge_details['merge_strategy'] = dict(
                 details={},
                 message=_('Merge strategy: rebase')
             )
         else:
             merge_details['merge_strategy'] = dict(
                 details={},
                 message=_('Merge strategy: explicit merge commit')
             )
 
         close_branch = model._close_branch_before_merging(pull_request)
         if close_branch:
             repo_type = pull_request.target_repo.repo_type
             close_msg = ''
             if repo_type == 'hg':
                 close_msg = _('Source branch will be closed before the merge.')
             elif repo_type == 'git':
                 close_msg = _('Source branch will be deleted after the merge.')
 
             merge_details['close_branch'] = dict(
                 details={},
                 message=close_msg
             )
 
         return merge_details
 
 
 ChangeTuple = collections.namedtuple(
     'ChangeTuple', ['added', 'common', 'removed', 'total'])
 
 FileChangeTuple = collections.namedtuple(
     'FileChangeTuple', ['added', 'modified', 'removed'])