fix(LFS): fixed pull_changes method to support a new (sync_large_objects) flag, updated all parts that will use a new flag. Fixes: RCCE-8

rhodecode/api/views/repo_api.py

@@ -1,2533 +1,2535 b''
 # Copyright (C) 2011-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import logging
 import time
 
 import rhodecode
 from rhodecode.api import (
     jsonrpc_method, JSONRPCError, JSONRPCForbidden, JSONRPCValidationError)
 from rhodecode.api.utils import (
     has_superadmin_permission, Optional, OAttr, get_repo_or_error,
     get_user_group_or_error, get_user_or_error, validate_repo_permissions,
     get_perm_or_error, parse_args, get_origin, build_commit_data,
     validate_set_owner_permissions)
 from rhodecode.lib import audit_logger, rc_cache, channelstream
 from rhodecode.lib import repo_maintenance
 from rhodecode.lib.auth import (
     HasPermissionAnyApi, HasUserGroupPermissionAnyApi,
     HasRepoPermissionAnyApi)
 from rhodecode.lib.celerylib.utils import get_task_id
 from rhodecode.lib.utils2 import (
     str2bool, time_to_datetime, safe_str, safe_int)
 from rhodecode.lib.ext_json import json
 from rhodecode.lib.exceptions import (
     StatusChangeOnClosedPullRequestError, CommentVersionMismatch)
 from rhodecode.lib.vcs import RepositoryError
 from rhodecode.lib.vcs.exceptions import NodeDoesNotExistError
 from rhodecode.model.changeset_status import ChangesetStatusModel
 from rhodecode.model.comment import CommentsModel
 from rhodecode.model.db import (
     Session, ChangesetStatus, RepositoryField, Repository, RepoGroup,
     ChangesetComment)
 from rhodecode.model.permission import PermissionModel
 from rhodecode.model.pull_request import PullRequestModel
 from rhodecode.model.repo import RepoModel
 from rhodecode.model.scm import ScmModel, RepoList
 from rhodecode.model.settings import SettingsModel, VcsSettingsModel
 from rhodecode.model import validation_schema
 from rhodecode.model.validation_schema.schemas import repo_schema
 
 log = logging.getLogger(__name__)
 
 
 @jsonrpc_method()
 def get_repo(request, apiuser, repoid, cache=Optional(True)):
     """
     Gets an existing repository by its name or repository_id.
 
     The members section so the output returns users groups or users
     associated with that repository.
 
     This command can only be run using an |authtoken| with admin rights,
     or users with at least read rights to the |repo|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository id.
     :type repoid: str or int
     :param cache: use the cached value for last changeset
     :type: cache: Optional(bool)
 
     Example output:
 
     .. code-block:: bash
 
         {
           "error": null,
           "id": <repo_id>,
           "result": {
             "clone_uri": null,
             "created_on": "timestamp",
             "description": "repo description",
             "enable_downloads": false,
             "enable_locking": false,
             "enable_statistics": false,
             "followers": [
               {
                 "active": true,
                 "admin": false,
                 "api_key": "****************************************",
                 "api_keys": [
                   "****************************************"
                 ],
                 "email": "[e-mail]",
                 "emails": [
                   "[e-mail]"
                 ],
                 "extern_name": "rhodecode",
                 "extern_type": "rhodecode",
                 "firstname": "username",
                 "ip_addresses": [],
                 "language": null,
                 "last_login": "2015-09-16T17:16:35.854",
                 "lastname": "surname",
                 "user_id": <user_id>,
                 "username": "name"
               }
             ],
             "fork_of": "parent-repo",
             "landing_rev": [
               "rev",
               "tip"
             ],
             "last_changeset": {
               "author": "User <[e-mail]>",
               "branch": "default",
               "date": "timestamp",
               "message": "last commit message",
               "parents": [
                 {
                   "raw_id": "commit-id"
                 }
               ],
               "raw_id": "commit-id",
               "revision": <revision number>,
               "short_id": "short id"
             },
             "lock_reason": null,
             "locked_by": null,
             "locked_date": null,
             "owner": "owner-name",
             "permissions": [
               {
                 "name": "super-admin-name",
                 "origin": "super-admin",
                 "permission": "repository.admin",
                 "type": "user"
               },
               {
                 "name": "owner-name",
                 "origin": "owner",
                 "permission": "repository.admin",
                 "type": "user"
               },
               {
                 "name": "user-group-name",
                 "origin": "permission",
                 "permission": "repository.write",
                 "type": "user_group"
               }
             ],
             "private": true,
             "repo_id": 676,
             "repo_name": "user-group/repo-name",
             "repo_type": "hg"
           }
         }
     """
 
     repo = get_repo_or_error(repoid)
     cache = Optional.extract(cache)
 
     include_secrets = False
     if has_superadmin_permission(apiuser):
         include_secrets = True
     else:
         # check if we have at least read permission for this repo !
         _perms = (
             'repository.admin', 'repository.write', 'repository.read',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     permissions = []
     for _user in repo.permissions():
         user_data = {
             'name': _user.username,
             'permission': _user.permission,
             'origin': get_origin(_user),
             'type': "user",
         }
         permissions.append(user_data)
 
     for _user_group in repo.permission_user_groups():
         user_group_data = {
             'name': _user_group.users_group_name,
             'permission': _user_group.permission,
             'origin': get_origin(_user_group),
             'type': "user_group",
         }
         permissions.append(user_group_data)
 
     following_users = [
         user.user.get_api_data(include_secrets=include_secrets)
         for user in repo.followers]
 
     if not cache:
         repo.update_commit_cache()
     data = repo.get_api_data(include_secrets=include_secrets)
     data['permissions'] = permissions
     data['followers'] = following_users
 
     return data
 
 
 @jsonrpc_method()
 def get_repos(request, apiuser, root=Optional(None), traverse=Optional(True)):
     """
     Lists all existing repositories.
 
     This command can only be run using an |authtoken| with admin rights,
     or users with at least read rights to |repos|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param root: specify root repository group to fetch repositories.
         filters the returned repositories to be members of given root group.
     :type root: Optional(None)
     :param traverse: traverse given root into subrepositories. With this flag
         set to False, it will only return top-level repositories from `root`.
         if root is empty it will return just top-level repositories.
     :type traverse: Optional(True)
 
 
     Example output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result: [
                   {
                     "repo_id" :          "<repo_id>",
                     "repo_name" :        "<reponame>"
                     "repo_type" :        "<repo_type>",
                     "clone_uri" :        "<clone_uri>",
                     "private": :         "<bool>",
                     "created_on" :       "<datetimecreated>",
                     "description" :      "<description>",
                     "landing_rev":       "<landing_rev>",
                     "owner":             "<repo_owner>",
                     "fork_of":           "<name_of_fork_parent>",
                     "enable_downloads":  "<bool>",
                     "enable_locking":    "<bool>",
                     "enable_statistics": "<bool>",
                   },
                   ...
                 ]
         error:  null
     """
 
     include_secrets = has_superadmin_permission(apiuser)
     _perms = ('repository.read', 'repository.write', 'repository.admin',)
     extras = {'user': apiuser}
 
     root = Optional.extract(root)
     traverse = Optional.extract(traverse, binary=True)
 
     if root:
         # verify parent existance, if it's empty return an error
         parent = RepoGroup.get_by_group_name(root)
         if not parent:
             raise JSONRPCError(
                 f'Root repository group `{root}` does not exist')
 
         if traverse:
             repos = RepoModel().get_repos_for_root(root=root, traverse=traverse)
         else:
             repos = RepoModel().get_repos_for_root(root=parent)
     else:
         if traverse:
             repos = RepoModel().get_all()
         else:
             # return just top-level
             repos = RepoModel().get_repos_for_root(root=None)
 
     repo_list = RepoList(repos, perm_set=_perms, extra_kwargs=extras)
     return [repo.get_api_data(include_secrets=include_secrets)
             for repo in repo_list]
 
 
 @jsonrpc_method()
 def get_repo_changeset(request, apiuser, repoid, revision,
                        details=Optional('basic')):
     """
     Returns information about a changeset.
 
     Additionally parameters define the amount of details returned by
     this function.
 
     This command can only be run using an |authtoken| with admin rights,
     or users with at least read rights to the |repo|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository id
     :type repoid: str or int
     :param revision: revision for which listing should be done
     :type revision: str
     :param details: details can be 'basic|extended|full' full gives diff
         info details like the diff itself, and number of changed files etc.
     :type details: Optional(str)
 
     """
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin', 'repository.write', 'repository.read',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     changes_details = Optional.extract(details)
     _changes_details_types = ['basic', 'extended', 'full']
     if changes_details not in _changes_details_types:
         raise JSONRPCError(
             'ret_type must be one of %s' % (
                 ','.join(_changes_details_types)))
 
     vcs_repo = repo.scm_instance()
     pre_load = ['author', 'branch', 'date', 'message', 'parents',
                 'status', '_commit', '_file_paths']
 
     try:
         commit = repo.get_commit(commit_id=revision, pre_load=pre_load)
     except TypeError as e:
         raise JSONRPCError(safe_str(e))
     _cs_json = commit.__json__()
     _cs_json['diff'] = build_commit_data(vcs_repo, commit, changes_details)
     if changes_details == 'full':
         _cs_json['refs'] = commit._get_refs()
     return _cs_json
 
 
 @jsonrpc_method()
 def get_repo_changesets(request, apiuser, repoid, start_rev, limit,
                         details=Optional('basic')):
     """
     Returns a set of commits limited by the number starting
     from the `start_rev` option.
 
     Additional parameters define the amount of details returned by this
     function.
 
     This command can only be run using an |authtoken| with admin rights,
     or users with at least read rights to |repos|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository ID.
     :type repoid: str or int
     :param start_rev: The starting revision from where to get changesets.
     :type start_rev: str
     :param limit: Limit the number of commits to this amount
     :type limit: str or int
     :param details: Set the level of detail returned. Valid option are:
         ``basic``, ``extended`` and ``full``.
     :type details: Optional(str)
 
     .. note::
 
        Setting the parameter `details` to the value ``full`` is extensive
        and returns details like the diff itself, and the number
        of changed files.
 
     """
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin', 'repository.write', 'repository.read',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     changes_details = Optional.extract(details)
     _changes_details_types = ['basic', 'extended', 'full']
     if changes_details not in _changes_details_types:
         raise JSONRPCError(
             'ret_type must be one of %s' % (
                 ','.join(_changes_details_types)))
 
     limit = int(limit)
     pre_load = ['author', 'branch', 'date', 'message', 'parents',
                 'status', '_commit', '_file_paths']
 
     vcs_repo = repo.scm_instance()
     # SVN needs a special case to distinguish its index and commit id
     if vcs_repo and vcs_repo.alias == 'svn' and (start_rev == '0'):
         start_rev = vcs_repo.commit_ids[0]
 
     try:
         commits = vcs_repo.get_commits(
             start_id=start_rev, pre_load=pre_load, translate_tags=False)
     except TypeError as e:
         raise JSONRPCError(safe_str(e))
     except Exception:
         log.exception('Fetching of commits failed')
         raise JSONRPCError('Error occurred during commit fetching')
 
     ret = []
     for cnt, commit in enumerate(commits):
         if cnt >= limit != -1:
             break
         _cs_json = commit.__json__()
         _cs_json['diff'] = build_commit_data(vcs_repo, commit, changes_details)
         if changes_details == 'full':
             _cs_json['refs'] = {
                 'branches': [commit.branch],
                 'bookmarks': getattr(commit, 'bookmarks', []),
                 'tags': commit.tags
             }
         ret.append(_cs_json)
     return ret
 
 
 @jsonrpc_method()
 def get_repo_nodes(request, apiuser, repoid, revision, root_path,
                    ret_type=Optional('all'), details=Optional('basic'),
                    max_file_bytes=Optional(None)):
     """
     Returns a list of nodes and children in a flat list for a given
     path at given revision.
 
     It's possible to specify ret_type to show only `files` or `dirs`.
 
     This command can only be run using an |authtoken| with admin rights,
     or users with at least read rights to |repos|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository ID.
     :type repoid: str or int
     :param revision: The revision for which listing should be done.
     :type revision: str
     :param root_path: The path from which to start displaying.
     :type root_path: str
     :param ret_type: Set the return type. Valid options are
         ``all`` (default), ``files`` and ``dirs``.
     :type ret_type: Optional(str)
     :param details: Returns extended information about nodes, such as
         md5, binary, and or content.
         The valid options are ``basic`` and ``full``.
     :type details: Optional(str)
     :param max_file_bytes: Only return file content under this file size bytes
     :type details: Optional(int)
 
     Example output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result: [
                     {
                       "binary": false,
                       "content": "File line",
                       "extension": "md",
                       "lines": 2,
                       "md5": "059fa5d29b19c0657e384749480f6422",
                       "mimetype": "text/x-minidsrc",
                       "name": "file.md",
                       "size": 580,
                       "type": "file"
                     },
                   ...
                 ]
         error:  null
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin', 'repository.write', 'repository.read',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     ret_type = Optional.extract(ret_type)
     details = Optional.extract(details)
     max_file_bytes = Optional.extract(max_file_bytes)
 
     _extended_types = ['basic', 'full']
     if details not in _extended_types:
         ret_types = ','.join(_extended_types)
         raise JSONRPCError(f'ret_type must be one of {ret_types}')
 
     extended_info = False
     content = False
     if details == 'basic':
         extended_info = True
 
     if details == 'full':
         extended_info = content = True
 
     _map = {}
     try:
         # check if repo is not empty by any chance, skip quicker if it is.
         _scm = repo.scm_instance()
         if _scm.is_empty():
             return []
 
         _d, _f = ScmModel().get_nodes(
             repo, revision, root_path, flat=False,
             extended_info=extended_info, content=content,
             max_file_bytes=max_file_bytes)
 
         _map = {
             'all': _d + _f,
             'files': _f,
             'dirs': _d,
         }
 
         return _map[ret_type]
     except KeyError:
         keys = ','.join(sorted(_map.keys()))
         raise JSONRPCError(f'ret_type must be one of {keys}')
     except Exception:
         log.exception("Exception occurred while trying to get repo nodes")
         raise JSONRPCError(f'failed to get repo: `{repo.repo_name}` nodes')
 
 
 @jsonrpc_method()
 def get_repo_file(request, apiuser, repoid, commit_id, file_path,
                   max_file_bytes=Optional(0), details=Optional('basic'),
                   cache=Optional(True)):
     """
     Returns a single file from repository at given revision.
 
     This command can only be run using an |authtoken| with admin rights,
     or users with at least read rights to |repos|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository ID.
     :type repoid: str or int
     :param commit_id: The revision for which listing should be done.
     :type commit_id: str
     :param file_path: The path from which to start displaying.
     :type file_path: str
     :param details: Returns different set of information about nodes.
         The valid options are ``minimal`` ``basic`` and ``full``.
     :type details: Optional(str)
     :param max_file_bytes: Only return file content under this file size bytes
     :type max_file_bytes: Optional(int)
     :param cache: Use internal caches for fetching files. If disabled fetching
         files is slower but more memory efficient
     :type cache: Optional(bool)
 
     Example output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result: {
             "binary": false,
             "extension": "py",
             "lines": 35,
             "content": "....",
             "md5": "76318336366b0f17ee249e11b0c99c41",
             "mimetype": "text/x-python",
             "name": "python.py",
             "size": 817,
             "type": "file",
         }
         error:  null
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin', 'repository.write', 'repository.read',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     cache = Optional.extract(cache, binary=True)
     details = Optional.extract(details)
     max_file_bytes = Optional.extract(max_file_bytes)
 
     _extended_types = ['minimal', 'minimal+search', 'basic', 'full']
     if details not in _extended_types:
         ret_types = ','.join(_extended_types)
         raise JSONRPCError(f'ret_type must be one of %s, got {ret_types}', details)
     extended_info = False
     content = False
 
     if details == 'minimal':
         extended_info = False
 
     elif details == 'basic':
         extended_info = True
 
     elif details == 'full':
         extended_info = content = True
 
     file_path = safe_str(file_path)
     try:
         # check if repo is not empty by any chance, skip quicker if it is.
         _scm = repo.scm_instance()
         if _scm.is_empty():
             return None
 
         node = ScmModel().get_node(
             repo, commit_id, file_path, extended_info=extended_info,
             content=content, max_file_bytes=max_file_bytes, cache=cache)
 
     except NodeDoesNotExistError:
         raise JSONRPCError(
             f'There is no file in repo: `{repo.repo_name}` at path `{file_path}` for commit: `{commit_id}`')
     except Exception:
         log.exception("Exception occurred while trying to get repo %s file",
                       repo.repo_name)
         raise JSONRPCError(f'failed to get repo: `{repo.repo_name}` file at path {file_path}')
 
     return node
 
 
 @jsonrpc_method()
 def get_repo_fts_tree(request, apiuser, repoid, commit_id, root_path):
     """
     Returns a list of tree nodes for path at given revision. This api is built
     strictly for usage in full text search building, and shouldn't be consumed
 
     This command can only be run using an |authtoken| with admin rights,
     or users with at least read rights to |repos|.
 
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin', 'repository.write', 'repository.read',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     repo_id = repo.repo_id
     cache_seconds = rhodecode.ConfigGet().get_int('rc_cache.cache_repo.expiration_time')
     cache_on = cache_seconds > 0
 
     cache_namespace_uid = f'repo.{rc_cache.FILE_TREE_CACHE_VER}.{repo_id}'
     rc_cache.get_or_create_region('cache_repo', cache_namespace_uid)
 
     def compute_fts_tree(repo_id, commit_id, root_path):
         return ScmModel().get_fts_data(repo_id, commit_id, root_path)
 
     try:
         # check if repo is not empty by any chance, skip quicker if it is.
         _scm = repo.scm_instance()
         if not _scm or _scm.is_empty():
             return []
     except RepositoryError:
         log.exception("Exception occurred while trying to get repo nodes")
         raise JSONRPCError(f'failed to get repo: `{repo.repo_name}` nodes')
 
     try:
         # we need to resolve commit_id to a FULL sha for cache to work correctly.
         # sending 'master' is a pointer that needs to be translated to current commit.
         commit_id = _scm.get_commit(commit_id=commit_id).raw_id
         log.debug(
             'Computing FTS REPO TREE for repo_id %s commit_id `%s` '
             'with caching: %s[TTL: %ss]' % (
                 repo_id, commit_id, cache_on, cache_seconds or 0))
 
         tree_files = compute_fts_tree(repo_id, commit_id, root_path)
 
         return tree_files
 
     except Exception:
         log.exception("Exception occurred while trying to get repo nodes")
         raise JSONRPCError('failed to get repo: `%s` nodes' % repo.repo_name)
 
 
 @jsonrpc_method()
 def get_repo_refs(request, apiuser, repoid):
     """
     Returns a dictionary of current references. It returns
     bookmarks, branches, closed_branches, and tags for given repository
 
     It's possible to specify ret_type to show only `files` or `dirs`.
 
     This command can only be run using an |authtoken| with admin rights,
     or users with at least read rights to |repos|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository ID.
     :type repoid: str or int
 
     Example output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         "result": {
             "bookmarks": {
               "dev": "5611d30200f4040ba2ab4f3d64e5b06408a02188",
               "master": "367f590445081d8ec8c2ea0456e73ae1f1c3d6cf"
             },
             "branches": {
               "default": "5611d30200f4040ba2ab4f3d64e5b06408a02188",
               "stable": "367f590445081d8ec8c2ea0456e73ae1f1c3d6cf"
             },
             "branches_closed": {},
             "tags": {
               "tip": "5611d30200f4040ba2ab4f3d64e5b06408a02188",
               "v4.4.0": "1232313f9e6adac5ce5399c2a891dc1e72b79022",
               "v4.4.1": "cbb9f1d329ae5768379cdec55a62ebdd546c4e27",
               "v4.4.2": "24ffe44a27fcd1c5b6936144e176b9f6dd2f3a17",
             }
         }
         error:  null
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin', 'repository.write', 'repository.read',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     try:
         # check if repo is not empty by any chance, skip quicker if it is.
         vcs_instance = repo.scm_instance()
         refs = vcs_instance.refs()
         return refs
     except Exception:
         log.exception("Exception occurred while trying to get repo refs")
         raise JSONRPCError(
             'failed to get repo: `%s` references' % repo.repo_name
         )
 
 
 @jsonrpc_method()
 def create_repo(
         request, apiuser, repo_name, repo_type,
         owner=Optional(OAttr('apiuser')),
         description=Optional(''),
         private=Optional(False),
         clone_uri=Optional(None),
         push_uri=Optional(None),
         landing_rev=Optional(None),
         enable_statistics=Optional(False),
         enable_locking=Optional(False),
         enable_downloads=Optional(False),
         copy_permissions=Optional(False)):
     """
     Creates a repository.
 
     * If the repository name contains "/", repository will be created inside
       a repository group or nested repository groups
 
       For example "foo/bar/repo1" will create |repo| called "repo1" inside
       group "foo/bar". You have to have permissions to access and write to
       the last repository group ("bar" in this example)
 
     This command can only be run using an |authtoken| with at least
     permissions to create repositories, or write permissions to
     parent repository groups.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repo_name: Set the repository name.
     :type repo_name: str
     :param repo_type: Set the repository type; 'hg','git', or 'svn'.
     :type repo_type: str
     :param owner: user_id or username
     :type owner: Optional(str)
     :param description: Set the repository description.
     :type description: Optional(str)
     :param private: set repository as private
     :type private: bool
     :param clone_uri: set clone_uri
     :type clone_uri: str
     :param push_uri: set push_uri
     :type push_uri: str
     :param landing_rev: <rev_type>:<rev>, e.g branch:default, book:dev, rev:abcd
     :type landing_rev: str
     :param enable_locking:
     :type enable_locking: bool
     :param enable_downloads:
     :type enable_downloads: bool
     :param enable_statistics:
     :type enable_statistics: bool
     :param copy_permissions: Copy permission from group in which the
         repository is being created.
     :type copy_permissions: bool
 
 
     Example output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result: {
                   "msg": "Created new repository `<reponame>`",
                   "success": true,
                   "task": "<celery task id or None if done sync>"
                 }
         error:  null
 
 
     Example error output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : null
       error :  {
          'failed to create repository `<repo_name>`'
       }
 
     """
 
     owner = validate_set_owner_permissions(apiuser, owner)
 
     description = Optional.extract(description)
     copy_permissions = Optional.extract(copy_permissions)
     clone_uri = Optional.extract(clone_uri)
     push_uri = Optional.extract(push_uri)
 
     defs = SettingsModel().get_default_repo_settings(strip_prefix=True)
     if isinstance(private, Optional):
         private = defs.get('repo_private') or Optional.extract(private)
     if isinstance(repo_type, Optional):
         repo_type = defs.get('repo_type')
     if isinstance(enable_statistics, Optional):
         enable_statistics = defs.get('repo_enable_statistics')
     if isinstance(enable_locking, Optional):
         enable_locking = defs.get('repo_enable_locking')
     if isinstance(enable_downloads, Optional):
         enable_downloads = defs.get('repo_enable_downloads')
 
     landing_ref, _label = ScmModel.backend_landing_ref(repo_type)
     ref_choices, _labels = ScmModel().get_repo_landing_revs(request.translate)
     ref_choices = list(set(ref_choices + [landing_ref]))
 
     landing_commit_ref = Optional.extract(landing_rev) or landing_ref
 
     schema = repo_schema.RepoSchema().bind(
         repo_type_options=rhodecode.BACKENDS.keys(),
         repo_ref_options=ref_choices,
         repo_type=repo_type,
         # user caller
         user=apiuser)
 
     try:
         schema_data = schema.deserialize(dict(
             repo_name=repo_name,
             repo_type=repo_type,
             repo_owner=owner.username,
             repo_description=description,
             repo_landing_commit_ref=landing_commit_ref,
             repo_clone_uri=clone_uri,
             repo_push_uri=push_uri,
             repo_private=private,
             repo_copy_permissions=copy_permissions,
             repo_enable_statistics=enable_statistics,
             repo_enable_downloads=enable_downloads,
             repo_enable_locking=enable_locking))
     except validation_schema.Invalid as err:
         raise JSONRPCValidationError(colander_exc=err)
 
     try:
         data = {
             'owner': owner,
             'repo_name': schema_data['repo_group']['repo_name_without_group'],
             'repo_name_full': schema_data['repo_name'],
             'repo_group': schema_data['repo_group']['repo_group_id'],
             'repo_type': schema_data['repo_type'],
             'repo_description': schema_data['repo_description'],
             'repo_private': schema_data['repo_private'],
             'clone_uri': schema_data['repo_clone_uri'],
             'push_uri': schema_data['repo_push_uri'],
             'repo_landing_rev': schema_data['repo_landing_commit_ref'],
             'enable_statistics': schema_data['repo_enable_statistics'],
             'enable_locking': schema_data['repo_enable_locking'],
             'enable_downloads': schema_data['repo_enable_downloads'],
             'repo_copy_permissions': schema_data['repo_copy_permissions'],
         }
 
         task = RepoModel().create(form_data=data, cur_user=owner.user_id)
         task_id = get_task_id(task)
         # no commit, it's done in RepoModel, or async via celery
         return {
             'msg': "Created new repository `{}`".format(schema_data['repo_name']),
             'success': True,  # cannot return the repo data here since fork
             # can be done async
             'task': task_id
         }
     except Exception:
         log.exception(
             "Exception while trying to create the repository %s",
             schema_data['repo_name'])
         raise JSONRPCError(
             'failed to create repository `{}`'.format(schema_data['repo_name']))
 
 
 @jsonrpc_method()
 def add_field_to_repo(request, apiuser, repoid, key, label=Optional(''),
                       description=Optional('')):
     """
     Adds an extra field to a repository.
 
     This command can only be run using an |authtoken| with at least
     write permissions to the |repo|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set the repository name or repository id.
     :type repoid: str or int
     :param key: Create a unique field key for this repository.
     :type key: str
     :param label:
     :type label: Optional(str)
     :param description:
     :type description: Optional(str)
     """
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     label = Optional.extract(label) or key
     description = Optional.extract(description)
 
     field = RepositoryField.get_by_key_name(key, repo)
     if field:
         raise JSONRPCError(f'Field with key `{key}` exists for repo `{repoid}`')
 
     try:
         RepoModel().add_repo_field(repo, key, field_label=label,
                                    field_desc=description)
         Session().commit()
         return {
             'msg': f"Added new repository field `{key}`",
             'success': True,
         }
     except Exception:
         log.exception("Exception occurred while trying to add field to repo")
         raise JSONRPCError(
             f'failed to create new field for repository `{repoid}`')
 
 
 @jsonrpc_method()
 def remove_field_from_repo(request, apiuser, repoid, key):
     """
     Removes an extra field from a repository.
 
     This command can only be run using an |authtoken| with at least
     write permissions to the |repo|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set the repository name or repository ID.
     :type repoid: str or int
     :param key: Set the unique field key for this repository.
     :type key: str
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     field = RepositoryField.get_by_key_name(key, repo)
     if not field:
         raise JSONRPCError('Field with key `%s` does not '
                            'exists for repo `%s`' % (key, repoid))
 
     try:
         RepoModel().delete_repo_field(repo, field_key=key)
         Session().commit()
         return {
             'msg': f"Deleted repository field `{key}`",
             'success': True,
         }
     except Exception:
         log.exception(
             "Exception occurred while trying to delete field from repo")
         raise JSONRPCError(
             f'failed to delete field for repository `{repoid}`')
 
 
 @jsonrpc_method()
 def update_repo(
         request, apiuser, repoid, repo_name=Optional(None),
         owner=Optional(OAttr('apiuser')), description=Optional(''),
         private=Optional(False),
         clone_uri=Optional(None), push_uri=Optional(None),
         landing_rev=Optional(None), fork_of=Optional(None),
         enable_statistics=Optional(False),
         enable_locking=Optional(False),
         enable_downloads=Optional(False), fields=Optional('')):
     r"""
     Updates a repository with the given information.
 
     This command can only be run using an |authtoken| with at least
     admin permissions to the |repo|.
 
     * If the repository name contains "/", repository will be updated
       accordingly with a repository group or nested repository groups
 
       For example repoid=repo-test name="foo/bar/repo-test" will update |repo|
       called "repo-test" and place it inside group "foo/bar".
       You have to have permissions to access and write to the last repository
       group ("bar" in this example)
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: repository name or repository ID.
     :type repoid: str or int
     :param repo_name: Update the |repo| name, including the
         repository group it's in.
     :type repo_name: str
     :param owner: Set the |repo| owner.
     :type owner: str
     :param fork_of: Set the |repo| as fork of another |repo|.
     :type fork_of: str
     :param description: Update the |repo| description.
     :type description: str
     :param private: Set the |repo| as private. (True | False)
     :type private: bool
     :param clone_uri: Update the |repo| clone URI.
     :type clone_uri: str
     :param landing_rev: Set the |repo| landing revision. e.g branch:default, book:dev, rev:abcd
     :type landing_rev: str
     :param enable_statistics: Enable statistics on the |repo|, (True | False).
     :type enable_statistics: bool
     :param enable_locking: Enable |repo| locking.
     :type enable_locking: bool
     :param enable_downloads: Enable downloads from the |repo|, (True | False).
     :type enable_downloads: bool
     :param fields: Add extra fields to the |repo|. Use the following
         example format: ``field_key=field_val,field_key2=fieldval2``.
         Escape ', ' with \,
     :type fields: str
     """
 
     repo = get_repo_or_error(repoid)
 
     include_secrets = False
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
     else:
         include_secrets = True
 
     updates = dict(
         repo_name=repo_name
         if not isinstance(repo_name, Optional) else repo.repo_name,
 
         fork_id=fork_of
         if not isinstance(fork_of, Optional) else repo.fork.repo_name if repo.fork else None,
 
         user=owner
         if not isinstance(owner, Optional) else repo.user.username,
 
         repo_description=description
         if not isinstance(description, Optional) else repo.description,
 
         repo_private=private
         if not isinstance(private, Optional) else repo.private,
 
         clone_uri=clone_uri
         if not isinstance(clone_uri, Optional) else repo.clone_uri,
 
         push_uri=push_uri
         if not isinstance(push_uri, Optional) else repo.push_uri,
 
         repo_landing_rev=landing_rev
         if not isinstance(landing_rev, Optional) else repo._landing_revision,
 
         repo_enable_statistics=enable_statistics
         if not isinstance(enable_statistics, Optional) else repo.enable_statistics,
 
         repo_enable_locking=enable_locking
         if not isinstance(enable_locking, Optional) else repo.enable_locking,
 
         repo_enable_downloads=enable_downloads
         if not isinstance(enable_downloads, Optional) else repo.enable_downloads)
 
     landing_ref, _label = ScmModel.backend_landing_ref(repo.repo_type)
     ref_choices, _labels = ScmModel().get_repo_landing_revs(
         request.translate, repo=repo)
     ref_choices = list(set(ref_choices + [landing_ref]))
 
     old_values = repo.get_api_data()
     repo_type = repo.repo_type
     schema = repo_schema.RepoSchema().bind(
         repo_type_options=rhodecode.BACKENDS.keys(),
         repo_ref_options=ref_choices,
         repo_type=repo_type,
         # user caller
         user=apiuser,
         old_values=old_values)
     try:
         schema_data = schema.deserialize(dict(
             # we save old value, users cannot change type
             repo_type=repo_type,
 
             repo_name=updates['repo_name'],
             repo_owner=updates['user'],
             repo_description=updates['repo_description'],
             repo_clone_uri=updates['clone_uri'],
             repo_push_uri=updates['push_uri'],
             repo_fork_of=updates['fork_id'],
             repo_private=updates['repo_private'],
             repo_landing_commit_ref=updates['repo_landing_rev'],
             repo_enable_statistics=updates['repo_enable_statistics'],
             repo_enable_downloads=updates['repo_enable_downloads'],
             repo_enable_locking=updates['repo_enable_locking']))
     except validation_schema.Invalid as err:
         raise JSONRPCValidationError(colander_exc=err)
 
     # save validated data back into the updates dict
     validated_updates = dict(
         repo_name=schema_data['repo_group']['repo_name_without_group'],
         repo_group=schema_data['repo_group']['repo_group_id'],
 
         user=schema_data['repo_owner'],
         repo_description=schema_data['repo_description'],
         repo_private=schema_data['repo_private'],
         clone_uri=schema_data['repo_clone_uri'],
         push_uri=schema_data['repo_push_uri'],
         repo_landing_rev=schema_data['repo_landing_commit_ref'],
         repo_enable_statistics=schema_data['repo_enable_statistics'],
         repo_enable_locking=schema_data['repo_enable_locking'],
         repo_enable_downloads=schema_data['repo_enable_downloads'],
     )
 
     if schema_data['repo_fork_of']:
         fork_repo = get_repo_or_error(schema_data['repo_fork_of'])
         validated_updates['fork_id'] = fork_repo.repo_id
 
     # extra fields
     fields = parse_args(Optional.extract(fields), key_prefix='ex_')
     if fields:
         validated_updates.update(fields)
 
     try:
         RepoModel().update(repo, **validated_updates)
         audit_logger.store_api(
             'repo.edit', action_data={'old_data': old_values},
             user=apiuser, repo=repo)
         Session().commit()
         return {
             'msg': f'updated repo ID:{repo.repo_id} {repo.repo_name}',
             'repository': repo.get_api_data(include_secrets=include_secrets)
         }
     except Exception:
         log.exception(
             "Exception while trying to update the repository %s",
             repoid)
         raise JSONRPCError('failed to update repo `%s`' % repoid)
 
 
 @jsonrpc_method()
 def fork_repo(request, apiuser, repoid, fork_name,
               owner=Optional(OAttr('apiuser')),
               description=Optional(''),
               private=Optional(False),
               clone_uri=Optional(None),
               landing_rev=Optional(None),
               copy_permissions=Optional(False)):
     """
     Creates a fork of the specified |repo|.
 
     * If the fork_name contains "/", fork will be created inside
       a repository group or nested repository groups
 
       For example "foo/bar/fork-repo" will create fork called "fork-repo"
       inside group "foo/bar". You have to have permissions to access and
       write to the last repository group ("bar" in this example)
 
     This command can only be run using an |authtoken| with minimum
     read permissions of the forked repo, create fork permissions for an user.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set repository name or repository ID.
     :type repoid: str or int
     :param fork_name: Set the fork name, including it's repository group membership.
     :type fork_name: str
     :param owner: Set the fork owner.
     :type owner: str
     :param description: Set the fork description.
     :type description: str
     :param copy_permissions: Copy permissions from parent |repo|. The
         default is False.
     :type copy_permissions: bool
     :param private: Make the fork private. The default is False.
     :type private: bool
     :param landing_rev: Set the landing revision. E.g branch:default, book:dev, rev:abcd
 
     Example output:
 
     .. code-block:: bash
 
         id : <id_for_response>
         api_key : "<api_key>"
         args:     {
                     "repoid" :          "<reponame or repo_id>",
                     "fork_name":        "<forkname>",
                     "owner":            "<username or user_id = Optional(=apiuser)>",
                     "description":      "<description>",
                     "copy_permissions": "<bool>",
                     "private":          "<bool>",
                     "landing_rev":      "<landing_rev>"
                   }
 
     Example error output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result: {
                   "msg": "Created fork of `<reponame>` as `<forkname>`",
                   "success": true,
                   "task": "<celery task id or None if done sync>"
                 }
         error:  null
 
     """
 
     repo = get_repo_or_error(repoid)
     repo_name = repo.repo_name
 
     if not has_superadmin_permission(apiuser):
         # check if we have at least read permission for
         # this repo that we fork !
         _perms = ('repository.admin', 'repository.write', 'repository.read')
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
         # check if the regular user has at least fork permissions as well
         if not HasPermissionAnyApi(PermissionModel.FORKING_ENABLED)(user=apiuser):
             raise JSONRPCForbidden()
 
     # check if user can set owner parameter
     owner = validate_set_owner_permissions(apiuser, owner)
 
     description = Optional.extract(description)
     copy_permissions = Optional.extract(copy_permissions)
     clone_uri = Optional.extract(clone_uri)
 
     landing_ref, _label = ScmModel.backend_landing_ref(repo.repo_type)
     ref_choices, _labels = ScmModel().get_repo_landing_revs(request.translate)
     ref_choices = list(set(ref_choices + [landing_ref]))
     landing_commit_ref = Optional.extract(landing_rev) or landing_ref
 
     private = Optional.extract(private)
 
     schema = repo_schema.RepoSchema().bind(
         repo_type_options=rhodecode.BACKENDS.keys(),
         repo_ref_options=ref_choices,
         repo_type=repo.repo_type,
         # user caller
         user=apiuser)
 
     try:
         schema_data = schema.deserialize(dict(
             repo_name=fork_name,
             repo_type=repo.repo_type,
             repo_owner=owner.username,
             repo_description=description,
             repo_landing_commit_ref=landing_commit_ref,
             repo_clone_uri=clone_uri,
             repo_private=private,
             repo_copy_permissions=copy_permissions))
     except validation_schema.Invalid as err:
         raise JSONRPCValidationError(colander_exc=err)
 
     try:
         data = {
             'fork_parent_id': repo.repo_id,
 
             'repo_name': schema_data['repo_group']['repo_name_without_group'],
             'repo_name_full': schema_data['repo_name'],
             'repo_group': schema_data['repo_group']['repo_group_id'],
             'repo_type': schema_data['repo_type'],
             'description': schema_data['repo_description'],
             'private': schema_data['repo_private'],
             'copy_permissions': schema_data['repo_copy_permissions'],
             'landing_rev': schema_data['repo_landing_commit_ref'],
         }
 
         task = RepoModel().create_fork(data, cur_user=owner.user_id)
         # no commit, it's done in RepoModel, or async via celery
         task_id = get_task_id(task)
 
         return {
             'msg': 'Created fork of `{}` as `{}`'.format(
                 repo.repo_name, schema_data['repo_name']),
             'success': True,  # cannot return the repo data here since fork
             # can be done async
             'task': task_id
         }
     except Exception:
         log.exception(
             "Exception while trying to create fork %s",
             schema_data['repo_name'])
         raise JSONRPCError(
             'failed to fork repository `{}` as `{}`'.format(
                 repo_name, schema_data['repo_name']))
 
 
 @jsonrpc_method()
 def delete_repo(request, apiuser, repoid, forks=Optional('')):
     """
     Deletes a repository.
 
     * When the `forks` parameter is set it's possible to detach or delete
       forks of deleted repository.
 
     This command can only be run using an |authtoken| with admin
     permissions on the |repo|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set the repository name or repository ID.
     :type repoid: str or int
     :param forks: Set to `detach` or `delete` forks from the |repo|.
     :type forks: Optional(str)
 
     Example error output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result: {
                   "msg": "Deleted repository `<reponame>`",
                   "success": true
                 }
         error:  null
     """
 
     repo = get_repo_or_error(repoid)
     repo_name = repo.repo_name
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     try:
         handle_forks = Optional.extract(forks)
         _forks_msg = ''
         _forks = [f for f in repo.forks]
         if handle_forks == 'detach':
             _forks_msg = ' ' + 'Detached %s forks' % len(_forks)
         elif handle_forks == 'delete':
             _forks_msg = ' ' + 'Deleted %s forks' % len(_forks)
         elif _forks:
             raise JSONRPCError(
                 'Cannot delete `%s` it still contains attached forks' %
                 (repo.repo_name,)
             )
         old_data = repo.get_api_data()
         RepoModel().delete(repo, forks=forks)
 
         repo = audit_logger.RepoWrap(repo_id=None,
                                      repo_name=repo.repo_name)
 
         audit_logger.store_api(
             'repo.delete', action_data={'old_data': old_data},
             user=apiuser, repo=repo)
 
         ScmModel().mark_for_invalidation(repo_name, delete=True)
         Session().commit()
         return {
             'msg': f'Deleted repository `{repo_name}`{_forks_msg}',
             'success': True
         }
     except Exception:
         log.exception("Exception occurred while trying to delete repo")
         raise JSONRPCError(
             f'failed to delete repository `{repo_name}`'
         )
 
 
 #TODO: marcink, change name ?
 @jsonrpc_method()
 def invalidate_cache(request, apiuser, repoid, delete_keys=Optional(False)):
     """
     Invalidates the cache for the specified repository.
 
     This command can only be run using an |authtoken| with admin rights to
     the specified repository.
 
     This command takes the following options:
 
     :param apiuser: This is filled automatically from |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Sets the repository name or repository ID.
     :type repoid: str or int
     :param delete_keys: This deletes the invalidated keys instead of
         just flagging them.
     :type delete_keys: Optional(``True`` | ``False``)
 
     Example output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : {
         'msg': Cache for repository `<repository name>` was invalidated,
         'repository': <repository name>
       }
       error :  null
 
     Example error output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : null
       error : {
          'Error occurred during cache invalidation action'
       }
 
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin', 'repository.write',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     delete = Optional.extract(delete_keys)
     try:
         ScmModel().mark_for_invalidation(repo.repo_name, delete=delete)
         return {
             'msg': f'Cache for repository `{repoid}` was invalidated',
             'repository': repo.repo_name
         }
     except Exception:
         log.exception(
             "Exception occurred while trying to invalidate repo cache")
         raise JSONRPCError(
             'Error occurred during cache invalidation action'
         )
 
 
 #TODO: marcink, change name ?
 @jsonrpc_method()
 def lock(request, apiuser, repoid, locked=Optional(None),
          userid=Optional(OAttr('apiuser'))):
     """
     Sets the lock state of the specified |repo| by the given user.
     From more information, see :ref:`repo-locking`.
 
     * If the ``userid`` option is not set, the repository is locked to the
       user who called the method.
     * If the ``locked`` parameter is not set, the current lock state of the
       repository is displayed.
 
     This command can only be run using an |authtoken| with admin rights to
     the specified repository.
 
     This command takes the following options:
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Sets the repository name or repository ID.
     :type repoid: str or int
     :param locked: Sets the lock state.
     :type locked: Optional(``True`` | ``False``)
     :param userid: Set the repository lock to this user.
     :type userid: Optional(str or int)
 
     Example error output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : {
         'repo': '<reponame>',
         'locked': <bool: lock state>,
         'locked_since': <int: lock timestamp>,
         'locked_by': <username of person who made the lock>,
         'lock_reason': <str: reason for locking>,
         'lock_state_changed': <bool: True if lock state has been changed in this request>,
         'msg': 'Repo `<reponame>` locked by `<username>` on <timestamp>.'
         or
         'msg': 'Repo `<repository name>` not locked.'
         or
         'msg': 'User `<user name>` set lock state for repo `<repository name>` to `<new lock state>`'
       }
       error :  null
 
     Example error output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : null
       error :  {
         'Error occurred locking repository `<reponame>`'
       }
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         # check if we have at least write permission for this repo !
         _perms = ('repository.admin', 'repository.write',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
         # make sure normal user does not pass someone else userid,
         # he is not allowed to do that
         if not isinstance(userid, Optional) and userid != apiuser.user_id:
             raise JSONRPCError('userid is not the same as your user')
 
     if isinstance(userid, Optional):
         userid = apiuser.user_id
 
     user = get_user_or_error(userid)
 
     if isinstance(locked, Optional):
         lockobj = repo.locked
 
         if lockobj[0] is None:
             _d = {
                 'repo': repo.repo_name,
                 'locked': False,
                 'locked_since': None,
                 'locked_by': None,
                 'lock_reason': None,
                 'lock_state_changed': False,
                 'msg': 'Repo `%s` not locked.' % repo.repo_name
             }
             return _d
         else:
             _user_id, _time, _reason = lockobj
             lock_user = get_user_or_error(userid)
             _d = {
                 'repo': repo.repo_name,
                 'locked': True,
                 'locked_since': _time,
                 'locked_by': lock_user.username,
                 'lock_reason': _reason,
                 'lock_state_changed': False,
                 'msg': ('Repo `%s` locked by `%s` on `%s`.'
                         % (repo.repo_name, lock_user.username,
                            json.dumps(time_to_datetime(_time))))
             }
             return _d
 
     # force locked state through a flag
     else:
         locked = str2bool(locked)
         lock_reason = Repository.LOCK_API
         try:
             if locked:
                 lock_time = time.time()
                 Repository.lock(repo, user.user_id, lock_time, lock_reason)
             else:
                 lock_time = None
                 Repository.unlock(repo)
             _d = {
                 'repo': repo.repo_name,
                 'locked': locked,
                 'locked_since': lock_time,
                 'locked_by': user.username,
                 'lock_reason': lock_reason,
                 'lock_state_changed': True,
                 'msg': ('User `%s` set lock state for repo `%s` to `%s`'
                         % (user.username, repo.repo_name, locked))
             }
             return _d
         except Exception:
             log.exception(
                 "Exception occurred while trying to lock repository")
             raise JSONRPCError(
                 'Error occurred locking repository `%s`' % repo.repo_name
             )
 
 
 @jsonrpc_method()
 def comment_commit(
         request, apiuser, repoid, commit_id, message, status=Optional(None),
         comment_type=Optional(ChangesetComment.COMMENT_TYPE_NOTE),
         resolves_comment_id=Optional(None), extra_recipients=Optional([]),
         userid=Optional(OAttr('apiuser')), send_email=Optional(True)):
     """
     Set a commit comment, and optionally change the status of the commit.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set the repository name or repository ID.
     :type repoid: str or int
     :param commit_id: Specify the commit_id for which to set a comment.
     :type commit_id: str
     :param message: The comment text.
     :type message: str
     :param status: (**Optional**) status of commit, one of: 'not_reviewed',
         'approved', 'rejected', 'under_review'
     :type status: str
     :param comment_type: Comment type, one of: 'note', 'todo'
     :type comment_type: Optional(str), default: 'note'
     :param resolves_comment_id: id of comment which this one will resolve
     :type resolves_comment_id: Optional(int)
     :param extra_recipients: list of user ids or usernames to add
         notifications for this comment. Acts like a CC for notification
     :type extra_recipients: Optional(list)
     :param userid: Set the user name of the comment creator.
     :type userid: Optional(str or int)
     :param send_email: Define if this comment should also send email notification
     :type send_email: Optional(bool)
 
     Example error output:
 
     .. code-block:: bash
 
         {
             "id" : <id_given_in_input>,
             "result" : {
                 "msg": "Commented on commit `<commit_id>` for repository `<repoid>`",
                 "status_change": null or <status>,
                 "success": true
             },
             "error" :  null
         }
 
     """
     _ = request.translate
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.read', 'repository.write', 'repository.admin')
         validate_repo_permissions(apiuser, repoid, repo, _perms)
     db_repo_name = repo.repo_name
 
     try:
         commit = repo.scm_instance().get_commit(commit_id=commit_id)
         commit_id = commit.raw_id
     except Exception as e:
         log.exception('Failed to fetch commit')
         raise JSONRPCError(safe_str(e))
 
     if isinstance(userid, Optional):
         userid = apiuser.user_id
 
     user = get_user_or_error(userid)
     status = Optional.extract(status)
     comment_type = Optional.extract(comment_type)
     resolves_comment_id = Optional.extract(resolves_comment_id)
     extra_recipients = Optional.extract(extra_recipients)
     send_email = Optional.extract(send_email, binary=True)
 
     allowed_statuses = [x[0] for x in ChangesetStatus.STATUSES]
     if status and status not in allowed_statuses:
         raise JSONRPCError('Bad status, must be on '
                            'of %s got %s' % (allowed_statuses, status,))
 
     if resolves_comment_id:
         comment = ChangesetComment.get(resolves_comment_id)
         if not comment:
             raise JSONRPCError(
                 'Invalid resolves_comment_id `%s` for this commit.'
                 % resolves_comment_id)
         if comment.comment_type != ChangesetComment.COMMENT_TYPE_TODO:
             raise JSONRPCError(
                 'Comment `%s` is wrong type for setting status to resolved.'
                 % resolves_comment_id)
 
     try:
         rc_config = SettingsModel().get_all_settings()
         renderer = rc_config.get('rhodecode_markup_renderer', 'rst')
         status_change_label = ChangesetStatus.get_status_lbl(status)
         comment = CommentsModel().create(
             message, repo, user, commit_id=commit_id,
             status_change=status_change_label,
             status_change_type=status,
             renderer=renderer,
             comment_type=comment_type,
             resolves_comment_id=resolves_comment_id,
             auth_user=apiuser,
             extra_recipients=extra_recipients,
             send_email=send_email
         )
         is_inline = comment.is_inline
 
         if status:
             # also do a status change
             try:
                 ChangesetStatusModel().set_status(
                     repo, status, user, comment, revision=commit_id,
                     dont_allow_on_closed_pull_request=True
                 )
             except StatusChangeOnClosedPullRequestError:
                 log.exception(
                     "Exception occurred while trying to change repo commit status")
                 msg = ('Changing status on a commit associated with '
                        'a closed pull request is not allowed')
                 raise JSONRPCError(msg)
 
         CommentsModel().trigger_commit_comment_hook(
             repo, apiuser, 'create',
             data={'comment': comment, 'commit': commit})
 
         Session().commit()
 
         comment_broadcast_channel = channelstream.comment_channel(
             db_repo_name, commit_obj=commit)
 
         comment_data = {'comment': comment, 'comment_id': comment.comment_id}
         comment_type = 'inline' if is_inline else 'general'
         channelstream.comment_channelstream_push(
             request, comment_broadcast_channel, apiuser,
             _('posted a new {} comment').format(comment_type),
             comment_data=comment_data)
 
         return {
             'msg': (
                 'Commented on commit `{}` for repository `{}`'.format(
                     comment.revision, repo.repo_name)),
             'status_change': status,
             'success': True,
         }
     except JSONRPCError:
         # catch any inside errors, and re-raise them to prevent from
         # below global catch to silence them
         raise
     except Exception:
         log.exception("Exception occurred while trying to comment on commit")
         raise JSONRPCError(
             f'failed to set comment on repository `{repo.repo_name}`'
         )
 
 
 @jsonrpc_method()
 def get_repo_comments(request, apiuser, repoid,
                       commit_id=Optional(None), comment_type=Optional(None),
                       userid=Optional(None)):
     """
     Get all comments for a repository
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set the repository name or repository ID.
     :type repoid: str or int
     :param commit_id: Optionally filter the comments by the commit_id
     :type commit_id: Optional(str), default: None
     :param comment_type: Optionally filter the comments by the comment_type
         one of: 'note', 'todo'
     :type comment_type: Optional(str), default: None
     :param userid: Optionally filter the comments by the author of comment
     :type userid: Optional(str or int), Default: None
 
     Example error output:
 
     .. code-block:: bash
 
         {
             "id" : <id_given_in_input>,
             "result" : [
                 {
                   "comment_author": <USER_DETAILS>,
                   "comment_created_on": "2017-02-01T14:38:16.309",
                   "comment_f_path": "file.txt",
                   "comment_id": 282,
                   "comment_lineno": "n1",
                   "comment_resolved_by": null,
                   "comment_status": [],
                   "comment_text": "This file needs a header",
                   "comment_type": "todo",
                   "comment_last_version: 0
                 }
             ],
             "error" :  null
         }
 
     """
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.read', 'repository.write', 'repository.admin')
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     commit_id = Optional.extract(commit_id)
 
     userid = Optional.extract(userid)
     if userid:
         user = get_user_or_error(userid)
     else:
         user = None
 
     comment_type = Optional.extract(comment_type)
     if comment_type and comment_type not in ChangesetComment.COMMENT_TYPES:
         raise JSONRPCError(
                 'comment_type must be one of `{}` got {}'.format(
                     ChangesetComment.COMMENT_TYPES, comment_type)
             )
 
     comments = CommentsModel().get_repository_comments(
         repo=repo, comment_type=comment_type, user=user, commit_id=commit_id)
     return comments
 
 
 @jsonrpc_method()
 def get_comment(request, apiuser, comment_id):
     """
     Get single comment from repository or pull_request
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param comment_id: comment id found in the URL of comment
     :type comment_id: str or int
 
     Example error output:
 
     .. code-block:: bash
 
         {
             "id" : <id_given_in_input>,
             "result" : {
                   "comment_author": <USER_DETAILS>,
                   "comment_created_on": "2017-02-01T14:38:16.309",
                   "comment_f_path": "file.txt",
                   "comment_id": 282,
                   "comment_lineno": "n1",
                   "comment_resolved_by": null,
                   "comment_status": [],
                   "comment_text": "This file needs a header",
                   "comment_type": "todo",
                   "comment_last_version: 0
             },
             "error" :  null
         }
 
     """
 
     comment = ChangesetComment.get(comment_id)
     if not comment:
         raise JSONRPCError(f'comment `{comment_id}` does not exist')
 
     perms = ('repository.read', 'repository.write', 'repository.admin')
     has_comment_perm = HasRepoPermissionAnyApi(*perms)\
         (user=apiuser, repo_name=comment.repo.repo_name)
 
     if not has_comment_perm:
         raise JSONRPCError(f'comment `{comment_id}` does not exist')
 
     return comment
 
 
 @jsonrpc_method()
 def edit_comment(request, apiuser, message, comment_id, version,
                  userid=Optional(OAttr('apiuser'))):
     """
     Edit comment on the pull request or commit,
     specified by the `comment_id` and version. Initially version should be 0
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param comment_id: Specify the comment_id for editing
     :type comment_id: int
     :param version: version of the comment that will be created, starts from 0
     :type version: int
     :param message: The text content of the comment.
     :type message: str
     :param userid: Comment on the pull request as this user
     :type userid: Optional(str or int)
 
     Example output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result : {
             "comment": "<comment data>",
             "version": "<Integer>",
         },
         error :  null
     """
 
     auth_user = apiuser
     comment = ChangesetComment.get(comment_id)
     if not comment:
         raise JSONRPCError(f'comment `{comment_id}` does not exist')
 
     is_super_admin = has_superadmin_permission(apiuser)
     is_repo_admin = HasRepoPermissionAnyApi('repository.admin')\
         (user=apiuser, repo_name=comment.repo.repo_name)
 
     if not isinstance(userid, Optional):
         if is_super_admin or is_repo_admin:
             apiuser = get_user_or_error(userid)
             auth_user = apiuser.AuthUser()
         else:
             raise JSONRPCError('userid is not the same as your user')
 
     comment_author = comment.author.user_id == auth_user.user_id
 
     if comment.immutable:
         raise JSONRPCError("Immutable comment cannot be edited")
 
     if not (is_super_admin or is_repo_admin or comment_author):
         raise JSONRPCError("you don't have access to edit this comment")
 
     try:
         comment_history = CommentsModel().edit(
             comment_id=comment_id,
             text=message,
             auth_user=auth_user,
             version=version,
         )
         Session().commit()
     except CommentVersionMismatch:
         raise JSONRPCError(
             f'comment ({comment_id}) version ({version}) mismatch'
         )
     if not comment_history and not message:
         raise JSONRPCError(
             f"comment ({comment_id}) can't be changed with empty string"
         )
 
     if comment.pull_request:
         pull_request = comment.pull_request
         PullRequestModel().trigger_pull_request_hook(
             pull_request, apiuser, 'comment_edit',
             data={'comment': comment})
     else:
         db_repo = comment.repo
         commit_id = comment.revision
         commit = db_repo.get_commit(commit_id)
         CommentsModel().trigger_commit_comment_hook(
             db_repo, apiuser, 'edit',
             data={'comment': comment, 'commit': commit})
 
     data = {
         'comment': comment,
         'version': comment_history.version if comment_history else None,
     }
     return data
 
 
 # TODO(marcink): write this with all required logic for deleting a comments in PR or commits
 # @jsonrpc_method()
 # def delete_comment(request, apiuser, comment_id):
 #     auth_user = apiuser
 #
 #     comment = ChangesetComment.get(comment_id)
 #     if not comment:
 #         raise JSONRPCError('comment `%s` does not exist' % (comment_id,))
 #
 #     is_super_admin = has_superadmin_permission(apiuser)
 #     is_repo_admin = HasRepoPermissionAnyApi('repository.admin')\
 #         (user=apiuser, repo_name=comment.repo.repo_name)
 #
 #     comment_author = comment.author.user_id == auth_user.user_id
 #     if not (comment.immutable is False and (is_super_admin or is_repo_admin) or comment_author):
 #         raise JSONRPCError("you don't have access to edit this comment")
 
 @jsonrpc_method()
 def grant_user_permission(request, apiuser, repoid, userid, perm):
     """
     Grant permissions for the specified user on the given repository,
     or update existing permissions if found.
 
     This command can only be run using an |authtoken| with admin
     permissions on the |repo|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set the repository name or repository ID.
     :type repoid: str or int
     :param userid: Set the user name.
     :type userid: str
     :param perm: Set the user permissions, using the following format
         ``(repository.(none|read|write|admin))``
     :type perm: str
 
     Example output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result: {
                   "msg" : "Granted perm: `<perm>` for user: `<username>` in repo: `<reponame>`",
                   "success": true
                 }
         error:  null
     """
 
     repo = get_repo_or_error(repoid)
     user = get_user_or_error(userid)
     perm = get_perm_or_error(perm)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     perm_additions = [[user.user_id, perm.permission_name, "user"]]
     try:
         changes = RepoModel().update_permissions(
             repo=repo, perm_additions=perm_additions, cur_user=apiuser)
 
         action_data = {
             'added': changes['added'],
             'updated': changes['updated'],
             'deleted': changes['deleted'],
         }
         audit_logger.store_api(
             'repo.edit.permissions', action_data=action_data, user=apiuser, repo=repo)
         Session().commit()
         PermissionModel().flush_user_permission_caches(changes)
 
         return {
             'msg': 'Granted perm: `{}` for user: `{}` in repo: `{}`'.format(
                 perm.permission_name, user.username, repo.repo_name
             ),
             'success': True
         }
     except Exception:
         log.exception("Exception occurred while trying edit permissions for repo")
         raise JSONRPCError(
             'failed to edit permission for user: `{}` in repo: `{}`'.format(
                 userid, repoid
             )
         )
 
 
 @jsonrpc_method()
 def revoke_user_permission(request, apiuser, repoid, userid):
     """
     Revoke permission for a user on the specified repository.
 
     This command can only be run using an |authtoken| with admin
     permissions on the |repo|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set the repository name or repository ID.
     :type repoid: str or int
     :param userid: Set the user name of revoked user.
     :type userid: str or int
 
     Example error output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result: {
                   "msg" : "Revoked perm for user: `<username>` in repo: `<reponame>`",
                   "success": true
                 }
         error:  null
     """
 
     repo = get_repo_or_error(repoid)
     user = get_user_or_error(userid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     perm_deletions = [[user.user_id, None, "user"]]
     try:
         changes = RepoModel().update_permissions(
             repo=repo, perm_deletions=perm_deletions, cur_user=user)
 
         action_data = {
             'added': changes['added'],
             'updated': changes['updated'],
             'deleted': changes['deleted'],
         }
         audit_logger.store_api(
             'repo.edit.permissions', action_data=action_data, user=apiuser, repo=repo)
         Session().commit()
         PermissionModel().flush_user_permission_caches(changes)
 
         return {
             'msg': 'Revoked perm for user: `{}` in repo: `{}`'.format(
                 user.username, repo.repo_name
             ),
             'success': True
         }
     except Exception:
         log.exception("Exception occurred while trying revoke permissions to repo")
         raise JSONRPCError(
             'failed to edit permission for user: `{}` in repo: `{}`'.format(
                 userid, repoid
             )
         )
 
 
 @jsonrpc_method()
 def grant_user_group_permission(request, apiuser, repoid, usergroupid, perm):
     """
     Grant permission for a user group on the specified repository,
     or update existing permissions.
 
     This command can only be run using an |authtoken| with admin
     permissions on the |repo|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set the repository name or repository ID.
     :type repoid: str or int
     :param usergroupid: Specify the ID of the user group.
     :type usergroupid: str or int
     :param perm: Set the user group permissions using the following
         format: (repository.(none|read|write|admin))
     :type perm: str
 
     Example output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : {
         "msg" : "Granted perm: `<perm>` for group: `<usersgroupname>` in repo: `<reponame>`",
         "success": true
 
       }
       error :  null
 
     Example error output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : null
       error :  {
         "failed to edit permission for user group: `<usergroup>` in repo `<repo>`'
       }
 
     """
 
     repo = get_repo_or_error(repoid)
     perm = get_perm_or_error(perm)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     user_group = get_user_group_or_error(usergroupid)
     if not has_superadmin_permission(apiuser):
         # check if we have at least read permission for this user group !
         _perms = ('usergroup.read', 'usergroup.write', 'usergroup.admin',)
         if not HasUserGroupPermissionAnyApi(*_perms)(
                 user=apiuser, user_group_name=user_group.users_group_name):
             raise JSONRPCError(
                 f'user group `{usergroupid}` does not exist')
 
     perm_additions = [[user_group.users_group_id, perm.permission_name, "user_group"]]
     try:
         changes = RepoModel().update_permissions(
             repo=repo, perm_additions=perm_additions, cur_user=apiuser)
         action_data = {
             'added': changes['added'],
             'updated': changes['updated'],
             'deleted': changes['deleted'],
         }
         audit_logger.store_api(
             'repo.edit.permissions', action_data=action_data, user=apiuser, repo=repo)
         Session().commit()
         PermissionModel().flush_user_permission_caches(changes)
 
         return {
             'msg': 'Granted perm: `%s` for user group: `%s` in '
                    'repo: `%s`' % (
                        perm.permission_name, user_group.users_group_name,
                        repo.repo_name
                    ),
             'success': True
         }
     except Exception:
         log.exception(
             "Exception occurred while trying change permission on repo")
         raise JSONRPCError(
             'failed to edit permission for user group: `%s` in '
             'repo: `%s`' % (
                 usergroupid, repo.repo_name
             )
         )
 
 
 @jsonrpc_method()
 def revoke_user_group_permission(request, apiuser, repoid, usergroupid):
     """
     Revoke the permissions of a user group on a given repository.
 
     This command can only be run using an |authtoken| with admin
     permissions on the |repo|.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: Set the repository name or repository ID.
     :type repoid: str or int
     :param usergroupid: Specify the user group ID.
     :type usergroupid: str or int
 
     Example output:
 
     .. code-block:: bash
 
         id : <id_given_in_input>
         result: {
                   "msg" : "Revoked perm for group: `<usersgroupname>` in repo: `<reponame>`",
                   "success": true
                 }
         error:  null
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     user_group = get_user_group_or_error(usergroupid)
     if not has_superadmin_permission(apiuser):
         # check if we have at least read permission for this user group !
         _perms = ('usergroup.read', 'usergroup.write', 'usergroup.admin',)
         if not HasUserGroupPermissionAnyApi(*_perms)(
                 user=apiuser, user_group_name=user_group.users_group_name):
             raise JSONRPCError(
                 f'user group `{usergroupid}` does not exist')
 
     perm_deletions = [[user_group.users_group_id, None, "user_group"]]
     try:
         changes = RepoModel().update_permissions(
             repo=repo, perm_deletions=perm_deletions, cur_user=apiuser)
         action_data = {
             'added': changes['added'],
             'updated': changes['updated'],
             'deleted': changes['deleted'],
         }
         audit_logger.store_api(
             'repo.edit.permissions', action_data=action_data, user=apiuser, repo=repo)
         Session().commit()
         PermissionModel().flush_user_permission_caches(changes)
 
         return {
             'msg': 'Revoked perm for user group: `{}` in repo: `{}`'.format(
                 user_group.users_group_name, repo.repo_name
             ),
             'success': True
         }
     except Exception:
         log.exception("Exception occurred while trying revoke "
                       "user group permission on repo")
         raise JSONRPCError(
             'failed to edit permission for user group: `%s` in '
             'repo: `%s`' % (
                 user_group.users_group_name, repo.repo_name
             )
         )
 
 
 @jsonrpc_method()
-def pull(request, apiuser, repoid, remote_uri=Optional(None)):
+def pull(request, apiuser, repoid, remote_uri=Optional(None), sync_large_objects=Optional(False)):
     """
     Triggers a pull on the given repository from a remote location. You
     can use this to keep remote repositories up-to-date.
 
     This command can only be run using an |authtoken| with admin
     rights to the specified repository. For more information,
     see :ref:`config-token-ref`.
 
     This command takes the following options:
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository ID.
     :type repoid: str or int
     :param remote_uri: Optional remote URI to pass in for pull
     :type remote_uri: str
+    :param sync_large_objects: Optional flag for pulling LFS objects.
+    :type sync_large_objects: bool
 
     Example output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : {
         "msg": "Pulled from url `<remote_url>` on repo `<repository name>`"
         "repository": "<repository name>"
       }
       error :  null
 
     Example error output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : null
       error :  {
         "Unable to push changes from `<remote_url>`"
       }
 
     """
 
     repo = get_repo_or_error(repoid)
     remote_uri = Optional.extract(remote_uri)
     remote_uri_display = remote_uri or repo.clone_uri_hidden
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     try:
         ScmModel().pull_changes(
-            repo.repo_name, apiuser.username, remote_uri=remote_uri)
+            repo.repo_name, apiuser.username, remote_uri=remote_uri, sync_large_objects=sync_large_objects)
         return {
             'msg': 'Pulled from url `{}` on repo `{}`'.format(
                 remote_uri_display, repo.repo_name),
             'repository': repo.repo_name
         }
     except Exception:
         log.exception("Exception occurred while trying to "
                       "pull changes from remote location")
         raise JSONRPCError(
             'Unable to pull changes from `%s`' % remote_uri_display
         )
 
 
 @jsonrpc_method()
 def strip(request, apiuser, repoid, revision, branch):
     """
     Strips the given revision from the specified repository.
 
     * This will remove the revision and all of its decendants.
 
     This command can only be run using an |authtoken| with admin rights to
     the specified repository.
 
     This command takes the following options:
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository ID.
     :type repoid: str or int
     :param revision: The revision you wish to strip.
     :type revision: str
     :param branch: The branch from which to strip the revision.
     :type branch: str
 
     Example output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : {
         "msg": "'Stripped commit <commit_hash> from repo `<repository name>`'"
         "repository": "<repository name>"
       }
       error :  null
 
     Example error output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : null
       error :  {
         "Unable to strip commit <commit_hash> from repo `<repository name>`"
       }
 
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     try:
         ScmModel().strip(repo, revision, branch)
         audit_logger.store_api(
             'repo.commit.strip', action_data={'commit_id': revision},
             repo=repo,
             user=apiuser, commit=True)
 
         return {
             'msg': 'Stripped commit {} from repo `{}`'.format(
                 revision, repo.repo_name),
             'repository': repo.repo_name
         }
     except Exception:
         log.exception("Exception while trying to strip")
         raise JSONRPCError(
             'Unable to strip commit {} from repo `{}`'.format(
                 revision, repo.repo_name)
         )
 
 
 @jsonrpc_method()
 def get_repo_settings(request, apiuser, repoid, key=Optional(None)):
     """
     Returns all settings for a repository. If key is given it only returns the
     setting identified by the key or null.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository id.
     :type repoid: str or int
     :param key: Key of the setting to return.
     :type: key: Optional(str)
 
     Example output:
 
     .. code-block:: bash
 
         {
             "error": null,
             "id": 237,
             "result": {
                 "extensions_largefiles": true,
                 "extensions_evolve": true,
                 "hooks_changegroup_push_logger": true,
                 "hooks_changegroup_repo_size": false,
                 "hooks_outgoing_pull_logger": true,
                 "phases_publish": "True",
                 "rhodecode_hg_use_rebase_for_merging": true,
                 "rhodecode_pr_merge_enabled": true,
                 "rhodecode_use_outdated_comments": true
             }
         }
     """
 
     # Restrict access to this api method to super-admins, and repo admins only.
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     try:
         settings_model = VcsSettingsModel(repo=repo)
         settings = settings_model.get_global_settings()
         settings.update(settings_model.get_repo_settings())
 
         # If only a single setting is requested fetch it from all settings.
         key = Optional.extract(key)
         if key is not None:
             settings = settings.get(key, None)
     except Exception:
         msg = f'Failed to fetch settings for repository `{repoid}`'
         log.exception(msg)
         raise JSONRPCError(msg)
 
     return settings
 
 
 @jsonrpc_method()
 def set_repo_settings(request, apiuser, repoid, settings):
     """
     Update repository settings. Returns true on success.
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository id.
     :type repoid: str or int
     :param settings: The new settings for the repository.
     :type: settings: dict
 
     Example output:
 
     .. code-block:: bash
 
         {
             "error": null,
             "id": 237,
             "result": true
         }
     """
     # Restrict access to this api method to super-admins, and repo admins only.
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     if type(settings) is not dict:
         raise JSONRPCError('Settings have to be a JSON Object.')
 
     try:
         settings_model = VcsSettingsModel(repo=repoid)
 
         # Merge global, repo and incoming settings.
         new_settings = settings_model.get_global_settings()
         new_settings.update(settings_model.get_repo_settings())
         new_settings.update(settings)
 
         # Update the settings.
         inherit_global_settings = new_settings.get(
             'inherit_global_settings', False)
         settings_model.create_or_update_repo_settings(
             new_settings, inherit_global_settings=inherit_global_settings)
         Session().commit()
     except Exception:
         msg = f'Failed to update settings for repository `{repoid}`'
         log.exception(msg)
         raise JSONRPCError(msg)
 
     # Indicate success.
     return True
 
 
 @jsonrpc_method()
 def maintenance(request, apiuser, repoid):
     """
     Triggers a maintenance on the given repository.
 
     This command can only be run using an |authtoken| with admin
     rights to the specified repository. For more information,
     see :ref:`config-token-ref`.
 
     This command takes the following options:
 
     :param apiuser: This is filled automatically from the |authtoken|.
     :type apiuser: AuthUser
     :param repoid: The repository name or repository ID.
     :type repoid: str or int
 
     Example output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : {
         "msg": "executed maintenance command",
         "executed_actions": [
            <action_message>, <action_message2>...
         ],
         "repository": "<repository name>"
       }
       error :  null
 
     Example error output:
 
     .. code-block:: bash
 
       id : <id_given_in_input>
       result : null
       error :  {
         "Unable to execute maintenance on `<reponame>`"
       }
 
     """
 
     repo = get_repo_or_error(repoid)
     if not has_superadmin_permission(apiuser):
         _perms = ('repository.admin',)
         validate_repo_permissions(apiuser, repoid, repo, _perms)
 
     try:
         maintenance = repo_maintenance.RepoMaintenance()
         executed_actions = maintenance.execute(repo)
 
         return {
             'msg': 'executed maintenance command',
             'executed_actions': executed_actions,
             'repository': repo.repo_name
         }
     except Exception:
         log.exception("Exception occurred while trying to run maintenance")
         raise JSONRPCError(
             'Unable to execute maintenance on `%s`' % repo.repo_name)

rhodecode/apps/repository/views/repo_settings_remote.py

@@ -1,62 +1,62 b''
 # Copyright (C) 2017-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import logging
 
 from pyramid.httpexceptions import HTTPFound
 
 
 from rhodecode.apps._base import RepoAppView
 from rhodecode.lib import helpers as h
 from rhodecode.lib.auth import (
     LoginRequired, CSRFRequired, HasRepoPermissionAnyDecorator)
 from rhodecode.model.scm import ScmModel
 
 log = logging.getLogger(__name__)
 
 
 class RepoSettingsRemoteView(RepoAppView):
     def load_default_context(self):
         c = self._get_local_tmpl_context()
         return c
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.admin')
     def repo_remote_edit_form(self):
         c = self.load_default_context()
         c.active = 'remote'
 
         return self._get_template_context(c)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.admin')
     @CSRFRequired()
     def repo_remote_pull_changes(self):
         _ = self.request.translate
         self.load_default_context()
 
         try:
             ScmModel().pull_changes(
-                self.db_repo_name, self._rhodecode_user.username)
+                self.db_repo_name, self._rhodecode_user.username, sync_large_objects=True)
             h.flash(_('Pulled from remote location'), category='success')
         except Exception:
             log.exception("Exception during pull from remote")
             h.flash(_('An error occurred during pull from remote location'),
                     category='error')
         raise HTTPFound(
             h.route_path('edit_repo_remote', repo_name=self.db_repo_name))

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

@@ -1,1053 +1,1053 b''
 # Copyright (C) 2014-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 """
 GIT repository module
 """
 
 import logging
 import os
 import re
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 from collections import OrderedDict
 from rhodecode.lib.datelib import (
     utcdate_fromtimestamp, makedate, date_astimestamp)
 from rhodecode.lib.hash_utils import 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 = f'branch:{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 = {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(f'cmd must be a list, got {type(cmd)} instead')
 
         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(
                 f"Cannot create repository at {self.path}, location already exist")
 
         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:
                 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(
                         f'Path "{self.path}" does not contain a Git repository')
 
         # 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)
         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_str(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.items():
             if ref.startswith(prefix):
                 ref_name = ref
                 if strip_prefix:
                     ref_name = ref[len(prefix):]
                 result.append((safe_str(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 f"Added tag {name} for commit {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.items():
             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 = f'refs/remotes/origin/{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 path:
         :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``.
         :param path1:
         """
         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')
 
         if not branch_name:
             raise ValueError(f'git strip requires a valid branch name, got {branch_name} instead')
 
         # parent is going to be the new head now
         commit = commit.parents[0]
         self._remote.update_refs(f'refs/heads/{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',
                  f'{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):
+    def fetch(self, url, commit_ids=None, **kwargs):
         """
         Fetch all git objects from external location.
         """
-        self._remote.sync_fetch(url, refs=commit_ids)
+        self._remote.sync_fetch(url, refs=commit_ids, **kwargs)
         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 = f'{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', f'user.name="{user_name}"',
                '-c', f'user.email={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 = [f'U {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),
                    f'{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 = f'pr_{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

@@ -1,1024 +1,1024 b''
 # Copyright (C) 2014-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 """
 HG repository module
 """
 import os
 import logging
 import binascii
 import configparser
 import urllib.request
 import urllib.parse
 import urllib.error
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 from collections import OrderedDict
 from rhodecode.lib.datelib import (
     date_to_timestamp_plus_offset, utcdate_fromtimestamp, makedate)
 from rhodecode.lib.str_utils import 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', '')])
 
         # NOTE(marcink): since python3 hgsubversion is deprecated.
         # From old installations we might still have this set enabled
         # we explicitly remove this now here to make sure it wont propagate further
         if config and config.get('extensions', 'hgsubversion') is not None:
             config.drop_option('extensions', 'hgsubversion')
 
         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 = {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 = [(n, 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 = [(n, 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 = f"Added tag {name} for commit {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 = [
             (n, 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=False)
 
         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(
                 f"Cannot create repository at {self.path}, location already exist")
 
         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_str(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_str(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.
         """
         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"
 
         # case here is no cached version, do an actual lookup instead
         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(f'branch("{branch_name}")')
         elif branch_name and branch_ancestors:
             commit_filter.append(f'ancestors(branch("{branch_name}"))')
 
         if start_date and not end_date:
             commit_filter.append(f'date(">{start_date}")')
         if end_date and not start_date:
             commit_filter.append(f'date("<{end_date}")')
         if start_date and end_date:
             commit_filter.append(
                 f'date(">{start_date}") and 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):
+    def fetch(self, url, commit_ids=None, **kwargs):
         """
         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 = f'rcbook{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(f'{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 f"Closing branch: `{source_ref.name}`"
         try:
             self._remote.commit(
                 message=safe_str(message),
                 username=safe_str(f'{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] +
                         [f'and {len(heads_all)-max_heads} more.'])
                 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')
                 metadata['unresolved_files'] = 'no unresolved files found'
 
                 if isinstance(e, UnresolvedFilesInRepo):
                     all_conflicts = list(e.args[0])
                     max_conflicts = 20
                     if len(all_conflicts) > max_conflicts:
                         conflicts = all_conflicts[:max_conflicts] \
                                     + [f'and {len(all_conflicts)-max_conflicts} more.']
                     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, str):
             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):
         if isinstance(commit_id, int):
             return self.repo.get_commit(
                 commit_idx=commit_id, pre_load=self.pre_load)
         else:
             return self.repo.get_commit(
                 commit_id=commit_id, pre_load=self.pre_load)

rhodecode/model/scm.py

@@ -1,1044 +1,1044 b''
 # Copyright (C) 2010-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 """
 Scm model for RhodeCode
 """
 
 import os.path
 import traceback
 import logging
 import io
 
 from sqlalchemy import func
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 import rhodecode
 from rhodecode.lib.str_utils import safe_bytes
 from rhodecode.lib.vcs import get_backend
 from rhodecode.lib.vcs.exceptions import RepositoryError, NodeNotChangedError
 from rhodecode.lib.vcs.nodes import FileNode
 from rhodecode.lib.vcs.backends.base import EmptyCommit
 from rhodecode.lib import helpers as h, rc_cache
 from rhodecode.lib.auth import (
     HasRepoPermissionAny, HasRepoGroupPermissionAny,
     HasUserGroupPermissionAny)
 from rhodecode.lib.exceptions import NonRelativePathError, IMCCommitError
 from rhodecode.lib import hooks_utils
 from rhodecode.lib.utils import (
     get_filesystem_repos, make_db_config)
 from rhodecode.lib.str_utils import safe_str
 from rhodecode.lib.system_info import get_system_info
 from rhodecode.model import BaseModel
 from rhodecode.model.db import (
     or_, false, null,
     Repository, CacheKey, UserFollowing, UserLog, User, RepoGroup,
     PullRequest, FileStore)
 from rhodecode.model.settings import VcsSettingsModel
 from rhodecode.model.validation_schema.validators import url_validator, InvalidCloneUrl
 
 log = logging.getLogger(__name__)
 
 
 class UserTemp(object):
     def __init__(self, user_id):
         self.user_id = user_id
 
     def __repr__(self):
         return "<{}('id:{}')>".format(self.__class__.__name__, self.user_id)
 
 
 class RepoTemp(object):
     def __init__(self, repo_id):
         self.repo_id = repo_id
 
     def __repr__(self):
         return "<{}('id:{}')>".format(self.__class__.__name__, self.repo_id)
 
 
 class SimpleCachedRepoList(object):
     """
     Lighter version of of iteration of repos without the scm initialisation,
     and with cache usage
     """
     def __init__(self, db_repo_list, repos_path, order_by=None, perm_set=None):
         self.db_repo_list = db_repo_list
         self.repos_path = repos_path
         self.order_by = order_by
         self.reversed = (order_by or '').startswith('-')
         if not perm_set:
             perm_set = ['repository.read', 'repository.write',
                         'repository.admin']
         self.perm_set = perm_set
 
     def __len__(self):
         return len(self.db_repo_list)
 
     def __repr__(self):
         return '<{} ({})>'.format(self.__class__.__name__, self.__len__())
 
     def __iter__(self):
         for dbr in self.db_repo_list:
             # check permission at this level
             has_perm = HasRepoPermissionAny(*self.perm_set)(
                 dbr.repo_name, 'SimpleCachedRepoList check')
             if not has_perm:
                 continue
 
             tmp_d = {
                 'name': dbr.repo_name,
                 'dbrepo': dbr.get_dict(),
                 'dbrepo_fork': dbr.fork.get_dict() if dbr.fork else {}
             }
             yield tmp_d
 
 
 class _PermCheckIterator(object):
 
     def __init__(
             self, obj_list, obj_attr, perm_set, perm_checker,
             extra_kwargs=None):
         """
         Creates iterator from given list of objects, additionally
         checking permission for them from perm_set var
 
         :param obj_list: list of db objects
         :param obj_attr: attribute of object to pass into perm_checker
         :param perm_set: list of permissions to check
         :param perm_checker: callable to check permissions against
         """
         self.obj_list = obj_list
         self.obj_attr = obj_attr
         self.perm_set = perm_set
         self.perm_checker = perm_checker(*self.perm_set)
         self.extra_kwargs = extra_kwargs or {}
 
     def __len__(self):
         return len(self.obj_list)
 
     def __repr__(self):
         return '<{} ({})>'.format(self.__class__.__name__, self.__len__())
 
     def __iter__(self):
         for db_obj in self.obj_list:
             # check permission at this level
             # NOTE(marcink): the __dict__.get() is ~4x faster then getattr()
             name = db_obj.__dict__.get(self.obj_attr, None)
             if not self.perm_checker(name, self.__class__.__name__, **self.extra_kwargs):
                 continue
 
             yield db_obj
 
 
 class RepoList(_PermCheckIterator):
 
     def __init__(self, db_repo_list, perm_set=None, extra_kwargs=None):
         if not perm_set:
             perm_set = ['repository.read', 'repository.write', 'repository.admin']
 
         super().__init__(
             obj_list=db_repo_list,
             obj_attr='_repo_name', perm_set=perm_set,
             perm_checker=HasRepoPermissionAny,
             extra_kwargs=extra_kwargs)
 
 
 class RepoGroupList(_PermCheckIterator):
 
     def __init__(self, db_repo_group_list, perm_set=None, extra_kwargs=None):
         if not perm_set:
             perm_set = ['group.read', 'group.write', 'group.admin']
 
         super().__init__(
             obj_list=db_repo_group_list,
             obj_attr='_group_name', perm_set=perm_set,
             perm_checker=HasRepoGroupPermissionAny,
             extra_kwargs=extra_kwargs)
 
 
 class UserGroupList(_PermCheckIterator):
 
     def __init__(self, db_user_group_list, perm_set=None, extra_kwargs=None):
         if not perm_set:
             perm_set = ['usergroup.read', 'usergroup.write', 'usergroup.admin']
 
         super().__init__(
             obj_list=db_user_group_list,
             obj_attr='users_group_name', perm_set=perm_set,
             perm_checker=HasUserGroupPermissionAny,
             extra_kwargs=extra_kwargs)
 
 
 class ScmModel(BaseModel):
     """
     Generic Scm Model
     """
 
     @LazyProperty
     def repos_path(self):
         """
         Gets the repositories root path from database
         """
 
         settings_model = VcsSettingsModel(sa=self.sa)
         return settings_model.get_repos_location()
 
     def repo_scan(self, repos_path=None):
         """
         Listing of repositories in given path. This path should not be a
         repository itself. Return a dictionary of repository objects
 
         :param repos_path: path to directory containing repositories
         """
 
         if repos_path is None:
             repos_path = self.repos_path
 
         log.info('scanning for repositories in %s', repos_path)
 
         config = make_db_config()
         config.set('extensions', 'largefiles', '')
         repos = {}
 
         for name, path in get_filesystem_repos(repos_path, recursive=True):
             # name need to be decomposed and put back together using the /
             # since this is internal storage separator for rhodecode
             name = Repository.normalize_repo_name(name)
 
             try:
                 if name in repos:
                     raise RepositoryError('Duplicate repository name %s '
                                           'found in %s' % (name, path))
                 elif path[0] in rhodecode.BACKENDS:
                     backend = get_backend(path[0])
                     repos[name] = backend(path[1], config=config,
                                           with_wire={"cache": False})
             except OSError:
                 continue
             except RepositoryError:
                 log.exception('Failed to create a repo')
                 continue
 
         log.debug('found %s paths with repositories', len(repos))
         return repos
 
     def get_repos(self, all_repos=None, sort_key=None):
         """
         Get all repositories from db and for each repo create it's
         backend instance and fill that backed with information from database
 
         :param all_repos: list of repository names as strings
             give specific repositories list, good for filtering
 
         :param sort_key: initial sorting of repositories
         """
         if all_repos is None:
             all_repos = self.sa.query(Repository)\
                 .filter(Repository.group_id == null())\
                 .order_by(func.lower(Repository.repo_name)).all()
         repo_iter = SimpleCachedRepoList(
             all_repos, repos_path=self.repos_path, order_by=sort_key)
         return repo_iter
 
     def get_repo_groups(self, all_groups=None):
         if all_groups is None:
             all_groups = RepoGroup.query()\
                 .filter(RepoGroup.group_parent_id == null()).all()
         return [x for x in RepoGroupList(all_groups)]
 
     def mark_for_invalidation(self, repo_name, delete=False):
         """
         Mark caches of this repo invalid in the database. `delete` flag
         removes the cache entries
 
         :param repo_name: the repo_name for which caches should be marked
             invalid, or deleted
         :param delete: delete the entry keys instead of setting bool
             flag on them, and also purge caches used by the dogpile
         """
         repo = Repository.get_by_repo_name(repo_name)
 
         if repo:
             invalidation_namespace = CacheKey.REPO_INVALIDATION_NAMESPACE.format(
                 repo_id=repo.repo_id)
             CacheKey.set_invalidate(invalidation_namespace, delete=delete)
 
             repo_id = repo.repo_id
             config = repo._config
             config.set('extensions', 'largefiles', '')
             repo.update_commit_cache(config=config, cs_cache=None)
             if delete:
                 cache_namespace_uid = f'cache_repo.{repo_id}'
                 rc_cache.clear_cache_namespace('cache_repo', cache_namespace_uid, method=rc_cache.CLEAR_INVALIDATE)
 
     def toggle_following_repo(self, follow_repo_id, user_id):
 
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_repo_id == follow_repo_id)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         if f is not None:
             try:
                 self.sa.delete(f)
                 return
             except Exception:
                 log.error(traceback.format_exc())
                 raise
 
         try:
             f = UserFollowing()
             f.user_id = user_id
             f.follows_repo_id = follow_repo_id
             self.sa.add(f)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def toggle_following_user(self, follow_user_id, user_id):
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_user_id == follow_user_id)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         if f is not None:
             try:
                 self.sa.delete(f)
                 return
             except Exception:
                 log.error(traceback.format_exc())
                 raise
 
         try:
             f = UserFollowing()
             f.user_id = user_id
             f.follows_user_id = follow_user_id
             self.sa.add(f)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def is_following_repo(self, repo_name, user_id, cache=False):
         r = self.sa.query(Repository)\
             .filter(Repository.repo_name == repo_name).scalar()
 
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_repository == r)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         return f is not None
 
     def is_following_user(self, username, user_id, cache=False):
         u = User.get_by_username(username)
 
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_user == u)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         return f is not None
 
     def get_followers(self, repo):
         repo = self._get_repo(repo)
 
         return self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_repository == repo).count()
 
     def get_forks(self, repo):
         repo = self._get_repo(repo)
         return self.sa.query(Repository)\
             .filter(Repository.fork == repo).count()
 
     def get_pull_requests(self, repo):
         repo = self._get_repo(repo)
         return self.sa.query(PullRequest)\
             .filter(PullRequest.target_repo == repo)\
             .filter(PullRequest.status != PullRequest.STATUS_CLOSED).count()
 
     def get_artifacts(self, repo):
         repo = self._get_repo(repo)
         return self.sa.query(FileStore)\
             .filter(FileStore.repo == repo)\
             .filter(or_(FileStore.hidden == null(), FileStore.hidden == false())).count()
 
     def mark_as_fork(self, repo, fork, user):
         repo = self._get_repo(repo)
         fork = self._get_repo(fork)
         if fork and repo.repo_id == fork.repo_id:
             raise Exception("Cannot set repository as fork of itself")
 
         if fork and repo.repo_type != fork.repo_type:
             raise RepositoryError(
                 "Cannot set repository as fork of repository with other type")
 
         repo.fork = fork
         self.sa.add(repo)
         return repo
 
-    def pull_changes(self, repo, username, remote_uri=None, validate_uri=True):
+    def pull_changes(self, repo, username, remote_uri=None, validate_uri=True, **kwargs):
         dbrepo = self._get_repo(repo)
         remote_uri = remote_uri or dbrepo.clone_uri
         if not remote_uri:
             raise Exception("This repository doesn't have a clone uri")
 
         repo = dbrepo.scm_instance(cache=False)
         repo.config.clear_section('hooks')
 
         try:
             # NOTE(marcink): add extra validation so we skip invalid urls
             # this is due this tasks can be executed via scheduler without
             # proper validation of remote_uri
             if validate_uri:
                 config = make_db_config(clear_session=False)
                 url_validator(remote_uri, dbrepo.repo_type, config)
         except InvalidCloneUrl:
             raise
 
         repo_name = dbrepo.repo_name
         try:
             # TODO: we need to make sure those operations call proper hooks !
-            repo.fetch(remote_uri)
+            repo.fetch(remote_uri, **kwargs)
 
             self.mark_for_invalidation(repo_name)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def push_changes(self, repo, username, remote_uri=None, validate_uri=True):
         dbrepo = self._get_repo(repo)
         remote_uri = remote_uri or dbrepo.push_uri
         if not remote_uri:
             raise Exception("This repository doesn't have a clone uri")
 
         repo = dbrepo.scm_instance(cache=False)
         repo.config.clear_section('hooks')
 
         try:
             # NOTE(marcink): add extra validation so we skip invalid urls
             # this is due this tasks can be executed via scheduler without
             # proper validation of remote_uri
             if validate_uri:
                 config = make_db_config(clear_session=False)
                 url_validator(remote_uri, dbrepo.repo_type, config)
         except InvalidCloneUrl:
             raise
 
         try:
             repo.push(remote_uri)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def commit_change(self, repo, repo_name, commit, user, author, message,
                       content: bytes, f_path: bytes, branch: str = None):
         """
         Commits changes
         """
         user = self._get_user(user)
 
         # message and author needs to be unicode
         # proper backend should then translate that into required type
         message = safe_str(message)
         author = safe_str(author)
         imc = repo.in_memory_commit
         imc.change(FileNode(f_path, content, mode=commit.get_file_mode(f_path)))
         try:
             # TODO: handle pre-push action !
             tip = imc.commit(
                 message=message, author=author, parents=[commit],
                 branch=branch or commit.branch)
         except Exception as e:
             log.error(traceback.format_exc())
             raise IMCCommitError(str(e))
         finally:
             # always clear caches, if commit fails we want fresh object also
             self.mark_for_invalidation(repo_name)
 
         # We trigger the post-push action
         hooks_utils.trigger_post_push_hook(
             username=user.username, action='push_local', hook_type='post_push',
             repo_name=repo_name, repo_type=repo.alias, commit_ids=[tip.raw_id])
         return tip
 
     def _sanitize_path(self, f_path: bytes):
         if f_path.startswith(b'/') or f_path.startswith(b'./') or b'../' in f_path:
             raise NonRelativePathError(b'%b is not an relative path' % f_path)
         if f_path:
             f_path = os.path.normpath(f_path)
         return f_path
 
     def get_dirnode_metadata(self, request, commit, dir_node):
         if not dir_node.is_dir():
             return []
 
         data = []
         for node in dir_node:
             if not node.is_file():
                 # we skip file-nodes
                 continue
 
             last_commit = node.last_commit
             last_commit_date = last_commit.date
             data.append({
                 'name': node.name,
                 'size': h.format_byte_size_binary(node.size),
                 'modified_at': h.format_date(last_commit_date),
                 'modified_ts': last_commit_date.isoformat(),
                 'revision': last_commit.revision,
                 'short_id': last_commit.short_id,
                 'message': h.escape(last_commit.message),
                 'author': h.escape(last_commit.author),
                 'user_profile': h.gravatar_with_user(
                     request, last_commit.author),
             })
 
         return data
 
     def get_nodes(self, repo_name, commit_id, root_path='/', flat=True,
                   extended_info=False, content=False, max_file_bytes=None):
         """
         recursive walk in root dir and return a set of all path in that dir
         based on repository walk function
 
         :param repo_name: name of repository
         :param commit_id: commit id for which to list nodes
         :param root_path: root path to list
         :param flat: return as a list, if False returns a dict with description
         :param extended_info: show additional info such as md5, binary, size etc
         :param content: add nodes content to the return data
         :param max_file_bytes: will not return file contents over this limit
 
         """
         _files = list()
         _dirs = list()
 
         try:
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
             root_path = root_path.lstrip('/')
 
             # get RootNode, inject pre-load options before walking
             top_node = commit.get_node(root_path)
             extended_info_pre_load = []
             if extended_info:
                 extended_info_pre_load += ['md5']
             top_node.default_pre_load = ['is_binary', 'size'] + extended_info_pre_load
 
             for __, dirs, files in commit.walk(top_node):
 
                 for f in files:
                     _content = None
                     _data = f_name = f.str_path
 
                     if not flat:
                         _data = {
                             "name": h.escape(f_name),
                             "type": "file",
                             }
                         if extended_info:
                             _data.update({
                                 "md5": f.md5,
                                 "binary": f.is_binary,
                                 "size": f.size,
                                 "extension": f.extension,
                                 "mimetype": f.mimetype,
                                 "lines": f.lines()[0]
                             })
 
                         if content:
                             over_size_limit = (max_file_bytes is not None
                                                and f.size > max_file_bytes)
                             full_content = None
                             if not f.is_binary and not over_size_limit:
                                 full_content = f.str_content
 
                             _data.update({
                                 "content": full_content,
                             })
                     _files.append(_data)
 
                 for d in dirs:
                     _data = d_name = d.str_path
                     if not flat:
                         _data = {
                             "name": h.escape(d_name),
                             "type": "dir",
                             }
                     if extended_info:
                         _data.update({
                             "md5": "",
                             "binary": False,
                             "size": 0,
                             "extension": "",
                         })
                     if content:
                         _data.update({
                             "content": None
                         })
                     _dirs.append(_data)
         except RepositoryError:
             log.exception("Exception in get_nodes")
             raise
 
         return _dirs, _files
 
     def get_quick_filter_nodes(self, repo_name, commit_id, root_path='/'):
         """
         Generate files for quick filter in files view
         """
 
         _files = list()
         _dirs = list()
         try:
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
             root_path = root_path.lstrip('/')
 
             top_node = commit.get_node(root_path)
             top_node.default_pre_load = []
 
             for __, dirs, files in commit.walk(top_node):
                 for f in files:
 
                     _data = {
                         "name": h.escape(f.str_path),
                         "type": "file",
                         }
 
                     _files.append(_data)
 
                 for d in dirs:
 
                     _data = {
                         "name": h.escape(d.str_path),
                         "type": "dir",
                         }
 
                     _dirs.append(_data)
         except RepositoryError:
             log.exception("Exception in get_quick_filter_nodes")
             raise
 
         return _dirs, _files
 
     def get_node(self, repo_name, commit_id, file_path,
                  extended_info=False, content=False, max_file_bytes=None, cache=True):
         """
         retrieve single node from commit
         """
 
         try:
 
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
 
             file_node = commit.get_node(file_path)
             if file_node.is_dir():
                 raise RepositoryError('The given path is a directory')
 
             _content = None
             f_name = file_node.str_path
 
             file_data = {
                 "name": h.escape(f_name),
                 "type": "file",
             }
 
             if extended_info:
                 file_data.update({
                     "extension": file_node.extension,
                     "mimetype": file_node.mimetype,
                 })
 
                 if cache:
                     md5 = file_node.md5
                     is_binary = file_node.is_binary
                     size = file_node.size
                 else:
                     is_binary, md5, size, _content = file_node.metadata_uncached()
 
                 file_data.update({
                     "md5": md5,
                     "binary": is_binary,
                     "size": size,
                 })
 
             if content and cache:
                 # get content + cache
                 size = file_node.size
                 over_size_limit = (max_file_bytes is not None and size > max_file_bytes)
                 full_content = None
                 all_lines = 0
                 if not file_node.is_binary and not over_size_limit:
                     full_content = safe_str(file_node.content)
                     all_lines, empty_lines = file_node.count_lines(full_content)
 
                 file_data.update({
                     "content": full_content,
                     "lines": all_lines
                 })
             elif content:
                 # get content *without* cache
                 if _content is None:
                     is_binary, md5, size, _content = file_node.metadata_uncached()
 
                 over_size_limit = (max_file_bytes is not None and size > max_file_bytes)
                 full_content = None
                 all_lines = 0
                 if not is_binary and not over_size_limit:
                     full_content = safe_str(_content)
                     all_lines, empty_lines = file_node.count_lines(full_content)
 
                 file_data.update({
                     "content": full_content,
                     "lines": all_lines
                 })
 
         except RepositoryError:
             log.exception("Exception in get_node")
             raise
 
         return file_data
 
     def get_fts_data(self, repo_name, commit_id, root_path='/'):
         """
         Fetch node tree for usage in full text search
         """
 
         tree_info = list()
 
         try:
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
             root_path = root_path.lstrip('/')
             top_node = commit.get_node(root_path)
             top_node.default_pre_load = []
 
             for __, dirs, files in commit.walk(top_node):
 
                 for f in files:
                     is_binary, md5, size, _content = f.metadata_uncached()
                     _data = {
                         "name": f.str_path,
                         "md5": md5,
                         "extension": f.extension,
                         "binary": is_binary,
                         "size": size
                     }
 
                     tree_info.append(_data)
 
         except RepositoryError:
             log.exception("Exception in get_nodes")
             raise
 
         return tree_info
 
     def create_nodes(self, user, repo, message, nodes, parent_commit=None,
                      author=None, trigger_push_hook=True):
         """
         Commits given multiple nodes into repo
 
         :param user: RhodeCode User object or user_id, the commiter
         :param repo: RhodeCode Repository object
         :param message: commit message
         :param nodes: mapping {filename:{'content':content},...}
         :param parent_commit: parent commit, can be empty than it's
            initial commit
         :param author: author of commit, cna be different that commiter
            only for git
         :param trigger_push_hook: trigger push hooks
 
         :returns: new committed commit
         """
 
         user = self._get_user(user)
         scm_instance = repo.scm_instance(cache=False)
 
         message = safe_str(message)
         commiter = user.full_contact
         author = safe_str(author) if author else commiter
 
         imc = scm_instance.in_memory_commit
 
         if not parent_commit:
             parent_commit = EmptyCommit(alias=scm_instance.alias)
 
         if isinstance(parent_commit, EmptyCommit):
             # EmptyCommit means we're editing empty repository
             parents = None
         else:
             parents = [parent_commit]
 
         upload_file_types = (io.BytesIO, io.BufferedRandom)
         processed_nodes = []
         for filename, content_dict in nodes.items():
             if not isinstance(filename, bytes):
                 raise ValueError(f'filename key in nodes needs to be bytes , or {upload_file_types}')
             content = content_dict['content']
             if not isinstance(content, upload_file_types + (bytes,)):
                 raise ValueError('content key value in nodes needs to be bytes')
 
         for f_path in nodes:
             f_path = self._sanitize_path(f_path)
             content = nodes[f_path]['content']
 
             # decoding here will force that we have proper encoded values
             # in any other case this will throw exceptions and deny commit
 
             if isinstance(content, bytes):
                 pass
             elif isinstance(content, upload_file_types):
                 content = content.read()
             else:
                 raise Exception(f'Content is of unrecognized type {type(content)}, expected {upload_file_types}')
             processed_nodes.append((f_path, content))
 
         # add multiple nodes
         for path, content in processed_nodes:
             imc.add(FileNode(path, content=content))
 
         # TODO: handle pre push scenario
         tip = imc.commit(message=message,
                          author=author,
                          parents=parents,
                          branch=parent_commit.branch)
 
         self.mark_for_invalidation(repo.repo_name)
         if trigger_push_hook:
             hooks_utils.trigger_post_push_hook(
                 username=user.username, action='push_local',
                 repo_name=repo.repo_name, repo_type=scm_instance.alias,
                 hook_type='post_push',
                 commit_ids=[tip.raw_id])
         return tip
 
     def update_nodes(self, user, repo, message, nodes, parent_commit=None,
                      author=None, trigger_push_hook=True):
         user = self._get_user(user)
         scm_instance = repo.scm_instance(cache=False)
 
         message = safe_str(message)
         commiter = user.full_contact
         author = safe_str(author) if author else commiter
 
         imc = scm_instance.in_memory_commit
 
         if not parent_commit:
             parent_commit = EmptyCommit(alias=scm_instance.alias)
 
         if isinstance(parent_commit, EmptyCommit):
             # EmptyCommit means we we're editing empty repository
             parents = None
         else:
             parents = [parent_commit]
 
         # add multiple nodes
         for _filename, data in nodes.items():
             # new filename, can be renamed from the old one, also sanitaze
             # the path for any hack around relative paths like ../../ etc.
             filename = self._sanitize_path(data['filename'])
             old_filename = self._sanitize_path(_filename)
             content = data['content']
             file_mode = data.get('mode')
             filenode = FileNode(old_filename, content=content, mode=file_mode)
             op = data['op']
             if op == 'add':
                 imc.add(filenode)
             elif op == 'del':
                 imc.remove(filenode)
             elif op == 'mod':
                 if filename != old_filename:
                     # TODO: handle renames more efficient, needs vcs lib changes
                     imc.remove(filenode)
                     imc.add(FileNode(filename, content=content, mode=file_mode))
                 else:
                     imc.change(filenode)
 
         try:
             # TODO: handle pre push scenario commit changes
             tip = imc.commit(message=message,
                              author=author,
                              parents=parents,
                              branch=parent_commit.branch)
         except NodeNotChangedError:
             raise
         except Exception as e:
             log.exception("Unexpected exception during call to imc.commit")
             raise IMCCommitError(str(e))
         finally:
             # always clear caches, if commit fails we want fresh object also
             self.mark_for_invalidation(repo.repo_name)
 
         if trigger_push_hook:
             hooks_utils.trigger_post_push_hook(
                 username=user.username, action='push_local', hook_type='post_push',
                 repo_name=repo.repo_name, repo_type=scm_instance.alias,
                 commit_ids=[tip.raw_id])
 
         return tip
 
     def delete_nodes(self, user, repo, message, nodes, parent_commit=None,
                      author=None, trigger_push_hook=True):
         """
         Deletes given multiple nodes into `repo`
 
         :param user: RhodeCode User object or user_id, the committer
         :param repo: RhodeCode Repository object
         :param message: commit message
         :param nodes: mapping {filename:{'content':content},...}
         :param parent_commit: parent commit, can be empty than it's initial
            commit
         :param author: author of commit, cna be different that commiter only
            for git
         :param trigger_push_hook: trigger push hooks
 
         :returns: new commit after deletion
         """
 
         user = self._get_user(user)
         scm_instance = repo.scm_instance(cache=False)
 
         processed_nodes = []
         for f_path in nodes:
             f_path = self._sanitize_path(f_path)
             # content can be empty but for compatibility it allows same dicts
             # structure as add_nodes
             content = nodes[f_path].get('content')
             processed_nodes.append((safe_bytes(f_path), content))
 
         message = safe_str(message)
         commiter = user.full_contact
         author = safe_str(author) if author else commiter
 
         imc = scm_instance.in_memory_commit
 
         if not parent_commit:
             parent_commit = EmptyCommit(alias=scm_instance.alias)
 
         if isinstance(parent_commit, EmptyCommit):
             # EmptyCommit means we we're editing empty repository
             parents = None
         else:
             parents = [parent_commit]
         # add multiple nodes
         for path, content in processed_nodes:
             imc.remove(FileNode(path, content=content))
 
         # TODO: handle pre push scenario
         tip = imc.commit(message=message,
                          author=author,
                          parents=parents,
                          branch=parent_commit.branch)
 
         self.mark_for_invalidation(repo.repo_name)
         if trigger_push_hook:
             hooks_utils.trigger_post_push_hook(
                 username=user.username, action='push_local', hook_type='post_push',
                 repo_name=repo.repo_name, repo_type=scm_instance.alias,
                 commit_ids=[tip.raw_id])
         return tip
 
     def strip(self, repo, commit_id, branch):
         scm_instance = repo.scm_instance(cache=False)
         scm_instance.config.clear_section('hooks')
         scm_instance.strip(commit_id, branch)
         self.mark_for_invalidation(repo.repo_name)
 
     def get_unread_journal(self):
         return self.sa.query(UserLog).count()
 
     @classmethod
     def backend_landing_ref(cls, repo_type):
         """
         Return a default landing ref based on a repository type.
         """
 
         landing_ref = {
             'hg': ('branch:default', 'default'),
             'git': ('branch:master', 'master'),
             'svn': ('rev:tip', 'latest tip'),
             'default': ('rev:tip', 'latest tip'),
         }
 
         return landing_ref.get(repo_type) or landing_ref['default']
 
     def get_repo_landing_revs(self, translator, repo=None):
         """
         Generates select option with tags branches and bookmarks (for hg only)
         grouped by type
 
         :param repo:
         """
         from rhodecode.lib.vcs.backends.git import GitRepository
 
         _ = translator
         repo = self._get_repo(repo)
 
         if repo:
             repo_type = repo.repo_type
         else:
             repo_type = 'default'
 
         default_landing_ref, landing_ref_lbl = self.backend_landing_ref(repo_type)
 
         default_ref_options = [
             [default_landing_ref, landing_ref_lbl]
         ]
         default_choices = [
             default_landing_ref
         ]
 
         if not repo:
             # presented at NEW repo creation
             return default_choices, default_ref_options
 
         repo = repo.scm_instance()
 
         ref_options = [(default_landing_ref, landing_ref_lbl)]
         choices = [default_landing_ref]
 
         # branches
         branch_group = [(f'branch:{safe_str(b)}', safe_str(b)) for b in repo.branches]
         if not branch_group:
             # new repo, or without maybe a branch?
             branch_group = default_ref_options
 
         branches_group = (branch_group, _("Branches"))
         ref_options.append(branches_group)
         choices.extend([x[0] for x in branches_group[0]])
 
         # bookmarks for HG
         if repo.alias == 'hg':
             bookmarks_group = (
                 [(f'book:{safe_str(b)}', safe_str(b))
                     for b in repo.bookmarks],
                 _("Bookmarks"))
             ref_options.append(bookmarks_group)
             choices.extend([x[0] for x in bookmarks_group[0]])
 
         # tags
         tags_group = (
             [(f'tag:{safe_str(t)}', safe_str(t))
                 for t in repo.tags],
             _("Tags"))
         ref_options.append(tags_group)
         choices.extend([x[0] for x in tags_group[0]])
 
         return choices, ref_options
 
     def get_server_info(self, environ=None):
         server_info = get_system_info(environ)
         return server_info