api: exposed modified added/modified/deleted functions of commit to return only paths.

This combined with full text search build for certain repositories that each commit
contains lots of file changes results in *huge* (10x in some cases) performance gain.

rhodecode/api/utils.py

@@ -1,453 +1,453 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2014-2019 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/
 
 """
 JSON RPC utils
 """
 
 import collections
 import logging
 
 from rhodecode.api.exc import JSONRPCError
 from rhodecode.lib.auth import (
     HasPermissionAnyApi, HasRepoPermissionAnyApi, HasRepoGroupPermissionAnyApi)
 from rhodecode.lib.utils import safe_unicode
 from rhodecode.lib.vcs.exceptions import RepositoryError
 from rhodecode.lib.view_utils import get_commit_from_ref_name
 from rhodecode.lib.utils2 import str2bool
 
 log = logging.getLogger(__name__)
 
 
 class OAttr(object):
     """
     Special Option that defines other attribute, and can default to them
 
     Example::
 
         def test(apiuser, userid=Optional(OAttr('apiuser')):
             user = Optional.extract(userid, evaluate_locals=local())
             #if we pass in userid, we get it, else it will default to apiuser
             #attribute
     """
 
     def __init__(self, attr_name):
         self.attr_name = attr_name
 
     def __repr__(self):
         return '<OptionalAttr:%s>' % self.attr_name
 
     def __call__(self):
         return self
 
 
 class Optional(object):
     """
     Defines an optional parameter::
 
         param = param.getval() if isinstance(param, Optional) else param
         param = param() if isinstance(param, Optional) else param
 
     is equivalent of::
 
         param = Optional.extract(param)
 
     """
 
     def __init__(self, type_):
         self.type_ = type_
 
     def __repr__(self):
         return '<Optional:%s>' % self.type_.__repr__()
 
     def __call__(self):
         return self.getval()
 
     def getval(self, evaluate_locals=None):
         """
         returns value from this Optional instance
         """
         if isinstance(self.type_, OAttr):
             param_name = self.type_.attr_name
             if evaluate_locals:
                 return evaluate_locals[param_name]
             # use params name
             return param_name
         return self.type_
 
     @classmethod
     def extract(cls, val, evaluate_locals=None, binary=None):
         """
         Extracts value from Optional() instance
 
         :param val:
         :return: original value if it's not Optional instance else
             value of instance
         """
         if isinstance(val, cls):
             val = val.getval(evaluate_locals)
 
         if binary:
             val = str2bool(val)
 
         return val
 
 
 def parse_args(cli_args, key_prefix=''):
     from rhodecode.lib.utils2 import (escape_split)
     kwargs = collections.defaultdict(dict)
     for el in escape_split(cli_args, ','):
         kv = escape_split(el, '=', 1)
         if len(kv) == 2:
             k, v = kv
             kwargs[key_prefix + k] = v
     return kwargs
 
 
 def get_origin(obj):
     """
     Get origin of permission from object.
 
     :param obj:
     """
     origin = 'permission'
 
     if getattr(obj, 'owner_row', '') and getattr(obj, 'admin_row', ''):
         # admin and owner case, maybe we should use dual string ?
         origin = 'owner'
     elif getattr(obj, 'owner_row', ''):
         origin = 'owner'
     elif getattr(obj, 'admin_row', ''):
         origin = 'super-admin'
     return origin
 
 
 def store_update(updates, attr, name):
     """
     Stores param in updates dict if it's not instance of Optional
     allows easy updates of passed in params
     """
     if not isinstance(attr, Optional):
         updates[name] = attr
 
 
 def has_superadmin_permission(apiuser):
     """
     Return True if apiuser is admin or return False
 
     :param apiuser:
     """
     if HasPermissionAnyApi('hg.admin')(user=apiuser):
         return True
     return False
 
 
 def validate_repo_permissions(apiuser, repoid, repo, perms):
     """
     Raise JsonRPCError if apiuser is not authorized or return True
 
     :param apiuser:
     :param repoid:
     :param repo:
     :param perms:
     """
     if not HasRepoPermissionAnyApi(*perms)(
             user=apiuser, repo_name=repo.repo_name):
         raise JSONRPCError(
             'repository `%s` does not exist' % repoid)
 
     return True
 
 
 def validate_repo_group_permissions(apiuser, repogroupid, repo_group, perms):
     """
     Raise JsonRPCError if apiuser is not authorized or return True
 
     :param apiuser:
     :param repogroupid: just the id of repository group
     :param repo_group: instance of repo_group
     :param perms:
     """
     if not HasRepoGroupPermissionAnyApi(*perms)(
             user=apiuser, group_name=repo_group.group_name):
         raise JSONRPCError(
             'repository group `%s` does not exist' % repogroupid)
 
     return True
 
 
 def validate_set_owner_permissions(apiuser, owner):
     if isinstance(owner, Optional):
         owner = get_user_or_error(apiuser.user_id)
     else:
         if has_superadmin_permission(apiuser):
             owner = get_user_or_error(owner)
         else:
             # forbid setting owner for non-admins
             raise JSONRPCError(
                 'Only RhodeCode super-admin can specify `owner` param')
     return owner
 
 
 def get_user_or_error(userid):
     """
     Get user by id or name or return JsonRPCError if not found
 
     :param userid:
     """
     from rhodecode.model.user import UserModel
     user_model = UserModel()
 
     if isinstance(userid, (int, long)):
         try:
             user = user_model.get_user(userid)
         except ValueError:
             user = None
     else:
         user = user_model.get_by_username(userid)
 
     if user is None:
         raise JSONRPCError(
             'user `%s` does not exist' % (userid,))
     return user
 
 
 def get_repo_or_error(repoid):
     """
     Get repo by id or name or return JsonRPCError if not found
 
     :param repoid:
     """
     from rhodecode.model.repo import RepoModel
     repo_model = RepoModel()
 
     if isinstance(repoid, (int, long)):
         try:
             repo = repo_model.get_repo(repoid)
         except ValueError:
             repo = None
     else:
         repo = repo_model.get_by_repo_name(repoid)
 
     if repo is None:
         raise JSONRPCError(
             'repository `%s` does not exist' % (repoid,))
     return repo
 
 
 def get_repo_group_or_error(repogroupid):
     """
     Get repo group by id or name or return JsonRPCError if not found
 
     :param repogroupid:
     """
     from rhodecode.model.repo_group import RepoGroupModel
     repo_group_model = RepoGroupModel()
 
     if isinstance(repogroupid, (int, long)):
         try:
             repo_group = repo_group_model._get_repo_group(repogroupid)
         except ValueError:
             repo_group = None
     else:
         repo_group = repo_group_model.get_by_group_name(repogroupid)
 
     if repo_group is None:
         raise JSONRPCError(
             'repository group `%s` does not exist' % (repogroupid,))
     return repo_group
 
 
 def get_user_group_or_error(usergroupid):
     """
     Get user group by id or name or return JsonRPCError if not found
 
     :param usergroupid:
     """
     from rhodecode.model.user_group import UserGroupModel
     user_group_model = UserGroupModel()
 
     if isinstance(usergroupid, (int, long)):
         try:
             user_group = user_group_model.get_group(usergroupid)
         except ValueError:
             user_group = None
     else:
         user_group = user_group_model.get_by_name(usergroupid)
 
     if user_group is None:
         raise JSONRPCError(
             'user group `%s` does not exist' % (usergroupid,))
     return user_group
 
 
 def get_perm_or_error(permid, prefix=None):
     """
     Get permission by id or name or return JsonRPCError if not found
 
     :param permid:
     """
     from rhodecode.model.permission import PermissionModel
 
     perm = PermissionModel.cls.get_by_key(permid)
     if perm is None:
         msg = 'permission `{}` does not exist.'.format(permid)
         if prefix:
             msg += ' Permission should start with prefix: `{}`'.format(prefix)
         raise JSONRPCError(msg)
 
     if prefix:
         if not perm.permission_name.startswith(prefix):
             raise JSONRPCError('permission `%s` is invalid, '
                                'should start with %s' % (permid, prefix))
     return perm
 
 
 def get_gist_or_error(gistid):
     """
     Get gist by id or gist_access_id or return JsonRPCError if not found
 
     :param gistid:
     """
     from rhodecode.model.gist import GistModel
 
     gist = GistModel.cls.get_by_access_id(gistid)
     if gist is None:
         raise JSONRPCError('gist `%s` does not exist' % (gistid,))
     return gist
 
 
 def get_pull_request_or_error(pullrequestid):
     """
     Get pull request by id or return JsonRPCError if not found
 
     :param pullrequestid:
     """
     from rhodecode.model.pull_request import PullRequestModel
 
     try:
         pull_request = PullRequestModel().get(int(pullrequestid))
     except ValueError:
         raise JSONRPCError('pullrequestid must be an integer')
     if not pull_request:
         raise JSONRPCError('pull request `%s` does not exist' % (
             pullrequestid,))
     return pull_request
 
 
 def build_commit_data(commit, detail_level):
     parsed_diff = []
     if detail_level == 'extended':
-        for f in commit.added:
-            parsed_diff.append(_get_commit_dict(filename=f.path, op='A'))
-        for f in commit.changed:
-            parsed_diff.append(_get_commit_dict(filename=f.path, op='M'))
-        for f in commit.removed:
-            parsed_diff.append(_get_commit_dict(filename=f.path, op='D'))
+        for f_path in commit.added_paths:
+            parsed_diff.append(_get_commit_dict(filename=f_path, op='A'))
+        for f_path in commit.changed_paths:
+            parsed_diff.append(_get_commit_dict(filename=f_path, op='M'))
+        for f_path in commit.removed_paths:
+            parsed_diff.append(_get_commit_dict(filename=f_path, op='D'))
 
     elif detail_level == 'full':
         from rhodecode.lib.diffs import DiffProcessor
         diff_processor = DiffProcessor(commit.diff())
         for dp in diff_processor.prepare():
             del dp['stats']['ops']
             _stats = dp['stats']
             parsed_diff.append(_get_commit_dict(
                 filename=dp['filename'], op=dp['operation'],
                 new_revision=dp['new_revision'],
                 old_revision=dp['old_revision'],
                 raw_diff=dp['raw_diff'], stats=_stats))
 
     return parsed_diff
 
 
 def get_commit_or_error(ref, repo):
     try:
         ref_type, _, ref_hash = ref.split(':')
     except ValueError:
         raise JSONRPCError(
             'Ref `{ref}` given in a wrong format. Please check the API'
             ' documentation for more details'.format(ref=ref))
     try:
         # TODO: dan: refactor this to use repo.scm_instance().get_commit()
         # once get_commit supports ref_types
         return get_commit_from_ref_name(repo, ref_hash)
     except RepositoryError:
         raise JSONRPCError('Ref `{ref}` does not exist'.format(ref=ref))
 
 
 def _get_ref_hash(repo, type_, name):
     vcs_repo = repo.scm_instance()
     if type_ in ['branch'] and vcs_repo.alias in ('hg', 'git'):
         return vcs_repo.branches[name]
     elif type_ in ['bookmark', 'book'] and vcs_repo.alias == 'hg':
         return vcs_repo.bookmarks[name]
     else:
         raise ValueError()
 
 
 def resolve_ref_or_error(ref, repo, allowed_ref_types=None):
     allowed_ref_types = allowed_ref_types or ['bookmark', 'book', 'tag', 'branch']
 
     def _parse_ref(type_, name, hash_=None):
         return type_, name, hash_
 
     try:
         ref_type, ref_name, ref_hash = _parse_ref(*ref.split(':'))
     except TypeError:
         raise JSONRPCError(
             'Ref `{ref}` given in a wrong format. Please check the API'
             ' documentation for more details'.format(ref=ref))
 
     if ref_type not in allowed_ref_types:
         raise JSONRPCError(
             'Ref `{ref}` type is not allowed. '
             'Only:{allowed_refs} are possible.'.format(
                 ref=ref, allowed_refs=allowed_ref_types))
 
     try:
         ref_hash = ref_hash or _get_ref_hash(repo, ref_type, ref_name)
     except (KeyError, ValueError):
         raise JSONRPCError(
             'The specified value:{type}:`{name}` does not exist, or is not allowed.'.format(
                 type=ref_type, name=ref_name))
 
     return ':'.join([ref_type, ref_name, ref_hash])
 
 
 def _get_commit_dict(
         filename, op, new_revision=None, old_revision=None,
         raw_diff=None, stats=None):
     if stats is None:
         stats = {
             "added": None,
             "binary": None,
             "deleted": None
         }
     return {
         "filename": safe_unicode(filename),
         "op": op,
 
         # extra details
         "new_revision": new_revision,
         "old_revision": old_revision,
 
         "raw_diff": raw_diff,
         "stats": stats
     }

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

@@ -1,484 +1,493 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2014-2019 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 """
 GIT commit module
 """
 
 import re
 import stat
 from itertools import chain
 from StringIO import StringIO
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 from rhodecode.lib.datelib import utcdate_fromtimestamp
 from rhodecode.lib.utils import safe_unicode, safe_str
 from rhodecode.lib.utils2 import safe_int
 from rhodecode.lib.vcs.conf import settings
 from rhodecode.lib.vcs.backends import base
 from rhodecode.lib.vcs.exceptions import CommitError, NodeDoesNotExistError
 from rhodecode.lib.vcs.nodes import (
     FileNode, DirNode, NodeKind, RootNode, SubModuleNode,
     ChangedFileNodesGenerator, AddedFileNodesGenerator,
     RemovedFileNodesGenerator, LargeFileNode)
 from rhodecode.lib.vcs.compat import configparser
 
 
 class GitCommit(base.BaseCommit):
     """
     Represents state of the repository at single commit id.
     """
 
     _filter_pre_load = [
         # done through a more complex tree walk on parents
         "affected_files",
         # done through subprocess not remote call
         "children",
         # done through a more complex tree walk on parents
         "status",
         # mercurial specific property not supported here
         "_file_paths",
         # mercurial specific property not supported here
         'obsolete',
         # mercurial specific property not supported here
         'phase',
         # mercurial specific property not supported here
         'hidden'
     ]
 
     def __init__(self, repository, raw_id, idx, pre_load=None):
         self.repository = repository
         self._remote = repository._remote
         # TODO: johbo: Tweak of raw_id should not be necessary
         self.raw_id = safe_str(raw_id)
         self.idx = idx
 
         self._set_bulk_properties(pre_load)
 
         # caches
         self._stat_modes = {}  # stat info for paths
         self._paths = {}  # path processed with parse_tree
         self.nodes = {}
         self._submodules = None
 
     def _set_bulk_properties(self, pre_load):
 
         if not pre_load:
             return
         pre_load = [entry for entry in pre_load
                     if entry not in self._filter_pre_load]
         if not pre_load:
             return
 
         result = self._remote.bulk_request(self.raw_id, pre_load)
         for attr, value in result.items():
             if attr in ["author", "message"]:
                 if value:
                     value = safe_unicode(value)
             elif attr == "date":
                 value = utcdate_fromtimestamp(*value)
             elif attr == "parents":
                 value = self._make_commits(value)
             elif attr == "branch":
                 value = value[0] if value else None
             self.__dict__[attr] = value
 
     @LazyProperty
     def _commit(self):
         return self._remote[self.raw_id]
 
     @LazyProperty
     def _tree_id(self):
         return self._remote[self._commit['tree']]['id']
 
     @LazyProperty
     def id(self):
         return self.raw_id
 
     @LazyProperty
     def short_id(self):
         return self.raw_id[:12]
 
     @LazyProperty
     def message(self):
         return safe_unicode(self._remote.message(self.id))
 
     @LazyProperty
     def committer(self):
         return safe_unicode(self._remote.author(self.id))
 
     @LazyProperty
     def author(self):
         return safe_unicode(self._remote.author(self.id))
 
     @LazyProperty
     def date(self):
         unix_ts, tz = self._remote.date(self.raw_id)
         return utcdate_fromtimestamp(unix_ts, tz)
 
     @LazyProperty
     def status(self):
         """
         Returns modified, added, removed, deleted files for current commit
         """
         return self.changed, self.added, self.removed
 
     @LazyProperty
     def tags(self):
         tags = [safe_unicode(name) for name,
                 commit_id in self.repository.tags.iteritems()
                 if commit_id == self.raw_id]
         return tags
 
     @LazyProperty
     def commit_branches(self):
         branches = []
         for name, commit_id in self.repository.branches.iteritems():
             if commit_id == self.raw_id:
                 branches.append(name)
         return branches
 
     @LazyProperty
     def branch(self):
         branches = self._remote.branch(self.raw_id)
 
         if branches:
             # actually commit can have multiple branches in git
             return safe_unicode(branches[0])
 
     def _get_tree_id_for_path(self, path):
         path = safe_str(path)
         if path in self._paths:
             return self._paths[path]
 
         tree_id = self._tree_id
 
         path = path.strip('/')
         if path == '':
             data = [tree_id, "tree"]
             self._paths[''] = data
             return data
 
         tree_id, tree_type, tree_mode = \
             self._remote.tree_and_type_for_path(self.raw_id, path)
         if tree_id is None:
             raise self.no_node_at_path(path)
 
         self._paths[path] = [tree_id, tree_type]
         self._stat_modes[path] = tree_mode
 
         if path not in self._paths:
             raise self.no_node_at_path(path)
 
         return self._paths[path]
 
     def _get_kind(self, path):
         tree_id, type_ = self._get_tree_id_for_path(path)
         if type_ == 'blob':
             return NodeKind.FILE
         elif type_ == 'tree':
             return NodeKind.DIR
         elif type_ == 'link':
             return NodeKind.SUBMODULE
         return None
 
     def _get_filectx(self, path):
         path = self._fix_path(path)
         if self._get_kind(path) != NodeKind.FILE:
             raise CommitError(
                 "File does not exist for commit %s at  '%s'" % (self.raw_id, path))
         return path
 
     def _get_file_nodes(self):
         return chain(*(t[2] for t in self.walk()))
 
     @LazyProperty
     def parents(self):
         """
         Returns list of parent commits.
         """
         parent_ids = self._remote.parents(self.id)
         return self._make_commits(parent_ids)
 
     @LazyProperty
     def children(self):
         """
         Returns list of child commits.
         """
 
         children = self._remote.children(self.raw_id)
         return self._make_commits(children)
 
     def _make_commits(self, commit_ids):
         def commit_maker(_commit_id):
             return self.repository.get_commit(commit_id=commit_id)
 
         return [commit_maker(commit_id) for commit_id in commit_ids]
 
     def get_file_mode(self, path):
         """
         Returns stat mode of the file at the given `path`.
         """
         path = safe_str(path)
         # ensure path is traversed
         self._get_tree_id_for_path(path)
         return self._stat_modes[path]
 
     def is_link(self, path):
         return stat.S_ISLNK(self.get_file_mode(path))
 
     def is_node_binary(self, path):
         tree_id, _ = self._get_tree_id_for_path(path)
         return self._remote.is_binary(tree_id)
 
     def get_file_content(self, path):
         """
         Returns content of the file at given `path`.
         """
         tree_id, _ = self._get_tree_id_for_path(path)
         return self._remote.blob_as_pretty_string(tree_id)
 
     def get_file_content_streamed(self, path):
         tree_id, _ = self._get_tree_id_for_path(path)
         stream_method = getattr(self._remote, 'stream:blob_as_pretty_string')
         return stream_method(tree_id)
 
     def get_file_size(self, path):
         """
         Returns size of the file at given `path`.
         """
         tree_id, _ = self._get_tree_id_for_path(path)
         return self._remote.blob_raw_length(tree_id)
 
     def get_path_history(self, path, limit=None, pre_load=None):
         """
         Returns history of file as reversed list of `GitCommit` objects for
         which file at given `path` has been modified.
         """
 
         path = self._get_filectx(path)
         hist = self._remote.node_history(self.raw_id, path, limit)
         return [
             self.repository.get_commit(commit_id=commit_id, pre_load=pre_load)
             for commit_id in hist]
 
     def get_file_annotate(self, path, pre_load=None):
         """
         Returns a generator of four element tuples with
             lineno, commit_id, commit lazy loader and line
         """
 
         result = self._remote.node_annotate(self.raw_id, path)
 
         for ln_no, commit_id, content in result:
             yield (
                 ln_no, commit_id,
                 lambda: self.repository.get_commit(commit_id=commit_id, pre_load=pre_load),
                 content)
 
     def get_nodes(self, path):
 
         if self._get_kind(path) != NodeKind.DIR:
             raise CommitError(
                 "Directory does not exist for commit %s at '%s'" % (self.raw_id, path))
         path = self._fix_path(path)
 
         tree_id, _ = self._get_tree_id_for_path(path)
 
         dirnodes = []
         filenodes = []
 
         # extracted tree ID gives us our files...
         for name, stat_, id_, type_ in self._remote.tree_items(tree_id):
             if type_ == 'link':
                 url = self._get_submodule_url('/'.join((path, name)))
                 dirnodes.append(SubModuleNode(
                     name, url=url, commit=id_, alias=self.repository.alias))
                 continue
 
             if path != '':
                 obj_path = '/'.join((path, name))
             else:
                 obj_path = name
             if obj_path not in self._stat_modes:
                 self._stat_modes[obj_path] = stat_
 
             if type_ == 'tree':
                 dirnodes.append(DirNode(obj_path, commit=self))
             elif type_ == 'blob':
                 filenodes.append(FileNode(obj_path, commit=self, mode=stat_))
             else:
                 raise CommitError(
                     "Requested object should be Tree or Blob, is %s", type_)
 
         nodes = dirnodes + filenodes
         for node in nodes:
             if node.path not in self.nodes:
                 self.nodes[node.path] = node
         nodes.sort()
         return nodes
 
     def get_node(self, path, pre_load=None):
         if isinstance(path, unicode):
             path = path.encode('utf-8')
         path = self._fix_path(path)
         if path not in self.nodes:
             try:
                 tree_id, type_ = self._get_tree_id_for_path(path)
             except CommitError:
                 raise NodeDoesNotExistError(
                     "Cannot find one of parents' directories for a given "
                     "path: %s" % path)
 
             if type_ in ['link', 'commit']:
                 url = self._get_submodule_url(path)
                 node = SubModuleNode(path, url=url, commit=tree_id,
                                      alias=self.repository.alias)
             elif type_ == 'tree':
                 if path == '':
                     node = RootNode(commit=self)
                 else:
                     node = DirNode(path, commit=self)
             elif type_ == 'blob':
                 node = FileNode(path, commit=self, pre_load=pre_load)
                 self._stat_modes[path] = node.mode
             else:
                 raise self.no_node_at_path(path)
 
             # cache node
             self.nodes[path] = node
 
         return self.nodes[path]
 
     def get_largefile_node(self, path):
         tree_id, _ = self._get_tree_id_for_path(path)
         pointer_spec = self._remote.is_large_file(tree_id)
 
         if pointer_spec:
             # content of that file regular FileNode is the hash of largefile
             file_id = pointer_spec.get('oid_hash')
             if self._remote.in_largefiles_store(file_id):
                 lf_path = self._remote.store_path(file_id)
                 return LargeFileNode(lf_path, commit=self, org_path=path)
 
     @LazyProperty
     def affected_files(self):
         """
         Gets a fast accessible file changes for given commit
         """
         added, modified, deleted = self._changes_cache
         return list(added.union(modified).union(deleted))
 
     @LazyProperty
     def _changes_cache(self):
         added = set()
         modified = set()
         deleted = set()
         _r = self._remote
 
         parents = self.parents
         if not self.parents:
             parents = [base.EmptyCommit()]
         for parent in parents:
             if isinstance(parent, base.EmptyCommit):
                 oid = None
             else:
                 oid = parent.raw_id
             changes = _r.tree_changes(oid, self.raw_id)
             for (oldpath, newpath), (_, _), (_, _) in changes:
                 if newpath and oldpath:
                     modified.add(newpath)
                 elif newpath and not oldpath:
                     added.add(newpath)
                 elif not newpath and oldpath:
                     deleted.add(oldpath)
         return added, modified, deleted
 
     def _get_paths_for_status(self, status):
         """
         Returns sorted list of paths for given ``status``.
 
         :param status: one of: *added*, *modified* or *deleted*
         """
         added, modified, deleted = self._changes_cache
         return sorted({
             'added': list(added),
             'modified': list(modified),
             'deleted': list(deleted)}[status]
         )
 
     @LazyProperty
     def added(self):
         """
         Returns list of added ``FileNode`` objects.
         """
         if not self.parents:
             return list(self._get_file_nodes())
-        return AddedFileNodesGenerator(
-            [n for n in self._get_paths_for_status('added')], self)
+        return AddedFileNodesGenerator(self.added_paths, self)
+
+    @LazyProperty
+    def added_paths(self):
+        return [n for n in self._get_paths_for_status('added')]
 
     @LazyProperty
     def changed(self):
         """
         Returns list of modified ``FileNode`` objects.
         """
         if not self.parents:
             return []
-        return ChangedFileNodesGenerator(
-            [n for n in self._get_paths_for_status('modified')], self)
+        return ChangedFileNodesGenerator(self.changed_paths, self)
+
+    @LazyProperty
+    def changed_paths(self):
+        return [n for n in self._get_paths_for_status('modified')]
 
     @LazyProperty
     def removed(self):
         """
         Returns list of removed ``FileNode`` objects.
         """
         if not self.parents:
             return []
-        return RemovedFileNodesGenerator(
-            [n for n in self._get_paths_for_status('deleted')], self)
+        return RemovedFileNodesGenerator(self.removed_paths, self)
+
+    @LazyProperty
+    def removed_paths(self):
+        return [n for n in self._get_paths_for_status('deleted')]
 
     def _get_submodule_url(self, submodule_path):
         git_modules_path = '.gitmodules'
 
         if self._submodules is None:
             self._submodules = {}
 
             try:
                 submodules_node = self.get_node(git_modules_path)
             except NodeDoesNotExistError:
                 return None
 
             # ConfigParser fails if there are whitespaces, also it needs an iterable
             # file like content
             def iter_content(_content):
                 for line in _content.splitlines():
                     yield line
 
             parser = configparser.RawConfigParser()
             parser.read_file(iter_content(submodules_node.content))
 
             for section in parser.sections():
                 path = parser.get(section, 'path')
                 url = parser.get(section, 'url')
                 if path and url:
                     self._submodules[path.strip('/')] = url
 
         return self._submodules.get(submodule_path.strip('/'))

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

@@ -1,389 +1,401 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2014-2019 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 commit module
 """
 
 import os
 
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 from rhodecode.lib.datelib import utcdate_fromtimestamp
 from rhodecode.lib.utils import safe_str, safe_unicode
 from rhodecode.lib.vcs import path as vcspath
 from rhodecode.lib.vcs.backends import base
 from rhodecode.lib.vcs.backends.hg.diff import MercurialDiff
 from rhodecode.lib.vcs.exceptions import CommitError
 from rhodecode.lib.vcs.nodes import (
     AddedFileNodesGenerator, ChangedFileNodesGenerator, DirNode, FileNode,
     NodeKind, RemovedFileNodesGenerator, RootNode, SubModuleNode,
     LargeFileNode, LARGEFILE_PREFIX)
 from rhodecode.lib.vcs.utils.paths import get_dirs_for_path
 
 
 class MercurialCommit(base.BaseCommit):
     """
     Represents state of the repository at the single commit.
     """
 
     _filter_pre_load = [
         # git specific property not supported here
         "_commit",
     ]
 
     def __init__(self, repository, raw_id, idx, pre_load=None):
         raw_id = safe_str(raw_id)
 
         self.repository = repository
         self._remote = repository._remote
 
         self.raw_id = raw_id
         self.idx = idx
 
         self._set_bulk_properties(pre_load)
 
         # caches
         self.nodes = {}
 
     def _set_bulk_properties(self, pre_load):
         if not pre_load:
             return
         pre_load = [entry for entry in pre_load
                     if entry not in self._filter_pre_load]
         if not pre_load:
             return
 
         result = self._remote.bulk_request(self.raw_id, pre_load)
         for attr, value in result.items():
             if attr in ["author", "branch", "message"]:
                 value = safe_unicode(value)
             elif attr == "affected_files":
                 value = map(safe_unicode, value)
             elif attr == "date":
                 value = utcdate_fromtimestamp(*value)
             elif attr in ["children", "parents"]:
                 value = self._make_commits(value)
             elif attr in ["phase"]:
                 value = self._get_phase_text(value)
             self.__dict__[attr] = value
 
     @LazyProperty
     def tags(self):
         tags = [name for name, commit_id in self.repository.tags.iteritems()
                 if commit_id == self.raw_id]
         return tags
 
     @LazyProperty
     def branch(self):
         return safe_unicode(self._remote.ctx_branch(self.raw_id))
 
     @LazyProperty
     def bookmarks(self):
         bookmarks = [
             name for name, commit_id in self.repository.bookmarks.iteritems()
             if commit_id == self.raw_id]
         return bookmarks
 
     @LazyProperty
     def message(self):
         return safe_unicode(self._remote.ctx_description(self.raw_id))
 
     @LazyProperty
     def committer(self):
         return safe_unicode(self.author)
 
     @LazyProperty
     def author(self):
         return safe_unicode(self._remote.ctx_user(self.raw_id))
 
     @LazyProperty
     def date(self):
         return utcdate_fromtimestamp(*self._remote.ctx_date(self.raw_id))
 
     @LazyProperty
     def status(self):
         """
         Returns modified, added, removed, deleted files for current commit
         """
         return self._remote.ctx_status(self.raw_id)
 
     @LazyProperty
     def _file_paths(self):
         return self._remote.ctx_list(self.raw_id)
 
     @LazyProperty
     def _dir_paths(self):
         p = list(set(get_dirs_for_path(*self._file_paths)))
         p.insert(0, '')
         return p
 
     @LazyProperty
     def _paths(self):
         return self._dir_paths + self._file_paths
 
     @LazyProperty
     def id(self):
         if self.last:
             return u'tip'
         return self.short_id
 
     @LazyProperty
     def short_id(self):
         return self.raw_id[:12]
 
     def _make_commits(self, commit_ids, pre_load=None):
         return [self.repository.get_commit(commit_id=commit_id, pre_load=pre_load)
                 for commit_id in commit_ids]
 
     @LazyProperty
     def parents(self):
         """
         Returns list of parent commits.
         """
         parents = self._remote.ctx_parents(self.raw_id)
         return self._make_commits(parents)
 
     def _get_phase_text(self, phase_id):
         return {
                    0: 'public',
                    1: 'draft',
                    2: 'secret',
                }.get(phase_id) or ''
 
     @LazyProperty
     def phase(self):
         phase_id = self._remote.ctx_phase(self.raw_id)
         phase_text = self._get_phase_text(phase_id)
 
         return safe_unicode(phase_text)
 
     @LazyProperty
     def obsolete(self):
         obsolete = self._remote.ctx_obsolete(self.raw_id)
         return obsolete
 
     @LazyProperty
     def hidden(self):
         hidden = self._remote.ctx_hidden(self.raw_id)
         return hidden
 
     @LazyProperty
     def children(self):
         """
         Returns list of child commits.
         """
         children = self._remote.ctx_children(self.raw_id)
         return self._make_commits(children)
 
     def _fix_path(self, path):
         """
         Mercurial keeps filenodes as str so we need to encode from unicode
         to str.
         """
         return safe_str(super(MercurialCommit, self)._fix_path(path))
 
     def _get_kind(self, path):
         path = self._fix_path(path)
         if path in self._file_paths:
             return NodeKind.FILE
         elif path in self._dir_paths:
             return NodeKind.DIR
         else:
             raise CommitError(
                 "Node does not exist at the given path '%s'" % (path, ))
 
     def _get_filectx(self, path):
         path = self._fix_path(path)
         if self._get_kind(path) != NodeKind.FILE:
             raise CommitError(
                 "File does not exist for idx %s at '%s'" % (self.raw_id, path))
         return path
 
     def get_file_mode(self, path):
         """
         Returns stat mode of the file at the given ``path``.
         """
         path = self._get_filectx(path)
         if 'x' in self._remote.fctx_flags(self.raw_id, path):
             return base.FILEMODE_EXECUTABLE
         else:
             return base.FILEMODE_DEFAULT
 
     def is_link(self, path):
         path = self._get_filectx(path)
         return 'l' in self._remote.fctx_flags(self.raw_id, path)
 
     def is_node_binary(self, path):
         path = self._get_filectx(path)
         return self._remote.is_binary(self.raw_id, path)
 
     def get_file_content(self, path):
         """
         Returns content of the file at given ``path``.
         """
         path = self._get_filectx(path)
         return self._remote.fctx_node_data(self.raw_id, path)
 
     def get_file_content_streamed(self, path):
         path = self._get_filectx(path)
         stream_method = getattr(self._remote, 'stream:fctx_node_data')
         return stream_method(self.raw_id, path)
 
     def get_file_size(self, path):
         """
         Returns size of the file at given ``path``.
         """
         path = self._get_filectx(path)
         return self._remote.fctx_size(self.raw_id, path)
 
     def get_path_history(self, path, limit=None, pre_load=None):
         """
         Returns history of file as reversed list of `MercurialCommit` objects
         for which file at given ``path`` has been modified.
         """
         path = self._get_filectx(path)
         hist = self._remote.node_history(self.raw_id, path, limit)
         return [
             self.repository.get_commit(commit_id=commit_id, pre_load=pre_load)
             for commit_id in hist]
 
     def get_file_annotate(self, path, pre_load=None):
         """
         Returns a generator of four element tuples with
             lineno, commit_id, commit lazy loader and line
         """
         result = self._remote.fctx_annotate(self.raw_id, path)
 
         for ln_no, commit_id, content in result:
             yield (
                 ln_no, commit_id,
                 lambda: self.repository.get_commit(commit_id=commit_id, pre_load=pre_load),
                 content)
 
     def get_nodes(self, path):
         """
         Returns combined ``DirNode`` and ``FileNode`` objects list representing
         state of commit at the given ``path``. If node at the given ``path``
         is not instance of ``DirNode``, CommitError would be raised.
         """
 
         if self._get_kind(path) != NodeKind.DIR:
             raise CommitError(
                 "Directory does not exist for idx %s at '%s'" % (self.raw_id, path))
         path = self._fix_path(path)
 
         filenodes = [
             FileNode(f, commit=self) for f in self._file_paths
             if os.path.dirname(f) == path]
         # TODO: johbo: Check if this can be done in a more obvious way
         dirs = path == '' and '' or [
             d for d in self._dir_paths
             if d and vcspath.dirname(d) == path]
         dirnodes = [
             DirNode(d, commit=self) for d in dirs
             if os.path.dirname(d) == path]
 
         alias = self.repository.alias
         for k, vals in self._submodules.iteritems():
             if vcspath.dirname(k) == path:
                 loc = vals[0]
                 commit = vals[1]
                 dirnodes.append(SubModuleNode(k, url=loc, commit=commit, alias=alias))
 
         nodes = dirnodes + filenodes
         for node in nodes:
             if node.path not in self.nodes:
                 self.nodes[node.path] = node
         nodes.sort()
 
         return nodes
 
     def get_node(self, path, pre_load=None):
         """
         Returns `Node` object from the given `path`. If there is no node at
         the given `path`, `NodeDoesNotExistError` would be raised.
         """
         path = self._fix_path(path)
 
         if path not in self.nodes:
             if path in self._file_paths:
                 node = FileNode(path, commit=self, pre_load=pre_load)
             elif path in self._dir_paths:
                 if path == '':
                     node = RootNode(commit=self)
                 else:
                     node = DirNode(path, commit=self)
             else:
                 raise self.no_node_at_path(path)
 
             # cache node
             self.nodes[path] = node
         return self.nodes[path]
 
     def get_largefile_node(self, path):
         pointer_spec = self._remote.is_large_file(self.raw_id, path)
         if pointer_spec:
             # content of that file regular FileNode is the hash of largefile
             file_id = self.get_file_content(path).strip()
 
             if self._remote.in_largefiles_store(file_id):
                 lf_path = self._remote.store_path(file_id)
                 return LargeFileNode(lf_path, commit=self, org_path=path)
             elif self._remote.in_user_cache(file_id):
                 lf_path = self._remote.store_path(file_id)
                 self._remote.link(file_id, path)
                 return LargeFileNode(lf_path, commit=self, org_path=path)
 
     @LazyProperty
     def _submodules(self):
         """
         Returns a dictionary with submodule information from substate file
         of hg repository.
         """
         return self._remote.ctx_substate(self.raw_id)
 
     @LazyProperty
     def affected_files(self):
         """
         Gets a fast accessible file changes for given commit
         """
         return self._remote.ctx_files(self.raw_id)
 
     @property
     def added(self):
         """
         Returns list of added ``FileNode`` objects.
         """
-        return AddedFileNodesGenerator([n for n in self.status[1]], self)
+        return AddedFileNodesGenerator(self.added_paths, self)
+
+    @LazyProperty
+    def added_paths(self):
+        return [n for n in self.status[1]]
 
     @property
     def changed(self):
         """
         Returns list of modified ``FileNode`` objects.
         """
-        return ChangedFileNodesGenerator([n for n in self.status[0]], self)
+        return ChangedFileNodesGenerator(self.changed_paths, self)
+
+    @LazyProperty
+    def changed_paths(self):
+        return [n for n in self.status[0]]
 
     @property
     def removed(self):
         """
         Returns list of removed ``FileNode`` objects.
         """
-        return RemovedFileNodesGenerator([n for n in self.status[2]], self)
+        return RemovedFileNodesGenerator(self.removed_paths, self)
+
+    @LazyProperty
+    def removed_paths(self):
+        return [n for n in self.status[2]]

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

@@ -1,245 +1,254 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2014-2019 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 """
 SVN commit module
 """
 
 
 import dateutil.parser
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 from rhodecode.lib.utils import safe_str, safe_unicode
 from rhodecode.lib.vcs import nodes, path as vcspath
 from rhodecode.lib.vcs.backends import base
 from rhodecode.lib.vcs.exceptions import CommitError, NodeDoesNotExistError
 
 
 _SVN_PROP_TRUE = '*'
 
 
 class SubversionCommit(base.BaseCommit):
     """
     Subversion specific implementation of commits
 
     .. attribute:: branch
 
        The Subversion backend does not support to assign branches to
        specific commits. This attribute has always the value `None`.
 
     """
 
     def __init__(self, repository, commit_id):
         self.repository = repository
         self.idx = self.repository._get_commit_idx(commit_id)
         self._svn_rev = self.idx + 1
         self._remote = repository._remote
         # TODO: handling of raw_id should be a method on repository itself,
         # which knows how to translate commit index and commit id
         self.raw_id = commit_id
         self.short_id = commit_id
         self.id = 'r%s' % (commit_id, )
 
         # TODO: Implement the following placeholder attributes
         self.nodes = {}
         self.tags = []
 
     @property
     def author(self):
         return safe_unicode(self._properties.get('svn:author'))
 
     @property
     def date(self):
         return _date_from_svn_properties(self._properties)
 
     @property
     def message(self):
         return safe_unicode(self._properties.get('svn:log'))
 
     @LazyProperty
     def _properties(self):
         return self._remote.revision_properties(self._svn_rev)
 
     @LazyProperty
     def parents(self):
         parent_idx = self.idx - 1
         if parent_idx >= 0:
             parent = self.repository.get_commit(commit_idx=parent_idx)
             return [parent]
         return []
 
     @LazyProperty
     def children(self):
         child_idx = self.idx + 1
         if child_idx < len(self.repository.commit_ids):
             child = self.repository.get_commit(commit_idx=child_idx)
             return [child]
         return []
 
     def get_file_mode(self, path):
         # Note: Subversion flags files which are executable with a special
         # property `svn:executable` which is set to the value ``"*"``.
         if self._get_file_property(path, 'svn:executable') == _SVN_PROP_TRUE:
             return base.FILEMODE_EXECUTABLE
         else:
             return base.FILEMODE_DEFAULT
 
     def is_link(self, path):
         # Note: Subversion has a flag for special files, the content of the
         # file contains the type of that file.
         if self._get_file_property(path, 'svn:special') == _SVN_PROP_TRUE:
             return self.get_file_content(path).startswith('link')
         return False
 
     def is_node_binary(self, path):
         path = self._fix_path(path)
         return self._remote.is_binary(self._svn_rev, safe_str(path))
 
     def _get_file_property(self, path, name):
         file_properties = self._remote.node_properties(
             safe_str(path), self._svn_rev)
         return file_properties.get(name)
 
     def get_file_content(self, path):
         path = self._fix_path(path)
         return self._remote.get_file_content(safe_str(path), self._svn_rev)
 
     def get_file_content_streamed(self, path):
         path = self._fix_path(path)
         stream_method = getattr(self._remote, 'stream:get_file_content')
         return stream_method(safe_str(path), self._svn_rev)
 
     def get_file_size(self, path):
         path = self._fix_path(path)
         return self._remote.get_file_size(safe_str(path), self._svn_rev)
 
     def get_path_history(self, path, limit=None, pre_load=None):
         path = safe_str(self._fix_path(path))
         history = self._remote.node_history(path, self._svn_rev, limit)
         return [
             self.repository.get_commit(commit_id=str(svn_rev))
             for svn_rev in history]
 
     def get_file_annotate(self, path, pre_load=None):
         result = self._remote.file_annotate(safe_str(path), self._svn_rev)
 
         for zero_based_line_no, svn_rev, content in result:
             commit_id = str(svn_rev)
             line_no = zero_based_line_no + 1
             yield (
                 line_no,
                 commit_id,
                 lambda: self.repository.get_commit(commit_id=commit_id),
                 content)
 
     def get_node(self, path, pre_load=None):
         path = self._fix_path(path)
         if path not in self.nodes:
 
             if path == '':
                 node = nodes.RootNode(commit=self)
             else:
                 node_type = self._remote.get_node_type(
                     safe_str(path), self._svn_rev)
                 if node_type == 'dir':
                     node = nodes.DirNode(path, commit=self)
                 elif node_type == 'file':
                     node = nodes.FileNode(path, commit=self, pre_load=pre_load)
                 else:
                     raise self.no_node_at_path(path)
 
             self.nodes[path] = node
         return self.nodes[path]
 
     def get_nodes(self, path):
         if self._get_kind(path) != nodes.NodeKind.DIR:
             raise CommitError(
                 "Directory does not exist for commit %s at "
                 " '%s'" % (self.raw_id, path))
         path = self._fix_path(path)
 
         path_nodes = []
         for name, kind in self._remote.get_nodes(
                 safe_str(path), revision=self._svn_rev):
             node_path = vcspath.join(path, name)
             if kind == 'dir':
                 node = nodes.DirNode(node_path, commit=self)
             elif kind == 'file':
                 node = nodes.FileNode(node_path, commit=self)
             else:
                 raise ValueError("Node kind %s not supported." % (kind, ))
             self.nodes[node_path] = node
             path_nodes.append(node)
 
         return path_nodes
 
     def _get_kind(self, path):
         path = self._fix_path(path)
         kind = self._remote.get_node_type(path, self._svn_rev)
         if kind == 'file':
             return nodes.NodeKind.FILE
         elif kind == 'dir':
             return nodes.NodeKind.DIR
         else:
             raise CommitError(
                 "Node does not exist at the given path '%s'" % (path, ))
 
     @LazyProperty
     def _changes_cache(self):
         return self._remote.revision_changes(self._svn_rev)
 
     @LazyProperty
     def affected_files(self):
         changed_files = set()
         for files in self._changes_cache.itervalues():
             changed_files.update(files)
         return list(changed_files)
 
     @LazyProperty
     def id(self):
         return self.raw_id
 
     @property
     def added(self):
-        return nodes.AddedFileNodesGenerator(
-            self._changes_cache['added'], self)
+        return nodes.AddedFileNodesGenerator(self.added_paths, self)
+
+    @LazyProperty
+    def added_paths(self):
+        return [n for n in self._changes_cache['added']]
 
     @property
     def changed(self):
-        return nodes.ChangedFileNodesGenerator(
-            self._changes_cache['changed'], self)
+        return nodes.ChangedFileNodesGenerator(self.changed_paths, self)
+
+    @LazyProperty
+    def changed_paths(self):
+        return [n for n in self._changes_cache['changed']]
 
     @property
     def removed(self):
-        return nodes.RemovedFileNodesGenerator(
-            self._changes_cache['removed'], self)
+        return nodes.RemovedFileNodesGenerator(self.removed_paths, self)
+
+    @LazyProperty
+    def removed_paths(self):
+        return [n for n in self._changes_cache['removed']]
 
 
 def _date_from_svn_properties(properties):
     """
     Parses the date out of given svn properties.
 
     :return: :class:`datetime.datetime` instance. The object is naive.
     """
 
     aware_date = dateutil.parser.parse(properties.get('svn:date'))
     # final_date = aware_date.astimezone(dateutil.tz.tzlocal())
     final_date = aware_date
     return final_date.replace(tzinfo=None)