##// END OF EJS Templates
processes: better handling of PID that are not part of RhodeCode processes....
processes: better handling of PID that are not part of RhodeCode processes. - in some cases for multiple gunicorn instances the PID sent is not from rhodecode which would lead to AccessDenied errors. We prevent from crashing the server on this type of errors.

File last commit:

r2487:fcee5614 default
r2661:042cb4c7 default
Show More
changeset_status.py
393 lines | 14.3 KiB | text/x-python | PythonLexer
# -*- coding: utf-8 -*-
# Copyright (C) 2010-2018 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 itertools
import logging
import collections
from rhodecode.model import BaseModel
from rhodecode.model.db import (
ChangesetStatus, ChangesetComment, PullRequest, Session)
from rhodecode.lib.exceptions import StatusChangeOnClosedPullRequestError
from rhodecode.lib.markup_renderer import (
DEFAULT_COMMENTS_RENDERER, RstTemplateRenderer)
log = logging.getLogger(__name__)
class ChangesetStatusModel(BaseModel):
cls = ChangesetStatus
def __get_changeset_status(self, changeset_status):
return self._get_instance(ChangesetStatus, changeset_status)
def __get_pull_request(self, pull_request):
return self._get_instance(PullRequest, pull_request)
def _get_status_query(self, repo, revision, pull_request,
with_revisions=False):
repo = self._get_repo(repo)
q = ChangesetStatus.query()\
.filter(ChangesetStatus.repo == repo)
if not with_revisions:
q = q.filter(ChangesetStatus.version == 0)
if revision:
q = q.filter(ChangesetStatus.revision == revision)
elif pull_request:
pull_request = self.__get_pull_request(pull_request)
# TODO: johbo: Think about the impact of this join, there must
# be a reason why ChangesetStatus and ChanagesetComment is linked
# to the pull request. Might be that we want to do the same for
# the pull_request_version_id.
q = q.join(ChangesetComment).filter(
ChangesetStatus.pull_request == pull_request,
ChangesetComment.pull_request_version_id == None)
else:
raise Exception('Please specify revision or pull_request')
q = q.order_by(ChangesetStatus.version.asc())
return q
def calculate_group_vote(self, group_id, group_statuses_by_reviewers,
trim_votes=True):
"""
Calculate status based on given group members, and voting rule
group1 - 4 members, 3 required for approval
user1 - approved
user2 - reject
user3 - approved
user4 - rejected
final_state: rejected, reasons not at least 3 votes
group1 - 4 members, 2 required for approval
user1 - approved
user2 - reject
user3 - approved
user4 - rejected
final_state: approved, reasons got at least 2 approvals
group1 - 4 members, ALL required for approval
user1 - approved
user2 - reject
user3 - approved
user4 - rejected
final_state: rejected, reasons not all approvals
group1 - 4 members, ALL required for approval
user1 - approved
user2 - approved
user3 - approved
user4 - approved
final_state: approved, reason all approvals received
group1 - 4 members, 5 required for approval
(approval should be shorted to number of actual members)
user1 - approved
user2 - approved
user3 - approved
user4 - approved
final_state: approved, reason all approvals received
"""
group_vote_data = {}
got_rule = False
members = collections.OrderedDict()
for review_obj, user, reasons, mandatory, statuses \
in group_statuses_by_reviewers:
if not got_rule:
group_vote_data = review_obj.rule_user_group_data()
got_rule = bool(group_vote_data)
members[user.user_id] = statuses
if not group_vote_data:
return []
required_votes = group_vote_data['vote_rule']
if required_votes == -1:
# -1 means all required, so we replace it with how many people
# are in the members
required_votes = len(members)
if trim_votes and required_votes > len(members):
# we require more votes than we have members in the group
# in this case we trim the required votes to the number of members
required_votes = len(members)
approvals = sum([
1 for statuses in members.values()
if statuses and
statuses[0][1].status == ChangesetStatus.STATUS_APPROVED])
calculated_votes = []
# we have all votes from users, now check if we have enough votes
# to fill other
fill_in = ChangesetStatus.STATUS_UNDER_REVIEW
if approvals >= required_votes:
fill_in = ChangesetStatus.STATUS_APPROVED
for member, statuses in members.items():
if statuses:
ver, latest = statuses[0]
if fill_in == ChangesetStatus.STATUS_APPROVED:
calculated_votes.append(fill_in)
else:
calculated_votes.append(latest.status)
else:
calculated_votes.append(fill_in)
return calculated_votes
def calculate_status(self, statuses_by_reviewers):
"""
Given the approval statuses from reviewers, calculates final approval
status. There can only be 3 results, all approved, all rejected. If
there is no consensus the PR is under review.
:param statuses_by_reviewers:
"""
def group_rule(element):
review_obj = element[0]
rule_data = review_obj.rule_user_group_data()
if rule_data and rule_data['id']:
return rule_data['id']
voting_groups = itertools.groupby(
sorted(statuses_by_reviewers, key=group_rule), group_rule)
voting_by_groups = [(x, list(y)) for x, y in voting_groups]
reviewers_number = len(statuses_by_reviewers)
votes = collections.defaultdict(int)
for group, group_statuses_by_reviewers in voting_by_groups:
if group:
# calculate how the "group" voted
for vote_status in self.calculate_group_vote(
group, group_statuses_by_reviewers):
votes[vote_status] += 1
else:
for review_obj, user, reasons, mandatory, statuses \
in group_statuses_by_reviewers:
# individual vote
if statuses:
ver, latest = statuses[0]
votes[latest.status] += 1
approved_votes_count = votes[ChangesetStatus.STATUS_APPROVED]
rejected_votes_count = votes[ChangesetStatus.STATUS_REJECTED]
# TODO(marcink): with group voting, how does rejected work,
# do we ever get rejected state ?
if approved_votes_count == reviewers_number:
return ChangesetStatus.STATUS_APPROVED
if rejected_votes_count == reviewers_number:
return ChangesetStatus.STATUS_REJECTED
return ChangesetStatus.STATUS_UNDER_REVIEW
def get_statuses(self, repo, revision=None, pull_request=None,
with_revisions=False):
q = self._get_status_query(repo, revision, pull_request,
with_revisions)
return q.all()
def get_status(self, repo, revision=None, pull_request=None, as_str=True):
"""
Returns latest status of changeset for given revision or for given
pull request. Statuses are versioned inside a table itself and
version == 0 is always the current one
:param repo:
:param revision: 40char hash or None
:param pull_request: pull_request reference
:param as_str: return status as string not object
"""
q = self._get_status_query(repo, revision, pull_request)
# need to use first here since there can be multiple statuses
# returned from pull_request
status = q.first()
if as_str:
status = status.status if status else status
st = status or ChangesetStatus.DEFAULT
return str(st)
return status
def _render_auto_status_message(
self, status, commit_id=None, pull_request=None):
"""
render the message using DEFAULT_COMMENTS_RENDERER (RST renderer),
so it's always looking the same disregarding on which default
renderer system is using.
:param status: status text to change into
:param commit_id: the commit_id we change the status for
:param pull_request: the pull request we change the status for
"""
new_status = ChangesetStatus.get_status_lbl(status)
params = {
'new_status_label': new_status,
'pull_request': pull_request,
'commit_id': commit_id,
}
renderer = RstTemplateRenderer()
return renderer.render('auto_status_change.mako', **params)
def set_status(self, repo, status, user, comment=None, revision=None,
pull_request=None, dont_allow_on_closed_pull_request=False):
"""
Creates new status for changeset or updates the old ones bumping their
version, leaving the current status at
:param repo:
:param revision:
:param status:
:param user:
:param comment:
:param dont_allow_on_closed_pull_request: don't allow a status change
if last status was for pull request and it's closed. We shouldn't
mess around this manually
"""
repo = self._get_repo(repo)
q = ChangesetStatus.query()
if revision:
q = q.filter(ChangesetStatus.repo == repo)
q = q.filter(ChangesetStatus.revision == revision)
elif pull_request:
pull_request = self.__get_pull_request(pull_request)
q = q.filter(ChangesetStatus.repo == pull_request.source_repo)
q = q.filter(ChangesetStatus.revision.in_(pull_request.revisions))
cur_statuses = q.all()
# if statuses exists and last is associated with a closed pull request
# we need to check if we can allow this status change
if (dont_allow_on_closed_pull_request and cur_statuses
and getattr(cur_statuses[0].pull_request, 'status', '')
== PullRequest.STATUS_CLOSED):
raise StatusChangeOnClosedPullRequestError(
'Changing status on closed pull request is not allowed'
)
# update all current statuses with older version
if cur_statuses:
for st in cur_statuses:
st.version += 1
Session().add(st)
def _create_status(user, repo, status, comment, revision, pull_request):
new_status = ChangesetStatus()
new_status.author = self._get_user(user)
new_status.repo = self._get_repo(repo)
new_status.status = status
new_status.comment = comment
new_status.revision = revision
new_status.pull_request = pull_request
return new_status
if not comment:
from rhodecode.model.comment import CommentsModel
comment = CommentsModel().create(
text=self._render_auto_status_message(
status, commit_id=revision, pull_request=pull_request),
repo=repo,
user=user,
pull_request=pull_request,
send_email=False, renderer=DEFAULT_COMMENTS_RENDERER
)
if revision:
new_status = _create_status(
user=user, repo=repo, status=status, comment=comment,
revision=revision, pull_request=pull_request)
Session().add(new_status)
return new_status
elif pull_request:
# pull request can have more than one revision associated to it
# we need to create new version for each one
new_statuses = []
repo = pull_request.source_repo
for rev in pull_request.revisions:
new_status = _create_status(
user=user, repo=repo, status=status, comment=comment,
revision=rev, pull_request=pull_request)
new_statuses.append(new_status)
Session().add(new_status)
return new_statuses
def reviewers_statuses(self, pull_request):
_commit_statuses = self.get_statuses(
pull_request.source_repo,
pull_request=pull_request,
with_revisions=True)
commit_statuses = collections.defaultdict(list)
for st in _commit_statuses:
commit_statuses[st.author.username] += [st]
pull_request_reviewers = []
def version(commit_status):
return commit_status.version
for obj in pull_request.reviewers:
if not obj.user:
continue
statuses = commit_statuses.get(obj.user.username, None)
if statuses:
status_groups = itertools.groupby(
sorted(statuses, key=version), version)
statuses = [(x, list(y)[0]) for x, y in status_groups]
pull_request_reviewers.append(
(obj, obj.user, obj.reasons, obj.mandatory, statuses))
return pull_request_reviewers
def calculated_review_status(self, pull_request, reviewers_statuses=None):
"""
calculate pull request status based on reviewers, it should be a list
of two element lists.
:param reviewers_statuses:
"""
reviewers = reviewers_statuses or self.reviewers_statuses(pull_request)
return self.calculate_status(reviewers)