Changeset - f713a37564c0
[Not reviewed]
kallithea/controllers/compare.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
# This program is free software: you can redistribute it and/or modify
 
# it under the terms of the GNU General Public License as published by
 
# the Free Software Foundation, either version 3 of the License, or
 
# (at your option) any later version.
 
#
 
# 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 General Public License
 
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
"""
 
kallithea.controllers.compare
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
compare controller showing differences between two
 
repos, branches, bookmarks or tips
 

	
 
This file was forked by the Kallithea project in July 2014.
 
Original author and date, and relevant copyright and licensing information is below:
 
:created_on: May 6, 2012
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 

	
 
import logging
 
import re
 

	
 
import mercurial.unionrepo
 
from tg import request
 
from tg import tmpl_context as c
 
from tg.i18n import ugettext as _
 
from webob.exc import HTTPBadRequest, HTTPFound, HTTPNotFound
 

	
 
from kallithea.config.routing import url
 
from kallithea.controllers.changeset import _context_url, _ignorews_url
 
from kallithea.lib import diffs
 
from kallithea.lib import helpers as h
 
from kallithea.lib.auth import HasRepoPermissionLevelDecorator, LoginRequired
 
from kallithea.lib.base import BaseRepoController, render
 
from kallithea.lib.graphmod import graph_data
 
from kallithea.lib.utils2 import ascii_bytes, ascii_str, safe_int, safe_str
 
from kallithea.lib.vcs.utils.hgcompat import unionrepo
 
from kallithea.model.db import Repository
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
class CompareController(BaseRepoController):
 

	
 
    def _before(self, *args, **kwargs):
 
        super(CompareController, self)._before(*args, **kwargs)
 

	
 
        # The base repository has already been retrieved.
 
        c.a_repo = c.db_repo
 

	
 
        # Retrieve the "changeset" repository (default: same as base).
 
        other_repo = request.GET.get('other_repo', None)
 
        if other_repo is None:
 
            c.cs_repo = c.a_repo
 
        else:
 
            c.cs_repo = Repository.get_by_repo_name(other_repo)
 
            if c.cs_repo is None:
 
                msg = _('Could not find other repository %s') % other_repo
 
                h.flash(msg, category='error')
 
                raise HTTPFound(location=url('compare_home', repo_name=c.a_repo.repo_name))
 

	
 
        # Verify that it's even possible to compare these two repositories.
 
        if c.a_repo.scm_instance.alias != c.cs_repo.scm_instance.alias:
 
            msg = _('Cannot compare repositories of different types')
 
            h.flash(msg, category='error')
 
            raise HTTPFound(location=url('compare_home', repo_name=c.a_repo.repo_name))
 

	
 
    @staticmethod
 
    def _get_changesets(alias, org_repo, org_rev, other_repo, other_rev):
 
        """
 
        Returns lists of changesets that can be merged from org_repo@org_rev
 
        to other_repo@other_rev
 
        ... and the other way
 
        ... and the ancestors that would be used for merge
 

	
 
        :param org_repo: repo object, that is most likely the original repo we forked from
 
        :param org_rev: the revision we want our compare to be made
 
        :param other_repo: repo object, most likely the fork of org_repo. It has
 
            all changesets that we need to obtain
 
        :param other_rev: revision we want out compare to be made on other_repo
 
        """
 
        ancestors = None
 
        if org_rev == other_rev:
 
            org_changesets = []
 
            other_changesets = []
 

	
 
        elif alias == 'hg':
 
            # case two independent repos
 
            if org_repo != other_repo:
 
                hgrepo = unionrepo.makeunionrepository(other_repo.baseui,
 
                hgrepo = mercurial.unionrepo.makeunionrepository(other_repo.baseui,
 
                                                       other_repo.path,
 
                                                       org_repo.path)
 
                # all ancestors of other_rev will be in other_repo and
 
                # rev numbers from hgrepo can be used in other_repo - org_rev ancestors cannot
 

	
 
            # no remote compare do it on the same repository
 
            else:
 
                hgrepo = other_repo._repo
 

	
 
            ancestors = [ascii_str(hgrepo[ancestor].hex()) for ancestor in
 
                         hgrepo.revs(b"id(%s) & ::id(%s)", ascii_bytes(other_rev), ascii_bytes(org_rev))]
 
            if ancestors:
 
                log.debug("shortcut found: %s is already an ancestor of %s", other_rev, org_rev)
 
            else:
 
                log.debug("no shortcut found: %s is not an ancestor of %s", other_rev, org_rev)
 
                ancestors = [ascii_str(hgrepo[ancestor].hex()) for ancestor in
 
                             hgrepo.revs(b"heads(::id(%s) & ::id(%s))", ascii_bytes(org_rev), ascii_bytes(other_rev))] # FIXME: expensive!
 

	
 
            other_changesets = [
 
                other_repo.get_changeset(rev)
 
                for rev in hgrepo.revs(
 
                    b"ancestors(id(%s)) and not ancestors(id(%s)) and not id(%s)",
 
                    ascii_bytes(other_rev), ascii_bytes(org_rev), ascii_bytes(org_rev))
 
            ]
 
            org_changesets = [
 
                org_repo.get_changeset(ascii_str(hgrepo[rev].hex()))
 
                for rev in hgrepo.revs(
 
                    b"ancestors(id(%s)) and not ancestors(id(%s)) and not id(%s)",
 
                    ascii_bytes(org_rev), ascii_bytes(other_rev), ascii_bytes(other_rev))
 
            ]
 

	
 
        elif alias == 'git':
 
            if org_repo != other_repo:
 
                from dulwich.repo import Repo
 
                from dulwich.client import SubprocessGitClient
 

	
 
                gitrepo = Repo(org_repo.path)
 
                SubprocessGitClient(thin_packs=False).fetch(safe_str(other_repo.path), gitrepo)
 

	
 
                gitrepo_remote = Repo(other_repo.path)
 
                SubprocessGitClient(thin_packs=False).fetch(safe_str(org_repo.path), gitrepo_remote)
 

	
 
                revs = [
 
                    ascii_str(x.commit.id)
 
                    for x in gitrepo_remote.get_walker(include=[ascii_bytes(other_rev)],
 
                                                       exclude=[ascii_bytes(org_rev)])
 
                ]
 
                other_changesets = [other_repo.get_changeset(rev) for rev in reversed(revs)]
 
                if other_changesets:
 
                    ancestors = [other_changesets[0].parents[0].raw_id]
 
                else:
 
                    # no changesets from other repo, ancestor is the other_rev
 
                    ancestors = [other_rev]
 

	
 
                gitrepo.close()
 
                gitrepo_remote.close()
 

	
 
            else:
 
                so = org_repo.run_git_command(
 
                    ['log', '--reverse', '--pretty=format:%H',
 
                     '-s', '%s..%s' % (org_rev, other_rev)]
 
                )
 
                other_changesets = [org_repo.get_changeset(cs)
 
                              for cs in re.findall(r'[0-9a-fA-F]{40}', so)]
 
                so = org_repo.run_git_command(
 
                    ['merge-base', org_rev, other_rev]
 
                )
 
                ancestors = [re.findall(r'[0-9a-fA-F]{40}', so)[0]]
 
            org_changesets = []
 

	
 
        else:
 
            raise Exception('Bad alias only git and hg is allowed')
 

	
 
        return other_changesets, org_changesets, ancestors
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def index(self, repo_name):
 
        c.compare_home = True
 
        c.a_ref_name = c.cs_ref_name = None
 
        return render('compare/compare_diff.html')
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def compare(self, repo_name, org_ref_type, org_ref_name, other_ref_type, other_ref_name):
 
        org_ref_name = org_ref_name.strip()
 
        other_ref_name = other_ref_name.strip()
 

	
 
        # If merge is True:
 
        #   Show what org would get if merged with other:
 
        #   List changesets that are ancestors of other but not of org.
 
        #   New changesets in org is thus ignored.
 
        #   Diff will be from common ancestor, and merges of org to other will thus be ignored.
 
        # If merge is False:
 
        #   Make a raw diff from org to other, no matter if related or not.
 
        #   Changesets in one and not in the other will be ignored
 
        merge = bool(request.GET.get('merge'))
 
        # fulldiff disables cut_off_limit
 
        fulldiff = request.GET.get('fulldiff')
 
        # partial uses compare_cs.html template directly
 
        partial = request.environ.get('HTTP_X_PARTIAL_XHR')
 
        # is_ajax_preview puts hidden input field with changeset revisions
 
        c.is_ajax_preview = partial and request.GET.get('is_ajax_preview')
 
        # swap url for compare_diff page - never partial and never is_ajax_preview
 
        c.swap_url = h.url('compare_url',
 
            repo_name=c.cs_repo.repo_name,
 
            org_ref_type=other_ref_type, org_ref_name=other_ref_name,
 
            other_repo=c.a_repo.repo_name,
 
            other_ref_type=org_ref_type, other_ref_name=org_ref_name,
 
            merge=merge or '')
 

	
 
        # set callbacks for generating markup for icons
 
        c.ignorews_url = _ignorews_url
 
        c.context_url = _context_url
 
        ignore_whitespace = request.GET.get('ignorews') == '1'
 
        line_context = safe_int(request.GET.get('context'), 3)
 

	
 
        c.a_rev = self._get_ref_rev(c.a_repo, org_ref_type, org_ref_name,
 
            returnempty=True)
 
        c.cs_rev = self._get_ref_rev(c.cs_repo, other_ref_type, other_ref_name)
 

	
 
        c.compare_home = False
 
        c.a_ref_name = org_ref_name
 
        c.a_ref_type = org_ref_type
 
        c.cs_ref_name = other_ref_name
 
        c.cs_ref_type = other_ref_type
 

	
 
        c.cs_ranges, c.cs_ranges_org, c.ancestors = self._get_changesets(
 
            c.a_repo.scm_instance.alias, c.a_repo.scm_instance, c.a_rev,
 
            c.cs_repo.scm_instance, c.cs_rev)
 
        raw_ids = [x.raw_id for x in c.cs_ranges]
 
        c.cs_comments = c.cs_repo.get_comments(raw_ids)
 
        c.cs_statuses = c.cs_repo.statuses(raw_ids)
 

	
 
        revs = [ctx.revision for ctx in reversed(c.cs_ranges)]
 
        c.jsdata = graph_data(c.cs_repo.scm_instance, revs)
 

	
 
        if partial:
 
            return render('compare/compare_cs.html')
 

	
 
        org_repo = c.a_repo
 
        other_repo = c.cs_repo
 

	
 
        if merge:
 
            rev1 = msg = None
 
            if not c.cs_ranges:
 
                msg = _('Cannot show empty diff')
 
            elif not c.ancestors:
 
                msg = _('No ancestor found for merge diff')
 
            elif len(c.ancestors) == 1:
 
                rev1 = c.ancestors[0]
 
            else:
 
                msg = _('Multiple merge ancestors found for merge compare')
 
            if rev1 is None:
 
                h.flash(msg, category='error')
 
                log.error(msg)
 
                raise HTTPNotFound
 

	
 
            # case we want a simple diff without incoming changesets,
 
            # previewing what will be merged.
 
            # Make the diff on the other repo (which is known to have other_rev)
 
            log.debug('Using ancestor %s as rev1 instead of %s',
 
                      rev1, c.a_rev)
 
            org_repo = other_repo
 
        else: # comparing tips, not necessarily linearly related
 
            if org_repo != other_repo:
 
                # TODO: we could do this by using hg unionrepo
 
                log.error('cannot compare across repos %s and %s', org_repo, other_repo)
 
                h.flash(_('Cannot compare repositories without using common ancestor'), category='error')
 
                raise HTTPBadRequest
 
            rev1 = c.a_rev
 

	
 
        diff_limit = None if fulldiff else self.cut_off_limit
 

	
 
        log.debug('running diff between %s and %s in %s',
 
                  rev1, c.cs_rev, org_repo.scm_instance.path)
 
        raw_diff = diffs.get_diff(org_repo.scm_instance, rev1=rev1, rev2=c.cs_rev,
 
                                      ignore_whitespace=ignore_whitespace,
 
                                      context=line_context)
 

	
 
        diff_processor = diffs.DiffProcessor(raw_diff, diff_limit=diff_limit)
 
        c.limited_diff = diff_processor.limited_diff
 
        c.file_diff_data = []
 
        c.lines_added = 0
 
        c.lines_deleted = 0
 
        for f in diff_processor.parsed:
 
            st = f['stats']
 
            c.lines_added += st['added']
 
            c.lines_deleted += st['deleted']
 
            filename = f['filename']
 
            fid = h.FID('', filename)
 
            html_diff = diffs.as_html(enable_comments=False, parsed_lines=[f])
 
            c.file_diff_data.append((fid, None, f['operation'], f['old_filename'], filename, html_diff, st))
 

	
 
        return render('compare/compare_diff.html')
kallithea/controllers/pullrequests.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
# This program is free software: you can redistribute it and/or modify
 
# it under the terms of the GNU General Public License as published by
 
# the Free Software Foundation, either version 3 of the License, or
 
# (at your option) any later version.
 
#
 
# 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 General Public License
 
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
"""
 
kallithea.controllers.pullrequests
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
pull requests controller for Kallithea for initializing pull requests
 

	
 
This file was forked by the Kallithea project in July 2014.
 
Original author and date, and relevant copyright and licensing information is below:
 
:created_on: May 7, 2012
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
import logging
 
import traceback
 

	
 
import formencode
 
import mercurial.unionrepo
 
from tg import request
 
from tg import tmpl_context as c
 
from tg.i18n import ugettext as _
 
from webob.exc import HTTPBadRequest, HTTPForbidden, HTTPFound, HTTPNotFound
 

	
 
from kallithea.config.routing import url
 
from kallithea.controllers.changeset import _context_url, _ignorews_url, create_cs_pr_comment, delete_cs_pr_comment
 
from kallithea.lib import diffs
 
from kallithea.lib import helpers as h
 
from kallithea.lib.auth import HasRepoPermissionLevelDecorator, LoginRequired
 
from kallithea.lib.base import BaseRepoController, jsonify, render
 
from kallithea.lib.graphmod import graph_data
 
from kallithea.lib.page import Page
 
from kallithea.lib.utils2 import ascii_bytes, safe_bytes, safe_int, safe_str
 
from kallithea.lib.vcs.exceptions import ChangesetDoesNotExistError, EmptyRepositoryError
 
from kallithea.lib.vcs.utils.hgcompat import unionrepo
 
from kallithea.model.changeset_status import ChangesetStatusModel
 
from kallithea.model.comment import ChangesetCommentsModel
 
from kallithea.model.db import ChangesetStatus, PullRequest, PullRequestReviewer, Repository, User
 
from kallithea.model.forms import PullRequestForm, PullRequestPostForm
 
from kallithea.model.meta import Session
 
from kallithea.model.pull_request import CreatePullRequestAction, CreatePullRequestIterationAction, PullRequestModel
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
def _get_reviewer(user_id):
 
    """Look up user by ID and validate it as a potential reviewer."""
 
    try:
 
        user = User.get(int(user_id))
 
    except ValueError:
 
        user = None
 

	
 
    if user is None or user.is_default_user:
 
        h.flash(_('Invalid reviewer "%s" specified') % user_id, category='error')
 
        raise HTTPBadRequest()
 

	
 
    return user
 

	
 

	
 
class PullrequestsController(BaseRepoController):
 

	
 
    def _get_repo_refs(self, repo, rev=None, branch=None, branch_rev=None):
 
        """return a structure with repo's interesting changesets, suitable for
 
        the selectors in pullrequest.html
 

	
 
        rev: a revision that must be in the list somehow and selected by default
 
        branch: a branch that must be in the list and selected by default - even if closed
 
        branch_rev: a revision of which peers should be preferred and available."""
 
        # list named branches that has been merged to this named branch - it should probably merge back
 
        peers = []
 

	
 
        if rev:
 
            rev = safe_str(rev)
 

	
 
        if branch:
 
            branch = safe_str(branch)
 

	
 
        if branch_rev:
 
            # a revset not restricting to merge() would be better
 
            # (especially because it would get the branch point)
 
            # ... but is currently too expensive
 
            # including branches of children could be nice too
 
            peerbranches = set()
 
            for i in repo._repo.revs(
 
                b"sort(parents(branch(id(%s)) and merge()) - branch(id(%s)), -rev)",
 
                ascii_bytes(branch_rev), ascii_bytes(branch_rev),
 
            ):
 
                for abranch in repo.get_changeset(i).branches:
 
                    if abranch not in peerbranches:
 
                        n = 'branch:%s:%s' % (abranch, repo.get_changeset(abranch).raw_id)
 
                        peers.append((n, abranch))
 
                        peerbranches.add(abranch)
 

	
 
        selected = None
 
        tiprev = repo.tags.get('tip')
 
        tipbranch = None
 

	
 
        branches = []
 
        for abranch, branchrev in repo.branches.iteritems():
 
            n = 'branch:%s:%s' % (abranch, branchrev)
 
            desc = abranch
 
            if branchrev == tiprev:
 
                tipbranch = abranch
 
                desc = '%s (current tip)' % desc
 
            branches.append((n, desc))
 
            if rev == branchrev:
 
                selected = n
 
            if branch == abranch:
 
                if not rev:
 
                    selected = n
 
                branch = None
 
        if branch:  # branch not in list - it is probably closed
 
            branchrev = repo.closed_branches.get(branch)
 
            if branchrev:
 
                n = 'branch:%s:%s' % (branch, branchrev)
 
                branches.append((n, _('%s (closed)') % branch))
 
                selected = n
 
                branch = None
 
            if branch:
 
                log.debug('branch %r not found in %s', branch, repo)
 

	
 
        bookmarks = []
 
        for bookmark, bookmarkrev in repo.bookmarks.iteritems():
 
            n = 'book:%s:%s' % (bookmark, bookmarkrev)
 
            bookmarks.append((n, bookmark))
 
            if rev == bookmarkrev:
 
                selected = n
 

	
 
        tags = []
 
        for tag, tagrev in repo.tags.iteritems():
 
            if tag == 'tip':
 
                continue
 
            n = 'tag:%s:%s' % (tag, tagrev)
 
            tags.append((n, tag))
 
            # note: even if rev == tagrev, don't select the static tag - it must be chosen explicitly
 

	
 
        # prio 1: rev was selected as existing entry above
 

	
 
        # prio 2: create special entry for rev; rev _must_ be used
 
        specials = []
 
        if rev and selected is None:
 
            selected = 'rev:%s:%s' % (rev, rev)
 
            specials = [(selected, '%s: %s' % (_("Changeset"), rev[:12]))]
 

	
 
        # prio 3: most recent peer branch
 
        if peers and not selected:
 
            selected = peers[0][0]
 

	
 
        # prio 4: tip revision
 
        if not selected:
 
            if h.is_hg(repo):
 
                if tipbranch:
 
                    selected = 'branch:%s:%s' % (tipbranch, tiprev)
 
                else:
 
                    selected = 'tag:null:' + repo.EMPTY_CHANGESET
 
                    tags.append((selected, 'null'))
 
            else:
 
                if 'master' in repo.branches:
 
                    selected = 'branch:master:%s' % repo.branches['master']
 
                else:
 
                    k, v = list(repo.branches.items())[0]
 
                    selected = 'branch:%s:%s' % (k, v)
 

	
 
        groups = [(specials, _("Special")),
 
                  (peers, _("Peer branches")),
 
                  (bookmarks, _("Bookmarks")),
 
                  (branches, _("Branches")),
 
                  (tags, _("Tags")),
 
                  ]
 
        return [g for g in groups if g[0]], selected
 

	
 
    def _is_allowed_to_change_status(self, pull_request):
 
        if pull_request.is_closed():
 
            return False
 

	
 
        owner = request.authuser.user_id == pull_request.owner_id
 
        reviewer = PullRequestReviewer.query() \
 
            .filter(PullRequestReviewer.pull_request == pull_request) \
 
            .filter(PullRequestReviewer.user_id == request.authuser.user_id) \
 
            .count() != 0
 

	
 
        return request.authuser.admin or owner or reviewer
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def show_all(self, repo_name):
 
        c.from_ = request.GET.get('from_') or ''
 
        c.closed = request.GET.get('closed') or ''
 
        url_params = {}
 
        if c.from_:
 
            url_params['from_'] = 1
 
        if c.closed:
 
            url_params['closed'] = 1
 
        p = safe_int(request.GET.get('page'), 1)
 

	
 
        q = PullRequest.query(include_closed=c.closed, sorted=True)
 
        if c.from_:
 
            q = q.filter_by(org_repo=c.db_repo)
 
        else:
 
            q = q.filter_by(other_repo=c.db_repo)
 
        c.pull_requests = q.all()
 

	
 
        c.pullrequests_pager = Page(c.pull_requests, page=p, items_per_page=100, **url_params)
 

	
 
        return render('/pullrequests/pullrequest_show_all.html')
 

	
 
    @LoginRequired()
 
    def show_my(self):
 
        c.closed = request.GET.get('closed') or ''
 

	
 
        c.my_pull_requests = PullRequest.query(
 
            include_closed=c.closed,
 
            sorted=True,
 
        ).filter_by(owner_id=request.authuser.user_id).all()
 

	
 
        c.participate_in_pull_requests = []
 
        c.participate_in_pull_requests_todo = []
 
        done_status = set([ChangesetStatus.STATUS_APPROVED, ChangesetStatus.STATUS_REJECTED])
 
        for pr in PullRequest.query(
 
            include_closed=c.closed,
 
            reviewer_id=request.authuser.user_id,
 
            sorted=True,
 
        ):
 
            status = pr.user_review_status(request.authuser.user_id) # very inefficient!!!
 
            if status in done_status:
 
                c.participate_in_pull_requests.append(pr)
 
            else:
 
                c.participate_in_pull_requests_todo.append(pr)
 

	
 
        return render('/pullrequests/pullrequest_show_my.html')
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('read')
 
    def index(self):
 
        org_repo = c.db_repo
 
        org_scm_instance = org_repo.scm_instance
 
        try:
 
            org_scm_instance.get_changeset()
 
        except EmptyRepositoryError as e:
 
            h.flash(_('There are no changesets yet'),
 
                    category='warning')
 
            raise HTTPFound(location=url('summary_home', repo_name=org_repo.repo_name))
 

	
 
        org_rev = request.GET.get('rev_end')
 
        # rev_start is not directly useful - its parent could however be used
 
        # as default for other and thus give a simple compare view
 
        rev_start = request.GET.get('rev_start')
 
        other_rev = None
 
        if rev_start:
 
            starters = org_repo.get_changeset(rev_start).parents
 
            if starters:
 
                other_rev = starters[0].raw_id
 
            else:
 
                other_rev = org_repo.scm_instance.EMPTY_CHANGESET
 
        branch = request.GET.get('branch')
 

	
 
        c.cs_repos = [(org_repo.repo_name, org_repo.repo_name)]
 
        c.default_cs_repo = org_repo.repo_name
 
        c.cs_refs, c.default_cs_ref = self._get_repo_refs(org_scm_instance, rev=org_rev, branch=branch)
 

	
 
        default_cs_ref_type, default_cs_branch, default_cs_rev = c.default_cs_ref.split(':')
 
        if default_cs_ref_type != 'branch':
 
            default_cs_branch = org_repo.get_changeset(default_cs_rev).branch
 

	
 
        # add org repo to other so we can open pull request against peer branches on itself
 
        c.a_repos = [(org_repo.repo_name, '%s (self)' % org_repo.repo_name)]
 

	
 
        if org_repo.parent:
 
            # add parent of this fork also and select it.
 
            # use the same branch on destination as on source, if available.
 
            c.a_repos.append((org_repo.parent.repo_name, '%s (parent)' % org_repo.parent.repo_name))
 
            c.a_repo = org_repo.parent
 
            c.a_refs, c.default_a_ref = self._get_repo_refs(
 
                    org_repo.parent.scm_instance, branch=default_cs_branch, rev=other_rev)
 

	
 
        else:
 
            c.a_repo = org_repo
 
            c.a_refs, c.default_a_ref = self._get_repo_refs(org_scm_instance, rev=other_rev)
 

	
 
        # gather forks and add to this list ... even though it is rare to
 
        # request forks to pull from their parent
 
        for fork in org_repo.forks:
 
            c.a_repos.append((fork.repo_name, fork.repo_name))
 

	
 
        return render('/pullrequests/pullrequest.html')
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('read')
 
    @jsonify
 
    def repo_info(self, repo_name):
 
        repo = c.db_repo
 
        refs, selected_ref = self._get_repo_refs(repo.scm_instance)
 
        return {
 
            'description': repo.description.split('\n', 1)[0],
 
            'selected_ref': selected_ref,
 
            'refs': refs,
 
            }
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('read')
 
    def create(self, repo_name):
 
        repo = c.db_repo
 
        try:
 
            _form = PullRequestForm(repo.repo_id)().to_python(request.POST)
 
        except formencode.Invalid as errors:
 
            log.error(traceback.format_exc())
 
            log.error(str(errors))
 
            msg = _('Error creating pull request: %s') % errors.msg
 
            h.flash(msg, 'error')
 
            raise HTTPBadRequest
 

	
 
        # heads up: org and other might seem backward here ...
 
        org_ref = _form['org_ref'] # will have merge_rev as rev but symbolic name
 
        org_repo = Repository.guess_instance(_form['org_repo'])
 

	
 
        other_ref = _form['other_ref'] # will have symbolic name and head revision
 
        other_repo = Repository.guess_instance(_form['other_repo'])
 

	
 
        reviewers = []
 

	
 
        title = _form['pullrequest_title']
 
        description = _form['pullrequest_desc'].strip()
 
        owner = User.get(request.authuser.user_id)
 

	
 
        try:
 
            cmd = CreatePullRequestAction(org_repo, other_repo, org_ref, other_ref, title, description, owner, reviewers)
 
        except CreatePullRequestAction.ValidationError as e:
 
            h.flash(e, category='error', logf=log.error)
 
            raise HTTPNotFound
 

	
 
        try:
 
            pull_request = cmd.execute()
 
            Session().commit()
 
        except Exception:
 
            h.flash(_('Error occurred while creating pull request'),
 
                    category='error')
 
            log.error(traceback.format_exc())
 
            raise HTTPFound(location=url('pullrequest_home', repo_name=repo_name))
 

	
 
        h.flash(_('Successfully opened new pull request'),
 
                category='success')
 
        raise HTTPFound(location=pull_request.url())
 

	
 
    def create_new_iteration(self, old_pull_request, new_rev, title, description, reviewers):
 
        owner = User.get(request.authuser.user_id)
 
        new_org_rev = self._get_ref_rev(old_pull_request.org_repo, 'rev', new_rev)
 
        new_other_rev = self._get_ref_rev(old_pull_request.other_repo, old_pull_request.other_ref_parts[0], old_pull_request.other_ref_parts[1])
 
        try:
 
            cmd = CreatePullRequestIterationAction(old_pull_request, new_org_rev, new_other_rev, title, description, owner, reviewers)
 
        except CreatePullRequestAction.ValidationError as e:
 
            h.flash(e, category='error', logf=log.error)
 
            raise HTTPNotFound
 

	
 
        try:
 
            pull_request = cmd.execute()
 
            Session().commit()
 
        except Exception:
 
            h.flash(_('Error occurred while creating pull request'),
 
                    category='error')
 
            log.error(traceback.format_exc())
 
            raise HTTPFound(location=old_pull_request.url())
 

	
 
        h.flash(_('New pull request iteration created'),
 
                category='success')
 
        raise HTTPFound(location=pull_request.url())
 

	
 
    # pullrequest_post for PR editing
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('read')
 
    def post(self, repo_name, pull_request_id):
 
        pull_request = PullRequest.get_or_404(pull_request_id)
 
        if pull_request.is_closed():
 
            raise HTTPForbidden()
 
        assert pull_request.other_repo.repo_name == repo_name
 
        # only owner or admin can update it
 
        owner = pull_request.owner_id == request.authuser.user_id
 
        repo_admin = h.HasRepoPermissionLevel('admin')(c.repo_name)
 
        if not (h.HasPermissionAny('hg.admin')() or repo_admin or owner):
 
            raise HTTPForbidden()
 

	
 
        _form = PullRequestPostForm()().to_python(request.POST)
 

	
 
        cur_reviewers = set(pull_request.get_reviewer_users())
 
        new_reviewers = set(_get_reviewer(s) for s in _form['review_members'])
 
        old_reviewers = set(_get_reviewer(s) for s in _form['org_review_members'])
 

	
 
        other_added = cur_reviewers - old_reviewers
 
        other_removed = old_reviewers - cur_reviewers
 

	
 
        if other_added:
 
            h.flash(_('Meanwhile, the following reviewers have been added: %s') %
 
                    (', '.join(u.username for u in other_added)),
 
                    category='warning')
 
        if other_removed:
 
            h.flash(_('Meanwhile, the following reviewers have been removed: %s') %
 
                    (', '.join(u.username for u in other_removed)),
 
                    category='warning')
 

	
 
        if _form['updaterev']:
 
            return self.create_new_iteration(pull_request,
 
                                      _form['updaterev'],
 
                                      _form['pullrequest_title'],
 
                                      _form['pullrequest_desc'],
 
                                      new_reviewers)
 

	
 
        added_reviewers = new_reviewers - old_reviewers - cur_reviewers
 
        removed_reviewers = (old_reviewers - new_reviewers) & cur_reviewers
 

	
 
        old_description = pull_request.description
 
        pull_request.title = _form['pullrequest_title']
 
        pull_request.description = _form['pullrequest_desc'].strip() or _('No description')
 
        pull_request.owner = User.get_by_username(_form['owner'])
 
        user = User.get(request.authuser.user_id)
 

	
 
        PullRequestModel().mention_from_description(user, pull_request, old_description)
 
        PullRequestModel().add_reviewers(user, pull_request, added_reviewers)
 
        PullRequestModel().remove_reviewers(user, pull_request, removed_reviewers)
 

	
 
        Session().commit()
 
        h.flash(_('Pull request updated'), category='success')
 

	
 
        raise HTTPFound(location=pull_request.url())
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('read')
 
    @jsonify
 
    def delete(self, repo_name, pull_request_id):
 
        pull_request = PullRequest.get_or_404(pull_request_id)
 
        # only owner can delete it !
 
        if pull_request.owner_id == request.authuser.user_id:
 
            PullRequestModel().delete(pull_request)
 
            Session().commit()
 
            h.flash(_('Successfully deleted pull request'),
 
                    category='success')
 
            raise HTTPFound(location=url('my_pullrequests'))
 
        raise HTTPForbidden()
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def show(self, repo_name, pull_request_id, extra=None):
 
        c.pull_request = PullRequest.get_or_404(pull_request_id)
 
        c.allowed_to_change_status = self._is_allowed_to_change_status(c.pull_request)
 
        cc_model = ChangesetCommentsModel()
 
        cs_model = ChangesetStatusModel()
 

	
 
        # pull_requests repo_name we opened it against
 
        # ie. other_repo must match
 
        if repo_name != c.pull_request.other_repo.repo_name:
 
            raise HTTPNotFound
 

	
 
        # load compare data into template context
 
        c.cs_repo = c.pull_request.org_repo
 
        (c.cs_ref_type,
 
         c.cs_ref_name,
 
         c.cs_rev) = c.pull_request.org_ref.split(':')
 

	
 
        c.a_repo = c.pull_request.other_repo
 
        (c.a_ref_type,
 
         c.a_ref_name,
 
         c.a_rev) = c.pull_request.other_ref.split(':') # a_rev is ancestor
 

	
 
        org_scm_instance = c.cs_repo.scm_instance # property with expensive cache invalidation check!!!
 
        try:
 
            c.cs_ranges = []
 
            for x in c.pull_request.revisions:
 
                c.cs_ranges.append(org_scm_instance.get_changeset(x))
 
        except ChangesetDoesNotExistError:
 
            c.cs_ranges = []
 
            h.flash(_('Revision %s not found in %s') % (x, c.cs_repo.repo_name),
 
                'error')
 
        c.cs_ranges_org = None # not stored and not important and moving target - could be calculated ...
 
        revs = [ctx.revision for ctx in reversed(c.cs_ranges)]
 
        c.jsdata = graph_data(org_scm_instance, revs)
 

	
 
        c.is_range = False
 
        try:
 
            if c.a_ref_type == 'rev': # this looks like a free range where target is ancestor
 
                cs_a = org_scm_instance.get_changeset(c.a_rev)
 
                root_parents = c.cs_ranges[0].parents
 
                c.is_range = cs_a in root_parents
 
                #c.merge_root = len(root_parents) > 1 # a range starting with a merge might deserve a warning
 
        except ChangesetDoesNotExistError: # probably because c.a_rev not found
 
            pass
 
        except IndexError: # probably because c.cs_ranges is empty, probably because revisions are missing
 
            pass
 

	
 
        avail_revs = set()
 
        avail_show = []
 
        c.cs_branch_name = c.cs_ref_name
 
        c.a_branch_name = None
 
        other_scm_instance = c.a_repo.scm_instance
 
        c.update_msg = ""
 
        c.update_msg_other = ""
 
        try:
 
            if not c.cs_ranges:
 
                c.update_msg = _('Error: changesets not found when displaying pull request from %s.') % c.cs_rev
 
            elif org_scm_instance.alias == 'hg' and c.a_ref_name != 'ancestor':
 
                if c.cs_ref_type != 'branch':
 
                    c.cs_branch_name = org_scm_instance.get_changeset(c.cs_ref_name).branch # use ref_type ?
 
                c.a_branch_name = c.a_ref_name
 
                if c.a_ref_type != 'branch':
 
                    try:
 
                        c.a_branch_name = other_scm_instance.get_changeset(c.a_ref_name).branch # use ref_type ?
 
                    except EmptyRepositoryError:
 
                        c.a_branch_name = 'null' # not a branch name ... but close enough
 
                # candidates: descendants of old head that are on the right branch
 
                #             and not are the old head itself ...
 
                #             and nothing at all if old head is a descendant of target ref name
 
                if not c.is_range and other_scm_instance._repo.revs('present(%s)::&%s', c.cs_ranges[-1].raw_id, c.a_branch_name):
 
                    c.update_msg = _('This pull request has already been merged to %s.') % c.a_branch_name
 
                elif c.pull_request.is_closed():
 
                    c.update_msg = _('This pull request has been closed and can not be updated.')
 
                else: # look for descendants of PR head on source branch in org repo
 
                    avail_revs = org_scm_instance._repo.revs('%s:: & branch(%s)',
 
                                                             revs[0], c.cs_branch_name)
 
                    if len(avail_revs) > 1: # more than just revs[0]
 
                        # also show changesets that not are descendants but would be merged in
 
                        targethead = other_scm_instance.get_changeset(c.a_branch_name).raw_id
 
                        if org_scm_instance.path != other_scm_instance.path:
 
                            # Note: org_scm_instance.path must come first so all
 
                            # valid revision numbers are 100% org_scm compatible
 
                            # - both for avail_revs and for revset results
 
                            hgrepo = unionrepo.makeunionrepository(org_scm_instance.baseui,
 
                            hgrepo = mercurial.unionrepo.makeunionrepository(org_scm_instance.baseui,
 
                                                                   org_scm_instance.path,
 
                                                                   other_scm_instance.path)
 
                        else:
 
                            hgrepo = org_scm_instance._repo
 
                        show = set(hgrepo.revs('::%ld & !::parents(%s) & !::%s',
 
                                               avail_revs, revs[0], targethead))
 
                        if show:
 
                            c.update_msg = _('The following additional changes are available on %s:') % c.cs_branch_name
 
                        else:
 
                            c.update_msg = _('No additional changesets found for iterating on this pull request.')
 
                    else:
 
                        show = set()
 
                        avail_revs = set() # drop revs[0]
 
                        c.update_msg = _('No additional changesets found for iterating on this pull request.')
 

	
 
                    # TODO: handle branch heads that not are tip-most
 
                    brevs = org_scm_instance._repo.revs('%s - %ld - %s', c.cs_branch_name, avail_revs, revs[0])
 
                    if brevs:
 
                        # also show changesets that are on branch but neither ancestors nor descendants
 
                        show.update(org_scm_instance._repo.revs('::%ld - ::%ld - ::%s', brevs, avail_revs, c.a_branch_name))
 
                        show.add(revs[0]) # make sure graph shows this so we can see how they relate
 
                        c.update_msg_other = _('Note: Branch %s has another head: %s.') % (c.cs_branch_name,
 
                            h.short_id(org_scm_instance.get_changeset((max(brevs))).raw_id))
 

	
 
                    avail_show = sorted(show, reverse=True)
 

	
 
            elif org_scm_instance.alias == 'git':
 
                c.cs_repo.scm_instance.get_changeset(c.cs_rev) # check it exists - raise ChangesetDoesNotExistError if not
 
                c.update_msg = _("Git pull requests don't support iterating yet.")
 
        except ChangesetDoesNotExistError:
 
            c.update_msg = _('Error: some changesets not found when displaying pull request from %s.') % c.cs_rev
 

	
 
        c.avail_revs = avail_revs
 
        c.avail_cs = [org_scm_instance.get_changeset(r) for r in avail_show]
 
        c.avail_jsdata = graph_data(org_scm_instance, avail_show)
 

	
 
        raw_ids = [x.raw_id for x in c.cs_ranges]
 
        c.cs_comments = c.cs_repo.get_comments(raw_ids)
 
        c.cs_statuses = c.cs_repo.statuses(raw_ids)
 

	
 
        ignore_whitespace = request.GET.get('ignorews') == '1'
 
        line_context = safe_int(request.GET.get('context'), 3)
 
        c.ignorews_url = _ignorews_url
 
        c.context_url = _context_url
 
        fulldiff = request.GET.get('fulldiff')
 
        diff_limit = None if fulldiff else self.cut_off_limit
 

	
 
        # we swap org/other ref since we run a simple diff on one repo
 
        log.debug('running diff between %s and %s in %s',
 
                  c.a_rev, c.cs_rev, org_scm_instance.path)
 
        try:
 
            raw_diff = diffs.get_diff(org_scm_instance, rev1=safe_str(c.a_rev), rev2=safe_str(c.cs_rev),
 
                                      ignore_whitespace=ignore_whitespace, context=line_context)
 
        except ChangesetDoesNotExistError:
 
            raw_diff = safe_bytes(_("The diff can't be shown - the PR revisions could not be found."))
 
        diff_processor = diffs.DiffProcessor(raw_diff, diff_limit=diff_limit)
 
        c.limited_diff = diff_processor.limited_diff
 
        c.file_diff_data = []
 
        c.lines_added = 0
 
        c.lines_deleted = 0
 

	
 
        for f in diff_processor.parsed:
 
            st = f['stats']
 
            c.lines_added += st['added']
 
            c.lines_deleted += st['deleted']
 
            filename = f['filename']
 
            fid = h.FID('', filename)
 
            html_diff = diffs.as_html(enable_comments=True, parsed_lines=[f])
 
            c.file_diff_data.append((fid, None, f['operation'], f['old_filename'], filename, html_diff, st))
 

	
 
        # inline comments
 
        c.inline_cnt = 0
 
        c.inline_comments = cc_model.get_inline_comments(
 
                                c.db_repo.repo_id,
 
                                pull_request=pull_request_id)
 
        # count inline comments
 
        for __, lines in c.inline_comments:
 
            for comments in lines.values():
 
                c.inline_cnt += len(comments)
 
        # comments
 
        c.comments = cc_model.get_comments(c.db_repo.repo_id, pull_request=pull_request_id)
 

	
 
        # (badly named) pull-request status calculation based on reviewer votes
 
        (c.pull_request_reviewers,
 
         c.pull_request_pending_reviewers,
 
         c.current_voting_result,
 
         ) = cs_model.calculate_pull_request_result(c.pull_request)
 
        c.changeset_statuses = ChangesetStatus.STATUSES
 

	
 
        c.is_ajax_preview = False
 
        c.ancestors = None # [c.a_rev] ... but that is shown in an other way
 
        return render('/pullrequests/pullrequest_show.html')
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('read')
 
    @jsonify
 
    def comment(self, repo_name, pull_request_id):
 
        pull_request = PullRequest.get_or_404(pull_request_id)
 
        allowed_to_change_status = self._is_allowed_to_change_status(pull_request)
 
        return create_cs_pr_comment(repo_name, pull_request=pull_request,
 
                allowed_to_change_status=allowed_to_change_status)
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('read')
 
    @jsonify
 
    def delete_comment(self, repo_name, comment_id):
 
        return delete_cs_pr_comment(repo_name, comment_id)
kallithea/lib/hooks.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
# This program is free software: you can redistribute it and/or modify
 
# it under the terms of the GNU General Public License as published by
 
# the Free Software Foundation, either version 3 of the License, or
 
# (at your option) any later version.
 
#
 
# 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 General Public License
 
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
"""
 
kallithea.lib.hooks
 
~~~~~~~~~~~~~~~~~~~
 

	
 
Hooks run by Kallithea
 

	
 
This file was forked by the Kallithea project in July 2014.
 
Original author and date, and relevant copyright and licensing information is below:
 
:created_on: Aug 6, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
import os
 
import time
 

	
 
import mercurial.scmutil
 

	
 
from kallithea.lib import helpers as h
 
from kallithea.lib.exceptions import UserCreationError
 
from kallithea.lib.utils import action_logger, make_ui, setup_cache_regions
 
from kallithea.lib.utils2 import ascii_str, get_hook_environment, safe_str, safe_unicode
 
from kallithea.lib.vcs.backends.base import EmptyChangeset
 
from kallithea.lib.vcs.utils.hgcompat import revrange
 
from kallithea.model.db import Repository, User
 

	
 

	
 
def _get_scm_size(alias, root_path):
 
    if not alias.startswith('.'):
 
        alias += '.'
 

	
 
    size_scm, size_root = 0, 0
 
    for path, dirs, files in os.walk(safe_str(root_path)):
 
        if path.find(alias) != -1:
 
            for f in files:
 
                try:
 
                    size_scm += os.path.getsize(os.path.join(path, f))
 
                except OSError:
 
                    pass
 
        else:
 
            for f in files:
 
                try:
 
                    size_root += os.path.getsize(os.path.join(path, f))
 
                except OSError:
 
                    pass
 

	
 
    size_scm_f = h.format_byte_size(size_scm)
 
    size_root_f = h.format_byte_size(size_root)
 
    size_total_f = h.format_byte_size(size_root + size_scm)
 

	
 
    return size_scm_f, size_root_f, size_total_f
 

	
 

	
 
def repo_size(ui, repo, hooktype=None, **kwargs):
 
    """Presents size of repository after push"""
 
    size_hg_f, size_root_f, size_total_f = _get_scm_size('.hg', repo.root)
 

	
 
    last_cs = repo[len(repo) - 1]
 

	
 
    msg = ('Repository size .hg: %s Checkout: %s Total: %s\n'
 
           'Last revision is now r%s:%s\n') % (
 
        size_hg_f, size_root_f, size_total_f, last_cs.rev(), ascii_str(last_cs.hex())[:12]
 
    )
 
    ui.status(msg)
 

	
 

	
 
def log_pull_action(ui, repo, **kwargs):
 
    """Logs user last pull action
 

	
 
    Called as Mercurial hook outgoing.pull_logger or from Kallithea before invoking Git.
 

	
 
    Does *not* use the action from the hook environment but is always 'pull'.
 
    """
 
    ex = get_hook_environment()
 

	
 
    user = User.get_by_username(ex.username)
 
    action = 'pull'
 
    action_logger(user, action, ex.repository, ex.ip, commit=True)
 
    # extension hook call
 
    from kallithea import EXTENSIONS
 
    callback = getattr(EXTENSIONS, 'PULL_HOOK', None)
 
    if callable(callback):
 
        kw = {}
 
        kw.update(ex)
 
        callback(**kw)
 

	
 
    return 0
 

	
 

	
 
def log_push_action(ui, repo, node, node_last, **kwargs):
 
    """
 
    Entry point for Mercurial hook changegroup.push_logger.
 

	
 
    The pushed changesets is given by the revset 'node:node_last'.
 

	
 
    Note: This hook is not only logging, but also the side effect invalidating
 
    cahes! The function should perhaps be renamed.
 
    """
 
    revs = [ascii_str(repo[r].hex()) for r in revrange(repo, [b'%s:%s' % (node, node_last)])]
 
    revs = [ascii_str(repo[r].hex()) for r in mercurial.scmutil.revrange(repo, [b'%s:%s' % (node, node_last)])]
 
    process_pushed_raw_ids(revs)
 
    return 0
 

	
 

	
 
def process_pushed_raw_ids(revs):
 
    """
 
    Register that changes have been added to the repo - log the action *and* invalidate caches.
 

	
 
    Called from  Mercurial changegroup.push_logger calling hook log_push_action,
 
    or from the Git post-receive hook calling handle_git_post_receive ...
 
    or from scm _handle_push.
 
    """
 
    ex = get_hook_environment()
 

	
 
    action = '%s:%s' % (ex.action, ','.join(revs))
 
    action_logger(ex.username, action, ex.repository, ex.ip, commit=True)
 

	
 
    from kallithea.model.scm import ScmModel
 
    ScmModel().mark_for_invalidation(ex.repository)
 

	
 
    # extension hook call
 
    from kallithea import EXTENSIONS
 
    callback = getattr(EXTENSIONS, 'PUSH_HOOK', None)
 
    if callable(callback):
 
        kw = {'pushed_revs': revs}
 
        kw.update(ex)
 
        callback(**kw)
 

	
 

	
 
def log_create_repository(repository_dict, created_by, **kwargs):
 
    """
 
    Post create repository Hook.
 

	
 
    :param repository: dict dump of repository object
 
    :param created_by: username who created repository
 

	
 
    available keys of repository_dict:
 

	
 
     'repo_type',
 
     'description',
 
     'private',
 
     'created_on',
 
     'enable_downloads',
 
     'repo_id',
 
     'owner_id',
 
     'enable_statistics',
 
     'clone_uri',
 
     'fork_id',
 
     'group_id',
 
     'repo_name'
 

	
 
    """
 
    from kallithea import EXTENSIONS
 
    callback = getattr(EXTENSIONS, 'CREATE_REPO_HOOK', None)
 
    if callable(callback):
 
        kw = {}
 
        kw.update(repository_dict)
 
        kw.update({'created_by': created_by})
 
        kw.update(kwargs)
 
        return callback(**kw)
 

	
 
    return 0
 

	
 

	
 
def check_allowed_create_user(user_dict, created_by, **kwargs):
 
    # pre create hooks
 
    from kallithea import EXTENSIONS
 
    callback = getattr(EXTENSIONS, 'PRE_CREATE_USER_HOOK', None)
 
    if callable(callback):
 
        allowed, reason = callback(created_by=created_by, **user_dict)
 
        if not allowed:
 
            raise UserCreationError(reason)
 

	
 

	
 
def log_create_user(user_dict, created_by, **kwargs):
 
    """
 
    Post create user Hook.
 

	
 
    :param user_dict: dict dump of user object
 

	
 
    available keys for user_dict:
 

	
 
     'username',
 
     'full_name_or_username',
 
     'full_contact',
 
     'user_id',
 
     'name',
 
     'firstname',
 
     'short_contact',
 
     'admin',
 
     'lastname',
 
     'ip_addresses',
 
     'ldap_dn',
 
     'email',
 
     'api_key',
 
     'last_login',
 
     'full_name',
 
     'active',
 
     'password',
 
     'emails',
 

	
 
    """
 
    from kallithea import EXTENSIONS
 
    callback = getattr(EXTENSIONS, 'CREATE_USER_HOOK', None)
 
    if callable(callback):
 
        return callback(created_by=created_by, **user_dict)
 

	
 
    return 0
 

	
 

	
 
def log_delete_repository(repository_dict, deleted_by, **kwargs):
 
    """
 
    Post delete repository Hook.
 

	
 
    :param repository: dict dump of repository object
 
    :param deleted_by: username who deleted the repository
 

	
 
    available keys of repository_dict:
 

	
 
     'repo_type',
 
     'description',
 
     'private',
 
     'created_on',
 
     'enable_downloads',
 
     'repo_id',
 
     'owner_id',
 
     'enable_statistics',
 
     'clone_uri',
 
     'fork_id',
 
     'group_id',
 
     'repo_name'
 

	
 
    """
 
    from kallithea import EXTENSIONS
 
    callback = getattr(EXTENSIONS, 'DELETE_REPO_HOOK', None)
 
    if callable(callback):
 
        kw = {}
 
        kw.update(repository_dict)
 
        kw.update({'deleted_by': deleted_by,
 
                   'deleted_on': time.time()})
 
        kw.update(kwargs)
 
        return callback(**kw)
 

	
 
    return 0
 

	
 

	
 
def log_delete_user(user_dict, deleted_by, **kwargs):
 
    """
 
    Post delete user Hook.
 

	
 
    :param user_dict: dict dump of user object
 

	
 
    available keys for user_dict:
 

	
 
     'username',
 
     'full_name_or_username',
 
     'full_contact',
 
     'user_id',
 
     'name',
 
     'firstname',
 
     'short_contact',
 
     'admin',
 
     'lastname',
 
     'ip_addresses',
 
     'ldap_dn',
 
     'email',
 
     'api_key',
 
     'last_login',
 
     'full_name',
 
     'active',
 
     'password',
 
     'emails',
 

	
 
    """
 
    from kallithea import EXTENSIONS
 
    callback = getattr(EXTENSIONS, 'DELETE_USER_HOOK', None)
 
    if callable(callback):
 
        return callback(deleted_by=deleted_by, **user_dict)
 

	
 
    return 0
 

	
 

	
 
def _hook_environment(repo_path):
 
    """
 
    Create a light-weight environment for stand-alone scripts and return an UI and the
 
    db repository.
 

	
 
    Git hooks are executed as subprocess of Git while Kallithea is waiting, and
 
    they thus need enough info to be able to create an app environment and
 
    connect to the database.
 
    """
 
    from paste.deploy import appconfig
 
    from sqlalchemy import engine_from_config
 
    from kallithea.config.environment import load_environment
 
    from kallithea.model.base import init_model
 

	
 
    extras = get_hook_environment()
 
    ini_file_path = extras['config']
 
    #logging.config.fileConfig(ini_file_path) # Note: we are in a different process - don't use configured logging
 
    app_conf = appconfig('config:%s' % ini_file_path)
 
    conf = load_environment(app_conf.global_conf, app_conf.local_conf)
 

	
 
    setup_cache_regions(conf)
 

	
 
    engine = engine_from_config(conf, 'sqlalchemy.')
 
    init_model(engine)
 

	
 
    repo_path = safe_unicode(repo_path)
 
    # fix if it's not a bare repo
 
    if repo_path.endswith(os.sep + '.git'):
 
        repo_path = repo_path[:-5]
 

	
 
    repo = Repository.get_by_full_path(repo_path)
 
    if not repo:
 
        raise OSError('Repository %s not found in database'
 
                      % (safe_str(repo_path)))
 

	
 
    baseui = make_ui()
 
    return baseui, repo
 

	
 

	
 
def handle_git_pre_receive(repo_path, git_stdin_lines):
 
    """Called from Git pre-receive hook"""
 
    # Currently unused. TODO: remove?
 
    return 0
 

	
 

	
 
def handle_git_post_receive(repo_path, git_stdin_lines):
 
    """Called from Git post-receive hook"""
 
    baseui, repo = _hook_environment(repo_path)
 

	
 
    # the post push hook should never use the cached instance
 
    scm_repo = repo.scm_instance_no_cache()
 

	
 
    rev_data = []
 
    for l in git_stdin_lines:
 
        old_rev, new_rev, ref = l.strip().split(' ')
 
        _ref_data = ref.split('/')
 
        if _ref_data[1] in ['tags', 'heads']:
 
            rev_data.append({'old_rev': old_rev,
 
                             'new_rev': new_rev,
 
                             'ref': ref,
 
                             'type': _ref_data[1],
 
                             'name': '/'.join(_ref_data[2:])})
 

	
 
    git_revs = []
 
    for push_ref in rev_data:
 
        _type = push_ref['type']
 
        if _type == 'heads':
 
            if push_ref['old_rev'] == EmptyChangeset().raw_id:
 
                # update the symbolic ref if we push new repo
 
                if scm_repo.is_empty():
 
                    scm_repo._repo.refs.set_symbolic_ref(
 
                        b'HEAD',
 
                        b'refs/heads/%s' % push_ref['name'])
 

	
 
                # build exclude list without the ref
 
                cmd = ['for-each-ref', '--format=%(refname)', 'refs/heads/*']
 
                stdout = scm_repo.run_git_command(cmd)
 
                ref = push_ref['ref']
 
                heads = [head for head in stdout.splitlines() if head != ref]
 
                # now list the git revs while excluding from the list
 
                cmd = ['log', push_ref['new_rev'], '--reverse', '--pretty=format:%H']
 
                cmd.append('--not')
 
                cmd.extend(heads) # empty list is ok
 
                stdout = scm_repo.run_git_command(cmd)
 
                git_revs += stdout.splitlines()
 

	
 
            elif push_ref['new_rev'] == EmptyChangeset().raw_id:
 
                # delete branch case
 
                git_revs += ['delete_branch=>%s' % push_ref['name']]
 
            else:
 
                cmd = ['log', '%(old_rev)s..%(new_rev)s' % push_ref,
 
                       '--reverse', '--pretty=format:%H']
 
                stdout = scm_repo.run_git_command(cmd)
 
                git_revs += stdout.splitlines()
 

	
 
        elif _type == 'tags':
 
            git_revs += ['tag=>%s' % push_ref['name']]
 

	
 
    process_pushed_raw_ids(git_revs)
 

	
 
    return 0
 

	
 

	
 
# Almost exactly like Mercurial contrib/hg-ssh:
 
def rejectpush(ui, **kwargs):
 
    """Mercurial hook to be installed as pretxnopen and prepushkey for read-only repos"""
 
    ex = get_hook_environment()
 
    ui.warn((b"Push access to %r denied\n") % safe_str(ex.repository))
 
    return 1
kallithea/lib/middleware/simplehg.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
# This program is free software: you can redistribute it and/or modify
 
# it under the terms of the GNU General Public License as published by
 
# the Free Software Foundation, either version 3 of the License, or
 
# (at your option) any later version.
 
#
 
# 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 General Public License
 
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
"""
 
kallithea.lib.middleware.simplehg
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
SimpleHg middleware for handling Mercurial protocol requests (push/clone etc.).
 
It's implemented with basic auth function
 

	
 
This file was forked by the Kallithea project in July 2014.
 
Original author and date, and relevant copyright and licensing information is below:
 
:created_on: Apr 28, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 

	
 
"""
 

	
 

	
 
import logging
 
import os
 
import urllib
 

	
 
import mercurial.hgweb
 

	
 
from kallithea.lib.base import BaseVCSController, get_path_info
 
from kallithea.lib.utils import make_ui
 
from kallithea.lib.utils2 import safe_str, safe_unicode
 
from kallithea.lib.vcs.utils.hgcompat import hgweb_mod
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
def get_header_hgarg(environ):
 
    """Decode the special Mercurial encoding of big requests over multiple headers.
 
    >>> get_header_hgarg({})
 
    ''
 
    >>> get_header_hgarg({'HTTP_X_HGARG_0': ' ', 'HTTP_X_HGARG_1': 'a','HTTP_X_HGARG_2': '','HTTP_X_HGARG_3': 'b+c %20'})
 
    'ab+c %20'
 
    """
 
    chunks = []
 
    i = 1
 
    while True:
 
        v = environ.get('HTTP_X_HGARG_%d' % i)
 
        if v is None:
 
            break
 
        chunks.append(v)
 
        i += 1
 
    return ''.join(chunks)
 

	
 

	
 
cmd_mapping = {
 
    # 'batch' is not in this list - it is handled explicitly
 
    'between': 'pull',
 
    'branches': 'pull',
 
    'branchmap': 'pull',
 
    'capabilities': 'pull',
 
    'changegroup': 'pull',
 
    'changegroupsubset': 'pull',
 
    'changesetdata': 'pull',
 
    'clonebundles': 'pull',
 
    'debugwireargs': 'pull',
 
    'filedata': 'pull',
 
    'getbundle': 'pull',
 
    'getlfile': 'pull',
 
    'heads': 'pull',
 
    'hello': 'pull',
 
    'known': 'pull',
 
    'lheads': 'pull',
 
    'listkeys': 'pull',
 
    'lookup': 'pull',
 
    'manifestdata': 'pull',
 
    'narrow_widen': 'pull',
 
    'protocaps': 'pull',
 
    'statlfile': 'pull',
 
    'stream_out': 'pull',
 
    'pushkey': 'push',
 
    'putlfile': 'push',
 
    'unbundle': 'push',
 
    }
 

	
 

	
 
class SimpleHg(BaseVCSController):
 

	
 
    scm_alias = 'hg'
 

	
 
    @classmethod
 
    def parse_request(cls, environ):
 
        http_accept = environ.get('HTTP_ACCEPT', '')
 
        if not http_accept.startswith('application/mercurial'):
 
            return None
 
        path_info = get_path_info(environ)
 
        if not path_info.startswith('/'): # it must!
 
            return None
 

	
 
        class parsed_request(object):
 
            repo_name = safe_unicode(path_info[1:].rstrip('/'))
 

	
 
            query_string = environ['QUERY_STRING']
 

	
 
            action = None
 
            for qry in query_string.split('&'):
 
                parts = qry.split('=', 1)
 
                if len(parts) == 2 and parts[0] == 'cmd':
 
                    cmd = parts[1]
 
                    if cmd == 'batch':
 
                        hgarg = get_header_hgarg(environ)
 
                        if not hgarg.startswith('cmds='):
 
                            action = 'push' # paranoid and safe
 
                            break
 
                        action = 'pull'
 
                        for cmd_arg in hgarg[5:].split(';'):
 
                            cmd, _args = urllib.unquote_plus(cmd_arg).split(' ', 1)
 
                            op = cmd_mapping.get(cmd, 'push')
 
                            if op != 'pull':
 
                                assert op == 'push'
 
                                action = 'push'
 
                                break
 
                    else:
 
                        action = cmd_mapping.get(cmd, 'push')
 
                    break # only process one cmd
 

	
 
        return parsed_request
 

	
 
    def _make_app(self, parsed_request):
 
        """
 
        Make an hgweb wsgi application.
 
        """
 
        str_repo_name = safe_str(parsed_request.repo_name)
 
        repo_path = os.path.join(safe_str(self.basepath), str_repo_name)
 
        baseui = make_ui(repo_path=repo_path)
 
        hgweb_app = hgweb_mod.hgweb(repo_path, name=str_repo_name, baseui=baseui)
 
        hgweb_app = mercurial.hgweb.hgweb(repo_path, name=str_repo_name, baseui=baseui)
 

	
 
        def wrapper_app(environ, start_response):
 
            environ['REPO_NAME'] = str_repo_name # used by hgweb_mod.hgweb
 
            environ['REPO_NAME'] = str_repo_name # used by mercurial.hgweb.hgweb
 
            return hgweb_app(environ, start_response)
 

	
 
        return wrapper_app
kallithea/lib/utils.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
# This program is free software: you can redistribute it and/or modify
 
# it under the terms of the GNU General Public License as published by
 
# the Free Software Foundation, either version 3 of the License, or
 
# (at your option) any later version.
 
#
 
# 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 General Public License
 
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
"""
 
kallithea.lib.utils
 
~~~~~~~~~~~~~~~~~~~
 

	
 
Utilities library for Kallithea
 

	
 
This file was forked by the Kallithea project in July 2014.
 
Original author and date, and relevant copyright and licensing information is below:
 
:created_on: Apr 18, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
import datetime
 
import logging
 
import os
 
import re
 
import sys
 
import traceback
 
from distutils.version import StrictVersion
 

	
 
import beaker
 
import mercurial.config
 
import mercurial.ui
 
from beaker.cache import _cache_decorate
 
from tg.i18n import ugettext as _
 

	
 
import kallithea.config.conf
 
from kallithea.lib.exceptions import HgsubversionImportError
 
from kallithea.lib.utils2 import ascii_bytes, aslist, get_current_authuser, safe_bytes, safe_str, safe_unicode
 
from kallithea.lib.vcs.backends.git.repository import GitRepository
 
from kallithea.lib.vcs.backends.hg.repository import MercurialRepository
 
from kallithea.lib.vcs.conf import settings
 
from kallithea.lib.vcs.exceptions import RepositoryError, VCSError
 
from kallithea.lib.vcs.utils.fakemod import create_module
 
from kallithea.lib.vcs.utils.helpers import get_scm
 
from kallithea.lib.vcs.utils.hgcompat import config, ui
 
from kallithea.model import meta
 
from kallithea.model.db import RepoGroup, Repository, Setting, Ui, User, UserGroup, UserLog
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 
REMOVED_REPO_PAT = re.compile(r'rm__\d{8}_\d{6}_\d{6}_.*')
 

	
 

	
 
#==============================================================================
 
# PERM DECORATOR HELPERS FOR EXTRACTING NAMES FOR PERM CHECKS
 
#==============================================================================
 
def get_repo_slug(request):
 
    _repo = request.environ['pylons.routes_dict'].get('repo_name')
 
    if _repo:
 
        _repo = _repo.rstrip('/')
 
    return _repo
 

	
 

	
 
def get_repo_group_slug(request):
 
    _group = request.environ['pylons.routes_dict'].get('group_name')
 
    if _group:
 
        _group = _group.rstrip('/')
 
    return _group
 

	
 

	
 
def get_user_group_slug(request):
 
    _group = request.environ['pylons.routes_dict'].get('id')
 
    _group = UserGroup.get(_group)
 
    if _group:
 
        return _group.users_group_name
 
    return None
 

	
 

	
 
def _get_permanent_id(s):
 
    """Helper for decoding stable URLs with repo ID. For a string like '_123'
 
    return 123.
 
    """
 
    by_id_match = re.match(r'^_(\d+)$', s)
 
    if by_id_match is None:
 
        return None
 
    return int(by_id_match.group(1))
 

	
 

	
 
def fix_repo_id_name(path):
 
    """
 
    Rewrite repo_name for _<ID> permanent URLs.
 

	
 
    Given a path, if the first path element is like _<ID>, return the path with
 
    this part expanded to the corresponding full repo name, else return the
 
    provided path.
 
    """
 
    first, rest = path, ''
 
    if '/' in path:
 
        first, rest_ = path.split('/', 1)
 
        rest = '/' + rest_
 
    repo_id = _get_permanent_id(first)
 
    if repo_id is not None:
 
        repo = Repository.get(repo_id)
 
        if repo is not None:
 
            return repo.repo_name + rest
 
    return path
 

	
 

	
 
def action_logger(user, action, repo, ipaddr='', commit=False):
 
    """
 
    Action logger for various actions made by users
 

	
 
    :param user: user that made this action, can be a unique username string or
 
        object containing user_id attribute
 
    :param action: action to log, should be on of predefined unique actions for
 
        easy translations
 
    :param repo: string name of repository or object containing repo_id,
 
        that action was made on
 
    :param ipaddr: optional IP address from what the action was made
 

	
 
    """
 

	
 
    # if we don't get explicit IP address try to get one from registered user
 
    # in tmpl context var
 
    if not ipaddr:
 
        ipaddr = getattr(get_current_authuser(), 'ip_addr', '')
 

	
 
    if getattr(user, 'user_id', None):
 
        user_obj = User.get(user.user_id)
 
    elif isinstance(user, basestring):
 
        user_obj = User.get_by_username(user)
 
    else:
 
        raise Exception('You have to provide a user object or a username')
 

	
 
    if getattr(repo, 'repo_id', None):
 
        repo_obj = Repository.get(repo.repo_id)
 
        repo_name = repo_obj.repo_name
 
    elif isinstance(repo, basestring):
 
        repo_name = repo.lstrip('/')
 
        repo_obj = Repository.get_by_repo_name(repo_name)
 
    else:
 
        repo_obj = None
 
        repo_name = u''
 

	
 
    user_log = UserLog()
 
    user_log.user_id = user_obj.user_id
 
    user_log.username = user_obj.username
 
    user_log.action = safe_unicode(action)
 

	
 
    user_log.repository = repo_obj
 
    user_log.repository_name = repo_name
 

	
 
    user_log.action_date = datetime.datetime.now()
 
    user_log.user_ip = ipaddr
 
    meta.Session().add(user_log)
 

	
 
    log.info('Logging action:%s on %s by user:%s ip:%s',
 
             action, safe_unicode(repo), user_obj, ipaddr)
 
    if commit:
 
        meta.Session().commit()
 

	
 

	
 
def get_filesystem_repos(path):
 
    """
 
    Scans given path for repos and return (name,(type,path)) tuple
 

	
 
    :param path: path to scan for repositories
 
    :param recursive: recursive search and return names with subdirs in front
 
    """
 

	
 
    # remove ending slash for better results
 
    path = safe_str(path.rstrip(os.sep))
 
    log.debug('now scanning in %s', path)
 

	
 
    def isdir(*n):
 
        return os.path.isdir(os.path.join(*n))
 

	
 
    for root, dirs, _files in os.walk(path):
 
        recurse_dirs = []
 
        for subdir in dirs:
 
            # skip removed repos
 
            if REMOVED_REPO_PAT.match(subdir):
 
                continue
 

	
 
            # skip .<something> dirs TODO: rly? then we should prevent creating them ...
 
            if subdir.startswith('.'):
 
                continue
 

	
 
            cur_path = os.path.join(root, subdir)
 
            if isdir(cur_path, '.git'):
 
                log.warning('ignoring non-bare Git repo: %s', cur_path)
 
                continue
 

	
 
            if (isdir(cur_path, '.hg') or
 
                isdir(cur_path, '.svn') or
 
                isdir(cur_path, 'objects') and (isdir(cur_path, 'refs') or
 
                                                os.path.isfile(os.path.join(cur_path, 'packed-refs')))):
 

	
 
                if not os.access(cur_path, os.R_OK) or not os.access(cur_path, os.X_OK):
 
                    log.warning('ignoring repo path without access: %s', cur_path)
 
                    continue
 

	
 
                if not os.access(cur_path, os.W_OK):
 
                    log.warning('repo path without write access: %s', cur_path)
 

	
 
                try:
 
                    scm_info = get_scm(cur_path)
 
                    assert cur_path.startswith(path)
 
                    repo_path = cur_path[len(path) + 1:]
 
                    yield repo_path, scm_info
 
                    continue # no recursion
 
                except VCSError:
 
                    # We should perhaps ignore such broken repos, but especially
 
                    # the bare git detection is unreliable so we dive into it
 
                    pass
 

	
 
            recurse_dirs.append(subdir)
 

	
 
        dirs[:] = recurse_dirs
 

	
 

	
 
def is_valid_repo_uri(repo_type, url, ui):
 
    """Check if the url seems like a valid remote repo location - raise an Exception if any problems"""
 
    if repo_type == 'hg':
 
        if url.startswith('http') or url.startswith('ssh'):
 
            # initially check if it's at least the proper URL
 
            # or does it pass basic auth
 
            MercurialRepository._check_url(url, ui)
 
        elif url.startswith('svn+http'):
 
            try:
 
                from hgsubversion.svnrepo import svnremoterepo
 
            except ImportError:
 
                raise HgsubversionImportError(_('Unable to activate hgsubversion support. '
 
                                                'The "hgsubversion" library is missing'))
 
            svnremoterepo(ui, url).svn.uuid
 
        elif url.startswith('git+http'):
 
            raise NotImplementedError()
 
        else:
 
            raise Exception('URI %s not allowed' % (url,))
 

	
 
    elif repo_type == 'git':
 
        if url.startswith('http') or url.startswith('git'):
 
            # initially check if it's at least the proper URL
 
            # or does it pass basic auth
 
            GitRepository._check_url(url)
 
        elif url.startswith('svn+http'):
 
            raise NotImplementedError()
 
        elif url.startswith('hg+http'):
 
            raise NotImplementedError()
 
        else:
 
            raise Exception('URI %s not allowed' % (url))
 

	
 

	
 
def is_valid_repo(repo_name, base_path, scm=None):
 
    """
 
    Returns True if given path is a valid repository False otherwise.
 
    If scm param is given also compare if given scm is the same as expected
 
    from scm parameter
 

	
 
    :param repo_name:
 
    :param base_path:
 
    :param scm:
 

	
 
    :return True: if given path is a valid repository
 
    """
 
    # TODO: paranoid security checks?
 
    full_path = os.path.join(safe_str(base_path), safe_str(repo_name))
 

	
 
    try:
 
        scm_ = get_scm(full_path)
 
        if scm:
 
            return scm_[0] == scm
 
        return True
 
    except VCSError:
 
        return False
 

	
 

	
 
def is_valid_repo_group(repo_group_name, base_path, skip_path_check=False):
 
    """
 
    Returns True if given path is a repository group False otherwise
 

	
 
    :param repo_name:
 
    :param base_path:
 
    """
 
    full_path = os.path.join(safe_str(base_path), safe_str(repo_group_name))
 

	
 
    # check if it's not a repo
 
    if is_valid_repo(repo_group_name, base_path):
 
        return False
 

	
 
    try:
 
        # we need to check bare git repos at higher level
 
        # since we might match branches/hooks/info/objects or possible
 
        # other things inside bare git repo
 
        get_scm(os.path.dirname(full_path))
 
        return False
 
    except VCSError:
 
        pass
 

	
 
    # check if it's a valid path
 
    if skip_path_check or os.path.isdir(full_path):
 
        return True
 

	
 
    return False
 

	
 

	
 
# propagated from mercurial documentation
 
ui_sections = ['alias', 'auth',
 
                'decode/encode', 'defaults',
 
                'diff', 'email',
 
                'extensions', 'format',
 
                'merge-patterns', 'merge-tools',
 
                'hooks', 'http_proxy',
 
                'smtp', 'patch',
 
                'paths', 'profiling',
 
                'server', 'trusted',
 
                'ui', 'web', ]
 

	
 

	
 
def make_ui(repo_path=None):
 
    """
 
    Create an Mercurial 'ui' object based on database Ui settings, possibly
 
    augmenting with content from a hgrc file.
 
    """
 
    baseui = ui.ui()
 
    baseui = mercurial.ui.ui()
 

	
 
    # clean the baseui object
 
    baseui._ocfg = config.config()
 
    baseui._ucfg = config.config()
 
    baseui._tcfg = config.config()
 
    baseui._ocfg = mercurial.config.config()
 
    baseui._ucfg = mercurial.config.config()
 
    baseui._tcfg = mercurial.config.config()
 

	
 
    sa = meta.Session()
 
    for ui_ in sa.query(Ui).all():
 
        if ui_.ui_active:
 
            log.debug('config from db: [%s] %s=%r', ui_.ui_section,
 
                      ui_.ui_key, ui_.ui_value)
 
            baseui.setconfig(ascii_bytes(ui_.ui_section), ascii_bytes(ui_.ui_key),
 
                             b'' if ui_.ui_value is None else safe_bytes(ui_.ui_value))
 

	
 
    # force set push_ssl requirement to False, Kallithea handles that
 
    baseui.setconfig(b'web', b'push_ssl', False)
 
    baseui.setconfig(b'web', b'allow_push', b'*')
 
    # prevent interactive questions for ssh password / passphrase
 
    ssh = baseui.config(b'ui', b'ssh', default=b'ssh')
 
    baseui.setconfig(b'ui', b'ssh', b'%s -oBatchMode=yes -oIdentitiesOnly=yes' % ssh)
 
    # push / pull hooks
 
    baseui.setconfig(b'hooks', b'changegroup.kallithea_log_push_action', b'python:kallithea.lib.hooks.log_push_action')
 
    baseui.setconfig(b'hooks', b'outgoing.kallithea_log_pull_action', b'python:kallithea.lib.hooks.log_pull_action')
 

	
 
    if repo_path is not None:
 
        hgrc_path = os.path.join(repo_path, '.hg', 'hgrc')
 
        if os.path.isfile(hgrc_path):
 
            log.debug('reading hgrc from %s', hgrc_path)
 
            cfg = config.config()
 
            cfg = mercurial.config.config()
 
            cfg.read(hgrc_path)
 
            for section in ui_sections:
 
                for k, v in cfg.items(section):
 
                    log.debug('config from file: [%s] %s=%s', section, k, v)
 
                    baseui.setconfig(ascii_bytes(section), ascii_bytes(k), safe_bytes(v))
 
        else:
 
            log.debug('hgrc file is not present at %s, skipping...', hgrc_path)
 

	
 
    return baseui
 

	
 

	
 
def set_app_settings(config):
 
    """
 
    Updates app config with new settings from database
 

	
 
    :param config:
 
    """
 
    hgsettings = Setting.get_app_settings()
 
    for k, v in hgsettings.items():
 
        config[k] = v
 

	
 

	
 
def set_vcs_config(config):
 
    """
 
    Patch VCS config with some Kallithea specific stuff
 

	
 
    :param config: kallithea.CONFIG
 
    """
 
    settings.BACKENDS = {
 
        'hg': 'kallithea.lib.vcs.backends.hg.MercurialRepository',
 
        'git': 'kallithea.lib.vcs.backends.git.GitRepository',
 
    }
 

	
 
    settings.GIT_EXECUTABLE_PATH = config.get('git_path', 'git')
 
    settings.GIT_REV_FILTER = config.get('git_rev_filter', '--all').strip()
 
    settings.DEFAULT_ENCODINGS = aslist(config.get('default_encoding',
 
                                                        'utf-8'), sep=',')
 

	
 

	
 
def set_indexer_config(config):
 
    """
 
    Update Whoosh index mapping
 

	
 
    :param config: kallithea.CONFIG
 
    """
 
    log.debug('adding extra into INDEX_EXTENSIONS')
 
    kallithea.config.conf.INDEX_EXTENSIONS.extend(re.split(r'\s+', config.get('index.extensions', '')))
 

	
 
    log.debug('adding extra into INDEX_FILENAMES')
 
    kallithea.config.conf.INDEX_FILENAMES.extend(re.split(r'\s+', config.get('index.filenames', '')))
 

	
 

	
 
def map_groups(path):
 
    """
 
    Given a full path to a repository, create all nested groups that this
 
    repo is inside. This function creates parent-child relationships between
 
    groups and creates default perms for all new groups.
 

	
 
    :param paths: full path to repository
 
    """
 
    from kallithea.model.repo_group import RepoGroupModel
 
    sa = meta.Session()
 
    groups = path.split(Repository.url_sep())
 
    parent = None
 
    group = None
 

	
 
    # last element is repo in nested groups structure
 
    groups = groups[:-1]
 
    rgm = RepoGroupModel()
 
    owner = User.get_first_admin()
 
    for lvl, group_name in enumerate(groups):
 
        group_name = u'/'.join(groups[:lvl] + [group_name])
 
        group = RepoGroup.get_by_group_name(group_name)
 
        desc = '%s group' % group_name
 

	
 
        # skip folders that are now removed repos
 
        if REMOVED_REPO_PAT.match(group_name):
 
            break
 

	
 
        if group is None:
 
            log.debug('creating group level: %s group_name: %s',
 
                      lvl, group_name)
 
            group = RepoGroup(group_name, parent)
 
            group.group_description = desc
 
            group.owner = owner
 
            sa.add(group)
 
            rgm._create_default_perms(group)
 
            sa.flush()
 

	
 
        parent = group
 
    return group
 

	
 

	
 
def repo2db_mapper(initial_repo_dict, remove_obsolete=False,
 
                   install_git_hooks=False, user=None, overwrite_git_hooks=False):
 
    """
 
    maps all repos given in initial_repo_dict, non existing repositories
 
    are created, if remove_obsolete is True it also check for db entries
 
    that are not in initial_repo_dict and removes them.
 

	
 
    :param initial_repo_dict: mapping with repositories found by scanning methods
 
    :param remove_obsolete: check for obsolete entries in database
 
    :param install_git_hooks: if this is True, also check and install git hook
 
        for a repo if missing
 
    :param overwrite_git_hooks: if this is True, overwrite any existing git hooks
 
        that may be encountered (even if user-deployed)
 
    """
 
    from kallithea.model.repo import RepoModel
 
    from kallithea.model.scm import ScmModel
 
    sa = meta.Session()
 
    repo_model = RepoModel()
 
    if user is None:
 
        user = User.get_first_admin()
 
    added = []
 

	
 
    # creation defaults
 
    defs = Setting.get_default_repo_settings(strip_prefix=True)
 
    enable_statistics = defs.get('repo_enable_statistics')
 
    enable_downloads = defs.get('repo_enable_downloads')
 
    private = defs.get('repo_private')
 

	
 
    for name, repo in initial_repo_dict.items():
 
        group = map_groups(name)
 
        unicode_name = safe_unicode(name)
 
        db_repo = repo_model.get_by_repo_name(unicode_name)
 
        # found repo that is on filesystem not in Kallithea database
 
        if not db_repo:
 
            log.info('repository %s not found, creating now', name)
 
            added.append(name)
 
            desc = (repo.description
 
                    if repo.description != 'unknown'
 
                    else '%s repository' % name)
 

	
 
            new_repo = repo_model._create_repo(
 
                repo_name=name,
 
                repo_type=repo.alias,
 
                description=desc,
 
                repo_group=getattr(group, 'group_id', None),
 
                owner=user,
 
                enable_downloads=enable_downloads,
 
                enable_statistics=enable_statistics,
 
                private=private,
 
                state=Repository.STATE_CREATED
 
            )
 
            sa.commit()
 
            # we added that repo just now, and make sure it has githook
 
            # installed, and updated server info
 
            if new_repo.repo_type == 'git':
 
                git_repo = new_repo.scm_instance
 
                ScmModel().install_git_hooks(git_repo)
 
                # update repository server-info
 
                log.debug('Running update server info')
 
                git_repo._update_server_info()
 
            new_repo.update_changeset_cache()
 
        elif install_git_hooks:
 
            if db_repo.repo_type == 'git':
 
                ScmModel().install_git_hooks(db_repo.scm_instance, force_create=overwrite_git_hooks)
 

	
 
    removed = []
 
    # remove from database those repositories that are not in the filesystem
 
    unicode_initial_repo_names = set(safe_unicode(name) for name in initial_repo_dict)
 
    for repo in sa.query(Repository).all():
 
        if repo.repo_name not in unicode_initial_repo_names:
 
            if remove_obsolete:
 
                log.debug("Removing non-existing repository found in db `%s`",
 
                          repo.repo_name)
 
                try:
 
                    RepoModel().delete(repo, forks='detach', fs_remove=False)
 
                    sa.commit()
 
                except Exception:
 
                    #don't hold further removals on error
 
                    log.error(traceback.format_exc())
 
                    sa.rollback()
 
            removed.append(repo.repo_name)
 
    return added, removed
 

	
 

	
 
def load_rcextensions(root_path):
 
    path = os.path.join(root_path, 'rcextensions', '__init__.py')
 
    if os.path.isfile(path):
 
        rcext = create_module('rc', path)
 
        EXT = kallithea.EXTENSIONS = rcext
 
        log.debug('Found rcextensions now loading %s...', rcext)
 

	
 
        # Additional mappings that are not present in the pygments lexers
 
        kallithea.config.conf.LANGUAGES_EXTENSIONS_MAP.update(getattr(EXT, 'EXTRA_MAPPINGS', {}))
 

	
 
        # OVERRIDE OUR EXTENSIONS FROM RC-EXTENSIONS (if present)
 

	
 
        if getattr(EXT, 'INDEX_EXTENSIONS', []):
 
            log.debug('settings custom INDEX_EXTENSIONS')
 
            kallithea.config.conf.INDEX_EXTENSIONS = getattr(EXT, 'INDEX_EXTENSIONS', [])
 

	
 
        # ADDITIONAL MAPPINGS
 
        log.debug('adding extra into INDEX_EXTENSIONS')
 
        kallithea.config.conf.INDEX_EXTENSIONS.extend(getattr(EXT, 'EXTRA_INDEX_EXTENSIONS', []))
 

	
 
        # auto check if the module is not missing any data, set to default if is
 
        # this will help autoupdate new feature of rcext module
 
        #from kallithea.config import rcextensions
 
        #for k in dir(rcextensions):
 
        #    if not k.startswith('_') and not hasattr(EXT, k):
 
        #        setattr(EXT, k, getattr(rcextensions, k))
 

	
 

	
 
#==============================================================================
 
# MISC
 
#==============================================================================
 

	
 
git_req_ver = StrictVersion('1.7.4')
 

	
 
def check_git_version():
 
    """
 
    Checks what version of git is installed on the system, and raise a system exit
 
    if it's too old for Kallithea to work properly.
 
    """
 
    if 'git' not in kallithea.BACKENDS:
 
        return None
 

	
 
    if not settings.GIT_EXECUTABLE_PATH:
 
        log.warning('No git executable configured - check "git_path" in the ini file.')
 
        return None
 

	
 
    try:
 
        stdout, stderr = GitRepository._run_git_command(['--version'])
 
    except RepositoryError as e:
 
        # message will already have been logged as error
 
        log.warning('No working git executable found - check "git_path" in the ini file.')
 
        return None
 

	
 
    if stderr:
 
        log.warning('Error/stderr from "%s --version":\n%s', settings.GIT_EXECUTABLE_PATH, stderr)
 

	
 
    if not stdout:
 
        log.warning('No working git executable found - check "git_path" in the ini file.')
 
        return None
 

	
 
    output = stdout.strip()
 
    m = re.search(r"\d+.\d+.\d+", output)
 
    if m:
 
        ver = StrictVersion(m.group(0))
 
        log.debug('Git executable: "%s", version %s (parsed from: "%s")',
 
                  settings.GIT_EXECUTABLE_PATH, ver, output)
 
        if ver < git_req_ver:
 
            log.error('Kallithea detected %s version %s, which is too old '
 
                      'for the system to function properly. '
 
                      'Please upgrade to version %s or later. '
 
                      'If you strictly need Mercurial repositories, you can '
 
                      'clear the "git_path" setting in the ini file.',
 
                      settings.GIT_EXECUTABLE_PATH, ver, git_req_ver)
 
            log.error("Terminating ...")
 
            sys.exit(1)
 
    else:
 
        ver = StrictVersion('0.0.0')
 
        log.warning('Error finding version number in "%s --version" stdout:\n%s',
 
                    settings.GIT_EXECUTABLE_PATH, output)
 

	
 
    return ver
 

	
 

	
 
#===============================================================================
 
# CACHE RELATED METHODS
 
#===============================================================================
 

	
 
# set cache regions for beaker so celery can utilise it
 
def setup_cache_regions(settings):
 
    # Create dict with just beaker cache configs with prefix stripped
 
    cache_settings = {'regions': None}
 
    prefix = 'beaker.cache.'
 
    for key in settings:
 
        if key.startswith(prefix):
 
            name = key[len(prefix):]
 
            cache_settings[name] = settings[key]
 
    # Find all regions, apply defaults, and apply to beaker
 
    if cache_settings['regions']:
 
        for region in cache_settings['regions'].split(','):
 
            region = region.strip()
 
            prefix = region + '.'
 
            region_settings = {}
 
            for key in cache_settings:
 
                if key.startswith(prefix):
 
                    name = key[len(prefix):]
 
                    region_settings[name] = cache_settings[key]
 
            region_settings.setdefault('expire',
 
                                       cache_settings.get('expire', '60'))
 
            region_settings.setdefault('lock_dir',
 
                                       cache_settings.get('lock_dir'))
 
            region_settings.setdefault('data_dir',
 
                                       cache_settings.get('data_dir'))
 
            region_settings.setdefault('type',
 
                                       cache_settings.get('type', 'memory'))
 
            beaker.cache.cache_regions[region] = region_settings
 

	
 

	
 
def conditional_cache(region, prefix, condition, func):
 
    """
 

	
 
    Conditional caching function use like::
 
        def _c(arg):
 
            #heavy computation function
 
            return data
 

	
 
        # depending from condition the compute is wrapped in cache or not
 
        compute = conditional_cache('short_term', 'cache_desc', condition=True, func=func)
 
        return compute(arg)
 

	
 
    :param region: name of cache region
 
    :param prefix: cache region prefix
 
    :param condition: condition for cache to be triggered, and return data cached
 
    :param func: wrapped heavy function to compute
 

	
 
    """
 
    wrapped = func
 
    if condition:
 
        log.debug('conditional_cache: True, wrapping call of '
 
                  'func: %s into %s region cache' % (region, func))
 
        wrapped = _cache_decorate((prefix,), None, None, region)(func)
 

	
 
    return wrapped
kallithea/lib/vcs/backends/git/repository.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
"""
 
    vcs.backends.git.repository
 
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
    Git repository implementation.
 

	
 
    :created_on: Apr 8, 2010
 
    :copyright: (c) 2010-2011 by Marcin Kuzminski, Lukasz Balcerzak.
 
"""
 

	
 
import errno
 
import logging
 
import os
 
import re
 
import time
 
import urllib
 
import urllib2
 
from collections import OrderedDict
 

	
 
import mercurial.url  # import httpbasicauthhandler, httpdigestauthhandler
 
import mercurial.util  # import url as hg_url
 
from dulwich.config import ConfigFile
 
from dulwich.objects import Tag
 
from dulwich.repo import NotGitRepository, Repo
 

	
 
from kallithea.lib.vcs import subprocessio
 
from kallithea.lib.vcs.backends.base import BaseRepository, CollectionGenerator
 
from kallithea.lib.vcs.conf import settings
 
from kallithea.lib.vcs.exceptions import (
 
    BranchDoesNotExistError, ChangesetDoesNotExistError, EmptyRepositoryError, RepositoryError, TagAlreadyExistError, TagDoesNotExistError)
 
from kallithea.lib.vcs.utils import ascii_str, date_fromtimestamp, makedate, safe_str, safe_unicode
 
from kallithea.lib.vcs.utils.hgcompat import hg_url, httpbasicauthhandler, httpdigestauthhandler
 
from kallithea.lib.vcs.utils.lazy import LazyProperty
 
from kallithea.lib.vcs.utils.paths import abspath, get_user_home
 

	
 
from .changeset import GitChangeset
 
from .inmemory import GitInMemoryChangeset
 
from .workdir import GitWorkdir
 

	
 

	
 
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 = 'master'
 
    scm = 'git'
 

	
 
    def __init__(self, repo_path, create=False, src_url=None,
 
                 update_after_clone=False, bare=False):
 

	
 
        self.path = safe_unicode(abspath(repo_path))
 
        self.repo = self._get_repo(create, src_url, update_after_clone, bare)
 
        self.bare = self.repo.bare
 

	
 
    @property
 
    def _config_files(self):
 
        return [
 
            self.bare and abspath(self.path, 'config')
 
                      or abspath(self.path, '.git', 'config'),
 
             abspath(get_user_home(), '.gitconfig'),
 
         ]
 

	
 
    @property
 
    def _repo(self):
 
        return self.repo
 

	
 
    @property
 
    def head(self):
 
        try:
 
            return self._repo.head()
 
        except KeyError:
 
            return None
 

	
 
    @property
 
    def _empty(self):
 
        """
 
        Checks if repository is empty ie. without any changesets
 
        """
 

	
 
        try:
 
            self.revisions[0]
 
        except (KeyError, IndexError):
 
            return True
 
        return False
 

	
 
    @LazyProperty
 
    def revisions(self):
 
        """
 
        Returns list of revisions' ids, in ascending order.  Being lazy
 
        attribute allows external tools to inject shas from cache.
 
        """
 
        return self._get_all_revisions()
 

	
 
    @classmethod
 
    def _run_git_command(cls, cmd, cwd=None):
 
        """
 
        Runs given ``cmd`` as git command and returns output bytes in a tuple
 
        (stdout, stderr) ... or raise RepositoryError.
 

	
 
        :param cmd: git command to be executed
 
        :param cwd: passed directly to subprocess
 
        """
 
        # need to clean fix GIT_DIR !
 
        gitenv = dict(os.environ)
 
        gitenv.pop('GIT_DIR', None)
 
        gitenv['GIT_CONFIG_NOGLOBAL'] = '1'
 

	
 
        assert isinstance(cmd, list), cmd
 
        cmd = [settings.GIT_EXECUTABLE_PATH, '-c', 'core.quotepath=false'] + cmd
 
        try:
 
            p = subprocessio.SubprocessIOChunker(cmd, cwd=cwd, env=gitenv, shell=False)
 
        except (EnvironmentError, OSError) as err:
 
            # output from the failing process is in str(EnvironmentError)
 
            msg = ("Couldn't run git command %s.\n"
 
                   "Subprocess failed with '%s': %s\n" %
 
                   (cmd, type(err).__name__, err)
 
            ).strip()
 
            log.error(msg)
 
            raise RepositoryError(msg)
 

	
 
        try:
 
            stdout = b''.join(p.output)
 
            stderr = b''.join(p.error)
 
        finally:
 
            p.close()
 
        # TODO: introduce option to make commands fail if they have any stderr output?
 
        if stderr:
 
            log.debug('stderr from %s:\n%s', cmd, stderr)
 
        else:
 
            log.debug('stderr from %s: None', cmd)
 
        return stdout, stderr
 

	
 
    def run_git_command(self, cmd):
 
        """
 
        Runs given ``cmd`` as git command with cwd set to current repo.
 
        Returns stdout as unicode str ... or raise RepositoryError.
 
        """
 
        cwd = None
 
        if os.path.isdir(self.path):
 
            cwd = self.path
 
        stdout, _stderr = self._run_git_command(cmd, cwd=cwd)
 
        return safe_unicode(stdout)
 

	
 
    @classmethod
 
    def _check_url(cls, url):
 
        """
 
        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 local url
 
        if os.path.isdir(url) or url.startswith('file:'):
 
            return True
 

	
 
        if url.startswith('git://'):
 
            return True
 

	
 
        if '+' in url[:url.find('://')]:
 
            url = url[url.find('+') + 1:]
 

	
 
        handlers = []
 
        url_obj = hg_url(url)
 
        url_obj = mercurial.util.url(url)
 
        test_uri, authinfo = url_obj.authinfo()
 
        if not test_uri.endswith('info/refs'):
 
            test_uri = test_uri.rstrip('/') + '/info/refs'
 

	
 
        url_obj.passwd = b'*****'
 
        cleaned_uri = str(url_obj)
 

	
 
        if authinfo:
 
            # create a password manager
 
            passmgr = urllib2.HTTPPasswordMgrWithDefaultRealm()
 
            passmgr.add_password(*authinfo)
 

	
 
            handlers.extend((httpbasicauthhandler(passmgr),
 
                             httpdigestauthhandler(passmgr)))
 
            handlers.extend((mercurial.url.httpbasicauthhandler(passmgr),
 
                             mercurial.url.httpdigestauthhandler(passmgr)))
 

	
 
        o = urllib2.build_opener(*handlers)
 
        o.addheaders = [('User-Agent', 'git/1.7.8.0')]  # fake some git
 

	
 
        q = {"service": 'git-upload-pack'}
 
        qs = '?%s' % urllib.urlencode(q)
 
        cu = "%s%s" % (test_uri, qs)
 
        req = urllib2.Request(cu, None, {})
 

	
 
        try:
 
            resp = o.open(req)
 
            if resp.code != 200:
 
                raise Exception('Return Code is not 200')
 
        except Exception as e:
 
            # means it cannot be cloned
 
            raise urllib2.URLError("[%s] org_exc: %s" % (cleaned_uri, e))
 

	
 
        # now detect if it's proper git repo
 
        gitdata = resp.read()
 
        if 'service=git-upload-pack' not in gitdata:
 
            raise urllib2.URLError(
 
                "url [%s] does not look like an git" % cleaned_uri)
 

	
 
        return True
 

	
 
    def _get_repo(self, create, src_url=None, update_after_clone=False,
 
                  bare=False):
 
        if create and os.path.exists(self.path):
 
            raise RepositoryError("Location already exist")
 
        if src_url and not create:
 
            raise RepositoryError("Create should be set to True if src_url is "
 
                                  "given (clone operation creates repository)")
 
        try:
 
            if create and src_url:
 
                GitRepository._check_url(src_url)
 
                self.clone(src_url, update_after_clone, bare)
 
                return Repo(self.path)
 
            elif create:
 
                os.makedirs(self.path)
 
                if bare:
 
                    return Repo.init_bare(self.path)
 
                else:
 
                    return Repo.init(self.path)
 
            else:
 
                return Repo(self.path)
 
        except (NotGitRepository, OSError) as err:
 
            raise RepositoryError(err)
 

	
 
    def _get_all_revisions(self):
 
        # 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
 
        try:
 
            self._repo.head()
 
        except KeyError:
 
            return []
 

	
 
        rev_filter = settings.GIT_REV_FILTER
 
        cmd = ['rev-list', rev_filter, '--reverse', '--date-order']
 
        try:
 
            so = self.run_git_command(cmd)
 
        except RepositoryError:
 
            # Can be raised for empty repositories
 
            return []
 
        return so.splitlines()
 

	
 
    def _get_all_revisions2(self):
 
        # alternate implementation using dulwich
 
        includes = [ascii_str(x[1][0]) for x in self._parsed_refs.iteritems()
 
                    if x[1][1] != b'T']
 
        return [c.commit.id for c in self._repo.get_walker(include=includes)]
 

	
 
    def _get_revision(self, revision):
 
        """
 
        Given any revision identifier, returns a 40 char string with revision hash.
 
        """
 
        if self._empty:
 
            raise EmptyRepositoryError("There are no changesets yet")
 

	
 
        if revision in (None, '', 'tip', 'HEAD', 'head', -1):
 
            revision = -1
 

	
 
        if isinstance(revision, int):
 
            try:
 
                return self.revisions[revision]
 
            except IndexError:
 
                msg = ("Revision %s does not exist for %s" % (revision, self))
 
                raise ChangesetDoesNotExistError(msg)
 

	
 
        if isinstance(revision, (str, unicode)):
 
            if revision.isdigit() and (len(revision) < 12 or len(revision) == revision.count('0')):
 
                try:
 
                    return self.revisions[int(revision)]
 
                except IndexError:
 
                    msg = "Revision %r does not exist for %s" % (revision, self)
 
                    raise ChangesetDoesNotExistError(msg)
 

	
 
            # get by branch/tag name
 
            _ref_revision = self._parsed_refs.get(revision)
 
            if _ref_revision:  # and _ref_revision[1] in [b'H', b'RH', b'T']:
 
                return ascii_str(_ref_revision[0])
 

	
 
            if revision in self.revisions:
 
                return revision
 

	
 
            # maybe it's a tag ? we don't have them in self.revisions
 
            if revision in self.tags.values():
 
                return revision
 

	
 
            if SHA_PATTERN.match(revision):
 
                msg = ("Revision %s does not exist for %s" % (revision, self))
 
                raise ChangesetDoesNotExistError(msg)
 

	
 
        raise ChangesetDoesNotExistError("Given revision %r not recognized" % revision)
 

	
 
    def get_ref_revision(self, ref_type, ref_name):
 
        """
 
        Returns ``GitChangeset`` object representing repository's
 
        changeset at the given ``revision``.
 
        """
 
        return self._get_revision(ref_name)
 

	
 
    def _get_archives(self, archive_name='tip'):
 

	
 
        for i in [('zip', '.zip'), ('gz', '.tar.gz'), ('bz2', '.tar.bz2')]:
 
            yield {"type": i[0], "extension": i[1], "node": archive_name}
 

	
 
    def _get_url(self, url):
 
        """
 
        Returns normalized url. If schema is not given, would fall to
 
        filesystem (``file:///``) schema.
 
        """
 
        url = safe_str(url)
 
        if url != 'default' and '://' not in url:
 
            url = ':///'.join(('file', url))
 
        return url
 

	
 
    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 name(self):
 
        return os.path.basename(self.path)
 

	
 
    @LazyProperty
 
    def last_change(self):
 
        """
 
        Returns last change made on this repository as datetime object
 
        """
 
        return date_fromtimestamp(self._get_mtime(), makedate()[1])
 

	
 
    def _get_mtime(self):
 
        try:
 
            return time.mktime(self.get_changeset().date.timetuple())
 
        except RepositoryError:
 
            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):
 
        return safe_unicode(self._repo.get_description() or b'unknown')
 

	
 
    @LazyProperty
 
    def contact(self):
 
        undefined_contact = u'Unknown'
 
        return undefined_contact
 

	
 
    @property
 
    def branches(self):
 
        if not self.revisions:
 
            return {}
 
        sortkey = lambda ctx: ctx[0]
 
        _branches = [(x[0], ascii_str(x[1][0]))
 
                     for x in self._parsed_refs.iteritems() if x[1][1] == b'H']
 
        return OrderedDict(sorted(_branches, key=sortkey, reverse=False))
 

	
 
    @LazyProperty
 
    def closed_branches(self):
 
        return {}
 

	
 
    @LazyProperty
 
    def tags(self):
 
        return self._get_tags()
 

	
 
    def _get_tags(self):
 
        if not self.revisions:
 
            return {}
 

	
 
        sortkey = lambda ctx: ctx[0]
 
        _tags = [(x[0], ascii_str(x[1][0]))
 
                 for x in self._parsed_refs.iteritems() if x[1][1] == b'T']
 
        return OrderedDict(sorted(_tags, key=sortkey, reverse=True))
 

	
 
    def tag(self, name, user, revision=None, message=None, date=None,
 
            **kwargs):
 
        """
 
        Creates and returns a tag for the given ``revision``.
 

	
 
        :param name: name for new tag
 
        :param user: full username, i.e.: "Joe Doe <joe.doe@example.com>"
 
        :param revision: changeset 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)
 
        changeset = self.get_changeset(revision)
 
        message = message or "Added tag %s for commit %s" % (name,
 
            changeset.raw_id)
 
        self._repo.refs[b"refs/tags/%s" % name] = changeset._commit.id
 

	
 
        self._parsed_refs = self._get_parsed_refs()
 
        self.tags = self._get_tags()
 
        return changeset
 

	
 
    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 <joe.doe@example.com>"
 
        :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._repo.refs is a DiskRefsContainer, and .path gives the full absolute path of '.git'
 
        tagpath = os.path.join(self._repo.refs.path, 'refs', 'tags', name)
 
        try:
 
            os.remove(tagpath)
 
            self._parsed_refs = self._get_parsed_refs()
 
            self.tags = self._get_tags()
 
        except OSError as e:
 
            raise RepositoryError(e.strerror)
 

	
 
    @LazyProperty
 
    def bookmarks(self):
 
        """
 
        Gets bookmarks for this repository
 
        """
 
        return {}
 

	
 
    @LazyProperty
 
    def _parsed_refs(self):
 
        return self._get_parsed_refs()
 

	
 
    def _get_parsed_refs(self):
 
        # cache the property
 
        _repo = self._repo
 
        refs = _repo.get_refs()
 
        keys = [(b'refs/heads/', b'H'),
 
                (b'refs/remotes/origin/', b'RH'),
 
                (b'refs/tags/', b'T')]
 
        _refs = {}
 
        for ref, sha in refs.iteritems():
 
            for k, type_ in keys:
 
                if ref.startswith(k):
 
                    _key = ref[len(k):]
 
                    if type_ == b'T':
 
                        obj = _repo.get_object(sha)
 
                        if isinstance(obj, Tag):
 
                            sha = _repo.get_object(sha).object[1]
 
                    _refs[_key] = [sha, type_]
 
                    break
 
        return _refs
 

	
 
    def _heads(self, reverse=False):
 
        refs = self._repo.get_refs()
 
        heads = {}
 

	
 
        for key, val in refs.items():
 
            for ref_key in [b'refs/heads/', b'refs/remotes/origin/']:
 
                if key.startswith(ref_key):
 
                    n = key[len(ref_key):]
 
                    if n not in [b'HEAD']:
 
                        heads[n] = val
 

	
 
        return heads if reverse else dict((y, x) for x, y in heads.iteritems())
 

	
 
    def get_changeset(self, revision=None):
 
        """
 
        Returns ``GitChangeset`` object representing commit from git repository
 
        at the given revision or head (most recent commit) if None given.
 
        """
 
        if isinstance(revision, GitChangeset):
 
            return revision
 
        revision = self._get_revision(revision)
 
        changeset = GitChangeset(repository=self, revision=revision)
 
        return changeset
 

	
 
    def get_changesets(self, start=None, end=None, start_date=None,
 
           end_date=None, branch_name=None, reverse=False, max_revisions=None):
 
        """
 
        Returns iterator of ``GitChangeset`` objects from start to end (both
 
        are inclusive), in ascending date order (unless ``reverse`` is set).
 

	
 
        :param start: changeset ID, as str; first returned changeset
 
        :param end: changeset ID, as str; last returned changeset
 
        :param start_date: if specified, changesets with commit date less than
 
          ``start_date`` would be filtered out from returned set
 
        :param end_date: if specified, changesets with commit date greater than
 
          ``end_date`` would be filtered out from returned set
 
        :param branch_name: if specified, changesets not reachable from given
 
          branch would be filtered out from returned set
 
        :param reverse: if ``True``, returned generator would be reversed
 
          (meaning that returned changesets would have descending date order)
 

	
 
        :raise BranchDoesNotExistError: If given ``branch_name`` does not
 
            exist.
 
        :raise ChangesetDoesNotExistError: If changeset for given ``start`` or
 
          ``end`` could not be found.
 

	
 
        """
 
        if branch_name and branch_name not in self.branches:
 
            raise BranchDoesNotExistError("Branch '%s' not found"
 
                                          % branch_name)
 
        # actually we should check now if it's not an empty repo to not spaw
 
        # subprocess commands
 
        if self._empty:
 
            raise EmptyRepositoryError("There are no changesets yet")
 

	
 
        # %H at format means (full) commit hash, initial hashes are retrieved
 
        # in ascending date order
 
        cmd = ['log', '--date-order', '--reverse', '--pretty=format:%H']
 
        if max_revisions:
 
            cmd += ['--max-count=%s' % max_revisions]
 
        if start_date:
 
            cmd += ['--since', start_date.strftime('%m/%d/%y %H:%M:%S')]
 
        if end_date:
 
            cmd += ['--until', end_date.strftime('%m/%d/%y %H:%M:%S')]
 
        if branch_name:
 
            cmd.append(branch_name)
 
        else:
 
            cmd.append(settings.GIT_REV_FILTER)
 

	
 
        revs = self.run_git_command(cmd).splitlines()
 
        start_pos = 0
 
        end_pos = len(revs)
 
        if start:
 
            _start = self._get_revision(start)
 
            try:
 
                start_pos = revs.index(_start)
 
            except ValueError:
 
                pass
 

	
 
        if end is not None:
 
            _end = self._get_revision(end)
 
            try:
 
                end_pos = revs.index(_end)
 
            except ValueError:
 
                pass
 

	
 
        if None not in [start, end] and start_pos > end_pos:
 
            raise RepositoryError('start cannot be after end')
 

	
 
        if end_pos is not None:
 
            end_pos += 1
 

	
 
        revs = revs[start_pos:end_pos]
 
        if reverse:
 
            revs.reverse()
 

	
 
        return CollectionGenerator(self, revs)
 

	
 
    def get_diff(self, rev1, rev2, path=None, ignore_whitespace=False,
 
                 context=3):
 
        """
 
        Returns (git like) *diff*, as plain bytes text. Shows changes
 
        introduced by ``rev2`` since ``rev1``.
 

	
 
        :param rev1: Entry point from which diff is shown. Can be
 
          ``self.EMPTY_CHANGESET`` - in this case, patch showing all
 
          the changes since empty state of the repository until ``rev2``
 
        :param rev2: Until which revision 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``. Due to limitations in Git, if
 
          value passed-in is greater than ``2**31-1``
 
          (``2147483647``), it will be set to ``2147483647``
 
          instead. If negative value is passed-in, it will be set to
 
          ``0`` instead.
 
        """
 

	
 
        # Git internally uses a signed long int for storing context
 
        # size (number of lines to show before and after the
 
        # differences). This can result in integer overflow, so we
 
        # ensure the requested context is smaller by one than the
 
        # number that would cause the overflow. It is highly unlikely
 
        # that a single file will contain that many lines, so this
 
        # kind of change should not cause any realistic consequences.
 
        overflowed_long_int = 2**31
 

	
 
        if context >= overflowed_long_int:
 
            context = overflowed_long_int - 1
 

	
 
        # Negative context values make no sense, and will result in
 
        # errors. Ensure this does not happen.
 
        if context < 0:
 
            context = 0
 

	
 
        flags = ['-U%s' % context, '--full-index', '--binary', '-p', '-M', '--abbrev=40']
 
        if ignore_whitespace:
 
            flags.append('-w')
 

	
 
        if hasattr(rev1, 'raw_id'):
 
            rev1 = getattr(rev1, 'raw_id')
 

	
 
        if hasattr(rev2, 'raw_id'):
 
            rev2 = getattr(rev2, 'raw_id')
 

	
 
        if rev1 == self.EMPTY_CHANGESET:
 
            rev2 = self.get_changeset(rev2).raw_id
 
            cmd = ['show'] + flags + [rev2]
 
        else:
 
            rev1 = self.get_changeset(rev1).raw_id
 
            rev2 = self.get_changeset(rev2).raw_id
 
            cmd = ['diff'] + flags + [rev1, rev2]
 

	
 
        if path:
 
            cmd += ['--', path]
 

	
 
        stdout, stderr = self._run_git_command(cmd, cwd=self.path)
 
        # If we used 'show' command, strip first few lines (until actual diff
 
        # starts)
 
        if rev1 == self.EMPTY_CHANGESET:
 
            parts = stdout.split(b'\ndiff ', 1)
 
            if len(parts) > 1:
 
                stdout = b'diff ' + parts[1]
 
        return stdout
 

	
 
    @LazyProperty
 
    def in_memory_changeset(self):
 
        """
 
        Returns ``GitInMemoryChangeset`` object for this repository.
 
        """
 
        return GitInMemoryChangeset(self)
 

	
 
    def clone(self, url, update_after_clone=True, bare=False):
 
        """
 
        Tries to clone changes from external location.
 

	
 
        :param update_after_clone: If set to ``False``, git won't checkout
 
          working directory
 
        :param bare: If set to ``True``, repository would be cloned into
 
          *bare* git repository (no working directory at all).
 
        """
 
        url = self._get_url(url)
 
        cmd = ['clone', '-q']
 
        if bare:
 
            cmd.append('--bare')
 
        elif not update_after_clone:
 
            cmd.append('--no-checkout')
 
        cmd += ['--', url, self.path]
 
        # If error occurs run_git_command raises RepositoryError already
 
        self.run_git_command(cmd)
 

	
 
    def pull(self, url):
 
        """
 
        Tries to pull changes from external location.
 
        """
 
        url = self._get_url(url)
 
        cmd = ['pull', '--ff-only', url]
 
        # If error occurs run_git_command raises RepositoryError already
 
        self.run_git_command(cmd)
 

	
 
    def fetch(self, url):
 
        """
 
        Tries to pull changes from external location.
 
        """
 
        url = self._get_url(url)
 
        so = self.run_git_command(['ls-remote', '-h', url])
 
        cmd = ['fetch', url, '--']
 
        for line in (x for x in so.splitlines()):
 
            sha, ref = line.split('\t')
 
            cmd.append('+%s:%s' % (ref, ref))
 
        self.run_git_command(cmd)
 

	
 
    def _update_server_info(self):
 
        """
 
        runs gits update-server-info command in this repo instance
 
        """
 
        from dulwich.server import update_server_info
 
        try:
 
            update_server_info(self._repo)
 
        except OSError as e:
 
            if e.errno not in [errno.ENOENT, errno.EROFS]:
 
                raise
 
            # Workaround for dulwich crashing on for example its own dulwich/tests/data/repos/simple_merge.git/info/refs.lock
 
            log.error('Ignoring %s running update-server-info: %s', type(e).__name__, e)
 

	
 
    @LazyProperty
 
    def workdir(self):
 
        """
 
        Returns ``Workdir`` instance for this repository.
 
        """
 
        return GitWorkdir(self)
 

	
 
    def get_config_value(self, section, name, config_file=None):
 
        """
 
        Returns configuration value for a given [``section``] and ``name``.
 

	
 
        :param section: Section we want to retrieve value from
 
        :param name: Name of configuration we want to retrieve
 
        :param config_file: A path to file which should be used to retrieve
 
          configuration from (might also be a list of file paths)
 
        """
 
        if config_file is None:
 
            config_file = []
 
        elif isinstance(config_file, basestring):
 
            config_file = [config_file]
 

	
 
        def gen_configs():
 
            for path in config_file + self._config_files:
 
                try:
 
                    yield ConfigFile.from_path(path)
 
                except (IOError, OSError, ValueError):
 
                    continue
 

	
 
        for config in gen_configs():
 
            try:
 
                return config.get(section, name)
 
            except KeyError:
 
                continue
 
        return None
 

	
 
    def get_user_name(self, config_file=None):
 
        """
 
        Returns user's name from global configuration file.
 

	
 
        :param config_file: A path to file which should be used to retrieve
 
          configuration from (might also be a list of file paths)
 
        """
 
        return self.get_config_value('user', 'name', config_file)
 

	
 
    def get_user_email(self, config_file=None):
 
        """
 
        Returns user's email from global configuration file.
 

	
 
        :param config_file: A path to file which should be used to retrieve
 
          configuration from (might also be a list of file paths)
 
        """
 
        return self.get_config_value('user', 'email', config_file)
kallithea/lib/vcs/backends/hg/__init__.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
"""
 
    vcs.backends.hg
 
    ~~~~~~~~~~~~~~~
 

	
 
    Mercurial backend implementation.
 

	
 
    :created_on: Apr 8, 2010
 
    :copyright: (c) 2010-2011 by Marcin Kuzminski, Lukasz Balcerzak.
 
"""
 

	
 
from kallithea.lib.vcs.utils import hgcompat
 

	
 
from .changeset import MercurialChangeset
 
from .inmemory import MercurialInMemoryChangeset
 
from .repository import MercurialRepository
 
from .workdir import MercurialWorkdir
 

	
 

	
 
__all__ = [
 
    'MercurialRepository', 'MercurialChangeset',
 
    'MercurialInMemoryChangeset', 'MercurialWorkdir',
 
]
 

	
 
hgcompat.monkey_do()
kallithea/lib/vcs/backends/hg/changeset.py
Show inline comments
 
import os
 
import posixpath
 

	
 
import mercurial.archival
 
import mercurial.node
 
import mercurial.obsutil
 

	
 
from kallithea.lib.vcs.backends.base import BaseChangeset
 
from kallithea.lib.vcs.conf import settings
 
from kallithea.lib.vcs.exceptions import ChangesetDoesNotExistError, ChangesetError, ImproperArchiveTypeError, NodeDoesNotExistError, VCSError
 
from kallithea.lib.vcs.nodes import (
 
    AddedFileNodesGenerator, ChangedFileNodesGenerator, DirNode, FileNode, NodeKind, RemovedFileNodesGenerator, RootNode, SubModuleNode)
 
from kallithea.lib.vcs.utils import ascii_bytes, ascii_str, date_fromtimestamp, safe_str, safe_unicode
 
from kallithea.lib.vcs.utils.hgcompat import archival, hex, obsutil
 
from kallithea.lib.vcs.utils.lazy import LazyProperty
 
from kallithea.lib.vcs.utils.paths import get_dirs_for_path
 

	
 

	
 
class MercurialChangeset(BaseChangeset):
 
    """
 
    Represents state of the repository at a revision.
 
    """
 

	
 
    def __init__(self, repository, revision):
 
        self.repository = repository
 
        assert isinstance(revision, basestring), repr(revision)
 
        self._ctx = repository._repo[ascii_bytes(revision)]
 
        self.raw_id = ascii_str(self._ctx.hex())
 
        self.revision = self._ctx._rev
 
        self.nodes = {}
 

	
 
    @LazyProperty
 
    def tags(self):
 
        return [safe_unicode(tag) for tag in self._ctx.tags()]
 

	
 
    @LazyProperty
 
    def branch(self):
 
        return safe_unicode(self._ctx.branch())
 

	
 
    @LazyProperty
 
    def branches(self):
 
        return [safe_unicode(self._ctx.branch())]
 

	
 
    @LazyProperty
 
    def closesbranch(self):
 
        return self._ctx.closesbranch()
 

	
 
    @LazyProperty
 
    def obsolete(self):
 
        return self._ctx.obsolete()
 

	
 
    @LazyProperty
 
    def bumped(self):
 
        return self._ctx.phasedivergent()
 

	
 
    @LazyProperty
 
    def divergent(self):
 
        return self._ctx.contentdivergent()
 

	
 
    @LazyProperty
 
    def extinct(self):
 
        return self._ctx.extinct()
 

	
 
    @LazyProperty
 
    def unstable(self):
 
        return self._ctx.orphan()
 

	
 
    @LazyProperty
 
    def phase(self):
 
        if(self._ctx.phase() == 1):
 
            return 'Draft'
 
        elif(self._ctx.phase() == 2):
 
            return 'Secret'
 
        else:
 
            return ''
 

	
 
    @LazyProperty
 
    def successors(self):
 
        successors = obsutil.successorssets(self._ctx._repo, self._ctx.node(), closest=True)
 
        successors = mercurial.obsutil.successorssets(self._ctx._repo, self._ctx.node(), closest=True)
 
        if successors:
 
            # flatten the list here handles both divergent (len > 1)
 
            # and the usual case (len = 1)
 
            successors = [hex(n)[:12] for sub in successors for n in sub if n != self._ctx.node()]
 
            successors = [mercurial.node.hex(n)[:12] for sub in successors for n in sub if n != self._ctx.node()]
 

	
 
        return successors
 

	
 
    @LazyProperty
 
    def predecessors(self):
 
        return [hex(n)[:12] for n in obsutil.closestpredecessors(self._ctx._repo, self._ctx.node())]
 
        return [mercurial.node.hex(n)[:12] for n in mercurial.obsutil.closestpredecessors(self._ctx._repo, self._ctx.node())]
 

	
 
    @LazyProperty
 
    def bookmarks(self):
 
        return [safe_unicode(bookmark) for bookmark in self._ctx.bookmarks()]
 

	
 
    @LazyProperty
 
    def message(self):
 
        return safe_unicode(self._ctx.description())
 

	
 
    @LazyProperty
 
    def committer(self):
 
        return safe_unicode(self.author)
 

	
 
    @LazyProperty
 
    def author(self):
 
        return safe_unicode(self._ctx.user())
 

	
 
    @LazyProperty
 
    def date(self):
 
        return date_fromtimestamp(*self._ctx.date())
 

	
 
    @LazyProperty
 
    def _timestamp(self):
 
        return self._ctx.date()[0]
 

	
 
    @LazyProperty
 
    def status(self):
 
        """
 
        Returns modified, added, removed, deleted files for current changeset
 
        """
 
        return self.repository._repo.status(self._ctx.p1().node(),
 
                                            self._ctx.node())
 

	
 
    @LazyProperty
 
    def _file_paths(self):
 
        return list(self._ctx)
 

	
 
    @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]
 

	
 
    @LazyProperty
 
    def parents(self):
 
        """
 
        Returns list of parents changesets.
 
        """
 
        return [self.repository.get_changeset(parent.rev())
 
                for parent in self._ctx.parents() if parent.rev() >= 0]
 

	
 
    @LazyProperty
 
    def children(self):
 
        """
 
        Returns list of children changesets.
 
        """
 
        return [self.repository.get_changeset(child.rev())
 
                for child in self._ctx.children() if child.rev() >= 0]
 

	
 
    def next(self, branch=None):
 
        if branch and self.branch != branch:
 
            raise VCSError('Branch option used on changeset not belonging '
 
                           'to that branch')
 

	
 
        cs = self
 
        while True:
 
            try:
 
                next_ = cs.repository.revisions.index(cs.raw_id) + 1
 
                next_rev = cs.repository.revisions[next_]
 
            except IndexError:
 
                raise ChangesetDoesNotExistError
 
            cs = cs.repository.get_changeset(next_rev)
 

	
 
            if not branch or branch == cs.branch:
 
                return cs
 

	
 
    def prev(self, branch=None):
 
        if branch and self.branch != branch:
 
            raise VCSError('Branch option used on changeset not belonging '
 
                           'to that branch')
 

	
 
        cs = self
 
        while True:
 
            try:
 
                prev_ = cs.repository.revisions.index(cs.raw_id) - 1
 
                if prev_ < 0:
 
                    raise IndexError
 
                prev_rev = cs.repository.revisions[prev_]
 
            except IndexError:
 
                raise ChangesetDoesNotExistError
 
            cs = cs.repository.get_changeset(prev_rev)
 

	
 
            if not branch or branch == cs.branch:
 
                return cs
 

	
 
    def diff(self):
 
        # Only used to feed diffstat
 
        return b''.join(self._ctx.diff())
 

	
 
    def _fix_path(self, path):
 
        """
 
        Paths are stored without trailing slash so we need to get rid off it if
 
        needed. Also mercurial keeps filenodes as str so we need to decode
 
        from unicode to str
 
        """
 
        if path.endswith('/'):
 
            path = path.rstrip('/')
 

	
 
        return safe_str(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 ChangesetError("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 ChangesetError("File does not exist for revision %s at "
 
                " '%s'" % (self.raw_id, path))
 
        return self._ctx.filectx(path)
 

	
 
    def _extract_submodules(self):
 
        """
 
        returns a dictionary with submodule information from substate file
 
        of hg repository
 
        """
 
        return self._ctx.substate
 

	
 
    def get_file_mode(self, path):
 
        """
 
        Returns stat mode of the file at the given ``path``.
 
        """
 
        fctx = self._get_filectx(path)
 
        if b'x' in fctx.flags():
 
            return 0o100755
 
        else:
 
            return 0o100644
 

	
 
    def get_file_content(self, path):
 
        """
 
        Returns content of the file at given ``path``.
 
        """
 
        fctx = self._get_filectx(path)
 
        return fctx.data()
 

	
 
    def get_file_size(self, path):
 
        """
 
        Returns size of the file at given ``path``.
 
        """
 
        fctx = self._get_filectx(path)
 
        return fctx.size()
 

	
 
    def get_file_changeset(self, path):
 
        """
 
        Returns last commit of the file at the given ``path``.
 
        """
 
        return self.get_file_history(path, limit=1)[0]
 

	
 
    def get_file_history(self, path, limit=None):
 
        """
 
        Returns history of file as reversed list of ``Changeset`` objects for
 
        which file at given ``path`` has been modified.
 
        """
 
        fctx = self._get_filectx(path)
 
        hist = []
 
        cnt = 0
 
        for cs in reversed([x for x in fctx.filelog()]):
 
            cnt += 1
 
            hist.append(hex(fctx.filectx(cs).node()))
 
            hist.append(mercurial.node.hex(fctx.filectx(cs).node()))
 
            if limit is not None and cnt == limit:
 
                break
 

	
 
        return [self.repository.get_changeset(node) for node in hist]
 

	
 
    def get_file_annotate(self, path):
 
        """
 
        Returns a generator of four element tuples with
 
            lineno, sha, changeset lazy loader and line
 
        """
 
        annotations = self._get_filectx(path).annotate()
 
        annotation_lines = [(annotateline.fctx, annotateline.text) for annotateline in annotations]
 
        for i, (fctx, l) in enumerate(annotation_lines):
 
            sha = ascii_str(fctx.hex())
 
            yield (i + 1, sha, lambda sha=sha, l=l: self.repository.get_changeset(sha), l)
 

	
 
    def fill_archive(self, stream=None, kind='tgz', prefix=None,
 
                     subrepos=False):
 
        """
 
        Fills up given stream.
 

	
 
        :param stream: file like object.
 
        :param kind: one of following: ``zip``, ``tgz`` or ``tbz2``.
 
            Default: ``tgz``.
 
        :param prefix: name of root directory in archive.
 
            Default is repository name and changeset's raw_id joined with dash
 
            (``repo-tip.<KIND>``).
 
        :param subrepos: include subrepos in this archive.
 

	
 
        :raise ImproperArchiveTypeError: If given kind is wrong.
 
        :raise VcsError: If given stream is None
 
        """
 
        allowed_kinds = settings.ARCHIVE_SPECS
 
        if kind not in allowed_kinds:
 
            raise ImproperArchiveTypeError('Archive kind not supported use one'
 
                'of %s' % ' '.join(allowed_kinds))
 

	
 
        if stream is None:
 
            raise VCSError('You need to pass in a valid stream for filling'
 
                           ' with archival data')
 

	
 
        if prefix is None:
 
            prefix = '%s-%s' % (self.repository.name, self.short_id)
 
        elif prefix.startswith('/'):
 
            raise VCSError("Prefix cannot start with leading slash")
 
        elif prefix.strip() == '':
 
            raise VCSError("Prefix cannot be empty")
 

	
 
        archival.archive(self.repository._repo, stream, ascii_bytes(self.raw_id),
 
        mercurial.archival.archive(self.repository._repo, stream, ascii_bytes(self.raw_id),
 
                         kind, prefix=prefix, subrepos=subrepos)
 

	
 
    def get_nodes(self, path):
 
        """
 
        Returns combined ``DirNode`` and ``FileNode`` objects list representing
 
        state of changeset at the given ``path``. If node at the given ``path``
 
        is not instance of ``DirNode``, ChangesetError would be raised.
 
        """
 

	
 
        if self._get_kind(path) != NodeKind.DIR:
 
            raise ChangesetError("Directory does not exist for revision %s at "
 
                " '%s'" % (self.revision, path))
 
        path = self._fix_path(path)
 

	
 
        filenodes = [FileNode(f, changeset=self) for f in self._file_paths
 
            if os.path.dirname(f) == path]
 
        dirs = path == '' and '' or [d for d in self._dir_paths
 
            if d and posixpath.dirname(d) == path]
 
        dirnodes = [DirNode(d, changeset=self) for d in dirs
 
            if os.path.dirname(d) == path]
 

	
 
        als = self.repository.alias
 
        for k, vals in self._extract_submodules().iteritems():
 
            #vals = url,rev,type
 
            loc = vals[0]
 
            cs = vals[1]
 
            dirnodes.append(SubModuleNode(k, url=loc, changeset=cs,
 
                                          alias=als))
 
        nodes = dirnodes + filenodes
 
        for node in nodes:
 
            self.nodes[node.path] = node
 
        nodes.sort()
 
        return nodes
 

	
 
    def get_node(self, path):
 
        """
 
        Returns ``Node`` object from the given ``path``. If there is no node at
 
        the given ``path``, ``ChangesetError`` would be raised.
 
        """
 
        path = self._fix_path(path)
 
        if path not in self.nodes:
 
            if path in self._file_paths:
 
                node = FileNode(path, changeset=self)
 
            elif path in self._dir_paths or path in self._dir_paths:
 
                if path == '':
 
                    node = RootNode(changeset=self)
 
                else:
 
                    node = DirNode(path, changeset=self)
 
            else:
 
                raise NodeDoesNotExistError("There is no file nor directory "
 
                    "at the given path: '%s' at revision %s"
 
                    % (path, self.short_id))
 
            # cache node
 
            self.nodes[path] = node
 
        return self.nodes[path]
 

	
 
    @LazyProperty
 
    def affected_files(self):
 
        """
 
        Gets a fast accessible file changes for given changeset
 
        """
 
        return self._ctx.files()
 

	
 
    @property
 
    def added(self):
 
        """
 
        Returns list of added ``FileNode`` objects.
 
        """
 
        return AddedFileNodesGenerator([n for n in self.status[1]], self)
 

	
 
    @property
 
    def changed(self):
 
        """
 
        Returns list of modified ``FileNode`` objects.
 
        """
 
        return ChangedFileNodesGenerator([n for n in self.status[0]], self)
 

	
 
    @property
 
    def removed(self):
 
        """
 
        Returns list of removed ``FileNode`` objects.
 
        """
 
        return RemovedFileNodesGenerator([n for n in self.status[2]], self)
 

	
 
    @LazyProperty
 
    def extra(self):
 
        return self._ctx.extra()
kallithea/lib/vcs/backends/hg/inmemory.py
Show inline comments
 
import datetime
 

	
 
import mercurial.context
 
import mercurial.node
 

	
 
from kallithea.lib.vcs.backends.base import BaseInMemoryChangeset
 
from kallithea.lib.vcs.exceptions import RepositoryError
 
from kallithea.lib.vcs.utils import ascii_str, safe_bytes
 
from kallithea.lib.vcs.utils.hgcompat import hex, memctx, memfilectx
 

	
 

	
 
class MercurialInMemoryChangeset(BaseInMemoryChangeset):
 

	
 
    def commit(self, message, author, parents=None, branch=None, date=None,
 
            **kwargs):
 
        """
 
        Performs in-memory commit (doesn't check workdir in any way) and
 
        returns newly created ``Changeset``. Updates repository's
 
        ``revisions``.
 

	
 
        :param message: message of the commit
 
        :param author: full username, i.e. "Joe Doe <joe.doe@example.com>"
 
        :param parents: single parent or sequence of parents from which commit
 
          would be derived
 
        :param date: ``datetime.datetime`` instance. Defaults to
 
          ``datetime.datetime.now()``.
 
        :param branch: branch name, as string. If none given, default backend's
 
          branch would be used.
 

	
 
        :raises ``CommitError``: if any error occurs while committing
 
        """
 
        self.check_integrity(parents)
 

	
 
        from .repository import MercurialRepository
 
        if not isinstance(message, unicode) or not isinstance(author, unicode):
 
            raise RepositoryError('Given message and author needs to be '
 
                                  'an <unicode> instance got %r & %r instead'
 
                                  % (type(message), type(author)))
 

	
 
        if branch is None:
 
            branch = MercurialRepository.DEFAULT_BRANCH_NAME
 
        kwargs[b'branch'] = branch
 

	
 
        def filectxfn(_repo, memctx, path):
 
            """
 
            Marks given path as added/changed/removed in a given _repo. This is
 
            for internal mercurial commit function.
 
            """
 

	
 
            # check if this path is removed
 
            if path in (node.path for node in self.removed):
 
                return None
 

	
 
            # check if this path is added
 
            for node in self.added:
 
                if node.path == path:
 
                    return memfilectx(_repo, memctx, path=node.path,
 
                    return mercurial.context.memfilectx(_repo, memctx, path=node.path,
 
                        data=node.content,
 
                        islink=False,
 
                        isexec=node.is_executable,
 
                        copysource=False)
 

	
 
            # or changed
 
            for node in self.changed:
 
                if node.path == path:
 
                    return memfilectx(_repo, memctx, path=node.path,
 
                    return mercurial.context.memfilectx(_repo, memctx, path=node.path,
 
                        data=node.content,
 
                        islink=False,
 
                        isexec=node.is_executable,
 
                        copysource=False)
 

	
 
            raise RepositoryError("Given path haven't been marked as added,"
 
                                  "changed or removed (%s)" % path)
 

	
 
        parents = [None, None]
 
        for i, parent in enumerate(self.parents):
 
            if parent is not None:
 
                parents[i] = parent._ctx.node()
 

	
 
        if date and isinstance(date, datetime.datetime):
 
            date = date.strftime('%a, %d %b %Y %H:%M:%S')
 

	
 
        commit_ctx = memctx(repo=self.repository._repo,
 
        commit_ctx = mercurial.context.memctx(
 
            repo=self.repository._repo,
 
            parents=parents,
 
            text=b'',
 
            files=self.get_paths(),
 
            filectxfn=filectxfn,
 
            user=author,
 
            date=date,
 
            extra=kwargs)
 

	
 
        # injecting given _repo params
 
        commit_ctx._text = safe_bytes(message)
 
        commit_ctx._user = safe_bytes(author)
 
        commit_ctx._date = date
 

	
 
        # TODO: Catch exceptions!
 
        n = self.repository._repo.commitctx(commit_ctx)
 
        # Returns mercurial node
 
        self._commit_ctx = commit_ctx  # For reference
 
        # Update vcs repository object & recreate mercurial _repo
 
        # new_ctx = self.repository._repo[node]
 
        # new_tip = ascii_str(self.repository.get_changeset(new_ctx.hex()))
 
        self.repository.revisions.append(ascii_str(hex(n)))
 
        self.repository.revisions.append(ascii_str(mercurial.node.hex(n)))
 
        self._repo = self.repository._get_repo(create=False)
 
        self.repository.branches = self.repository._get_branches()
 
        tip = self.repository.get_changeset()
 
        self.reset()
 
        return tip
kallithea/lib/vcs/backends/hg/repository.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
"""
 
    vcs.backends.hg.repository
 
    ~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
    Mercurial repository implementation.
 

	
 
    :created_on: Apr 8, 2010
 
    :copyright: (c) 2010-2011 by Marcin Kuzminski, Lukasz Balcerzak.
 
"""
 

	
 
import datetime
 
import logging
 
import os
 
import time
 
import urllib
 
import urllib2
 
from collections import OrderedDict
 

	
 
import mercurial.commands
 
import mercurial.error
 
import mercurial.exchange
 
import mercurial.hg
 
import mercurial.hgweb
 
import mercurial.httppeer
 
import mercurial.match
 
import mercurial.mdiff
 
import mercurial.node
 
import mercurial.patch
 
import mercurial.scmutil
 
import mercurial.sshpeer
 
import mercurial.tags
 
import mercurial.ui
 
import mercurial.url
 
import mercurial.util
 

	
 
from kallithea.lib.vcs.backends.base import BaseRepository, CollectionGenerator
 
from kallithea.lib.vcs.exceptions import (
 
    BranchDoesNotExistError, ChangesetDoesNotExistError, EmptyRepositoryError, RepositoryError, TagAlreadyExistError, TagDoesNotExistError, VCSError)
 
from kallithea.lib.vcs.utils import ascii_str, author_email, author_name, date_fromtimestamp, makedate, safe_bytes, safe_str, safe_unicode
 
from kallithea.lib.vcs.utils.hgcompat import (
 
    Abort, RepoError, RepoLookupError, clone, diffopts, get_contact, hex, hg_url, httpbasicauthhandler, httpdigestauthhandler, httppeer, localrepo, match_exact, nullid, patch, peer, scmutil, sshpeer, tag, ui)
 
from kallithea.lib.vcs.utils.lazy import LazyProperty
 
from kallithea.lib.vcs.utils.paths import abspath
 

	
 
from .changeset import MercurialChangeset
 
from .inmemory import MercurialInMemoryChangeset
 
from .workdir import MercurialWorkdir
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
class MercurialRepository(BaseRepository):
 
    """
 
    Mercurial repository backend
 
    """
 
    DEFAULT_BRANCH_NAME = 'default'
 
    scm = 'hg'
 

	
 
    def __init__(self, repo_path, create=False, baseui=None, src_url=None,
 
                 update_after_clone=False):
 
        """
 
        Raises RepositoryError if repository could not be find at the given
 
        ``repo_path``.
 

	
 
        :param repo_path: local path of the repository
 
        :param create=False: if set to True, would try to create repository if
 
           it does not exist rather than raising exception
 
        :param baseui=None: user data
 
        :param src_url=None: would try to clone repository from given location
 
        :param update_after_clone=False: sets update of working copy after
 
          making a clone
 
        """
 

	
 
        if not isinstance(repo_path, str):
 
            raise VCSError('Mercurial backend requires repository path to '
 
                           'be instance of <str> got %s instead' %
 
                           type(repo_path))
 

	
 
        self.path = abspath(repo_path)
 
        self.baseui = baseui or ui.ui()
 
        self.baseui = baseui or mercurial.ui.ui()
 
        # We've set path and ui, now we can set _repo itself
 
        self._repo = self._get_repo(create, src_url, update_after_clone)
 

	
 
    @property
 
    def _empty(self):
 
        """
 
        Checks if repository is empty ie. without any changesets
 
        """
 
        # TODO: Following raises errors when using InMemoryChangeset...
 
        # return len(self._repo.changelog) == 0
 
        return len(self.revisions) == 0
 

	
 
    @LazyProperty
 
    def revisions(self):
 
        """
 
        Returns list of revisions' ids, in ascending order.  Being lazy
 
        attribute allows external tools to inject shas from cache.
 
        """
 
        return self._get_all_revisions()
 

	
 
    @LazyProperty
 
    def name(self):
 
        return os.path.basename(self.path)
 

	
 
    @LazyProperty
 
    def branches(self):
 
        return self._get_branches()
 

	
 
    @LazyProperty
 
    def closed_branches(self):
 
        return self._get_branches(normal=False, closed=True)
 

	
 
    @LazyProperty
 
    def allbranches(self):
 
        """
 
        List all branches, including closed branches.
 
        """
 
        return self._get_branches(closed=True)
 

	
 
    def _get_branches(self, normal=True, closed=False):
 
        """
 
        Gets branches for this repository
 
        Returns only not closed branches by default
 

	
 
        :param closed: return also closed branches for mercurial
 
        :param normal: return also normal branches
 
        """
 

	
 
        if self._empty:
 
            return {}
 

	
 
        bt = OrderedDict()
 
        for bn, _heads, node, isclosed in sorted(self._repo.branchmap().iterbranches()):
 
            if isclosed:
 
                if closed:
 
                    bt[safe_unicode(bn)] = ascii_str(hex(node))
 
                    bt[safe_unicode(bn)] = ascii_str(mercurial.node.hex(node))
 
            else:
 
                if normal:
 
                    bt[safe_unicode(bn)] = ascii_str(hex(node))
 
                    bt[safe_unicode(bn)] = ascii_str(mercurial.node.hex(node))
 
        return bt
 

	
 
    @LazyProperty
 
    def tags(self):
 
        """
 
        Gets tags for this repository
 
        """
 
        return self._get_tags()
 

	
 
    def _get_tags(self):
 
        if self._empty:
 
            return {}
 

	
 
        return OrderedDict(sorted(
 
            ((safe_unicode(n), ascii_str(hex(h))) for n, h in self._repo.tags().items()),
 
            ((safe_unicode(n), ascii_str(mercurial.node.hex(h))) for n, h in self._repo.tags().items()),
 
            reverse=True,
 
            key=lambda x: x[0],  # sort by name
 
        ))
 

	
 
    def tag(self, name, user, revision=None, message=None, date=None,
 
            **kwargs):
 
        """
 
        Creates and returns a tag for the given ``revision``.
 

	
 
        :param name: name for new tag
 
        :param user: full username, i.e.: "Joe Doe <joe.doe@example.com>"
 
        :param revision: changeset 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)
 
        changeset = self.get_changeset(revision)
 
        local = kwargs.setdefault('local', False)
 

	
 
        if message is None:
 
            message = "Added tag %s for changeset %s" % (name,
 
                changeset.short_id)
 

	
 
        if date is None:
 
            date = datetime.datetime.now().strftime('%a, %d %b %Y %H:%M:%S')
 

	
 
        try:
 
            tag(self._repo, name, changeset._ctx.node(), message, local, user, date)
 
        except Abort as e:
 
            mercurial.tags.tag(self._repo, name, changeset._ctx.node(), message, local, user, date)
 
        except mercurial.error.Abort as e:
 
            raise RepositoryError(e.message)
 

	
 
        # Reinitialize tags
 
        self.tags = self._get_tags()
 
        tag_id = self.tags[name]
 

	
 
        return self.get_changeset(revision=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 <joe.doe@example.com>"
 
        :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
 
        if date is None:
 
            date = datetime.datetime.now().strftime('%a, %d %b %Y %H:%M:%S')
 
        local = False
 

	
 
        try:
 
            tag(self._repo, name, nullid, message, local, user, date)
 
            mercurial.tags.tag(self._repo, name, mercurial.commands.nullid, message, local, user, date)
 
            self.tags = self._get_tags()
 
        except Abort as e:
 
        except mercurial.error.Abort as e:
 
            raise RepositoryError(e.message)
 

	
 
    @LazyProperty
 
    def bookmarks(self):
 
        """
 
        Gets bookmarks for this repository
 
        """
 
        return self._get_bookmarks()
 

	
 
    def _get_bookmarks(self):
 
        if self._empty:
 
            return {}
 

	
 
        return OrderedDict(sorted(
 
            ((safe_unicode(n), ascii_str(h)) for n, h in self._repo._bookmarks.items()),
 
            reverse=True,
 
            key=lambda x: x[0],  # sort by name
 
        ))
 

	
 
    def _get_all_revisions(self):
 
        return [ascii_str(self._repo[x].hex()) for x in self._repo.filtered(b'visible').changelog.revs()]
 

	
 
    def get_diff(self, rev1, rev2, path='', ignore_whitespace=False,
 
                  context=3):
 
        """
 
        Returns (git like) *diff*, as plain text. Shows changes introduced by
 
        ``rev2`` since ``rev1``.
 

	
 
        :param rev1: Entry point from which diff is shown. Can be
 
          ``self.EMPTY_CHANGESET`` - in this case, patch showing all
 
          the changes since empty state of the repository until ``rev2``
 
        :param rev2: Until which revision 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``. If negative value is passed-in, it will be
 
          set to ``0`` instead.
 
        """
 

	
 
        # Negative context values make no sense, and will result in
 
        # errors. Ensure this does not happen.
 
        if context < 0:
 
            context = 0
 

	
 
        if hasattr(rev1, 'raw_id'):
 
            rev1 = getattr(rev1, 'raw_id')
 

	
 
        if hasattr(rev2, 'raw_id'):
 
            rev2 = getattr(rev2, 'raw_id')
 

	
 
        # Check if given revisions are present at repository (may raise
 
        # ChangesetDoesNotExistError)
 
        if rev1 != self.EMPTY_CHANGESET:
 
            self.get_changeset(rev1)
 
        self.get_changeset(rev2)
 
        if path:
 
            file_filter = match_exact(path)
 
            file_filter = mercurial.match.exact(path)
 
        else:
 
            file_filter = None
 

	
 
        return b''.join(patch.diff(self._repo, rev1, rev2, match=file_filter,
 
                          opts=diffopts(git=True,
 
        return b''.join(mercurial.patch.diff(self._repo, rev1, rev2, match=file_filter,
 
                          opts=mercurial.mdiff.diffopts(git=True,
 
                                        showfunc=True,
 
                                        ignorews=ignore_whitespace,
 
                                        context=context)))
 

	
 
    @classmethod
 
    def _check_url(cls, url, repoui=None):
 
        """
 
        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(b'file:'):
 
            return True
 

	
 
        if url.startswith(b'ssh:'):
 
            # in case of invalid uri or authentication issues, sshpeer will
 
            # throw an exception.
 
            sshpeer.instance(repoui or ui.ui(), url, False).lookup(b'tip')
 
            mercurial.sshpeer.instance(repoui or mercurial.ui.ui(), url, False).lookup(b'tip')
 
            return True
 

	
 
        url_prefix = None
 
        if b'+' in url[:url.find(b'://')]:
 
            url_prefix, url = url.split(b'+', 1)
 

	
 
        handlers = []
 
        url_obj = hg_url(url)
 
        url_obj = mercurial.util.url(url)
 
        test_uri, authinfo = url_obj.authinfo()
 
        url_obj.passwd = b'*****'
 
        cleaned_uri = str(url_obj)
 

	
 
        if authinfo:
 
            # create a password manager
 
            passmgr = urllib2.HTTPPasswordMgrWithDefaultRealm()
 
            passmgr.add_password(*authinfo)
 

	
 
            handlers.extend((httpbasicauthhandler(passmgr),
 
                             httpdigestauthhandler(passmgr)))
 
            handlers.extend((mercurial.url.httpbasicauthhandler(passmgr),
 
                             mercurial.url.httpdigestauthhandler(passmgr)))
 

	
 
        o = urllib2.build_opener(*handlers)
 
        o.addheaders = [('Content-Type', 'application/mercurial-0.1'),
 
                        ('Accept', 'application/mercurial-0.1')]
 

	
 
        q = {"cmd": 'between'}
 
        q.update({'pairs': "%s-%s" % ('0' * 40, '0' * 40)})
 
        qs = '?%s' % urllib.urlencode(q)
 
        cu = "%s%s" % (test_uri, qs)
 
        req = urllib2.Request(cu, None, {})
 

	
 
        try:
 
            resp = o.open(req)
 
            if resp.code != 200:
 
                raise Exception('Return Code is not 200')
 
        except Exception as e:
 
            # means it cannot be cloned
 
            raise urllib2.URLError("[%s] org_exc: %s" % (cleaned_uri, e))
 

	
 
        if not url_prefix: # skip svn+http://... (and git+... too)
 
            # now check if it's a proper hg repo
 
            try:
 
                httppeer.instance(repoui or ui.ui(), url, False).lookup(b'tip')
 
                mercurial.httppeer.instance(repoui or mercurial.ui.ui(), url, False).lookup(b'tip')
 
            except Exception as e:
 
                raise urllib2.URLError(
 
                    "url [%s] does not look like an hg repo org_exc: %s"
 
                    % (cleaned_uri, e))
 

	
 
        return True
 

	
 
    def _get_repo(self, create, src_url=None, update_after_clone=False):
 
        """
 
        Function will check for mercurial repository in given path and return
 
        a localrepo object. 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 and returned.
 
        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 ``update_after_clone`` flag
 
        """
 

	
 
        try:
 
            if src_url:
 
                url = safe_bytes(self._get_url(src_url))
 
                opts = {}
 
                if not update_after_clone:
 
                    opts.update({'noupdate': True})
 
                MercurialRepository._check_url(url, self.baseui)
 
                clone(self.baseui, url, self.path, **opts)
 
                mercurial.commands.clone(self.baseui, url, self.path, **opts)
 

	
 
                # Don't try to create if we've already cloned repo
 
                create = False
 
            return localrepo.instance(self.baseui, self.path, create=create)
 
        except (Abort, RepoError) as err:
 
            return mercurial.localrepo.instance(self.baseui, self.path, create=create)
 
        except (mercurial.error.Abort, mercurial.error.RepoError) as err:
 
            if create:
 
                msg = "Cannot create repository at %s. Original error was %s" \
 
                    % (self.path, err)
 
            else:
 
                msg = "Not valid repository at %s. Original error was %s" \
 
                    % (self.path, err)
 
            raise RepositoryError(msg)
 

	
 
    @LazyProperty
 
    def in_memory_changeset(self):
 
        return MercurialInMemoryChangeset(self)
 

	
 
    @LazyProperty
 
    def description(self):
 
        _desc = self._repo.ui.config(b'web', b'description', None, untrusted=True)
 
        return safe_unicode(_desc or b'unknown')
 

	
 
    @LazyProperty
 
    def contact(self):
 
        return safe_unicode(get_contact(self._repo.ui.config)
 
        return safe_unicode(mercurial.hgweb.common.get_contact(self._repo.ui.config)
 
                            or b'Unknown')
 

	
 
    @LazyProperty
 
    def last_change(self):
 
        """
 
        Returns last change made on this repository as datetime object
 
        """
 
        return date_fromtimestamp(self._get_mtime(), makedate()[1])
 

	
 
    def _get_mtime(self):
 
        try:
 
            return time.mktime(self.get_changeset().date.timetuple())
 
        except RepositoryError:
 
            # 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_revision(self, revision):
 
        """
 
        Given any revision identifier, returns a 40 char string with revision hash.
 

	
 
        :param revision: str or int or None
 
        """
 
        if self._empty:
 
            raise EmptyRepositoryError("There are no changesets yet")
 

	
 
        if revision in [-1, None]:
 
            revision = b'tip'
 
        elif isinstance(revision, unicode):
 
            revision = safe_bytes(revision)
 

	
 
        try:
 
            if isinstance(revision, int):
 
                return ascii_str(self._repo[revision].hex())
 
            return ascii_str(scmutil.revsymbol(self._repo, revision).hex())
 
        except (IndexError, ValueError, RepoLookupError, TypeError):
 
            return ascii_str(mercurial.scmutil.revsymbol(self._repo, revision).hex())
 
        except (IndexError, ValueError, mercurial.error.RepoLookupError, TypeError):
 
            msg = ("Revision %s does not exist for %s" % (revision, self))
 
            raise ChangesetDoesNotExistError(msg)
 
        except (LookupError, ):
 
            msg = ("Ambiguous identifier `%s` for %s" % (revision, self))
 
            raise ChangesetDoesNotExistError(msg)
 

	
 
    def get_ref_revision(self, ref_type, ref_name):
 
        """
 
        Returns revision number for the given reference.
 
        """
 
        ref_name = safe_str(ref_name)
 
        if ref_type == 'rev' and not ref_name.strip('0'):
 
            return self.EMPTY_CHANGESET
 
        # lookup up the exact node id
 
        _revset_predicates = {
 
                'branch': 'branch',
 
                'book': 'bookmark',
 
                'tag': 'tag',
 
                'rev': 'id',
 
            }
 
        # avoid expensive branch(x) iteration over whole repo
 
        rev_spec = "%%s & %s(%%s)" % _revset_predicates[ref_type]
 
        try:
 
            revs = self._repo.revs(rev_spec, ref_name, ref_name)
 
        except LookupError:
 
            msg = ("Ambiguous identifier %s:%s for %s" % (ref_type, ref_name, self.name))
 
            raise ChangesetDoesNotExistError(msg)
 
        except RepoLookupError:
 
        except mercurial.error.RepoLookupError:
 
            msg = ("Revision %s:%s does not exist for %s" % (ref_type, ref_name, self.name))
 
            raise ChangesetDoesNotExistError(msg)
 
        if revs:
 
            revision = revs.last()
 
        else:
 
            # TODO: just report 'not found'?
 
            revision = ref_name
 

	
 
        return self._get_revision(revision)
 

	
 
    def _get_archives(self, archive_name='tip'):
 
        allowed = self.baseui.configlist(b"web", b"allow_archive",
 
                                         untrusted=True)
 
        for name, ext in [(b'zip', '.zip'), (b'gz', '.tar.gz'), (b'bz2', '.tar.bz2')]:
 
            if name in allowed or self._repo.ui.configbool(b"web",
 
                                                           b"allow" + name,
 
                                                           untrusted=True):
 
                yield {"type": name, "extension": ext, "node": archive_name}
 

	
 
    def _get_url(self, url):
 
        """
 
        Returns normalized url. If schema is not given, would fall
 
        to filesystem
 
        (``file:///``) schema.
 
        """
 
        url = safe_str(url)
 
        if url != 'default' and '://' not in url:
 
            url = "file:" + urllib.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_changeset(self, revision=None):
 
        """
 
        Returns ``MercurialChangeset`` object representing repository's
 
        changeset at the given ``revision``.
 
        """
 
        revision = self._get_revision(revision)
 
        changeset = MercurialChangeset(repository=self, revision=revision)
 
        return changeset
 

	
 
    def get_changesets(self, start=None, end=None, start_date=None,
 
                       end_date=None, branch_name=None, reverse=False, max_revisions=None):
 
        """
 
        Returns iterator of ``MercurialChangeset`` objects from start to end
 
        (both are inclusive)
 

	
 
        :param start: None, str, int or mercurial lookup format
 
        :param end:  None, str, int or mercurial lookup format
 
        :param start_date:
 
        :param end_date:
 
        :param branch_name:
 
        :param reversed: return changesets in reversed order
 
        """
 
        start_raw_id = self._get_revision(start)
 
        start_pos = None if start is None else self.revisions.index(start_raw_id)
 
        end_raw_id = self._get_revision(end)
 
        end_pos = None if end is None else self.revisions.index(end_raw_id)
 

	
 
        if start_pos is not None and end_pos is not None and start_pos > end_pos:
 
            raise RepositoryError("Start revision '%s' cannot be "
 
                                  "after end revision '%s'" % (start, end))
 

	
 
        if branch_name and branch_name not in self.allbranches:
 
            msg = ("Branch %s not found in %s" % (branch_name, self))
 
            raise BranchDoesNotExistError(msg)
 
        if end_pos is not None:
 
            end_pos += 1
 
        # filter branches
 
        filter_ = []
 
        if branch_name:
 
            filter_.append(b'branch("%s")' % safe_bytes(branch_name))
 
        if start_date:
 
            filter_.append(b'date(">%s")' % start_date)
 
        if end_date:
 
            filter_.append(b'date("<%s")' % end_date)
 
        if filter_ or max_revisions:
 
            if filter_:
 
                revspec = b' and '.join(filter_)
 
            else:
 
                revspec = b'all()'
 
            if max_revisions:
 
                revspec = b'limit(%s, %d)' % (revspec, max_revisions)
 
            revisions = scmutil.revrange(self._repo, [revspec])
 
            revisions = mercurial.scmutil.revrange(self._repo, [revspec])
 
        else:
 
            revisions = self.revisions
 

	
 
        # this is very much a hack to turn this into a list; a better solution
 
        # would be to get rid of this function entirely and use revsets
 
        revs = list(revisions)[start_pos:end_pos]
 
        if reverse:
 
            revs.reverse()
 

	
 
        return CollectionGenerator(self, revs)
 

	
 
    def pull(self, url):
 
        """
 
        Tries to pull changes from external location.
 
        """
 
        url = self._get_url(url)
 
        other = peer(self._repo, {}, url)
 
        other = mercurial.hg.peer(self._repo, {}, url)
 
        try:
 
            from mercurial import exchange
 
            exchange.pull(self._repo, other, heads=None, force=None)
 
        except Abort as err:
 
            mercurial.exchange.pull(self._repo, other, heads=None, force=None)
 
        except mercurial.error.Abort as err:
 
            # Propagate error but with vcs's type
 
            raise RepositoryError(str(err))
 

	
 
    @LazyProperty
 
    def workdir(self):
 
        """
 
        Returns ``Workdir`` instance for this repository.
 
        """
 
        return MercurialWorkdir(self)
 

	
 
    def get_config_value(self, section, name=None, config_file=None):
 
        """
 
        Returns configuration value for a given [``section``] and ``name``.
 

	
 
        :param section: Section we want to retrieve value from
 
        :param name: Name of configuration we want to retrieve
 
        :param config_file: A path to file which should be used to retrieve
 
          configuration from (might also be a list of file paths)
 
        """
 
        if config_file is None:
 
            config_file = []
 
        elif isinstance(config_file, basestring):
 
            config_file = [config_file]
 

	
 
        config = self._repo.ui
 
        if config_file:
 
            config = ui.ui()
 
            config = mercurial.ui.ui()
 
            for path in config_file:
 
                config.readconfig(path)
 
        return config.config(section, name)
 

	
 
    def get_user_name(self, config_file=None):
 
        """
 
        Returns user's name from global configuration file.
 

	
 
        :param config_file: A path to file which should be used to retrieve
 
          configuration from (might also be a list of file paths)
 
        """
 
        username = self.get_config_value('ui', 'username', config_file=config_file)
 
        if username:
 
            return author_name(username)
 
        return None
 

	
 
    def get_user_email(self, config_file=None):
 
        """
 
        Returns user's email from global configuration file.
 

	
 
        :param config_file: A path to file which should be used to retrieve
 
          configuration from (might also be a list of file paths)
 
        """
 
        username = self.get_config_value('ui', 'username', config_file=config_file)
 
        if username:
 
            return author_email(username)
 
        return None
kallithea/lib/vcs/backends/hg/ssh.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
# This program is free software: you can redistribute it and/or modify
 
# it under the terms of the GNU General Public License as published by
 
# the Free Software Foundation, either version 3 of the License, or
 
# (at your option) any later version.
 
#
 
# 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 General Public License
 
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
 

	
 
import logging
 

	
 
from mercurial import hg
 
from mercurial.wireprotoserver import sshserver
 
import mercurial.hg
 
import mercurial.wireprotoserver
 

	
 
from kallithea.lib.utils import make_ui
 
from kallithea.lib.vcs.backends.ssh import BaseSshHandler
 
from kallithea.lib.vcs.utils import safe_bytes, safe_unicode
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
class MercurialSshHandler(BaseSshHandler):
 
    vcs_type = 'hg'
 

	
 
    @classmethod
 
    def make(cls, ssh_command_parts):
 
        r"""
 
        >>> import shlex
 

	
 
        >>> MercurialSshHandler.make(shlex.split('hg -R "foo bar" serve --stdio')).repo_name
 
        u'foo bar'
 
        >>> MercurialSshHandler.make(shlex.split(' hg -R blåbærgrød serve --stdio ')).repo_name
 
        u'bl\xe5b\xe6rgr\xf8d'
 
        >>> MercurialSshHandler.make(shlex.split('''hg -R 'foo"bar' serve --stdio''')).repo_name
 
        u'foo"bar'
 

	
 
        >>> MercurialSshHandler.make(shlex.split('/bin/hg -R "foo" serve --stdio'))
 
        >>> MercurialSshHandler.make(shlex.split('''hg -R "foo"bar" serve --stdio''')) # ssh-serve will report: Error parsing SSH command "...": invalid syntax
 
        Traceback (most recent call last):
 
        ValueError: No closing quotation
 
        >>> MercurialSshHandler.make(shlex.split('git-upload-pack "/foo"')) # not handled here
 
        """
 
        if ssh_command_parts[:2] == ['hg', '-R'] and ssh_command_parts[3:] == ['serve', '--stdio']:
 
            return cls(safe_unicode(ssh_command_parts[2]))
 

	
 
        return None
 

	
 
    def __init__(self, repo_name):
 
        self.repo_name = repo_name
 

	
 
    def _serve(self):
 
        # Note: we want a repo with config based on .hg/hgrc and can thus not use self.db_repo.scm_instance._repo.ui
 
        baseui = make_ui(repo_path=self.db_repo.repo_full_path)
 
        if not self.allow_push:
 
            baseui.setconfig(b'hooks', b'pretxnopen._ssh_reject', b'python:kallithea.lib.hooks.rejectpush')
 
            baseui.setconfig(b'hooks', b'prepushkey._ssh_reject', b'python:kallithea.lib.hooks.rejectpush')
 

	
 
        repo = hg.repository(baseui, safe_bytes(self.db_repo.repo_full_path))
 
        repo = mercurial.hg.repository(baseui, safe_bytes(self.db_repo.repo_full_path))
 
        log.debug("Starting Mercurial sshserver for %s", self.db_repo.repo_full_path)
 
        sshserver(baseui, repo).serve_forever()
 
        mercurial.wireprotoserver.sshserver(baseui, repo).serve_forever()
kallithea/lib/vcs/backends/hg/workdir.py
Show inline comments
 
import mercurial.merge
 

	
 
from kallithea.lib.vcs.backends.base import BaseWorkdir
 
from kallithea.lib.vcs.exceptions import BranchDoesNotExistError
 
from kallithea.lib.vcs.utils import ascii_bytes, ascii_str
 
from kallithea.lib.vcs.utils.hgcompat import hg_merge
 

	
 

	
 
class MercurialWorkdir(BaseWorkdir):
 

	
 
    def get_branch(self):
 
        return self.repository._repo.dirstate.branch()
 

	
 
    def get_changeset(self):
 
        wk_dir_id = ascii_str(self.repository._repo[None].parents()[0].hex())
 
        return self.repository.get_changeset(wk_dir_id)
 

	
 
    def checkout_branch(self, branch=None):
 
        if branch is None:
 
            branch = self.repository.DEFAULT_BRANCH_NAME
 
        if branch not in self.repository.branches:
 
            raise BranchDoesNotExistError
 

	
 
        raw_id = self.repository.branches[branch]
 
        hg_merge.update(self.repository._repo, ascii_bytes(raw_id), False, False, None)
 
        mercurial.merge.update(self.repository._repo, ascii_bytes(raw_id), False, False, None)
kallithea/lib/vcs/utils/hgcompat.py
Show inline comments
 
"""
 
Mercurial libs compatibility
 
"""
 

	
 
import mercurial
 
from mercurial import archival, config, demandimport, discovery, httppeer, localrepo
 
from mercurial import merge as hg_merge
 
from mercurial import obsutil, patch, scmutil, sshpeer, ui, unionrepo
 
from mercurial.commands import clone, nullid, pull
 
from mercurial.context import memctx, memfilectx
 
from mercurial.discovery import findcommonoutgoing
 
from mercurial.error import Abort, RepoError, RepoLookupError
 
from mercurial.hg import peer
 
from mercurial.hgweb import hgweb_mod
 
from mercurial.hgweb.common import get_contact
 
from mercurial.match import exact as match_exact
 
from mercurial.match import match
 
from mercurial.mdiff import diffopts
 
from mercurial.node import hex, nullrev
 
from mercurial.scmutil import revrange
 
from mercurial.tags import tag
 
from mercurial.url import httpbasicauthhandler, httpdigestauthhandler
 
from mercurial.util import url as hg_url
 
import mercurial.localrepo
 

	
 

	
 
def monkey_do():
 
    """Apply some Mercurial monkey patching"""
 
# workaround for 3.3 94ac64bcf6fe and not calling largefiles reposetup correctly, and test_archival failing
 
localrepo.localrepository._lfstatuswriters = [lambda *msg, **opts: None]
 
    mercurial.localrepo.localrepository._lfstatuswriters = [lambda *msg, **opts: None]
 
# 3.5 7699d3212994 added the invariant that repo.lfstatus must exist before hitting overridearchive
 
localrepo.localrepository.lfstatus = False
 
    mercurial.localrepo.localrepository.lfstatus = False
kallithea/tests/vcs/test_hg.py
Show inline comments
 
import os
 

	
 
import mock
 
import pytest
 

	
 
from kallithea.lib.vcs.backends.hg import MercurialChangeset, MercurialRepository
 
from kallithea.lib.vcs.exceptions import NodeDoesNotExistError, RepositoryError, VCSError
 
from kallithea.lib.vcs.nodes import NodeKind, NodeState
 
from kallithea.lib.vcs.utils import safe_str
 
from kallithea.tests.vcs.conf import TEST_HG_REPO, TEST_HG_REPO_CLONE, TEST_HG_REPO_PULL, TESTS_TMP_PATH
 

	
 

	
 
class TestMercurialRepository(object):
 

	
 
    def __check_for_existing_repo(self):
 
        if os.path.exists(TEST_HG_REPO_CLONE):
 
            pytest.fail('Cannot test mercurial clone repo as location %s already '
 
                      'exists. You should manually remove it first.'
 
                      % TEST_HG_REPO_CLONE)
 

	
 
    def setup_method(self):
 
        self.repo = MercurialRepository(safe_str(TEST_HG_REPO))
 

	
 
    def test_wrong_repo_path(self):
 
        wrong_repo_path = os.path.join(TESTS_TMP_PATH, 'errorrepo')
 
        with pytest.raises(RepositoryError):
 
            MercurialRepository(wrong_repo_path)
 

	
 
    def test_unicode_path_repo(self):
 
        with pytest.raises(VCSError):
 
            MercurialRepository(u'iShouldFail')
 

	
 
    def test_repo_clone(self):
 
        self.__check_for_existing_repo()
 
        repo = MercurialRepository(safe_str(TEST_HG_REPO))
 
        repo_clone = MercurialRepository(TEST_HG_REPO_CLONE,
 
            src_url=TEST_HG_REPO, update_after_clone=True)
 
        assert len(repo.revisions) == len(repo_clone.revisions)
 
        # Checking hashes of changesets should be enough
 
        for changeset in repo.get_changesets():
 
            raw_id = changeset.raw_id
 
            assert raw_id == repo_clone.get_changeset(raw_id).raw_id
 

	
 
    def test_repo_clone_with_update(self):
 
        repo = MercurialRepository(safe_str(TEST_HG_REPO))
 
        repo_clone = MercurialRepository(TEST_HG_REPO_CLONE + '_w_update',
 
            src_url=TEST_HG_REPO, update_after_clone=True)
 
        assert len(repo.revisions) == len(repo_clone.revisions)
 

	
 
        # check if current workdir was updated
 
        assert os.path.isfile(
 
            os.path.join(
 
                TEST_HG_REPO_CLONE + '_w_update', 'MANIFEST.in'
 
            )
 
        )
 

	
 
    def test_repo_clone_without_update(self):
 
        repo = MercurialRepository(safe_str(TEST_HG_REPO))
 
        repo_clone = MercurialRepository(TEST_HG_REPO_CLONE + '_wo_update',
 
            src_url=TEST_HG_REPO, update_after_clone=False)
 
        assert len(repo.revisions) == len(repo_clone.revisions)
 
        assert not os.path.isfile(
 
            os.path.join(
 
                TEST_HG_REPO_CLONE + '_wo_update', 'MANIFEST.in'
 
            )
 
        )
 

	
 
    def test_pull(self):
 
        if os.path.exists(TEST_HG_REPO_PULL):
 
            pytest.fail('Cannot test mercurial pull command as location %s '
 
                      'already exists. You should manually remove it first'
 
                      % TEST_HG_REPO_PULL)
 
        repo_new = MercurialRepository(TEST_HG_REPO_PULL, create=True)
 
        assert len(self.repo.revisions) > len(repo_new.revisions)
 

	
 
        repo_new.pull(self.repo.path)
 
        repo_new = MercurialRepository(TEST_HG_REPO_PULL)
 
        assert len(self.repo.revisions) == len(repo_new.revisions)
 

	
 
    def test_revisions(self):
 
        # there are 21 revisions at bitbucket now
 
        # so we can assume they would be available from now on
 
        subset = set(['b986218ba1c9b0d6a259fac9b050b1724ed8e545',
 
                 '3d8f361e72ab303da48d799ff1ac40d5ac37c67e',
 
                 '6cba7170863a2411822803fa77a0a264f1310b35',
 
                 '56349e29c2af3ac913b28bde9a2c6154436e615b',
 
                 '2dda4e345facb0ccff1a191052dd1606dba6781d',
 
                 '6fff84722075f1607a30f436523403845f84cd9e',
 
                 '7d4bc8ec6be56c0f10425afb40b6fc315a4c25e7',
 
                 '3803844fdbd3b711175fc3da9bdacfcd6d29a6fb',
 
                 'dc5d2c0661b61928834a785d3e64a3f80d3aad9c',
 
                 'be90031137367893f1c406e0a8683010fd115b79',
 
                 'db8e58be770518cbb2b1cdfa69146e47cd481481',
 
                 '84478366594b424af694a6c784cb991a16b87c21',
 
                 '17f8e105dddb9f339600389c6dc7175d395a535c',
 
                 '20a662e756499bde3095ffc9bc0643d1def2d0eb',
 
                 '2e319b85e70a707bba0beff866d9f9de032aa4f9',
 
                 '786facd2c61deb9cf91e9534735124fb8fc11842',
 
                 '94593d2128d38210a2fcd1aabff6dda0d6d9edf8',
 
                 'aa6a0de05b7612707db567078e130a6cd114a9a7',
 
                 'eada5a770da98ab0dd7325e29d00e0714f228d09'
 
                ])
 
        assert subset.issubset(set(self.repo.revisions))
 

	
 
        # check if we have the proper order of revisions
 
        org = ['b986218ba1c9b0d6a259fac9b050b1724ed8e545',
 
                '3d8f361e72ab303da48d799ff1ac40d5ac37c67e',
 
                '6cba7170863a2411822803fa77a0a264f1310b35',
 
                '56349e29c2af3ac913b28bde9a2c6154436e615b',
 
                '2dda4e345facb0ccff1a191052dd1606dba6781d',
 
                '6fff84722075f1607a30f436523403845f84cd9e',
 
                '7d4bc8ec6be56c0f10425afb40b6fc315a4c25e7',
 
                '3803844fdbd3b711175fc3da9bdacfcd6d29a6fb',
 
                'dc5d2c0661b61928834a785d3e64a3f80d3aad9c',
 
                'be90031137367893f1c406e0a8683010fd115b79',
 
                'db8e58be770518cbb2b1cdfa69146e47cd481481',
 
                '84478366594b424af694a6c784cb991a16b87c21',
 
                '17f8e105dddb9f339600389c6dc7175d395a535c',
 
                '20a662e756499bde3095ffc9bc0643d1def2d0eb',
 
                '2e319b85e70a707bba0beff866d9f9de032aa4f9',
 
                '786facd2c61deb9cf91e9534735124fb8fc11842',
 
                '94593d2128d38210a2fcd1aabff6dda0d6d9edf8',
 
                'aa6a0de05b7612707db567078e130a6cd114a9a7',
 
                'eada5a770da98ab0dd7325e29d00e0714f228d09',
 
                '2c1885c735575ca478bf9e17b0029dca68824458',
 
                'd9bcd465040bf869799b09ad732c04e0eea99fe9',
 
                '469e9c847fe1f6f7a697b8b25b4bc5b48780c1a7',
 
                '4fb8326d78e5120da2c7468dcf7098997be385da',
 
                '62b4a097164940bd66030c4db51687f3ec035eed',
 
                '536c1a19428381cfea92ac44985304f6a8049569',
 
                '965e8ab3c44b070cdaa5bf727ddef0ada980ecc4',
 
                '9bb326a04ae5d98d437dece54be04f830cf1edd9',
 
                'f8940bcb890a98c4702319fbe36db75ea309b475',
 
                'ff5ab059786ebc7411e559a2cc309dfae3625a3b',
 
                '6b6ad5f82ad5bb6190037671bd254bd4e1f4bf08',
 
                'ee87846a61c12153b51543bf860e1026c6d3dcba', ]
 
        assert org == self.repo.revisions[:31]
 

	
 
    def test_iter_slice(self):
 
        sliced = list(self.repo[:10])
 
        itered = list(self.repo)[:10]
 
        assert sliced == itered
 

	
 
    def test_slicing(self):
 
        # 4 1 5 10 95
 
        for sfrom, sto, size in [(0, 4, 4), (1, 2, 1), (10, 15, 5),
 
                                 (10, 20, 10), (5, 100, 95)]:
 
            revs = list(self.repo[sfrom:sto])
 
            assert len(revs) == size
 
            assert revs[0] == self.repo.get_changeset(sfrom)
 
            assert revs[-1] == self.repo.get_changeset(sto - 1)
 

	
 
    def test_branches(self):
 
        # TODO: Need more tests here
 

	
 
        # active branches
 
        assert 'default' in self.repo.branches
 
        assert 'stable' in self.repo.branches
 

	
 
        # closed
 
        assert 'git' in self.repo._get_branches(closed=True)
 
        assert 'web' in self.repo._get_branches(closed=True)
 

	
 
        for name, id in self.repo.branches.items():
 
            assert isinstance(self.repo.get_changeset(id), MercurialChangeset)
 

	
 
    def test_tip_in_tags(self):
 
        # tip is always a tag
 
        assert 'tip' in self.repo.tags
 

	
 
    def test_tip_changeset_in_tags(self):
 
        tip = self.repo.get_changeset()
 
        assert self.repo.tags['tip'] == tip.raw_id
 

	
 
    def test_initial_changeset(self):
 

	
 
        init_chset = self.repo.get_changeset(0)
 
        assert init_chset.message == 'initial import'
 
        assert init_chset.author == 'Marcin Kuzminski <marcin@python-blog.com>'
 
        assert sorted(init_chset._file_paths) == sorted([
 
            'vcs/__init__.py',
 
            'vcs/backends/BaseRepository.py',
 
            'vcs/backends/__init__.py',
 
        ])
 

	
 
        assert sorted(init_chset._dir_paths) == sorted(['', 'vcs', 'vcs/backends'])
 

	
 
        with pytest.raises(NodeDoesNotExistError):
 
            init_chset.get_node(path='foobar')
 

	
 
        node = init_chset.get_node('vcs/')
 
        assert hasattr(node, 'kind')
 
        assert node.kind == NodeKind.DIR
 

	
 
        node = init_chset.get_node('vcs')
 
        assert hasattr(node, 'kind')
 
        assert node.kind == NodeKind.DIR
 

	
 
        node = init_chset.get_node('vcs/__init__.py')
 
        assert hasattr(node, 'kind')
 
        assert node.kind == NodeKind.FILE
 

	
 
    def test_not_existing_changeset(self):
 
        # rawid
 
        with pytest.raises(RepositoryError):
 
            self.repo.get_changeset('abcd' * 10)
 
        # shortid
 
        with pytest.raises(RepositoryError):
 
            self.repo.get_changeset('erro' * 4)
 
        # numeric
 
        with pytest.raises(RepositoryError):
 
            self.repo.get_changeset(self.repo.count() + 1)
 

	
 
        # Small chance we ever get to this one
 
        revision = pow(2, 30)
 
        with pytest.raises(RepositoryError):
 
            self.repo.get_changeset(revision)
 

	
 
    def test_changeset10(self):
 

	
 
        chset10 = self.repo.get_changeset(10)
 
        readme = b"""===
 
VCS
 
===
 

	
 
Various Version Control System management abstraction layer for Python.
 

	
 
Introduction
 
------------
 

	
 
TODO: To be written...
 

	
 
"""
 
        node = chset10.get_node('README.rst')
 
        assert node.kind == NodeKind.FILE
 
        assert node.content == readme
 

	
 
    @mock.patch('kallithea.lib.vcs.backends.hg.repository.diffopts')
 
    @mock.patch('mercurial.mdiff.diffopts')
 
    def test_get_diff_does_not_sanitize_zero_context(self, mock_diffopts):
 
        zero_context = 0
 

	
 
        self.repo.get_diff(0, 1, 'foo', context=zero_context)
 

	
 
        mock_diffopts.assert_called_once_with(git=True, showfunc=True, ignorews=False, context=zero_context)
 

	
 
    @mock.patch('kallithea.lib.vcs.backends.hg.repository.diffopts')
 
    @mock.patch('mercurial.mdiff.diffopts')
 
    def test_get_diff_sanitizes_negative_context(self, mock_diffopts):
 
        negative_context = -10
 
        zero_context = 0
 

	
 
        self.repo.get_diff(0, 1, 'foo', context=negative_context)
 

	
 
        mock_diffopts.assert_called_once_with(git=True, showfunc=True, ignorews=False, context=zero_context)
 

	
 

	
 
class TestMercurialChangeset(object):
 

	
 
    def setup_method(self):
 
        self.repo = MercurialRepository(safe_str(TEST_HG_REPO))
 

	
 
    def _test_equality(self, changeset):
 
        revision = changeset.revision
 
        assert changeset == self.repo.get_changeset(revision)
 

	
 
    def test_equality(self):
 
        revs = [0, 10, 20]
 
        changesets = [self.repo.get_changeset(rev) for rev in revs]
 
        for changeset in changesets:
 
            self._test_equality(changeset)
 

	
 
    def test_default_changeset(self):
 
        tip = self.repo.get_changeset('tip')
 
        assert tip == self.repo.get_changeset()
 
        assert tip == self.repo.get_changeset(revision=None)
 
        assert tip == list(self.repo[-1:])[0]
 

	
 
    def test_root_node(self):
 
        tip = self.repo.get_changeset('tip')
 
        assert tip.root is tip.get_node('')
 

	
 
    def test_lazy_fetch(self):
 
        """
 
        Test if changeset's nodes expands and are cached as we walk through
 
        the revision. This test is somewhat hard to write as order of tests
 
        is a key here. Written by running command after command in a shell.
 
        """
 
        chset = self.repo.get_changeset(45)
 
        assert len(chset.nodes) == 0
 
        root = chset.root
 
        assert len(chset.nodes) == 1
 
        assert len(root.nodes) == 8
 
        # accessing root.nodes updates chset.nodes
 
        assert len(chset.nodes) == 9
 

	
 
        docs = root.get_node('docs')
 
        # we haven't yet accessed anything new as docs dir was already cached
 
        assert len(chset.nodes) == 9
 
        assert len(docs.nodes) == 8
 
        # accessing docs.nodes updates chset.nodes
 
        assert len(chset.nodes) == 17
 

	
 
        assert docs is chset.get_node('docs')
 
        assert docs is root.nodes[0]
 
        assert docs is root.dirs[0]
 
        assert docs is chset.get_node('docs')
 

	
 
    def test_nodes_with_changeset(self):
 
        chset = self.repo.get_changeset(45)
 
        root = chset.root
 
        docs = root.get_node('docs')
 
        assert docs is chset.get_node('docs')
 
        api = docs.get_node('api')
 
        assert api is chset.get_node('docs/api')
 
        index = api.get_node('index.rst')
 
        assert index is chset.get_node('docs/api/index.rst')
 
        assert index is chset.get_node('docs').get_node('api').get_node('index.rst')
 

	
 
    def test_branch_and_tags(self):
 
        chset0 = self.repo.get_changeset(0)
 
        assert chset0.branch == 'default'
 
        assert chset0.branches == ['default']
 
        assert chset0.tags == []
 

	
 
        chset10 = self.repo.get_changeset(10)
 
        assert chset10.branch == 'default'
 
        assert chset10.branches == ['default']
 
        assert chset10.tags == []
 

	
 
        chset44 = self.repo.get_changeset(44)
 
        assert chset44.branch == 'web'
 
        assert chset44.branches == ['web']
 

	
 
        tip = self.repo.get_changeset('tip')
 
        assert 'tip' in tip.tags
 

	
 
    def _test_file_size(self, revision, path, size):
 
        node = self.repo.get_changeset(revision).get_node(path)
 
        assert node.is_file()
 
        assert node.size == size
 

	
 
    def test_file_size(self):
 
        to_check = (
 
            (10, 'setup.py', 1068),
 
            (20, 'setup.py', 1106),
 
            (60, 'setup.py', 1074),
 

	
 
            (10, 'vcs/backends/base.py', 2921),
 
            (20, 'vcs/backends/base.py', 3936),
 
            (60, 'vcs/backends/base.py', 6189),
 
        )
 
        for revision, path, size in to_check:
 
            self._test_file_size(revision, path, size)
 

	
 
    def _test_dir_size(self, revision, path, size):
 
        node = self.repo.get_changeset(revision).get_node(path)
 
        assert not node.is_file()
 
        assert node.size == size
 

	
 
    def test_dir_size(self):
 
        to_check = (
 
            ('96507bd11ecc', '/', 682421),
 
            ('a53d9201d4bc', '/', 682410),
 
            ('90243de06161', '/', 682006),
 
        )
 
        for revision, path, size in to_check:
 
            self._test_dir_size(revision, path, size)
 

	
 
    def test_repo_size(self):
 
        assert self.repo.size == 682421
 

	
 
    def test_file_history(self):
 
        # we can only check if those revisions are present in the history
 
        # as we cannot update this test every time file is changed
 
        files = {
 
            'setup.py': [7, 18, 45, 46, 47, 69, 77],
 
            'vcs/nodes.py': [7, 8, 24, 26, 30, 45, 47, 49, 56, 57, 58, 59, 60,
 
                61, 73, 76],
 
            'vcs/backends/hg.py': [4, 5, 6, 11, 12, 13, 14, 15, 16, 21, 22, 23,
 
                26, 27, 28, 30, 31, 33, 35, 36, 37, 38, 39, 40, 41, 44, 45, 47,
 
                48, 49, 53, 54, 55, 58, 60, 61, 67, 68, 69, 70, 73, 77, 78, 79,
 
                82],
 
        }
 
        for path, revs in files.items():
 
            tip = self.repo.get_changeset(revs[-1])
 
            node = tip.get_node(path)
 
            node_revs = [chset.revision for chset in node.history]
 
            assert set(revs).issubset(set(node_revs)), \
 
                "We assumed that %s is subset of revisions for which file %s " \
 
                "has been changed, and history of that node returned: %s" \
 
                % (revs, path, node_revs)
 

	
 
    def test_file_annotate(self):
 
        files = {
 
                 'vcs/backends/__init__.py':
 
                  {89: {'lines_no': 31,
 
                        'changesets': [32, 32, 61, 32, 32, 37, 32, 32, 32, 44,
 
                                       37, 37, 37, 37, 45, 37, 44, 37, 37, 37,
 
                                       32, 32, 32, 32, 37, 32, 37, 37, 32,
 
                                       32, 32]},
 
                   20: {'lines_no': 1,
 
                        'changesets': [4]},
 
                   55: {'lines_no': 31,
 
                        'changesets': [32, 32, 45, 32, 32, 37, 32, 32, 32, 44,
 
                                       37, 37, 37, 37, 45, 37, 44, 37, 37, 37,
 
                                       32, 32, 32, 32, 37, 32, 37, 37, 32,
 
                                       32, 32]}},
 
                 'vcs/exceptions.py':
 
                 {89: {'lines_no': 18,
 
                       'changesets': [16, 16, 16, 16, 16, 16, 16, 16, 16, 16,
 
                                      16, 16, 17, 16, 16, 18, 18, 18]},
 
                  20: {'lines_no': 18,
 
                       'changesets': [16, 16, 16, 16, 16, 16, 16, 16, 16, 16,
 
                                      16, 16, 17, 16, 16, 18, 18, 18]},
 
                  55: {'lines_no': 18, 'changesets': [16, 16, 16, 16, 16, 16,
 
                                                      16, 16, 16, 16, 16, 16,
 
                                                      17, 16, 16, 18, 18, 18]}},
 
                 'MANIFEST.in': {89: {'lines_no': 5,
 
                                      'changesets': [7, 7, 7, 71, 71]},
 
                                 20: {'lines_no': 3,
 
                                      'changesets': [7, 7, 7]},
 
                                 55: {'lines_no': 3,
 
                                     'changesets': [7, 7, 7]}}}
 

	
 
        for fname, revision_dict in files.items():
 
            for rev, data in revision_dict.items():
 
                cs = self.repo.get_changeset(rev)
 
                l1_1 = [x[1] for x in cs.get_file_annotate(fname)]
 
                l1_2 = [x[2]().raw_id for x in cs.get_file_annotate(fname)]
 
                assert l1_1 == l1_2
 
                l1 = l1_2 = [x[2]().revision for x in cs.get_file_annotate(fname)]
 
                l2 = files[fname][rev]['changesets']
 
                assert l1 == l2, "The lists of revision for %s@rev%s" \
 
                    "from annotation list should match each other," \
 
                    "got \n%s \nvs \n%s " % (fname, rev, l1, l2)
 

	
 
    def test_changeset_state(self):
 
        """
 
        Tests which files have been added/changed/removed at particular revision
 
        """
 

	
 
        # rev 46ad32a4f974:
 
        # hg st --rev 46ad32a4f974
 
        #    changed: 13
 
        #    added:   20
 
        #    removed: 1
 
        changed = set(['.hgignore'
 
            , 'README.rst', 'docs/conf.py', 'docs/index.rst', 'setup.py'
 
            , 'tests/test_hg.py', 'tests/test_nodes.py', 'vcs/__init__.py'
 
            , 'vcs/backends/__init__.py', 'vcs/backends/base.py'
 
            , 'vcs/backends/hg.py', 'vcs/nodes.py', 'vcs/utils/__init__.py'])
 

	
 
        added = set(['docs/api/backends/hg.rst'
 
            , 'docs/api/backends/index.rst', 'docs/api/index.rst'
 
            , 'docs/api/nodes.rst', 'docs/api/web/index.rst'
 
            , 'docs/api/web/simplevcs.rst', 'docs/installation.rst'
 
            , 'docs/quickstart.rst', 'setup.cfg', 'vcs/utils/baseui_config.py'
 
            , 'vcs/utils/web.py', 'vcs/web/__init__.py', 'vcs/web/exceptions.py'
 
            , 'vcs/web/simplevcs/__init__.py', 'vcs/web/simplevcs/exceptions.py'
 
            , 'vcs/web/simplevcs/middleware.py', 'vcs/web/simplevcs/models.py'
 
            , 'vcs/web/simplevcs/settings.py', 'vcs/web/simplevcs/utils.py'
 
            , 'vcs/web/simplevcs/views.py'])
 

	
 
        removed = set(['docs/api.rst'])
 

	
 
        chset64 = self.repo.get_changeset('46ad32a4f974')
 
        assert set((node.path for node in chset64.added)) == added
 
        assert set((node.path for node in chset64.changed)) == changed
 
        assert set((node.path for node in chset64.removed)) == removed
 

	
 
        # rev b090f22d27d6:
 
        # hg st --rev b090f22d27d6
 
        #    changed: 13
 
        #    added:   20
 
        #    removed: 1
 
        chset88 = self.repo.get_changeset('b090f22d27d6')
 
        assert set((node.path for node in chset88.added)) == set()
 
        assert set((node.path for node in chset88.changed)) == set(['.hgignore'])
 
        assert set((node.path for node in chset88.removed)) == set()
 

	
 
        # 85:
 
        #    added:   2 ['vcs/utils/diffs.py', 'vcs/web/simplevcs/views/diffs.py']
 
        #    changed: 4 ['vcs/web/simplevcs/models.py', ...]
 
        #    removed: 1 ['vcs/utils/web.py']
 
        chset85 = self.repo.get_changeset(85)
 
        assert set((node.path for node in chset85.added)) == set([
 
            'vcs/utils/diffs.py',
 
            'vcs/web/simplevcs/views/diffs.py'
 
        ])
 

	
 
        assert set((node.path for node in chset85.changed)) == set([
 
            'vcs/web/simplevcs/models.py',
 
            'vcs/web/simplevcs/utils.py',
 
            'vcs/web/simplevcs/views/__init__.py',
 
            'vcs/web/simplevcs/views/repository.py',
 
        ])
 

	
 
        assert set((node.path for node in chset85.removed)) == set([
 
            'vcs/utils/web.py'
 
        ])
 

	
 

	
 
    def test_files_state(self):
 
        """
 
        Tests state of FileNodes.
 
        """
 
        chset = self.repo.get_changeset(85)
 
        node = chset.get_node('vcs/utils/diffs.py')
 
        assert node.state, NodeState.ADDED
 
        assert node.added
 
        assert not node.changed
 
        assert not node.not_changed
 
        assert not node.removed
 

	
 
        chset = self.repo.get_changeset(88)
 
        node = chset.get_node('.hgignore')
 
        assert node.state, NodeState.CHANGED
 
        assert not node.added
 
        assert node.changed
 
        assert not node.not_changed
 
        assert not node.removed
 

	
 
        chset = self.repo.get_changeset(85)
 
        node = chset.get_node('setup.py')
 
        assert node.state, NodeState.NOT_CHANGED
 
        assert not node.added
 
        assert not node.changed
 
        assert node.not_changed
 
        assert not node.removed
 

	
 
        # If node has REMOVED state then trying to fetch it would raise
 
        # ChangesetError exception
 
        chset = self.repo.get_changeset(2)
 
        path = 'vcs/backends/BaseRepository.py'
 
        with pytest.raises(NodeDoesNotExistError):
 
            chset.get_node(path)
 
        # but it would be one of ``removed`` (changeset's attribute)
 
        assert path in [rf.path for rf in chset.removed]
 

	
 
    def test_commit_message_is_unicode(self):
 
        for cm in self.repo:
 
            assert isinstance(cm.message, unicode)
 

	
 
    def test_changeset_author_is_unicode(self):
 
        for cm in self.repo:
 
            assert isinstance(cm.author, unicode)
 

	
 
    def test_repo_files_content_is_bytes(self):
 
        test_changeset = self.repo.get_changeset(100)
 
        for node in test_changeset.get_node('/'):
 
            if node.is_file():
 
                assert isinstance(node.content, bytes)
 

	
 
    def test_wrong_path(self):
 
        # There is 'setup.py' in the root dir but not there:
 
        path = 'foo/bar/setup.py'
 
        with pytest.raises(VCSError):
 
            self.repo.get_changeset().get_node(path)
 

	
 
    def test_archival_file(self):
 
        # TODO:
 
        pass
 

	
 
    def test_archival_as_generator(self):
 
        # TODO:
 
        pass
 

	
 
    def test_archival_wrong_kind(self):
 
        tip = self.repo.get_changeset()
 
        with pytest.raises(VCSError):
 
            tip.fill_archive(kind='error')
 

	
 
    def test_archival_empty_prefix(self):
 
        # TODO:
 
        pass
 

	
 
    def test_author_email(self):
 
        assert 'marcin@python-blog.com' == self.repo.get_changeset('b986218ba1c9').author_email
 
        assert 'lukasz.balcerzak@python-center.pl' == self.repo.get_changeset('3803844fdbd3').author_email
 
        assert '' == self.repo.get_changeset('84478366594b').author_email
 

	
 
    def test_author_username(self):
 
        assert 'Marcin Kuzminski' == self.repo.get_changeset('b986218ba1c9').author_name
 
        assert 'Lukasz Balcerzak' == self.repo.get_changeset('3803844fdbd3').author_name
 
        assert 'marcink' == self.repo.get_changeset('84478366594b').author_name
 

	
 
    def test_successors(self):
 
        init_chset = self.repo.get_changeset(0)
 
        assert init_chset.successors == []
 

	
 
    def test_predecessors(self):
 
        init_chset = self.repo.get_changeset(0)
 
        assert len(init_chset.predecessors) == 0
0 comments (0 inline, 0 general)