# -*- 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.feed

~~~~~~~~~~~~~~~~~~~~~~~~~~

Feed controller 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 23, 2010

:author: marcink

:copyright: (c) 2013 RhodeCode GmbH, and others.

:license: GPLv3, see LICENSE.md for more details.

"""

import logging

from tg import response, tmpl_context as c

from tg.i18n import ugettext as _

from beaker.cache import cache_region, region_invalidate

from webhelpers.feedgenerator import Atom1Feed, Rss201rev2Feed

from kallithea import CONFIG

from kallithea.lib import helpers as h

from kallithea.lib.auth import LoginRequired, HasRepoPermissionLevelDecorator

from kallithea.lib.base import BaseRepoController

from kallithea.lib.diffs import DiffProcessor

from kallithea.model.db import CacheInvalidation

from kallithea.lib.utils2 import safe_int, str2bool, safe_unicode

log = logging.getLogger(__name__)

language = 'en-us'

ttl = "5"

class FeedController(BaseRepoController):

    @LoginRequired(api_access=True, allow_default_user=True)

    @HasRepoPermissionLevelDecorator('read')

    def _before(self, *args, **kwargs):

        super(FeedController, self)._before(*args, **kwargs)

    def _get_title(self, cs):

        return h.shorter(cs.message, 160)

    def __get_desc(self, cs):

        desc_msg = [(_('%s committed on %s')

                     % (h.person(cs.author), h.fmt_date(cs.date))) + '<br/>']

        # branches, tags, bookmarks

        for book in cs.bookmarks:

            desc_msg.append('bookmark: %s<br/>' % book)

        for tag in cs.tags:

            desc_msg.append('tag: %s<br/>' % tag)

        changes = []

        diff_limit = safe_int(CONFIG.get('rss_cut_off_limit', 32 * 1024))

        raw_diff = cs.diff()

        diff_processor = DiffProcessor(raw_diff,

                                       diff_limit=diff_limit,

                                       inline_diff=False)

        for st in diff_processor.parsed:

            st.update({'added': st['stats']['added'],

                       'removed': st['stats']['deleted']})

            changes.append('\n %(operation)s %(filename)s '

                           '(%(added)s lines added, %(removed)s lines removed)'

                            % st)

        if diff_processor.limited_diff:

            changes = changes + ['\n ' +

                                 _('Changeset was too big and was cut off...')]

        # rev link

        _url = h.canonical_url('changeset_home', repo_name=c.db_repo.repo_name,

                   revision=cs.raw_id)

        desc_msg.append('changeset: <a href="%s">%s</a>' % (_url, cs.raw_id[:8]))

        desc_msg.append('<pre>')

        desc_msg.append(h.urlify_text(cs.message))

        desc_msg.append('\n')

        desc_msg.extend(changes)

        if str2bool(CONFIG.get('rss_include_diff', False)):

            desc_msg.append('\n\n')

            desc_msg.append(raw_diff)

        desc_msg.append('</pre>')

        return map(safe_unicode, desc_msg)

    def atom(self, repo_name):

        """Produce an atom-1.0 feed via feedgenerator module"""

        @cache_region('long_term', '_get_feed_from_cache')

        def _get_feed_from_cache(key, kind):

            feed = Atom1Feed(

                title=_('%s %s feed') % (c.site_name, repo_name),

                link=h.canonical_url('summary_home', repo_name=repo_name),

                description=_('Changes on %s repository') % repo_name,

                language=language,

                ttl=ttl

            )

            rss_items_per_page = safe_int(CONFIG.get('rss_items_per_page', 20))

            for cs in reversed(list(c.db_repo_scm_instance[-rss_items_per_page:])):

                feed.add_item(title=self._get_title(cs),

                              link=h.canonical_url('changeset_home', repo_name=repo_name,

                                       revision=cs.raw_id),

                              author_name=cs.author,

                              description=''.join(self.__get_desc(cs)),

                              pubdate=cs.date,

                              )

            response.content_type = feed.mime_type

            return feed.writeString('utf-8')

        kind = 'ATOM'

        valid = CacheInvalidation.test_and_set_valid(repo_name, kind)

        if not valid:

            region_invalidate(_get_feed_from_cache, None, '_get_feed_from_cache', repo_name, kind)

        return _get_feed_from_cache(repo_name, kind)

    def rss(self, repo_name):

        """Produce an rss2 feed via feedgenerator module"""

        @cache_region('long_term', '_get_feed_from_cache')

        def _get_feed_from_cache(key, kind):

            feed = Rss201rev2Feed(

                title=_('%s %s feed') % (c.site_name, repo_name),

                link=h.canonical_url('summary_home', repo_name=repo_name),

                description=_('Changes on %s repository') % repo_name,

                language=language,

                ttl=ttl

            )

            rss_items_per_page = safe_int(CONFIG.get('rss_items_per_page', 20))

            for cs in reversed(list(c.db_repo_scm_instance[-rss_items_per_page:])):

                feed.add_item(title=self._get_title(cs),

                              link=h.canonical_url('changeset_home', repo_name=repo_name,

                                       revision=cs.raw_id),

                              author_name=cs.author,

                              description=''.join(self.__get_desc(cs)),

                              pubdate=cs.date,

                             )

            response.content_type = feed.mime_type

            return feed.writeString('utf-8')

        kind = 'RSS'

        except (ChangesetDoesNotExistError, LookupError):

            msg = _('Such revision does not exist for this repository')

            h.flash(msg, category='error')

            raise HTTPNotFound()

        except RepositoryError as e:

            h.flash(safe_str(e), category='error')

            raise HTTPNotFound()

    def __get_filenode(self, cs, path):

        """

        Returns file_node or raise HTTP error.

        :param cs: given changeset

        :param path: path to lookup

        """

        try:

            file_node = cs.get_node(path)

            if file_node.is_dir():

                raise RepositoryError('given path is a directory')

        except ChangesetDoesNotExistError:

            msg = _('Such revision does not exist for this repository')

            h.flash(msg, category='error')

            raise HTTPNotFound()

        except RepositoryError as e:

            h.flash(safe_str(e), category='error')

            raise HTTPNotFound()

        return file_node

    @LoginRequired(allow_default_user=True)

    @HasRepoPermissionLevelDecorator('read')

    def index(self, repo_name, revision, f_path, annotate=False):

        # redirect to given revision from form if given

        post_revision = request.POST.get('at_rev', None)

        if post_revision:

            cs = self.__get_cs(post_revision) # FIXME - unused!

        c.revision = revision

        c.changeset = self.__get_cs(revision)

        c.branch = request.GET.get('branch', None)

        c.f_path = f_path

        c.annotate = annotate

        cur_rev = c.changeset.revision

        # used in files_source.html:

        c.cut_off_limit = self.cut_off_limit

        c.fulldiff = request.GET.get('fulldiff')

        # prev link

        try:

            prev_rev = c.db_repo_scm_instance.get_changeset(cur_rev).prev(c.branch)

            c.url_prev = url('files_home', repo_name=c.repo_name,

                         revision=prev_rev.raw_id, f_path=f_path)

            if c.branch:

                c.url_prev += '?branch=%s' % c.branch

        except (ChangesetDoesNotExistError, VCSError):

            c.url_prev = '#'

        # next link

        try:

            next_rev = c.db_repo_scm_instance.get_changeset(cur_rev).next(c.branch)

            c.url_next = url('files_home', repo_name=c.repo_name,

                     revision=next_rev.raw_id, f_path=f_path)

            if c.branch:

                c.url_next += '?branch=%s' % c.branch

        except (ChangesetDoesNotExistError, VCSError):

            c.url_next = '#'

        # files or dirs

        try:

            c.file = c.changeset.get_node(f_path)

            if c.file.is_file():

                c.load_full_history = False

                # determine if we're on branch head

                _branches = c.db_repo_scm_instance.branches

                c.on_branch_head = revision in _branches.keys() + _branches.values()

                _hist = []

                c.file_history = []

                if c.load_full_history:

                    c.file_history, _hist = self._get_node_history(c.changeset, f_path)

                c.authors = []

                for a in set([x.author for x in _hist]):

                    c.authors.append((h.email(a), h.person(a)))

            else:

                c.authors = c.file_history = []

        except RepositoryError as e:

            h.flash(safe_str(e), category='error')

            raise HTTPNotFound()

        if request.environ.get('HTTP_X_PARTIAL_XHR'):

            return render('files/files_ypjax.html')

        # TODO: tags and bookmarks?

        c.revision_options = [(c.changeset.raw_id,

            [(n, b) for b, n in c.db_repo_scm_instance.branches.items()]

        if c.db_repo_scm_instance.closed_branches:

            prefix = _('(closed)') + ' '

            c.revision_options += [('-', '-')] + \

                [(n, prefix + b) for b, n in c.db_repo_scm_instance.closed_branches.items()]

        return render('files/files.html')

    @LoginRequired(allow_default_user=True)

    @HasRepoPermissionLevelDecorator('read')

    @jsonify

    def history(self, repo_name, revision, f_path):

        changeset = self.__get_cs(revision)

        _file = changeset.get_node(f_path)

        if _file.is_file():

            file_history, _hist = self._get_node_history(changeset, f_path)

            res = []

            for obj in file_history:

                res.append({

                    'text': obj[1],

                    'children': [{'id': o[0], 'text': o[1]} for o in obj[0]]

                })

            data = {

                'more': False,

                'results': res

            }

            return data

    @LoginRequired(allow_default_user=True)

    @HasRepoPermissionLevelDecorator('read')

    def authors(self, repo_name, revision, f_path):

        changeset = self.__get_cs(revision)

        _file = changeset.get_node(f_path)

        if _file.is_file():

            file_history, _hist = self._get_node_history(changeset, f_path)

            c.authors = []

            for a in set([x.author for x in _hist]):

                c.authors.append((h.email(a), h.person(a)))

            return render('files/files_history_box.html')

    @LoginRequired(allow_default_user=True)

    @HasRepoPermissionLevelDecorator('read')

    def rawfile(self, repo_name, revision, f_path):

        cs = self.__get_cs(revision)

        file_node = self.__get_filenode(cs, f_path)

        response.content_disposition = 'attachment; filename=%s' % \

            safe_str(f_path.split(Repository.url_sep())[-1])

        response.content_type = file_node.mimetype

        return file_node.content

    @LoginRequired(allow_default_user=True)

    @HasRepoPermissionLevelDecorator('read')

    def raw(self, repo_name, revision, f_path):

        cs = self.__get_cs(revision)

        file_node = self.__get_filenode(cs, f_path)

        raw_mimetype_mapping = {

            # map original mimetype to a mimetype used for "show as raw"

            # you can also provide a content-disposition to override the

            # default "attachment" disposition.

            # orig_type: (new_type, new_dispo)

            # show images inline:

            'image/x-icon': ('image/x-icon', 'inline'),

            'image/png': ('image/png', 'inline'),

            'image/gif': ('image/gif', 'inline'),

            'image/jpeg': ('image/jpeg', 'inline'),

            'image/svg+xml': ('image/svg+xml', 'inline'),

        }

        mimetype = file_node.mimetype

        try:

            mimetype, dispo = raw_mimetype_mapping[mimetype]

        except KeyError:

            # we don't know anything special about this, handle it safely

            if file_node.is_binary:

                # do same as download raw for binary files

                mimetype, dispo = 'application/octet-stream', 'attachment'

            else:

                # do not just use the original mimetype, but force text/plain,

                # otherwise it would serve text/html and that might be unsafe.

                # Note: underlying vcs library fakes text/plain mimetype if the

                # mimetype can not be determined and it thinks it is not

                # binary.This might lead to erroneous text display in some

                # cases, but helps in other cases, like with text files

                # without extension.

                mimetype, dispo = 'text/plain', 'inline'

        if dispo == 'attachment':

            dispo = 'attachment; filename=%s' % \

                        safe_str(f_path.split(os.sep)[-1])

            return raw_diff

        elif c.action == 'raw':

            raw_diff = diffs.get_gitdiff(node1, node2,

                                      ignore_whitespace=ignore_whitespace,

                                      context=line_context)

            response.content_type = 'text/plain'

            return raw_diff

        else:

            fid = h.FID(diff2, node2.path)

            line_context_lcl = get_line_ctx(fid, request.GET)

            ign_whitespace_lcl = get_ignore_ws(fid, request.GET)

            diff_limit = None if fulldiff else self.cut_off_limit

            c.a_rev, c.cs_rev, a_path, diff, st, op = diffs.wrapped_diff(filenode_old=node1,

                                         filenode_new=node2,

                                         diff_limit=diff_limit,

                                         ignore_whitespace=ign_whitespace_lcl,

                                         line_context=line_context_lcl,

                                         enable_comments=False)

            c.file_diff_data = [(fid, fid, op, a_path, node2.path, diff, st)]

            return render('files/file_diff.html')

    @LoginRequired(allow_default_user=True)

    @HasRepoPermissionLevelDecorator('read')

    def diff_2way(self, repo_name, f_path):

        diff1 = request.GET.get('diff1', '')

        diff2 = request.GET.get('diff2', '')

        try:

            if diff1 not in ['', None, 'None', '0' * 12, '0' * 40]:

                c.changeset_1 = c.db_repo_scm_instance.get_changeset(diff1)

                try:

                    node1 = c.changeset_1.get_node(f_path)

                    if node1.is_dir():

                        raise NodeError('%s path is a %s not a file'

                                        % (node1, type(node1)))

                except NodeDoesNotExistError:

                    c.changeset_1 = EmptyChangeset(cs=diff1,

                                                   revision=c.changeset_1.revision,

                                                   repo=c.db_repo_scm_instance)

                    node1 = FileNode(f_path, '', changeset=c.changeset_1)

            else:

                c.changeset_1 = EmptyChangeset(repo=c.db_repo_scm_instance)

                node1 = FileNode(f_path, '', changeset=c.changeset_1)

            if diff2 not in ['', None, 'None', '0' * 12, '0' * 40]:

                c.changeset_2 = c.db_repo_scm_instance.get_changeset(diff2)

                try:

                    node2 = c.changeset_2.get_node(f_path)

                    if node2.is_dir():

                        raise NodeError('%s path is a %s not a file'

                                        % (node2, type(node2)))

                except NodeDoesNotExistError:

                    c.changeset_2 = EmptyChangeset(cs=diff2,

                                                   revision=c.changeset_2.revision,

                                                   repo=c.db_repo_scm_instance)

                    node2 = FileNode(f_path, '', changeset=c.changeset_2)

            else:

                c.changeset_2 = EmptyChangeset(repo=c.db_repo_scm_instance)

                node2 = FileNode(f_path, '', changeset=c.changeset_2)

        except ChangesetDoesNotExistError as e:

            msg = _('Such revision does not exist for this repository')

            h.flash(msg, category='error')

            raise HTTPNotFound()

        c.node1 = node1

        c.node2 = node2

        c.cs1 = c.changeset_1

        c.cs2 = c.changeset_2

        return render('files/diff_2way.html')

    def _get_node_history(self, cs, f_path, changesets=None):

        """

        get changesets history for given node

        :param cs: changeset to calculate history

        :param f_path: path for node to calculate history for

        :param changesets: if passed don't calculate history and take

            changesets defined in this list

        """

        # calculate history based on tip

        tip_cs = c.db_repo_scm_instance.get_changeset()

        if changesets is None:

            try:

                changesets = tip_cs.get_file_history(f_path)

            except (NodeDoesNotExistError, ChangesetError):

                # this node is not present at tip !

                changesets = cs.get_file_history(f_path)

        hist_l = []

        changesets_group = ([], _("Changesets"))

        branches_group = ([], _("Branches"))

        tags_group = ([], _("Tags"))

        for chs in changesets:

            _branch = chs.branch

            n_desc = '%s (%s)' % (h.show_id(chs), _branch)

            changesets_group[0].append((chs.raw_id, n_desc,))

        hist_l.append(changesets_group)

        for name, chs in c.db_repo_scm_instance.branches.items():

            branches_group[0].append((chs, name),)

        hist_l.append(branches_group)

        for name, chs in c.db_repo_scm_instance.tags.items():

            tags_group[0].append((chs, name),)

        hist_l.append(tags_group)

        return hist_l, changesets

    @LoginRequired(allow_default_user=True)

    @HasRepoPermissionLevelDecorator('read')

    @jsonify

    def nodelist(self, repo_name, revision, f_path):

        if request.environ.get('HTTP_X_PARTIAL_XHR'):

            cs = self.__get_cs(revision)

            _d, _f = ScmModel().get_nodes(repo_name, cs.raw_id, f_path,

                                          flat=False)

            return {'nodes': _d + _f}

# 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

from tg import request, tmpl_context as c

from tg.i18n import ugettext as _

from webob.exc import HTTPFound, HTTPNotFound, HTTPForbidden, HTTPBadRequest

from kallithea.config.routing import url

from kallithea.lib import helpers as h

from kallithea.lib import diffs

from kallithea.lib.auth import LoginRequired, HasRepoPermissionLevelDecorator

from kallithea.lib.base import BaseRepoController, render, jsonify

from kallithea.lib.page import Page

from kallithea.lib.utils import action_logger

from kallithea.lib.vcs.exceptions import EmptyRepositoryError, ChangesetDoesNotExistError

from kallithea.lib.vcs.utils import safe_str

from kallithea.lib.vcs.utils.hgcompat import unionrepo

from kallithea.model.db import PullRequest, ChangesetStatus, ChangesetComment, \

    PullRequestReviewer, Repository, User

from kallithea.model.pull_request import CreatePullRequestAction, CreatePullRequestIterationAction, PullRequestModel

from kallithea.model.meta import Session

from kallithea.model.repo import RepoModel

from kallithea.model.comment import ChangesetCommentsModel

from kallithea.model.changeset_status import ChangesetStatusModel

from kallithea.model.forms import PullRequestForm, PullRequestPostForm

from kallithea.lib.utils2 import safe_int

from kallithea.controllers.changeset import _ignorews_url, _context_url, \

    create_comment

from kallithea.controllers.compare import CompareController

from kallithea.lib.graphmod import graph_data

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:

            branch_rev = safe_str(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(

                "sort(parents(branch(id(%s)) and merge()) - branch(id(%s)), -rev)",

                branch_rev, branch_rev):

        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 = 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 _get_is_allowed_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 ''

        if self.changed and not parents:

            raise NodeDoesNotExistError(str(self.changed[0].path))

        for p in parents:

            for node in self.changed:

                try:

                    old = p.get_node(node.path)

                    missing.remove(node)

                    # if content actually changed, remove node from unchanged

                    if old.content != node.content:

                        not_changed.remove(node)

                except NodeDoesNotExistError:

                    pass

        if self.changed and missing:

            raise NodeDoesNotExistError("Node at %s is missing "

                "(parents: %s)" % (node.path, parents))

        if self.changed and not_changed:

            raise NodeNotChangedError("Node at %s wasn't actually changed "

                "since parents' changesets: %s" % (not_changed.pop().path,

                    parents)

            )

        # Check nodes marked as removed

        if self.removed and not parents:

            raise NodeDoesNotExistError("Cannot remove node at %s as there "

                "were no parents specified" % self.removed[0].path)

        really_removed = set()

        for p in parents:

            for node in self.removed:

                try:

                    p.get_node(node.path)

                    really_removed.add(node)

                except ChangesetError:

                    pass

        not_removed = set(self.removed) - really_removed

        if not_removed:

            raise NodeDoesNotExistError("Cannot remove node at %s from "

                "following parents: %s" % (not_removed[0], parents))

    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``.

        .. note::

            While overriding this method each backend's should call

            ``self.check_integrity(parents)`` in the first place.

        :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

        """

        raise NotImplementedError

class EmptyChangeset(BaseChangeset):

    """

    An dummy empty changeset. It's possible to pass hash when creating

    an EmptyChangeset

    """

    def __init__(self, cs='0' * 40, repo=None, requested_revision=None,

                 alias=None, revision=-1, message='', author='', date=None):

        self._empty_cs = cs

        self.revision = revision

        self.message = message

        self.author = author

        self.date = date or datetime.datetime.fromtimestamp(0)

        self.repository = repo

        self.requested_revision = requested_revision

        self.alias = alias

    @LazyProperty

    def raw_id(self):

        """

        Returns raw string identifying this changeset, useful for web

        representation.

        """

        return self._empty_cs

    @LazyProperty

    def branch(self):

        from kallithea.lib.vcs.backends import get_backend

        return get_backend(self.alias).DEFAULT_BRANCH_NAME

    @LazyProperty

    def short_id(self):

        return self.raw_id[:12]

    def get_file_changeset(self, path):

        return self

    def get_file_content(self, path):

        return u''

    def get_file_size(self, path):

        return 0

class CollectionGenerator(object):

    def __init__(self, repo, revs):

        self.repo = repo

        self.revs = revs

    def __len__(self):

        return len(self.revs)

    def __iter__(self):

        for rev in self.revs:

            yield self.repo.get_changeset(rev)

    def __getitem__(self, what):

        """Return either a single element by index, or a sliced collection."""

        if isinstance(what, slice):

            return CollectionGenerator(self.repo, self.revs[what])

        else:

            # single item

            return self.repo.get_changeset(self.revs[what])

    def __repr__(self):

        return '<CollectionGenerator[len:%s]>' % (len(self))

import re

from itertools import chain

from dulwich import objects

from subprocess import Popen, PIPE

from kallithea.lib.vcs.conf import settings

from kallithea.lib.vcs.backends.base import BaseChangeset, EmptyChangeset

from kallithea.lib.vcs.exceptions import (

    RepositoryError, ChangesetError, NodeDoesNotExistError, VCSError,

    ChangesetDoesNotExistError, ImproperArchiveTypeError

)

from kallithea.lib.vcs.nodes import (

    FileNode, DirNode, NodeKind, RootNode, SubModuleNode,

    ChangedFileNodesGenerator, AddedFileNodesGenerator, RemovedFileNodesGenerator

)

from kallithea.lib.vcs.utils import (

    safe_unicode, safe_str, safe_int, date_fromtimestamp

)

from kallithea.lib.vcs.utils.lazy import LazyProperty

class GitChangeset(BaseChangeset):

    """

    Represents state of the repository at single revision.

    """

    def __init__(self, repository, revision):

        self._stat_modes = {}