Changeset - 75b128508fa8
[Not reviewed]
default
1 6 0
Mads Kiilerich - 6 years ago 2019-07-21 19:12:20
mads@kiilerich.com
compat: use collections.OrderedDict - we only support Python 2.7 and no longer need our own implementation
7 files changed with 6 insertions and 388 deletions:
0 comments (0 inline, 0 general)
kallithea/controllers/changeset.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.changeset
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
changeset controller showing changes between revisions
 

	
 
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 25, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
import logging
 
import traceback
 
from collections import defaultdict
 
from collections import defaultdict, OrderedDict
 

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

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

	
 
import kallithea.lib.helpers as h
 
from kallithea.lib.auth import LoginRequired, HasRepoPermissionLevelDecorator
 
from kallithea.lib.base import BaseRepoController, render, jsonify
 
from kallithea.lib.utils import action_logger
 
from kallithea.lib.compat import OrderedDict
 
from kallithea.lib import diffs
 
from kallithea.model.db import ChangesetComment, ChangesetStatus
 
from kallithea.model.comment import ChangesetCommentsModel
 
from kallithea.model.changeset_status import ChangesetStatusModel
 
from kallithea.model.meta import Session
 
from kallithea.model.pull_request import PullRequestModel
 
from kallithea.model.repo import RepoModel
 
from kallithea.lib.vcs.backends.base import EmptyChangeset
 
from kallithea.lib.utils2 import safe_unicode
 
from kallithea.lib.graphmod import graph_data
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
def _update_with_GET(params, GET):
 
    for k in ['diff1', 'diff2', 'diff']:
 
        params[k] += GET.getall(k)
 

	
 

	
 
def anchor_url(revision, path, GET):
 
    fid = h.FID(revision, path)
 
    return h.url.current(anchor=fid, **dict(GET))
 

	
 

	
 
def get_ignore_ws(fid, GET):
 
    ig_ws_global = GET.get('ignorews')
 
    ig_ws = filter(lambda k: k.startswith('WS'), GET.getall(fid))
 
    if ig_ws:
 
        try:
 
            return int(ig_ws[0].split(':')[-1])
 
        except ValueError:
 
            raise HTTPBadRequest()
 
    return ig_ws_global
 

	
 

	
 
def _ignorews_url(GET, fileid=None):
 
    fileid = str(fileid) if fileid else None
 
    params = defaultdict(list)
 
    _update_with_GET(params, GET)
 
    lbl = _('Show whitespace')
 
    ig_ws = get_ignore_ws(fileid, GET)
 
    ln_ctx = get_line_ctx(fileid, GET)
 
    # global option
 
    if fileid is None:
 
        if ig_ws is None:
 
            params['ignorews'] += [1]
 
            lbl = _('Ignore whitespace')
 
        ctx_key = 'context'
 
        ctx_val = ln_ctx
 
    # per file options
 
    else:
 
        if ig_ws is None:
 
            params[fileid] += ['WS:1']
 
            lbl = _('Ignore whitespace')
 

	
 
        ctx_key = fileid
 
        ctx_val = 'C:%s' % ln_ctx
 
    # if we have passed in ln_ctx pass it along to our params
 
    if ln_ctx:
 
        params[ctx_key] += [ctx_val]
 

	
 
    params['anchor'] = fileid
 
    icon = h.literal('<i class="icon-strike"></i>')
 
    return h.link_to(icon, h.url.current(**params), title=lbl, **{'data-toggle': 'tooltip'})
 

	
 

	
 
def get_line_ctx(fid, GET):
 
    ln_ctx_global = GET.get('context')
 
    if fid:
 
        ln_ctx = filter(lambda k: k.startswith('C'), GET.getall(fid))
 
    else:
 
        _ln_ctx = filter(lambda k: k.startswith('C'), GET)
 
        ln_ctx = GET.get(_ln_ctx[0]) if _ln_ctx else ln_ctx_global
 
        if ln_ctx:
 
            ln_ctx = [ln_ctx]
 

	
 
    if ln_ctx:
 
        retval = ln_ctx[0].split(':')[-1]
 
    else:
 
        retval = ln_ctx_global
 

	
 
    try:
 
        return int(retval)
 
    except Exception:
 
        return 3
 

	
 

	
 
def _context_url(GET, fileid=None):
 
    """
 
    Generates url for context lines
 

	
 
    :param fileid:
 
    """
 

	
 
    fileid = str(fileid) if fileid else None
 
    ig_ws = get_ignore_ws(fileid, GET)
 
    ln_ctx = (get_line_ctx(fileid, GET) or 3) * 2
 

	
 
    params = defaultdict(list)
 
    _update_with_GET(params, GET)
 

	
 
    # global option
 
    if fileid is None:
 
        if ln_ctx > 0:
 
            params['context'] += [ln_ctx]
 

	
 
        if ig_ws:
 
            ig_ws_key = 'ignorews'
 
            ig_ws_val = 1
 

	
 
    # per file option
 
    else:
 
        params[fileid] += ['C:%s' % ln_ctx]
 
        ig_ws_key = fileid
 
        ig_ws_val = 'WS:%s' % 1
 

	
 
    if ig_ws:
 
        params[ig_ws_key] += [ig_ws_val]
 

	
 
    lbl = _('Increase diff context to %(num)s lines') % {'num': ln_ctx}
 

	
 
    params['anchor'] = fileid
 
    icon = h.literal('<i class="icon-sort"></i>')
 
    return h.link_to(icon, h.url.current(**params), title=lbl, **{'data-toggle': 'tooltip'})
 

	
 

	
 
def create_cs_pr_comment(repo_name, revision=None, pull_request=None, allowed_to_change_status=True):
 
    """
 
    Add a comment to the specified changeset or pull request, using POST values
 
    from the request.
 

	
 
    Comments can be inline (when a file path and line number is specified in
 
    POST) or general comments.
 
    A comment can be accompanied by a review status change (accepted, rejected,
 
    etc.). Pull requests can be closed or deleted.
 

	
 
    Parameter 'allowed_to_change_status' is used for both status changes and
 
    closing of pull requests. For deleting of pull requests, more specific
 
    checks are done.
 
    """
 

	
 
    assert request.environ.get('HTTP_X_PARTIAL_XHR')
 
    if pull_request:
 
        pull_request_id = pull_request.pull_request_id
 
    else:
 
        pull_request_id = None
 

	
 
    status = request.POST.get('changeset_status')
 
    close_pr = request.POST.get('save_close')
 
    delete = request.POST.get('save_delete')
 
    f_path = request.POST.get('f_path')
 
    line_no = request.POST.get('line')
 

	
 
    if (status or close_pr or delete) and (f_path or line_no):
 
        # status votes and closing is only possible in general comments
 
        raise HTTPBadRequest()
 

	
 
    if not allowed_to_change_status:
 
        if status or close_pr:
 
            h.flash(_('No permission to change status'), 'error')
 
            raise HTTPForbidden()
 

	
 
    if pull_request and delete == "delete":
 
        if (pull_request.owner_id == request.authuser.user_id or
 
            h.HasPermissionAny('hg.admin')() or
 
            h.HasRepoPermissionLevel('admin')(pull_request.org_repo.repo_name) or
 
            h.HasRepoPermissionLevel('admin')(pull_request.other_repo.repo_name)
 
            ) and not pull_request.is_closed():
 
            PullRequestModel().delete(pull_request)
 
            Session().commit()
 
            h.flash(_('Successfully deleted pull request %s') % pull_request_id,
 
                    category='success')
 
            return {
 
               'location': h.url('my_pullrequests'), # or repo pr list?
 
            }
 
            raise HTTPFound(location=h.url('my_pullrequests')) # or repo pr list?
 
        raise HTTPForbidden()
 

	
 
    text = request.POST.get('text', '').strip()
 

	
 
    comment = ChangesetCommentsModel().create(
 
        text=text,
 
        repo=c.db_repo.repo_id,
 
        author=request.authuser.user_id,
 
        revision=revision,
 
        pull_request=pull_request_id,
 
        f_path=f_path or None,
 
        line_no=line_no or None,
 
        status_change=ChangesetStatus.get_status_lbl(status) if status else None,
 
        closing_pr=close_pr,
 
    )
 

	
 
    if status:
 
        ChangesetStatusModel().set_status(
 
            c.db_repo.repo_id,
 
            status,
 
            request.authuser.user_id,
 
            comment,
 
            revision=revision,
 
            pull_request=pull_request_id,
 
        )
 

	
 
    if pull_request:
 
        action = 'user_commented_pull_request:%s' % pull_request_id
 
    else:
 
        action = 'user_commented_revision:%s' % revision
 
    action_logger(request.authuser, action, c.db_repo, request.ip_addr)
 

	
 
    if pull_request and close_pr:
 
        PullRequestModel().close_pull_request(pull_request_id)
 
        action_logger(request.authuser,
 
                      'user_closed_pull_request:%s' % pull_request_id,
 
                      c.db_repo, request.ip_addr)
 

	
 
    Session().commit()
 

	
 
    data = {
 
       'target_id': h.safeid(h.safe_unicode(request.POST.get('f_path'))),
 
    }
 
    if comment is not None:
 
        c.comment = comment
 
        data.update(comment.get_dict())
 
        data.update({'rendered_text':
 
                     render('changeset/changeset_comment_block.html')})
 

	
 
    return data
 

	
 
def delete_cs_pr_comment(repo_name, comment_id):
 
    """Delete a comment from a changeset or pull request"""
 
    co = ChangesetComment.get_or_404(comment_id)
 
    if co.repo.repo_name != repo_name:
 
        raise HTTPNotFound()
 
    if co.pull_request and co.pull_request.is_closed():
 
        # don't allow deleting comments on closed pull request
 
        raise HTTPForbidden()
 

	
 
    owner = co.author_id == request.authuser.user_id
 
    repo_admin = h.HasRepoPermissionLevel('admin')(repo_name)
 
    if h.HasPermissionAny('hg.admin')() or repo_admin or owner:
 
        ChangesetCommentsModel().delete(comment=co)
 
        Session().commit()
 
        return True
 
    else:
 
        raise HTTPForbidden()
 

	
 
class ChangesetController(BaseRepoController):
 

	
 
    def _before(self, *args, **kwargs):
 
        super(ChangesetController, self)._before(*args, **kwargs)
 
        c.affected_files_cut_off = 60
 

	
 
    def _index(self, revision, method):
 
        c.pull_request = None
 
        c.anchor_url = anchor_url
 
        c.ignorews_url = _ignorews_url
 
        c.context_url = _context_url
 
        c.fulldiff = request.GET.get('fulldiff') # for reporting number of changed files
 
        # get ranges of revisions if preset
 
        rev_range = revision.split('...')[:2]
 
        enable_comments = True
 
        c.cs_repo = c.db_repo
 
        try:
 
            if len(rev_range) == 2:
 
                enable_comments = False
 
                rev_start = rev_range[0]
 
                rev_end = rev_range[1]
 
                rev_ranges = c.db_repo_scm_instance.get_changesets(start=rev_start,
 
                                                             end=rev_end)
 
            else:
 
                rev_ranges = [c.db_repo_scm_instance.get_changeset(revision)]
 

	
 
            c.cs_ranges = list(rev_ranges)
 
            if not c.cs_ranges:
 
                raise RepositoryError('Changeset range returned empty result')
 

	
 
        except (ChangesetDoesNotExistError, EmptyRepositoryError):
 
            log.debug(traceback.format_exc())
 
            msg = _('Such revision does not exist for this repository')
 
            h.flash(msg, category='error')
 
            raise HTTPNotFound()
 

	
 
        c.changes = OrderedDict()
 

	
 
        c.lines_added = 0  # count of lines added
 
        c.lines_deleted = 0  # count of lines removes
 

	
 
        c.changeset_statuses = ChangesetStatus.STATUSES
 
        comments = dict()
 
        c.statuses = []
 
        c.inline_comments = []
 
        c.inline_cnt = 0
 

	
 
        # Iterate over ranges (default changeset view is always one changeset)
 
        for changeset in c.cs_ranges:
 
            if method == 'show':
 
                c.statuses.extend([ChangesetStatusModel().get_status(
 
                            c.db_repo.repo_id, changeset.raw_id)])
 

	
 
                # Changeset comments
 
                comments.update((com.comment_id, com)
 
                                for com in ChangesetCommentsModel()
 
                                .get_comments(c.db_repo.repo_id,
 
                                              revision=changeset.raw_id))
 

	
 
                # Status change comments - mostly from pull requests
 
                comments.update((st.comment_id, st.comment)
 
                                for st in ChangesetStatusModel()
 
                                .get_statuses(c.db_repo.repo_id,
 
                                              changeset.raw_id, with_revisions=True)
 
                                if st.comment_id is not None)
 

	
 
                inlines = ChangesetCommentsModel() \
 
                            .get_inline_comments(c.db_repo.repo_id,
 
                                                 revision=changeset.raw_id)
 
                c.inline_comments.extend(inlines)
 

	
 
            cs2 = changeset.raw_id
 
            cs1 = changeset.parents[0].raw_id if changeset.parents else EmptyChangeset().raw_id
 
            context_lcl = get_line_ctx('', request.GET)
 
            ign_whitespace_lcl = get_ignore_ws('', request.GET)
 

	
 
            raw_diff = diffs.get_diff(c.db_repo_scm_instance, cs1, cs2,
 
                ignore_whitespace=ign_whitespace_lcl, context=context_lcl)
 
            diff_limit = None if c.fulldiff else self.cut_off_limit
 
            file_diff_data = []
 
            if method == 'show':
 
                diff_processor = diffs.DiffProcessor(raw_diff,
 
                                                     vcs=c.db_repo_scm_instance.alias,
 
                                                     diff_limit=diff_limit)
 
                c.limited_diff = diff_processor.limited_diff
 
                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(changeset.raw_id, filename)
 
                    url_fid = h.FID('', filename)
 
                    html_diff = diffs.as_html(enable_comments=enable_comments, parsed_lines=[f])
 
                    file_diff_data.append((fid, url_fid, f['operation'], f['old_filename'], filename, html_diff, st))
 
            else:
 
                # downloads/raw we only need RAW diff nothing else
 
                file_diff_data.append(('', None, None, None, raw_diff, None))
 
            c.changes[changeset.raw_id] = (cs1, cs2, file_diff_data)
 

	
 
        # sort comments in creation order
 
        c.comments = [com for com_id, com in sorted(comments.items())]
 

	
 
        # count inline comments
 
        for __, lines in c.inline_comments:
 
            for comments in lines.values():
 
                c.inline_cnt += len(comments)
 

	
 
        if len(c.cs_ranges) == 1:
 
            c.changeset = c.cs_ranges[0]
 
            c.parent_tmpl = ''.join(['# Parent  %s\n' % x.raw_id
 
                                     for x in c.changeset.parents])
 
        if method == 'download':
 
            response.content_type = 'text/plain'
 
            response.content_disposition = 'attachment; filename=%s.diff' \
 
                                            % revision[:12]
 
            return raw_diff
 
        elif method == 'patch':
 
            response.content_type = 'text/plain'
 
            c.diff = safe_unicode(raw_diff)
 
            return render('changeset/patch_changeset.html')
 
        elif method == 'raw':
 
            response.content_type = 'text/plain'
 
            return raw_diff
 
        elif method == 'show':
 
            if len(c.cs_ranges) == 1:
 
                return render('changeset/changeset.html')
 
            else:
 
                c.cs_ranges_org = None
 
                c.cs_comments = {}
 
                revs = [ctx.revision for ctx in reversed(c.cs_ranges)]
 
                c.jsdata = graph_data(c.db_repo_scm_instance, revs)
 
                return render('changeset/changeset_range.html')
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def index(self, revision, method='show'):
 
        return self._index(revision, method=method)
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def changeset_raw(self, revision):
 
        return self._index(revision, method='raw')
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def changeset_patch(self, revision):
 
        return self._index(revision, method='patch')
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def changeset_download(self, revision):
 
        return self._index(revision, method='download')
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('read')
 
    @jsonify
 
    def comment(self, repo_name, revision):
 
        return create_cs_pr_comment(repo_name, revision=revision)
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('read')
 
    @jsonify
 
    def delete_comment(self, repo_name, comment_id):
 
        return delete_cs_pr_comment(repo_name, comment_id)
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    @jsonify
 
    def changeset_info(self, repo_name, revision):
 
        if request.is_xhr:
 
            try:
 
                return c.db_repo_scm_instance.get_changeset(revision)
 
            except ChangesetDoesNotExistError as e:
 
                return EmptyChangeset(message=str(e))
 
        else:
 
            raise HTTPBadRequest()
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    @jsonify
 
    def changeset_children(self, repo_name, revision):
 
        if request.is_xhr:
 
            changeset = c.db_repo_scm_instance.get_changeset(revision)
 
            result = {"results": []}
 
            if changeset.children:
 
                result = {"results": changeset.children}
 
            return result
 
        else:
 
            raise HTTPBadRequest()
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    @jsonify
 
    def changeset_parents(self, repo_name, revision):
 
        if request.is_xhr:
 
            changeset = c.db_repo_scm_instance.get_changeset(revision)
 
            result = {"results": []}
 
            if changeset.parents:
 
                result = {"results": changeset.parents}
 
            return result
 
        else:
 
            raise HTTPBadRequest()
kallithea/controllers/files.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.files
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
Files 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 21, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
import os
 
import posixpath
 
import logging
 
import traceback
 
import tempfile
 
import shutil
 
from collections import OrderedDict
 

	
 
from tg import request, response, tmpl_context as c
 
from tg.i18n import ugettext as _
 
from webob.exc import HTTPFound
 

	
 
from kallithea.config.routing import url
 
from kallithea.lib.utils import action_logger
 
from kallithea.lib import diffs
 
from kallithea.lib import helpers as h
 

	
 
from kallithea.lib.compat import OrderedDict
 
from kallithea.lib.utils2 import convert_line_endings, detect_mode, safe_str, \
 
    str2bool, safe_int
 
from kallithea.lib.auth import LoginRequired, HasRepoPermissionLevelDecorator
 
from kallithea.lib.base import BaseRepoController, render, jsonify
 
from kallithea.lib.vcs.backends.base import EmptyChangeset
 
from kallithea.lib.vcs.conf import settings
 
from kallithea.lib.vcs.exceptions import RepositoryError, \
 
    ChangesetDoesNotExistError, EmptyRepositoryError, \
 
    ImproperArchiveTypeError, VCSError, NodeAlreadyExistsError, \
 
    NodeDoesNotExistError, ChangesetError, NodeError
 
from kallithea.lib.vcs.nodes import FileNode
 

	
 
from kallithea.model.repo import RepoModel
 
from kallithea.model.scm import ScmModel
 
from kallithea.model.db import Repository
 

	
 
from kallithea.controllers.changeset import anchor_url, _ignorews_url, \
 
    _context_url, get_line_ctx, get_ignore_ws
 
from webob.exc import HTTPNotFound
 
from kallithea.lib.exceptions import NonRelativePathError
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
class FilesController(BaseRepoController):
 

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

	
 
    def __get_cs(self, rev, silent_empty=False):
 
        """
 
        Safe way to get changeset if error occur it redirects to tip with
 
        proper message
 

	
 
        :param rev: revision to fetch
 
        :silent_empty: return None if repository is empty
 
        """
 

	
 
        try:
 
            return c.db_repo_scm_instance.get_changeset(rev)
 
        except EmptyRepositoryError as e:
 
            if silent_empty:
 
                return None
 
            url_ = url('files_add_home',
 
                       repo_name=c.repo_name,
 
                       revision=0, f_path='', anchor='edit')
 
            add_new = h.link_to(_('Click here to add new file'), url_, class_="alert-link")
 
            h.flash(_('There are no files yet.') + ' ' + add_new, category='warning')
 
            raise HTTPNotFound()
 
        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_submodule():
 
                raise HTTPFound(location=c.file.url)
 
            elif 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,
 
                              _('%s at %s') % (b, h.short_id(c.changeset.raw_id))) for b in c.changeset.branches] + \
 
            [(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])
 

	
 
        response.content_disposition = dispo
 
        response.content_type = mimetype
 
        return file_node.content
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('write')
 
    def delete(self, repo_name, revision, f_path):
 
        repo = c.db_repo
 
        # check if revision is a branch identifier- basically we cannot
 
        # create multiple heads via file editing
 
        _branches = repo.scm_instance.branches
 
        # check if revision is a branch name or branch hash
 
        if revision not in _branches.keys() + _branches.values():
 
            h.flash(_('You can only delete files with revision '
 
                      'being a valid branch'), category='warning')
 
            raise HTTPFound(location=h.url('files_home',
 
                                  repo_name=repo_name, revision='tip',
 
                                  f_path=f_path))
 

	
 
        r_post = request.POST
 

	
 
        c.cs = self.__get_cs(revision)
 
        c.file = self.__get_filenode(c.cs, f_path)
 

	
 
        c.default_message = _('Deleted file %s via Kallithea') % (f_path)
 
        c.f_path = f_path
 
        node_path = f_path
 
        author = request.authuser.full_contact
 

	
 
        if r_post:
 
            message = r_post.get('message') or c.default_message
 

	
 
            try:
 
                nodes = {
 
                    node_path: {
 
                        'content': ''
 
                    }
 
                }
 
                self.scm_model.delete_nodes(
 
                    user=request.authuser.user_id,
 
                    ip_addr=request.ip_addr,
 
                    repo=c.db_repo,
 
                    message=message,
 
                    nodes=nodes,
 
                    parent_cs=c.cs,
 
                    author=author,
 
                )
 

	
 
                h.flash(_('Successfully deleted file %s') % f_path,
 
                        category='success')
 
            except Exception:
 
                log.error(traceback.format_exc())
 
                h.flash(_('Error occurred during commit'), category='error')
 
            raise HTTPFound(location=url('changeset_home',
 
                                repo_name=c.repo_name, revision='tip'))
 

	
 
        return render('files/files_delete.html')
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('write')
 
    def edit(self, repo_name, revision, f_path):
 
        repo = c.db_repo
 
        # check if revision is a branch identifier- basically we cannot
 
        # create multiple heads via file editing
 
        _branches = repo.scm_instance.branches
 
        # check if revision is a branch name or branch hash
 
        if revision not in _branches.keys() + _branches.values():
 
            h.flash(_('You can only edit files with revision '
 
                      'being a valid branch'), category='warning')
 
            raise HTTPFound(location=h.url('files_home',
 
                                  repo_name=repo_name, revision='tip',
 
                                  f_path=f_path))
 

	
 
        r_post = request.POST
 

	
 
        c.cs = self.__get_cs(revision)
 
        c.file = self.__get_filenode(c.cs, f_path)
 

	
 
        if c.file.is_binary:
 
            raise HTTPFound(location=url('files_home', repo_name=c.repo_name,
 
                            revision=c.cs.raw_id, f_path=f_path))
 
        c.default_message = _('Edited file %s via Kallithea') % (f_path)
 
        c.f_path = f_path
 

	
 
        if r_post:
 

	
 
            old_content = c.file.content
 
            sl = old_content.splitlines(1)
 
            first_line = sl[0] if sl else ''
 
            # modes:  0 - Unix, 1 - Mac, 2 - DOS
 
            mode = detect_mode(first_line, 0)
 
            content = convert_line_endings(r_post.get('content', ''), mode)
 

	
 
            message = r_post.get('message') or c.default_message
 
            author = request.authuser.full_contact
 

	
 
            if content == old_content:
 
                h.flash(_('No changes'), category='warning')
 
                raise HTTPFound(location=url('changeset_home', repo_name=c.repo_name,
 
                                    revision='tip'))
 
            try:
 
                self.scm_model.commit_change(repo=c.db_repo_scm_instance,
 
                                             repo_name=repo_name, cs=c.cs,
 
                                             user=request.authuser.user_id,
 
                                             ip_addr=request.ip_addr,
 
                                             author=author, message=message,
 
                                             content=content, f_path=f_path)
 
                h.flash(_('Successfully committed to %s') % f_path,
 
                        category='success')
 
            except Exception:
 
                log.error(traceback.format_exc())
 
                h.flash(_('Error occurred during commit'), category='error')
 
            raise HTTPFound(location=url('changeset_home',
 
                                repo_name=c.repo_name, revision='tip'))
 

	
 
        return render('files/files_edit.html')
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionLevelDecorator('write')
 
    def add(self, repo_name, revision, f_path):
 

	
 
        repo = c.db_repo
 
        r_post = request.POST
 
        c.cs = self.__get_cs(revision, silent_empty=True)
 
        if c.cs is None:
 
            c.cs = EmptyChangeset(alias=c.db_repo_scm_instance.alias)
 
        c.default_message = (_('Added file via Kallithea'))
 
        c.f_path = f_path
 

	
 
        if r_post:
 
            unix_mode = 0
 
            content = convert_line_endings(r_post.get('content', ''), unix_mode)
 

	
 
            message = r_post.get('message') or c.default_message
 
            filename = r_post.get('filename')
 
            location = r_post.get('location', '')
 
            file_obj = r_post.get('upload_file', None)
 

	
 
            if file_obj is not None and hasattr(file_obj, 'filename'):
 
                filename = file_obj.filename
 
                content = file_obj.file
 

	
 
                if hasattr(content, 'file'):
 
                    # non posix systems store real file under file attr
 
                    content = content.file
 

	
 
            if not content:
 
                h.flash(_('No content'), category='warning')
 
                raise HTTPFound(location=url('changeset_home', repo_name=c.repo_name,
 
                                    revision='tip'))
 
            if not filename:
 
                h.flash(_('No filename'), category='warning')
 
                raise HTTPFound(location=url('changeset_home', repo_name=c.repo_name,
 
                                    revision='tip'))
 
            # strip all crap out of file, just leave the basename
 
            filename = os.path.basename(filename)
 
            node_path = posixpath.join(location, filename)
 
            author = request.authuser.full_contact
 

	
 
            try:
 
                nodes = {
 
                    node_path: {
 
                        'content': content
 
                    }
 
                }
 
                self.scm_model.create_nodes(
 
                    user=request.authuser.user_id,
 
                    ip_addr=request.ip_addr,
 
                    repo=c.db_repo,
 
                    message=message,
 
                    nodes=nodes,
 
                    parent_cs=c.cs,
 
                    author=author,
 
                )
 

	
 
                h.flash(_('Successfully committed to %s') % node_path,
 
                        category='success')
 
            except NonRelativePathError as e:
 
                h.flash(_('Location must be relative path and must not '
 
                          'contain .. in path'), category='warning')
 
                raise HTTPFound(location=url('changeset_home', repo_name=c.repo_name,
 
                                    revision='tip'))
 
            except (NodeError, NodeAlreadyExistsError) as e:
 
                h.flash(_(e), category='error')
 
            except Exception:
 
                log.error(traceback.format_exc())
 
                h.flash(_('Error occurred during commit'), category='error')
 
            raise HTTPFound(location=url('changeset_home',
 
                                repo_name=c.repo_name, revision='tip'))
 

	
 
        return render('files/files_add.html')
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def archivefile(self, repo_name, fname):
 
        fileformat = None
 
        revision = None
 
        ext = None
 
        subrepos = request.GET.get('subrepos') == 'true'
 

	
 
        for a_type, ext_data in settings.ARCHIVE_SPECS.items():
 
            archive_spec = fname.split(ext_data[1])
 
            if len(archive_spec) == 2 and archive_spec[1] == '':
 
                fileformat = a_type or ext_data[1]
 
                revision = archive_spec[0]
 
                ext = ext_data[1]
 

	
 
        try:
 
            dbrepo = RepoModel().get_by_repo_name(repo_name)
 
            if not dbrepo.enable_downloads:
 
                return _('Downloads disabled') # TODO: do something else?
 

	
 
            if c.db_repo_scm_instance.alias == 'hg':
 
                # patch and reset hooks section of UI config to not run any
 
                # hooks on fetching archives with subrepos
 
                for k, v in c.db_repo_scm_instance._repo.ui.configitems('hooks'):
 
                    c.db_repo_scm_instance._repo.ui.setconfig('hooks', k, None)
 

	
 
            cs = c.db_repo_scm_instance.get_changeset(revision)
 
            content_type = settings.ARCHIVE_SPECS[fileformat][0]
 
        except ChangesetDoesNotExistError:
 
            return _('Unknown revision %s') % revision
 
        except EmptyRepositoryError:
 
            return _('Empty repository')
 
        except (ImproperArchiveTypeError, KeyError):
 
            return _('Unknown archive type')
 

	
 
        from kallithea import CONFIG
 
        rev_name = cs.raw_id[:12]
 
        archive_name = '%s-%s%s' % (safe_str(repo_name.replace('/', '_')),
 
                                    safe_str(rev_name), ext)
 

	
 
        archive_path = None
 
        cached_archive_path = None
 
        archive_cache_dir = CONFIG.get('archive_cache_dir')
 
        if archive_cache_dir and not subrepos: # TODO: subrepo caching?
 
            if not os.path.isdir(archive_cache_dir):
 
                os.makedirs(archive_cache_dir)
 
            cached_archive_path = os.path.join(archive_cache_dir, archive_name)
 
            if os.path.isfile(cached_archive_path):
 
                log.debug('Found cached archive in %s', cached_archive_path)
 
                archive_path = cached_archive_path
 
            else:
 
                log.debug('Archive %s is not yet cached', archive_name)
 

	
 
        if archive_path is None:
 
            # generate new archive
 
            fd, archive_path = tempfile.mkstemp()
 
            log.debug('Creating new temp archive in %s', archive_path)
 
            with os.fdopen(fd, 'wb') as stream:
 
                cs.fill_archive(stream=stream, kind=fileformat, subrepos=subrepos)
 
                # stream (and thus fd) has been closed by cs.fill_archive
 
            if cached_archive_path is not None:
 
                # we generated the archive - move it to cache
 
                log.debug('Storing new archive in %s', cached_archive_path)
 
                shutil.move(archive_path, cached_archive_path)
 
                archive_path = cached_archive_path
 

	
 
        def get_chunked_archive(archive_path):
 
            stream = open(archive_path, 'rb')
 
            while True:
 
                data = stream.read(16 * 1024)
 
                if not data:
 
                    break
 
                yield data
 
            stream.close()
 
            if archive_path != cached_archive_path:
 
                log.debug('Destroying temp archive %s', archive_path)
 
                os.remove(archive_path)
 

	
 
        action_logger(user=request.authuser,
 
                      action='user_downloaded_archive:%s' % (archive_name),
 
                      repo=repo_name, ipaddr=request.ip_addr, commit=True)
 

	
 
        response.content_disposition = str('attachment; filename=%s' % (archive_name))
 
        response.content_type = str(content_type)
 
        return get_chunked_archive(archive_path)
 

	
 
    @LoginRequired(allow_default_user=True)
 
    @HasRepoPermissionLevelDecorator('read')
 
    def diff(self, repo_name, f_path):
 
        ignore_whitespace = request.GET.get('ignorews') == '1'
 
        line_context = safe_int(request.GET.get('context'), 3)
 
        diff2 = request.GET.get('diff2', '')
 
        diff1 = request.GET.get('diff1', '') or diff2
 
        c.action = request.GET.get('diff')
 
        c.no_changes = diff1 == diff2
 
        c.f_path = f_path
 
        c.big_diff = False
 
        fulldiff = request.GET.get('fulldiff')
 
        c.anchor_url = anchor_url
 
        c.ignorews_url = _ignorews_url
 
        c.context_url = _context_url
 
        c.changes = OrderedDict()
 
        c.changes[diff2] = []
 

	
 
        # special case if we want a show rev only, it's impl here
 
        # to reduce JS and callbacks
 

	
 
        if request.GET.get('show_rev'):
 
            if str2bool(request.GET.get('annotate', 'False')):
 
                _url = url('files_annotate_home', repo_name=c.repo_name,
 
                           revision=diff1, f_path=c.f_path)
 
            else:
 
                _url = url('files_home', repo_name=c.repo_name,
 
                           revision=diff1, f_path=c.f_path)
 

	
 
            raise HTTPFound(location=_url)
 
        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 (RepositoryError, NodeError):
 
            log.error(traceback.format_exc())
 
            raise HTTPFound(location=url('files_home', repo_name=c.repo_name,
 
                                f_path=f_path))
 

	
 
        if c.action == 'download':
 
            raw_diff = diffs.get_gitdiff(node1, node2,
 
                                      ignore_whitespace=ignore_whitespace,
 
                                      context=line_context)
 
            diff_name = '%s_vs_%s.diff' % (diff1, diff2)
 
            response.content_type = 'text/plain'
 
            response.content_disposition = (
 
                'attachment; filename=%s' % diff_name
 
            )
 
            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:
 
            # TODO: loop over chs.branches ... but that will not give all the bogus None branches for Git ...
 
            _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}
kallithea/lib/celerylib/tasks.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.celerylib.tasks
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
Kallithea task modules, containing all task that suppose to be run
 
by celery daemon
 

	
 
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: Oct 6, 2010
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 
import os
 
import traceback
 
import logging
 
import rfc822
 

	
 
from collections import OrderedDict
 
from time import mktime
 
from operator import itemgetter
 
from string import lower
 

	
 
from tg import config
 

	
 
from kallithea import CELERY_ON
 
from kallithea.lib import celerylib
 
from kallithea.lib.helpers import person
 
from kallithea.lib.rcmail.smtp_mailer import SmtpMailer
 
from kallithea.lib.utils import action_logger
 
from kallithea.lib.utils2 import str2bool
 
from kallithea.lib.vcs.utils import author_email
 
from kallithea.lib.compat import json, OrderedDict
 
from kallithea.lib.compat import json
 
from kallithea.lib.hooks import log_create_repository
 

	
 
from kallithea.model.db import Statistics, RepoGroup, Repository, User
 

	
 

	
 
__all__ = ['whoosh_index', 'get_commits_stats', 'send_email']
 

	
 

	
 
log = logging.getLogger(__name__)
 

	
 

	
 
@celerylib.task
 
@celerylib.locked_task
 
@celerylib.dbsession
 
def whoosh_index(repo_location, full_index):
 
    from kallithea.lib.indexers.daemon import WhooshIndexingDaemon
 
    celerylib.get_session() # initialize database connection
 

	
 
    index_location = config['index_dir']
 
    WhooshIndexingDaemon(index_location=index_location,
 
                         repo_location=repo_location) \
 
                         .run(full_index=full_index)
 

	
 

	
 
@celerylib.task
 
@celerylib.dbsession
 
def get_commits_stats(repo_name, ts_min_y, ts_max_y, recurse_limit=100):
 
    DBS = celerylib.get_session()
 
    lockkey = celerylib.__get_lockkey('get_commits_stats', repo_name, ts_min_y,
 
                            ts_max_y)
 
    lockkey_path = config['app_conf']['cache_dir']
 

	
 
    log.info('running task with lockkey %s', lockkey)
 

	
 
    try:
 
        lock = celerylib.DaemonLock(os.path.join(lockkey_path, lockkey))
 

	
 
        # for js data compatibility cleans the key for person from '
 
        akc = lambda k: person(k).replace('"', "")
 

	
 
        co_day_auth_aggr = {}
 
        commits_by_day_aggregate = {}
 
        repo = Repository.get_by_repo_name(repo_name)
 
        if repo is None:
 
            return True
 

	
 
        repo = repo.scm_instance
 
        repo_size = repo.count()
 
        # return if repo have no revisions
 
        if repo_size < 1:
 
            lock.release()
 
            return True
 

	
 
        skip_date_limit = True
 
        parse_limit = int(config['app_conf'].get('commit_parse_limit'))
 
        last_rev = None
 
        last_cs = None
 
        timegetter = itemgetter('time')
 

	
 
        dbrepo = DBS.query(Repository) \
 
            .filter(Repository.repo_name == repo_name).scalar()
 
        cur_stats = DBS.query(Statistics) \
 
            .filter(Statistics.repository == dbrepo).scalar()
 

	
 
        if cur_stats is not None:
 
            last_rev = cur_stats.stat_on_revision
 

	
 
        if last_rev == repo.get_changeset().revision and repo_size > 1:
 
            # pass silently without any work if we're not on first revision or
 
            # current state of parsing revision(from db marker) is the
 
            # last revision
 
            lock.release()
 
            return True
 

	
 
        if cur_stats:
 
            commits_by_day_aggregate = OrderedDict(json.loads(
 
                                        cur_stats.commit_activity_combined))
 
            co_day_auth_aggr = json.loads(cur_stats.commit_activity)
 

	
 
        log.debug('starting parsing %s', parse_limit)
 
        lmktime = mktime
 

	
 
        last_rev = last_rev + 1 if last_rev >= 0 else 0
 
        log.debug('Getting revisions from %s to %s',
 
             last_rev, last_rev + parse_limit
 
        )
 
        for cs in repo[last_rev:last_rev + parse_limit]:
 
            log.debug('parsing %s', cs)
 
            last_cs = cs  # remember last parsed changeset
 
            k = lmktime([cs.date.timetuple()[0], cs.date.timetuple()[1],
 
                          cs.date.timetuple()[2], 0, 0, 0, 0, 0, 0])
 

	
 
            if akc(cs.author) in co_day_auth_aggr:
 
                try:
 
                    l = [timegetter(x) for x in
 
                         co_day_auth_aggr[akc(cs.author)]['data']]
 
                    time_pos = l.index(k)
 
                except ValueError:
 
                    time_pos = None
 

	
 
                if time_pos >= 0 and time_pos is not None:
 

	
 
                    datadict = \
 
                        co_day_auth_aggr[akc(cs.author)]['data'][time_pos]
 

	
 
                    datadict["commits"] += 1
 
                    datadict["added"] += len(cs.added)
 
                    datadict["changed"] += len(cs.changed)
 
                    datadict["removed"] += len(cs.removed)
 

	
 
                else:
 
                    if k >= ts_min_y and k <= ts_max_y or skip_date_limit:
 

	
 
                        datadict = {"time": k,
 
                                    "commits": 1,
 
                                    "added": len(cs.added),
 
                                    "changed": len(cs.changed),
 
                                    "removed": len(cs.removed),
 
                                   }
 
                        co_day_auth_aggr[akc(cs.author)]['data'] \
 
                            .append(datadict)
 

	
 
            else:
 
                if k >= ts_min_y and k <= ts_max_y or skip_date_limit:
 
                    co_day_auth_aggr[akc(cs.author)] = {
 
                                        "label": akc(cs.author),
 
                                        "data": [{"time":k,
 
                                                 "commits":1,
 
                                                 "added":len(cs.added),
 
                                                 "changed":len(cs.changed),
 
                                                 "removed":len(cs.removed),
 
                                                 }],
 
                                        "schema": ["commits"],
 
                                        }
 

	
 
            # gather all data by day
 
            if k in commits_by_day_aggregate:
 
                commits_by_day_aggregate[k] += 1
 
            else:
 
                commits_by_day_aggregate[k] = 1
 

	
 
        overview_data = sorted(commits_by_day_aggregate.items(),
 
                               key=itemgetter(0))
 

	
 
        if not co_day_auth_aggr:
 
            co_day_auth_aggr[akc(repo.contact)] = {
 
                "label": akc(repo.contact),
 
                "data": [0, 1],
 
                "schema": ["commits"],
 
            }
 

	
 
        stats = cur_stats if cur_stats else Statistics()
 
        stats.commit_activity = json.dumps(co_day_auth_aggr)
 
        stats.commit_activity_combined = json.dumps(overview_data)
 

	
 
        log.debug('last revision %s', last_rev)
 
        leftovers = len(repo.revisions[last_rev:])
 
        log.debug('revisions to parse %s', leftovers)
 

	
 
        if last_rev == 0 or leftovers < parse_limit:
 
            log.debug('getting code trending stats')
 
            stats.languages = json.dumps(__get_codes_stats(repo_name))
 

	
 
        try:
 
            stats.repository = dbrepo
 
            stats.stat_on_revision = last_cs.revision if last_cs else 0
 
            DBS.add(stats)
 
            DBS.commit()
 
        except:
 
            log.error(traceback.format_exc())
 
            DBS.rollback()
 
            lock.release()
 
            return False
 

	
 
        # final release
 
        lock.release()
 

	
 
        # execute another task if celery is enabled
 
        if len(repo.revisions) > 1 and CELERY_ON and recurse_limit > 0:
 
            get_commits_stats(repo_name, ts_min_y, ts_max_y, recurse_limit - 1)
 
        elif recurse_limit <= 0:
 
            log.debug('Not recursing - limit has been reached')
 
        else:
 
            log.debug('Not recursing')
 
    except celerylib.LockHeld:
 
        log.info('Task with key %s already running', lockkey)
 
        return 'Task with key %s already running' % lockkey
 

	
 

	
 
@celerylib.task
 
@celerylib.dbsession
 
def send_email(recipients, subject, body='', html_body='', headers=None, author=None):
 
    """
 
    Sends an email with defined parameters from the .ini files.
 

	
 
    :param recipients: list of recipients, if this is None, the defined email
 
        address from field 'email_to' and all admins is used instead
 
    :param subject: subject of the mail
 
    :param body: body of the mail
 
    :param html_body: html version of body
 
    :param headers: dictionary of prepopulated e-mail headers
 
    :param author: User object of the author of this mail, if known and relevant
 
    """
 
    assert isinstance(recipients, list), recipients
 
    if headers is None:
 
        headers = {}
 
    else:
 
        # do not modify the original headers object passed by the caller
 
        headers = headers.copy()
 

	
 
    email_config = config
 
    email_prefix = email_config.get('email_prefix', '')
 
    if email_prefix:
 
        subject = "%s %s" % (email_prefix, subject)
 

	
 
    if not recipients:
 
        # if recipients are not defined we send to email_config + all admins
 
        recipients = [u.email for u in User.query()
 
                      .filter(User.admin == True).all()]
 
        if email_config.get('email_to') is not None:
 
            recipients += email_config.get('email_to').split(',')
 

	
 
        # If there are still no recipients, there are no admins and no address
 
        # configured in email_to, so return.
 
        if not recipients:
 
            log.error("No recipients specified and no fallback available.")
 
            return False
 

	
 
        log.warning("No recipients specified for '%s' - sending to admins %s", subject, ' '.join(recipients))
 

	
 
    # SMTP sender
 
    envelope_from = email_config.get('app_email_from', 'Kallithea')
 
    # 'From' header
 
    if author is not None:
 
        # set From header based on author but with a generic e-mail address
 
        # In case app_email_from is in "Some Name <e-mail>" format, we first
 
        # extract the e-mail address.
 
        envelope_addr = author_email(envelope_from)
 
        headers['From'] = '"%s" <%s>' % (
 
            rfc822.quote('%s (no-reply)' % author.full_name_or_username),
 
            envelope_addr)
 

	
 
    user = email_config.get('smtp_username')
 
    passwd = email_config.get('smtp_password')
 
    mail_server = email_config.get('smtp_server')
 
    mail_port = email_config.get('smtp_port')
 
    tls = str2bool(email_config.get('smtp_use_tls'))
 
    ssl = str2bool(email_config.get('smtp_use_ssl'))
 
    debug = str2bool(email_config.get('debug'))
 
    smtp_auth = email_config.get('smtp_auth')
 

	
 
    logmsg = ("Mail details:\n"
 
              "recipients: %s\n"
 
              "headers: %s\n"
 
              "subject: %s\n"
 
              "body:\n%s\n"
 
              "html:\n%s\n"
 
              % (' '.join(recipients), headers, subject, body, html_body))
 

	
 
    if mail_server:
 
        log.debug("Sending e-mail. " + logmsg)
 
    else:
 
        log.error("SMTP mail server not configured - cannot send e-mail.")
 
        log.warning(logmsg)
 
        return False
 

	
 
    try:
 
        m = SmtpMailer(envelope_from, user, passwd, mail_server, smtp_auth,
 
                       mail_port, ssl, tls, debug=debug)
 
        m.send(recipients, subject, body, html_body, headers=headers)
 
    except:
 
        log.error('Mail sending failed')
 
        log.error(traceback.format_exc())
 
        return False
 
    return True
 

	
 

	
 
@celerylib.task
 
@celerylib.dbsession
 
def create_repo(form_data, cur_user):
 
    from kallithea.model.repo import RepoModel
 
    from kallithea.model.db import Setting
 

	
 
    DBS = celerylib.get_session()
 

	
 
    cur_user = User.guess_instance(cur_user)
 

	
 
    owner = cur_user
 
    repo_name = form_data['repo_name']
 
    repo_name_full = form_data['repo_name_full']
 
    repo_type = form_data['repo_type']
 
    description = form_data['repo_description']
 
    private = form_data['repo_private']
 
    clone_uri = form_data.get('clone_uri')
 
    repo_group = form_data['repo_group']
 
    landing_rev = form_data['repo_landing_rev']
 
    copy_fork_permissions = form_data.get('copy_permissions')
 
    copy_group_permissions = form_data.get('repo_copy_permissions')
 
    fork_of = form_data.get('fork_parent_id')
 
    state = form_data.get('repo_state', Repository.STATE_PENDING)
 

	
 
    # repo creation defaults, private and repo_type are filled in form
 
    defs = Setting.get_default_repo_settings(strip_prefix=True)
 
    enable_statistics = defs.get('repo_enable_statistics')
 
    enable_downloads = defs.get('repo_enable_downloads')
 

	
 
    try:
 
        repo = RepoModel()._create_repo(
 
            repo_name=repo_name_full,
 
            repo_type=repo_type,
 
            description=description,
 
            owner=owner,
 
            private=private,
 
            clone_uri=clone_uri,
 
            repo_group=repo_group,
 
            landing_rev=landing_rev,
 
            fork_of=fork_of,
 
            copy_fork_permissions=copy_fork_permissions,
 
            copy_group_permissions=copy_group_permissions,
 
            enable_statistics=enable_statistics,
 
            enable_downloads=enable_downloads,
 
            state=state
 
        )
 

	
 
        action_logger(cur_user, 'user_created_repo',
 
                      form_data['repo_name_full'], '')
 

	
 
        DBS.commit()
 
        # now create this repo on Filesystem
 
        RepoModel()._create_filesystem_repo(
 
            repo_name=repo_name,
 
            repo_type=repo_type,
 
            repo_group=RepoGroup.guess_instance(repo_group),
 
            clone_uri=clone_uri,
 
        )
 
        repo = Repository.get_by_repo_name(repo_name_full)
 
        log_create_repository(repo.get_dict(), created_by=owner.username)
 

	
 
        # update repo changeset caches initially
 
        repo.update_changeset_cache()
 

	
 
        # set new created state
 
        repo.set_state(Repository.STATE_CREATED)
 
        DBS.commit()
 
    except Exception as e:
 
        log.warning('Exception %s occurred when forking repository, '
 
                    'doing cleanup...' % e)
 
        # rollback things manually !
 
        repo = Repository.get_by_repo_name(repo_name_full)
 
        if repo:
 
            Repository.delete(repo.repo_id)
 
            DBS.commit()
 
            RepoModel()._delete_filesystem_repo(repo)
 
        raise
 

	
 
    return True
 

	
 

	
 
@celerylib.task
 
@celerylib.dbsession
 
def create_repo_fork(form_data, cur_user):
 
    """
 
    Creates a fork of repository using interval VCS methods
 

	
 
    :param form_data:
 
    :param cur_user:
 
    """
 
    from kallithea.model.repo import RepoModel
 

	
 
    DBS = celerylib.get_session()
 

	
 
    base_path = Repository.base_path()
 
    cur_user = User.guess_instance(cur_user)
 

	
 
    repo_name = form_data['repo_name']  # fork in this case
 
    repo_name_full = form_data['repo_name_full']
 

	
 
    repo_type = form_data['repo_type']
 
    owner = cur_user
 
    private = form_data['private']
 
    clone_uri = form_data.get('clone_uri')
 
    repo_group = form_data['repo_group']
 
    landing_rev = form_data['landing_rev']
 
    copy_fork_permissions = form_data.get('copy_permissions')
 

	
 
    try:
 
        fork_of = Repository.guess_instance(form_data.get('fork_parent_id'))
 

	
 
        RepoModel()._create_repo(
 
            repo_name=repo_name_full,
 
            repo_type=repo_type,
 
            description=form_data['description'],
 
            owner=owner,
 
            private=private,
 
            clone_uri=clone_uri,
 
            repo_group=repo_group,
 
            landing_rev=landing_rev,
 
            fork_of=fork_of,
 
            copy_fork_permissions=copy_fork_permissions
 
        )
 
        action_logger(cur_user, 'user_forked_repo:%s' % repo_name_full,
 
                      fork_of.repo_name, '')
 
        DBS.commit()
 

	
 
        source_repo_path = os.path.join(base_path, fork_of.repo_name)
 

	
 
        # now create this repo on Filesystem
 
        RepoModel()._create_filesystem_repo(
 
            repo_name=repo_name,
 
            repo_type=repo_type,
 
            repo_group=RepoGroup.guess_instance(repo_group),
 
            clone_uri=source_repo_path,
 
        )
 
        repo = Repository.get_by_repo_name(repo_name_full)
 
        log_create_repository(repo.get_dict(), created_by=owner.username)
 

	
 
        # update repo changeset caches initially
 
        repo.update_changeset_cache()
 

	
 
        # set new created state
 
        repo.set_state(Repository.STATE_CREATED)
 
        DBS.commit()
 
    except Exception as e:
 
        log.warning('Exception %s occurred when forking repository, '
 
                    'doing cleanup...' % e)
 
        # rollback things manually !
 
        repo = Repository.get_by_repo_name(repo_name_full)
 
        if repo:
 
            Repository.delete(repo.repo_id)
 
            DBS.commit()
 
            RepoModel()._delete_filesystem_repo(repo)
 
        raise
 

	
 
    return True
 

	
 

	
 
def __get_codes_stats(repo_name):
 
    from kallithea.config.conf import LANGUAGES_EXTENSIONS_MAP
 
    repo = Repository.get_by_repo_name(repo_name).scm_instance
 

	
 
    tip = repo.get_changeset()
 
    code_stats = {}
 

	
 
    def aggregate(cs):
 
        for f in cs[2]:
 
            ext = lower(f.extension)
 
            if ext in LANGUAGES_EXTENSIONS_MAP.keys() and not f.is_binary:
 
                if ext in code_stats:
 
                    code_stats[ext] += 1
 
                else:
 
                    code_stats[ext] = 1
 

	
 
    map(aggregate, tip.walk('/'))
 

	
 
    return code_stats or {}
kallithea/lib/compat.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.compat
 
~~~~~~~~~~~~~~~~~~~~
 

	
 
Python backward compatibility functions and common libs
 

	
 

	
 
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: Oct 7, 2011
 
:author: marcink
 
:copyright: (c) 2013 RhodeCode GmbH, and others.
 
:license: GPLv3, see LICENSE.md for more details.
 
"""
 

	
 

	
 
import sys
 
import os
 
import functools
 

	
 
#==============================================================================
 
# json
 
#==============================================================================
 
from kallithea.lib.ext_json import json
 

	
 
# alias for formatted json
 
formatted_json = functools.partial(json.dumps, indent=4, sort_keys=True)
 

	
 

	
 
#==============================================================================
 
# unittest
 
#==============================================================================
 
if sys.version_info >= (2, 7):
 
    import unittest
 
else:
 
    import unittest2 as unittest
 

	
 

	
 
#==============================================================================
 
# OrderedDict - Python 2.7 could perhaps use collections.OrderedDict
 
#==============================================================================
 

	
 
# Python Software Foundation License
 

	
 

	
 
# XXX: it feels like using the class with "is" and "is not" instead of "==" and
 
# "!=" should be faster.
 
class _Nil(object):
 

	
 
    def __repr__(self):
 
        return "nil"
 

	
 
    def __eq__(self, other):
 
        if (isinstance(other, _Nil)):
 
            return True
 
        else:
 
            return NotImplemented
 

	
 
    def __ne__(self, other):
 
        if (isinstance(other, _Nil)):
 
            return False
 
        else:
 
            return NotImplemented
 

	
 

	
 
_nil = _Nil()
 

	
 

	
 
class OrderedDict(dict):
 
    """Ordered dict data structure, with O(1) complexity for dict operations
 
    that modify one element.
 

	
 
    Overwriting values doesn't change their original sequential order.
 
    """
 

	
 
    def __init__(self, data=(), **kwds):
 
        """This doesn't accept keyword initialization as normal dicts to avoid
 
        a trap - inside a function or method the keyword args are accessible
 
        only as a dict, without a defined order, so their original order is
 
        lost.
 
        """
 
        if kwds:
 
            raise TypeError("__init__() of ordered dict takes no keyword "
 
                            "arguments to avoid an ordering trap.")
 
        dict.__init__(self)
 
        # If you give a normal dict, then the order of elements is undefined
 
        if hasattr(data, "iteritems"):
 
            for key, val in data.iteritems():
 
                self[key] = val
 
        else:
 
            for key, val in data:
 
                self[key] = val
 

	
 
    # Double-linked list header
 
    def _get_lh(self):
 
        if not hasattr(self, '_lh'):
 
            dict.__setattr__(self, '_lh', _nil)
 
        return dict.__getattribute__(self, '_lh')
 

	
 
    def _set_lh(self, val):
 
        dict.__setattr__(self, '_lh', val)
 

	
 
    lh = property(_get_lh, _set_lh)
 

	
 
    # Double-linked list tail
 
    def _get_lt(self):
 
        if not hasattr(self, '_lt'):
 
            dict.__setattr__(self, '_lt', _nil)
 
        return dict.__getattribute__(self, '_lt')
 

	
 
    def _set_lt(self, val):
 
        dict.__setattr__(self, '_lt', val)
 

	
 
    lt = property(_get_lt, _set_lt)
 

	
 
    def __getitem__(self, key):
 
        return dict.__getitem__(self, key)[1]
 

	
 
    def __setitem__(self, key, val):
 
        try:
 
            dict.__getitem__(self, key)[1] = val
 
        except KeyError:
 
            new = [dict.__getattribute__(self, 'lt'), val, _nil]
 
            dict.__setitem__(self, key, new)
 
            if dict.__getattribute__(self, 'lt') == _nil:
 
                dict.__setattr__(self, 'lh', key)
 
            else:
 
                dict.__getitem__(
 
                    self, dict.__getattribute__(self, 'lt'))[2] = key
 
            dict.__setattr__(self, 'lt', key)
 

	
 
    def __delitem__(self, key):
 
        pred, _, succ = dict.__getitem__(self, key)
 
        if pred == _nil:
 
            dict.__setattr__(self, 'lh', succ)
 
        else:
 
            dict.__getitem__(self, pred)[2] = succ
 
        if succ == _nil:
 
            dict.__setattr__(self, 'lt', pred)
 
        else:
 
            dict.__getitem__(self, succ)[0] = pred
 
        dict.__delitem__(self, key)
 

	
 
    def __contains__(self, key):
 
        return key in self.keys()
 

	
 
    def __len__(self):
 
        return len(self.keys())
 

	
 
    def __str__(self):
 
        pairs = ("%r: %r" % (k, v) for k, v in self.iteritems())
 
        return "{%s}" % ", ".join(pairs)
 

	
 
    def __repr__(self):
 
        if self:
 
            pairs = ("(%r, %r)" % (k, v) for k, v in self.iteritems())
 
            return "odict([%s])" % ", ".join(pairs)
 
        else:
 
            return "odict()"
 

	
 
    def get(self, k, x=None):
 
        if k in self:
 
            return dict.__getitem__(self, k)[1]
 
        else:
 
            return x
 

	
 
    def __iter__(self):
 
        curr_key = dict.__getattribute__(self, 'lh')
 
        while curr_key != _nil:
 
            yield curr_key
 
            curr_key = dict.__getitem__(self, curr_key)[2]
 

	
 
    iterkeys = __iter__
 

	
 
    def keys(self):
 
        return list(self.iterkeys())
 

	
 
    def itervalues(self):
 
        curr_key = dict.__getattribute__(self, 'lh')
 
        while curr_key != _nil:
 
            _, val, curr_key = dict.__getitem__(self, curr_key)
 
            yield val
 

	
 
    def values(self):
 
        return list(self.itervalues())
 

	
 
    def iteritems(self):
 
        curr_key = dict.__getattribute__(self, 'lh')
 
        while curr_key != _nil:
 
            _, val, next_key = dict.__getitem__(self, curr_key)
 
            yield curr_key, val
 
            curr_key = next_key
 

	
 
    def items(self):
 
        return list(self.iteritems())
 

	
 
    def sort(self, cmp=None, key=None, reverse=False):
 
        items = [(k, v) for k, v in self.items()]
 
        if cmp is not None:
 
            items = sorted(items, cmp=cmp)
 
        elif key is not None:
 
            items = sorted(items, key=key)
 
        else:
 
            items = sorted(items, key=lambda x: x[1])
 
        if reverse:
 
            items.reverse()
 
        self.clear()
 
        self.__init__(items)
 

	
 
    def clear(self):
 
        dict.clear(self)
 
        dict.__setattr__(self, 'lh', _nil)
 
        dict.__setattr__(self, 'lt', _nil)
 

	
 
    def copy(self):
 
        return self.__class__(self)
 

	
 
    def update(self, data=(), **kwds):
 
        if kwds:
 
            raise TypeError("update() of ordered dict takes no keyword "
 
                            "arguments to avoid an ordering trap.")
 
        if hasattr(data, "iteritems"):
 
            data = data.iteritems()
 
        for key, val in data:
 
            self[key] = val
 

	
 
    def setdefault(self, k, x=None):
 
        try:
 
            return self[k]
 
        except KeyError:
 
            self[k] = x
 
            return x
 

	
 
    def pop(self, k, x=_nil):
 
        try:
 
            val = self[k]
 
            del self[k]
 
            return val
 
        except KeyError:
 
            if x == _nil:
 
                raise
 
            return x
 

	
 
    def popitem(self):
 
        try:
 
            key = dict.__getattribute__(self, 'lt')
 
            return key, self.pop(key)
 
        except KeyError:
 
            raise KeyError("'popitem(): ordered dictionary is empty'")
 

	
 
    def riterkeys(self):
 
        """To iterate on keys in reversed order.
 
        """
 
        curr_key = dict.__getattribute__(self, 'lt')
 
        while curr_key != _nil:
 
            yield curr_key
 
            curr_key = dict.__getitem__(self, curr_key)[0]
 

	
 
    __reversed__ = riterkeys
 

	
 
    def rkeys(self):
 
        """List of the keys in reversed order.
 
        """
 
        return list(self.riterkeys())
 

	
 
    def ritervalues(self):
 
        """To iterate on values in reversed order.
 
        """
 
        curr_key = dict.__getattribute__(self, 'lt')
 
        while curr_key != _nil:
 
            curr_key, val, _ = dict.__getitem__(self, curr_key)
 
            yield val
 

	
 
    def rvalues(self):
 
        """List of the values in reversed order.
 
        """
 
        return list(self.ritervalues())
 

	
 
    def riteritems(self):
 
        """To iterate on (key, value) in reversed order.
 
        """
 
        curr_key = dict.__getattribute__(self, 'lt')
 
        while curr_key != _nil:
 
            pred_key, val, _ = dict.__getitem__(self, curr_key)
 
            yield curr_key, val
 
            curr_key = pred_key
 

	
 
    def ritems(self):
 
        """List of the (key, value) in reversed order.
 
        """
 
        return list(self.riteritems())
 

	
 
    def firstkey(self):
 
        if self:
 
            return dict.__getattribute__(self, 'lh')
 
        else:
 
            raise KeyError("'firstkey(): ordered dictionary is empty'")
 

	
 
    def lastkey(self):
 
        if self:
 
            return dict.__getattribute__(self, 'lt')
 
        else:
 
            raise KeyError("'lastkey(): ordered dictionary is empty'")
 

	
 
    def as_dict(self):
 
        return dict(self.items())
 

	
 
    def _repr(self):
 
        """_repr(): low level repr of the whole data contained in the odict.
 
        Useful for debugging.
 
        """
 
        form = "odict low level repr lh,lt,data: %r, %r, %s"
 
        return form % (dict.__getattribute__(self, 'lh'),
 
                       dict.__getattribute__(self, 'lt'),
 
                       dict.__repr__(self))
 

	
 

	
 
#==============================================================================
 
# OrderedSet
 
#==============================================================================
 
from sqlalchemy.util import OrderedSet
 

	
 

	
 
#==============================================================================
 
# Hybrid property/method
 
#==============================================================================
 
from sqlalchemy.ext.hybrid import hybrid_property
 

	
 

	
 
#==============================================================================
 
# kill
 
#==============================================================================
 
if os.name == 'nt': # Windows
 
    import ctypes
 

	
 
    def kill(pid, sig):
 
        """kill function for Win32"""
 
        kernel32 = ctypes.windll.kernel32
 
        handle = kernel32.OpenProcess(1, 0, pid)
 
        return (0 != kernel32.TerminateProcess(handle, 0))
 

	
 
else:
 
    kill = os.kill
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 os
 
import re
 
import time
 
import errno
 
import urllib
 
import urllib2
 
import logging
 
import posixpath
 
from collections import OrderedDict
 

	
 
from dulwich.objects import Tag
 
from dulwich.repo import Repo, NotGitRepository
 
from dulwich.config import ConfigFile
 

	
 
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 safe_str, safe_unicode, makedate, date_fromtimestamp
 
from kallithea.lib.vcs.utils.lazy import LazyProperty
 
from kallithea.lib.vcs.utils.ordered_dict import OrderedDict
 
from kallithea.lib.vcs.utils.paths import abspath, get_user_home
 

	
 
from kallithea.lib.vcs.utils.hgcompat import (
 
    hg_url, httpbasicauthhandler, httpdigestauthhandler
 
)
 

	
 
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, **opts):
 
        """
 
        Runs given ``cmd`` as git command and returns tuple
 
        (stdout, stderr).
 

	
 
        :param cmd: git command to be executed
 
        :param opts: env options to pass into Subprocess command
 
        """
 

	
 
        if '_bare' in opts:
 
            _copts = []
 
            del opts['_bare']
 
        else:
 
            _copts = ['-c', 'core.quotepath=false', ]
 
        safe_call = False
 
        if '_safe' in opts:
 
            # no exc on failure
 
            del opts['_safe']
 
            safe_call = True
 

	
 
        assert isinstance(cmd, list), cmd
 

	
 
        gitenv = os.environ
 
        # need to clean fix GIT_DIR !
 
        if 'GIT_DIR' in gitenv:
 
            del gitenv['GIT_DIR']
 
        gitenv['GIT_CONFIG_NOGLOBAL'] = '1'
 

	
 
        _git_path = settings.GIT_EXECUTABLE_PATH
 
        cmd = [_git_path] + _copts + cmd
 

	
 
        try:
 
            _opts = dict(
 
                env=gitenv,
 
                shell=False,
 
            )
 
            _opts.update(opts)
 
            p = subprocessio.SubprocessIOChunker(cmd, **_opts)
 
        except (EnvironmentError, OSError) as err:
 
            tb_err = ("Couldn't run git command (%s).\n"
 
                      "Original error was:%s\n" % (cmd, err))
 
            log.error(tb_err)
 
            if safe_call:
 
                return '', err
 
            else:
 
                raise RepositoryError(tb_err)
 

	
 
        return ''.join(p.output), ''.join(p.error)
 

	
 
    def run_git_command(self, cmd):
 
        opts = {}
 
        if os.path.isdir(self.path):
 
            opts['cwd'] = self.path
 
        return self._run_git_command(cmd, **opts)
 

	
 
    @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)
 
        test_uri, authinfo = url_obj.authinfo()
 
        url_obj.passwd = '*****'
 
        cleaned_uri = str(url_obj)
 

	
 
        if not test_uri.endswith('info/refs'):
 
            test_uri = test_uri.rstrip('/') + '/info/refs'
 

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

	
 
            handlers.extend((httpbasicauthhandler(passmgr),
 
                             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, se = 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 = [x[1][0] for x in self._parsed_refs.iteritems()
 
                    if x[1][1] != 'T']
 
        return [c.commit.id for c in self._repo.get_walker(include=includes)]
 

	
 
    def _get_revision(self, revision):
 
        """
 
        For git backend we always return integer here. This way we ensure
 
        that changeset's revision attribute would become integer.
 
        """
 

	
 
        is_null = lambda o: len(o) == revision.count('0')
 

	
 
        if self._empty:
 
            raise EmptyRepositoryError("There are no changesets yet")
 

	
 
        if revision in (None, '', 'tip', 'HEAD', 'head', -1):
 
            return self.revisions[-1]
 

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

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

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

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

	
 
        # Ensure we return full id
 
        if not SHA_PATTERN.match(str(revision)):
 
            raise ChangesetDoesNotExistError("Given revision %s not recognized"
 
                % revision)
 
        return revision
 

	
 
    def get_ref_revision(self, ref_type, ref_name):
 
        """
 
        Returns ``MercurialChangeset`` 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):
 
        undefined_description = u'unknown'
 
        _desc = self._repo.get_description()
 
        return safe_unicode(_desc or undefined_description)
 

	
 
    @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], x[1][0])
 
                     for x in self._parsed_refs.iteritems() if x[1][1] == '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], x[1][0])
 
                 for x in self._parsed_refs.iteritems() if x[1][1] == '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["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)
 
        tagpath = posixpath.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 = [('refs/heads/', 'H'),
 
                ('refs/remotes/origin/', 'RH'),
 
                ('refs/tags/', 'T')]
 
        _refs = {}
 
        for ref, sha in refs.iteritems():
 
            for k, type_ in keys:
 
                if ref.startswith(k):
 
                    _key = ref[len(k):]
 
                    if type_ == '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 ['refs/heads/', 'refs/remotes/origin/']:
 
                if key.startswith(ref_key):
 
                    n = key[len(ref_key):]
 
                    if n not in ['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)[0].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 = reversed(revs)
 
        return CollectionGenerator(self, revs)
 

	
 
    def get_diff(self, rev1, rev2, path=None, 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``. 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)
 
        # TODO: don't ignore stderr
 
        # If we used 'show' command, strip first few lines (until actual diff
 
        # starts)
 
        if rev1 == self.EMPTY_CHANGESET:
 
            parts = stdout.split('\ndiff ', 1)
 
            if len(parts) > 1:
 
                stdout = '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, se = 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/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 os
 
import time
 
import urllib
 
import urllib2
 
import logging
 
import datetime
 
from collections import OrderedDict
 

	
 
from kallithea.lib.vcs.backends.base import BaseRepository, CollectionGenerator
 

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

	
 
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()
 
        # 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, tip, isclosed in sorted(self._repo.branchmap().iterbranches()):
 
            if isclosed:
 
                if closed:
 
                    bt[safe_unicode(bn)] = hex(tip)
 
            else:
 
                if normal:
 
                    bt[safe_unicode(bn)] = hex(tip)
 

	
 
        return bt
 

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

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

	
 
        sortkey = lambda ctx: ctx[0]  # sort by name
 
        _tags = [(safe_unicode(n), hex(h),) for n, h in
 
                 self._repo.tags().items()]
 

	
 
        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)
 
        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:
 
            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)
 
            self.tags = self._get_tags()
 
        except 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 {}
 

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

	
 
    def _get_all_revisions(self):
 

	
 
        return [self._repo[x].hex() for x in self._repo.filtered('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)
 
        else:
 
            file_filter = None
 

	
 
        return ''.join(patch.diff(self._repo, rev1, rev2, match=file_filter,
 
                          opts=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('file:'):
 
            return True
 

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

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

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

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

	
 
            handlers.extend((httpbasicauthhandler(passmgr),
 
                             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('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_str(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)
 

	
 
                # 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:
 
            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):
 
        undefined_description = u'unknown'
 
        _desc = self._repo.ui.config('web', 'description', None, untrusted=True)
 
        return safe_unicode(_desc or undefined_description)
 

	
 
    @LazyProperty
 
    def contact(self):
 
        undefined_contact = u'Unknown'
 
        return safe_unicode(get_contact(self._repo.ui.config)
 
                            or undefined_contact)
 

	
 
    @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):
 
        """
 
        Gets an ID revision given as str. This will always return a full
 
        40 char revision number
 

	
 
        :param revision: str or int or None
 
        """
 
        if isinstance(revision, unicode):
 
            revision = safe_str(revision)
 

	
 
        if self._empty:
 
            raise EmptyRepositoryError("There are no changesets yet")
 

	
 
        if revision in [-1, None]:
 
            revision = 'tip'
 

	
 
        try:
 
            if isinstance(revision, int):
 
                return self._repo[revision].hex()
 
            try:
 
                return scmutil.revsymbol(self._repo, revision).hex()
 
            except AttributeError: # revsymbol was introduced in Mercurial 4.6
 
                return self._repo[revision].hex()
 
        except (IndexError, ValueError, 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:
 
            msg = ("Revision %s:%s does not exist for %s" % (ref_type, ref_name, self.name))
 
            raise ChangesetDoesNotExistError(msg)
 
        if revs:
 
            try:
 
                revision = revs.last()
 
            except AttributeError:
 
                # removed in hg 3.2
 
                revision = revs[-1]
 
        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("web", "allow_archive",
 
                                         untrusted=True)
 
        for i in [('zip', '.zip'), ('gz', '.tar.gz'), ('bz2', '.tar.bz2')]:
 
            if i[0] in allowed or self._repo.ui.configbool("web",
 
                                                           "allow" + i[0],
 
                                                           untrusted=True):
 
                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 = "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 = self.revisions.index(start_raw_id) if start else None
 
        end_raw_id = self._get_revision(end)
 
        end_pos = self.revisions.index(end_raw_id) if end else None
 

	
 
        if None not in [start, end] 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.keys():
 
            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('branch("%s")' % safe_str(branch_name))
 
        if start_date:
 
            filter_.append('date(">%s")' % start_date)
 
        if end_date:
 
            filter_.append('date("<%s")' % end_date)
 
        if filter_ or max_revisions:
 
            if filter_:
 
                revspec = ' and '.join(filter_)
 
            else:
 
                revspec = 'all()'
 
            if max_revisions:
 
                revspec = 'limit(%s, %s)' % (revspec, max_revisions)
 
            revisions = 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 = reversed(revs)
 

	
 
        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)
 
        try:
 
            # hg 3.2 moved push / pull to exchange module
 
            from mercurial import exchange
 
            exchange.pull(self._repo, other, heads=None, force=None)
 
        except ImportError:
 
            self._repo.pull(other, heads=None, force=None)
 
        except 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()
 
            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/utils/ordered_dict.py
Show inline comments
 
deleted file
0 comments (0 inline, 0 general)