kallithea Changeset - e7b6a0ce123c

Changeset - e7b6a0ce123c

Parent rev.

Child rev.

[Not reviewed]

default

0 6 0

Mads Kiilerich - 6 years ago 2019-12-27 01:43:46
mads@kiilerich.com

Grafted from: ac456490cdda

cleanup: minor formatting

6 files changed with 5 insertions and 9 deletions:

kallithea/lib/utils.py

kallithea/lib/utils2.py

kallithea/lib/vcs/backends/git/inmemory.py

kallithea/lib/vcs/backends/git/repository.py

kallithea/lib/vcs/subprocessio.py

kallithea/lib/vcs/utils/__init__.py

0 comments (0 inline, 0 general)

kallithea/lib/utils.py

➞

Show inline comments

 # -*- coding: utf-8 -*-
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
 # the Free Software Foundation, either version 3 of the License, or
 # (at your option) any later version.
+#
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
+#
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 """
 kallithea.lib.utils
 ~~~~~~~~~~~~~~~~~~~
 Utilities library for Kallithea
 This file was forked by the Kallithea project in July 2014.
 Original author and date, and relevant copyright and licensing information is below:
 :created_on: Apr 18, 2010
 :author: marcink
 :copyright: (c) 2013 RhodeCode GmbH, and others.
 :license: GPLv3, see LICENSE.md for more details.
 """
 import datetime
 import logging
 import os
 import re
 import sys
 import traceback
 from distutils.version import StrictVersion
 import beaker
 from beaker.cache import _cache_decorate
 from tg.i18n import ugettext as _
 import kallithea.config.conf
 from kallithea.lib.exceptions import HgsubversionImportError
 from kallithea.lib.utils2 import aslist, get_current_authuser, safe_str, safe_unicode
 from kallithea.lib.vcs.backends.git.repository import GitRepository
 from kallithea.lib.vcs.backends.hg.repository import MercurialRepository
 from kallithea.lib.vcs.conf import settings
 from kallithea.lib.vcs.exceptions import VCSError
 from kallithea.lib.vcs.utils.fakemod import create_module
 from kallithea.lib.vcs.utils.helpers import get_scm
 from kallithea.lib.vcs.utils.hgcompat import config, ui
 from kallithea.model import meta
 from kallithea.model.db import RepoGroup, Repository, Setting, Ui, User, UserGroup, UserLog
 log = logging.getLogger(__name__)
 REMOVED_REPO_PAT = re.compile(r'rm__\d{8}_\d{6}_\d{6}_.*')
 #==============================================================================
 # PERM DECORATOR HELPERS FOR EXTRACTING NAMES FOR PERM CHECKS
 #==============================================================================
 def get_repo_slug(request):
     _repo = request.environ['pylons.routes_dict'].get('repo_name')
     if _repo:
         _repo = _repo.rstrip('/')
     return _repo
 def get_repo_group_slug(request):
     _group = request.environ['pylons.routes_dict'].get('group_name')
     if _group:
         _group = _group.rstrip('/')
     return _group
 def get_user_group_slug(request):
     _group = request.environ['pylons.routes_dict'].get('id')
     _group = UserGroup.get(_group)
     if _group:
         return _group.users_group_name
     return None
 def _get_permanent_id(s):
     """Helper for decoding stable URLs with repo ID. For a string like '_123'
     return 123.
     """
     by_id_match = re.match(r'^_(\d+)$', s)
     if by_id_match is None:
         return None
     return int(by_id_match.group(1))
 def fix_repo_id_name(path):
     """
     Rewrite repo_name for _<ID> permanent URLs.
     Given a path, if the first path element is like _<ID>, return the path with
     this part expanded to the corresponding full repo name, else return the
     provided path.
     """
     first, rest = path, ''
     if '/' in path:
         first, rest_ = path.split('/', 1)
         rest = '/' + rest_
     repo_id = _get_permanent_id(first)
     if repo_id is not None:
         repo = Repository.get(repo_id)
         if repo is not None:
             return repo.repo_name + rest
     return path
 def action_logger(user, action, repo, ipaddr='', commit=False):
     """
     Action logger for various actions made by users
     :param user: user that made this action, can be a unique username string or
         object containing user_id attribute
     :param action: action to log, should be on of predefined unique actions for
         easy translations
     :param repo: string name of repository or object containing repo_id,
         that action was made on
     :param ipaddr: optional IP address from what the action was made
     """
     # if we don't get explicit IP address try to get one from registered user
     # in tmpl context var
     if not ipaddr:
         ipaddr = getattr(get_current_authuser(), 'ip_addr', '')
     if getattr(user, 'user_id', None):
         user_obj = User.get(user.user_id)
     elif isinstance(user, basestring):
         user_obj = User.get_by_username(user)
     else:
         raise Exception('You have to provide a user object or a username')
     if getattr(repo, 'repo_id', None):
         repo_obj = Repository.get(repo.repo_id)
         repo_name = repo_obj.repo_name
     elif isinstance(repo, basestring):
         repo_name = repo.lstrip('/')
         repo_obj = Repository.get_by_repo_name(repo_name)
     else:
         repo_obj = None
         repo_name = u''
     user_log = UserLog()
     user_log.user_id = user_obj.user_id
     user_log.username = user_obj.username
     user_log.action = safe_unicode(action)
     user_log.repository = repo_obj
     user_log.repository_name = repo_name
     user_log.action_date = datetime.datetime.now()
     user_log.user_ip = ipaddr
     meta.Session().add(user_log)
     log.info('Logging action:%s on %s by user:%s ip:%s',
              action, safe_unicode(repo), user_obj, ipaddr)
     if commit:
         meta.Session().commit()
 def get_filesystem_repos(path):
     """
     Scans given path for repos and return (name,(type,path)) tuple
     :param path: path to scan for repositories
     :param recursive: recursive search and return names with subdirs in front
     """
     # remove ending slash for better results
     path = safe_str(path.rstrip(os.sep))
     log.debug('now scanning in %s', path)
     def isdir(*n):
         return os.path.isdir(os.path.join(*n))
     for root, dirs, _files in os.walk(path):
         recurse_dirs = []
         for subdir in dirs:
             # skip removed repos
             if REMOVED_REPO_PAT.match(subdir):
                 continue
             # skip .<something> dirs TODO: rly? then we should prevent creating them ...
             if subdir.startswith('.'):
                 continue
             cur_path = os.path.join(root, subdir)
             if isdir(cur_path, '.git'):
                 log.warning('ignoring non-bare Git repo: %s', cur_path)
                 continue
             if (isdir(cur_path, '.hg') or
                 isdir(cur_path, '.svn') or
                 isdir(cur_path, 'objects') and (isdir(cur_path, 'refs') or
                                                 os.path.isfile(os.path.join(cur_path, 'packed-refs')))):
                 if not os.access(cur_path, os.R_OK) or not os.access(cur_path, os.X_OK):
                     log.warning('ignoring repo path without access: %s', cur_path)
                     continue
                 if not os.access(cur_path, os.W_OK):
                     log.warning('repo path without write access: %s', cur_path)
                 try:
                     scm_info = get_scm(cur_path)
                     assert cur_path.startswith(path)
                     repo_path = cur_path[len(path) + 1:]
                     yield repo_path, scm_info
                     continue # no recursion
                 except VCSError:
                     # We should perhaps ignore such broken repos, but especially
                     # the bare git detection is unreliable so we dive into it
                     pass
             recurse_dirs.append(subdir)
         dirs[:] = recurse_dirs
 def is_valid_repo_uri(repo_type, url, ui):
     """Check if the url seems like a valid remote repo location - raise an Exception if any problems"""
     if repo_type == 'hg':
         if url.startswith('http') or url.startswith('ssh'):
             # initially check if it's at least the proper URL
             # or does it pass basic auth
             MercurialRepository._check_url(url, ui)
         elif url.startswith('svn+http'):
             try:
                 from hgsubversion.svnrepo import svnremoterepo
             except ImportError:
                 raise HgsubversionImportError(_('Unable to activate hgsubversion support. '
                                                 'The "hgsubversion" library is missing'))
             svnremoterepo(ui, url).svn.uuid
         elif url.startswith('git+http'):
             raise NotImplementedError()
         else:
             raise Exception('URI %s not allowed' % (url,))
     elif repo_type == 'git':
         if url.startswith('http') or url.startswith('git'):
             # initially check if it's at least the proper URL
             # or does it pass basic auth
             GitRepository._check_url(url)
         elif url.startswith('svn+http'):
             raise NotImplementedError()
         elif url.startswith('hg+http'):
             raise NotImplementedError()
         else:
             raise Exception('URI %s not allowed' % (url))
 def is_valid_repo(repo_name, base_path, scm=None):
     """
     Returns True if given path is a valid repository False otherwise.
     If scm param is given also compare if given scm is the same as expected
     from scm parameter
     :param repo_name:
     :param base_path:
     :param scm:
     :return True: if given path is a valid repository
     """
     # TODO: paranoid security checks?
     full_path = os.path.join(safe_str(base_path), safe_str(repo_name))
     try:
         scm_ = get_scm(full_path)
         if scm:
             return scm_[0] == scm
         return True
     except VCSError:
         return False
 def is_valid_repo_group(repo_group_name, base_path, skip_path_check=False):
     """
     Returns True if given path is a repository group False otherwise
     :param repo_name:
     :param base_path:
     """
     full_path = os.path.join(safe_str(base_path), safe_str(repo_group_name))
     # check if it's not a repo
     if is_valid_repo(repo_group_name, base_path):
         return False
     try:
         # we need to check bare git repos at higher level
         # since we might match branches/hooks/info/objects or possible
         # other things inside bare git repo
         get_scm(os.path.dirname(full_path))
         return False
     except VCSError:
         pass
     # check if it's a valid path
     if skip_path_check or os.path.isdir(full_path):
         return True
     return False
 # propagated from mercurial documentation
 ui_sections = ['alias', 'auth',
                 'decode/encode', 'defaults',
                 'diff', 'email',
                 'extensions', 'format',
                 'merge-patterns', 'merge-tools',
                 'hooks', 'http_proxy',
                 'smtp', 'patch',
                 'paths', 'profiling',
                 'server', 'trusted',
                 'ui', 'web', ]
 def make_ui(repo_path=None):
     """
     Create an Mercurial 'ui' object based on database Ui settings, possibly
     augmenting with content from a hgrc file.
     """
     baseui = ui.ui()
     # clean the baseui object
     baseui._ocfg = config.config()
     baseui._ucfg = config.config()
     baseui._tcfg = config.config()
     sa = meta.Session()
     for ui_ in sa.query(Ui).all():
         if ui_.ui_active:
             ui_val = '' if ui_.ui_value is None else safe_str(ui_.ui_value)
             log.debug('config from db: [%s] %s=%r', ui_.ui_section,
                       ui_.ui_key, ui_val)
             baseui.setconfig(safe_str(ui_.ui_section), safe_str(ui_.ui_key),
                              ui_val)
     # force set push_ssl requirement to False, Kallithea handles that
     baseui.setconfig('web', 'push_ssl', False)
     baseui.setconfig('web', 'allow_push', '*')
     # prevent interactive questions for ssh password / passphrase
     ssh = baseui.config('ui', 'ssh', default='ssh')
     baseui.setconfig('ui', 'ssh', '%s -oBatchMode=yes -oIdentitiesOnly=yes' % ssh)
     # push / pull hooks
     baseui.setconfig('hooks', 'changegroup.kallithea_log_push_action', 'python:kallithea.lib.hooks.log_push_action')
     baseui.setconfig('hooks', 'outgoing.kallithea_log_pull_action', 'python:kallithea.lib.hooks.log_pull_action')
     if repo_path is not None:
         hgrc_path = os.path.join(repo_path, '.hg', 'hgrc')
         if os.path.isfile(hgrc_path):
             log.debug('reading hgrc from %s', hgrc_path)
             cfg = config.config()
             cfg.read(hgrc_path)
             for section in ui_sections:
                 for k, v in cfg.items(section):
                     log.debug('config from file: [%s] %s=%s', section, k, v)
                     baseui.setconfig(safe_str(section), safe_str(k), safe_str(v))
         else:
             log.debug('hgrc file is not present at %s, skipping...', hgrc_path)
     return baseui
 def set_app_settings(config):
     """
     Updates app config with new settings from database
     :param config:
     """
     hgsettings = Setting.get_app_settings()
     for k, v in hgsettings.items():
         config[k] = v
 def set_vcs_config(config):
     """
     Patch VCS config with some Kallithea specific stuff
     :param config: kallithea.CONFIG
     """
     settings.BACKENDS = {
         'hg': 'kallithea.lib.vcs.backends.hg.MercurialRepository',
         'git': 'kallithea.lib.vcs.backends.git.GitRepository',
+    }
     settings.GIT_EXECUTABLE_PATH = config.get('git_path', 'git')
     settings.GIT_REV_FILTER = config.get('git_rev_filter', '--all').strip()
     settings.DEFAULT_ENCODINGS = aslist(config.get('default_encoding',
                                                         'utf-8'), sep=',')
 def set_indexer_config(config):
     """
     Update Whoosh index mapping
     :param config: kallithea.CONFIG
     """
     log.debug('adding extra into INDEX_EXTENSIONS')
     kallithea.config.conf.INDEX_EXTENSIONS.extend(re.split(r'\s+', config.get('index.extensions', '')))
     log.debug('adding extra into INDEX_FILENAMES')
     kallithea.config.conf.INDEX_FILENAMES.extend(re.split(r'\s+', config.get('index.filenames', '')))
 def map_groups(path):
     """
     Given a full path to a repository, create all nested groups that this
     repo is inside. This function creates parent-child relationships between
     groups and creates default perms for all new groups.
     :param paths: full path to repository
     """
     from kallithea.model.repo_group import RepoGroupModel
     sa = meta.Session()
     groups = path.split(Repository.url_sep())
     parent = None
     group = None
     # last element is repo in nested groups structure
     groups = groups[:-1]
     rgm = RepoGroupModel()
     owner = User.get_first_admin()
     for lvl, group_name in enumerate(groups):
         group_name = u'/'.join(groups[:lvl] + [group_name])
         group = RepoGroup.get_by_group_name(group_name)
         desc = '%s group' % group_name
         # skip folders that are now removed repos
         if REMOVED_REPO_PAT.match(group_name):
             break
         if group is None:
             log.debug('creating group level: %s group_name: %s',
                       lvl, group_name)
             group = RepoGroup(group_name, parent)
             group.group_description = desc
             group.owner = owner
             sa.add(group)
             rgm._create_default_perms(group)
             sa.flush()
         parent = group
     return group
 def repo2db_mapper(initial_repo_list, remove_obsolete=False,
                    install_git_hooks=False, user=None, overwrite_git_hooks=False):
     """
     maps all repos given in initial_repo_list, non existing repositories
     are created, if remove_obsolete is True it also check for db entries
     that are not in initial_repo_list and removes them.
     :param initial_repo_list: list of repositories found by scanning methods
     :param remove_obsolete: check for obsolete entries in database
     :param install_git_hooks: if this is True, also check and install git hook
         for a repo if missing
     :param overwrite_git_hooks: if this is True, overwrite any existing git hooks
         that may be encountered (even if user-deployed)
     """
     from kallithea.model.repo import RepoModel
     from kallithea.model.scm import ScmModel
     sa = meta.Session()
     repo_model = RepoModel()
     if user is None:
         user = User.get_first_admin()
     added = []
     # creation defaults
     defs = Setting.get_default_repo_settings(strip_prefix=True)
     enable_statistics = defs.get('repo_enable_statistics')
     enable_downloads = defs.get('repo_enable_downloads')
     private = defs.get('repo_private')
     for name, repo in initial_repo_list.items():
         group = map_groups(name)
         unicode_name = safe_unicode(name)
         db_repo = repo_model.get_by_repo_name(unicode_name)
         # found repo that is on filesystem not in Kallithea database
         if not db_repo:
             log.info('repository %s not found, creating now', name)
             added.append(name)
             desc = (repo.description
                     if repo.description != 'unknown'
                     else '%s repository' % name)
             new_repo = repo_model._create_repo(
                 repo_name=name,
                 repo_type=repo.alias,
                 description=desc,
                 repo_group=getattr(group, 'group_id', None),
                 owner=user,
                 enable_downloads=enable_downloads,
                 enable_statistics=enable_statistics,
                 private=private,
                 state=Repository.STATE_CREATED
+            )
             sa.commit()
             # we added that repo just now, and make sure it has githook
             # installed, and updated server info
             if new_repo.repo_type == 'git':
                 git_repo = new_repo.scm_instance
                 ScmModel().install_git_hooks(git_repo)
                 # update repository server-info
                 log.debug('Running update server info')
                 git_repo._update_server_info()
             new_repo.update_changeset_cache()
         elif install_git_hooks:
             if db_repo.repo_type == 'git':
                 ScmModel().install_git_hooks(db_repo.scm_instance, force_create=overwrite_git_hooks)
     removed = []
     # remove from database those repositories that are not in the filesystem
     unicode_initial_repo_list = set(safe_unicode(name) for name in initial_repo_list)
     for repo in sa.query(Repository).all():
         if repo.repo_name not in unicode_initial_repo_list:
             if remove_obsolete:
                 log.debug("Removing non-existing repository found in db `%s`",
                           repo.repo_name)
                 try:
                     RepoModel().delete(repo, forks='detach', fs_remove=False)
                     sa.commit()
                 except Exception:
                     #don't hold further removals on error
                     log.error(traceback.format_exc())
                     sa.rollback()
             removed.append(repo.repo_name)
     return added, removed
 def load_rcextensions(root_path):
     path = os.path.join(root_path, 'rcextensions', '__init__.py')
     if os.path.isfile(path):
         rcext = create_module('rc', path)
         EXT = kallithea.EXTENSIONS = rcext
         log.debug('Found rcextensions now loading %s...', rcext)
         # Additional mappings that are not present in the pygments lexers
         kallithea.config.conf.LANGUAGES_EXTENSIONS_MAP.update(getattr(EXT, 'EXTRA_MAPPINGS', {}))
         # OVERRIDE OUR EXTENSIONS FROM RC-EXTENSIONS (if present)
         if getattr(EXT, 'INDEX_EXTENSIONS', []):
             log.debug('settings custom INDEX_EXTENSIONS')
             kallithea.config.conf.INDEX_EXTENSIONS = getattr(EXT, 'INDEX_EXTENSIONS', [])
         # ADDITIONAL MAPPINGS
         log.debug('adding extra into INDEX_EXTENSIONS')
         kallithea.config.conf.INDEX_EXTENSIONS.extend(getattr(EXT, 'EXTRA_INDEX_EXTENSIONS', []))
         # auto check if the module is not missing any data, set to default if is
         # this will help autoupdate new feature of rcext module
         #from kallithea.config import rcextensions
         #for k in dir(rcextensions):
         #    if not k.startswith('_') and not hasattr(EXT, k):
         #        setattr(EXT, k, getattr(rcextensions, k))
 #==============================================================================
 # MISC
 #==============================================================================
 git_req_ver = StrictVersion('1.7.4')
 def check_git_version():
     """
     Checks what version of git is installed on the system, and raise a system exit
     if it's too old for Kallithea to work properly.
     """
     if 'git' not in kallithea.BACKENDS:
         return None
     if not settings.GIT_EXECUTABLE_PATH:
         log.warning('No git executable configured - check "git_path" in the ini file.')
         return None
     stdout, stderr = GitRepository._run_git_command(['--version'], _bare=True,
                                                     _safe=True)
     if stderr:
         log.warning('Error/stderr from "%s --version": %r', settings.GIT_EXECUTABLE_PATH, stderr)
     m = re.search(r"\d+.\d+.\d+", stdout)
     if m:
         ver = StrictVersion(m.group(0))
         log.debug('Git executable: "%s", version %s (parsed from: "%s")',
                   settings.GIT_EXECUTABLE_PATH, ver, stdout.strip())
         if ver < git_req_ver:
             log.error('Kallithea detected %s version %s, which is too old '
                       'for the system to function properly. '
                       'Please upgrade to version %s or later. '
                       'If you strictly need Mercurial repositories, you can '
                       'clear the "git_path" setting in the ini file.',
                       settings.GIT_EXECUTABLE_PATH, ver, git_req_ver)
             log.error("Terminating ...")
             sys.exit(1)
     else:
         ver = StrictVersion('0.0.0')
         log.warning('Error finding version number in "%s --version" stdout: %r',
                     settings.GIT_EXECUTABLE_PATH, stdout.strip())
     return ver
 #===============================================================================
 # CACHE RELATED METHODS
 #===============================================================================
 # set cache regions for beaker so celery can utilise it
 def setup_cache_regions(settings):
     # Create dict with just beaker cache configs with prefix stripped
     cache_settings = {'regions': None}
     prefix = 'beaker.cache.'
     for key in settings:
         if key.startswith(prefix):
             name = key[len(prefix):]
             cache_settings[name] = settings[key]
     # Find all regions, apply defaults, and apply to beaker
     if cache_settings['regions']:
         for region in cache_settings['regions'].split(','):
             region = region.strip()
             prefix = region + '.'
             region_settings = {}
             for key in cache_settings:
                 if key.startswith(prefix):
                     name = key[len(prefix):]
                     region_settings[name] = cache_settings[key]
             region_settings.setdefault('expire',
                                        cache_settings.get('expire', '60'))
             region_settings.setdefault('lock_dir',
                                        cache_settings.get('lock_dir'))
             region_settings.setdefault('data_dir',
                                        cache_settings.get('data_dir'))
             region_settings.setdefault('type',
                                        cache_settings.get('type', 'memory'))
             beaker.cache.cache_regions[region] = region_settings
 def conditional_cache(region, prefix, condition, func):
     """
     Conditional caching function use like::
         def _c(arg):
             #heavy computation function
             return data
         # depending from condition the compute is wrapped in cache or not
         compute = conditional_cache('short_term', 'cache_desc', condition=True, func=func)
         return compute(arg)
     :param region: name of cache region
     :param prefix: cache region prefix
     :param condition: condition for cache to be triggered, and return data cached
     :param func: wrapped heavy function to compute
     """
     wrapped = func
     if condition:
         log.debug('conditional_cache: True, wrapping call of '
                   'func: %s into %s region cache' % (region, func))
         wrapped = _cache_decorate((prefix,), None, None, region)(func)
     return wrapped

kallithea/lib/utils2.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.utils2
 ~~~~~~~~~~~~~~~~~~~~
 Some simple helper functions.
 Note: all these functions should be independent of Kallithea classes, i.e.
 models, controllers, etc.  to prevent import cycles.
 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: Jan 5, 2011
 :author: marcink
 :copyright: (c) 2013 RhodeCode GmbH, and others.
 :license: GPLv3, see LICENSE.md for more details.
 """
 from __future__ import print_function
 import binascii
 import datetime
 import os
 import pwd
 import re
 import time
 import urllib
 import urlobject
 from tg.i18n import ugettext as _
 from tg.i18n import ungettext
 from webhelpers2.text import collapse, remove_formatting, strip_tags
 from kallithea.lib.compat import json
 from kallithea.lib.vcs.utils import safe_str, safe_unicode  # re-export
 from kallithea.lib.vcs.utils.lazy import LazyProperty
 def str2bool(_str):
     """
     returns True/False value from given string, it tries to translate the
     string into boolean
     :param _str: string value to translate into boolean
     :rtype: boolean
     :returns: boolean from given string
     """
     if _str is None:
         return False
     if _str in (True, False):
         return _str
     _str = str(_str).strip().lower()
     return _str in ('t', 'true', 'y', 'yes', 'on', '1')
 def aslist(obj, sep=None, strip=True):
     """
     Returns given string separated by sep as list
     :param obj:
     :param sep:
     :param strip:
     """
     if isinstance(obj, (basestring)):
         lst = obj.split(sep)
         if strip:
             lst = [v.strip() for v in lst]
         return lst
     elif isinstance(obj, (list, tuple)):
         return obj
     elif obj is None:
         return []
     else:
         return [obj]
 def convert_line_endings(line, mode):
     """
     Converts a given line  "line end" according to given mode
     Available modes are::
 - Unix
 - Mac
 - DOS
     :param line: given line to convert
     :param mode: mode to convert to
     :rtype: str
     :return: converted line according to mode
     """
     from string import replace
     if mode == 0:
         line = replace(line, '\r\n', '\n')
         line = replace(line, '\r', '\n')
     elif mode == 1:
         line = replace(line, '\r\n', '\r')
         line = replace(line, '\n', '\r')
     elif mode == 2:
         line = re.sub("\r(?!\n)|(?<!\r)\n", "\r\n", line)
     return line
 def detect_mode(line, default):
     """
     Detects line break for given line, if line break couldn't be found
     given default value is returned
     :param line: str line
     :param default: default
     :rtype: int
     :return: value of line end on of 0 - Unix, 1 - Mac, 2 - DOS
     """
     if line.endswith('\r\n'):
         return 2
     elif line.endswith('\n'):
         return 0
     elif line.endswith('\r'):
         return 1
     else:
         return default
 def generate_api_key():
     """
     Generates a random (presumably unique) API key.
     This value is used in URLs and "Bearer" HTTP Authorization headers,
     which in practice means it should only contain URL-safe characters
     (RFC 3986):
         unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
     """
     # Hexadecimal certainly qualifies as URL-safe.
     return binascii.hexlify(os.urandom(20))
 def safe_int(val, default=None):
     """
     Returns int() of val if val is not convertable to int use default
     instead
     :param val:
     :param default:
     """
     try:
         val = int(val)
     except (ValueError, TypeError):
         val = default
     return val
 def remove_suffix(s, suffix):
     if s.endswith(suffix):
         s = s[:-1 * len(suffix)]
     return s
 def remove_prefix(s, prefix):
     if s.startswith(prefix):
         s = s[len(prefix):]
     return s
 def age(prevdate, show_short_version=False, now=None):
     """
     turns a datetime into an age string.
     If show_short_version is True, then it will generate a not so accurate but shorter string,
     example: 2days ago, instead of 2 days and 23 hours ago.
     :param prevdate: datetime object
     :param show_short_version: if it should approximate the date and return a shorter string
     :rtype: unicode
     :returns: unicode words describing age
     """
     now = now or datetime.datetime.now()
     order = ['year', 'month', 'day', 'hour', 'minute', 'second']
     deltas = {}
     future = False
     if prevdate > now:
         now, prevdate = prevdate, now
         future = True
     if future:
         prevdate = prevdate.replace(microsecond=0)
     # Get date parts deltas
     from dateutil import relativedelta
     for part in order:
         d = relativedelta.relativedelta(now, prevdate)
         deltas[part] = getattr(d, part + 's')
     # Fix negative offsets (there is 1 second between 10:59:59 and 11:00:00,
     # not 1 hour, -59 minutes and -59 seconds)
     for num, length in [(5, 60), (4, 60), (3, 24)]:  # seconds, minutes, hours
         part = order[num]
         carry_part = order[num - 1]
         if deltas[part] < 0:
             deltas[part] += length
             deltas[carry_part] -= 1
     # Same thing for days except that the increment depends on the (variable)
     # number of days in the month
     month_lengths = [31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31]
     if deltas['day'] < 0:
         if prevdate.month == 2 and (prevdate.year % 4 == 0 and
             (prevdate.year % 100 != 0 or prevdate.year % 400 == 0)
         ):
             deltas['day'] += 29
         else:
             deltas['day'] += month_lengths[prevdate.month - 1]
         deltas['month'] -= 1
     if deltas['month'] < 0:
         deltas['month'] += 12
         deltas['year'] -= 1
     # In short version, we want nicer handling of ages of more than a year
     if show_short_version:
         if deltas['year'] == 1:
             # ages between 1 and 2 years: show as months
             deltas['month'] += 12
             deltas['year'] = 0
         if deltas['year'] >= 2:
             # ages 2+ years: round
             if deltas['month'] > 6:
                 deltas['year'] += 1
                 deltas['month'] = 0
     # Format the result
     fmt_funcs = {
         'year': lambda d: ungettext(u'%d year', '%d years', d) % d,
         'month': lambda d: ungettext(u'%d month', '%d months', d) % d,
         'day': lambda d: ungettext(u'%d day', '%d days', d) % d,
         'hour': lambda d: ungettext(u'%d hour', '%d hours', d) % d,
         'minute': lambda d: ungettext(u'%d minute', '%d minutes', d) % d,
         'second': lambda d: ungettext(u'%d second', '%d seconds', d) % d,
+    }
     for i, part in enumerate(order):
         value = deltas[part]
         if value == 0:
             continue
         if i < 5:
             sub_part = order[i + 1]
             sub_value = deltas[sub_part]
         else:
             sub_value = 0
         if sub_value == 0 or show_short_version:
             if future:
                 return _('in %s') % fmt_funcs[part](value)
             else:
                 return _('%s ago') % fmt_funcs[part](value)
         if future:
             return _('in %s and %s') % (fmt_funcs[part](value),
                 fmt_funcs[sub_part](sub_value))
         else:
             return _('%s and %s ago') % (fmt_funcs[part](value),
                 fmt_funcs[sub_part](sub_value))
     return _('just now')
 def uri_filter(uri):
     """
     Removes user:password from given url string
     :param uri:
     :rtype: unicode
     :returns: filtered list of strings
     """
     if not uri:
         return ''
     proto = ''
     for pat in ('https://', 'http://', 'git://'):
         if uri.startswith(pat):
             uri = uri[len(pat):]
             proto = pat
             break
     # remove passwords and username
     uri = uri[uri.find('@') + 1:]
     # get the port
     cred_pos = uri.find(':')
     if cred_pos == -1:
         host, port = uri, None
     else:
         host, port = uri[:cred_pos], uri[cred_pos + 1:]
     return [_f for _f in [proto, host, port] if _f]
 def credentials_filter(uri):
     """
     Returns a url with removed credentials
     :param uri:
     """
     uri = uri_filter(uri)
     # check if we have port
     if len(uri) > 2 and uri[2]:
         uri[2] = ':' + uri[2]
     return ''.join(uri)
 def get_clone_url(clone_uri_tmpl, prefix_url, repo_name, repo_id, username=None):
     parsed_url = urlobject.URLObject(prefix_url)
     prefix = safe_unicode(urllib.unquote(parsed_url.path.rstrip('/')))
     try:
         system_user = pwd.getpwuid(os.getuid()).pw_name
     except Exception: # TODO: support all systems - especially Windows
         system_user = 'kallithea' # hardcoded default value ...
     args = {
         'scheme': parsed_url.scheme,
         'user': safe_unicode(urllib.quote(safe_str(username or ''))),
         'netloc': parsed_url.netloc + prefix,  # like "hostname:port/prefix" (with optional ":port" and "/prefix")
         'prefix': prefix, # undocumented, empty or starting with /
         'repo': repo_name,
         'repoid': str(repo_id),
         'system_user': safe_unicode(system_user),
         'hostname': parsed_url.hostname,
+    }
     url = re.sub('{([^{}]+)}', lambda m: args.get(m.group(1), m.group(0)), clone_uri_tmpl)
     # remove leading @ sign if it's present. Case of empty user
     url_obj = urlobject.URLObject(url)
     if not url_obj.username:
         url_obj = url_obj.with_username(None)
     return safe_unicode(url_obj)
 def get_changeset_safe(repo, rev):
     """
     Safe version of get_changeset if this changeset doesn't exists for a
     repo it returns a Dummy one instead
     :param repo:
     :param rev:
     """
     from kallithea.lib.vcs.backends.base import BaseRepository
     from kallithea.lib.vcs.exceptions import RepositoryError
     from kallithea.lib.vcs.backends.base import EmptyChangeset
     if not isinstance(repo, BaseRepository):
         raise Exception('You must pass an Repository '
                         'object as first argument got %s' % type(repo))
     try:
         cs = repo.get_changeset(rev)
     except (RepositoryError, LookupError):
         cs = EmptyChangeset(requested_revision=rev)
     return cs
 def datetime_to_time(dt):
     if dt:
         return time.mktime(dt.timetuple())
 def time_to_datetime(tm):
     if tm:
         if isinstance(tm, basestring):
             try:
                 tm = float(tm)
             except ValueError:
                 return
         return datetime.datetime.fromtimestamp(tm)
 # Must match regexp in kallithea/public/js/base.js MentionsAutoComplete()
 # Check char before @ - it must not look like we are in an email addresses.
 # Matching is greedy so we don't have to look beyond the end.
 MENTIONS_REGEX = re.compile(r'(?:^|(?<=[^a-zA-Z0-9]))@([a-zA-Z0-9][-_.a-zA-Z0-9]*[a-zA-Z0-9])')
 def extract_mentioned_usernames(text):
     r"""
     Returns list of (possible) usernames @mentioned in given text.
     >>> extract_mentioned_usernames('@1-2.a_X,@1234 not@not @ddd@not @n @ee @ff @gg, @gg;@hh @n\n@zz,')
     ['1-2.a_X', '1234', 'ddd', 'ee', 'ff', 'gg', 'gg', 'hh', 'zz']
     """
     return MENTIONS_REGEX.findall(text)
 def extract_mentioned_users(text):
     """ Returns set of actual database Users @mentioned in given text. """
     from kallithea.model.db import User
     result = set()
     for name in extract_mentioned_usernames(text):
         user = User.get_by_username(name, case_insensitive=True)
         if user is not None and not user.is_default_user:
             result.add(user)
     return result
 class AttributeDict(dict):
     def __getattr__(self, attr):
         return self.get(attr, None)
     __setattr__ = dict.__setitem__
     __delattr__ = dict.__delitem__
 def obfuscate_url_pw(engine):
     from sqlalchemy.engine import url as sa_url
     from sqlalchemy.exc import ArgumentError
     try:
         _url = sa_url.make_url(engine or '')
     except ArgumentError:
         return engine
     if _url.password:
         _url.password = 'XXXXX'
     return str(_url)
 def get_hook_environment():
     """
     Get hook context by deserializing the global KALLITHEA_EXTRAS environment
     variable.
     Called early in Git out-of-process hooks to get .ini config path so the
     basic environment can be configured properly. Also used in all hooks to get
     information about the action that triggered it.
     """
     try:
         extras = json.loads(os.environ['KALLITHEA_EXTRAS'])
     except KeyError:
         raise Exception("Environment variable KALLITHEA_EXTRAS not found")
     try:
         for k in ['username', 'repository', 'scm', 'action', 'ip']:
             extras[k]
     except KeyError:
         raise Exception('Missing key %s in KALLITHEA_EXTRAS %s' % (k, extras))
     return AttributeDict(extras)
 def set_hook_environment(username, ip_addr, repo_name, repo_alias, action=None):
     """Prepare global context for running hooks by serializing data in the
     global KALLITHEA_EXTRAS environment variable.
     Most importantly, this allow Git hooks to do proper logging and updating of
     caches after pushes.
     Must always be called before anything with hooks are invoked.
     """
     from kallithea import CONFIG
     extras = {
         'ip': ip_addr, # used in log_push/pull_action action_logger
         'username': username,
         'action': action or 'push_local', # used in log_push_action_raw_ids action_logger
         'repository': repo_name,
         'scm': repo_alias, # used to pick hack in log_push_action_raw_ids
         'config': CONFIG['__file__'], # used by git hook to read config
+    }
     os.environ['KALLITHEA_EXTRAS'] = json.dumps(extras)
 def get_current_authuser():
     """
     Gets kallithea user from threadlocal tmpl_context variable if it's
     defined, else returns None.
     """
     from tg import tmpl_context
     if hasattr(tmpl_context, 'authuser'):
         return tmpl_context.authuser
     return None
 class OptionalAttr(object):
     """
     Special Optional Option that defines other attribute. Example::
         def test(apiuser, userid=Optional(OAttr('apiuser')):
             user = Optional.extract(userid)
             # calls
     """
     def __init__(self, attr_name):
         self.attr_name = attr_name
     def __repr__(self):
         return '<OptionalAttr:%s>' % self.attr_name
     def __call__(self):
         return self
 # alias
 OAttr = OptionalAttr
 class Optional(object):
     """
     Defines an optional parameter::
         param = param.getval() if isinstance(param, Optional) else param
         param = param() if isinstance(param, Optional) else param
     is equivalent of::
         param = Optional.extract(param)
     """
     def __init__(self, type_):
         self.type_ = type_
     def __repr__(self):
         return '<Optional:%s>' % self.type_.__repr__()
     def __call__(self):
         return self.getval()
     def getval(self):
         """
         returns value from this Optional instance
         """
         if isinstance(self.type_, OAttr):
             # use params name
             return self.type_.attr_name

kallithea/lib/vcs/backends/git/inmemory.py

➞

Show inline comments

 import datetime
 import posixpath
 import stat
 import time
 from dulwich import objects
 from kallithea.lib.vcs.backends.base import BaseInMemoryChangeset
 from kallithea.lib.vcs.exceptions import RepositoryError
 from kallithea.lib.vcs.utils import safe_str
 class GitInMemoryChangeset(BaseInMemoryChangeset):
     def commit(self, message, author, parents=None, branch=None, date=None,
                **kwargs):
         """
         Performs in-memory commit (doesn't check workdir in any way) and
         returns newly created ``Changeset``. Updates repository's
         ``revisions``.
         :param message: message of the commit
         :param author: full username, i.e. "Joe Doe <joe.doe@example.com>"
         :param parents: single parent or sequence of parents from which commit
           would be derived
         :param date: ``datetime.datetime`` instance. Defaults to
           ``datetime.datetime.now()``.
         :param branch: branch name, as string. If none given, default backend's
           branch would be used.
         :raises ``CommitError``: if any error occurs while committing
         """
         self.check_integrity(parents)
         from .repository import GitRepository
         if branch is None:
             branch = GitRepository.DEFAULT_BRANCH_NAME
         repo = self.repository._repo
         object_store = repo.object_store
         ENCODING = "UTF-8"
         # Create tree and populates it with blobs
         commit_tree = self.parents[0] and repo[self.parents[0]._commit.tree] or \
             objects.Tree()
         for node in self.added + self.changed:
             # Compute subdirs if needed
             dirpath, nodename = posixpath.split(node.path)
             dirnames = safe_str(dirpath).split('/') if dirpath else []
             parent = commit_tree
             ancestors = [('', parent)]
             # Tries to dig for the deepest existing tree
             while dirnames:
                 curdir = dirnames.pop(0)
                 try:
                     dir_id = parent[curdir][1]
                 except KeyError:
                     # put curdir back into dirnames and stops
                     dirnames.insert(0, curdir)
                     break
                 else:
                     # If found, updates parent
                     parent = self.repository._repo[dir_id]
                     ancestors.append((curdir, parent))
             # Now parent is deepest existing tree and we need to create subtrees
             # for dirnames (in reverse order) [this only applies for nodes from added]
             new_trees = []
             if not node.is_binary:
                 content = node.content.encode(ENCODING)
             else:
                 content = node.content
             blob = objects.Blob.from_string(content)
             node_path = node.name.encode(ENCODING)
             if dirnames:
                 # If there are trees which should be created we need to build
                 # them now (in reverse order)
                 reversed_dirnames = list(reversed(dirnames))
                 curtree = objects.Tree()
                 curtree[node_path] = node.mode, blob.id
                 new_trees.append(curtree)
                 for dirname in reversed_dirnames[:-1]:
                     newtree = objects.Tree()
                     #newtree.add(stat.S_IFDIR, dirname, curtree.id)
                     newtree[dirname] = stat.S_IFDIR, curtree.id
                     new_trees.append(newtree)
                     curtree = newtree
                 parent[reversed_dirnames[-1]] = stat.S_IFDIR, curtree.id
             else:
                 parent.add(name=node_path, mode=node.mode, hexsha=blob.id)
             new_trees.append(parent)
             # Update ancestors
             for parent, tree, path in reversed([(a[1], b[1], b[0]) for a, b in
                 zip(ancestors, ancestors[1:])]
             ):
                 parent[path] = stat.S_IFDIR, tree.id
                 object_store.add_object(tree)
             object_store.add_object(blob)
             for tree in new_trees:
                 object_store.add_object(tree)
         for node in self.removed:
             paths = node.path.split('/')
             tree = commit_tree
             trees = [tree]
             # Traverse deep into the forest...
             for path in paths:
                 try:
                     obj = self.repository._repo[tree[path][1]]
                     if isinstance(obj, objects.Tree):
                         trees.append(obj)
                         tree = obj
                 except KeyError:
                     break
             # Cut down the blob and all rotten trees on the way back...
             for path, tree in reversed(zip(paths, trees)):
                 del tree[path]
                 if tree:
                     # This tree still has elements - don't remove it or any
                     # of it's parents
                     break
         object_store.add_object(commit_tree)
         # Create commit
         commit = objects.Commit()
         commit.tree = commit_tree.id
         commit.parents = [p._commit.id for p in self.parents if p]
         commit.author = commit.committer = safe_str(author)
         commit.encoding = ENCODING
         commit.message = safe_str(message)
         # Compute date
         if date is None:
             date = time.time()
         elif isinstance(date, datetime.datetime):
             date = time.mktime(date.timetuple())
         author_time = kwargs.pop('author_time', date)
         commit.commit_time = int(date)
         commit.author_time = int(author_time)
         tz = time.timezone
         author_tz = kwargs.pop('author_timezone', tz)
         commit.commit_timezone = tz
         commit.author_timezone = author_tz
         object_store.add_object(commit)
         ref = 'refs/heads/%s' % branch
         repo.refs[ref] = commit.id
         # Update vcs repository object & recreate dulwich repo
         self.repository.revisions.append(commit.id)
         # invalidate parsed refs after commit
         self.repository._parsed_refs = self.repository._get_parsed_refs()
         tip = self.repository.get_changeset()
         self.reset()
         return tip
     def _get_missing_trees(self, path, root_tree):
         """
         Creates missing ``Tree`` objects for the given path.
         :param path: path given as a string. It may be a path to a file node
           (i.e. ``foo/bar/baz.txt``) or directory path - in that case it must
           end with slash (i.e. ``foo/bar/``).
         :param root_tree: ``dulwich.objects.Tree`` object from which we start
           traversing (should be commit's root tree)
         """
         dirpath = posixpath.split(path)[0]
         dirs = dirpath.split('/')
         if not dirs or dirs == ['']:
             return []
         def get_tree_for_dir(tree, dirname):
             for name, mode, id in tree.iteritems():
                 if name == dirname:
                     obj = self.repository._repo[id]
                     if isinstance(obj, objects.Tree):
                         return obj
                     else:
                         raise RepositoryError("Cannot create directory %s "
                         "at tree %s as path is occupied and is not a "
                         "Tree" % (dirname, tree))
                             "at tree %s as path is occupied and is not a "
                             "Tree" % (dirname, tree))
             return None
         trees = []
         parent = root_tree
         for dirname in dirs:
             tree = get_tree_for_dir(parent, dirname)
             if tree is None:
                 tree = objects.Tree()
                 parent.add(stat.S_IFDIR, dirname, tree.id)
                 parent = tree
             # Always append tree
             trees.append(tree)
         return trees

kallithea/lib/vcs/backends/git/repository.py

➞

Show inline comments

 # -*- coding: utf-8 -*-
 """
     vcs.backends.git.repository
     ~~~~~~~~~~~~~~~~~~~~~~~~~~~
     Git repository implementation.
     :created_on: Apr 8, 2010
     :copyright: (c) 2010-2011 by Marcin Kuzminski, Lukasz Balcerzak.
 """
 import errno
 import logging
 import os
 import posixpath
 import re
 import time
 import urllib
 import urllib2
 from collections import OrderedDict
 from dulwich.config import ConfigFile
 from dulwich.objects import Tag
 from dulwich.repo import NotGitRepository, Repo
 from kallithea.lib.vcs import subprocessio
 from kallithea.lib.vcs.backends.base import BaseRepository, CollectionGenerator
 from kallithea.lib.vcs.conf import settings
 from kallithea.lib.vcs.exceptions import (
     BranchDoesNotExistError, ChangesetDoesNotExistError, EmptyRepositoryError, RepositoryError, TagAlreadyExistError, TagDoesNotExistError)
 from kallithea.lib.vcs.utils import date_fromtimestamp, makedate, safe_str, safe_unicode
 from kallithea.lib.vcs.utils.hgcompat import hg_url, httpbasicauthhandler, httpdigestauthhandler
 from kallithea.lib.vcs.utils.lazy import LazyProperty
 from kallithea.lib.vcs.utils.paths import abspath, get_user_home
 from .changeset import GitChangeset
 from .inmemory import GitInMemoryChangeset
 from .workdir import GitWorkdir
 SHA_PATTERN = re.compile(r'^([0-9a-fA-F]{12}|[0-9a-fA-F]{40})$')
 log = logging.getLogger(__name__)
 class GitRepository(BaseRepository):
     """
     Git repository backend.
     """
     DEFAULT_BRANCH_NAME = 'master'
     scm = 'git'
     def __init__(self, repo_path, create=False, src_url=None,
                  update_after_clone=False, bare=False):
         self.path = safe_unicode(abspath(repo_path))
         self.repo = self._get_repo(create, src_url, update_after_clone, bare)
         self.bare = self.repo.bare
     @property
     def _config_files(self):
         return [
             self.bare and abspath(self.path, 'config')
                       or abspath(self.path, '.git', 'config'),
              abspath(get_user_home(), '.gitconfig'),
+         ]
     @property
     def _repo(self):
         return self.repo
     @property
     def head(self):
         try:
             return self._repo.head()
         except KeyError:
             return None
     @property
     def _empty(self):
         """
         Checks if repository is empty ie. without any changesets
         """
         try:
             self.revisions[0]
         except (KeyError, IndexError):
             return True
         return False
     @LazyProperty
     def revisions(self):
         """
         Returns list of revisions' ids, in ascending order.  Being lazy
         attribute allows external tools to inject shas from cache.
         """
         return self._get_all_revisions()
     @classmethod
     def _run_git_command(cls, cmd, **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)
         try:
             return ''.join(p.output), ''.join(p.error)
         finally:
             p.close()
     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))
                 "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.reverse()
         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

kallithea/lib/vcs/subprocessio.py

➞

Show inline comments

 """
 Module provides a class allowing to wrap communication over subprocess.Popen
 input, output, error streams into a meaningful, non-blocking, concurrent
 stream processor exposing the output data as an iterator fitting to be a
 return value passed by a WSGI application to a WSGI server per PEP 3333.
 Copyright (c) 2011  Daniel Dotsenko <dotsa[at]hotmail.com>
 This file is part of git_http_backend.py Project.
 git_http_backend.py Project is free software: you can redistribute it and/or
 modify it under the terms of the GNU Lesser General Public License as
 published by the Free Software Foundation, either version 2.1 of the License,
 or (at your option) any later version.
 git_http_backend.py Project 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 Lesser General Public License for more details.
 You should have received a copy of the GNU Lesser General Public License
 along with git_http_backend.py Project.
 If not, see <http://www.gnu.org/licenses/>.
 """
 import collections
 import os
 import subprocess
 import threading
 class StreamFeeder(threading.Thread):
     """
     Normal writing into pipe-like is blocking once the buffer is filled.
     This thread allows a thread to seep data from a file-like into a pipe
     without blocking the main thread.
     We close inpipe once the end of the source stream is reached.
     """
     def __init__(self, source):
         super(StreamFeeder, self).__init__()
         self.daemon = True
         filelike = False
         self.bytes = bytes()
         if type(source) in (type(''), bytes, bytearray):  # string-like
             self.bytes = bytes(source)
         else:  # can be either file pointer or file-like
             if type(source) in (int, long):  # file pointer it is
                 # converting file descriptor (int) stdin into file-like
                 source = os.fdopen(source, 'rb', 16384)
             # let's see if source is file-like by now
             filelike = hasattr(source, 'read')
         if not filelike and not self.bytes:
             raise TypeError("StreamFeeder's source object must be a readable "
                             "file-like, a file descriptor, or a string-like.")
         self.source = source
         self.readiface, self.writeiface = os.pipe()
     def run(self):
         t = self.writeiface
         if self.bytes:
             os.write(t, self.bytes)
         else:
             s = self.source
             b = s.read(4096)
             while b:
                 os.write(t, b)
                 b = s.read(4096)
         os.close(t)
     @property
     def output(self):
         return self.readiface
 class InputStreamChunker(threading.Thread):
     def __init__(self, source, target, buffer_size, chunk_size):
         super(InputStreamChunker, self).__init__()
         self.daemon = True  # die die die.
         self.source = source
         self.target = target
         self.chunk_count_max = int(buffer_size / chunk_size) + 1
         self.chunk_size = chunk_size
         self.data_added = threading.Event()
         self.data_added.clear()
         self.keep_reading = threading.Event()
         self.keep_reading.set()
         self.EOF = threading.Event()
         self.EOF.clear()
         self.go = threading.Event()
         self.go.set()
     def stop(self):
         self.go.clear()
         self.EOF.set()
         try:
             # this is not proper, but is done to force the reader thread let
             # go of the input because, if successful, .close() will send EOF
             # down the pipe.
             self.source.close()
         except:
             pass
     def run(self):
         s = self.source
         t = self.target
         cs = self.chunk_size
         ccm = self.chunk_count_max
         kr = self.keep_reading
         da = self.data_added
         go = self.go
         try:
             b = s.read(cs)
         except ValueError:
             b = ''
         while b and go.is_set():
             if len(t) > ccm:
                 kr.clear()
                 kr.wait(2)
                 # # this only works on 2.7.x and up
                 # if not kr.wait(10):
                 #     raise Exception("Timed out while waiting for input to be read.")
                 # instead we'll use this
                 if len(t) > ccm + 3:
                     raise IOError(
                         "Timed out while waiting for input from subprocess.")
             t.append(b)
             da.set()
             try:
                 b = s.read(cs)
             except ValueError: # probably "I/O operation on closed file"
                 b = ''
         self.EOF.set()
         da.set()  # for cases when done but there was no input.
 class BufferedGenerator(object):
     """
     Class behaves as a non-blocking, buffered pipe reader.
     Reads chunks of data (through a thread)
     from a blocking pipe, and attaches these to an array (Deque) of chunks.
     Reading is halted in the thread when max chunks is internally buffered.
     The .next() may operate in blocking or non-blocking fashion by yielding
     '' if no data is ready
     to be sent or by not returning until there is some data to send
     When we get EOF from underlying source pipe we raise the marker to raise
     StopIteration after the last chunk of data is yielded.
     """
     def __init__(self, source, buffer_size=65536, chunk_size=4096,
                  starting_values=None, bottomless=False):
         starting_values = starting_values or []
         if bottomless:
             maxlen = int(buffer_size / chunk_size)
         else:
             maxlen = None
         self.data = collections.deque(starting_values, maxlen)
         self.worker = InputStreamChunker(source, self.data, buffer_size,
                                          chunk_size)
         if starting_values:
             self.worker.data_added.set()
         self.worker.start()
     ####################
     # Generator's methods
     ####################
     def __iter__(self):
         return self
     def next(self):
         while not len(self.data) and not self.worker.EOF.is_set():
             self.worker.data_added.clear()
             self.worker.data_added.wait(0.2)
         if len(self.data):
             self.worker.keep_reading.set()
             return bytes(self.data.popleft())
         elif self.worker.EOF.is_set():
             raise StopIteration
     def throw(self, type, value=None, traceback=None):
         if not self.worker.EOF.is_set():
             raise type(value)
     def start(self):
         self.worker.start()
     def stop(self):
         self.worker.stop()
     def close(self):
         try:
             self.worker.stop()
             self.throw(GeneratorExit)
         except (GeneratorExit, StopIteration):
             pass
     ####################
     # Threaded reader's infrastructure.
     ####################
     @property
     def input(self):
         return self.worker.w
     @property
     def data_added_event(self):
         return self.worker.data_added
     @property
     def data_added(self):
         return self.worker.data_added.is_set()
     @property
     def reading_paused(self):
         return not self.worker.keep_reading.is_set()
     @property
     def done_reading_event(self):
         """
         Done_reading does not mean that the iterator's buffer is empty.
         Iterator might have done reading from underlying source, but the read
         chunks might still be available for serving through .next() method.
         :returns: An threading.Event class instance.
         """
         return self.worker.EOF
     @property
     def done_reading(self):
         """
         Done_reading does not mean that the iterator's buffer is empty.
         Iterator might have done reading from underlying source, but the read
         chunks might still be available for serving through .next() method.
         :returns: An Bool value.
         """
         return self.worker.EOF.is_set()
     @property
     def length(self):
         """
         returns int.
         This is the length of the queue of chunks, not the length of
         the combined contents in those chunks.
         __len__() cannot be meaningfully implemented because this
         reader is just flying through a bottomless pit content and
         can only know the length of what it already saw.
         If __len__() on WSGI server per PEP 3333 returns a value,
         the response's length will be set to that. In order not to
         confuse WSGI PEP3333 servers, we will not implement __len__
         at all.
         """
         return len(self.data)
     def prepend(self, x):
         self.data.appendleft(x)
     def append(self, x):
         self.data.append(x)
     def extend(self, o):
         self.data.extend(o)
     def __getitem__(self, i):
         return self.data[i]
 class SubprocessIOChunker(object):
     """
     Processor class wrapping handling of subprocess IO.
     In a way, this is a "communicate()" replacement with a twist.
     - We are multithreaded. Writing in and reading out, err are all sep threads.
     - We support concurrent (in and out) stream processing.
     - The output is not a stream. It's a queue of read string (bytes, not unicode)
       chunks. The object behaves as an iterable. You can "for chunk in obj:" us.
     - We are non-blocking in more respects than communicate()
       (reading from subprocess out pauses when internal buffer is full, but
        does not block the parent calling code. On the flip side, reading from
        slow-yielding subprocess may block the iteration until data shows up. This
        does not block the parallel inpipe reading occurring parallel thread.)
     The purpose of the object is to allow us to wrap subprocess interactions into
     an iterable that can be passed to a WSGI server as the application's return
     value. Because of stream-processing-ability, WSGI does not have to read ALL
     of the subprocess's output and buffer it, before handing it to WSGI server for
     HTTP response. Instead, the class initializer reads just a bit of the stream
     to figure out if error occurred or likely to occur and if not, just hands the
     further iteration over subprocess output to the server for completion of HTTP
     response.
     The real or perceived subprocess error is trapped and raised as one of
     EnvironmentError family of exceptions
     Example usage:
     #    try:
     #        answer = SubprocessIOChunker(
     #            cmd,
     #            input,
     #            buffer_size = 65536,
     #            chunk_size = 4096
     #            )
     #    except (EnvironmentError) as e:
     #        print str(e)
     #        raise e
+    #
     #    return answer
     """
     def __init__(self, cmd, inputstream=None, buffer_size=65536,
                  chunk_size=4096, starting_values=None, **kwargs):
         """
         Initializes SubprocessIOChunker
         :param cmd: A Subprocess.Popen style "cmd". Can be string or array of strings
         :param inputstream: (Default: None) A file-like, string, or file pointer.
         :param buffer_size: (Default: 65536) A size of total buffer per stream in bytes.
         :param chunk_size: (Default: 4096) A max size of a chunk. Actual chunk may be smaller.
         :param starting_values: (Default: []) An array of strings to put in front of output que.
         """
         starting_values = starting_values or []
         if inputstream:
             input_streamer = StreamFeeder(inputstream)
             input_streamer.start()
             inputstream = input_streamer.output
         # Note: fragile cmd mangling has been removed for use in Kallithea
         assert isinstance(cmd, list), cmd
         _p = subprocess.Popen(cmd, bufsize=-1,
                               stdin=inputstream,
                               stdout=subprocess.PIPE,
                               stderr=subprocess.PIPE,
                               **kwargs)
         bg_out = BufferedGenerator(_p.stdout, buffer_size, chunk_size,
                                    starting_values)
         bg_err = BufferedGenerator(_p.stderr, 16000, 1, bottomless=True)
         while not bg_out.done_reading and not bg_out.reading_paused:
             # doing this until we reach either end of file, or end of buffer.
             bg_out.data_added_event.wait(1)
             bg_out.data_added_event.clear()
         # at this point it's still ambiguous if we are done reading or just full buffer.
         # Either way, if error (returned by ended process, or implied based on
         # presence of stuff in stderr output) we error out.
         # Else, we are happy.
         returncode = _p.poll()
         if (returncode is not None # process has terminated
             and returncode != 0
         ): # and it failed
             bg_out.stop()
             out = ''.join(bg_out)
             bg_err.stop()
             err = ''.join(bg_err)
             if (err.strip() == 'fatal: The remote end hung up unexpectedly' and
                 out.startswith('0034shallow ')
             ):
                 # hack inspired by https://github.com/schacon/grack/pull/7
                 bg_out = iter([out])
                 _p = None
             elif err:
                 raise EnvironmentError(
                     "Subprocess exited due to an error:\n" + err)
                 raise EnvironmentError("Subprocess exited due to an error: %s" % err)
             else:
                 raise EnvironmentError(
                     "Subprocess exited with non 0 ret code: %s" % returncode)
         self.process = _p
         self.output = bg_out
         self.error = bg_err
         self.inputstream = inputstream
     def __iter__(self):
         return self
     def next(self):
         if self.process:
             returncode = self.process.poll()
             if (returncode is not None # process has terminated
                 and returncode != 0
             ): # and it failed
                 self.output.stop()
                 self.error.stop()
                 err = ''.join(self.error)
                 raise EnvironmentError("Subprocess exited due to an error:\n" + err)
         return self.output.next()
     def throw(self, type, value=None, traceback=None):
         if self.output.length or not self.output.done_reading:
             raise type(value)
     def close(self):
         try:
             self.process.terminate()
         except:
             pass
         try:
             self.output.close()
         except:
             pass
         try:
             self.error.close()
         except:
             pass
         try:
             os.close(self.inputstream)
         except:
             pass

kallithea/lib/vcs/utils/__init__.py

➞

Show inline comments

 """
 This module provides some useful tools for ``vcs`` like annotate/diff html
 output. It also includes some internal helpers.
 """
 import datetime
 import re
 import time
 def makedate():
     lt = time.localtime()
     if lt[8] == 1 and time.daylight:
         tz = time.altzone
     else:
         tz = time.timezone
     return time.mktime(lt), tz
 def aslist(obj, sep=None, strip=True):
     """
     Returns given string separated by sep as list
     :param obj:
     :param sep:
     :param strip:
     """
-    if isinstance(obj, (basestring)):
     if isinstance(obj, basestring):
         lst = obj.split(sep)
         if strip:
             lst = [v.strip() for v in lst]
         return lst
     elif isinstance(obj, (list, tuple)):
         return obj
     elif obj is None:
         return []
     else:
         return [obj]
 def date_fromtimestamp(unixts, tzoffset=0):
     """
     Makes a local datetime object out of unix timestamp
     :param unixts:
     :param tzoffset:
     """
     return datetime.datetime.fromtimestamp(float(unixts))
 def safe_int(val, default=None):
     """
     Returns int() of val if val is not convertible to int use default
     instead
     :param val:
     :param default:
     """
     try:
         val = int(val)
     except (ValueError, TypeError):
         val = default
     return val
 def safe_unicode(s):
     """
     Safe unicode function. Use a few tricks to turn s into unicode string:
     In case of UnicodeDecodeError with configured default encodings, try to
     detect encoding with chardet library, then fall back to first encoding with
     errors replaced.
     """
     if isinstance(s, unicode):
         return s
     if not isinstance(s, str):  # use __str__ / __unicode__ and don't expect UnicodeDecodeError
         return unicode(s)
     from kallithea.lib.vcs.conf import settings
     for enc in settings.DEFAULT_ENCODINGS:
         try:
             return unicode(s, enc)
         except UnicodeDecodeError:
             pass
     try:
         import chardet
         encoding = chardet.detect(s)['encoding']
         if encoding is not None:
             return s.decode(encoding)
     except (ImportError, UnicodeDecodeError):
         pass
     return unicode(s, settings.DEFAULT_ENCODINGS[0], 'replace')
 def safe_str(s):
     """
     Safe str function. Use a few tricks to turn s into bytes string:
     In case of UnicodeEncodeError with configured default encodings, fall back
     to first configured encoding with errors replaced.
     """
     if isinstance(s, str):
         return s
     assert isinstance(s, unicode), s  # don't use safe_str to coerce non-strings
     from kallithea.lib.vcs.conf import settings
     for enc in settings.DEFAULT_ENCODINGS:
         try:
             return s.encode(enc)
         except UnicodeEncodeError:
             pass
     return s.encode(settings.DEFAULT_ENCODINGS[0], 'replace')
 # Regex taken from http://www.regular-expressions.info/email.html
 email_re = re.compile(
     r"""[a-z0-9!#$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#$%&'*+/=?^_`{|}~-]+)*@"""
     r"""(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?""",
     re.IGNORECASE)
 def author_email(author):
     """
     Returns email address of given author string.
     If author contains <> brackets, only look inside that.
     If any RFC valid email address is found, return that.
     Else, return empty string.
     """
     if not author:
         return ''
     l = author.find('<') + 1
     if l != 0:
         r = author.find('>', l)
         if r != -1:
             author = author[l:r]
     m = email_re.search(author)
     if m is None:
         return ''
     return safe_str(m.group(0))
 def author_name(author):
     """
     get name of author, or else username.
     It'll try to find an email in the author string and just cut it off
     to get the username
     """
     if not author:
         return ''
     if '@' not in author:
         return author
     return author.replace(author_email(author), '').replace('<', '') \
         .replace('>', '').strip()

0 comments (0 inline, 0 general)