Changeset - e7b6a0ce123c
[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
5 files changed with 3 insertions and 7 deletions:
0 comments (0 inline, 0 general)
kallithea/lib/utils.py
Show inline comments
 
@@ -187,385 +187,384 @@ def get_filesystem_repos(path):
 
            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')
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::
 
        0 - Unix
 
        1 - Mac
 
        2 - 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):
kallithea/lib/vcs/backends/git/repository.py
Show inline comments
 
@@ -25,385 +25,385 @@ from dulwich.repo import NotGitRepositor
 

	
 
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``.
 

	
kallithea/lib/vcs/subprocessio.py
Show inline comments
 
@@ -188,238 +188,237 @@ class BufferedGenerator(object):
 
        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)