audit-logs: improve user-agent normalizer, and added tests

rhodecode/lib/utils2.py

@@ -1,1165 +1,1173 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2011-2020 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # 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 Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 
 """
 Some simple helper functions
 """
 
 import collections
 import datetime
 import dateutil.relativedelta
 import hashlib
 import logging
 import re
 import sys
 import time
 import urllib
 import urlobject
 import uuid
 import getpass
 from functools import update_wrapper, partial, wraps
 
 import pygments.lexers
 import sqlalchemy
 import sqlalchemy.engine.url
 import sqlalchemy.exc
 import sqlalchemy.sql
 import webob
 import pyramid.threadlocal
 from pyramid import compat
 from pyramid.settings import asbool
 
 import rhodecode
 from rhodecode.translation import _, _pluralize
 
 
 def md5(s):
     return hashlib.md5(s).hexdigest()
 
 
 def md5_safe(s):
     return md5(safe_str(s))
 
 
 def sha1(s):
     return hashlib.sha1(s).hexdigest()
 
 
 def sha1_safe(s):
     return sha1(safe_str(s))
 
 
 def __get_lem(extra_mapping=None):
     """
     Get language extension map based on what's inside pygments lexers
     """
     d = collections.defaultdict(lambda: [])
 
     def __clean(s):
         s = s.lstrip('*')
         s = s.lstrip('.')
 
         if s.find('[') != -1:
             exts = []
             start, stop = s.find('['), s.find(']')
 
             for suffix in s[start + 1:stop]:
                 exts.append(s[:s.find('[')] + suffix)
             return [e.lower() for e in exts]
         else:
             return [s.lower()]
 
     for lx, t in sorted(pygments.lexers.LEXERS.items()):
         m = map(__clean, t[-2])
         if m:
             m = reduce(lambda x, y: x + y, m)
             for ext in m:
                 desc = lx.replace('Lexer', '')
                 d[ext].append(desc)
 
     data = dict(d)
 
     extra_mapping = extra_mapping or {}
     if extra_mapping:
         for k, v in extra_mapping.items():
             if k not in data:
                 # register new mapping2lexer
                 data[k] = [v]
 
     return data
 
 
 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" accordingly 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
     """
     if mode == 0:
         line = line.replace('\r\n', '\n')
         line = line.replace('\r', '\n')
     elif mode == 1:
         line = line.replace('\r\n', '\r')
         line = line.replace('\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 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 safe_unicode(str_, from_encoding=None, use_chardet=False):
     """
     safe unicode function. Does few trick to turn str_ into unicode
 
     In case of UnicodeDecode error, we try to return it with encoding detected
     by chardet library if it fails fallback to unicode with errors replaced
 
     :param str_: string to decode
     :rtype: unicode
     :returns: unicode object
     """
     if isinstance(str_, unicode):
         return str_
 
     if not from_encoding:
         DEFAULT_ENCODINGS = aslist(rhodecode.CONFIG.get('default_encoding',
                                                         'utf8'), sep=',')
         from_encoding = DEFAULT_ENCODINGS
 
     if not isinstance(from_encoding, (list, tuple)):
         from_encoding = [from_encoding]
 
     try:
         return unicode(str_)
     except UnicodeDecodeError:
         pass
 
     for enc in from_encoding:
         try:
             return unicode(str_, enc)
         except UnicodeDecodeError:
             pass
 
     if use_chardet:
         try:
             import chardet
             encoding = chardet.detect(str_)['encoding']
             if encoding is None:
                 raise Exception()
             return str_.decode(encoding)
         except (ImportError, UnicodeDecodeError, Exception):
             return unicode(str_, from_encoding[0], 'replace')
     else:
         return unicode(str_, from_encoding[0], 'replace')
 
 def safe_str(unicode_, to_encoding=None, use_chardet=False):
     """
     safe str function. Does few trick to turn unicode_ into string
 
     In case of UnicodeEncodeError, we try to return it with encoding detected
     by chardet library if it fails fallback to string with errors replaced
 
     :param unicode_: unicode to encode
     :rtype: str
     :returns: str object
     """
 
     # if it's not basestr cast to str
     if not isinstance(unicode_, compat.string_types):
         return str(unicode_)
 
     if isinstance(unicode_, str):
         return unicode_
 
     if not to_encoding:
         DEFAULT_ENCODINGS = aslist(rhodecode.CONFIG.get('default_encoding',
                                                         'utf8'), sep=',')
         to_encoding = DEFAULT_ENCODINGS
 
     if not isinstance(to_encoding, (list, tuple)):
         to_encoding = [to_encoding]
 
     for enc in to_encoding:
         try:
             return unicode_.encode(enc)
         except UnicodeEncodeError:
             pass
 
     if use_chardet:
         try:
             import chardet
             encoding = chardet.detect(unicode_)['encoding']
             if encoding is None:
                 raise UnicodeEncodeError()
 
             return unicode_.encode(encoding)
         except (ImportError, UnicodeEncodeError):
             return unicode_.encode(to_encoding[0], 'replace')
     else:
         return unicode_.encode(to_encoding[0], 'replace')
 
 
 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 find_calling_context(ignore_modules=None):
     """
     Look through the calling stack and return the frame which called
     this function and is part of core module ( ie. rhodecode.* )
 
     :param ignore_modules: list of modules to ignore eg. ['rhodecode.lib']
     """
 
     ignore_modules = ignore_modules or []
 
     f = sys._getframe(2)
     while f.f_back is not None:
         name = f.f_globals.get('__name__')
         if name and name.startswith(__name__.split('.')[0]):
             if name not in ignore_modules:
                 return f
         f = f.f_back
     return None
 
 
 def ping_connection(connection, branch):
     if branch:
         # "branch" refers to a sub-connection of a connection,
         # we don't want to bother pinging on these.
         return
 
     # turn off "close with result".  This flag is only used with
     # "connectionless" execution, otherwise will be False in any case
     save_should_close_with_result = connection.should_close_with_result
     connection.should_close_with_result = False
 
     try:
         # run a SELECT 1.   use a core select() so that
         # the SELECT of a scalar value without a table is
         # appropriately formatted for the backend
         connection.scalar(sqlalchemy.sql.select([1]))
     except sqlalchemy.exc.DBAPIError as err:
         # catch SQLAlchemy's DBAPIError, which is a wrapper
         # for the DBAPI's exception.  It includes a .connection_invalidated
         # attribute which specifies if this connection is a "disconnect"
         # condition, which is based on inspection of the original exception
         # by the dialect in use.
         if err.connection_invalidated:
             # run the same SELECT again - the connection will re-validate
             # itself and establish a new connection.  The disconnect detection
             # here also causes the whole connection pool to be invalidated
             # so that all stale connections are discarded.
             connection.scalar(sqlalchemy.sql.select([1]))
         else:
             raise
     finally:
         # restore "close with result"
         connection.should_close_with_result = save_should_close_with_result
 
 
 def engine_from_config(configuration, prefix='sqlalchemy.', **kwargs):
     """Custom engine_from_config functions."""
     log = logging.getLogger('sqlalchemy.engine')
     use_ping_connection = asbool(configuration.pop('sqlalchemy.db1.ping_connection', None))
     debug = asbool(configuration.pop('sqlalchemy.db1.debug_query', None))
 
     engine = sqlalchemy.engine_from_config(configuration, prefix, **kwargs)
 
     def color_sql(sql):
         color_seq = '\033[1;33m'  # This is yellow: code 33
         normal = '\x1b[0m'
         return ''.join([color_seq, sql, normal])
 
     if use_ping_connection:
         log.debug('Adding ping_connection on the engine config.')
         sqlalchemy.event.listen(engine, "engine_connect", ping_connection)
 
     if debug:
         # attach events only for debug configuration
         def before_cursor_execute(conn, cursor, statement,
                                   parameters, context, executemany):
             setattr(conn, 'query_start_time', time.time())
             log.info(color_sql(">>>>> STARTING QUERY >>>>>"))
             calling_context = find_calling_context(ignore_modules=[
                 'rhodecode.lib.caching_query',
                 'rhodecode.model.settings',
             ])
             if calling_context:
                 log.info(color_sql('call context %s:%s' % (
                     calling_context.f_code.co_filename,
                     calling_context.f_lineno,
                 )))
 
         def after_cursor_execute(conn, cursor, statement,
                                  parameters, context, executemany):
             delattr(conn, 'query_start_time')
 
         sqlalchemy.event.listen(engine, "before_cursor_execute", before_cursor_execute)
         sqlalchemy.event.listen(engine, "after_cursor_execute", after_cursor_execute)
 
     return engine
 
 
 def get_encryption_key(config):
     secret = config.get('rhodecode.encrypted_values.secret')
     default = config['beaker.session.secret']
     return secret or default
 
 
 def age(prevdate, now=None, show_short_version=False, show_suffix=True,
         short_format=False):
     """
     Turns a datetime into an age string.
     If show_short_version is True, this generates a shorter string with
     an approximate age; ex. '1 day ago', rather than '1 day and 23 hours ago'.
 
     * IMPORTANT*
     Code of this function is written in special way so it's easier to
     backport it to javascript. If you mean to update it, please also update
     `jquery.timeago-extension.js` file
 
     :param prevdate: datetime object
     :param now: get current time, if not define we use
         `datetime.datetime.now()`
     :param show_short_version: if it should approximate the date and
         return a shorter string
     :param show_suffix:
     :param short_format: show short format, eg 2D instead of 2 days
     :rtype: unicode
     :returns: unicode words describing age
     """
 
     def _get_relative_delta(now, prevdate):
         base = dateutil.relativedelta.relativedelta(now, prevdate)
         return {
             'year': base.years,
             'month': base.months,
             'day': base.days,
             'hour': base.hours,
             'minute': base.minutes,
             'second': base.seconds,
         }
 
     def _is_leap_year(year):
         return year % 4 == 0 and (year % 100 != 0 or year % 400 == 0)
 
     def get_month(prevdate):
         return prevdate.month
 
     def get_year(prevdate):
         return prevdate.year
 
     now = now or datetime.datetime.now()
     order = ['year', 'month', 'day', 'hour', 'minute', 'second']
     deltas = {}
     future = False
 
     if prevdate > now:
         now_old = now
         now = prevdate
         prevdate = now_old
         future = True
     if future:
         prevdate = prevdate.replace(microsecond=0)
     # Get date parts deltas
     for part in order:
         rel_delta = _get_relative_delta(now, prevdate)
         deltas[part] = rel_delta[part]
 
     # Fix negative offsets (there is 1 second between 10:59:59 and 11:00:00,
     # not 1 hour, -59 minutes and -59 seconds)
     offsets = [[5, 60], [4, 60], [3, 24]]
     for element in offsets:  # seconds, minutes, hours
         num = element[0]
         length = element[1]
 
         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 get_month(prevdate) == 2 and _is_leap_year(get_year(prevdate)):
             deltas['day'] += 29
         else:
             deltas['day'] += month_lengths[get_month(prevdate) - 1]
 
         deltas['month'] -= 1
 
     if deltas['month'] < 0:
         deltas['month'] += 12
         deltas['year'] -= 1
 
     # Format the result
     if short_format:
         fmt_funcs = {
             'year': lambda d: u'%dy' % d,
             'month': lambda d: u'%dm' % d,
             'day': lambda d: u'%dd' % d,
             'hour': lambda d: u'%dh' % d,
             'minute': lambda d: u'%dmin' % d,
             'second': lambda d: u'%dsec' % d,
         }
     else:
         fmt_funcs = {
             'year': lambda d: _pluralize(u'${num} year', u'${num} years', d, mapping={'num': d}).interpolate(),
             'month': lambda d: _pluralize(u'${num} month', u'${num} months', d, mapping={'num': d}).interpolate(),
             'day': lambda d: _pluralize(u'${num} day', u'${num} days', d, mapping={'num': d}).interpolate(),
             'hour': lambda d: _pluralize(u'${num} hour', u'${num} hours', d, mapping={'num': d}).interpolate(),
             'minute': lambda d: _pluralize(u'${num} minute', u'${num} minutes', d, mapping={'num': d}).interpolate(),
             'second': lambda d: _pluralize(u'${num} second', u'${num} seconds', d, mapping={'num': d}).interpolate(),
         }
 
     i = 0
     for part in order:
         value = deltas[part]
         if value != 0:
 
             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:
                 _val = fmt_funcs[part](value)
                 if future:
                     if show_suffix:
                         return _(u'in ${ago}', mapping={'ago': _val})
                     else:
                         return _(_val)
 
                 else:
                     if show_suffix:
                         return _(u'${ago} ago', mapping={'ago': _val})
                     else:
                         return _(_val)
 
             val = fmt_funcs[part](value)
             val_detail = fmt_funcs[sub_part](sub_value)
             mapping = {'val': val, 'detail': val_detail}
 
             if short_format:
                 datetime_tmpl = _(u'${val}, ${detail}', mapping=mapping)
                 if show_suffix:
                     datetime_tmpl = _(u'${val}, ${detail} ago', mapping=mapping)
                     if future:
                         datetime_tmpl = _(u'in ${val}, ${detail}', mapping=mapping)
             else:
                 datetime_tmpl = _(u'${val} and ${detail}', mapping=mapping)
                 if show_suffix:
                     datetime_tmpl = _(u'${val} and ${detail} ago', mapping=mapping)
                     if future:
                         datetime_tmpl = _(u'in ${val} and ${detail}', mapping=mapping)
 
             return datetime_tmpl
         i += 1
     return _(u'just now')
 
 
 def age_from_seconds(seconds):
     seconds = safe_int(seconds) or 0
     prevdate = time_to_datetime(time.time() + seconds)
     return age(prevdate, show_suffix=False, show_short_version=True)
 
 
 def cleaned_uri(uri):
     """
     Quotes '[' and ']' from uri if there is only one of them.
         according to RFC3986 we cannot use such chars in uri
     :param uri:
     :return: uri without this chars
     """
     return urllib.quote(uri, safe='@$:/')
 
 
 def credentials_filter(uri):
     """
     Returns a url with removed credentials
 
     :param uri:
     """
     import urlobject
     if isinstance(uri, rhodecode.lib.encrypt.InvalidDecryptedValue):
         return 'InvalidDecryptionKey'
 
     url_obj = urlobject.URLObject(cleaned_uri(uri))
     url_obj = url_obj.without_password().without_username()
 
     return url_obj
 
 
 def get_host_info(request):
     """
     Generate host info, to obtain full url e.g https://server.com
     use this
     `{scheme}://{netloc}`
     """
     if not request:
         return {}
 
     qualified_home_url = request.route_url('home')
     parsed_url = urlobject.URLObject(qualified_home_url)
     decoded_path = safe_unicode(urllib.unquote(parsed_url.path.rstrip('/')))
 
     return {
         'scheme': parsed_url.scheme,
         'netloc': parsed_url.netloc+decoded_path,
         'hostname': parsed_url.hostname,
     }
 
 
 def get_clone_url(request, uri_tmpl, repo_name, repo_id, repo_type, **override):
     qualified_home_url = request.route_url('home')
     parsed_url = urlobject.URLObject(qualified_home_url)
     decoded_path = safe_unicode(urllib.unquote(parsed_url.path.rstrip('/')))
 
     args = {
         'scheme': parsed_url.scheme,
         'user': '',
         'sys_user': getpass.getuser(),
         # path if we use proxy-prefix
         'netloc': parsed_url.netloc+decoded_path,
         'hostname': parsed_url.hostname,
         'prefix': decoded_path,
         'repo': repo_name,
         'repoid': str(repo_id),
         'repo_type': repo_type
     }
     args.update(override)
     args['user'] = urllib.quote(safe_str(args['user']))
 
     for k, v in args.items():
         uri_tmpl = uri_tmpl.replace('{%s}' % k, v)
 
     # special case for SVN clone url
     if repo_type == 'svn':
         uri_tmpl = uri_tmpl.replace('ssh://', 'svn+ssh://')
 
     # remove leading @ sign if it's present. Case of empty user
     url_obj = urlobject.URLObject(uri_tmpl)
     url = url_obj.with_netloc(url_obj.netloc.lstrip('@'))
 
     return safe_unicode(url)
 
 
 def get_commit_safe(repo, commit_id=None, commit_idx=None, pre_load=None,
                     maybe_unreachable=False, reference_obj=None):
     """
     Safe version of get_commit if this commit doesn't exists for a
     repository it returns a Dummy one instead
 
     :param repo: repository instance
     :param commit_id: commit id as str
     :param commit_idx: numeric commit index
     :param pre_load: optional list of commit attributes to load
     :param maybe_unreachable: translate unreachable commits on git repos
     :param reference_obj: explicitly search via a reference obj in git. E.g "branch:123" would mean branch "123"
     """
     # TODO(skreft): remove these circular imports
     from rhodecode.lib.vcs.backends.base import BaseRepository, EmptyCommit
     from rhodecode.lib.vcs.exceptions import RepositoryError
     if not isinstance(repo, BaseRepository):
         raise Exception('You must pass an Repository '
                         'object as first argument got %s', type(repo))
 
     try:
         commit = repo.get_commit(
             commit_id=commit_id, commit_idx=commit_idx, pre_load=pre_load,
             maybe_unreachable=maybe_unreachable, reference_obj=reference_obj)
     except (RepositoryError, LookupError):
         commit = EmptyCommit()
     return commit
 
 
 def datetime_to_time(dt):
     if dt:
         return time.mktime(dt.timetuple())
 
 
 def time_to_datetime(tm):
     if tm:
         if isinstance(tm, compat.string_types):
             try:
                 tm = float(tm)
             except ValueError:
                 return
         return datetime.datetime.fromtimestamp(tm)
 
 
 def time_to_utcdatetime(tm):
     if tm:
         if isinstance(tm, compat.string_types):
             try:
                 tm = float(tm)
             except ValueError:
                 return
         return datetime.datetime.utcfromtimestamp(tm)
 
 
 MENTIONS_REGEX = re.compile(
     # ^@ or @ without any special chars in front
     r'(?:^@|[^a-zA-Z0-9\-\_\.]@)'
     # main body starts with letter, then can be . - _
     r'([a-zA-Z0-9]{1}[a-zA-Z0-9\-\_\.]+)',
     re.VERBOSE | re.MULTILINE)
 
 
 def extract_mentioned_users(s):
     """
     Returns unique usernames from given string s that have @mention
 
     :param s: string to get mentions
     """
     usrs = set()
     for username in MENTIONS_REGEX.findall(s):
         usrs.add(username)
 
     return sorted(list(usrs), key=lambda k: k.lower())
 
 
 class AttributeDictBase(dict):
     def __getstate__(self):
         odict = self.__dict__  # get attribute dictionary
         return odict
 
     def __setstate__(self, dict):
         self.__dict__ = dict
 
     __setattr__ = dict.__setitem__
     __delattr__ = dict.__delitem__
 
 
 class StrictAttributeDict(AttributeDictBase):
     """
     Strict Version of Attribute dict which raises an Attribute error when
     requested attribute is not set
     """
     def __getattr__(self, attr):
         try:
             return self[attr]
         except KeyError:
             raise AttributeError('%s object has no attribute %s' % (
                 self.__class__, attr))
 
 
 class AttributeDict(AttributeDictBase):
     def __getattr__(self, attr):
         return self.get(attr, None)
 
 
 
 class OrderedDefaultDict(collections.OrderedDict, collections.defaultdict):
     def __init__(self, default_factory=None, *args, **kwargs):
         # in python3 you can omit the args to super
         super(OrderedDefaultDict, self).__init__(*args, **kwargs)
         self.default_factory = default_factory
 
 
 def fix_PATH(os_=None):
     """
     Get current active python path, and append it to PATH variable to fix
     issues of subprocess calls and different python versions
     """
     if os_ is None:
         import os
     else:
         os = os_
 
     cur_path = os.path.split(sys.executable)[0]
     if not os.environ['PATH'].startswith(cur_path):
         os.environ['PATH'] = '%s:%s' % (cur_path, os.environ['PATH'])
 
 
 def obfuscate_url_pw(engine):
     _url = engine or ''
     try:
         _url = sqlalchemy.engine.url.make_url(engine)
         if _url.password:
             _url.password = 'XXXXX'
     except Exception:
         pass
     return unicode(_url)
 
 
 def get_server_url(environ):
     req = webob.Request(environ)
     return req.host_url + req.script_name
 
 
 def unique_id(hexlen=32):
     alphabet = "23456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghjklmnpqrstuvwxyz"
     return suuid(truncate_to=hexlen, alphabet=alphabet)
 
 
 def suuid(url=None, truncate_to=22, alphabet=None):
     """
     Generate and return a short URL safe UUID.
 
     If the url parameter is provided, set the namespace to the provided
     URL and generate a UUID.
 
     :param url to get the uuid for
     :truncate_to: truncate the basic 22 UUID to shorter version
 
     The IDs won't be universally unique any longer, but the probability of
     a collision will still be very low.
     """
     # Define our alphabet.
     _ALPHABET = alphabet or "23456789ABCDEFGHJKLMNPQRSTUVWXYZ"
 
     # If no URL is given, generate a random UUID.
     if url is None:
         unique_id = uuid.uuid4().int
     else:
         unique_id = uuid.uuid3(uuid.NAMESPACE_URL, url).int
 
     alphabet_length = len(_ALPHABET)
     output = []
     while unique_id > 0:
         digit = unique_id % alphabet_length
         output.append(_ALPHABET[digit])
         unique_id = int(unique_id / alphabet_length)
     return "".join(output)[:truncate_to]
 
 
 def get_current_rhodecode_user(request=None):
     """
     Gets rhodecode user from request
     """
     pyramid_request = request or pyramid.threadlocal.get_current_request()
 
     # web case
     if pyramid_request and hasattr(pyramid_request, 'user'):
         return pyramid_request.user
 
     # api case
     if pyramid_request and hasattr(pyramid_request, 'rpc_user'):
         return pyramid_request.rpc_user
 
     return None
 
 
 def action_logger_generic(action, namespace=''):
     """
     A generic logger for actions useful to the system overview, tries to find
     an acting user for the context of the call otherwise reports unknown user
 
     :param action: logging message eg 'comment 5 deleted'
     :param type: string
 
     :param namespace: namespace of the logging message eg. 'repo.comments'
     :param type: string
 
     """
 
     logger_name = 'rhodecode.actions'
 
     if namespace:
         logger_name += '.' + namespace
 
     log = logging.getLogger(logger_name)
 
     # get a user if we can
     user = get_current_rhodecode_user()
 
     logfunc = log.info
 
     if not user:
         user = '<unknown user>'
         logfunc = log.warning
 
     logfunc('Logging action by {}: {}'.format(user, action))
 
 
 def escape_split(text, sep=',', maxsplit=-1):
     r"""
     Allows for escaping of the separator: e.g. arg='foo\, bar'
 
     It should be noted that the way bash et. al. do command line parsing, those
     single quotes are required.
     """
     escaped_sep = r'\%s' % sep
 
     if escaped_sep not in text:
         return text.split(sep, maxsplit)
 
     before, _mid, after = text.partition(escaped_sep)
     startlist = before.split(sep, maxsplit)  # a regular split is fine here
     unfinished = startlist[-1]
     startlist = startlist[:-1]
 
     # recurse because there may be more escaped separators
     endlist = escape_split(after, sep, maxsplit)
 
     # finish building the escaped value. we use endlist[0] becaue the first
     # part of the string sent in recursion is the rest of the escaped value.
     unfinished += sep + endlist[0]
 
     return startlist + [unfinished] + endlist[1:]  # put together all the parts
 
 
 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
         return self.type_
 
     @classmethod
     def extract(cls, val):
         """
         Extracts value from Optional() instance
 
         :param val:
         :return: original value if it's not Optional instance else
             value of instance
         """
         if isinstance(val, cls):
             return val.getval()
         return val
 
 
 def glob2re(pat):
     """
     Translate a shell PATTERN to a regular expression.
 
     There is no way to quote meta-characters.
     """
 
     i, n = 0, len(pat)
     res = ''
     while i < n:
         c = pat[i]
         i = i+1
         if c == '*':
             #res = res + '.*'
             res = res + '[^/]*'
         elif c == '?':
             #res = res + '.'
             res = res + '[^/]'
         elif c == '[':
             j = i
             if j < n and pat[j] == '!':
                 j = j+1
             if j < n and pat[j] == ']':
                 j = j+1
             while j < n and pat[j] != ']':
                 j = j+1
             if j >= n:
                 res = res + '\\['
             else:
                 stuff = pat[i:j].replace('\\','\\\\')
                 i = j+1
                 if stuff[0] == '!':
                     stuff = '^' + stuff[1:]
                 elif stuff[0] == '^':
                     stuff = '\\' + stuff
                 res = '%s[%s]' % (res, stuff)
         else:
             res = res + re.escape(c)
     return res + '\Z(?ms)'
 
 
 def parse_byte_string(size_str):
     match = re.match(r'(\d+)(MB|KB)', size_str, re.IGNORECASE)
     if not match:
         raise ValueError('Given size:%s is invalid, please make sure '
                          'to use format of <num>(MB|KB)' % size_str)
 
     _parts = match.groups()
     num, type_ = _parts
     return long(num) * {'mb': 1024*1024, 'kb': 1024}[type_.lower()]
 
 
 class CachedProperty(object):
     """
     Lazy Attributes. With option to invalidate the cache by running a method
 
     >>> class Foo(object):
     ...
     ...    @CachedProperty
     ...    def heavy_func(self):
     ...        return 'super-calculation'
     ...
     ... foo = Foo()
     ... foo.heavy_func() # first computation
     ... foo.heavy_func() # fetch from cache
     ... foo._invalidate_prop_cache('heavy_func')
 
     # at this point calling foo.heavy_func() will be re-computed
     """
 
     def __init__(self, func, func_name=None):
 
         if func_name is None:
             func_name = func.__name__
         self.data = (func, func_name)
         update_wrapper(self, func)
 
     def __get__(self, inst, class_):
         if inst is None:
             return self
 
         func, func_name = self.data
         value = func(inst)
         inst.__dict__[func_name] = value
         if '_invalidate_prop_cache' not in inst.__dict__:
             inst.__dict__['_invalidate_prop_cache'] = partial(
                 self._invalidate_prop_cache, inst)
         return value
 
     def _invalidate_prop_cache(self, inst, name):
         inst.__dict__.pop(name, None)
 
 
 def retry(func=None, exception=Exception, n_tries=5, delay=5, backoff=1, logger=True):
     """
     Retry decorator with exponential backoff.
 
     Parameters
     ----------
     func : typing.Callable, optional
         Callable on which the decorator is applied, by default None
     exception : Exception or tuple of Exceptions, optional
         Exception(s) that invoke retry, by default Exception
     n_tries : int, optional
         Number of tries before giving up, by default 5
     delay : int, optional
         Initial delay between retries in seconds, by default 5
     backoff : int, optional
         Backoff multiplier e.g. value of 2 will double the delay, by default 1
     logger : bool, optional
         Option to log or print, by default False
 
     Returns
     -------
     typing.Callable
         Decorated callable that calls itself when exception(s) occur.
 
     Examples
     --------
     >>> import random
     >>> @retry(exception=Exception, n_tries=3)
     ... def test_random(text):
     ...    x = random.random()
     ...    if x < 0.5:
     ...        raise Exception("Fail")
     ...    else:
     ...        print("Success: ", text)
     >>> test_random("It works!")
     """
 
     if func is None:
         return partial(
             retry,
             exception=exception,
             n_tries=n_tries,
             delay=delay,
             backoff=backoff,
             logger=logger,
         )
 
     @wraps(func)
     def wrapper(*args, **kwargs):
         _n_tries, n_delay = n_tries, delay
         log = logging.getLogger('rhodecode.retry')
 
         while _n_tries > 1:
             try:
                 return func(*args, **kwargs)
             except exception as e:
                 e_details = repr(e)
                 msg = "Exception on calling func {func}: {e}, " \
                       "Retrying in {n_delay} seconds..."\
                     .format(func=func, e=e_details, n_delay=n_delay)
                 if logger:
                     log.warning(msg)
                 else:
                     print(msg)
                 time.sleep(n_delay)
                 _n_tries -= 1
                 n_delay *= backoff
 
         return func(*args, **kwargs)
 
     return wrapper
 
 
-def user_agent_normalizer(user_agent_raw):
+def user_agent_normalizer(user_agent_raw, safe=True):
     log = logging.getLogger('rhodecode.user_agent_normalizer')
     ua = (user_agent_raw or '').strip().lower()
+    ua = ua.replace('"', '')
 
     try:
         if 'mercurial/proto-1.0' in ua:
             ua = ua.replace('mercurial/proto-1.0', '')
             ua = ua.replace('(', '').replace(')', '').strip()
             ua = ua.replace('mercurial ', 'mercurial/')
         elif ua.startswith('git'):
-            pass
+            parts = ua.split(' ')
+            if parts:
+                ua = parts[0]
+            ua = re.sub('\.windows\.\d', '', ua).strip()
+
+        return ua
     except Exception:
         log.exception('Failed to parse scm user-agent')
+        if not safe:
+            raise
 
     return ua

rhodecode/tests/lib/test_utils.py

@@ -1,449 +1,478 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2010-2020 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # 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 Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import json
 import multiprocessing
 import os
 
 import mock
 import py
 import pytest
 
 from rhodecode.lib import caching_query
 from rhodecode.lib import utils
 from rhodecode.lib.utils2 import md5
 from rhodecode.model import settings
 from rhodecode.model import db
 from rhodecode.model import meta
 from rhodecode.model.repo import RepoModel
 from rhodecode.model.repo_group import RepoGroupModel
 from rhodecode.model.scm import ScmModel
 from rhodecode.model.settings import UiSetting, SettingsModel
 from rhodecode.tests.fixture import Fixture
 
 
 fixture = Fixture()
 
 
 def extract_hooks(config):
     """Return a dictionary with the hook entries of the given config."""
     hooks = {}
     config_items = config.serialize()
     for section, name, value in config_items:
         if section != 'hooks':
             continue
         hooks[name] = value
 
     return hooks
 
 
 def disable_hooks(request, hooks):
     """Disables the given hooks from the UI settings."""
     session = meta.Session()
 
     model = SettingsModel()
     for hook_key in hooks:
         sett = model.get_ui_by_key(hook_key)
         sett.ui_active = False
         session.add(sett)
 
     # Invalidate cache
     ui_settings = session.query(db.RhodeCodeUi).options(
         caching_query.FromCache('sql_cache_short', 'get_hg_ui_settings'))
     ui_settings.invalidate()
 
     ui_settings = session.query(db.RhodeCodeUi).options(
         caching_query.FromCache('sql_cache_short', 'get_hook_settings'))
     ui_settings.invalidate()
 
     @request.addfinalizer
     def rollback():
         session.rollback()
 
 
 HOOK_PRE_PUSH = db.RhodeCodeUi.HOOK_PRE_PUSH
 HOOK_PRETX_PUSH = db.RhodeCodeUi.HOOK_PRETX_PUSH
 HOOK_PUSH = db.RhodeCodeUi.HOOK_PUSH
 HOOK_PRE_PULL = db.RhodeCodeUi.HOOK_PRE_PULL
 HOOK_PULL = db.RhodeCodeUi.HOOK_PULL
 HOOK_REPO_SIZE = db.RhodeCodeUi.HOOK_REPO_SIZE
 HOOK_PUSH_KEY = db.RhodeCodeUi.HOOK_PUSH_KEY
 
 HG_HOOKS = frozenset(
     (HOOK_PRE_PULL, HOOK_PULL, HOOK_PRE_PUSH, HOOK_PRETX_PUSH, HOOK_PUSH,
      HOOK_REPO_SIZE, HOOK_PUSH_KEY))
 
 
 @pytest.mark.parametrize('disabled_hooks,expected_hooks', [
     ([], HG_HOOKS),
     (HG_HOOKS, []),
 
     ([HOOK_PRE_PUSH, HOOK_PRETX_PUSH, HOOK_REPO_SIZE, HOOK_PUSH_KEY], [HOOK_PRE_PULL, HOOK_PULL, HOOK_PUSH]),
 
     # When a pull/push hook is disabled, its pre-pull/push counterpart should
     # be disabled too.
     ([HOOK_PUSH], [HOOK_PRE_PULL, HOOK_PULL, HOOK_REPO_SIZE]),
     ([HOOK_PULL], [HOOK_PRE_PUSH, HOOK_PRETX_PUSH, HOOK_PUSH, HOOK_REPO_SIZE,
                    HOOK_PUSH_KEY]),
 ])
 def test_make_db_config_hg_hooks(baseapp, request, disabled_hooks,
                                  expected_hooks):
     disable_hooks(request, disabled_hooks)
 
     config = utils.make_db_config()
     hooks = extract_hooks(config)
 
     assert set(hooks.iterkeys()).intersection(HG_HOOKS) == set(expected_hooks)
 
 
 @pytest.mark.parametrize('disabled_hooks,expected_hooks', [
     ([], ['pull', 'push']),
     ([HOOK_PUSH], ['pull']),
     ([HOOK_PULL], ['push']),
     ([HOOK_PULL, HOOK_PUSH], []),
 ])
 def test_get_enabled_hook_classes(disabled_hooks, expected_hooks):
     hook_keys = (HOOK_PUSH, HOOK_PULL)
     ui_settings = [
         ('hooks', key, 'some value', key not in disabled_hooks)
         for key in hook_keys]
 
     result = utils.get_enabled_hook_classes(ui_settings)
     assert sorted(result) == expected_hooks
 
 
 def test_get_filesystem_repos_finds_repos(tmpdir, baseapp):
     _stub_git_repo(tmpdir.ensure('repo', dir=True))
     repos = list(utils.get_filesystem_repos(str(tmpdir)))
     assert repos == [('repo', ('git', tmpdir.join('repo')))]
 
 
 def test_get_filesystem_repos_skips_directories(tmpdir, baseapp):
     tmpdir.ensure('not-a-repo', dir=True)
     repos = list(utils.get_filesystem_repos(str(tmpdir)))
     assert repos == []
 
 
 def test_get_filesystem_repos_skips_directories_with_repos(tmpdir, baseapp):
     _stub_git_repo(tmpdir.ensure('subdir/repo', dir=True))
     repos = list(utils.get_filesystem_repos(str(tmpdir)))
     assert repos == []
 
 
 def test_get_filesystem_repos_finds_repos_in_subdirectories(tmpdir, baseapp):
     _stub_git_repo(tmpdir.ensure('subdir/repo', dir=True))
     repos = list(utils.get_filesystem_repos(str(tmpdir), recursive=True))
     assert repos == [('subdir/repo', ('git', tmpdir.join('subdir', 'repo')))]
 
 
 def test_get_filesystem_repos_skips_names_starting_with_dot(tmpdir):
     _stub_git_repo(tmpdir.ensure('.repo', dir=True))
     repos = list(utils.get_filesystem_repos(str(tmpdir)))
     assert repos == []
 
 
 def test_get_filesystem_repos_skips_files(tmpdir):
     tmpdir.ensure('test-file')
     repos = list(utils.get_filesystem_repos(str(tmpdir)))
     assert repos == []
 
 
 def test_get_filesystem_repos_skips_removed_repositories(tmpdir):
     removed_repo_name = 'rm__00000000_000000_000000__.stub'
     assert utils.REMOVED_REPO_PAT.match(removed_repo_name)
     _stub_git_repo(tmpdir.ensure(removed_repo_name, dir=True))
     repos = list(utils.get_filesystem_repos(str(tmpdir)))
     assert repos == []
 
 
 def _stub_git_repo(repo_path):
     """
     Make `repo_path` look like a Git repository.
     """
     repo_path.ensure('.git', dir=True)
 
 
 @pytest.mark.parametrize('str_class', [str, unicode], ids=['str', 'unicode'])
 def test_get_dirpaths_returns_all_paths(tmpdir, str_class):
     tmpdir.ensure('test-file')
     dirpaths = utils._get_dirpaths(str_class(tmpdir))
     assert dirpaths == ['test-file']
 
 
 def test_get_dirpaths_returns_all_paths_bytes(
         tmpdir, platform_encodes_filenames):
     if platform_encodes_filenames:
         pytest.skip("This platform seems to encode filenames.")
     tmpdir.ensure('repo-a-umlaut-\xe4')
     dirpaths = utils._get_dirpaths(str(tmpdir))
     assert dirpaths == ['repo-a-umlaut-\xe4']
 
 
 def test_get_dirpaths_skips_paths_it_cannot_decode(
         tmpdir, platform_encodes_filenames):
     if platform_encodes_filenames:
         pytest.skip("This platform seems to encode filenames.")
     path_with_latin1 = 'repo-a-umlaut-\xe4'
     tmpdir.ensure(path_with_latin1)
     dirpaths = utils._get_dirpaths(unicode(tmpdir))
     assert dirpaths == []
 
 
 @pytest.fixture(scope='session')
 def platform_encodes_filenames():
     """
     Boolean indicator if the current platform changes filename encodings.
     """
     path_with_latin1 = 'repo-a-umlaut-\xe4'
     tmpdir = py.path.local.mkdtemp()
     tmpdir.ensure(path_with_latin1)
     read_path = tmpdir.listdir()[0].basename
     tmpdir.remove()
     return path_with_latin1 != read_path
 
 
 
 
 def test_repo2db_mapper_groups(repo_groups):
     session = meta.Session()
     zombie_group, parent_group, child_group = repo_groups
     zombie_path = os.path.join(
         RepoGroupModel().repos_path, zombie_group.full_path)
     os.rmdir(zombie_path)
 
     # Avoid removing test repos when calling repo2db_mapper
     repo_list = {
         repo.repo_name: 'test' for repo in session.query(db.Repository).all()
     }
     utils.repo2db_mapper(repo_list, remove_obsolete=True)
 
     groups_in_db = session.query(db.RepoGroup).all()
     assert child_group in groups_in_db
     assert parent_group in groups_in_db
     assert zombie_path not in groups_in_db
 
 
 def test_repo2db_mapper_enables_largefiles(backend):
     repo = backend.create_repo()
     repo_list = {repo.repo_name: 'test'}
     with mock.patch('rhodecode.model.db.Repository.scm_instance') as scm_mock:
         utils.repo2db_mapper(repo_list, remove_obsolete=False)
         _, kwargs = scm_mock.call_args
         assert kwargs['config'].get('extensions', 'largefiles') == ''
 
 
 @pytest.mark.backends("git", "svn")
 def test_repo2db_mapper_installs_hooks_for_repos_in_db(backend):
     repo = backend.create_repo()
     repo_list = {repo.repo_name: 'test'}
     utils.repo2db_mapper(repo_list, remove_obsolete=False)
 
 
 @pytest.mark.backends("git", "svn")
 def test_repo2db_mapper_installs_hooks_for_newly_added_repos(backend):
     repo = backend.create_repo()
     RepoModel().delete(repo, fs_remove=False)
     meta.Session().commit()
     repo_list = {repo.repo_name: repo.scm_instance()}
     utils.repo2db_mapper(repo_list, remove_obsolete=False)
 
 
 class TestPasswordChanged(object):
     def setup(self):
         self.session = {
             'rhodecode_user': {
                 'password': '0cc175b9c0f1b6a831c399e269772661'
             }
         }
         self.auth_user = mock.Mock()
         self.auth_user.userame = 'test'
         self.auth_user.password = 'abc123'
 
     def test_returns_false_for_default_user(self):
         self.auth_user.username = db.User.DEFAULT_USER
         result = utils.password_changed(self.auth_user, self.session)
         assert result is False
 
     def test_returns_false_if_password_was_not_changed(self):
         self.session['rhodecode_user']['password'] = md5(
             self.auth_user.password)
         result = utils.password_changed(self.auth_user, self.session)
         assert result is False
 
     def test_returns_true_if_password_was_changed(self):
         result = utils.password_changed(self.auth_user, self.session)
         assert result is True
 
     def test_returns_true_if_auth_user_password_is_empty(self):
         self.auth_user.password = None
         result = utils.password_changed(self.auth_user, self.session)
         assert result is True
 
     def test_returns_true_if_session_password_is_empty(self):
         self.session['rhodecode_user'].pop('password')
         result = utils.password_changed(self.auth_user, self.session)
         assert result is True
 
 
 class TestReadOpenSourceLicenses(object):
     def test_success(self):
         utils._license_cache = None
         json_data = '''
         {
             "python2.7-pytest-2.7.1": {"UNKNOWN": null},
             "python2.7-Markdown-2.6.2": {
                 "BSD-3-Clause": "http://spdx.org/licenses/BSD-3-Clause"
             }
         }
         '''
         resource_string_patch = mock.patch.object(
             utils.pkg_resources, 'resource_string', return_value=json_data)
         with resource_string_patch:
             result = utils.read_opensource_licenses()
         assert result == json.loads(json_data)
 
     def test_caching(self):
         utils._license_cache = {
             "python2.7-pytest-2.7.1": {
                 "UNKNOWN": None
             },
             "python2.7-Markdown-2.6.2": {
                 "BSD-3-Clause": "http://spdx.org/licenses/BSD-3-Clause"
             }
         }
         resource_patch = mock.patch.object(
             utils.pkg_resources, 'resource_string', side_effect=Exception)
         json_patch = mock.patch.object(
             utils.json, 'loads', side_effect=Exception)
 
         with resource_patch as resource_mock, json_patch as json_mock:
             result = utils.read_opensource_licenses()
 
         assert resource_mock.call_count == 0
         assert json_mock.call_count == 0
         assert result == utils._license_cache
 
     def test_licenses_file_contains_no_unknown_licenses(self):
         utils._license_cache = None
         result = utils.read_opensource_licenses()
 
         for license_data in result:
             if isinstance(license_data["license"], list):
                 for lic_data in license_data["license"]:
                     assert 'UNKNOWN' not in lic_data["fullName"]
             else:
                 full_name = license_data.get("fullName") or license_data
                 assert 'UNKNOWN' not in full_name
 
 
 class TestMakeDbConfig(object):
     def test_data_from_config_data_from_db_returned(self):
         test_data = [
             ('section1', 'option1', 'value1'),
             ('section2', 'option2', 'value2'),
             ('section3', 'option3', 'value3'),
         ]
         with mock.patch.object(utils, 'config_data_from_db') as config_mock:
             config_mock.return_value = test_data
             kwargs = {'clear_session': False, 'repo': 'test_repo'}
             result = utils.make_db_config(**kwargs)
         config_mock.assert_called_once_with(**kwargs)
         for section, option, expected_value in test_data:
             value = result.get(section, option)
             assert value == expected_value
 
 
 class TestConfigDataFromDb(object):
     def test_config_data_from_db_returns_active_settings(self):
         test_data = [
             UiSetting('section1', 'option1', 'value1', True),
             UiSetting('section2', 'option2', 'value2', True),
             UiSetting('section3', 'option3', 'value3', False),
         ]
         repo_name = 'test_repo'
 
         model_patch = mock.patch.object(settings, 'VcsSettingsModel')
         hooks_patch = mock.patch.object(
             utils, 'get_enabled_hook_classes',
             return_value=['pull', 'push', 'repo_size'])
         with model_patch as model_mock, hooks_patch:
             instance_mock = mock.Mock()
             model_mock.return_value = instance_mock
             instance_mock.get_ui_settings.return_value = test_data
             result = utils.config_data_from_db(
                 clear_session=False, repo=repo_name)
 
         self._assert_repo_name_passed(model_mock, repo_name)
 
         expected_result = [
             ('section1', 'option1', 'value1'),
             ('section2', 'option2', 'value2'),
         ]
         assert result == expected_result
 
     def _assert_repo_name_passed(self, model_mock, repo_name):
         assert model_mock.call_count == 1
         call_args, call_kwargs = model_mock.call_args
         assert call_kwargs['repo'] == repo_name
 
 
 class TestIsDirWritable(object):
     def test_returns_false_when_not_writable(self):
         with mock.patch('__builtin__.open', side_effect=OSError):
             assert not utils._is_dir_writable('/stub-path')
 
     def test_returns_true_when_writable(self, tmpdir):
         assert utils._is_dir_writable(str(tmpdir))
 
     def test_is_safe_against_race_conditions(self, tmpdir):
         workers = multiprocessing.Pool()
         directories = [str(tmpdir)] * 10
         workers.map(utils._is_dir_writable, directories)
 
 
 class TestGetEnabledHooks(object):
     def test_only_active_hooks_are_enabled(self):
         ui_settings = [
             UiSetting('hooks', db.RhodeCodeUi.HOOK_PUSH, 'value', True),
             UiSetting('hooks', db.RhodeCodeUi.HOOK_REPO_SIZE, 'value', True),
             UiSetting('hooks', db.RhodeCodeUi.HOOK_PULL, 'value', False)
         ]
         result = utils.get_enabled_hook_classes(ui_settings)
         assert result == ['push', 'repo_size']
 
     def test_all_hooks_are_enabled(self):
         ui_settings = [
             UiSetting('hooks', db.RhodeCodeUi.HOOK_PUSH, 'value', True),
             UiSetting('hooks', db.RhodeCodeUi.HOOK_REPO_SIZE, 'value', True),
             UiSetting('hooks', db.RhodeCodeUi.HOOK_PULL, 'value', True)
         ]
         result = utils.get_enabled_hook_classes(ui_settings)
         assert result == ['push', 'repo_size', 'pull']
 
     def test_no_enabled_hooks_when_no_hook_settings_are_found(self):
         ui_settings = []
         result = utils.get_enabled_hook_classes(ui_settings)
         assert result == []
 
 
 def test_obfuscate_url_pw():
     from rhodecode.lib.utils2 import obfuscate_url_pw
     engine = u'/home/repos/malmö'
-    assert obfuscate_url_pw(engine)
  No newline at end of file
+    assert obfuscate_url_pw(engine)
+
+
[e-mail]("test_ua, expected", [
+    ("", ""),
+    ('"quoted"', 'quoted'),
+    ('internal-merge', 'internal-merge'),
+    ('hg/internal-merge', 'hg/internal-merge'),
+    ('git/internal-merge', 'git/internal-merge'),
+
+    # git
+    ('git/2.10.1 (Apple Git-78)', 'git/2.10.1'),
+    ('GiT/2.37.2.windows.2', 'git/2.37.2'),
+    ('git/2.35.1 (Microsoft Windows NT 10.0.19044.0; Win32NT x64) CLR/4.0.30319 VS16/16.0.0', 'git/2.35.1'),
+    ('ssh-user-agent', 'ssh-user-agent'),
+    ('git/ssh-user-agent', 'git/ssh-user-agent'),
+
+
+    # hg
+    ('mercurial/proto-1.0 (Mercurial 4.2)', 'mercurial/4.2'),
+    ('mercurial/proto-1.0', ''),
+    ('mercurial/proto-1.0 (Mercurial 3.9.2)', 'mercurial/3.9.2'),
+    ('mercurial/ssh-user-agent', 'mercurial/ssh-user-agent'),
+    ('mercurial/proto-1.0 (Mercurial 5.8rc0)', 'mercurial/5.8rc0'),
+
+
+])
+def test_user_agent_normalizer(test_ua, expected):
+    from rhodecode.lib.utils2 import user_agent_normalizer
+    assert user_agent_normalizer(test_ua, safe=False) == expected