markup: make relative links pint to raw files for images and to standard files as links.

- allows building relative MD/RST links that go to rendered content
- still allows to embed images/video with relative paths

rhodecode/lib/helpers.py

@@ -1,2045 +1,2045 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2010-2017 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/
 
 """
 Helper functions
 
 Consists of functions to typically be used within templates, but also
 available to Controllers. This module is available to both as 'h'.
 """
 
 import random
 import hashlib
 import StringIO
 import urllib
 import math
 import logging
 import re
 import urlparse
 import time
 import string
 import hashlib
 from collections import OrderedDict
 
 import pygments
 import itertools
 import fnmatch
 
 from datetime import datetime
 from functools import partial
 from pygments.formatters.html import HtmlFormatter
 from pygments import highlight as code_highlight
 from pygments.lexers import (
     get_lexer_by_name, get_lexer_for_filename, get_lexer_for_mimetype)
 
 from pyramid.threadlocal import get_current_request
 
 from webhelpers.html import literal, HTML, escape
 from webhelpers.html.tools import *
 from webhelpers.html.builder import make_tag
 from webhelpers.html.tags import auto_discovery_link, checkbox, css_classes, \
     end_form, file, form as wh_form, hidden, image, javascript_link, link_to, \
     link_to_if, link_to_unless, ol, required_legend, select, stylesheet_link, \
     submit, text, password, textarea, title, ul, xml_declaration, radio
 from webhelpers.html.tools import auto_link, button_to, highlight, \
     js_obfuscate, mail_to, strip_links, strip_tags, tag_re
 from webhelpers.pylonslib import Flash as _Flash
 from webhelpers.text import chop_at, collapse, convert_accented_entities, \
     convert_misc_entities, lchop, plural, rchop, remove_formatting, \
     replace_whitespace, urlify, truncate, wrap_paragraphs
 from webhelpers.date import time_ago_in_words
 from webhelpers.paginate import Page as _Page
 from webhelpers.html.tags import _set_input_attrs, _set_id_attr, \
     convert_boolean_attrs, NotGiven, _make_safe_id_component
 from webhelpers2.number import format_byte_size
 
 from rhodecode.lib.action_parser import action_parser
 from rhodecode.lib.ext_json import json
 from rhodecode.lib.utils import repo_name_slug, get_custom_lexer
 from rhodecode.lib.utils2 import str2bool, safe_unicode, safe_str, \
     get_commit_safe, datetime_to_time, time_to_datetime, time_to_utcdatetime, \
     AttributeDict, safe_int, md5, md5_safe
 from rhodecode.lib.markup_renderer import MarkupRenderer, relative_links
 from rhodecode.lib.vcs.exceptions import CommitDoesNotExistError
 from rhodecode.lib.vcs.backends.base import BaseChangeset, EmptyCommit
 from rhodecode.config.conf import DATE_FORMAT, DATETIME_FORMAT
 from rhodecode.model.changeset_status import ChangesetStatusModel
 from rhodecode.model.db import Permission, User, Repository
 from rhodecode.model.repo_group import RepoGroupModel
 from rhodecode.model.settings import IssueTrackerSettingsModel
 
 log = logging.getLogger(__name__)
 
 
 DEFAULT_USER = User.DEFAULT_USER
 DEFAULT_USER_EMAIL = User.DEFAULT_USER_EMAIL
 
 
 def url(*args, **kw):
     from pylons import url as pylons_url
     return pylons_url(*args, **kw)
 
 
 def pylons_url_current(*args, **kw):
     """
     This function overrides pylons.url.current() which returns the current
     path so that it will also work from a pyramid only context. This
     should be removed once port to pyramid is complete.
     """
     from pylons import url as pylons_url
     if not args and not kw:
         request = get_current_request()
         return request.path
     return pylons_url.current(*args, **kw)
 
 url.current = pylons_url_current
 
 
 def url_replace(**qargs):
     """ Returns the current request url while replacing query string args """
 
     request = get_current_request()
     new_args = request.GET.mixed()
     new_args.update(qargs)
     return url('', **new_args)
 
 
 def asset(path, ver=None, **kwargs):
     """
     Helper to generate a static asset file path for rhodecode assets
 
     eg. h.asset('images/image.png', ver='3923')
 
     :param path: path of asset
     :param ver: optional version query param to append as ?ver=
     """
     request = get_current_request()
     query = {}
     query.update(kwargs)
     if ver:
         query = {'ver': ver}
     return request.static_path(
         'rhodecode:public/{}'.format(path), _query=query)
 
 
 default_html_escape_table = {
     ord('&'): u'&amp;',
     ord('<'): u'&lt;',
     ord('>'): u'&gt;',
     ord('"'): u'&quot;',
     ord("'"): u'&#39;',
 }
 
 
 def html_escape(text, html_escape_table=default_html_escape_table):
     """Produce entities within text."""
     return text.translate(html_escape_table)
 
 
 def chop_at_smart(s, sub, inclusive=False, suffix_if_chopped=None):
     """
     Truncate string ``s`` at the first occurrence of ``sub``.
 
     If ``inclusive`` is true, truncate just after ``sub`` rather than at it.
     """
     suffix_if_chopped = suffix_if_chopped or ''
     pos = s.find(sub)
     if pos == -1:
         return s
 
     if inclusive:
         pos += len(sub)
 
     chopped = s[:pos]
     left = s[pos:].strip()
 
     if left and suffix_if_chopped:
         chopped += suffix_if_chopped
 
     return chopped
 
 
 def shorter(text, size=20):
     postfix = '...'
     if len(text) > size:
         return text[:size - len(postfix)] + postfix
     return text
 
 
 def _reset(name, value=None, id=NotGiven, type="reset", **attrs):
     """
     Reset button
     """
     _set_input_attrs(attrs, type, name, value)
     _set_id_attr(attrs, id, name)
     convert_boolean_attrs(attrs, ["disabled"])
     return HTML.input(**attrs)
 
 reset = _reset
 safeid = _make_safe_id_component
 
 
 def branding(name, length=40):
     return truncate(name, length, indicator="")
 
 
 def FID(raw_id, path):
     """
     Creates a unique ID for filenode based on it's hash of path and commit
     it's safe to use in urls
 
     :param raw_id:
     :param path:
     """
 
     return 'c-%s-%s' % (short_id(raw_id), md5_safe(path)[:12])
 
 
 class _GetError(object):
     """Get error from form_errors, and represent it as span wrapped error
     message
 
     :param field_name: field to fetch errors for
     :param form_errors: form errors dict
     """
 
     def __call__(self, field_name, form_errors):
         tmpl = """<span class="error_msg">%s</span>"""
         if form_errors and field_name in form_errors:
             return literal(tmpl % form_errors.get(field_name))
 
 get_error = _GetError()
 
 
 class _ToolTip(object):
 
     def __call__(self, tooltip_title, trim_at=50):
         """
         Special function just to wrap our text into nice formatted
         autowrapped text
 
         :param tooltip_title:
         """
         tooltip_title = escape(tooltip_title)
         tooltip_title = tooltip_title.replace('<', '&lt;').replace('>', '&gt;')
         return tooltip_title
 tooltip = _ToolTip()
 
 
 def files_breadcrumbs(repo_name, commit_id, file_path):
     if isinstance(file_path, str):
         file_path = safe_unicode(file_path)
 
     # TODO: johbo: Is this always a url like path, or is this operating
     # system dependent?
     path_segments = file_path.split('/')
 
     repo_name_html = escape(repo_name)
     if len(path_segments) == 1 and path_segments[0] == '':
         url_segments = [repo_name_html]
     else:
         url_segments = [
             link_to(
                 repo_name_html,
                 route_path(
                     'repo_files',
                     repo_name=repo_name,
                     commit_id=commit_id,
                     f_path=''),
                 class_='pjax-link')]
 
     last_cnt = len(path_segments) - 1
     for cnt, segment in enumerate(path_segments):
         if not segment:
             continue
         segment_html = escape(segment)
 
         if cnt != last_cnt:
             url_segments.append(
                 link_to(
                     segment_html,
                     route_path(
                         'repo_files',
                         repo_name=repo_name,
                         commit_id=commit_id,
                         f_path='/'.join(path_segments[:cnt + 1])),
                     class_='pjax-link'))
         else:
             url_segments.append(segment_html)
 
     return literal('/'.join(url_segments))
 
 
 class CodeHtmlFormatter(HtmlFormatter):
     """
     My code Html Formatter for source codes
     """
 
     def wrap(self, source, outfile):
         return self._wrap_div(self._wrap_pre(self._wrap_code(source)))
 
     def _wrap_code(self, source):
         for cnt, it in enumerate(source):
             i, t = it
             t = '<div id="L%s">%s</div>' % (cnt + 1, t)
             yield i, t
 
     def _wrap_tablelinenos(self, inner):
         dummyoutfile = StringIO.StringIO()
         lncount = 0
         for t, line in inner:
             if t:
                 lncount += 1
             dummyoutfile.write(line)
 
         fl = self.linenostart
         mw = len(str(lncount + fl - 1))
         sp = self.linenospecial
         st = self.linenostep
         la = self.lineanchors
         aln = self.anchorlinenos
         nocls = self.noclasses
         if sp:
             lines = []
 
             for i in range(fl, fl + lncount):
                 if i % st == 0:
                     if i % sp == 0:
                         if aln:
                             lines.append('<a href="#%s%d" class="special">%*d</a>' %
                                          (la, i, mw, i))
                         else:
                             lines.append('<span class="special">%*d</span>' % (mw, i))
                     else:
                         if aln:
                             lines.append('<a href="#%s%d">%*d</a>' % (la, i, mw, i))
                         else:
                             lines.append('%*d' % (mw, i))
                 else:
                     lines.append('')
             ls = '\n'.join(lines)
         else:
             lines = []
             for i in range(fl, fl + lncount):
                 if i % st == 0:
                     if aln:
                         lines.append('<a href="#%s%d">%*d</a>' % (la, i, mw, i))
                     else:
                         lines.append('%*d' % (mw, i))
                 else:
                     lines.append('')
             ls = '\n'.join(lines)
 
         # in case you wonder about the seemingly redundant <div> here: since the
         # content in the other cell also is wrapped in a div, some browsers in
         # some configurations seem to mess up the formatting...
         if nocls:
             yield 0, ('<table class="%stable">' % self.cssclass +
                       '<tr><td><div class="linenodiv" '
                       'style="background-color: #f0f0f0; padding-right: 10px">'
                       '<pre style="line-height: 125%">' +
                       ls + '</pre></div></td><td id="hlcode" class="code">')
         else:
             yield 0, ('<table class="%stable">' % self.cssclass +
                       '<tr><td class="linenos"><div class="linenodiv"><pre>' +
                       ls + '</pre></div></td><td id="hlcode" class="code">')
         yield 0, dummyoutfile.getvalue()
         yield 0, '</td></tr></table>'
 
 
 class SearchContentCodeHtmlFormatter(CodeHtmlFormatter):
     def __init__(self, **kw):
         # only show these line numbers if set
         self.only_lines = kw.pop('only_line_numbers', [])
         self.query_terms = kw.pop('query_terms', [])
         self.max_lines = kw.pop('max_lines', 5)
         self.line_context = kw.pop('line_context', 3)
         self.url = kw.pop('url', None)
 
         super(CodeHtmlFormatter, self).__init__(**kw)
 
     def _wrap_code(self, source):
         for cnt, it in enumerate(source):
             i, t = it
             t = '<pre>%s</pre>' % t
             yield i, t
 
     def _wrap_tablelinenos(self, inner):
         yield 0, '<table class="code-highlight %stable">' % self.cssclass
 
         last_shown_line_number = 0
         current_line_number = 1
 
         for t, line in inner:
             if not t:
                 yield t, line
                 continue
 
             if current_line_number in self.only_lines:
                 if last_shown_line_number + 1 != current_line_number:
                     yield 0, '<tr>'
                     yield 0, '<td class="line">...</td>'
                     yield 0, '<td id="hlcode" class="code"></td>'
                     yield 0, '</tr>'
 
                 yield 0, '<tr>'
                 if self.url:
                     yield 0, '<td class="line"><a href="%s#L%i">%i</a></td>' % (
                         self.url, current_line_number, current_line_number)
                 else:
                     yield 0, '<td class="line"><a href="">%i</a></td>' % (
                         current_line_number)
                 yield 0, '<td id="hlcode" class="code">' + line + '</td>'
                 yield 0, '</tr>'
 
                 last_shown_line_number = current_line_number
 
             current_line_number += 1
 
 
         yield 0, '</table>'
 
 
 def extract_phrases(text_query):
     """
     Extracts phrases from search term string making sure phrases
     contained in double quotes are kept together - and discarding empty values
     or fully whitespace values eg.
 
     'some   text "a phrase" more' => ['some', 'text', 'a phrase', 'more']
 
     """
 
     in_phrase = False
     buf = ''
     phrases = []
     for char in text_query:
         if in_phrase:
             if char == '"': # end phrase
                 phrases.append(buf)
                 buf = ''
                 in_phrase = False
                 continue
             else:
                 buf += char
                 continue
         else:
             if char == '"': # start phrase
                 in_phrase = True
                 phrases.append(buf)
                 buf = ''
                 continue
             elif char == ' ':
                 phrases.append(buf)
                 buf = ''
                 continue
             else:
                 buf += char
 
     phrases.append(buf)
     phrases = [phrase.strip() for phrase in phrases if phrase.strip()]
     return phrases
 
 
 def get_matching_offsets(text, phrases):
     """
     Returns a list of string offsets in `text` that the list of `terms` match
 
     >>> get_matching_offsets('some text here', ['some', 'here'])
     [(0, 4), (10, 14)]
 
     """
     offsets = []
     for phrase in phrases:
         for match in re.finditer(phrase, text):
             offsets.append((match.start(), match.end()))
 
     return offsets
 
 
 def normalize_text_for_matching(x):
     """
     Replaces all non alnum characters to spaces and lower cases the string,
     useful for comparing two text strings without punctuation
     """
     return re.sub(r'[^\w]', ' ', x.lower())
 
 
 def get_matching_line_offsets(lines, terms):
     """ Return a set of `lines` indices (starting from 1) matching a
     text search query, along with `context` lines above/below matching lines
 
     :param lines: list of strings representing lines
     :param terms: search term string to match in lines eg. 'some text'
     :param context: number of lines above/below a matching line to add to result
     :param max_lines: cut off for lines of interest
      eg.
 
     text = '''
     words words words
     words words words
     some text some
     words words words
     words words words
     text here what
     '''
     get_matching_line_offsets(text, 'text', context=1)
     {3: [(5, 9)], 6: [(0, 4)]]
 
     """
     matching_lines = {}
     phrases = [normalize_text_for_matching(phrase)
                for phrase in extract_phrases(terms)]
 
     for line_index, line in enumerate(lines, start=1):
         match_offsets = get_matching_offsets(
             normalize_text_for_matching(line), phrases)
         if match_offsets:
             matching_lines[line_index] = match_offsets
 
     return matching_lines
 
 
 def hsv_to_rgb(h, s, v):
     """ Convert hsv color values to rgb """
 
     if s == 0.0:
         return v, v, v
     i = int(h * 6.0)  # XXX assume int() truncates!
     f = (h * 6.0) - i
     p = v * (1.0 - s)
     q = v * (1.0 - s * f)
     t = v * (1.0 - s * (1.0 - f))
     i = i % 6
     if i == 0:
         return v, t, p
     if i == 1:
         return q, v, p
     if i == 2:
         return p, v, t
     if i == 3:
         return p, q, v
     if i == 4:
         return t, p, v
     if i == 5:
         return v, p, q
 
 
 def unique_color_generator(n=10000, saturation=0.10, lightness=0.95):
     """
     Generator for getting n of evenly distributed colors using
     hsv color and golden ratio. It always return same order of colors
 
     :param n: number of colors to generate
     :param saturation: saturation of returned colors
     :param lightness: lightness of returned colors
     :returns: RGB tuple
     """
 
     golden_ratio = 0.618033988749895
     h = 0.22717784590367374
 
     for _ in xrange(n):
         h += golden_ratio
         h %= 1
         HSV_tuple = [h, saturation, lightness]
         RGB_tuple = hsv_to_rgb(*HSV_tuple)
         yield map(lambda x: str(int(x * 256)), RGB_tuple)
 
 
 def color_hasher(n=10000, saturation=0.10, lightness=0.95):
     """
     Returns a function which when called with an argument returns a unique
     color for that argument, eg.
 
     :param n: number of colors to generate
     :param saturation: saturation of returned colors
     :param lightness: lightness of returned colors
     :returns: css RGB string
 
     >>> color_hash = color_hasher()
     >>> color_hash('hello')
     'rgb(34, 12, 59)'
     >>> color_hash('hello')
     'rgb(34, 12, 59)'
     >>> color_hash('other')
     'rgb(90, 224, 159)'
     """
 
     color_dict = {}
     cgenerator = unique_color_generator(
         saturation=saturation, lightness=lightness)
 
     def get_color_string(thing):
         if thing in color_dict:
             col = color_dict[thing]
         else:
             col = color_dict[thing] = cgenerator.next()
         return "rgb(%s)" % (', '.join(col))
 
     return get_color_string
 
 
 def get_lexer_safe(mimetype=None, filepath=None):
     """
     Tries to return a relevant pygments lexer using mimetype/filepath name,
     defaulting to plain text if none could be found
     """
     lexer = None
     try:
         if mimetype:
             lexer = get_lexer_for_mimetype(mimetype)
         if not lexer:
             lexer = get_lexer_for_filename(filepath)
     except pygments.util.ClassNotFound:
         pass
 
     if not lexer:
         lexer = get_lexer_by_name('text')
 
     return lexer
 
 
 def get_lexer_for_filenode(filenode):
     lexer = get_custom_lexer(filenode.extension) or filenode.lexer
     return lexer
 
 
 def pygmentize(filenode, **kwargs):
     """
     pygmentize function using pygments
 
     :param filenode:
     """
     lexer = get_lexer_for_filenode(filenode)
     return literal(code_highlight(filenode.content, lexer,
                                   CodeHtmlFormatter(**kwargs)))
 
 
 def is_following_repo(repo_name, user_id):
     from rhodecode.model.scm import ScmModel
     return ScmModel().is_following_repo(repo_name, user_id)
 
 
 class _Message(object):
     """A message returned by ``Flash.pop_messages()``.
 
     Converting the message to a string returns the message text. Instances
     also have the following attributes:
 
     * ``message``: the message text.
     * ``category``: the category specified when the message was created.
     """
 
     def __init__(self, category, message):
         self.category = category
         self.message = message
 
     def __str__(self):
         return self.message
 
     __unicode__ = __str__
 
     def __html__(self):
         return escape(safe_unicode(self.message))
 
 
 class Flash(_Flash):
 
     def pop_messages(self, request=None):
         """Return all accumulated messages and delete them from the session.
 
         The return value is a list of ``Message`` objects.
         """
         messages = []
 
         if request:
             session = request.session
         else:
             from pylons import session
 
         # Pop the 'old' pylons flash messages. They are tuples of the form
         # (category, message)
         for cat, msg in session.pop(self.session_key, []):
             messages.append(_Message(cat, msg))
 
         # Pop the 'new' pyramid flash messages for each category as list
         # of strings.
         for cat in self.categories:
             for msg in session.pop_flash(queue=cat):
                 messages.append(_Message(cat, msg))
         # Map messages from the default queue to the 'notice' category.
         for msg in session.pop_flash():
             messages.append(_Message('notice', msg))
 
         session.save()
         return messages
 
     def json_alerts(self, request=None):
         payloads = []
         messages = flash.pop_messages(request=request)
         if messages:
             for message in messages:
                 subdata = {}
                 if hasattr(message.message, 'rsplit'):
                     flash_data = message.message.rsplit('|DELIM|', 1)
                     org_message = flash_data[0]
                     if len(flash_data) > 1:
                         subdata = json.loads(flash_data[1])
                 else:
                     org_message = message.message
                 payloads.append({
                     'message': {
                         'message': u'{}'.format(org_message),
                         'level': message.category,
                         'force': True,
                         'subdata': subdata
                     }
                 })
         return json.dumps(payloads)
 
 flash = Flash()
 
 #==============================================================================
 # SCM FILTERS available via h.
 #==============================================================================
 from rhodecode.lib.vcs.utils import author_name, author_email
 from rhodecode.lib.utils2 import credentials_filter, age as _age
 from rhodecode.model.db import User, ChangesetStatus
 
 age = _age
 capitalize = lambda x: x.capitalize()
 email = author_email
 short_id = lambda x: x[:12]
 hide_credentials = lambda x: ''.join(credentials_filter(x))
 
 
 def age_component(datetime_iso, value=None, time_is_local=False):
     title = value or format_date(datetime_iso)
     tzinfo = '+00:00'
 
     # detect if we have a timezone info, otherwise, add it
     if isinstance(datetime_iso, datetime) and not datetime_iso.tzinfo:
         if time_is_local:
             tzinfo = time.strftime("+%H:%M",
                 time.gmtime(
                     (datetime.now() - datetime.utcnow()).seconds + 1
                 )
             )
 
     return literal(
         '<time class="timeago tooltip" '
         'title="{1}{2}" datetime="{0}{2}">{1}</time>'.format(
             datetime_iso, title, tzinfo))
 
 
 def _shorten_commit_id(commit_id):
     from rhodecode import CONFIG
     def_len = safe_int(CONFIG.get('rhodecode_show_sha_length', 12))
     return commit_id[:def_len]
 
 
 def show_id(commit):
     """
     Configurable function that shows ID
     by default it's r123:fffeeefffeee
 
     :param commit: commit instance
     """
     from rhodecode import CONFIG
     show_idx = str2bool(CONFIG.get('rhodecode_show_revision_number', True))
 
     raw_id = _shorten_commit_id(commit.raw_id)
     if show_idx:
         return 'r%s:%s' % (commit.idx, raw_id)
     else:
         return '%s' % (raw_id, )
 
 
 def format_date(date):
     """
     use a standardized formatting for dates used in RhodeCode
 
     :param date: date/datetime object
     :return: formatted date
     """
 
     if date:
         _fmt = "%a, %d %b %Y %H:%M:%S"
         return safe_unicode(date.strftime(_fmt))
 
     return u""
 
 
 class _RepoChecker(object):
 
     def __init__(self, backend_alias):
         self._backend_alias = backend_alias
 
     def __call__(self, repository):
         if hasattr(repository, 'alias'):
             _type = repository.alias
         elif hasattr(repository, 'repo_type'):
             _type = repository.repo_type
         else:
             _type = repository
         return _type == self._backend_alias
 
 is_git = _RepoChecker('git')
 is_hg = _RepoChecker('hg')
 is_svn = _RepoChecker('svn')
 
 
 def get_repo_type_by_name(repo_name):
     repo = Repository.get_by_repo_name(repo_name)
     return repo.repo_type
 
 
 def is_svn_without_proxy(repository):
     if is_svn(repository):
         from rhodecode.model.settings import VcsSettingsModel
         conf = VcsSettingsModel().get_ui_settings_as_config_obj()
         return not str2bool(conf.get('vcs_svn_proxy', 'http_requests_enabled'))
     return False
 
 
 def discover_user(author):
     """
     Tries to discover RhodeCode User based on the autho string. Author string
     is typically `FirstName LastName <[e-mail]>`
     """
 
     # if author is already an instance use it for extraction
     if isinstance(author, User):
         return author
 
     # Valid email in the attribute passed, see if they're in the system
     _email = author_email(author)
     if _email != '':
         user = User.get_by_email(_email, case_insensitive=True, cache=True)
         if user is not None:
             return user
 
     # Maybe it's a username, we try to extract it and fetch by username ?
     _author = author_name(author)
     user = User.get_by_username(_author, case_insensitive=True, cache=True)
     if user is not None:
         return user
 
     return None
 
 
 def email_or_none(author):
     # extract email from the commit string
     _email = author_email(author)
 
     # If we have an email, use it, otherwise
     # see if it contains a username we can get an email from
     if _email != '':
         return _email
     else:
         user = User.get_by_username(
             author_name(author), case_insensitive=True, cache=True)
 
     if user is not None:
             return user.email
 
     # No valid email, not a valid user in the system, none!
     return None
 
 
 def link_to_user(author, length=0, **kwargs):
     user = discover_user(author)
     # user can be None, but if we have it already it means we can re-use it
     # in the person() function, so we save 1 intensive-query
     if user:
         author = user
 
     display_person = person(author, 'username_or_name_or_email')
     if length:
         display_person = shorter(display_person, length)
 
     if user:
         return link_to(
             escape(display_person),
             route_path('user_profile', username=user.username),
             **kwargs)
     else:
         return escape(display_person)
 
 
 def person(author, show_attr="username_and_name"):
     user = discover_user(author)
     if user:
         return getattr(user, show_attr)
     else:
         _author = author_name(author)
         _email = email(author)
         return _author or _email
 
 
 def author_string(email):
     if email:
         user = User.get_by_email(email, case_insensitive=True, cache=True)
         if user:
             if user.first_name or user.last_name:
                 return '%s %s &lt;%s&gt;' % (
                     user.first_name, user.last_name, email)
             else:
                 return email
         else:
             return email
     else:
         return None
 
 
 def person_by_id(id_, show_attr="username_and_name"):
     # attr to return from fetched user
     person_getter = lambda usr: getattr(usr, show_attr)
 
     #maybe it's an ID ?
     if str(id_).isdigit() or isinstance(id_, int):
         id_ = int(id_)
         user = User.get(id_)
         if user is not None:
             return person_getter(user)
     return id_
 
 
 def gravatar_with_user(request, author, show_disabled=False):
     _render = request.get_partial_renderer('base/base.mako')
     return _render('gravatar_with_user', author, show_disabled=show_disabled)
 
 
 def desc_stylize(value):
     """
     converts tags from value into html equivalent
 
     :param value:
     """
     if not value:
         return ''
 
     value = re.sub(r'\[see\ \=\>\ *([a-zA-Z0-9\/\=\?\&\ \:\/\.\-]*)\]',
                    '<div class="metatag" tag="see">see =&gt; \\1 </div>', value)
     value = re.sub(r'\[license\ \=\>\ *([a-zA-Z0-9\/\=\?\&\ \:\/\.\-]*)\]',
                    '<div class="metatag" tag="license"><a href="http:\/\/www.opensource.org/licenses/\\1">\\1</a></div>', value)
     value = re.sub(r'\[(requires|recommends|conflicts|base)\ \=\>\ *([a-zA-Z0-9\-\/]*)\]',
                    '<div class="metatag" tag="\\1">\\1 =&gt; <a href="/\\2">\\2</a></div>', value)
     value = re.sub(r'\[(lang|language)\ \=\>\ *([a-zA-Z\-\/\#\+]*)\]',
                    '<div class="metatag" tag="lang">\\2</div>', value)
     value = re.sub(r'\[([a-z]+)\]',
                    '<div class="metatag" tag="\\1">\\1</div>', value)
 
     return value
 
 
 def escaped_stylize(value):
     """
     converts tags from value into html equivalent, but escaping its value first
     """
     if not value:
         return ''
 
     # Using default webhelper escape method, but has to force it as a
     # plain unicode instead of a markup tag to be used in regex expressions
     value = unicode(escape(safe_unicode(value)))
 
     value = re.sub(r'\[see\ \=\&gt;\ *([a-zA-Z0-9\/\=\?\&amp;\ \:\/\.\-]*)\]',
                    '<div class="metatag" tag="see">see =&gt; \\1 </div>', value)
     value = re.sub(r'\[license\ \=\&gt;\ *([a-zA-Z0-9\/\=\?\&amp;\ \:\/\.\-]*)\]',
                    '<div class="metatag" tag="license"><a href="http:\/\/www.opensource.org/licenses/\\1">\\1</a></div>', value)
     value = re.sub(r'\[(requires|recommends|conflicts|base)\ \=\&gt;\ *([a-zA-Z0-9\-\/]*)\]',
                    '<div class="metatag" tag="\\1">\\1 =&gt; <a href="/\\2">\\2</a></div>', value)
     value = re.sub(r'\[(lang|language)\ \=\&gt;\ *([a-zA-Z\-\/\#\+]*)\]',
                    '<div class="metatag" tag="lang">\\2</div>', value)
     value = re.sub(r'\[([a-z]+)\]',
                    '<div class="metatag" tag="\\1">\\1</div>', value)
 
     return value
 
 
 def bool2icon(value):
     """
     Returns boolean value of a given value, represented as html element with
     classes that will represent icons
 
     :param value: given value to convert to html node
     """
 
     if value:  # does bool conversion
         return HTML.tag('i', class_="icon-true")
     else:  # not true as bool
         return HTML.tag('i', class_="icon-false")
 
 
 #==============================================================================
 # PERMS
 #==============================================================================
 from rhodecode.lib.auth import HasPermissionAny, HasPermissionAll, \
 HasRepoPermissionAny, HasRepoPermissionAll, HasRepoGroupPermissionAll, \
 HasRepoGroupPermissionAny, HasRepoPermissionAnyApi, get_csrf_token, \
 csrf_token_key
 
 
 #==============================================================================
 # GRAVATAR URL
 #==============================================================================
 class InitialsGravatar(object):
     def __init__(self, email_address, first_name, last_name, size=30,
                  background=None, text_color='#fff'):
         self.size = size
         self.first_name = first_name
         self.last_name = last_name
         self.email_address = email_address
         self.background = background or self.str2color(email_address)
         self.text_color = text_color
 
     def get_color_bank(self):
         """
         returns a predefined list of colors that gravatars can use.
         Those are randomized distinct colors that guarantee readability and
         uniqueness.
 
         generated with: http://phrogz.net/css/distinct-colors.html
         """
         return [
             '#bf3030', '#a67f53', '#00ff00', '#5989b3', '#392040', '#d90000',
             '#402910', '#204020', '#79baf2', '#a700b3', '#bf6060', '#7f5320',
             '#008000', '#003059', '#ee00ff', '#ff0000', '#8c4b00', '#007300',
             '#005fb3', '#de73e6', '#ff4040', '#ffaa00', '#3df255', '#203140',
             '#47004d', '#591616', '#664400', '#59b365', '#0d2133', '#83008c',
             '#592d2d', '#bf9f60', '#73e682', '#1d3f73', '#73006b', '#402020',
             '#b2862d', '#397341', '#597db3', '#e600d6', '#a60000', '#736039',
             '#00b318', '#79aaf2', '#330d30', '#ff8080', '#403010', '#16591f',
             '#002459', '#8c4688', '#e50000', '#ffbf40', '#00732e', '#102340',
             '#bf60ac', '#8c4646', '#cc8800', '#00a642', '#1d3473', '#b32d98',
             '#660e00', '#ffd580', '#80ffb2', '#7391e6', '#733967', '#d97b6c',
             '#8c5e00', '#59b389', '#3967e6', '#590047', '#73281d', '#665200',
             '#00e67a', '#2d50b3', '#8c2377', '#734139', '#b2982d', '#16593a',
             '#001859', '#ff00aa', '#a65e53', '#ffcc00', '#0d3321', '#2d3959',
             '#731d56', '#401610', '#4c3d00', '#468c6c', '#002ca6', '#d936a3',
             '#d94c36', '#403920', '#36d9a3', '#0d1733', '#592d4a', '#993626',
             '#cca300', '#00734d', '#46598c', '#8c005e', '#7f1100', '#8c7000',
             '#00a66f', '#7382e6', '#b32d74', '#d9896c', '#ffe680', '#1d7362',
             '#364cd9', '#73003d', '#d93a00', '#998a4d', '#59b3a1', '#5965b3',
             '#e5007a', '#73341d', '#665f00', '#00b38f', '#0018b3', '#59163a',
             '#b2502d', '#bfb960', '#00ffcc', '#23318c', '#a6537f', '#734939',
             '#b2a700', '#104036', '#3d3df2', '#402031', '#e56739', '#736f39',
             '#79f2ea', '#000059', '#401029', '#4c1400', '#ffee00', '#005953',
             '#101040', '#990052', '#402820', '#403d10', '#00ffee', '#0000d9',
             '#ff80c4', '#a66953', '#eeff00', '#00ccbe', '#8080ff', '#e673a1',
             '#a62c00', '#474d00', '#1a3331', '#46468c', '#733950', '#662900',
             '#858c23', '#238c85', '#0f0073', '#b20047', '#d9986c', '#becc00',
             '#396f73', '#281d73', '#ff0066', '#ff6600', '#dee673', '#59adb3',
             '#6559b3', '#590024', '#b2622d', '#98b32d', '#36ced9', '#332d59',
             '#40001a', '#733f1d', '#526600', '#005359', '#242040', '#bf6079',
             '#735039', '#cef23d', '#007780', '#5630bf', '#66001b', '#b24700',
             '#acbf60', '#1d6273', '#25008c', '#731d34', '#a67453', '#50592d',
             '#00ccff', '#6600ff', '#ff0044', '#4c1f00', '#8a994d', '#79daf2',
             '#a173e6', '#d93662', '#402310', '#aaff00', '#2d98b3', '#8c40ff',
             '#592d39', '#ff8c40', '#354020', '#103640', '#1a0040', '#331a20',
             '#331400', '#334d00', '#1d5673', '#583973', '#7f0022', '#4c3626',
             '#88cc00', '#36a3d9', '#3d0073', '#d9364c', '#33241a', '#698c23',
             '#5995b3', '#300059', '#e57382', '#7f3300', '#366600', '#00aaff',
             '#3a1659', '#733941', '#663600', '#74b32d', '#003c59', '#7f53a6',
             '#73000f', '#ff8800', '#baf279', '#79caf2', '#291040', '#a6293a',
             '#b2742d', '#587339', '#0077b3', '#632699', '#400009', '#d9a66c',
             '#294010', '#2d4a59', '#aa00ff', '#4c131b', '#b25f00', '#5ce600',
             '#267399', '#a336d9', '#990014', '#664e33', '#86bf60', '#0088ff',
             '#7700b3', '#593a16', '#073300', '#1d4b73', '#ac60bf', '#e59539',
             '#4f8c46', '#368dd9', '#5c0073'
         ]
 
     def rgb_to_hex_color(self, rgb_tuple):
         """
         Converts an rgb_tuple passed to an hex color.
 
         :param rgb_tuple: tuple with 3 ints represents rgb color space
         """
         return '#' + ("".join(map(chr, rgb_tuple)).encode('hex'))
 
     def email_to_int_list(self, email_str):
         """
         Get every byte of the hex digest value of email and turn it to integer.
         It's going to be always between 0-255
         """
         digest = md5_safe(email_str.lower())
         return [int(digest[i * 2:i * 2 + 2], 16) for i in range(16)]
 
     def pick_color_bank_index(self, email_str, color_bank):
         return self.email_to_int_list(email_str)[0] % len(color_bank)
 
     def str2color(self, email_str):
         """
         Tries to map in a stable algorithm an email to color
 
         :param email_str:
         """
         color_bank = self.get_color_bank()
         # pick position (module it's length so we always find it in the
         # bank even if it's smaller than 256 values
         pos = self.pick_color_bank_index(email_str, color_bank)
         return color_bank[pos]
 
     def normalize_email(self, email_address):
         import unicodedata
         # default host used to fill in the fake/missing email
         default_host = u'localhost'
 
         if not email_address:
             email_address = u'%s@%s' % (User.DEFAULT_USER, default_host)
 
         email_address = safe_unicode(email_address)
 
         if u'@' not in email_address:
             email_address = u'%s@%s' % (email_address, default_host)
 
         if email_address.endswith(u'@'):
             email_address = u'%s%s' % (email_address, default_host)
 
         email_address = unicodedata.normalize('NFKD', email_address)\
             .encode('ascii', 'ignore')
         return email_address
 
     def get_initials(self):
         """
         Returns 2 letter initials calculated based on the input.
         The algorithm picks first given email address, and takes first letter
         of part before @, and then the first letter of server name. In case
         the part before @ is in a format of `somestring.somestring2` it replaces
         the server letter with first letter of somestring2
 
         In case function was initialized with both first and lastname, this
         overrides the extraction from email by first letter of the first and
         last name. We add special logic to that functionality, In case Full name
         is compound, like Guido Von Rossum, we use last part of the last name
         (Von Rossum) picking `R`.
 
         Function also normalizes the non-ascii characters to they ascii
         representation, eg Ą => A
         """
         import unicodedata
         # replace non-ascii to ascii
         first_name = unicodedata.normalize(
             'NFKD', safe_unicode(self.first_name)).encode('ascii', 'ignore')
         last_name = unicodedata.normalize(
             'NFKD', safe_unicode(self.last_name)).encode('ascii', 'ignore')
 
         # do NFKD encoding, and also make sure email has proper format
         email_address = self.normalize_email(self.email_address)
 
         # first push the email initials
         prefix, server = email_address.split('@', 1)
 
         # check if prefix is maybe a 'first_name.last_name' syntax
         _dot_split = prefix.rsplit('.', 1)
         if len(_dot_split) == 2:
             initials = [_dot_split[0][0], _dot_split[1][0]]
         else:
             initials = [prefix[0], server[0]]
 
         # then try to replace either first_name or last_name
         fn_letter = (first_name or " ")[0].strip()
         ln_letter = (last_name.split(' ', 1)[-1] or " ")[0].strip()
 
         if fn_letter:
             initials[0] = fn_letter
 
         if ln_letter:
             initials[1] = ln_letter
 
         return ''.join(initials).upper()
 
     def get_img_data_by_type(self, font_family, img_type):
         default_user = """
         <svg xmlns="http://www.w3.org/2000/svg"
         version="1.1" x="0px" y="0px" width="{size}" height="{size}"
         viewBox="-15 -10 439.165 429.164"
 
         xml:space="preserve"
         style="background:{background};" >
 
         <path d="M204.583,216.671c50.664,0,91.74-48.075,
                  91.74-107.378c0-82.237-41.074-107.377-91.74-107.377
                  c-50.668,0-91.74,25.14-91.74,107.377C112.844,
                  168.596,153.916,216.671,
                  204.583,216.671z" fill="{text_color}"/>
         <path d="M407.164,374.717L360.88,
                  270.454c-2.117-4.771-5.836-8.728-10.465-11.138l-71.83-37.392
                  c-1.584-0.823-3.502-0.663-4.926,0.415c-20.316,
                  15.366-44.203,23.488-69.076,23.488c-24.877,
                  0-48.762-8.122-69.078-23.488
                  c-1.428-1.078-3.346-1.238-4.93-0.415L58.75,
                  259.316c-4.631,2.41-8.346,6.365-10.465,11.138L2.001,374.717
                  c-3.191,7.188-2.537,15.412,1.75,22.005c4.285,
                  6.592,11.537,10.526,19.4,10.526h362.861c7.863,0,15.117-3.936,
                  19.402-10.527 C409.699,390.129,
                  410.355,381.902,407.164,374.717z" fill="{text_color}"/>
         </svg>""".format(
             size=self.size,
             background='#979797',  # @grey4
             text_color=self.text_color,
             font_family=font_family)
 
         return {
             "default_user": default_user
         }[img_type]
 
     def get_img_data(self, svg_type=None):
         """
         generates the svg metadata for image
         """
 
         font_family = ','.join([
             'proximanovaregular',
             'Proxima Nova Regular',
             'Proxima Nova',
             'Arial',
             'Lucida Grande',
             'sans-serif'
             ])
         if svg_type:
             return self.get_img_data_by_type(font_family, svg_type)
 
         initials = self.get_initials()
         img_data = """
         <svg xmlns="http://www.w3.org/2000/svg" pointer-events="none"
              width="{size}" height="{size}"
              style="width: 100%; height: 100%; background-color: {background}"
              viewBox="0 0 {size} {size}">
             <text text-anchor="middle" y="50%" x="50%" dy="0.35em"
                   pointer-events="auto" fill="{text_color}"
                   font-family="{font_family}"
                   style="font-weight: 400; font-size: {f_size}px;">{text}
             </text>
         </svg>""".format(
             size=self.size,
             f_size=self.size/1.85,  # scale the text inside the box nicely
             background=self.background,
             text_color=self.text_color,
             text=initials.upper(),
             font_family=font_family)
 
         return img_data
 
     def generate_svg(self, svg_type=None):
         img_data = self.get_img_data(svg_type)
         return "data:image/svg+xml;base64,%s" % img_data.encode('base64')
 
 
 def initials_gravatar(email_address, first_name, last_name, size=30):
     svg_type = None
     if email_address == User.DEFAULT_USER_EMAIL:
         svg_type = 'default_user'
     klass = InitialsGravatar(email_address, first_name, last_name, size)
     return klass.generate_svg(svg_type=svg_type)
 
 
 def gravatar_url(email_address, size=30, request=None):
     request = get_current_request()
     if request and hasattr(request, 'call_context'):
         _use_gravatar = request.call_context.visual.use_gravatar
         _gravatar_url = request.call_context.visual.gravatar_url
     else:
         # doh, we need to re-import those to mock it later
         from pylons import tmpl_context as c
 
         _use_gravatar = c.visual.use_gravatar
         _gravatar_url = c.visual.gravatar_url
 
     _gravatar_url = _gravatar_url or User.DEFAULT_GRAVATAR_URL
 
     email_address = email_address or User.DEFAULT_USER_EMAIL
     if isinstance(email_address, unicode):
         # hashlib crashes on unicode items
         email_address = safe_str(email_address)
 
     # empty email or default user
     if not email_address or email_address == User.DEFAULT_USER_EMAIL:
         return initials_gravatar(User.DEFAULT_USER_EMAIL, '', '', size=size)
 
     if _use_gravatar:
         # TODO: Disuse pyramid thread locals. Think about another solution to
         # get the host and schema here.
         request = get_current_request()
         tmpl = safe_str(_gravatar_url)
         tmpl = tmpl.replace('{email}', email_address)\
                    .replace('{md5email}', md5_safe(email_address.lower())) \
                    .replace('{netloc}', request.host)\
                    .replace('{scheme}', request.scheme)\
                    .replace('{size}', safe_str(size))
         return tmpl
     else:
         return initials_gravatar(email_address, '', '', size=size)
 
 
 class Page(_Page):
     """
     Custom pager to match rendering style with paginator
     """
 
     def _get_pos(self, cur_page, max_page, items):
         edge = (items / 2) + 1
         if (cur_page <= edge):
             radius = max(items / 2, items - cur_page)
         elif (max_page - cur_page) < edge:
             radius = (items - 1) - (max_page - cur_page)
         else:
             radius = items / 2
 
         left = max(1, (cur_page - (radius)))
         right = min(max_page, cur_page + (radius))
         return left, cur_page, right
 
     def _range(self, regexp_match):
         """
         Return range of linked pages (e.g. '1 2 [3] 4 5 6 7 8').
 
         Arguments:
 
         regexp_match
             A "re" (regular expressions) match object containing the
             radius of linked pages around the current page in
             regexp_match.group(1) as a string
 
         This function is supposed to be called as a callable in
         re.sub.
 
         """
         radius = int(regexp_match.group(1))
 
         # Compute the first and last page number within the radius
         # e.g. '1 .. 5 6 [7] 8 9 .. 12'
         # -> leftmost_page  = 5
         # -> rightmost_page = 9
         leftmost_page, _cur, rightmost_page = self._get_pos(self.page,
                                                             self.last_page,
                                                             (radius * 2) + 1)
         nav_items = []
 
         # Create a link to the first page (unless we are on the first page
         # or there would be no need to insert '..' spacers)
         if self.page != self.first_page and self.first_page < leftmost_page:
             nav_items.append(self._pagerlink(self.first_page, self.first_page))
 
         # Insert dots if there are pages between the first page
         # and the currently displayed page range
         if leftmost_page - self.first_page > 1:
             # Wrap in a SPAN tag if nolink_attr is set
             text = '..'
             if self.dotdot_attr:
                 text = HTML.span(c=text, **self.dotdot_attr)
             nav_items.append(text)
 
         for thispage in xrange(leftmost_page, rightmost_page + 1):
             # Hilight the current page number and do not use a link
             if thispage == self.page:
                 text = '%s' % (thispage,)
                 # Wrap in a SPAN tag if nolink_attr is set
                 if self.curpage_attr:
                     text = HTML.span(c=text, **self.curpage_attr)
                 nav_items.append(text)
             # Otherwise create just a link to that page
             else:
                 text = '%s' % (thispage,)
                 nav_items.append(self._pagerlink(thispage, text))
 
         # Insert dots if there are pages between the displayed
         # page numbers and the end of the page range
         if self.last_page - rightmost_page > 1:
             text = '..'
             # Wrap in a SPAN tag if nolink_attr is set
             if self.dotdot_attr:
                 text = HTML.span(c=text, **self.dotdot_attr)
             nav_items.append(text)
 
         # Create a link to the very last page (unless we are on the last
         # page or there would be no need to insert '..' spacers)
         if self.page != self.last_page and rightmost_page < self.last_page:
             nav_items.append(self._pagerlink(self.last_page, self.last_page))
 
         ## prerender links
         #_page_link = url.current()
         #nav_items.append(literal('<link rel="prerender" href="%s?page=%s">' % (_page_link, str(int(self.page)+1))))
         #nav_items.append(literal('<link rel="prefetch" href="%s?page=%s">' % (_page_link, str(int(self.page)+1))))
         return self.separator.join(nav_items)
 
     def pager(self, format='~2~', page_param='page', partial_param='partial',
         show_if_single_page=False, separator=' ', onclick=None,
         symbol_first='<<', symbol_last='>>',
         symbol_previous='<', symbol_next='>',
         link_attr={'class': 'pager_link', 'rel': 'prerender'},
         curpage_attr={'class': 'pager_curpage'},
         dotdot_attr={'class': 'pager_dotdot'}, **kwargs):
 
         self.curpage_attr = curpage_attr
         self.separator = separator
         self.pager_kwargs = kwargs
         self.page_param = page_param
         self.partial_param = partial_param
         self.onclick = onclick
         self.link_attr = link_attr
         self.dotdot_attr = dotdot_attr
 
         # Don't show navigator if there is no more than one page
         if self.page_count == 0 or (self.page_count == 1 and not show_if_single_page):
             return ''
 
         from string import Template
         # Replace ~...~ in token format by range of pages
         result = re.sub(r'~(\d+)~', self._range, format)
 
         # Interpolate '%' variables
         result = Template(result).safe_substitute({
             'first_page': self.first_page,
             'last_page': self.last_page,
             'page': self.page,
             'page_count': self.page_count,
             'items_per_page': self.items_per_page,
             'first_item': self.first_item,
             'last_item': self.last_item,
             'item_count': self.item_count,
             'link_first': self.page > self.first_page and \
                     self._pagerlink(self.first_page, symbol_first) or '',
             'link_last': self.page < self.last_page and \
                     self._pagerlink(self.last_page, symbol_last) or '',
             'link_previous': self.previous_page and \
                     self._pagerlink(self.previous_page, symbol_previous) \
                     or HTML.span(symbol_previous, class_="pg-previous disabled"),
             'link_next': self.next_page and \
                     self._pagerlink(self.next_page, symbol_next) \
                     or HTML.span(symbol_next, class_="pg-next disabled")
         })
 
         return literal(result)
 
 
 #==============================================================================
 # REPO PAGER, PAGER FOR REPOSITORY
 #==============================================================================
 class RepoPage(Page):
 
     def __init__(self, collection, page=1, items_per_page=20,
                  item_count=None, url=None, **kwargs):
 
         """Create a "RepoPage" instance. special pager for paging
         repository
         """
         self._url_generator = url
 
         # Safe the kwargs class-wide so they can be used in the pager() method
         self.kwargs = kwargs
 
         # Save a reference to the collection
         self.original_collection = collection
 
         self.collection = collection
 
         # The self.page is the number of the current page.
         # The first page has the number 1!
         try:
             self.page = int(page)  # make it int() if we get it as a string
         except (ValueError, TypeError):
             self.page = 1
 
         self.items_per_page = items_per_page
 
         # Unless the user tells us how many items the collections has
         # we calculate that ourselves.
         if item_count is not None:
             self.item_count = item_count
         else:
             self.item_count = len(self.collection)
 
         # Compute the number of the first and last available page
         if self.item_count > 0:
             self.first_page = 1
             self.page_count = int(math.ceil(float(self.item_count) /
                                             self.items_per_page))
             self.last_page = self.first_page + self.page_count - 1
 
             # Make sure that the requested page number is the range of
             # valid pages
             if self.page > self.last_page:
                 self.page = self.last_page
             elif self.page < self.first_page:
                 self.page = self.first_page
 
             # Note: the number of items on this page can be less than
             #       items_per_page if the last page is not full
             self.first_item = max(0, (self.item_count) - (self.page *
                                                           items_per_page))
             self.last_item = ((self.item_count - 1) - items_per_page *
                               (self.page - 1))
 
             self.items = list(self.collection[self.first_item:self.last_item + 1])
 
             # Links to previous and next page
             if self.page > self.first_page:
                 self.previous_page = self.page - 1
             else:
                 self.previous_page = None
 
             if self.page < self.last_page:
                 self.next_page = self.page + 1
             else:
                 self.next_page = None
 
         # No items available
         else:
             self.first_page = None
             self.page_count = 0
             self.last_page = None
             self.first_item = None
             self.last_item = None
             self.previous_page = None
             self.next_page = None
             self.items = []
 
         # This is a subclass of the 'list' type. Initialise the list now.
         list.__init__(self, reversed(self.items))
 
 
 def breadcrumb_repo_link(repo):
     """
     Makes a breadcrumbs path link to repo
 
     ex::
         group >> subgroup >> repo
 
     :param repo: a Repository instance
     """
 
     path = [
         link_to(group.name, route_path('repo_group_home', repo_group_name=group.group_name))
         for group in repo.groups_with_parents
     ] + [
         link_to(repo.just_name, route_path('repo_summary', repo_name=repo.repo_name))
     ]
 
     return literal(' &raquo; '.join(path))
 
 
 def format_byte_size_binary(file_size):
     """
     Formats file/folder sizes to standard.
     """
     if file_size is None:
         file_size = 0
 
     formatted_size = format_byte_size(file_size, binary=True)
     return formatted_size
 
 
 def urlify_text(text_, safe=True):
     """
     Extrac urls from text and make html links out of them
 
     :param text_:
     """
 
     url_pat = re.compile(r'''(http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@#.&+]'''
                          '''|[!*\(\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+)''')
 
     def url_func(match_obj):
         url_full = match_obj.groups()[0]
         return '<a href="%(url)s">%(url)s</a>' % ({'url': url_full})
     _newtext = url_pat.sub(url_func, text_)
     if safe:
         return literal(_newtext)
     return _newtext
 
 
 def urlify_commits(text_, repository):
     """
     Extract commit ids from text and make link from them
 
     :param text_:
     :param repository: repo name to build the URL with
     """
 
     URL_PAT = re.compile(r'(^|\s)([0-9a-fA-F]{12,40})($|\s)')
 
     def url_func(match_obj):
         commit_id = match_obj.groups()[1]
         pref = match_obj.groups()[0]
         suf = match_obj.groups()[2]
 
         tmpl = (
             '%(pref)s<a class="%(cls)s" href="%(url)s">'
             '%(commit_id)s</a>%(suf)s'
         )
         return tmpl % {
             'pref': pref,
             'cls': 'revision-link',
             'url': route_url('repo_commit', repo_name=repository,
                              commit_id=commit_id),
             'commit_id': commit_id,
             'suf': suf
         }
 
     newtext = URL_PAT.sub(url_func, text_)
 
     return newtext
 
 
 def _process_url_func(match_obj, repo_name, uid, entry,
                       return_raw_data=False, link_format='html'):
     pref = ''
     if match_obj.group().startswith(' '):
         pref = ' '
 
     issue_id = ''.join(match_obj.groups())
 
     if link_format == 'html':
         tmpl = (
             '%(pref)s<a class="%(cls)s" href="%(url)s">'
             '%(issue-prefix)s%(id-repr)s'
             '</a>')
     elif link_format == 'rst':
         tmpl = '`%(issue-prefix)s%(id-repr)s <%(url)s>`_'
     elif link_format == 'markdown':
         tmpl = '[%(issue-prefix)s%(id-repr)s](%(url)s)'
     else:
         raise ValueError('Bad link_format:{}'.format(link_format))
 
     (repo_name_cleaned,
      parent_group_name) = RepoGroupModel().\
         _get_group_name_and_parent(repo_name)
 
     # variables replacement
     named_vars = {
         'id': issue_id,
         'repo': repo_name,
         'repo_name': repo_name_cleaned,
         'group_name': parent_group_name
     }
     # named regex variables
     named_vars.update(match_obj.groupdict())
     _url = string.Template(entry['url']).safe_substitute(**named_vars)
 
     data = {
         'pref': pref,
         'cls': 'issue-tracker-link',
         'url': _url,
         'id-repr': issue_id,
         'issue-prefix': entry['pref'],
         'serv': entry['url'],
     }
     if return_raw_data:
         return {
             'id': issue_id,
             'url': _url
         }
     return tmpl % data
 
 
 def process_patterns(text_string, repo_name, link_format='html'):
     allowed_formats = ['html', 'rst', 'markdown']
     if link_format not in allowed_formats:
         raise ValueError('Link format can be only one of:{} got {}'.format(
                          allowed_formats, link_format))
 
     repo = None
     if repo_name:
         # Retrieving repo_name to avoid invalid repo_name to explode on
         # IssueTrackerSettingsModel but still passing invalid name further down
         repo = Repository.get_by_repo_name(repo_name, cache=True)
 
     settings_model = IssueTrackerSettingsModel(repo=repo)
     active_entries = settings_model.get_settings(cache=True)
 
     issues_data = []
     newtext = text_string
 
     for uid, entry in active_entries.items():
         log.debug('found issue tracker entry with uid %s' % (uid,))
 
         if not (entry['pat'] and entry['url']):
             log.debug('skipping due to missing data')
             continue
 
         log.debug('issue tracker entry: uid: `%s` PAT:%s URL:%s PREFIX:%s'
                   % (uid, entry['pat'], entry['url'], entry['pref']))
 
         try:
             pattern = re.compile(r'%s' % entry['pat'])
         except re.error:
             log.exception(
                 'issue tracker pattern: `%s` failed to compile',
                 entry['pat'])
             continue
 
         data_func = partial(
             _process_url_func, repo_name=repo_name, entry=entry, uid=uid,
             return_raw_data=True)
 
         for match_obj in pattern.finditer(text_string):
             issues_data.append(data_func(match_obj))
 
         url_func = partial(
             _process_url_func, repo_name=repo_name, entry=entry, uid=uid,
             link_format=link_format)
 
         newtext = pattern.sub(url_func, newtext)
         log.debug('processed prefix:uid `%s`' % (uid,))
 
     return newtext, issues_data
 
 
 def urlify_commit_message(commit_text, repository=None):
     """
     Parses given text message and makes proper links.
     issues are linked to given issue-server, and rest is a commit link
 
     :param commit_text:
     :param repository:
     """
     from pylons import url  # doh, we need to re-import url to mock it later
 
     def escaper(string):
         return string.replace('<', '&lt;').replace('>', '&gt;')
 
     newtext = escaper(commit_text)
 
     # extract http/https links and make them real urls
     newtext = urlify_text(newtext, safe=False)
 
     # urlify commits - extract commit ids and make link out of them, if we have
     # the scope of repository present.
     if repository:
         newtext = urlify_commits(newtext, repository)
 
     # process issue tracker patterns
     newtext, issues = process_patterns(newtext, repository or '')
 
     return literal(newtext)
 
 
 def render_binary(repo_name, file_obj):
     """
     Choose how to render a binary file
     """
     filename = file_obj.name
 
     # images
     for ext in ['*.png', '*.jpg', '*.ico', '*.gif']:
         if fnmatch.fnmatch(filename, pat=ext):
             alt = filename
             src = route_path(
                 'repo_file_raw', repo_name=repo_name,
                 commit_id=file_obj.commit.raw_id, f_path=file_obj.path)
             return literal('<img class="rendered-binary" alt="{}" src="{}">'.format(alt, src))
 
 
 def renderer_from_filename(filename, exclude=None):
     """
     choose a renderer based on filename, this works only for text based files
     """
 
     # ipython
     for ext in ['*.ipynb']:
         if fnmatch.fnmatch(filename, pat=ext):
             return 'jupyter'
 
     is_markup = MarkupRenderer.renderer_from_filename(filename, exclude=exclude)
     if is_markup:
         return is_markup
     return None
 
 
-def render(source, renderer='rst', mentions=False, relative_url=None,
+def render(source, renderer='rst', mentions=False, relative_urls=None,
            repo_name=None):
 
     def maybe_convert_relative_links(html_source):
-        if relative_url:
-            return relative_links(html_source, relative_url)
+        if relative_urls:
+            return relative_links(html_source, relative_urls)
         return html_source
 
     if renderer == 'rst':
         if repo_name:
             # process patterns on comments if we pass in repo name
             source, issues = process_patterns(
                 source, repo_name, link_format='rst')
 
         return literal(
             '<div class="rst-block">%s</div>' %
             maybe_convert_relative_links(
                 MarkupRenderer.rst(source, mentions=mentions)))
     elif renderer == 'markdown':
         if repo_name:
             # process patterns on comments if we pass in repo name
             source, issues = process_patterns(
                 source, repo_name, link_format='markdown')
 
         return literal(
             '<div class="markdown-block">%s</div>' %
             maybe_convert_relative_links(
                 MarkupRenderer.markdown(source, flavored=True,
                                         mentions=mentions)))
     elif renderer == 'jupyter':
         return literal(
             '<div class="ipynb">%s</div>' %
             maybe_convert_relative_links(
                 MarkupRenderer.jupyter(source)))
 
     # None means just show the file-source
     return None
 
 
 def commit_status(repo, commit_id):
     return ChangesetStatusModel().get_status(repo, commit_id)
 
 
 def commit_status_lbl(commit_status):
     return dict(ChangesetStatus.STATUSES).get(commit_status)
 
 
 def commit_time(repo_name, commit_id):
     repo = Repository.get_by_repo_name(repo_name)
     commit = repo.get_commit(commit_id=commit_id)
     return commit.date
 
 
 def get_permission_name(key):
     return dict(Permission.PERMS).get(key)
 
 
 def journal_filter_help(request):
     _ = request.translate
 
     return _(
         'Example filter terms:\n' +
         '     repository:vcs\n' +
         '     username:marcin\n' +
         '     username:(NOT marcin)\n' +
         '     action:*push*\n' +
         '     ip:127.0.0.1\n' +
         '     date:20120101\n' +
         '     date:[20120101100000 TO 20120102]\n' +
         '\n' +
         'Generate wildcards using \'*\' character:\n' +
         '     "repository:vcs*" - search everything starting with \'vcs\'\n' +
         '     "repository:*vcs*" - search for repository containing \'vcs\'\n' +
         '\n' +
         'Optional AND / OR operators in queries\n' +
         '     "repository:vcs OR repository:test"\n' +
         '     "username:test AND repository:test*"\n'
     )
 
 
 def search_filter_help(searcher, request):
     _ = request.translate
 
     terms = ''
     return _(
         'Example filter terms for `{searcher}` search:\n' +
         '{terms}\n' +
         'Generate wildcards using \'*\' character:\n' +
         '     "repo_name:vcs*" - search everything starting with \'vcs\'\n' +
         '     "repo_name:*vcs*" - search for repository containing \'vcs\'\n' +
         '\n' +
         'Optional AND / OR operators in queries\n' +
         '     "repo_name:vcs OR repo_name:test"\n' +
         '     "owner:test AND repo_name:test*"\n' +
         'More: {search_doc}'
     ).format(searcher=searcher.name,
              terms=terms, search_doc=searcher.query_lang_doc)
 
 
 def not_mapped_error(repo_name):
     from rhodecode.translation import _
     flash(_('%s repository is not mapped to db perhaps'
             ' it was created or renamed from the filesystem'
             ' please run the application again'
             ' in order to rescan repositories') % repo_name, category='error')
 
 
 def ip_range(ip_addr):
     from rhodecode.model.db import UserIpMap
     s, e = UserIpMap._get_ip_range(ip_addr)
     return '%s - %s' % (s, e)
 
 
 def form(url, method='post', needs_csrf_token=True, **attrs):
     """Wrapper around webhelpers.tags.form to prevent CSRF attacks."""
     if method.lower() != 'get' and needs_csrf_token:
         raise Exception(
             'Forms to POST/PUT/DELETE endpoints should have (in general) a ' +
             'CSRF token. If the endpoint does not require such token you can ' +
             'explicitly set the parameter needs_csrf_token to false.')
 
     return wh_form(url, method=method, **attrs)
 
 
 def secure_form(url, method="POST", multipart=False, **attrs):
     """Start a form tag that points the action to an url. This
     form tag will also include the hidden field containing
     the auth token.
 
     The url options should be given either as a string, or as a
     ``url()`` function. The method for the form defaults to POST.
 
     Options:
 
     ``multipart``
         If set to True, the enctype is set to "multipart/form-data".
     ``method``
         The method to use when submitting the form, usually either
         "GET" or "POST". If "PUT", "DELETE", or another verb is used, a
         hidden input with name _method is added to simulate the verb
         over POST.
 
     """
     from webhelpers.pylonslib.secure_form import insecure_form
     form = insecure_form(url, method, multipart, **attrs)
 
     session = None
     # TODO(marcink): after pyramid migration require request variable ALWAYS
     if 'request' in attrs:
         session = attrs['request'].session
 
     token = literal(
         '<input type="hidden" id="{}" name="{}" value="{}">'.format(
         csrf_token_key, csrf_token_key, get_csrf_token(session)))
 
     return literal("%s\n%s" % (form, token))
 
 
 def dropdownmenu(name, selected, options, enable_filter=False, **attrs):
     select_html = select(name, selected, options, **attrs)
     select2 = """
         <script>
             $(document).ready(function() {
                   $('#%s').select2({
                       containerCssClass: 'drop-menu',
                       dropdownCssClass: 'drop-menu-dropdown',
                       dropdownAutoWidth: true%s
                   });
             });
         </script>
     """
     filter_option = """,
                         minimumResultsForSearch: -1
     """
     input_id = attrs.get('id') or name
     filter_enabled = "" if enable_filter else filter_option
     select_script = literal(select2 % (input_id, filter_enabled))
 
     return literal(select_html+select_script)
 
 
 def get_visual_attr(tmpl_context_var, attr_name):
     """
     A safe way to get a variable from visual variable of template context
 
     :param tmpl_context_var: instance of tmpl_context, usually present as `c`
     :param attr_name: name of the attribute we fetch from the c.visual
     """
     visual = getattr(tmpl_context_var, 'visual', None)
     if not visual:
         return
     else:
         return getattr(visual, attr_name, None)
 
 
 def get_last_path_part(file_node):
     if not file_node.path:
         return u''
 
     path = safe_unicode(file_node.path.split('/')[-1])
     return u'../' + path
 
 
 def route_url(*args, **kwargs):
     """
     Wrapper around pyramids `route_url` (fully qualified url) function. 
     It is used to generate URLs from within pylons views or templates. 
     This will be removed when pyramid migration if finished.
     """
     req = get_current_request()
     return req.route_url(*args, **kwargs)
 
 
 def route_path(*args, **kwargs):
     """
     Wrapper around pyramids `route_path` function. It is used to generate
     URLs from within pylons views or templates. This will be removed when
     pyramid migration if finished.
     """
     req = get_current_request()
     return req.route_path(*args, **kwargs)
 
 
 def route_path_or_none(*args, **kwargs):
     try:
         return route_path(*args, **kwargs)
     except KeyError:
         return None
 
 
 def static_url(*args, **kwds):
     """
     Wrapper around pyramids `route_path` function. It is used to generate
     URLs from within pylons views or templates. This will be removed when
     pyramid migration if finished.
     """
     req = get_current_request()
     return req.static_url(*args, **kwds)
 
 
 def resource_path(*args, **kwds):
     """
     Wrapper around pyramids `route_path` function. It is used to generate
     URLs from within pylons views or templates. This will be removed when
     pyramid migration if finished.
     """
     req = get_current_request()
     return req.resource_path(*args, **kwds)
 
 
 def api_call_example(method, args):
     """
     Generates an API call example via CURL
     """
     args_json = json.dumps(OrderedDict([
         ('id', 1),
         ('auth_token', 'SECRET'),
         ('method', method),
         ('args', args)
     ]))
     return literal(
         "curl {api_url} -X POST -H 'content-type:text/plain' --data-binary '{data}'"
         "<br/><br/>SECRET can be found in <a href=\"{token_url}\">auth-tokens</a> page, "
         "and needs to be of `api calls` role."
         .format(
             api_url=route_url('apiv2'),
             token_url=route_url('my_account_auth_tokens'),
             data=args_json))
 
 
 def notification_description(notification, request):
     """
     Generate notification human readable description based on notification type
     """
     from rhodecode.model.notification import NotificationModel
     return NotificationModel().make_description(
         notification, translate=request.translate)

rhodecode/lib/markup_renderer.py

@@ -1,495 +1,499 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2011-2017 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/
 
 
 """
 Renderer for markup languages with ability to parse using rst or markdown
 """
 
 import re
 import os
 import lxml
 import logging
 import urlparse
 
 from mako.lookup import TemplateLookup
 from mako.template import Template as MakoTemplate
 
 from docutils.core import publish_parts
 from docutils.parsers.rst import directives
 from docutils import writers
 from docutils.writers import html4css1
 import markdown
 
 from rhodecode.lib.markdown_ext import GithubFlavoredMarkdownExtension
 from rhodecode.lib.utils2 import (
     safe_str, safe_unicode, md5_safe, MENTIONS_REGEX)
 
 log = logging.getLogger(__name__)
 
 # default renderer used to generate automated comments
 DEFAULT_COMMENTS_RENDERER = 'rst'
 
 
 class CustomHTMLTranslator(writers.html4css1.HTMLTranslator):
     """
     Custom HTML Translator used for sandboxing potential
     JS injections in ref links
     """
 
     def visit_reference(self, node):
         if 'refuri' in node.attributes:
             refuri = node['refuri']
             if ':' in refuri:
                 prefix, link = refuri.lstrip().split(':', 1)
                 if prefix == 'javascript':
                     # we don't allow javascript type of refs...
                     node['refuri'] = 'javascript:alert("SandBoxedJavascript")'
 
         # old style class requires this...
         return html4css1.HTMLTranslator.visit_reference(self, node)
 
 
 class RhodeCodeWriter(writers.html4css1.Writer):
     def __init__(self):
         writers.Writer.__init__(self)
         self.translator_class = CustomHTMLTranslator
 
 
-def relative_links(html_source, server_path):
+def relative_links(html_source, server_paths):
     if not html_source:
         return html_source
 
     try:
         from lxml.html import fromstring
         from lxml.html import tostring
     except ImportError:
         log.exception('Failed to import lxml')
         return html_source
 
     try:
         doc = lxml.html.fromstring(html_source)
     except Exception:
         return html_source
 
     for el in doc.cssselect('img, video'):
         src = el.attrib.get('src')
         if src:
-            el.attrib['src'] = relative_path(src, server_path)
+            el.attrib['src'] = relative_path(src, server_paths['raw'])
 
     for el in doc.cssselect('a:not(.gfm)'):
         src = el.attrib.get('href')
         if src:
-            el.attrib['href'] = relative_path(src, server_path)
+            raw_mode = el.attrib['href'].endswith('?raw=1')
+            if raw_mode:
+                el.attrib['href'] = relative_path(src, server_paths['raw'])
+            else:
+                el.attrib['href'] = relative_path(src, server_paths['standard'])
 
     return lxml.html.tostring(doc)
 
 
 def relative_path(path, request_path, is_repo_file=None):
     """
     relative link support, path is a rel path, and request_path is current
     server path (not absolute)
 
     e.g.
 
     path = '../logo.png'
     request_path= '/repo/files/path/file.md'
     produces: '/repo/files/logo.png'
     """
     # TODO(marcink): unicode/str support ?
     # maybe=> safe_unicode(urllib.quote(safe_str(final_path), '/:'))
 
     def dummy_check(p):
         return True  # assume default is a valid file path
 
     is_repo_file = is_repo_file or dummy_check
     if not path:
         return request_path
 
     path = safe_unicode(path)
     request_path = safe_unicode(request_path)
 
     if path.startswith((u'data:', u'javascript:', u'#', u':')):
         # skip data, anchor, invalid links
         return path
 
     is_absolute = bool(urlparse.urlparse(path).netloc)
     if is_absolute:
         return path
 
     if not request_path:
         return path
 
     if path.startswith(u'/'):
         path = path[1:]
 
     if path.startswith(u'./'):
         path = path[2:]
 
     parts = request_path.split('/')
     # compute how deep we need to traverse the request_path
     depth = 0
 
     if is_repo_file(request_path):
         # if request path is a VALID file, we use a relative path with
         # one level up
         depth += 1
 
     while path.startswith(u'../'):
         depth += 1
         path = path[3:]
 
     if depth > 0:
         parts = parts[:-depth]
 
     parts.append(path)
     final_path = u'/'.join(parts).lstrip(u'/')
 
     return u'/' + final_path
 
 
 class MarkupRenderer(object):
     RESTRUCTUREDTEXT_DISALLOWED_DIRECTIVES = ['include', 'meta', 'raw']
 
     MARKDOWN_PAT = re.compile(r'\.(md|mkdn?|mdown|markdown)$', re.IGNORECASE)
     RST_PAT = re.compile(r'\.re?st$', re.IGNORECASE)
     JUPYTER_PAT = re.compile(r'\.(ipynb)$', re.IGNORECASE)
     PLAIN_PAT = re.compile(r'^readme$', re.IGNORECASE)
 
     extensions = ['codehilite', 'extra', 'def_list', 'sane_lists']
     markdown_renderer = markdown.Markdown(
         extensions, safe_mode=True, enable_attributes=False)
 
     markdown_renderer_flavored = markdown.Markdown(
         extensions + [GithubFlavoredMarkdownExtension()], safe_mode=True,
         enable_attributes=False)
 
     # extension together with weights. Lower is first means we control how
     # extensions are attached to readme names with those.
     PLAIN_EXTS = [
         # prefer no extension
         ('', 0),  # special case that renders READMES names without extension
         ('.text', 2), ('.TEXT', 2),
         ('.txt', 3), ('.TXT', 3)
     ]
 
     RST_EXTS = [
         ('.rst', 1), ('.rest', 1),
         ('.RST', 2), ('.REST', 2)
     ]
 
     MARKDOWN_EXTS = [
         ('.md', 1), ('.MD', 1),
         ('.mkdn', 2), ('.MKDN', 2),
         ('.mdown', 3), ('.MDOWN', 3),
         ('.markdown', 4), ('.MARKDOWN', 4)
     ]
 
     def _detect_renderer(self, source, filename=None):
         """
         runs detection of what renderer should be used for generating html
         from a markup language
 
         filename can be also explicitly a renderer name
 
         :param source:
         :param filename:
         """
 
         if MarkupRenderer.MARKDOWN_PAT.findall(filename):
             detected_renderer = 'markdown'
         elif MarkupRenderer.RST_PAT.findall(filename):
             detected_renderer = 'rst'
         elif MarkupRenderer.JUPYTER_PAT.findall(filename):
             detected_renderer = 'jupyter'
         elif MarkupRenderer.PLAIN_PAT.findall(filename):
             detected_renderer = 'plain'
         else:
             detected_renderer = 'plain'
 
         return getattr(MarkupRenderer, detected_renderer)
 
     @classmethod
     def renderer_from_filename(cls, filename, exclude):
         """
         Detect renderer markdown/rst from filename and optionally use exclude
         list to remove some options. This is mostly used in helpers.
         Returns None when no renderer can be detected.
         """
         def _filter(elements):
             if isinstance(exclude, (list, tuple)):
                 return [x for x in elements if x not in exclude]
             return elements
 
         if filename.endswith(
                 tuple(_filter([x[0] for x in cls.MARKDOWN_EXTS if x[0]]))):
             return 'markdown'
         if filename.endswith(tuple(_filter([x[0] for x in cls.RST_EXTS if x[0]]))):
             return 'rst'
 
         return None
 
     def render(self, source, filename=None):
         """
         Renders a given filename using detected renderer
         it detects renderers based on file extension or mimetype.
         At last it will just do a simple html replacing new lines with <br/>
 
         :param file_name:
         :param source:
         """
 
         renderer = self._detect_renderer(source, filename)
         readme_data = renderer(source)
         return readme_data
 
     @classmethod
     def _flavored_markdown(cls, text):
         """
         Github style flavored markdown
 
         :param text:
         """
 
         # Extract pre blocks.
         extractions = {}
 
         def pre_extraction_callback(matchobj):
             digest = md5_safe(matchobj.group(0))
             extractions[digest] = matchobj.group(0)
             return "{gfm-extraction-%s}" % digest
         pattern = re.compile(r'<pre>.*?</pre>', re.MULTILINE | re.DOTALL)
         text = re.sub(pattern, pre_extraction_callback, text)
 
         # Prevent foo_bar_baz from ending up with an italic word in the middle.
         def italic_callback(matchobj):
             s = matchobj.group(0)
             if list(s).count('_') >= 2:
                 return s.replace('_', r'\_')
             return s
         text = re.sub(r'^(?! {4}|\t)\w+_\w+_\w[\w_]*', italic_callback, text)
 
         # Insert pre block extractions.
         def pre_insert_callback(matchobj):
             return '\n\n' + extractions[matchobj.group(1)]
         text = re.sub(r'\{gfm-extraction-([0-9a-f]{32})\}',
                       pre_insert_callback, text)
 
         return text
 
     @classmethod
     def urlify_text(cls, text):
         url_pat = re.compile(r'(http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@.&+]'
                              r'|[!*\(\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+)')
 
         def url_func(match_obj):
             url_full = match_obj.groups()[0]
             return '<a href="%(url)s">%(url)s</a>' % ({'url': url_full})
 
         return url_pat.sub(url_func, text)
 
     @classmethod
     def plain(cls, source, universal_newline=True):
         source = safe_unicode(source)
         if universal_newline:
             newline = '\n'
             source = newline.join(source.splitlines())
 
         source = cls.urlify_text(source)
         return '<br />' + source.replace("\n", '<br />')
 
     @classmethod
     def markdown(cls, source, safe=True, flavored=True, mentions=False):
         # It does not allow to insert inline HTML. In presence of HTML tags, it
         # will replace them instead with [HTML_REMOVED]. This is controlled by
         # the safe_mode=True parameter of the markdown method.
 
         if flavored:
             markdown_renderer = cls.markdown_renderer_flavored
         else:
             markdown_renderer = cls.markdown_renderer
 
         if mentions:
             mention_pat = re.compile(MENTIONS_REGEX)
 
             def wrapp(match_obj):
                 uname = match_obj.groups()[0]
                 return ' **@%(uname)s** ' % {'uname': uname}
             mention_hl = mention_pat.sub(wrapp, source).strip()
             # we extracted mentions render with this using Mentions false
             return cls.markdown(mention_hl, safe=safe, flavored=flavored,
                                 mentions=False)
 
         source = safe_unicode(source)
         try:
             if flavored:
                 source = cls._flavored_markdown(source)
             return markdown_renderer.convert(source)
         except Exception:
             log.exception('Error when rendering Markdown')
             if safe:
                 log.debug('Fallback to render in plain mode')
                 return cls.plain(source)
             else:
                 raise
 
     @classmethod
     def rst(cls, source, safe=True, mentions=False):
         if mentions:
             mention_pat = re.compile(MENTIONS_REGEX)
 
             def wrapp(match_obj):
                 uname = match_obj.groups()[0]
                 return ' **@%(uname)s** ' % {'uname': uname}
             mention_hl = mention_pat.sub(wrapp, source).strip()
             # we extracted mentions render with this using Mentions false
             return cls.rst(mention_hl, safe=safe, mentions=False)
 
         source = safe_unicode(source)
         try:
             docutils_settings = dict(
                 [(alias, None) for alias in
                  cls.RESTRUCTUREDTEXT_DISALLOWED_DIRECTIVES])
 
             docutils_settings.update({'input_encoding': 'unicode',
                                       'report_level': 4})
 
             for k, v in docutils_settings.iteritems():
                 directives.register_directive(k, v)
 
             parts = publish_parts(source=source,
                                   writer=RhodeCodeWriter(),
                                   settings_overrides=docutils_settings)
 
             return parts['html_title'] + parts["fragment"]
         except Exception:
             log.exception('Error when rendering RST')
             if safe:
                 log.debug('Fallbacking to render in plain mode')
                 return cls.plain(source)
             else:
                 raise
 
     @classmethod
     def jupyter(cls, source, safe=True):
         from rhodecode.lib import helpers
 
         from traitlets.config import Config
         import nbformat
         from nbconvert import HTMLExporter
         from nbconvert.preprocessors import Preprocessor
 
         class CustomHTMLExporter(HTMLExporter):
             def _template_file_default(self):
                 return 'basic'
 
         class Sandbox(Preprocessor):
 
             def preprocess(self, nb, resources):
                 sandbox_text = 'SandBoxed(IPython.core.display.Javascript object)'
                 for cell in nb['cells']:
                     if safe and 'outputs' in cell:
                         for cell_output in cell['outputs']:
                             if 'data' in cell_output:
                                 if 'application/javascript' in cell_output['data']:
                                     cell_output['data']['text/plain'] = sandbox_text
                                     cell_output['data'].pop('application/javascript', None)
                 return nb, resources
 
         def _sanitize_resources(resources):
             """
             Skip/sanitize some of the CSS generated and included in jupyter
             so it doesn't messes up UI so much
             """
 
             # TODO(marcink): probably we should replace this with whole custom
             # CSS set that doesn't screw up, but jupyter generated html has some
             # special markers, so it requires Custom HTML exporter template with
             # _default_template_path_default, to achieve that
 
             # strip the reset CSS
             resources[0] = resources[0][resources[0].find('/*! Source'):]
             return resources
 
         def as_html(notebook):
             conf = Config()
             conf.CustomHTMLExporter.preprocessors = [Sandbox]
             html_exporter = CustomHTMLExporter(config=conf)
 
             (body, resources) = html_exporter.from_notebook_node(notebook)
             header = '<!-- ## IPYTHON NOTEBOOK RENDERING ## -->'
             js = MakoTemplate(r'''
             <!-- Load mathjax -->
                 <!-- MathJax configuration -->
                 <script type="text/x-mathjax-config">
                 MathJax.Hub.Config({
                     jax: ["input/TeX","output/HTML-CSS", "output/PreviewHTML"],
                     extensions: ["tex2jax.js","MathMenu.js","MathZoom.js", "fast-preview.js", "AssistiveMML.js", "[Contrib]/a11y/accessibility-menu.js"],
                     TeX: {
                         extensions: ["AMSmath.js","AMSsymbols.js","noErrors.js","noUndefined.js"]
                     },
                     tex2jax: {
                         inlineMath: [ ['$','$'], ["\\(","\\)"] ],
                         displayMath: [ ['$$','$$'], ["\\[","\\]"] ],
                         processEscapes: true,
                         processEnvironments: true
                     },
                     // Center justify equations in code and markdown cells. Elsewhere
                     // we use CSS to left justify single line equations in code cells.
                     displayAlign: 'center',
                     "HTML-CSS": {
                         styles: {'.MathJax_Display': {"margin": 0}},
                         linebreaks: { automatic: true },
                         availableFonts: ["STIX", "TeX"]
                     },
                     showMathMenu: false
                 });
                 </script>
                 <!-- End of mathjax configuration -->
                 <script src="${h.asset('js/src/math_jax/MathJax.js')}"></script>
             ''').render(h=helpers)
 
             css = '<style>{}</style>'.format(
                 ''.join(_sanitize_resources(resources['inlining']['css'])))
 
             body = '\n'.join([header, css, js, body])
             return body, resources
 
         notebook = nbformat.reads(source, as_version=4)
         (body, resources) = as_html(notebook)
         return body
 
 
 class RstTemplateRenderer(object):
 
     def __init__(self):
         base = os.path.abspath(os.path.dirname(os.path.dirname(__file__)))
         rst_template_dirs = [os.path.join(base, 'templates', 'rst_templates')]
         self.template_store = TemplateLookup(
             directories=rst_template_dirs,
             input_encoding='utf-8',
             imports=['from rhodecode.lib import helpers as h'])
 
     def _get_template(self, templatename):
         return self.template_store.get_template(templatename)
 
     def render(self, template_name, **kwargs):
         template = self._get_template(template_name)
         return template.render(**kwargs)

rhodecode/templates/files/files_source.mako

@@ -1,98 +1,104 b''
 <%namespace name="sourceblock" file="/codeblocks/source.mako"/>
 
 <div id="codeblock" class="codeblock">
     <div class="codeblock-header">
       <div class="stats">
             <span>
                 <strong>
                     <i class="icon-file-text"></i>
                     ${c.file}
                 </strong>
             </span>
             % if c.lf_node:
             <span title="${_('This file is a pointer to large binary file')}"> | ${_('LargeFile')} ${h.format_byte_size_binary(c.lf_node.size)} </span>
             % endif
             <span> | ${c.file.lines()[0]} ${_ungettext('line', 'lines', c.file.lines()[0])}</span>
             <span> | ${h.format_byte_size_binary(c.file.size)}</span>
             <span> | ${c.file.mimetype} </span>
             <span> | ${h.get_lexer_for_filenode(c.file).__class__.__name__}</span>
             <span class="item last"> | <i class="tooltip icon-clipboard clipboard-action" data-clipboard-text="${c.file.path}" title="${_('Copy the full path')}"></i></span>
       </div>
       <div class="buttons">
         <a id="file_history_overview" href="#">
             ${_('History')}
         </a>
         <a id="file_history_overview_full" style="display: none" href="${h.route_path('repo_changelog_file',repo_name=c.repo_name, commit_id=c.commit.raw_id, f_path=c.f_path)}">
            ${_('Show Full History')}
         </a> |
         %if c.annotate:
           ${h.link_to(_('Source'), h.route_path('repo_files', repo_name=c.repo_name,commit_id=c.commit.raw_id,f_path=c.f_path))}
         %else:
           ${h.link_to(_('Annotation'), h.route_path('repo_files:annotated',repo_name=c.repo_name,commit_id=c.commit.raw_id,f_path=c.f_path))}
         %endif
          | ${h.link_to(_('Raw'), h.route_path('repo_file_raw',repo_name=c.repo_name,commit_id=c.commit.raw_id,f_path=c.f_path))}
          |
           % if c.lf_node:
               <a href="${h.route_path('repo_file_download',repo_name=c.repo_name,commit_id=c.commit.raw_id,f_path=c.f_path, _query=dict(lf=1))}">
               ${_('Download largefile')}
               </a>
           % else:
               <a href="${h.route_path('repo_file_download',repo_name=c.repo_name,commit_id=c.commit.raw_id,f_path=c.f_path)}">
               ${_('Download')}
               </a>
           % endif
 
         %if h.HasRepoPermissionAny('repository.write','repository.admin')(c.repo_name):
            |
          %if c.on_branch_head and c.branch_or_raw_id and not c.file.is_binary:
             <a href="${h.route_path('repo_files_edit_file',repo_name=c.repo_name,commit_id=c.branch_or_raw_id,f_path=c.f_path, _anchor='edit')}">
               ${_('Edit on Branch:{}').format(c.branch_name)}
             </a>
             | <a class="btn-danger btn-link" href="${h.route_path('repo_files_remove_file',repo_name=c.repo_name,commit_id=c.branch_or_raw_id,f_path=c.f_path, _anchor='edit')}">${_('Delete')}
             </a>
          %elif c.on_branch_head and c.branch_or_raw_id and c.file.is_binary:
           ${h.link_to(_('Edit'), '#', class_="btn btn-link disabled tooltip", title=_('Editing binary files not allowed'))}
            | ${h.link_to(_('Delete'), h.route_path('repo_files_remove_file',repo_name=c.repo_name,commit_id=c.branch_or_raw_id,f_path=c.f_path, _anchor='edit'),class_="btn-danger btn-link")}
          %else:
           ${h.link_to(_('Edit'), '#', class_="btn btn-link disabled tooltip", title=_('Editing files allowed only when on branch head commit'))}
            | ${h.link_to(_('Delete'), '#', class_="btn btn-danger btn-link disabled tooltip", title=_('Deleting files allowed only when on branch head commit'))}
          %endif
         %endif
       </div>
     </div>
     <div id="file_history_container"></div>
     <div class="code-body">
      %if c.file.is_binary:
            <% rendered_binary = h.render_binary(c.repo_name, c.file)%>
            % if rendered_binary:
                ${rendered_binary}
            % else:
                <div>
                 ${_('Binary file (%s)') % c.file.mimetype}
                </div>
            % endif
      %else:
         % if c.file.size < c.visual.cut_off_limit_file:
             %if c.renderer and not c.annotate:
-                ${h.render(c.file.content, renderer=c.renderer, relative_url=h.route_path('repo_file_raw',repo_name=c.repo_name,commit_id=c.commit.raw_id,f_path=c.f_path))}
+                <%
+                    relative_urls = {
+                        'raw': h.route_path('repo_file_raw',repo_name=c.repo_name,commit_id=c.commit.raw_id,f_path=c.f_path),
+                        'standard': h.route_path('repo_files',repo_name=c.repo_name,commit_id=c.commit.raw_id,f_path=c.f_path),
+                    }
+                %>
+                ${h.render(c.file.content, renderer=c.renderer, relative_urls=relative_urls)}
             %else:
                 <table class="cb codehilite">
                 %if c.annotate:
                   <% color_hasher = h.color_hasher() %>
                   %for annotation, lines in c.annotated_lines:
                     ${sourceblock.render_annotation_lines(annotation, lines, color_hasher)}
                   %endfor
                 %else:
                   %for line_num, tokens in enumerate(c.lines, 1):
                     ${sourceblock.render_line(line_num, tokens)}
                   %endfor
                 %endif
                 </table>
             %endif
         %else:
             ${_('File size {} is bigger then allowed limit {}. ').format(h.format_byte_size_binary(c.file.size), h.format_byte_size_binary(c.visual.cut_off_limit_file))} ${h.link_to(_('Show as raw'),
             h.route_path('repo_file_raw',repo_name=c.repo_name,commit_id=c.commit.raw_id,f_path=c.f_path))}
         %endif
      %endif
     </div>
 </div>
  No newline at end of file

rhodecode/tests/lib/test_markup_renderer.py

@@ -1,255 +1,258 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2010-2017 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 pytest
 
 from rhodecode.lib.markup_renderer import (
     MarkupRenderer, RstTemplateRenderer, relative_path, relative_links)
 
 
 @pytest.mark.parametrize(
     "filename, expected_renderer",
     [
         ('readme.md', 'markdown'),
         ('readme.Md', 'markdown'),
         ('readme.MdoWn', 'markdown'),
         ('readme.rst', 'rst'),
         ('readme.Rst', 'rst'),
         ('readme.rest', 'rst'),
         ('readme.rest', 'rst'),
 
         ('markdown.xml', 'plain'),
         ('rest.xml', 'plain'),
         ('readme.xml', 'plain'),
 
         ('readme', 'plain'),
         ('README', 'plain'),
         ('readme.mdx', 'plain'),
         ('readme.rstx', 'plain'),
         ('readmex', 'plain'),
     ])
 def test_detect_renderer(filename, expected_renderer):
     detected_renderer = MarkupRenderer()._detect_renderer(
         '', filename=filename).__name__
     assert expected_renderer == detected_renderer
 
 
 def test_markdown_xss_link():
     xss_md = "[link](javascript:alert('XSS: pwned!'))"
     rendered_html = MarkupRenderer.markdown(xss_md)
     assert 'href="javascript:alert(\'XSS: pwned!\')"' not in rendered_html
 
 
 def test_markdown_xss_inline_html():
     xss_md = '\n'.join([
         '> <a name="n"',
         '> href="javascript:alert(\'XSS: pwned!\')">link</a>'])
     rendered_html = MarkupRenderer.markdown(xss_md)
     assert 'href="javascript:alert(\'XSS: pwned!\')">' not in rendered_html
 
 
 def test_markdown_inline_html():
     xss_md = '\n'.join(['> <a name="n"',
                         '> href="https://rhodecode.com">link</a>'])
     rendered_html = MarkupRenderer.markdown(xss_md)
     assert '[HTML_REMOVED]link[HTML_REMOVED]' in rendered_html
 
 
 def test_rst_xss_link():
     xss_rst = "`Link<javascript:alert('XSS: pwned!')>`_"
     rendered_html = MarkupRenderer.rst(xss_rst)
     assert "href=javascript:alert('XSS: pwned!')" not in rendered_html
 
 
 @pytest.mark.xfail(reason='Bug in docutils. Waiting answer from the author')
 def test_rst_xss_inline_html():
     xss_rst = '<a href="javascript:alert(\'XSS: pwned!\')">link</a>'
     rendered_html = MarkupRenderer.rst(xss_rst)
     assert 'href="javascript:alert(' not in rendered_html
 
 
 def test_rst_xss_raw_directive():
     xss_rst = '\n'.join([
         '.. raw:: html',
         '',
         '  <a href="javascript:alert(\'XSS: pwned!\')">link</a>'])
     rendered_html = MarkupRenderer.rst(xss_rst)
     assert 'href="javascript:alert(' not in rendered_html
 
 
 def test_render_rst_template_without_files():
     expected = u'''\
 Pull request updated. Auto status change to |under_review|
 
 .. role:: added
 .. role:: removed
 .. parsed-literal::
 
   Changed commits:
     * :added:`2 added`
     * :removed:`3 removed`
 
   No file changes found
 
 .. |under_review| replace:: *"NEW STATUS"*'''
 
     params = {
         'under_review_label': 'NEW STATUS',
         'added_commits': ['a', 'b'],
         'removed_commits': ['a', 'b', 'c'],
         'changed_files': [],
         'added_files': [],
         'modified_files': [],
         'removed_files': [],
     }
     renderer = RstTemplateRenderer()
     rendered = renderer.render('pull_request_update.mako', **params)
     assert expected == rendered
 
 
 def test_render_rst_template_with_files():
     expected = u'''\
 Pull request updated. Auto status change to |under_review|
 
 .. role:: added
 .. role:: removed
 .. parsed-literal::
 
   Changed commits:
     * :added:`1 added`
     * :removed:`3 removed`
 
   Changed files:
     * `A /path/a.py <#a_c--68ed34923b68>`_
     * `A /path/b.js <#a_c--64f90608b607>`_
     * `M /path/d.js <#a_c--85842bf30c6e>`_
     * `M /path/ę.py <#a_c--d713adf009cd>`_
     * R /path/ź.py
 
 .. |under_review| replace:: *"NEW STATUS"*'''
 
     added = ['/path/a.py', '/path/b.js']
     modified = ['/path/d.js', u'/path/ę.py']
     removed = [u'/path/ź.py']
 
     params = {
         'under_review_label': 'NEW STATUS',
         'added_commits': ['a'],
         'removed_commits': ['a', 'b', 'c'],
         'changed_files': added + modified + removed,
         'added_files': added,
         'modified_files': modified,
         'removed_files': removed,
     }
     renderer = RstTemplateRenderer()
     rendered = renderer.render('pull_request_update.mako', **params)
 
     assert expected == rendered
 
 
 def test_render_rst_auto_status_template():
     expected = u'''\
 Auto status change to |new_status|
 
 .. |new_status| replace:: *"NEW STATUS"*'''
 
     params = {
         'new_status_label': 'NEW STATUS',
         'pull_request': None,
         'commit_id': None,
     }
     renderer = RstTemplateRenderer()
     rendered = renderer.render('auto_status_change.mako', **params)
     assert expected == rendered
 
 
 @pytest.mark.parametrize(
     "src_path, server_path, is_path, expected",
     [
         ('source.png', '/repo/files/path', lambda p: False,
          '/repo/files/path/source.png'),
 
         ('source.png', 'mk/git/blob/master/README.md', lambda p: True,
          '/mk/git/blob/master/source.png'),
 
         ('./source.png', 'mk/git/blob/master/README.md', lambda p: True,
          '/mk/git/blob/master/source.png'),
 
         ('/source.png', 'mk/git/blob/master/README.md', lambda p: True,
          '/mk/git/blob/master/source.png'),
 
         ('./source.png', 'repo/files/path/source.md', lambda p: True,
          '/repo/files/path/source.png'),
 
         ('./source.png', '/repo/files/path/file.md', lambda p: True,
          '/repo/files/path/source.png'),
 
         ('../source.png', '/repo/files/path/file.md', lambda p: True,
          '/repo/files/source.png'),
 
         ('./../source.png', '/repo/files/path/file.md', lambda p: True,
          '/repo/files/source.png'),
 
         ('./source.png', '/repo/files/path/file.md', lambda p: True,
          '/repo/files/path/source.png'),
 
         ('../../../source.png', 'path/file.md', lambda p: True,
          '/source.png'),
 
         ('../../../../../source.png', '/path/file.md', None,
          '/source.png'),
 
         ('../../../../../source.png', 'files/path/file.md', None,
          '/source.png'),
 
         ('../../../../../https://google.com/image.png', 'files/path/file.md', None,
          '/https://google.com/image.png'),
 
         ('https://google.com/image.png', 'files/path/file.md', None,
          'https://google.com/image.png'),
 
         ('://foo', '/files/path/file.md', None,
          '://foo'),
 
         (u'한글.png', '/files/path/file.md', None,
          u'/files/path/한글.png'),
 
         ('my custom image.png', '/files/path/file.md', None,
          '/files/path/my custom image.png'),
     ])
 def test_relative_path(src_path, server_path, is_path, expected):
     path = relative_path(src_path, server_path, is_path)
     assert path == expected
 
 
 @pytest.mark.parametrize(
     "src_html, expected_html",
     [
         ('<div></div>', '<div></div>'),
         ('<img src="/file.png"></img>', '<img src="/path/raw/file.png">'),
         ('<img src="data:abcd"/>', '<img src="data:abcd">'),
-        ('<a href="/file.png"></a>', '<a href="/path/raw/file.png"></a>'),
+        ('<a href="/file.png?raw=1"></a>', '<a href="/path/raw/file.png?raw=1"></a>'),
+        ('<a href="/file.png"></a>', '<a href="/path/file.png"></a>'),
         ('<a href="#anchor"></a>', '<a href="#anchor"></a>'),
-        ('<a href="./README.md"></a>', '<a href="/path/raw/README.md"></a>'),
-        ('<a href="../README.md"></a>', '<a href="/path/README.md"></a>'),
+        ('<a href="./README.md?raw=1"></a>', '<a href="/path/raw/README.md?raw=1"></a>'),
+        ('<a href="./README.md"></a>', '<a href="/path/README.md"></a>'),
+        ('<a href="../README.md"></a>', '<a href="/README.md"></a>'),
 
     ])
 def test_relative_links(src_html, expected_html):
-    assert relative_links(src_html, '/path/raw/file.md') == expected_html
+    server_paths = {'raw': '/path/raw/file.md', 'standard': '/path/file.md'}
+    assert relative_links(src_html, server_paths=server_paths) == expected_html