lib: use bleach to sanitize HTML generated from markdown - fix XSS issue when repo front page shows README.md

Reported by Bob Hogg <[e-mail]> .

kallithea/lib/markup_renderer.py

@@ -1,214 +1,227 b''
 # -*- coding: utf-8 -*-
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
 # the Free Software Foundation, either version 3 of the License, or
 # (at your option) any later version.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 """
 kallithea.lib.markup_renderer
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 Renderer for markup languages with ability to parse using rst or markdown
 
 This file was forked by the Kallithea project in July 2014.
 Original author and date, and relevant copyright and licensing information is below:
 :created_on: Oct 27, 2011
 :author: marcink
 :copyright: (c) 2013 RhodeCode GmbH, and others.
 :license: GPLv3, see LICENSE.md for more details.
 """
 
 
 import re
 import logging
 import traceback
 
 import markdown as markdown_mod
+import bleach
 
 from kallithea.lib.utils2 import safe_unicode, MENTIONS_REGEX
 
 log = logging.getLogger(__name__)
 
 
 url_re = re.compile(r'''(\bhttps?://(?:[\da-zA-Z0-9@:.-]+)'''
                     r'''(?:[/a-zA-Z0-9_=@#~&+%.,:;?!*()-]*[/a-zA-Z0-9_=@#~])?)''')
 
 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)
     PLAIN_PAT = re.compile(r'readme', re.IGNORECASE)
 
     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.PLAIN_PAT.findall(filename):
             detected_renderer = 'rst'
         else:
             detected_renderer = 'plain'
 
         return getattr(MarkupRenderer, detected_renderer)
 
     @classmethod
     def _flavored_markdown(cls, text):
         """
         Github style flavored markdown
 
         :param text:
         """
         from hashlib import md5
 
         # Extract pre blocks.
         extractions = {}
         def pre_extraction_callback(matchobj):
             digest = md5(matchobj.group(0)).hexdigest()
             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('_', '\_')
             return s
         text = re.sub(r'^(?! {4}|\t)\w+_\w+_\w[\w_]*', italic_callback, text)
 
         # In very clear cases, let newlines become <br /> tags.
         def newline_callback(matchobj):
             if len(matchobj.group(1)) == 1:
                 return matchobj.group(0).rstrip() + '  \n'
             else:
                 return matchobj.group(0)
         pattern = re.compile(r'^[\w\<][^\n]*(\n+)', re.MULTILINE)
         text = re.sub(pattern, newline_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
 
     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 plain(cls, source, universal_newline=True):
         source = safe_unicode(source)
         if universal_newline:
             newline = '\n'
             source = newline.join(source.splitlines())
 
         def url_func(match_obj):
             url_full = match_obj.groups()[0]
             return '<a href="%(url)s">%(url)s</a>' % ({'url': url_full})
         source = url_re.sub(url_func, source)
         return '<br />' + source.replace("\n", '<br />')
 
     @classmethod
     def markdown(cls, source, safe=True, flavored=False):
         """
-        Convert Markdown (possibly GitHub Flavored) to HTML, possibly
+        Convert Markdown (possibly GitHub Flavored) to XSS safe HTML, possibly
         with "safe" fall-back to plaintext.
 
         >>> MarkupRenderer.markdown('''<img id="a" style="margin-top:-1000px;color:red" src="http://example.com/test.jpg">''')
-        u'<p><img id="a" style="margin-top:-1000px;color:red" src="http://example.com/test.jpg"></p>'
+        u'<p><img id="a" src="http://example.com/test.jpg" style="color: red;"></p>'
         >>> MarkupRenderer.markdown('''<img class="c d" src="file://localhost/test.jpg">''')
-        u'<p><img class="c d" src="file://localhost/test.jpg"></p>'
+        u'<p><img class="c d"></p>'
         >>> MarkupRenderer.markdown('''<a href="foo">foo</a>''')
         u'<p><a href="foo">foo</a></p>'
         >>> MarkupRenderer.markdown('''<script>alert(1)</script>''')
-        u'<script>alert(1)</script>'
+        u'&lt;script&gt;alert(1)&lt;/script&gt;'
         >>> MarkupRenderer.markdown('''<div onclick="alert(2)">yo</div>''')
-        u'<div onclick="alert(2)">yo</div>'
+        u'<div>yo</div>'
         >>> MarkupRenderer.markdown('''<a href="javascript:alert(3)">yo</a>''')
-        u'<p><a href="javascript:alert(3)">yo</a></p>'
+        u'<p><a>yo</a></p>'
         """
         source = safe_unicode(source)
         try:
             if flavored:
                 source = cls._flavored_markdown(source)
             markdown_html = markdown_mod.markdown(source, ['codehilite', 'extra'])
-            return markdown_html
+            # Allow most HTML, while preventing XSS issues:
+            # no <script> tags, no onclick attributes, no javascript
+            # "protocol", and also limit styling to prevent defacing.
+            return bleach.clean(markdown_html,
+                tags=['a', 'abbr', 'b', 'blockquote', 'br', 'code', 'dd',
+                      'div', 'dl', 'dt', 'em', 'h1', 'h2', 'h3', 'h4', 'h5',
+                      'h6', 'hr', 'i', 'img', 'li', 'ol', 'p', 'pre', 'span',
+                      'strong', 'sub', 'sup', 'table', 'tbody', 'td', 'th',
+                      'thead', 'tr', 'ul'],
+                attributes=['class', 'id', 'style', 'label', 'title', 'alt', 'href', 'src'],
+                styles=['color'],
+                protocols=['http', 'https', 'mailto'],
+                )
         except Exception:
             log.error(traceback.format_exc())
             if safe:
                 log.debug('Falling back to render in plain mode')
                 return cls.plain(source)
             else:
                 raise
 
     @classmethod
     def rst(cls, source, safe=True):
         source = safe_unicode(source)
         try:
             from docutils.core import publish_parts
             from docutils.parsers.rst import directives
             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_name="html4css1",
                                   settings_overrides=docutils_settings)
 
             return parts['html_title'] + parts["fragment"]
         except ImportError:
             log.warning('Install docutils to use this function')
             return cls.plain(source)
         except Exception:
             log.error(traceback.format_exc())
             if safe:
                 log.debug('Falling back to render in plain mode')
                 return cls.plain(source)
             else:
                 raise
 
     @classmethod
     def rst_with_mentions(cls, source):
         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()
         return cls.rst(mention_hl)

setup.py

@@ -1,192 +1,193 b''
 #!/usr/bin/env python2
 # -*- coding: utf-8 -*-
 import os
 import sys
 import platform
 
 if sys.version_info < (2, 6):
     raise Exception('Kallithea requires python 2.6 or 2.7')
 
 
 here = os.path.abspath(os.path.dirname(__file__))
 
 
 def _get_meta_var(name, data, callback_handler=None):
     import re
     matches = re.compile(r'(?:%s)\s*=\s*(.*)' % name).search(data)
     if matches:
         if not callable(callback_handler):
             callback_handler = lambda v: v
 
         return callback_handler(eval(matches.groups()[0]))
 
 _meta = open(os.path.join(here, 'kallithea', '__init__.py'), 'rb')
 _metadata = _meta.read()
 _meta.close()
 
 callback = lambda V: ('.'.join(map(str, V[:3])) + '.'.join(V[3:]))
 __version__ = _get_meta_var('VERSION', _metadata, callback)
 __license__ = _get_meta_var('__license__', _metadata)
 __author__ = _get_meta_var('__author__', _metadata)
 __url__ = _get_meta_var('__url__', _metadata)
 # defines current platform
 __platform__ = platform.system()
 
 is_windows = __platform__ in ['Windows']
 
 requirements = [
     "setuptools<34", # setuptools==34 has an undeclared requirement of pyparsing >=2.1, but celery<2.3 requires pyparsing<2
     "waitress==0.8.8",
     "webob>=1.0.8,<=1.1.1",
     "webtest==1.4.3",
     "Pylons>=1.0.0,<=1.0.3",
     "Beaker==1.6.4",
     "WebHelpers==1.3",
     "formencode>=1.2.4,<=1.2.6",
     "SQLAlchemy==0.7.10",
     "Mako>=0.9.0,<=1.0.0",
     "pygments>=1.5",
     "whoosh>=2.4.0,<=2.5.7",
     "celery>=2.2.5,<2.3",
     "babel>=0.9.6,<=1.3",
     "python-dateutil>=1.5.0,<2.0.0",
     "markdown==2.2.1",
     "docutils>=0.8.1,<=0.11",
     "mock",
     "URLObject==2.3.4",
     "Routes==1.13",
     "dulwich>=0.9.9,<=0.9.9",
     "mercurial>=2.9,<4.3",
+    "bleach >= 3.0, < 3.1",
 ]
 
 if sys.version_info < (2, 7):
     requirements.append("importlib==1.0.1")
     requirements.append("unittest2")
     requirements.append("argparse")
 
 if not is_windows:
     requirements.append("py-bcrypt>=0.3.0,<=0.4")
 
 
 dependency_links = [
 ]
 
 classifiers = [
     'Development Status :: 4 - Beta',
     'Environment :: Web Environment',
     'Framework :: Pylons',
     'Intended Audience :: Developers',
     'License :: OSI Approved :: GNU General Public License (GPL)',
     'Operating System :: OS Independent',
     'Programming Language :: Python',
     'Programming Language :: Python :: 2.6',
     'Programming Language :: Python :: 2.7',
     'Topic :: Software Development :: Version Control',
 ]
 
 
 # additional files from project that goes somewhere in the filesystem
 # relative to sys.prefix
 data_files = []
 
 # additional files that goes into package itself
 package_data = {'kallithea': ['i18n/*/LC_MESSAGES/*.mo', ], }
 
 description = ('Kallithea is a fast and powerful management tool '
                'for Mercurial and Git with a built in push/pull server, '
                'full text search and code-review.')
 
 keywords = ' '.join([
     'kallithea', 'mercurial', 'git', 'code review',
     'repo groups', 'ldap', 'repository management', 'hgweb replacement',
     'hgwebdir', 'gitweb replacement', 'serving hgweb',
 ])
 
 # long description
 README_FILE = 'README.rst'
 CHANGELOG_FILE = 'docs/changelog.rst'
 try:
     long_description = open(README_FILE).read() + '\n\n' + \
         open(CHANGELOG_FILE).read()
 
 except IOError as err:
     sys.stderr.write(
         "[WARNING] Cannot find file specified as long_description (%s)\n or "
         "changelog (%s) skipping that file" % (README_FILE, CHANGELOG_FILE)
     )
     long_description = description
 
 try:
     from setuptools import setup, find_packages
 except ImportError:
     from ez_setup import use_setuptools
     use_setuptools()
     from setuptools import setup, find_packages
 
 # monkey patch setuptools to use distutils owner/group functionality
 from setuptools.command import sdist
 sdist_org = sdist.sdist
 class sdist_new(sdist_org):
     def initialize_options(self):
         sdist_org.initialize_options(self)
         self.owner = self.group = 'root'
 sdist.sdist = sdist_new
 
 # packages
 packages = find_packages(exclude=['ez_setup'])
 
 setup(
     name='Kallithea',
     version=__version__,
     description=description,
     long_description=long_description,
     keywords=keywords,
     license=__license__,
     author=__author__,
     author_email='[e-mail]',
     dependency_links=dependency_links,
     url=__url__,
     install_requires=requirements,
     classifiers=classifiers,
     setup_requires=["PasteScript>=1.6.3"],
     data_files=data_files,
     packages=packages,
     include_package_data=True,
     test_suite='nose.collector',
     package_data=package_data,
     message_extractors={'kallithea': [
             ('**.py', 'python', None),
             ('templates/**.mako', 'mako', {'input_encoding': 'utf-8'}),
             ('templates/**.html', 'mako', {'input_encoding': 'utf-8'}),
             ('public/**', 'ignore', None)]},
     zip_safe=False,
     paster_plugins=['PasteScript', 'Pylons'],
     entry_points="""
     [console_scripts]
     kallithea-api =    kallithea.bin.kallithea_api:main
     kallithea-gist =   kallithea.bin.kallithea_gist:main
     kallithea-config = kallithea.bin.kallithea_config:main
 
     [paste.app_factory]
     main = kallithea.config.middleware:make_app
 
     [paste.app_install]
     main = pylons.util:PylonsInstaller
 
     [paste.global_paster_command]
     setup-db=kallithea.lib.paster_commands.setup_db:Command
     cleanup-repos=kallithea.lib.paster_commands.cleanup:Command
     update-repoinfo=kallithea.lib.paster_commands.update_repoinfo:Command
     make-rcext=kallithea.lib.paster_commands.make_rcextensions:Command
     repo-scan=kallithea.lib.paster_commands.repo_scan:Command
     cache-keys=kallithea.lib.paster_commands.cache_keys:Command
     ishell=kallithea.lib.paster_commands.ishell:Command
     make-index=kallithea.lib.paster_commands.make_index:Command
     upgrade-db=kallithea.lib.dbmigrate:UpgradeDb
     celeryd=kallithea.lib.celerypylons.commands:CeleryDaemonCommand
     install-iis=kallithea.lib.paster_commands.install_iis:Command
 
     [nose.plugins]
     pylons = pylons.test:PylonsPlugin
     """,
 )