This file is indexed.

/usr/share/pyshared/bleach/sanitizer.py is in python-bleach 1.2.2-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
import re
from xml.sax.saxutils import escape, unescape

from html5lib.constants import tokenTypes
from html5lib.sanitizer import HTMLSanitizerMixin
from html5lib.tokenizer import HTMLTokenizer


PROTOS = HTMLSanitizerMixin.acceptable_protocols
PROTOS.remove('feed')


class BleachSanitizerMixin(HTMLSanitizerMixin):
    """Mixin to replace sanitize_token() and sanitize_css()."""

    allowed_svg_properties = []
    # TODO: When the next html5lib version comes out, nuke this.
    attr_val_is_uri = HTMLSanitizerMixin.attr_val_is_uri + ['poster']

    def sanitize_token(self, token):
        """Sanitize a token either by HTML-encoding or dropping.

        Unlike HTMLSanitizerMixin.sanitize_token, allowed_attributes can be
        a dict of {'tag': ['attribute', 'pairs'], 'tag': callable}.

        Here callable is a function with two arguments of attribute name
        and value. It should return true of false.

        Also gives the option to strip tags instead of encoding.

        """
        if (getattr(self, 'wildcard_attributes', None) is None and
            isinstance(self.allowed_attributes, dict)):
            self.wildcard_attributes = self.allowed_attributes.get('*', [])

        if token['type'] in (tokenTypes['StartTag'], tokenTypes['EndTag'],
                             tokenTypes['EmptyTag']):
            if token['name'] in self.allowed_elements:
                if 'data' in token:
                    if isinstance(self.allowed_attributes, dict):
                        allowed_attributes = self.allowed_attributes.get(
                            token['name'], [])
                        if not callable(allowed_attributes):
                            allowed_attributes += self.wildcard_attributes
                    else:
                        allowed_attributes = self.allowed_attributes
                    attrs = dict([(name, val) for name, val in
                                  token['data'][::-1]
                                  if (allowed_attributes(name, val)
                                      if callable(allowed_attributes)
                                      else name in allowed_attributes)])
                    for attr in self.attr_val_is_uri:
                        if not attr in attrs:
                            continue
                        val_unescaped = re.sub("[`\000-\040\177-\240\s]+", '',
                                               unescape(attrs[attr])).lower()
                        # Remove replacement characters from unescaped
                        # characters.
                        val_unescaped = val_unescaped.replace(u"\ufffd", "")
                        if (re.match(r'^[a-z0-9][-+.a-z0-9]*:', val_unescaped)
                            and (val_unescaped.split(':')[0] not in
                                 self.allowed_protocols)):
                            del attrs[attr]
                    for attr in self.svg_attr_val_allows_ref:
                        if attr in attrs:
                            attrs[attr] = re.sub(r'url\s*\(\s*[^#\s][^)]+?\)',
                                                 ' ',
                                                 unescape(attrs[attr]))
                    if (token['name'] in self.svg_allow_local_href and
                        'xlink:href' in attrs and
                        re.search(r'^\s*[^#\s].*', attrs['xlink:href'])):
                        del attrs['xlink:href']
                    if 'style' in attrs:
                        attrs['style'] = self.sanitize_css(attrs['style'])
                    token['data'] = [(name, val) for name, val in
                                     attrs.items()]
                return token
            elif self.strip_disallowed_elements:
                pass
            else:
                if token['type'] == tokenTypes['EndTag']:
                    token['data'] = '</%s>' % token['name']
                elif token['data']:
                    attrs = ''.join([' %s="%s"' % (k, escape(v)) for k, v in
                                    token['data']])
                    token['data'] = '<%s%s>' % (token['name'], attrs)
                else:
                    token['data'] = '<%s>' % token['name']
                if token['selfClosing']:
                    token['data'] = token['data'][:-1] + '/>'
                token['type'] = tokenTypes['Characters']
                del token["name"]
                return token
        elif token['type'] == tokenTypes['Comment']:
            if not self.strip_html_comments:
                return token
        else:
            return token

    def sanitize_css(self, style):
        """HTMLSanitizerMixin.sanitize_css replacement.

        HTMLSanitizerMixin.sanitize_css always whitelists background-*,
        border-*, margin-*, and padding-*. We only whitelist what's in
        the whitelist.

        """
        # disallow urls
        style = re.compile('url\s*\(\s*[^\s)]+?\s*\)\s*').sub(' ', style)

        # gauntlet
        # TODO: Make sure this does what it's meant to - I *think* it wants to
        # validate style attribute contents.
        parts = style.split(';')
        gauntlet = re.compile("""^([-/:,#%.'"\sa-zA-Z0-9!]|\w-\w|'[\s\w]+'\s*"""
                              """|"[\s\w]+"|\([\d,%\.\s]+\))*$""")
        for part in parts:
            if not gauntlet.match(part):
                return ''

        if not re.match("^\s*([-\w]+\s*:[^:;]*(;\s*|$))*$", style):
            return ''

        clean = []
        for prop, value in re.findall('([-\w]+)\s*:\s*([^:;]*)', style):
            if not value:
                continue
            if prop.lower() in self.allowed_css_properties:
                clean.append(prop + ': ' + value + ';')
            elif prop.lower() in self.allowed_svg_properties:
                clean.append(prop + ': ' + value + ';')

        return ' '.join(clean)


class BleachSanitizer(HTMLTokenizer, BleachSanitizerMixin):
    def __init__(self, stream, encoding=None, parseMeta=True, useChardet=True,
                 lowercaseElementName=True, lowercaseAttrName=True, **kwargs):
        HTMLTokenizer.__init__(self, stream, encoding, parseMeta, useChardet,
                               lowercaseElementName, lowercaseAttrName,
                               **kwargs)

    def __iter__(self):
        for token in HTMLTokenizer.__iter__(self):
            token = self.sanitize_token(token)
            if token:
                yield token