Mercurial > genshi > genshi-test
annotate genshi/filters/html.py @ 908:5fd4a1e28351
Fix for bug with the `HTMLFormFiller` in the handling of textareas. Thanks to Trevor Morgan for pointing this out on the mailing list.
author | cmlenz |
---|---|
date | Mon, 10 May 2010 14:02:55 +0000 |
parents | 1e2be9fb3348 |
children | 585fdbd30e05 |
rev | line source |
---|---|
1 | 1 # -*- coding: utf-8 -*- |
2 # | |
854
0d9e87c6cf6e
More work on reducing the size of the diff produced by 2to3.
cmlenz
parents:
853
diff
changeset
|
3 # Copyright (C) 2006-2009 Edgewall Software |
1 | 4 # All rights reserved. |
5 # | |
6 # This software is licensed as described in the file COPYING, which | |
7 # you should have received as part of this distribution. The terms | |
230 | 8 # are also available at http://genshi.edgewall.org/wiki/License. |
1 | 9 # |
10 # This software consists of voluntary contributions made by many | |
11 # individuals. For the exact contribution history, see the revision | |
230 | 12 # history and logs, available at http://genshi.edgewall.org/log/. |
1 | 13 |
14 """Implementation of a number of stream filters.""" | |
15 | |
856 | 16 try: |
17 any | |
18 except NameError: | |
19 from genshi.util import any | |
1 | 20 import re |
21 | |
403
32b283e1d310
Remove some magic/overhead from `Attrs` creation and manipulation by not automatically wrapping attribute names in `QName`.
cmlenz
parents:
363
diff
changeset
|
22 from genshi.core import Attrs, QName, stripentities |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
23 from genshi.core import END, START, TEXT, COMMENT |
1 | 24 |
363
caf7b68ab5dc
Parse template includes at parse time to avoid some runtime overhead.
cmlenz
parents:
345
diff
changeset
|
25 __all__ = ['HTMLFormFiller', 'HTMLSanitizer'] |
425
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
26 __docformat__ = 'restructuredtext en' |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
27 |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
28 |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
29 class HTMLFormFiller(object): |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
30 """A stream filter that can populate HTML forms from a dictionary of values. |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
31 |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
32 >>> from genshi.input import HTML |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
33 >>> html = HTML('''<form> |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
34 ... <p><input type="text" name="foo" /></p> |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
35 ... </form>''') |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
36 >>> filler = HTMLFormFiller(data={'foo': 'bar'}) |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
844
diff
changeset
|
37 >>> print(html | filler) |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
38 <form> |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
39 <p><input type="text" name="foo" value="bar"/></p> |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
40 </form> |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
41 """ |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
42 # TODO: only select the first radio button, and the first select option |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
43 # (if not in a multiple-select) |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
44 # TODO: only apply to elements in the XHTML namespace (or no namespace)? |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
45 |
841
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
46 def __init__(self, name=None, id=None, data=None, passwords=False): |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
47 """Create the filter. |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
48 |
425
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
49 :param name: The name of the form that should be populated. If this |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
50 parameter is given, only forms where the ``name`` attribute |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
51 value matches the parameter are processed. |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
52 :param id: The ID of the form that should be populated. If this |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
53 parameter is given, only forms where the ``id`` attribute |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
54 value matches the parameter are processed. |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
55 :param data: The dictionary of form values, where the keys are the names |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
56 of the form fields, and the values are the values to fill |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
57 in. |
841
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
58 :param passwords: Whether password input fields should be populated. |
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
59 This is off by default for security reasons (for |
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
60 example, a password may end up in the browser cache) |
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
61 :note: Changed in 0.5.2: added the `passwords` option |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
62 """ |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
63 self.name = name |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
64 self.id = id |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
65 if data is None: |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
66 data = {} |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
67 self.data = data |
841
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
68 self.passwords = passwords |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
69 |
439
b84c11a49ad2
Add support for adding custom template filters by passing a custom callback function to the `TemplateLoader`. Closes #89 (see added unit test).
cmlenz
parents:
431
diff
changeset
|
70 def __call__(self, stream): |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
71 """Apply the filter to the given stream. |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
72 |
425
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
73 :param stream: the markup event stream to filter |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
74 """ |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
75 in_form = in_select = in_option = in_textarea = False |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
76 select_value = option_value = textarea_value = None |
584
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
77 option_start = None |
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
78 option_text = [] |
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
79 no_option_value = False |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
80 |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
81 for kind, data, pos in stream: |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
82 |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
83 if kind is START: |
345 | 84 tag, attrs = data |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
85 tagname = tag.localname |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
86 |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
87 if tagname == 'form' and ( |
345 | 88 self.name and attrs.get('name') == self.name or |
89 self.id and attrs.get('id') == self.id or | |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
90 not (self.id or self.name)): |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
91 in_form = True |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
92 |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
93 elif in_form: |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
94 if tagname == 'input': |
844
0af77f663a65
Fix two instances of using None, which would cause an AttributeError.
jruigrok
parents:
841
diff
changeset
|
95 type = attrs.get('type', '').lower() |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
96 if type in ('checkbox', 'radio'): |
345 | 97 name = attrs.get('name') |
471
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
98 if name and name in self.data: |
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
99 value = self.data[name] |
345 | 100 declval = attrs.get('value') |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
101 checked = False |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
102 if isinstance(value, (list, tuple)): |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
103 if declval: |
584
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
104 checked = declval in [unicode(v) for v |
415
c267061c961f
`HTMLFormFiller` now correctly deals with non-string values in the data dictionary for select/checkbox/radio controls.
cmlenz
parents:
408
diff
changeset
|
105 in value] |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
106 else: |
856 | 107 checked = any(value) |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
108 else: |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
109 if declval: |
584
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
110 checked = declval == unicode(value) |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
111 elif type == 'checkbox': |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
112 checked = bool(value) |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
113 if checked: |
403
32b283e1d310
Remove some magic/overhead from `Attrs` creation and manipulation by not automatically wrapping attribute names in `QName`.
cmlenz
parents:
363
diff
changeset
|
114 attrs |= [(QName('checked'), 'checked')] |
345 | 115 elif 'checked' in attrs: |
116 attrs -= 'checked' | |
844
0af77f663a65
Fix two instances of using None, which would cause an AttributeError.
jruigrok
parents:
841
diff
changeset
|
117 elif type in ('', 'hidden', 'text') \ |
841
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
118 or type == 'password' and self.passwords: |
345 | 119 name = attrs.get('name') |
471
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
120 if name and name in self.data: |
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
121 value = self.data[name] |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
122 if isinstance(value, (list, tuple)): |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
123 value = value[0] |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
124 if value is not None: |
841
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
125 attrs |= [ |
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
126 (QName('value'), unicode(value)) |
67ec9a402bdc
Added an option to the `HTMLFiller` to also populate password fields.
cmlenz
parents:
840
diff
changeset
|
127 ] |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
128 elif tagname == 'select': |
345 | 129 name = attrs.get('name') |
471
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
130 if name in self.data: |
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
131 select_value = self.data[name] |
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
132 in_select = True |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
133 elif tagname == 'textarea': |
345 | 134 name = attrs.get('name') |
471
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
135 if name in self.data: |
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
136 textarea_value = self.data.get(name) |
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
137 if isinstance(textarea_value, (list, tuple)): |
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
138 textarea_value = textarea_value[0] |
17ce8bf006d7
The `HTMLFormFiller` stream filter no longer alters form elements for which the data element contains no corresponding item.
cmlenz
parents:
446
diff
changeset
|
139 in_textarea = True |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
140 elif in_select and tagname == 'option': |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
141 option_start = kind, data, pos |
345 | 142 option_value = attrs.get('value') |
584
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
143 if option_value is None: |
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
144 no_option_value = True |
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
145 option_value = '' |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
146 in_option = True |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
147 continue |
345 | 148 yield kind, (tag, attrs), pos |
149 | |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
150 elif in_form and kind is TEXT: |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
151 if in_select and in_option: |
584
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
152 if no_option_value: |
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
153 option_value += data |
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
154 option_text.append((kind, data, pos)) |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
155 continue |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
156 elif in_textarea: |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
157 continue |
345 | 158 yield kind, data, pos |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
159 |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
160 elif in_form and kind is END: |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
161 tagname = data.localname |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
162 if tagname == 'form': |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
163 in_form = False |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
164 elif tagname == 'select': |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
165 in_select = False |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
166 select_value = None |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
167 elif in_select and tagname == 'option': |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
168 if isinstance(select_value, (tuple, list)): |
584
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
169 selected = option_value in [unicode(v) for v |
415
c267061c961f
`HTMLFormFiller` now correctly deals with non-string values in the data dictionary for select/checkbox/radio controls.
cmlenz
parents:
408
diff
changeset
|
170 in select_value] |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
171 else: |
584
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
172 selected = option_value == unicode(select_value) |
345 | 173 okind, (tag, attrs), opos = option_start |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
174 if selected: |
403
32b283e1d310
Remove some magic/overhead from `Attrs` creation and manipulation by not automatically wrapping attribute names in `QName`.
cmlenz
parents:
363
diff
changeset
|
175 attrs |= [(QName('selected'), 'selected')] |
345 | 176 elif 'selected' in attrs: |
177 attrs -= 'selected' | |
178 yield okind, (tag, attrs), opos | |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
179 if option_text: |
584
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
180 for event in option_text: |
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
181 yield event |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
182 in_option = False |
584
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
183 no_option_value = False |
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
184 option_start = option_value = None |
84137a71a4ca
Fixed a few cases where HTMLFormFiller didn't work well with option elements:
jonas
parents:
576
diff
changeset
|
185 option_text = [] |
908
5fd4a1e28351
Fix for bug with the `HTMLFormFiller` in the handling of textareas. Thanks to Trevor Morgan for pointing this out on the mailing list.
cmlenz
parents:
856
diff
changeset
|
186 elif in_textarea and tagname == 'textarea': |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
187 if textarea_value: |
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
188 yield TEXT, unicode(textarea_value), pos |
908
5fd4a1e28351
Fix for bug with the `HTMLFormFiller` in the handling of textareas. Thanks to Trevor Morgan for pointing this out on the mailing list.
cmlenz
parents:
856
diff
changeset
|
189 textarea_value = None |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
190 in_textarea = False |
345 | 191 yield kind, data, pos |
275
7f24dd6fb904
Integrated `HTMLFormFiller` filter initially presented as a [wiki:FormFilling#Usingatemplatefilter recipe].
cmlenz
parents:
230
diff
changeset
|
192 |
345 | 193 else: |
194 yield kind, data, pos | |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
195 |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
196 |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
197 class HTMLSanitizer(object): |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
198 """A filter that removes potentially dangerous HTML tags and attributes |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
199 from the stream. |
431
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
200 |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
201 >>> from genshi import HTML |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
202 >>> html = HTML('<div><script>alert(document.cookie)</script></div>') |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
844
diff
changeset
|
203 >>> print(html | HTMLSanitizer()) |
431
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
204 <div/> |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
205 |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
206 The default set of safe tags and attributes can be modified when the filter |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
207 is instantiated. For example, to allow inline ``style`` attributes, the |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
208 following instantation would work: |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
209 |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
210 >>> html = HTML('<div style="background: #000"></div>') |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
211 >>> sanitizer = HTMLSanitizer(safe_attrs=HTMLSanitizer.SAFE_ATTRS | set(['style'])) |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
844
diff
changeset
|
212 >>> print(html | sanitizer) |
431
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
213 <div style="background: #000"/> |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
214 |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
215 Note that even in this case, the filter *does* attempt to remove dangerous |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
216 constructs from style attributes: |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
217 |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
218 >>> html = HTML('<div style="background: url(javascript:void); color: #000"></div>') |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
844
diff
changeset
|
219 >>> print(html | sanitizer) |
431
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
220 <div style="color: #000"/> |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
221 |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
222 This handles HTML entities, unicode escapes in CSS and Javascript text, as |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
223 well as a lot of other things. However, the style tag is still excluded by |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
224 default because it is very hard for such sanitizing to be completely safe, |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
225 especially considering how much error recovery current web browsers perform. |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
226 |
840
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
227 It also does some basic filtering of CSS properties that may be used for |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
228 typical phishing attacks. For more sophisticated filtering, this class |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
229 provides a couple of hooks that can be overridden in sub-classes. |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
230 |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
231 :warn: Note that this special processing of CSS is currently only applied to |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
232 style attributes, **not** style elements. |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
233 """ |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
234 |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
235 SAFE_TAGS = frozenset(['a', 'abbr', 'acronym', 'address', 'area', 'b', |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
236 'big', 'blockquote', 'br', 'button', 'caption', 'center', 'cite', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
237 'code', 'col', 'colgroup', 'dd', 'del', 'dfn', 'dir', 'div', 'dl', 'dt', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
238 'em', 'fieldset', 'font', 'form', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
239 'hr', 'i', 'img', 'input', 'ins', 'kbd', 'label', 'legend', 'li', 'map', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
240 'menu', 'ol', 'optgroup', 'option', 'p', 'pre', 'q', 's', 'samp', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
241 'select', 'small', 'span', 'strike', 'strong', 'sub', 'sup', 'table', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
242 'tbody', 'td', 'textarea', 'tfoot', 'th', 'thead', 'tr', 'tt', 'u', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
243 'ul', 'var']) |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
244 |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
245 SAFE_ATTRS = frozenset(['abbr', 'accept', 'accept-charset', 'accesskey', |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
246 'action', 'align', 'alt', 'axis', 'bgcolor', 'border', 'cellpadding', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
247 'cellspacing', 'char', 'charoff', 'charset', 'checked', 'cite', 'class', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
248 'clear', 'cols', 'colspan', 'color', 'compact', 'coords', 'datetime', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
249 'dir', 'disabled', 'enctype', 'for', 'frame', 'headers', 'height', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
250 'href', 'hreflang', 'hspace', 'id', 'ismap', 'label', 'lang', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
251 'longdesc', 'maxlength', 'media', 'method', 'multiple', 'name', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
252 'nohref', 'noshade', 'nowrap', 'prompt', 'readonly', 'rel', 'rev', |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
253 'rows', 'rowspan', 'rules', 'scope', 'selected', 'shape', 'size', |
431
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
254 'span', 'src', 'start', 'summary', 'tabindex', 'target', 'title', |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
255 'type', 'usemap', 'valign', 'value', 'vspace', 'width']) |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
256 |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
257 SAFE_SCHEMES = frozenset(['file', 'ftp', 'http', 'https', 'mailto', None]) |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
258 |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
259 URI_ATTRS = frozenset(['action', 'background', 'dynsrc', 'href', 'lowsrc', |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
260 'src']) |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
261 |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
262 def __init__(self, safe_tags=SAFE_TAGS, safe_attrs=SAFE_ATTRS, |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
263 safe_schemes=SAFE_SCHEMES, uri_attrs=URI_ATTRS): |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
264 """Create the sanitizer. |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
265 |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
266 The exact set of allowed elements and attributes can be configured. |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
267 |
425
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
268 :param safe_tags: a set of tag names that are considered safe |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
269 :param safe_attrs: a set of attribute names that are considered safe |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
270 :param safe_schemes: a set of URI schemes that are considered safe |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
271 :param uri_attrs: a set of names of attributes that contain URIs |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
272 """ |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
273 self.safe_tags = safe_tags |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
274 "The set of tag names that are considered safe." |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
275 self.safe_attrs = safe_attrs |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
276 "The set of attribute names that are considered safe." |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
277 self.uri_attrs = uri_attrs |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
278 "The set of names of attributes that may contain URIs." |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
279 self.safe_schemes = safe_schemes |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
280 "The set of URI schemes that are considered safe." |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
281 |
439
b84c11a49ad2
Add support for adding custom template filters by passing a custom callback function to the `TemplateLoader`. Closes #89 (see added unit test).
cmlenz
parents:
431
diff
changeset
|
282 def __call__(self, stream): |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
283 """Apply the filter to the given stream. |
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
284 |
425
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
415
diff
changeset
|
285 :param stream: the markup event stream to filter |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
286 """ |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
287 waiting_for = None |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
288 |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
289 for kind, data, pos in stream: |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
290 if kind is START: |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
291 if waiting_for: |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
292 continue |
345 | 293 tag, attrs = data |
840
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
294 if not self.is_safe_elem(tag, attrs): |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
295 waiting_for = tag |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
296 continue |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
297 |
345 | 298 new_attrs = [] |
299 for attr, value in attrs: | |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
300 value = stripentities(value) |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
301 if attr not in self.safe_attrs: |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
302 continue |
277
c9fd81953169
The `HTMLSanitizer` now lets you override the default set of tag and attribute names that are considered safe.
cmlenz
parents:
275
diff
changeset
|
303 elif attr in self.uri_attrs: |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
304 # Don't allow URI schemes such as "javascript:" |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
305 if not self.is_safe_uri(value): |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
306 continue |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
307 elif attr == 'style': |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
308 # Remove dangerous CSS declarations from inline styles |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
309 decls = self.sanitize_css(value) |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
310 if not decls: |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
311 continue |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
312 value = '; '.join(decls) |
345 | 313 new_attrs.append((attr, value)) |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
314 |
345 | 315 yield kind, (tag, Attrs(new_attrs)), pos |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
316 |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
317 elif kind is END: |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
318 tag = data |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
319 if waiting_for: |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
320 if waiting_for == tag: |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
321 waiting_for = None |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
322 else: |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
323 yield kind, data, pos |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
324 |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
325 elif kind is not COMMENT: |
123
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
326 if not waiting_for: |
93bbdcf9428b
Fix for #18: whitespace in space-sensitive elements such as `<pre>` and `<textarea>` is now preserved.
cmlenz
parents:
113
diff
changeset
|
327 yield kind, data, pos |
431
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
328 |
840
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
329 def is_safe_css(self, propname, value): |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
330 """Determine whether the given css property declaration is to be |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
331 considered safe for inclusion in the output. |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
332 |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
333 :param propname: the CSS property name |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
334 :param value: the value of the property |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
335 :return: whether the property value should be considered safe |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
336 :rtype: bool |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
337 :since: version 0.6 |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
338 """ |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
339 if propname == 'position': |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
340 return False |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
341 if propname.startswith('margin') and '-' in value: |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
342 # Negative margins can be used for phishing |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
343 return False |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
344 return True |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
345 |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
346 def is_safe_elem(self, tag, attrs): |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
347 """Determine whether the given element should be considered safe for |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
348 inclusion in the output. |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
349 |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
350 :param tag: the tag name of the element |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
351 :type tag: QName |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
352 :param attrs: the element attributes |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
353 :type attrs: Attrs |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
354 :return: whether the element should be considered safe |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
355 :rtype: bool |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
356 :since: version 0.6 |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
357 """ |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
358 if tag not in self.safe_tags: |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
359 return False |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
360 if tag.localname == 'input': |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
361 input_type = attrs.get('type', '').lower() |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
362 if input_type == 'password': |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
363 return False |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
364 return True |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
365 |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
366 def is_safe_uri(self, uri): |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
367 """Determine whether the given URI is to be considered safe for |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
368 inclusion in the output. |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
369 |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
370 The default implementation checks whether the scheme of the URI is in |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
371 the set of allowed URIs (`safe_schemes`). |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
372 |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
373 >>> sanitizer = HTMLSanitizer() |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
374 >>> sanitizer.is_safe_uri('http://example.org/') |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
375 True |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
376 >>> sanitizer.is_safe_uri('javascript:alert(document.cookie)') |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
377 False |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
378 |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
379 :param uri: the URI to check |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
380 :return: `True` if the URI can be considered safe, `False` otherwise |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
381 :rtype: `bool` |
576 | 382 :since: version 0.4.3 |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
383 """ |
837 | 384 if '#' in uri: |
385 uri = uri.split('#', 1)[0] # Strip out the fragment identifier | |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
386 if ':' not in uri: |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
387 return True # This is a relative URI |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
388 chars = [char for char in uri.split(':', 1)[0] if char.isalnum()] |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
389 return ''.join(chars).lower() in self.safe_schemes |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
390 |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
391 def sanitize_css(self, text): |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
392 """Remove potentially dangerous property declarations from CSS code. |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
393 |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
394 In particular, properties using the CSS ``url()`` function with a scheme |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
395 that is not considered safe are removed: |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
396 |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
397 >>> sanitizer = HTMLSanitizer() |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
398 >>> sanitizer.sanitize_css(u''' |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
399 ... background: url(javascript:alert("foo")); |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
400 ... color: #000; |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
401 ... ''') |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
402 [u'color: #000'] |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
403 |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
404 Also, the proprietary Internet Explorer function ``expression()`` is |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
405 always stripped: |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
406 |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
407 >>> sanitizer.sanitize_css(u''' |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
408 ... background: #fff; |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
409 ... color: #000; |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
410 ... width: e/**/xpression(alert("foo")); |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
411 ... ''') |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
412 [u'background: #fff', u'color: #000'] |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
413 |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
414 :param text: the CSS text; this is expected to be `unicode` and to not |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
415 contain any character or numeric references |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
416 :return: a list of declarations that are considered safe |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
417 :rtype: `list` |
576 | 418 :since: version 0.4.3 |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
419 """ |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
420 decls = [] |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
421 text = self._strip_css_comments(self._replace_unicode_escapes(text)) |
856 | 422 for decl in text.split(';'): |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
423 decl = decl.strip() |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
424 if not decl: |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
425 continue |
840
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
426 try: |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
427 propname, value = decl.split(':', 1) |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
428 except ValueError: |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
429 continue |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
430 if not self.is_safe_css(propname.strip().lower(), value.strip()): |
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
431 continue |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
432 is_evil = False |
840
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
433 if 'expression' in value: |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
434 is_evil = True |
840
878306a5b465
Ported some of the HTML sanitization improvements from Trac (see [T7658]).
cmlenz
parents:
837
diff
changeset
|
435 for match in re.finditer(r'url\s*\(([^)]+)', value): |
571
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
436 if not self.is_safe_uri(match.group(1)): |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
437 is_evil = True |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
438 break |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
439 if not is_evil: |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
440 decls.append(decl.strip()) |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
441 return decls |
5815ad5f75a4
* Cleaned up the implementation of the `HTMLSanitizer`.
cmlenz
parents:
556
diff
changeset
|
442 |
431
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
443 _NORMALIZE_NEWLINES = re.compile(r'\r\n').sub |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
444 _UNICODE_ESCAPE = re.compile(r'\\([0-9a-fA-F]{1,6})\s?').sub |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
445 |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
446 def _replace_unicode_escapes(self, text): |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
447 def _repl(match): |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
448 return unichr(int(match.group(1), 16)) |
747baa1cd597
* Don't allow `style` attributes by default in the `HTMLSanitizer`. Closes #97.
cmlenz
parents:
425
diff
changeset
|
449 return self._UNICODE_ESCAPE(_repl, self._NORMALIZE_NEWLINES('\n', text)) |
556
d5cb5c200045
The HTML sanitizer now strips any CSS comments in style attributes, which could previously be used to hide malicious property values.
cmlenz
parents:
471
diff
changeset
|
450 |
d5cb5c200045
The HTML sanitizer now strips any CSS comments in style attributes, which could previously be used to hide malicious property values.
cmlenz
parents:
471
diff
changeset
|
451 _CSS_COMMENTS = re.compile(r'/\*.*?\*/').sub |
d5cb5c200045
The HTML sanitizer now strips any CSS comments in style attributes, which could previously be used to hide malicious property values.
cmlenz
parents:
471
diff
changeset
|
452 |
d5cb5c200045
The HTML sanitizer now strips any CSS comments in style attributes, which could previously be used to hide malicious property values.
cmlenz
parents:
471
diff
changeset
|
453 def _strip_css_comments(self, text): |
d5cb5c200045
The HTML sanitizer now strips any CSS comments in style attributes, which could previously be used to hide malicious property values.
cmlenz
parents:
471
diff
changeset
|
454 return self._CSS_COMMENTS('', text) |