annotate src/test/test_plugin_misc_text_syntaxes.py @ 2025:43516c915ca6

misc: updated Chteufleur contributions in README
author Goffi <goffi@goffi.org>
date Tue, 02 Aug 2016 23:21:13 +0200
parents 2daf7b4c6756
children 8b37a62336c3
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
1934
2daf7b4c6756 use of /usr/bin/env instead of /usr/bin/python in shebang
Goffi <goffi@goffi.org>
parents: 1910
diff changeset
1 #!/usr/bin/env python2
694
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
2 # -*- coding: utf-8 -*-
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
3
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
4 # SAT: a jabber client
1766
d17772b0fe22 copyright update
Goffi <goffi@goffi.org>
parents: 1412
diff changeset
5 # Copyright (C) 2009-2016 Jérôme Poisson (goffi@goffi.org)
694
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
6
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
7 # This program is free software: you can redistribute it and/or modify
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
8 # it under the terms of the GNU Affero General Public License as published by
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
9 # the Free Software Foundation, either version 3 of the License, or
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
10 # (at your option) any later version.
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
11
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
12 # This program is distributed in the hope that it will be useful,
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
13 # but WITHOUT ANY WARRANTY; without even the implied warranty of
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
15 # GNU Affero General Public License for more details.
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
16
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
17 # You should have received a copy of the GNU Affero General Public License
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
18 # along with this program. If not, see <http://www.gnu.org/licenses/>.
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
19
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
20 """ Plugin text syntaxes tests """
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
21
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
22 from sat.test import helpers
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
23 from sat.plugins import plugin_misc_text_syntaxes
856
7ea7053dda88 test: skip the test for markdown syntax if the module is not installed
souliane <souliane@mailoo.org>
parents: 841
diff changeset
24 from twisted.trial.unittest import SkipTest
832
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
25 import re
694
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
26
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
27
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
28 class SanitisationTest(helpers.SatTestCase):
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
29
832
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
30 EVIL_HTML1 = """
694
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
31 <html>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
32 <head>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
33 <script type="text/javascript" src="evil-site"></script>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
34 <link rel="alternate" type="text/rss" src="evil-rss">
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
35 <style>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
36 body {background-image: url(javascript:do_evil)};
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
37 div {color: expression(evil)};
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
38 </style>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
39 </head>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
40 <body onload="evil_function()">
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
41 <!-- I am interpreted for EVIL! -->
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
42 <a href="javascript:evil_function()">a link</a>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
43 <a href="#" onclick="evil_function()">another link</a>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
44 <p onclick="evil_function()">a paragraph</p>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
45 <div style="display: none">secret EVIL!</div>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
46 <object> of EVIL! </object>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
47 <iframe src="evil-site"></iframe>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
48 <form action="evil-site">
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
49 Password: <input type="password" name="password">
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
50 </form>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
51 <blink>annoying EVIL!</blink>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
52 <a href="evil-site">spam spam SPAM!</a>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
53 <image src="evil!">
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
54 </body>
780
9810f22ba733 test: store the constants in constants.py + better PEP8 compliance
souliane <souliane@mailoo.org>
parents: 694
diff changeset
55 </html>""" # example from lxml: /usr/share/doc/python-lxml-doc/html/lxmlhtml.html#cleaning-up-html
694
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
56
832
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
57 EVIL_HTML2 = """<p style='display: None; test: blah; background: url(: alert()); color: blue;'>test <strong>retest</strong><br><span style="background-color: (alert('bouh')); titi; color: #cf2828; font-size: 3px; direction: !important; color: red; color: red !important; font-size: 100px !important; font-size: 100px ! important; font-size: 100%; font-size: 100ox; font-size: 100px; font-size: 100;;;; font-size: 100 %; color: 100 px 1.7em; color: rgba(0, 0, 0, 0.1); color: rgb(35,79,255); background-color: no-repeat; background-color: :alert(1); color: (alert('XSS')); color: (window.location='http://example.org/'); color: url(:window.location='http://example.org/'); "> toto </span></p>"""
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
58
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
59 def setUp(self):
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
60 self.host = helpers.FakeSAT()
1910
55440ee00905 test: update some tests
souliane <souliane@mailoo.org>
parents: 1812
diff changeset
61 reload(plugin_misc_text_syntaxes) # reload the plugin to avoid conflict error
832
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
62 self.text_syntaxes = plugin_misc_text_syntaxes.TextSyntaxes(self.host)
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
63
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
64 def test_xhtml_sanitise(self):
1412
979210da778a test: fix the tests
souliane <souliane@mailoo.org>
parents: 1396
diff changeset
65 expected = u"""<div>
694
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
66 <style>/* deleted */</style>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
67 <body>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
68 <a href="">a link</a>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
69 <a href="#">another link</a>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
70 <p>a paragraph</p>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
71 <div style="">secret EVIL!</div>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
72 of EVIL!
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
73 Password:
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
74 annoying EVIL!
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
75 <a href="evil-site">spam spam SPAM!</a>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
76 <img src="evil!">
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
77 </img></body>
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
78 </div>"""
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
79
1812
160b0d4c6520 plugin XEP-0071, XEP-0277: method clean_xhtml has been renamed to cleanXHTML
souliane <souliane@mailoo.org>
parents: 1809
diff changeset
80 d = self.text_syntaxes.cleanXHTML(self.EVIL_HTML1)
694
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
81 d.addCallback(self.assertEqualXML, expected, ignore_blank=True)
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
82 return d
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
83
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
84 def test_styles_sanitise(self):
1412
979210da778a test: fix the tests
souliane <souliane@mailoo.org>
parents: 1396
diff changeset
85 expected = u"""<p style="color: blue">test <strong>retest</strong><br/><span style="color: #cf2828; font-size: 3px; color: red; color: red !important; font-size: 100px !important; font-size: 100%; font-size: 100px; font-size: 100; font-size: 100 %; color: rgba(0, 0, 0, 0.1); color: rgb(35,79,255); background-color: no-repeat"> toto </span></p>"""
694
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
86
1812
160b0d4c6520 plugin XEP-0071, XEP-0277: method clean_xhtml has been renamed to cleanXHTML
souliane <souliane@mailoo.org>
parents: 1809
diff changeset
87 d = self.text_syntaxes.cleanXHTML(self.EVIL_HTML2)
694
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
88 d.addCallback(self.assertEqualXML, expected)
4284b6ad8aa3 tests: plugin text syntaxes sanitisation tests
Goffi <goffi@goffi.org>
parents:
diff changeset
89 return d
832
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
90
841
831f208b4ea3 plugin text_syntaxes: html2text was breaking the long URLs
souliane <souliane@mailoo.org>
parents: 832
diff changeset
91 def test_html2text(self):
831f208b4ea3 plugin text_syntaxes: html2text was breaking the long URLs
souliane <souliane@mailoo.org>
parents: 832
diff changeset
92 """Check that html2text is not inserting \n in the middle of that link.
831f208b4ea3 plugin text_syntaxes: html2text was breaking the long URLs
souliane <souliane@mailoo.org>
parents: 832
diff changeset
93 By default lines are truncated after the 79th characters."""
831f208b4ea3 plugin text_syntaxes: html2text was breaking the long URLs
souliane <souliane@mailoo.org>
parents: 832
diff changeset
94 source = "<img src=\"http://sat.goffi.org/static/images/screenshots/libervia/libervia_discussions.png\" alt=\"sat\"/>"
831f208b4ea3 plugin text_syntaxes: html2text was breaking the long URLs
souliane <souliane@mailoo.org>
parents: 832
diff changeset
95 expected = "![sat](http://sat.goffi.org/static/images/screenshots/libervia/libervia_discussions.png)"
856
7ea7053dda88 test: skip the test for markdown syntax if the module is not installed
souliane <souliane@mailoo.org>
parents: 841
diff changeset
96 try:
7ea7053dda88 test: skip the test for markdown syntax if the module is not installed
souliane <souliane@mailoo.org>
parents: 841
diff changeset
97 d = self.text_syntaxes.convert(source, self.text_syntaxes.SYNTAX_XHTML, self.text_syntaxes.SYNTAX_MARKDOWN)
7ea7053dda88 test: skip the test for markdown syntax if the module is not installed
souliane <souliane@mailoo.org>
parents: 841
diff changeset
98 except plugin_misc_text_syntaxes.UnknownSyntax:
7ea7053dda88 test: skip the test for markdown syntax if the module is not installed
souliane <souliane@mailoo.org>
parents: 841
diff changeset
99 raise SkipTest("Markdown syntax is not available.")
841
831f208b4ea3 plugin text_syntaxes: html2text was breaking the long URLs
souliane <souliane@mailoo.org>
parents: 832
diff changeset
100 d.addCallback(self.assertEqual, expected)
831f208b4ea3 plugin text_syntaxes: html2text was breaking the long URLs
souliane <souliane@mailoo.org>
parents: 832
diff changeset
101 return d
831f208b4ea3 plugin text_syntaxes: html2text was breaking the long URLs
souliane <souliane@mailoo.org>
parents: 832
diff changeset
102
832
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
103 def test_removeXHTMLMarkups(self):
1809
821c77574ad9 test, setup: requires lxml >= 3.1.0 (fixes html cleaning issue)
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
104 expected = u""" a link another link a paragraph secret EVIL! of EVIL! Password: annoying EVIL! spam spam SPAM! """
832
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
105 result = self.text_syntaxes._removeMarkups(self.EVIL_HTML1)
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
106 self.assertEqual(re.sub(r"\s+", " ", result).rstrip(), expected.rstrip())
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
107
1412
979210da778a test: fix the tests
souliane <souliane@mailoo.org>
parents: 1396
diff changeset
108 expected = u"""test retest toto"""
832
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
109 result = self.text_syntaxes._removeMarkups(self.EVIL_HTML2)
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
110 self.assertEqual(re.sub(r"\s+", " ", result).rstrip(), expected.rstrip())
c4b22aedb7d7 plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents: 811
diff changeset
111