Mercurial > libervia-backend
view src/test/test_plugin_misc_text_syntaxes.py @ 2532:772447ec070f
jp: pubsub options refactoring:
There is now only "use_pubsub", and specification are set using "pubsub_flags" argument when instantiating CommandBase.
Options are more Python Zen compliant by using explicit arguments for item, draft, url instead of trying to guess with magic keyword and type detection.
Pubsub node and item are now always using respecively "-n" and "-i" even when required, this way shell history can be used to change command more easily, and it's globally less confusing for user.
if --pubsub-url is used, elements can be overwritten with individual option (e.g. change item id with --item).
New "use_draft" argument in CommandBase, to re-use current draft or open a file path as draft.
Item can now be specified when using a draft. If it already exists, its content will be added to current draft (with a separator), to avoid loosing data.
common.BaseEdit.getItemPath could be simplified thanks to those changes.
Pubsub URI handling has been moved to base.py.
author | Goffi <goffi@goffi.org> |
---|---|
date | Wed, 21 Mar 2018 19:13:22 +0100 |
parents | 0046283a285d |
children |
line wrap: on
line source
#!/usr/bin/env python2 # -*- coding: utf-8 -*- # SAT: a jabber client # Copyright (C) 2009-2018 Jérôme Poisson (goffi@goffi.org) # This program is free software: you can redistribute it and/or modify # it under the terms of the GNU Affero General Public License as published by # the Free Software Foundation, either version 3 of the License, or # (at your option) any later version. # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU Affero General Public License for more details. # You should have received a copy of the GNU Affero General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>. """ Plugin text syntaxes tests """ from sat.test import helpers from sat.plugins import plugin_misc_text_syntaxes from twisted.trial.unittest import SkipTest import re class SanitisationTest(helpers.SatTestCase): EVIL_HTML1 = """ <html> <head> <script type="text/javascript" src="evil-site"></script> <link rel="alternate" type="text/rss" src="evil-rss"> <style> body {background-image: url(javascript:do_evil)}; div {color: expression(evil)}; </style> </head> <body onload="evil_function()"> <!-- I am interpreted for EVIL! --> <a href="javascript:evil_function()">a link</a> <a href="#" onclick="evil_function()">another link</a> <p onclick="evil_function()">a paragraph</p> <div style="display: none">secret EVIL!</div> <object> of EVIL! </object> <iframe src="evil-site"></iframe> <form action="evil-site"> Password: <input type="password" name="password"> </form> <blink>annoying EVIL!</blink> <a href="evil-site">spam spam SPAM!</a> <image src="evil!"> </body> </html>""" # example from lxml: /usr/share/doc/python-lxml-doc/html/lxmlhtml.html#cleaning-up-html 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>""" def setUp(self): self.host = helpers.FakeSAT() reload(plugin_misc_text_syntaxes) # reload the plugin to avoid conflict error self.text_syntaxes = plugin_misc_text_syntaxes.TextSyntaxes(self.host) def test_xhtml_sanitise(self): expected = u"""<div> <style>/* deleted */</style> <body> <a href="">a link</a> <a href="#">another link</a> <p>a paragraph</p> <div style="">secret EVIL!</div> of EVIL! Password: annoying EVIL! <a href="evil-site">spam spam SPAM!</a> <img src="evil!"> </img></body> </div>""" d = self.text_syntaxes.cleanXHTML(self.EVIL_HTML1) d.addCallback(self.assertEqualXML, expected, ignore_blank=True) return d def test_styles_sanitise(self): 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>""" d = self.text_syntaxes.cleanXHTML(self.EVIL_HTML2) d.addCallback(self.assertEqualXML, expected) return d def test_html2text(self): """Check that html2text is not inserting \n in the middle of that link. By default lines are truncated after the 79th characters.""" source = "<img src=\"http://sat.goffi.org/static/images/screenshots/libervia/libervia_discussions.png\" alt=\"sat\"/>" expected = "" try: d = self.text_syntaxes.convert(source, self.text_syntaxes.SYNTAX_XHTML, self.text_syntaxes.SYNTAX_MARKDOWN) except plugin_misc_text_syntaxes.UnknownSyntax: raise SkipTest("Markdown syntax is not available.") d.addCallback(self.assertEqual, expected) return d def test_removeXHTMLMarkups(self): expected = u""" a link another link a paragraph secret EVIL! of EVIL! Password: annoying EVIL! spam spam SPAM! """ result = self.text_syntaxes._removeMarkups(self.EVIL_HTML1) self.assertEqual(re.sub(r"\s+", " ", result).rstrip(), expected.rstrip()) expected = u"""test retest toto""" result = self.text_syntaxes._removeMarkups(self.EVIL_HTML2) self.assertEqual(re.sub(r"\s+", " ", result).rstrip(), expected.rstrip())