[Lxml-checkins] r49019 - lxml/trunk/src/lxml/html/tests
scoder at codespeak.net
scoder at codespeak.net
Sat Nov 24 11:00:38 CET 2007
Author: scoder
Date: Sat Nov 24 11:00:38 2007
New Revision: 49019
Added:
lxml/trunk/src/lxml/html/tests/test_clean_embed.txt
Modified:
lxml/trunk/src/lxml/html/tests/test_clean.py
lxml/trunk/src/lxml/html/tests/test_clean.txt
Log:
split of 'clean embed' test case which fails in libxml2 2.6.29/30
Modified: lxml/trunk/src/lxml/html/tests/test_clean.py
==============================================================================
--- lxml/trunk/src/lxml/html/tests/test_clean.py (original)
+++ lxml/trunk/src/lxml/html/tests/test_clean.py Sat Nov 24 11:00:38 2007
@@ -1,7 +1,10 @@
import unittest
from lxml.tests.common_imports import doctest
+from lxml.etree import LIBXML_VERSION
def test_suite():
suite = unittest.TestSuite()
suite.addTests([doctest.DocFileSuite('test_clean.txt')])
+ if LIBXML_VERSION <= (2,6,28):
+ suite.addTests([doctest.DocFileSuite('test_clean_embed.txt')])
return suite
Modified: lxml/trunk/src/lxml/html/tests/test_clean.txt
==============================================================================
--- lxml/trunk/src/lxml/html/tests/test_clean.txt (original)
+++ lxml/trunk/src/lxml/html/tests/test_clean.txt Sat Nov 24 11:00:38 2007
@@ -117,29 +117,3 @@
<img src="evil!">
</body>
</html>
-
->>> doc_embed = '''<div>
-... <embed src="http://www.youtube.com/v/183tVH1CZpA" type="application/x-shockwave-flash"></embed>
-... <embed src="http://anothersite.com/v/another"></embed>
-... <script src="http://www.youtube.com/example.js"></script>
-... <script src="/something-else.js"></script>
-... </div>'''
->>> print tostring(fromstring(doc_embed))
-<div>
-<embed src="http://www.youtube.com/v/183tVH1CZpA" type="application/x-shockwave-flash"></embed>
-<embed src="http://anothersite.com/v/another"></embed>
-<script src="http://www.youtube.com/example.js"></script>
-<script src="/something-else.js"></script>
-</div>
->>> print Cleaner().clean_html(doc_embed)
-<div>
-</div>
->>> print Cleaner(host_whitelist=['www.youtube.com']).clean_html(doc_embed)
-<div>
-<embed src="http://www.youtube.com/v/183tVH1CZpA" type="application/x-shockwave-flash"></embed>
-</div>
->>> print Cleaner(host_whitelist=['www.youtube.com'], whitelist_tags=None).clean_html(doc_embed)
-<div>
-<embed src="http://www.youtube.com/v/183tVH1CZpA" type="application/x-shockwave-flash"></embed>
-<script src="http://www.youtube.com/example.js"></script>
-</div>
Added: lxml/trunk/src/lxml/html/tests/test_clean_embed.txt
==============================================================================
--- (empty file)
+++ lxml/trunk/src/lxml/html/tests/test_clean_embed.txt Sat Nov 24 11:00:38 2007
@@ -0,0 +1,32 @@
+THIS FAILS IN libxml2 2.6.29 AND 2.6.30 !!
+
+
+>>> from lxml.html import fromstring, tostring
+>>> from lxml.html.clean import clean, clean_html, Cleaner
+>>> from lxml.html import usedoctest
+
+>>> doc_embed = '''<div>
+... <embed src="http://www.youtube.com/v/183tVH1CZpA" type="application/x-shockwave-flash"></embed>
+... <embed src="http://anothersite.com/v/another"></embed>
+... <script src="http://www.youtube.com/example.js"></script>
+... <script src="/something-else.js"></script>
+... </div>'''
+>>> print tostring(fromstring(doc_embed))
+<div>
+<embed src="http://www.youtube.com/v/183tVH1CZpA" type="application/x-shockwave-flash"></embed>
+<embed src="http://anothersite.com/v/another"></embed>
+<script src="http://www.youtube.com/example.js"></script>
+<script src="/something-else.js"></script>
+</div>
+>>> print Cleaner().clean_html(doc_embed)
+<div>
+</div>
+>>> print Cleaner(host_whitelist=['www.youtube.com']).clean_html(doc_embed)
+<div>
+<embed src="http://www.youtube.com/v/183tVH1CZpA" type="application/x-shockwave-flash"></embed>
+</div>
+>>> print Cleaner(host_whitelist=['www.youtube.com'], whitelist_tags=None).clean_html(doc_embed)
+<div>
+<embed src="http://www.youtube.com/v/183tVH1CZpA" type="application/x-shockwave-flash"></embed>
+<script src="http://www.youtube.com/example.js"></script>
+</div>
More information about the lxml-checkins
mailing list