After all these years, two separate Python runtime library bugs affected me in one day. This one caused the Feed Validator to blow up, but now has a workaround. Here’s a simple test case:
from xml.dom import minidom
minidom.parseString('<foo:bar xmlns:foo="http://example.com/\nboom"/>')
And in unrelated news, the SYO top100 planet saw it’s first non-wellformedness error this morning, from an unexpected direction. Take a look at Chris Pirillo’s page... elements with both HREFandhref attributes. The UFP lowercased the attribute names and voilà... duplicate attributes. This, too, has since been fixed.
Your links on that particular entry have both HREF and href attributes, which to my knowledge doesn’t cause anybody any problems except for those who attempt to convert your page from HTML to XHTML... and I’m probably the only one in the world who is insane enough to attempt that, and I’ve already made a fix.
I’m not sure where the unit tests live. maybe somewhere in
www.hixie.ch/tests/
You won’t find HTML5 parsing tests there. The largest collections of HTML5 test cases are fantasai’s tests (slightly modified by me) and Anne van Kesteren’s WF 2.0 tests. Neither test suite tests the error recovery features of the HTML5 parsing algorithm.
If there are other test cases, please let me know.