Consider this example:
<!-- watch this! --><foobar>
Dolor sit amet
</foobar>
I’m not sure what the expected output here is according to the spec, but I certainly would expect to see well-formed XML.
However, the actual output is not well-formed:
<!-- watch this! -->
<foobar>
<p>Dolor sit amet</foobar>
</p>
[ Edit: Same result with span
substituted for foobar
, but not with div
: This could be related to the block/inline classification distinction when parsing HTML “blocks” vs “tags”. ]
[ This is quite possibly a rather trivial bug in the implementation, but I haven’t investigated yet. ]