[webkit-gtk] XHTML files being processed as HTML files

[Michael Catanzaro]

Hi,

In bugs #152455, #152821, and #152823 I found XHTML files that
WebKitGTK+ is processing as HTML rather than XHTML, leading to various
problems.

I noticed nautilus is making the same mistake with the files, so I
suspect shared-mime-info.

Renaming the file extension from .xht to .xhtml causes both WebKitGTK+
and nautilus to treat the files as XHTML files. The Apple WebKit ports
treat both .xht and .xhtml as XHTML, and many of W3C's layout tests
expect this behavior, so surely WebKitGTK+ is wrong to treat them as
HTML.

Questions:

 - Does WebKitGTK+ rely on shared-mime-info to determine the MIME type
of the files it loads?
 - If so, can we simply teach shared-mime-info that .xht means .XHTML?
(If so, we need to add an appropriately-patched shared-mime-info to our
jhbuild environment.)
 - If not, where is the code that needs fixed?

Please, let's track this in bug #152455.

Michael