- From: Simon Pieters <zcorpan@hotmail.com>
- Date: Wed, 19 Jul 2006 16:13:22 +0000
Hi, From: Ian Hickson <ian@hixie.ch> >On Mon, 17 Jul 2006, Simon Pieters wrote: > > As for an algorithm for how to do that, I think that an extra flag would > > be sufficient. If the parser hits <!-- while in RCDATA or CDATA, the > > flag is set to true. Then, if the parser hits --> the flag sets to > > false. Initially the flag is false. While the flag is true the element > > can't be closed. > >It's slightly more complicated than that due to the whole problem with >things like "<!--->", but yes. You're right. I forgot about that. I've added more test cases (008-014, and 003-004 in rcdata)[1]. Opera never treats <!--> as a standalone pseudo-comment. Firefox treats <!--> as a standalone pseudo-comment for script, but not for title and textarea. IE always treats <!--> as a standalone pseudo-comment. Safari treats <!--> as a standalone pseudo-comment for style and script, but not for noscript, noembed and noframes. Now, I think that <!--> should always be treated as a standalone pseudo-comment if <!--> will be treated as a standalone real comment (in PCDATA), otherwise never. (If pseudo-comments really are needed, that is.) Speaking of PCDATA <!--> comments, it came to me that some pages (such as [2] and [3]) might use <!--> in context of IE conditional comments, like so: <!--[if !IE]> <--> ... <!--> <![endif]--> ...which doesn't actually rely on <!--> being parsed as one comment. [1] http://simon.html5.org/test/html/parsing/pseudo-comments/ [2] http://damowmow.com/playground/demos/flash/001.html [3] http://juicystudio.com/ Regards, Simon Pieters
Received on Wednesday, 19 July 2006 09:13:22 UTC