On Thu, 2011-04-28 at 21:32 +0200, Chr. Rossmanith wrote: > Hi, > > my fix for bug 34666 has been too conservative. Now the test for > > HTML_UNKNOWNCONTROL_ON is removed and > > <title /> is treated correctly (c.f. bug 36390). I'm not sure whether > > there should be a test if the token is a "ONOFF" token?
So, the original html parser code we've inherited there is really hacky and I rather poke out my eyes than read more of its horrors, but lets take it as a given for the moment and work within its boundaries. I think, yes, we could do with an additional ONOFF token test in there. That would be... if (HTML_TOKEN_ONOFF & nToken) ... > > And I'm not sure if bug 36080 should be listed in the comment as well, but > > the > > strange behaviour is mentioned in the discussion of 36080. Sure, that makes sense. Add all the relevant bugids to help the next misfortunate trying to make sense of it. > Just to emphasize: I didn't push yet because I'd like to have the > questions answered first... So, with -if (aToken.Len() >= 1 && '/' == aToken.GetChar(aToken.Len()-1)) { +if (aToken.Len() >= 1 && '/' == aToken.GetChar(aToken.Len()-1) && (nRet & HTML_TOKEN_ONOFF)) { it looks sane to me. C. _______________________________________________ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice