- From: olivier Thereaux <ot@w3.org>
- Date: Fri, 18 May 2007 09:58:33 +0900
- To: QA-dev Dev <public-qa-dev@w3.org>
- Cc: Shane McCarron <shane@aptest.com>, Yoshio Fukushige <fukushige.yoshio@jp.panasonic.com>
Hi Ville, On Apr 28, 2007, at 21:05 , Ville Skyttä wrote: > I'm running the HEAD validator locally (Fedora Core 6 x86_64, > OpenSP 1.5.2 > from Fedora Core, SGML::Parser::OpenSP 0.99 locally built), and I'm > getting > non SGML char errors where validator-test nor qa-dev's HEAD setup > shows them. > > Ideas where to look for the problem? I saw > http://www.w3.org/Bugs/Public/show_bug.cgi?id=3164 but I gather it's > supposedly already fixed in SPO 0.99, and I'm not sure if it's even > the same > issue. I think I found the solution. At first, we fixed the SGML declaration of XML, basically telling the validator "the characters 128-159 are OK in XML dialects". That was good, but actually, the issue remained for e.g HTML <= 4.01. And besides, the error was triggered for a number of unicode characters that have nothing to do with 129-159 control characters. Denmark, rotten, something. Yoshio's message earlier today, forwarded by Shane, prompted me to give a serious new look at our transcoding. I eventually found that when fixing http://www.w3.org/Bugs/Public/show_bug.cgi?id=4474 I completely broke our transcoding system, but (and that's what confused me for a while) only for some versions of the Encode library. The patch I applied a few minutes ago http://lists.w3.org/Archives/Public/www-validator-cvs/2007May/0064.html should hopefully clean up everything. Thanks all for the reports/info, it helped. Cheers, -- olivier
Received on Friday, 18 May 2007 00:58:38 UTC