Access Keys 01:24 http://lists.w3.org/Archives/Public/public-qa-dev/2005Mar/0013.html 01:25 was my proposal 01:25 1- Home 01:25 2- Skip Navigation 01:25 3- Documentation 01:25 4- Feedback 01:25 (0- Accesskeys) 01:25 anyone thinks good or ill of that? 01:26 accesskeys are definitely in then? 01:27 niq: you mean, as opposed to getting rid of them altogether? 01:27 + niq hates "skip navigation". I want to know where the link *does* go, not where it *doesn't* go. 01:27 yod, yes 01:28 but no strong opinions either way 01:28 well, in checklink, "skip navigation" is "go to the results", I guess 01:29 scop++ 01:29 my understanding was that this set was mostly harmless, and if we document it well, a pretty good thing 01:29 we could have it documented as "go to results" or "go to TOC" for documentation 01:30 provided we consistently have a toc for all documentation pages 01:30 a toc can be degenerate (one entry) 01:30 interface -> go to start of input fields", results -> "go to results", documentation -> "toc" 01:31 sounds fine ... 01:31 works for me too 01:31 although the first two can be incompatible if the results page have a revalidation fields, I think it's acceptable 01:32 any opinion on whether it makes sense to have an access key for access keys? 01:33 + yod thinks it's useless, but IANAExpert 01:33 + niq likewise 01:33 + scop shrugs 01:35 ok, let's take yod's 1-4 above (2 as clarified here), and drop 0 01:35 +1 01:35 we can still reconsider if necessary 01:36 I will take the ACTION to implement accesskeys 1-4 in check/wmvs 01:36 can do in wlc too, unless you have time for it, scop? 01:37 go ahead (not that it'd take too much time but...) 01:37 ok 01:37 thanks 01:37 ACTION: olivier to implement accesskeys in wmvs, wlc, document it 01:38 I will document it in wmvs docs space, we can then copy/link to that from wlc's Port Restriction 01:42 port restrictions: it's not used in production in CSS validator yet, I think Yves was not happy with the change, although I did not get him yet to voice his concerns clearly 01:43 he said there was a history/risk of use for DOS, I am not quite convinced, but we'll see 01:46 next item on my list is Custom DTDs (0.7.0 currently broken in this regard(?)) 01:47 I've fixed a number of small warnings/template bugs in the past week or so, and in the process noticed that 0.7.0dev does not work as 0.6.7 did for "non-obvious" document types 01:48 hmmm, someone's been DOSing Link Valet recently .... 01:49 i think anything that follows links is rather high risk of that 01:52 Any risk you have in that sense with a given port is also present on port 80 through CGI. 01:53 I agree, definitely, I don't think that has anything to do with port restriction 01:54 anyway, I'll followup with Yves Xhtml+MathML+SVG, and custom DTD validation 01:55 on the topic of "document types that used to be supported but don't seem to anymore", a couple examples: 01:55 http://qa-dev.w3.org/wmvs/HEAD/check?uri=http://www.w3.org/TR/XHTMLplusMathMLplusSVG/sample.xhtml 01:56 http://qa-dev.w3.org/wmvs/HEAD/check?uri=http%3A%2F%2Fwww.w3.org%2FStyle%2F 01:57 Nice namespace list... 01:57 very :) 01:57 Separate bug; the list isn't being pruned of duplicates. 01:58 what's wrong with the result for XHTMLplusMathMLplusSVG? 01:58 xover: yes, I've had that on my radar for a while 01:59 bjoern_: does not match http://validator.w3.org/check?uri=http://www.w3.org/TR/XHTMLplusMathMLplusSVG/sample.xhtml 01:59 and "This page is not Valid 1.1! 01:59 is weird, at best 02:00 + yod thinks there is something fishy with Doctype detection/parsing 02:00 but this is an area I am not familiar with yet 02:00 wondering if anyone had an idea, or wanted to give it a look 02:00 I've no idea what's going on there; will need to investigate further. 02:01 Do these two issues have BZ #s? 02:01 xover: not yet, I wanted to make sure this was not something I was stupidly overlooking, or a well-known issue 02:01 I will enter them, link them to the 0.7.0 metabug 02:01 Hmmm. Three, actually. 02:04 anyway, xover++, if that's not obvious/trivial, let's add to bugzilla and investigate 02:04 Well, I'd suspect that's a result of xover's 1.13 of xml.soc 02:04 marked as "may break some edgecases horribly" 02:05 http://lists.w3.org/Archives/Public/public-qa-dev/2005Feb/0028 02:05 this? 02:05 huh? 02:05 I don't think any changes were made wrt to 0028 02:06 Random Idea: Make "regressions/" subdir of "dev/tests/" and check in a test that triggers any given bug, named after the Bugzilla number for that bug. Possibly require it before closing a bug. 02:06 I found the diff, that's different, yes 02:07 PUBLIC "-//W3C//DTD XHTML 1.1 plus MathML 2.0 plus SVG 1.1//EN" "REC-MathML2-20031021/xhtml-math11-f.dtd" 02:07 that does not seem right... 02:08 scop? 02:09 http://dev.w3.org/cvsweb/validator/htdocs/sgml-lib/xml.soc.diff?r1=1.10&r2=1.11&f=h 02:10 + scop investigates, hold on... 02:11 bjoern_: what seems wrong about that? 02:12 xhtml-math11-f.dtd suggests it's for xhtml+math not xhtml+math+svg 02:12 It's not the X+M+S DTD, it's the X+M DTD. 02:12 http://qa-dev.w3.org/wmvs/HEAD/check?uri=http://www.w3.org/TR/XHTMLplusMathMLplusSVG/sample.xhtml 02:13 gives same result, but the test in /dev/tests should be fixed, indeed 02:14 ACTION olivier to fix X+M+S sample doc in /dev/tests 02:14 I remember reading the diffs while doing the above "cleanup", and dimly remember deciding xhtml-math-svg-flat-20020809.dtd is equivalent to xhtml-math11-f.dtd 02:14 xhtml-math11-f.dtd says: XHTML 1.1 plus MathML 2.0 plus SVG 1.1 DTD 02:15 xover: I think I like the idea of dev/tests/regression, though I suspect it will be a hassle to maintain 02:15 It may (hopefully) be more completely maintained than /dev/tests/ is, and fulfill the same purpose (roughly). 02:16 xhtml-math11-f.dtd has CR CR LF line endings here... 02:16 xover: fair enough 02:16 yep, I already diffed with -wbB... 02:18 oh, double CR, hmph. looks like "CR LF" here though. 02:19 yeah, that's due to LF => CR LF conversion on my system... 02:20 + xover does 'mac2unix xhtml-math11-f.dtd && cvs ci xhtml-math11-f.dtd'... 02:26 + scop is confused 02:26 + yod cvs updates on local installation, doesn't seem to see any change 02:28 xhtml-math11-f.dtd has fine CR LF line endings now... 02:31 bjoern_: ok, thanks for the confirmation 02:36 ideas? 02:37 I suggest we be zealously bureaucratic, file and move on if we don't find a cause quickly 02:37 ++ 02:38 scop: Look at the you mean for example 02:39 yup 02:40 crap. so, restore xhtml-math-svg-flat.dtd? 02:41 + bjoern_ -> off 02:41 That's probably the easiest way out, yeah. 02:41 by the way, what IS xhtml-math11-f.dtd, then? 02:42 X+M? 02:42 X+M IIRC. 02:42 http://lists.w3.org/Archives/Public/www-archive/2005Feb/0029.html can help to find out... 02:44 perhaps the MathML folks should be poked, suggesting they fix the description/comments in xhtml-math11-f.dtd 02:45 anyway, I'll resurrect the original flattened X+M+S DTD, let's move on... 02:46 most likely an old version of xhtml-math-svg-flat.dtd 02:47 confusing, rev history at end of xhtml-math11-f.dtd and the original suggests that the former supercedes the latter