15:04:00 RRSAgent has joined #htmlt
15:04:00 logging to http://www.w3.org/2012/06/19-htmlt-irc
15:04:23 Meeting will be on IRC, unless someone wants to dial in...
15:05:06 Agenda: http://lists.w3.org/Archives/Public/public-html-testsuite/2012Jun/0008.html
15:05:44 Item #1 Bugs on approved tests
15:05:54 shorturl to bugzilla -> http://tinyurl.com/6mvghxx
15:06:44 One new bug -> https://www.w3.org/Bugs/Public/show_bug.cgi?id=17141
15:08:14 Actually this is not a new bug we discussed this back on 5/22/2012
15:08:21 http://www.w3.org/2012/05/22-htmlt-irc
15:08:56 not a new bug but ms2ger is proposed a fix and is asking for review
15:09:24 No one understands why this was changed in the spec
15:10:08 Back on 5/22 when we discussed it seemed odd since no browser actually threw in this case
15:10:34 Well, changed
15:10:41 More like "actually defined now"
15:10:42 is there a bug against 2d context then?
15:10:47 So for interop it would seem best to just have the html5 spec match all current browsers (FF, Opera, Chrome, IE)
15:10:58 because http://www.w3.org/TR/2dcontext/#dom-context-2d-createpattern is clear
15:11:03 I saw the change was just landed in Gecko, fwiw
15:11:22 so I would say that the test needs to follow the spec, or someone needs to raise a bug against the spec
15:12:42 I think a spec bug makes the most sense
15:13:30 You're welcome to file one, of course
15:14:21 OK
15:14:45 Note the change looks good https://dvcs.w3.org/hg/html/rev/4e230b0f13dd
15:15:54 Next Agenda Item New Test Submissions
15:16:20 Mathias submission -> http://w3c-test.org/html/tests/submission/MathiasBynens/named-character-references/
15:17:18 this seems a fine test to me. any reason not to approve it?
15:17:41 Fine by me,,,
15:18:01 Though one oddity is that he hinted that he found some bug
15:18:10 s/bug/bugs/
15:18:26 ..in his initial post
15:18:28 see http://mathias.html5.org/tests/html/named-character-references/
15:18:29 He filed some bugs on webkit that got fixed, IIRC
15:18:48 And he mentioned there was something weird with nbsp in IE
15:18:55 But when I run them they pass?
15:19:23 All the better
15:19:31 Anyhow unless someone objects we can approve them
15:20:33 +1
15:21:14 http://dvcs.w3.org/hg/html/shortlog has more new updates as well (beyond Mathias submission)
15:22:54 http://w3c-test.org/html/tests/submission/Ms2ger/the-elements-of-html/scripting/the-script-element/script-for-event.html
15:22:58 from ms2ger
15:23:39 http://w3c-test.org/html/tests/submission/Ms2ger/events/event-handler-javascript.html
15:24:23 http://w3c-test.org/html/tests/submission/Ms2ger/the-elements-of-html/forms/the-form-element/form-elements-matches.html
15:25:06 script/@event isn't the spec as far as I can tell. is that a new proposal?
15:25:36 It is
15:26:01 imho, until it gets in the spec, I don't think we can approve it
15:26:10 http://w3c-test.org/html/tests/submission/Ms2ger/the-elements-of-html/tabular-data/the-table-element/insertRow-method-01.html
15:26:14 Step 12 of the "prepare a script" algorithm
15:26:21 http://w3c-test.org/html/tests/submission/Ms2ger/the-elements-of-html/tabular-data/the-table-element/insertRow-method-02.html
15:28:29 for the event-handler-javascript, I wonder if the test reports a proper fail if the javascript handler isn't executed at all
15:28:51 Those insertRow ones have been there since forever
15:29:08 plh, times out
15:29:18 Looking at FF (event handler) I see 4 failures
15:29:33 IE passes them all
15:29:39 for-event, you mean? That's correct
15:29:47 Let me see how other browsers look
15:31:08 Chrome fail 3 cases - of course different ones that FF
15:32:54 Opera 12 also has a number of failures
15:33:56 Having browsers run script the same is a very key scenario
15:34:05 apologizes for asking but which part of the spec defines the behavior of break javascript; can't find it so far
15:35:12 break javascript?
15:35:42 http://w3c-test.org/html/tests/submission/Ms2ger/events/event-handler-javascript.html
15:36:29 http://es5.github.com/#x12.12
15:36:43 Nice test :)
15:37:17 I wonder if it's a ES5 test, a JSURL test, or an HTML test...
15:37:40 It's a HTML test mor or less
15:37:51 I'm fine with approving (http://w3c-test.org/html/tests/submission/Ms2ger/events/event-handler-javascript.html)
15:37:58 It tests that event handler attributes are treated as a function body
15:38:05 ok
15:38:10 It's not a JSURL test; there's no JSURL involved :)
15:38:10 (and not as a javascript url)
15:38:41 jgraham, thanks to bz :)
15:40:13 Yes, indeed
15:40:24 (very easy to get confused there; I have)
15:41:20 form-elements-matches.html, insertRow-method-01.html, insertRow-method-02.html all look fine as well
15:41:54 These all work fine in IE, FF, Opera, Chrome
15:44:13 Looking at the canvas changes...
15:45:55 http://w3c-test.org/html/tests/submission/PhilipTaylor/canvas/2d.gradient.interpolate.overlap.html
15:46:55 Gradient testing/verififcation is pretty tough
15:47:14 I suspec this is why the variance is +/- 16
15:48:51 IE fails - though it looks visually correct ( *i think*)
15:51:47 s/suspec/suspect
15:52:40 Seems like the test should have a bigger tollerance
15:53:36 Good to see all the browser render the same (from a human's eyeball)
15:54:45 Ms2Ger is their some visual interop bug with 2d.gradient.interpolate.overlap.html?
15:56:13 I dunno, did I touch that test?
16:00:34 Looking deeper in the log, it looks like you just added 2d.gradient.interpolate.zerosize.stroke.html , 2d.gradient.interpolate.zerosize.strokeRect.html, 2d.gradient.interpolate.zerosize.fillRect.html, 2d.gradient.interpolate.zerosize.fill.html
16:00:58 The rest of the changes we done to update the various index pages
16:02:28 Though I'm getting 404s when I try to load them up (e.g. http://w3c-test.org/html/tests/submission/PhilipTaylor/canvas/2d.gradient.interpolate.zerosize.fill.html)
16:03:29 Ms2Ger was this your intention to add these 4 new tests?
16:04:31 Not sure, I'd need to look what happened there
16:04:36 Need to run now, though
16:04:51 OK, let's adjourn the meeting
16:07:59 RRSAgent, make logs public