15:59:16 <RRSAgent> RRSAgent has joined #htmlt
15:59:16 <RRSAgent> logging to http://www.w3.org/2011/11/29-htmlt-irc
16:00:49 <krisk> Meeting: HTML5 Testing Task Force
16:00:57 <krisk> Agenda: http://lists.w3.org/Archives/Public/public-html-testsuite/2011Nov/0021.html
16:01:13 <krisk> I'll wait a few more minutes in case someone else shows up....
16:01:42 <krisk> If you want to talk on the phone - speak up...else we'll do the meeting on IRC
16:01:56 <plh> plh has joined #htmlt
16:06:18 <krisk> Let's get going...
16:08:48 <krisk> First agenda item - new bugs on approved tests
16:08:50 <krisk> http://www.w3.org/Bugs/Public/buglist.cgi?product=HTML%20WG&component=testsuite&resolution=---
16:11:43 <krisk> A few new bugs...
16:12:08 <krisk> One asking to test toDataURL("image/jpg")
16:12:41 <krisk> ...another to test http://html5.org/r/6635 
16:14:27 <krisk> They have not bueen assigned...if someone is intrested in fixing the bugs...
16:14:55 <krisk> Next agenda item - New Test Submissions
16:15:49 <krisk> Looking at Hg http://dvcs.w3.org/hg/html/shortlog/ a good amount of activity
16:17:25 <krisk> Looks like the idlharness.js is the solution to test webIDL without breaking testharness.js
16:19:27 <jgraham> Well not quite
16:19:29 <krisk> Which is used by some preliminary html webidl tests http://www.w3c-test.org/html/tests/submission/AryehGregor/interfaces.html
16:19:54 <jgraham> The main idea is that by pasting in the IDL from the spec you will get a bunch of simple tests generated
16:20:30 <jgraham> leaving humans to think of more complex things
16:21:00 <krisk> Fine direction
16:21:56 <krisk> It looks like the idl is actually pasted into the test file?
16:23:31 <krisk> Seems like it would be better to have the the idl file be part of the spec deliverable
16:23:40 <krisk> ...then that could just be consumed by the harness
16:23:52 <jgraham> Yeah, it uses a web idl parser to read the IDL
16:25:29 <krisk> Intresting...
16:26:12 <krisk> Seems like a better target would be to use this in webapps
16:26:16 <jgraham> Right, one could require authors to have the IDL in some seperate section but that has a few big downsides
16:26:31 <jgraham> Notably that it will get out of sync
16:26:40 <jgraham> Sure, using this in non-HTML specs is also a goal
16:27:03 <krisk> I would think given the whole purpose of webidl
16:28:00 <jgraham> Yeah
16:28:07 <krisk> thanks for addressing the type - NOT_SUPPOTED_ERR: 9 in testharness.js
16:28:15 <jgraham> In theory anywhere WebIDL is used this would be good
16:28:17 <jgraham> No problem
16:31:30 <krisk> Seems like testing the webidl parser is a key part...
16:31:50 <krisk> ...else we'll end up with a bunch of issues later
16:32:51 <krisk> have you looked at the actual tests?
16:32:57 <jgraham> Yes, that is true of course
16:33:10 <jgraham> Not much yet. Aryeh has done all the work so far
16:35:42 <krisk> Looks like he pulled the idl from the whatwg and not the w3c spec..
16:36:01 <krisk> Though intresting is the events all seem to fail across all browsers
16:36:58 <krisk> ..even an old event like onload
16:37:34 <jgraham> Could be that webidl specifies something that doesn't actually work in browsers
16:38:05 <plh> if you guys are talking about Canvas2D, I'd suggest to hold before making decision on tests that might need to be reverted in a few weeks. the html chairs are on the issue of canvas2d
16:38:18 <krisk> OK
16:39:04 <plh> the fact is that the w3c canvas2d spec hasn't been updated on our side is a problem that the chairs need to address with the editor
16:39:42 <krisk> We looking at the recent submissions to Hg
16:40:56 <krisk> The most intresting is the idlharness.js (http://www.w3c-test.org/resources/idlharness.js)
16:41:31 <krisk> * thank goodness for script debuggers....
16:43:31 <krisk> In case you are catching up...
16:43:56 <krisk> The idea is to copy/paste the idl from a spec into a script tag...
16:44:21 <krisk> Then you can parse the idl and confirm various simple checks...
16:44:35 <krisk> for example read-only, short, long, etc...
16:46:03 <krisk> I was expecting to see some more bidi tests...
16:47:02 <krisk> james have you had time to take a peek at some of the tests Microsoft submitted?
16:47:13 <jgraham> I am really sorry
16:47:25 <jgraham> But I have been really busy
16:47:31 <krisk> OK
16:47:39 <jgraham> If I haven't done it by next time yell at me or something
16:48:00 <krisk> I'll take a peek at some more tests as well
16:48:13 <krisk> ms2ger updated a few tests from some of my feedback
16:48:47 <plh> speaking of canvas2d, are the tests using testharness.js now?
16:48:58 <krisk> No one has that much spare time...
16:49:14 <plh> hum...
16:49:32 <krisk> Everyone agrees that we should do the work
16:49:48 <plh> I might be able to ask someone to help here
16:49:52 <krisk> ...though it doesn't seem to be a top priority
16:50:01 <jgraham> Everyone hopes Philip will do it :)
16:50:02 <plh> I'll look around
16:50:12 <krisk> ..presume that the tests works 'OK' as is
16:53:50 <krisk> My view is that I'd rather see people provide feedback on tests and submit new tests rather than update the canvas tests to use testharness.js
16:55:55 <krisk> James do you still need to make some ES5 changes to testharness.js?  or will the work that arey is doing satisfy this need?
16:59:29 <krisk> No need to answer right away...
16:59:34 <krisk> Just curious...
16:59:37 <jgraham> Changes to testharness.js will happen where necessary to support Aryeh's work (along with that of other test authors)
16:59:48 <krisk> OK
16:59:53 <krisk> shall we adjourn?
17:00:24 <jgraham> But Aryeh's work makes it less important that testharness.js expose a high-level API for writing tests that match WebIDL
17:00:35 <jgraham> Since he has provided a very-high-level API
17:02:15 <krisk> makes sense
17:02:38 <krisk> feel free to speak up, else we can adjourn
17:06:18 <krisk> Let's adjourn
17:06:40 <krisk> rrsagent, generate minutes
17:06:40 <RRSAgent> I have made the request to generate http://www.w3.org/2011/11/29-htmlt-minutes.html krisk
17:07:01 <krisk> rrsagent, make log public