IRC log of htmlt on 2012-04-10

Timestamps are in UTC.

15:02:29 [RRSAgent]
RRSAgent has joined #htmlt
15:02:29 [RRSAgent]
logging to http://www.w3.org/2012/04/10-htmlt-irc
15:03:42 [krisk]
Let's get going...
15:03:47 [krisk]
Agenda -> http://lists.w3.org/Archives/Public/public-html-testsuite/2012Apr/0004.html
15:04:05 [krisk]
Agenda Item #1 Check for new bugs on approved tests
15:04:07 [Ms2ger]
Oh, a meeting
15:04:42 [krisk]
Tinyurl to bugzilla http://tinyurl.com/6mvghxx
15:05:18 [krisk]
No new bugs
15:05:34 [krisk]
Last meeting we agreed that https://www.w3.org/Bugs/Public/show_bug.cgi?id=15925 was correct
15:06:59 [krisk]
Patch exists
15:07:10 [krisk]
..in the bug as well
15:07:55 [krisk]
I can apply this, unless Ms2Ger you want to update this since you have a patch going already
15:08:13 [Ms2ger]
Either way is fine
15:08:46 [plh]
plh has joined #htmlt
15:08:49 [krisk]
I wonder if 'Qiankun Miao' wants to participate more
15:09:00 [krisk]
since they created a patch...
15:13:01 [krisk]
Now we also have bugs 14240 (move to typed array), 14421 that are fixed in Ms2Ger patch
15:14:17 [krisk]
So let's just move to discuss the patch Ms2Ger sent to the list (http://lists.w3.org/Archives/Public/public-html-testsuite/2012Mar/0000.html)
15:14:50 [krisk]
jgraham/gsnedders did you look at the diff Ms2Ger sent to the list?
15:15:25 [krisk]
Here is my notes looking at the patch...
15:15:48 [krisk]
2d.imageData.get.source.outside.html looks fine
15:16:39 [krisk]
Ms2Ger unless the Opera folks comment feel free to update this test in the approved folder
15:17:11 [Ms2ger]
Will do
15:17:23 [krisk]
Then 4 tests were updated to match the spec change last December - moving to Uint8ClampedArray rather than CanvasPixelArray
15:18:05 [krisk]
2d.imageData.get.type.html, 2d.imageData.get.length.html, 2d.imageData.create1.type.html, 2d.imageData.create2.type.html
15:18:35 [krisk]
These match the spec, so unless someone objects Ms2Ger can update these as well
15:19:32 [krisk]
Though I do wonder when each of the browsers will move to type...
15:19:55 [krisk]
..since the performance gains may not actually work well on all CPU types
15:21:05 [krisk]
IMHO it should not impact compat with the web today that has canvas with the former type (canvaspixelarray)
15:21:50 [krisk]
Next update in the patch deals with truncating and not rounding e.g. 0.5 -> 0.0 2d.imageData.object.round.html
15:22:15 [krisk]
Which is correct per the spec, but not much interop currently
15:23:12 [krisk]
Ms2Ger, unless someone objects this test can also be updated as well
15:23:41 [krisk]
Next is 2d.drawImage.broken.html
15:23:56 [krisk]
I *think* this update is incorrect
15:24:12 [krisk]
Basically the test fills the canvas with green
15:24:34 [krisk]
then draws the image on the canvas, the image is red
15:25:09 [krisk]
then it checks that at 50,25 the pixle is green...
15:25:27 [krisk]
The test should check that the pixle is red
15:25:31 [krisk]
comments?
15:30:03 [krisk]
Ms2Ger do you follow?
15:30:33 [plh]
maybe because it's not the right image?
15:31:37 [krisk]
The image loads up fine (HTTP200) so I'm not sure why the canvas would not change to red
15:31:59 [Ms2ger]
Philip made that change
15:33:33 [krisk]
Ok then let's not update this test until we get more information
15:35:04 [krisk]
Next is some exception updates - more webidl than Canvas
15:35:11 [krisk]
2d.imageData.get.nonfinite.html and 2d.imageData.create2.nonfinite.html
15:35:29 [krisk]
These look fine and can be updated unless someone objects
15:36:40 [krisk]
Next is 2d.text.font.parse.invalid.html which maps to bugzilla #14356
15:38:38 [krisk]
Did you make this change Ms2Ger or did it come from Philip?
15:39:00 [Ms2ger]
That was mine, probably
15:39:01 [krisk]
Looking at other browsers for interop...
15:39:36 [krisk]
IE10 fails two assertions, Opera fails two assertions, FireFox fails two assertions and Chrome fails 5 assertions..
15:42:58 [krisk]
A most browsers fail this assertion...
15:42:59 [krisk]
•Failed assertion ctx.font === '20px serif' (got 10px default[string], expected 20px serif[string])
15:43:59 [krisk]
which if I understand the tests intent... set ctx.font = '20px serif'; then set it to a bous value ctx.font = '10px default';
15:44:32 [krisk]
Then attem to access the ctx.font and make sure that the bogus value was tossed and serif was kept
15:45:06 [krisk]
s/attem/attempt/
15:45:10 [Ms2ger]
Right
15:45:29 [Ms2ger]
Those are unrelated to the change, right?
15:45:48 [krisk]
I'm just looking at the test overall
15:46:12 [krisk]
IMHO the test looks correct, just no interop today in browsers...
15:46:23 [krisk]
...so feel free to update this test as well
15:47:06 [krisk]
2d.path.stroke.prune.arc.html and 2d.composite.globalAlpha.invalid.html look fine as well
15:48:06 [krisk]
Ms2Ger feel free to update these as well and then we can resolve bugzilla bug #14240, #14421 and #14356
15:48:19 [Ms2ger]
Great, thanks
15:50:09 [krisk]
Next Agenda Item 'New Test Submissions'
15:50:28 [krisk]
Microsoft submitted a few progress and range tests
15:50:41 [krisk]
http://dvcs.w3.org/hg/html/rev/68d420c0525e
15:51:32 [krisk]
Ms2Ger updated doc.open (http://dvcs.w3.org/hg/html/rev/b01ecaedc906)
15:54:11 [krisk]
Other Business
15:55:51 [krisk]
I'm going to put out a RfR for some of the Microsoft tests we submitted as part of our IE10 development cycle
15:56:14 [plh]
sounds good
15:56:35 [plh]
I noticed that our fonts directory doesn't contain a woff equivalent
15:56:57 [plh]
would be good to convert it at some point
15:56:59 [krisk]
For example History, AppCache, Sandbox, structuredclone, drag and drop...
15:57:26 [krisk]
If people think these tests are wrong, please respond to the list
15:57:54 [krisk]
Since these tests map to IE10s planned implementation
15:58:38 [krisk]
One other note - GIT project
15:59:11 [krisk]
http://lists.w3.org/Archives/Public/public-html-testsuite/2012Apr/0003.html
16:00:15 [krisk]
If it's a mirror that is not an issue
16:00:36 [krisk]
though it can't become a RW and have tests submitted from GIT -> Hg
16:01:13 [plh]
so, originally, I believe it was created to facilitate submissions
16:01:19 [krisk]
Shall we adjourn?
16:01:46 [plh]
so, if someone wants to submit the tests into the git instead of the mercurial, we should be able to accommodate
16:02:15 [plh]
it just means that we need to be aware of them and replicate them into the mercurial
16:02:44 [plh]
but I agree that mercurial is still the preferred mode
16:03:21 [krisk]
Their are legal issues having people look at source code on a server that is not part of the w3c
16:03:39 [plh]
what are those?
16:04:14 [plh]
as long as individuals agree for us to put the tests in the mercurial under the proper license, we should be fine
16:04:38 [plh]
I'
16:04:58 [plh]
m happy to take this discussion offline if you prefer but I'm curious to know more about the legal issues
16:05:29 [krisk]
When the tests are placed in mecurial and submitted to the w3c the test creator agrees to License the test
16:05:30 [krisk]
http://www.w3.org/html/wg/wiki/Testing/Submission/
16:05:37 [plh]
correct
16:05:56 [plh]
there are legal in people submitting tests into the git
16:06:00 [krisk]
None of this exists for tests submitted to GIT
16:06:12 [plh]
but that doesn't affect looking at least
16:06:39 [plh]
if someone put tests into the git, we'll need to ensure that we're covered indeed
16:07:00 [krisk]
With all the complexity we have already, I don't think enabling another source management server will help
16:07:30 [plh]
well, so far no one has been using it, but the original goal was to increase the number of submitted tests
16:08:01 [plh]
since we didn't receive any into the git, that's not an issue at the moment
16:08:29 [krisk]
Read-Only works fine, Read-Write seems like it has multiple issues...
16:08:56 [plh]
I guess we can dive into those when the time comes, if it ever comes
16:10:02 [krisk]
ok - let's just have agreement if we decided to move GIT from RO -> RW
16:10:31 [plh]
ok
16:10:45 [krisk]
Thanks for understanding my point of view
16:10:56 [krisk]
let's adjourn
16:11:02 [plh]
you're welcome
16:11:12 [plh]
have a good day
16:11:38 [krisk]
* krisk waves goodbye
16:11:40 [krisk]
RRSAgent, make logs public