15:02:29 RRSAgent has joined #htmlt
15:02:29 logging to http://www.w3.org/2011/04/19-htmlt-irc
15:02:45 zakim, this is htmlt
15:02:45 sorry, krisk, I do not see a conference named 'htmlt' in progress or scheduled at this time
15:02:50 zakim, this is #htmlt
15:02:50 sorry, krisk, I do not see a conference named '#htmlt' in progress or scheduled at this time
15:03:45 still no conf call...
15:04:05 I suspect that this will not be a problem given who is on IRC
15:04:58 Agenda: http://lists.w3.org/Archives/Public/public-html-testsuite/2011Apr/0010.html
15:05:28 Agenda Item #1 Check for new bugs
15:05:41 tiny url link http://tinyurl.com/3jgnk98
15:05:58 We have 3 bugs total..
15:06:29 jgraham are you going to do anything with bug #11321?
15:06:38 it looks like it could be closed
15:06:56 plh has joined #htmlt
15:07:27 No conf call is setup...
15:09:23 Ooh conf
15:09:55 plh if you dial/setup the call I can dial in as well...
15:10:03 though IRC works just fine
15:11:10 I don't think the essensial style of testharness.js is likely to change at this point
15:11:25 But I will change the component
15:11:31 so then you can resolve it as won't fix
15:13:33 I'm ok with irc
15:13:36 I have moved it to the testharness.js component. I will keep it open for now
15:14:13 are you to change the interface between the harness and the framework?
15:14:17 s//going/
15:14:41 The callback API?
15:14:45 yes
15:15:31 I imagine that has to be at least somewhat stable now. I have considered using something like postMessage but maybe that would have to be a better supplementary API
15:16:12 Are there reasons to change it?
15:16:30 (I am assuming you are asking "is it stable" but maybe not)
15:16:54 yes, I'm asking how stable it is
15:17:03 I don't know any reason to change it
15:17:07 I would not change it....
15:17:15 Well it is a real pain for me if I change it since I have things that depoend on it :)
15:17:15 it's an ask that is not strictly needed
15:17:40 So at the very least I don't expect the current API to disappear
15:17:54 good
15:18:26 If someone really wants this then they can add to the harness and not break the current API
15:19:10 let's move on to the second agenda item
15:19:33 Mike5 has joined #htmlt
15:19:46 It more of an FYI rather than an agenda item
15:19:55 Zakim, call Mike-goog
15:19:55 sorry, Mike5, I don't know what conference this is
15:19:59 oh
15:20:05 Note the WebApps thread about CFC for test approval (http://lists.w3.org/Archives/Public/public-webapps/2011AprJun/0219.html)
15:20:06 we're on irc only Mike
15:20:09 OK
15:20:29 I just went through the thread and didn't anything dramatic there
15:21:05 there is still the approved/submitted debate which, imho, depend on the group and the stage of the spec
15:21:12 Feedback ends soon...april 20th
15:21:35 We (Microsoft) will be sending some feedback
15:21:45 ok
15:21:54 Is Opera OK with this proposal?
15:22:01 I think Aryeh has a point about the organisation of the HTML repo not being optimal
15:22:20 Sure I think we are happy with the webapps proposal
15:22:35 I like the CFC for test approval
15:23:18 Ensures that everyone has a voice on approved tests
15:24:21 I'm ok switching to CFC, but asking the html wg to do CFC on test suite might be premature imho
15:24:59 I was thinking we could drive the group using CFCs for the test suite once the spec is in CR
15:25:50 FWIW what I think we really need is a code review system that would allow us to track which tests have actually been looked at
15:25:53 if we want to do CFC for the test suite before CR, I would suggest talking to the chairs first. we don't want to overburden the group too much between the call for proposals, reviews, etc.
15:26:22 I'm not suggesting that we switch to this for the HTMLWG
15:26:45 ]I am not yet really worried about how the end game pans out for the HTML testsuite because we are so far from that point
15:27:21 I think it is much more profitable to worry about how to get more tests
15:27:28 +high quality
15:27:48 Though having multiple people agree that a test is 'approved' is better than having just one person 'approving'
15:28:33 If you have not noticed ms2ger has added a number of more tests into Hg
15:28:35 see http://w3c-test.org/html/tests/submission/Ms2ger/
15:28:47 so, the code review system would be for each test?
15:28:55 Right, being able to track who reviewed each commit (or even each line in a commit) would be perfect
15:29:18 I have been looking for mercurial code review systems that do this, but have drawn a blank so far, sadly
15:29:46 So I guess we will have to improvise something
15:30:40 do we have a page to track the submitted tests for html somewhere?
15:31:06 I just run a script and watch submissions
15:31:38 Sounds like a good task for the cross testing working group
15:31:59 Since this is not WG specific and I have heard this ask from other WG (for example CSS)
15:32:09 we can look into that sure. it's more than just watching submissions however
15:32:16 it's also watching the review process
15:32:32 It's a simple scale problem
15:33:20 it would be helpful to have a list of requirements/needs from this group however
15:33:36 if we're going to develop something, better make sure it's something you guys want
15:33:39 sure
15:34:08 it goes under http://www.w3.org/wiki/TestInfra/goals#Test-case_management.2Ftracking
15:34:41 we haven't developed the requirements a lot for this so far
15:35:36 It is not trivial to work out what the requirments are
15:36:04 My ntural instinct is to favour something more like traditional code review
15:36:08 *natural
15:36:24 Where people review parts of specific commits
15:36:57 Rather than trying to encode some idea of a test in the system
15:37:31 we actually have a code review system in place between our current mercurial, to do reviews of php files
15:37:41 may be it can be extended...
15:37:46 how does that work
15:37:56 Interesting. Are you using something widely avaliable or something homegrown?
15:38:10 I think it's homegrown
15:38:28 it sends emails to a few team folks if someone commit a pho file
15:38:30 php
15:38:53 and the php file only gets replicated to w3c-test.org is one of us approves it
15:39:19 email is not good for tracking
15:39:23 just a url to click on, pretty easy
15:39:48 well, we could have a page maybe instead of emails
15:40:08 Yeah I worry that doing this "properly" is quite a lot of effort
15:40:36 I'm open to alternatives :)
15:41:13 Me too :)
15:41:42 btw, where is your code to run the parser within testharness? did you publish it somewhere?
15:41:50 s/parser/parser tests/
15:42:02 It's the the repo under submissions/Opera/
15:42:18 thanks!
15:42:32 It has some issues still
15:42:53 Plh you can subscribe to the RSS feed for the HTML Hg repository
15:42:55 Mike5_ has joined #htmlt
15:43:15 Kris, I am already :(
15:43:54 Let's move on to the last agenda item
15:45:10 I have not seen any feedback for the metadata chunk of the reflection tests http://w3c-test.org/html/tests/submission/AryehGregor/reflection/reflection-metadata.html
15:45:36 We said last time we would pick a date if we didn't get any feedback
15:45:57 so how about we pick May 17th as the date?
15:46:08 why so far in the future?
15:46:41 So that people have time to actually look at the test and give feedback
15:47:19 If not one responds back it seems fair given the time
15:47:26 I'm fine with that. I think reviewing one set of the tests basically means reviewing all of them due to the way they are structured
15:47:42 I'm ok with May 17th
15:48:33 It also means that Aryeh needs to stop changing the tests
15:49:02 (this is why working with *commits* is better than tests)
15:49:53 I flexible either way as long as the person is 'done' with a commit or test
15:50:06 and doesn't later need to do some type of 'cleanup
15:50:17 (well really I suppose one would have to ask for review of a series of commits)
15:50:42 (and then any later changes would be marked unreviewed and reviewed seperately)
15:51:57 Phl as a side note - the notes that get generated from the RRSAgent always have bad permissions
15:52:13 is their a trick to getting the permissions set?
15:52:14 RRSAgent, make minutes public
15:52:14 I'm logging. I don't understand 'make minutes public', gsnedders. Try /msg RRSAgent help
15:52:43 RRSAgent, make logs public
15:52:52 got it
15:53:36 Shall we adjourn?
15:54:03 fine by me
15:54:10 Sure
15:54:33 rrsagent, generate minutes
15:54:33 I have made the request to generate http://www.w3.org/2011/04/19-htmlt-minutes.html krisk
16:36:06 Mike5 has joined #htmlt
17:35:13 Zakim has left #htmlt
17:59:06 plh has left #htmlt