16:02:09 <RRSAgent> RRSAgent has joined #htmlt
16:02:09 <RRSAgent> logging to http://www.w3.org/2013/01/15-htmlt-irc
16:03:55 <krisk> Let's get going, if someone shows up we can revisit, etc...
16:04:06 <krisk> http://lists.w3.org/Archives/Public/public-html-testsuite/2013Jan/0002.html
16:04:17 <krisk> Was what I posted to the list
16:04:33 <krisk> Seems like a fine place to start given the recent holiday break
16:05:23 <Mike5> Mike5 has joined #htmlt
16:06:10 <krisk> Hello Mike5!
16:06:28 <krisk> Would you like to introduce yourself?
16:07:30 <plh> plh has joined #htmlt
16:07:36 <Mike5> I am a malicious bot sent to mine this channel for test cases to sell to competing standards bodies
16:08:04 <plh> do they know those test cases are free? :)
16:08:21 <Mike5> . /whois Mike5
16:08:24 <Mike5> Michael[tm] Smith <mike@w3.org>
16:08:35 <Mike5> plh: sshhh
16:08:50 <krisk> Hi Mike!
16:09:34 <Mike5> hey krisk
16:09:44 <Mike5> Happy New Year and all
16:10:14 <krisk> So before the holidays we agreed to have tests get submitted/edited in Git.
16:11:32 <krisk> I was wondering about the how the sync work is going between git and www.w3c-test.org?
16:11:32 <mdyck_> mdyck_ has joined #htmlt
16:12:54 <krisk> Mike was this on your radar as a task?
16:13:00 <Mike5> yes
16:13:31 <Mike5> I am working on it.. sorta. Thinking bout it at least
16:13:53 <Mike5> will try to do something concrete this week
16:14:34 <darobin> krisk: on that topic, we just had a long discussion with Mike5
16:14:42 <darobin> I think we have most of it mapped out
16:14:57 <krisk> Will this include having tests that are in a branch?
16:14:58 <Mike5> yeah, agreed
16:15:07 <Mike5> krisk, yes
16:15:18 <Mike5> hmm
16:15:24 <darobin> yes, once you have the content, exposing stuff that's in branches isn't too complicated
16:15:31 <darobin> (since everything's in a branch)
16:15:45 <krisk> I think having this would make it easier to go fix cases, since without this a person has to setup their own webserver
16:15:55 <darobin> we probably shouldn't expose any random branch, but we can have a simple config that lists the branches to use
16:16:15 <darobin> or we could in fact expose any random branch, it probably doesn't matter too much
16:16:48 <krisk> Maybe I missed something...but I assumed that as long as the branch started with submission/ it would be proped to www.w3c-test.org?
16:16:58 <krisk> Like what we have in place today?
16:17:04 <darobin> the primary difficulty is ensuring code review for executable content (like PHP that's needed for testing)
16:17:07 <krisk> ..or had?
16:17:12 <darobin> mmmmm
16:17:28 <darobin> we want to push master and CR to w3c-test.org for sure
16:17:39 <darobin> I don't recall discussion pushing submission/*
16:17:55 <darobin> but so long as we're doing branches, doing two or doing twelve isn't a big difference
16:18:09 <darobin> (I should think)
16:18:13 <krisk> Open to having one branch for submissions...
16:18:28 <darobin> well it's not one branch, each submission is a branch
16:18:31 <darobin> but that's all manageable
16:18:49 <krisk> In my experience people do need an actual webserver, or stuff gets quite messy.
16:19:12 <darobin> that's how http://www.w3.org/html/wg/drafts/ is set up, each of the links in there maps to a branch in the HTML WG spec repo
16:19:14 <krisk> Especially when you have to do tests that do any X-domain testing
16:19:54 <darobin> krisk: well, if you're going to develop tests before submitting them, then you're better off having a local web server irrespective of our setup
16:20:12 <darobin> because a number of subtle things can be different when you load from file: anyway
16:20:21 <darobin> (or in some cases not so subtle)
16:21:57 <krisk> At the test the web forward event in Paris this was quite confusing to people
16:22:54 <krisk> since they didn't push submissions to Hg people had to setup a web server which didn't work so well.
16:23:32 <darobin> mmmmm, I didn't notice that
16:24:00 <darobin> but in any case, we can certainly mirror submissions automatically (so long as they don't have PHP or the such, which needs review)
16:24:36 <darobin> it does however beat me how one could develop correct tests and submit them without having run them locally first
16:25:56 <darobin> Ms2ger: thanks for the review!
16:26:00 <Ms2ger> Np
16:26:12 <darobin> in related new, Ms2ger and I have processed 13 pull requests to the repo: https://github.com/w3c/html-testsuite/pulls?direction=desc&page=1&sort=created&state=closed
16:27:00 <darobin> there are still old submissions left to process though: https://github.com/w3c/html-testsuite/tree/master/old-tests/submission
16:27:17 <jgraham> Oooh, meeting
16:27:37 <darobin> if you are Google, Opera, Microsoft, or Infraware I would very much appreciate if you were to go through your old submissions and tell me which ones are still valid
16:27:48 <darobin> better still, if you could reformulate them as pull requests
16:28:03 <darobin> in exchange for that, you will get relatively speedy review
16:28:18 <darobin> (certainly relative to previous speed)
16:28:25 <krisk> I was hoping to look at the tests once they were on the w3c-test.org and see which ones still work or got busted in the move
16:29:48 <darobin> ok, fair enough
16:29:52 <jgraham> darobin: I believe all our old submissions are still valid, but I am snowed under with other work and I think zcorpan is on parental leave, so I doubt anyone will clean up our test submissions soon
16:30:04 <darobin> jgraham: ok
16:30:27 <darobin> I'll make my way through them progressively, but since it involves understanding where code I didn't write is supposed to go it's not always very fast
16:31:01 <jgraham> darobin: If you ask questions I will tey to answer them, but perhaps with slightly higher than normal latency
16:31:05 <jgraham> *try
16:31:32 <darobin> we have a bunch of stuff that's more or less busted, including broken links and non-automated tests that aren't flagged as such — I plan to write a linting tool to check the repo with
16:31:46 <darobin> jgraham: latency is fine; thanks for the offer
16:34:48 <krisk> darobin: do you have a text file for xls that has a list of what tests you know work and what have not been looked at or are broken due to the move?
16:35:37 <darobin> krisk: no, at this point whenever I've seen a broken test I've simply fixed it
16:35:58 <darobin> but for a list of those I haven't looked at, there's the whole content of old-tests/submission
16:36:30 <darobin> I have a script that goes through all the tests and fixes common link errors (like not linking to testharness properly)
16:36:54 <darobin> I will run it again soon, notably after I take the new canvas tests PR since I know it has broken links
16:39:56 <krisk> Ok
16:40:39 <krisk> The other question I had was that once a person asks for a pull request how do we know that they have agreed to the license?
16:41:27 <krisk> In the past the person had to give explicit permission (http://www.w3.org/2002/09/wbs/1/testgrants2-200409/)
16:42:02 <krisk> From http://www.w3.org/html/wg/wiki/Testing/Submission/
16:42:06 <darobin> krisk: so the case hasn't yet presented itself of someone making a PR whom we didn't already know had committed
16:42:30 <darobin> on the day that it does, I reckon that duty falls to the chair and/or W3C Team
16:42:51 <krisk> Can we not just put this on the git site?
16:43:04 <darobin> I think the way to do it is a simple matter of asking the person in the PR comments
16:43:09 <darobin> sure, is it not there?
16:43:35 <darobin> there's some text here https://github.com/w3c/html-testsuite/#contributing
16:43:57 <darobin> feel free to change it
16:44:10 <krisk> I see no link that has the person to fillout the W3C's license grant form on the w3c site
16:44:21 <darobin> (or send me better text that I will make the change with)
16:44:41 <darobin> you want a link to http://www.w3.org/2002/09/wbs/1/testgrants2-200409/ ?
16:45:17 <krisk> See the 'How to License Your Contribution' on the wiki http://www.w3.org/html/wg/wiki/Testing/Submission/
16:45:35 <krisk> You should be able to just use this text
16:45:50 <darobin> ok, gimme a sec
16:46:38 <krisk> You don't have to do right now :)
16:48:01 <darobin> krisk: updated
16:48:11 <darobin> heh, it's a do-ocracy man!
16:48:55 <krisk> That is all I had for today on the agenda
16:49:06 <krisk> any other agenda items?
16:50:00 <darobin> all good for me
16:50:27 <jgraham> Nothing here
16:53:07 <krisk> OK let's adjourn
16:53:18 <krisk>  RRSAgent, make logs public
16:53:23 <krisk> rrsagent, generate minutes
16:53:23 <RRSAgent> I have made the request to generate http://www.w3.org/2013/01/15-htmlt-minutes.html krisk
16:54:57 <darobin> thanks! cya all
16:57:17 <mdyck_> mdyck_ has left #htmlt