15:58:25 RRSAgent has joined #htmlt
15:58:25 logging to http://www.w3.org/2012/12/18-htmlt-irc
15:59:04 Looks like a few people are on IRC already
15:59:27 mydck - looks like you are a new participant!
15:59:51 Do you want to share some information about yourself?
16:01:39 I'm currently an invited expert in the xquery wg.
16:02:16 in fact, there's a simultaneous xquery telcon, so my attention here will be intermittent)
16:02:28 That is fine, welcome Michael Dyck
16:03:00 plh has joined #htmlt
16:03:27 anyhow, lately i've become more interested in web platform specs
16:04:03 We'll wait a few more minutes of plh and darobin
16:04:20 I'm around
16:04:27 sorry for being late
16:04:38 OK - just waiting for darobin
16:05:25 odinho has joined #htmlt
16:06:10 OK let's start in 5 minutes
16:06:51 MikeSmith has joined #HTMLT
16:08:43 sorry, needed biobreak after previous meeting
16:09:04 plh, well done :)
16:10:34 Marcos has joined #HTMLT
16:10:44 That is fine
16:10:54 welcome Marcos!
16:11:03 hallo :)
16:11:08 OK let's get rolling!
16:12:08 Agenda: dvcs/cvs/git/github and branches
16:12:46 To make better progress, how about everyone list their requirements
16:13:37 I require a) git, b) something with a large community, c) something that supports pull requests :)
16:14:18 for the VC: ease of maintenance, good review tools, ease of getting contributions
16:14:30 What is VC: ?
16:14:39 version control system
16:14:40 version control?
16:14:50 should have used VCS I guess
16:15:20 Microsoft's requirements are
16:15:27 and btw, I believe we'd be better off with github, with a mirroring on dvcs
16:15:43 A) All contributors needs to agree to the following w3c test licences
16:15:44 http://www.w3.org/Consortium/Legal/2008/04-testsuite-license.html
16:15:51 http://www.w3.org/Consortium/Legal/2008/03-bsd-license.html
16:15:54 for branches, it seems to me that it's easier to adopt the same branches as the html spec itself
16:16:29 krisk: TS licensing requirements is not even something we can discuss, it's just something we have to stick to
16:16:33 for A) I would add: or use a license which is compatible with those two.
16:16:42 right
16:17:11 I strongly doubt that we could ever use anything incompatible with that (or that we would want to)
16:17:25 people submitting tests can't not be anonymous
16:17:54 I presume s/not//
16:18:15 Correct
16:18:24 but they can be pseudonymous imho
16:18:25 Oh?
16:18:27 in keeping with the practice we have for specs, I reckon we should accept pseudonymous so long as we have the right commitment
16:18:32 so we can keep Ms2ger
16:19:19 or we can rename Ms2ger into John Smith
16:19:22 :)
16:19:29 I'm not sure who Ms2Ger is though someone in the w3c knows who ms2ger is which is fine
16:19:45 ok
16:20:39 B) Microsoft will continue to make submits to dvcs.w3.org and cvs (for old stuff) for tests
16:21:01 C) Tests will still get proped to http://www.w3-test.org/
16:21:17 does that mean if we switch to github, Microsoft will stop contributing tests?
16:21:45 I agree with C btw
16:22:16 I think we're all agreed that tests will get propagated to w3c-test.org
16:22:21 D) Clear approved vs submitted/work in progress tests (how this is done is not a requirements,e.g. metadata, text file, folder structure)
16:22:33 ok
16:22:39 I don't understand the why of (B), it feels very unilateral
16:23:37 yeah, seems like more work for them too
16:24:05 Does Microsoft insist on just having a repository in w3.org space, or does it also require not having to push to github?
16:24:06 it's also not a requirement
16:24:09 I'm not saying you can't use another version control system you just have to figure how to get all the code to get moved/branched etc..
16:24:39 krisk: I thought the idea of consensus is that we work out the best solution, not have one of the participants dictate it?
16:24:47 ah, yes. I think we do want to maintain a copy on dvcs.w3.org
16:25:03 Like a mirror?
16:25:04 I'd like to here everyone requirements, so that we can make progress
16:25:12 odinho, yes
16:25:26 we certainly will maintain a copy if we use something outside of W3C, but it doesn't have to be a two-way replica (which is much harder to set up)
16:25:28 I
16:25:41 I've been told it's easy to set up on our side
16:26:22 krisk, does Microsoft require being able to *push* to w3.org besides having a copy there?
16:26:25 we can easily set things up so that we can keep an up to date copy of something in GitHub
16:26:44 it is much harder to have someone pushing to an incompatible system and keep that in sync with what everyone else is using
16:27:24 I'd also like to understand why Microsoft is requiring dvcs.w3 for testing but happy with GitHub for spec development? it's hard to grasp unsubstantiated requirements
16:27:38 especially when they're solutions and not requirements to start with
16:27:42 Yes Microsoft still requires to push to w3.org
16:28:03 Why, if I may ask?
16:28:48 I'm not saying the the VCS has to be the master branch
16:29:13 krisk: sure, but it's gonna get really messy
16:29:18 if we have two systems
16:29:24 github looks great today, but the w3c is something that will span many, many, many years
16:29:55 tmpsantos has joined #htmlt
16:30:04 it's going to be very painful to manage if we have two systems, it seems hard to do that for just one contributor
16:30:05 krisk, yes, we have heard that all before (and we all agree). But git is a distrubuted system, github is just one interface
16:30:17 How about we put up a read-only copy on w3.org, and when/if github shuts down, we can make the w3.org one writable?
16:30:23 Just like sourceforge.net was the cool place a few years ago..
16:30:37 krisk: what Ms2ger said.
16:31:04 what Ms2ger said
16:31:17 that's what we do for a bunch of other things already
16:31:31 it's why there's a D in DVCS in the first place
16:32:31 We'll this occur for all WGs?
16:32:43 a lot of WGs are using git already
16:32:47 including the HTML WG
16:32:53 in fact we're using GitHub
16:33:18 it's very important to understand here that with a DVCS like git there is *no* primary server
16:33:32 we just by convention agree on one, as participants
16:33:49 Having it as a requirement that everything on GitHub W3C is also pushed to W3C servers, so that they all can be enabled writable once GitHub becomes SourceForge, seems like a easy and nice way. :-)
16:34:26 if that one disappears, then shrug — anything else works the same
16:34:26 GitHub happens to be very convenient because it has tooling around that
16:35:25 indeed. in addition, W3C doesn't have the capacity to replicate all the services available on github
16:35:48 As long as we ensure we don't lose data if github suddenly dies, I don't really see the issue
16:36:11 Does w3c pay github to run this service?
16:36:32 we are recognized as a non profit organization by them
16:36:48 and get dedicated treatment for that
16:37:16 if we need more, we could certainly look into doing a contract with them
16:37:25 darobin_ has joined #htmlt
16:37:35 I've been on the phone with them in the past on that point
16:37:49 darobin has joined #htmlt
16:37:54 if they start selling our tests for advertizing purposes, we'll switch :)
16:38:12 nah, lets pocket the money! :)
16:39:05 I have made the request to generate http://www.w3.org/2012/12/18-htmlt-minutes.html plh
16:39:22 RRSAgent, make logs public
16:41:34 Ok then if you can't enable a two way sync and we can still agree to A) C) and D) then I can agree to having tests pushed to github
16:42:00 works for me
16:42:17 Sounds fine from here too.
16:42:32 Sounds good
16:42:46 sounds good to me!
16:42:59 (also agree with what darobin and plh wrote as requirements earlier)
16:45:47 so I guess this leaves us with the question on branches?
16:45:48 Does anyone else have any requirments?
16:46:32 darobin_ has joined #htmlt
16:47:08 ok, unfortunately I have to dash
16:47:21 please be sure to stick to my dahut/unicorn proposal for the branch naming :)
16:47:51 If not then we need to start to discuss how to get all the mechanical parts running to make this work
16:49:08 my thinking is to give an action item to Robin to create the repository, work with Mike on the proper replication, and start moving the tests around
16:49:45 plh can you start with getting a process in place that covers getting an account created once they agree to the licenses?
16:49:57 sure
16:52:13 How long will it take when one submits a test to have it appear on http://www.w3c-test.org/ ?
16:52:57 once the pull request is approved, that should be automatic
16:53:33 I think we'll look into deploying several branches on w3c-test
16:53:34 The web hooks run straight away, so should be as quick as today. Maybe a few seconds extra. But I would guess it'll be very well under half a minute.
16:53:57 (only technically from doing a similar thing earlier)
16:54:47 I just want to make sure that if someone submits a test it will appear on http://w3c-test.org/ with needing a review
16:55:13 So that people can test their tests before getting a test 'approved'
16:55:46 yes, we'll look at deploying the submitted tests as well
16:56:03 to facilitate the review
16:56:12 yeah this is all doable and I'm very happy to put time into getting it set up
16:56:47 OK let's do this...
16:57:58 Between now and Jan 8th (since the next two tuesdays are on x-mas and new years) someone from the w3c send a proposal to the list how this will work mechanically
16:58:10 submitted.w3c-test.org/ and w3c-test.org can track different branches.
16:58:25 might take us until January 14
16:58:33 since the tests will need to move around
16:58:39 AIUI, the idea was to have a branch for each submission
16:58:58 Ms2ger: Makes much sense.
16:59:00 Which sounds like it'd need some automatic merging
16:59:14 I think getting all the tests moved around and in the new directory structure will take a longer
16:59:32 I think we just need to make it clear how all of this is going to mechanically work.
16:59:34 I can spend some time on moving tests around
16:59:38 Ms2ger: Or just branchname.submitted.w3c-test.org/ :P Or something similar. I guess someone will think about it :]
16:59:48 Then once we understand how this will work we can go move all the tests around, etc..
17:00:48 ok
17:01:15 Happy to have this done sooner, just not sure what the w3c folks bandwidth is in the next two weeks.
17:01:41 Robin is away next week and I won't have a lot of bandwidth either
17:01:54 we'll see what can get done before next week
17:03:23 sounds good
17:05:25 I suspect people will need to think about how this will work and have feedback
17:05:33 Shall we adjorn?
17:06:17 Sure
17:06:29 :-) We never had the master naming talk :P
17:06:43 RRSAgent, make logs public
17:07:38 RRSAgent, generate minutes
17:07:38 I have made the request to generate http://www.w3.org/2012/12/18-htmlt-minutes.html Ms2ger
17:07:58 I'm sure that and how branches will be used and how the approval stuff will work will come up on the list
17:07:59 rrsagent, generate minutes
17:07:59 I have made the request to generate http://www.w3.org/2012/12/18-htmlt-minutes.html krisk
17:08:05 meeting adjorned!
17:08:09 rrsagent, generate minutes
17:08:09 I have made the request to generate http://www.w3.org/2012/12/18-htmlt-minutes.html krisk
17:09:15 Nice nice.
17:47:28 hober has joined #htmlt
18:09:38 plh has left #htmlt
18:12:09 Oh look, I entirely forgot
19:58:24 Ms2ger has joined #HTMLT
20:34:23 mdyck has left #htmlt