/irc-logs / w3c / #webapps / 2013-06-10 / end
Options:
- # Session Start: Mon Jun 10 00:00:00 2013
- # Session Ident: #webapps
- # [01:40] * Quits: smaug (~chatzilla@public.cloak) (Ping timeout: 180 seconds)
- # [01:57] * Joins: dom (dom@public.cloak)
- # [01:59] * Joins: schuki (~quassel@public.cloak)
- # [02:12] * Joins: chaals (~Adium@public.cloak)
- # [02:37] * Quits: hober (~ted@public.cloak) (Ping timeout: 180 seconds)
- # [02:38] * Joins: tobie (tobie@public.cloak)
- # [03:22] * Quits: marcosc (~marcosc@public.cloak) (Client closed connection)
- # [03:22] * Quits: tobie (tobie@public.cloak)
- # [03:54] * Quits: dom (dom@public.cloak) (Ping timeout: 180 seconds)
- # [04:11] * Joins: dom (dom@public.cloak)
- # [05:58] * Quits: chaals (~Adium@public.cloak) ("Leaving.")
- # [06:03] * Quits: dom (dom@public.cloak) (Ping timeout: 180 seconds)
- # [06:20] * Joins: chaals (~Adium@public.cloak)
- # [06:40] * Joins: dom (dom@public.cloak)
- # [07:02] * Quits: chaals (~Adium@public.cloak) ("Leaving.")
- # [07:20] * Joins: tobie (tobie@public.cloak)
- # [07:25] * Joins: chaals (~Adium@public.cloak)
- # [07:34] * Joins: richt (~richt@public.cloak)
- # [08:11] * Quits: tobie (tobie@public.cloak)
- # [08:37] * Quits: dom (dom@public.cloak) (Ping timeout: 180 seconds)
- # [08:41] * Quits: richt (~richt@public.cloak) (Client closed connection)
- # [08:41] * Joins: richt (~richt@public.cloak)
- # [08:46] * Joins: Ms2ger (~Ms2ger@public.cloak)
- # [08:49] * Quits: richt (~richt@public.cloak) (Ping timeout: 180 seconds)
- # [09:02] * Joins: dom (dom@public.cloak)
- # [09:26] * Quits: dom (dom@public.cloak) (Ping timeout: 180 seconds)
- # [09:29] * Joins: darobin (rberjon@public.cloak)
- # [09:37] * Joins: dom (dom@public.cloak)
- # [09:54] * Joins: richt (~richt@public.cloak)
- # [10:08] * Joins: richt_ (~richt@public.cloak)
- # [10:14] * Quits: richt (~richt@public.cloak) (Ping timeout: 180 seconds)
- # [10:20] * Joins: richt (~richt@public.cloak)
- # [10:26] * Quits: richt_ (~richt@public.cloak) (Ping timeout: 180 seconds)
- # [10:35] * Quits: Ms2ger (~Ms2ger@public.cloak) ("Leaving")
- # [10:35] * Joins: Ms2ger (~Ms2ger@public.cloak)
- # [10:50] * Quits: chaals (~Adium@public.cloak) ("Leaving.")
- # [10:50] * Joins: chaals (~Adium@public.cloak)
- # [10:51] * Quits: chaals (~Adium@public.cloak) ("Leaving.")
- # [10:51] * Quits: richt (~richt@public.cloak) (Client closed connection)
- # [10:52] * Joins: richt (~richt@public.cloak)
- # [10:56] * Quits: dom (dom@public.cloak) (Ping timeout: 180 seconds)
- # [10:59] * Quits: richt (~richt@public.cloak) (Ping timeout: 180 seconds)
- # [11:09] * Joins: smaug (~chatzilla@public.cloak)
- # [11:09] * Joins: chaals (~Adium@public.cloak)
- # [11:10] * Quits: chaals (~Adium@public.cloak) ("Leaving.")
- # [11:43] * Joins: marcosc (~marcosc@public.cloak)
- # [11:58] * Joins: abarsto (~abarsto@public.cloak)
- # [11:58] * abarsto is now known as ArtB
- # [12:32] * Joins: hallvors (~hallvord@public.cloak)
- # [12:42] * Joins: richt (~richt@public.cloak)
- # [12:43] * Joins: richt_ (~richt@public.cloak)
- # [12:50] * Quits: richt (~richt@public.cloak) (Ping timeout: 180 seconds)
- # [12:52] * Quits: marcosc (~marcosc@public.cloak) (Ping timeout: 180 seconds)
- # [12:54] * Quits: ArtB (~abarsto@public.cloak) ("Leaving.")
- # [13:32] * Joins: abarsto (~abarsto@public.cloak)
- # [13:32] * abarsto is now known as ArtB
- # [13:52] <hallvors> jgraham: I need some Git help again.. Around?
- # [13:52] <hallvors> (or maybe other Git gurus are around)
- # [13:53] <hallvors> saying git fetch gives me an error:
- # [13:53] <hallvors> fatal: refs/remotes/origin/pr/203 tracks both refs/heads/pr/203 and refs/pull/203/head
- # [13:53] <hallvors> Unexpected end of command stream
- # [13:54] <hallvors> And the reason may be found in this part of the .git/config file:
- # [13:54] <hallvors> [remote "origin"]
- # [13:54] <hallvors> fetch = +refs/heads/*:refs/remotes/origin/*
- # [13:54] <hallvors> url = https://github.com/w3c/web-platform-tests.git
- # [13:54] <hallvors> fetch = +refs/pull/*/head:refs/remotes/origin/pr/*
- # [13:54] <hallvors> Why are there two "fetch" lines for this entry?
- # [13:54] <darobin> hallvors: you've been checking out PRs, right?
- # [13:55] <hallvors> yes
- # [13:55] <darobin> in my experience it doesn't work all that well
- # [13:55] <darobin> at least, it seems to always create issues, especially if you touch the branches
- # [13:55] <hallvors> well..
- # [13:55] <darobin> that second fetch line in there is the one for PRs
- # [13:56] <hallvors> so the answer is basically "don't do this"?
- # [13:56] <darobin> if you yank it you'll likely make fetch happy again
- # [13:56] <darobin> well
- # [13:56] <darobin> the proper answer is: there may be a trick to doing this right
- # [13:56] <darobin> but I don't know it
- # [13:56] <hallvors> OK :)
- # [13:56] <darobin> whenever I've tried I've ended up with annoying errors
- # [13:56] <darobin> it also seems to interact poorly with git up
- # [13:57] <darobin> when I do when I need to tinker with PRs is that I do that in another local clone
- # [13:57] <darobin> hallvors: note that for the web-platform-tests repo, most if not all PRs should be from branches that are already in the repo anyway
- # [13:57] <darobin> given that, there's no need to pull out the PR branches
- # [13:58] * Ms2ger blames jgraham
- # [13:58] <hallvors> OK
- # [13:58] <Ms2ger> That said, I've never had such issues with hg ;)
- # [13:58] <darobin> it may be that jgraham knows how to do this right since I think critic does something like that, too
- # [13:58] <darobin> Ms2ger: can you even do that with hg? :)
- # [13:59] <Ms2ger> Well, I dunno what "that" actually is ;)
- # [14:00] <darobin> collaborating with other users?
- # [14:00] <Ms2ger> Hah
- # [14:00] * Ms2ger feeds darobin
- # [14:00] <hallvors> A more general question: should I rebase my local branch for PR 128 to get all that exciting action from last weekend in? Can I do that safely without confusing the pull request and Critic? Or is it safer to clone the repo again and create another pull request for any changes I want to add to the new tests?
- # [14:01] <hallvors> (that was three general questions actually)
- # [14:01] <darobin> hallvors: I haven't been looking at that PR, sorry
- # [14:02] <darobin> updating a PR can be done by pushing to the branch from whence it came
- # [14:02] * darobin looks
- # [14:02] <hallvors> darobin: no problem - some people seem to be active in the review :)
- # [14:02] <jgraham> Critic actually uses temporary branches, and I don't really know how to do this right. That said I've never had problems
- # [14:02] <hallvors> Critic says it's 29% done which sounds like it's getting somewhere
- # [14:02] <darobin> whoa that's a lot of commits
- # [14:02] <jgraham> Also, most PRs aren't in the repo already; they are mostly in other people's forks of the repo
- # [14:03] <darobin> really? lots of PRs have come from the repo itself (certainly not all though)
- # [14:03] <darobin> but 128 certainly comes from outside
- # [14:03] <darobin> that said, it's an outside that hallvors has write access to :)
- # [14:03] <jgraham> Well I don't know about "most", but after TestTWF I expect "most" and certainly going forward I expect "most"
- # [14:04] <darobin> hallvors: I'm not sure what your rebasing question is — do you want to rebase hallvors/XHRtestBugFixes2 with the current master?
- # [14:04] <darobin> if so, you can fetch the master upstream, rebase, and you should just be good to go
- # [14:04] <darobin> no need to much around with the PR branch
- # [14:04] <jgraham> If you want to rebase, you need to tell critic about it
- # [14:04] <darobin> jgraham: oh sure, going forward I would expect so
- # [14:05] <hallvors> Yes - so I should I say "git pull --rebase" or something and expect Git to bring in the new tests from this weekend?
- # [14:05] <darobin> hallvors: you mean the upstream ones?
- # [14:05] <hallvors> and jgraham: there is a "prepare rebase" button, that's what you meant?
- # [14:05] <jgraham> Generally it's not clear why you would need to rebase test branches
- # [14:05] * darobin doesn't know if hallvors has set things up to sync with upstream properly
- # [14:05] <jgraham> Because they typically won't conflict
- # [14:06] <darobin> you could indeed likely do a vanilla merge
- # [14:06] <hallvors> darobin: I mean the stuff merged into master right now
- # [14:06] <hallvors> jgraham: I'm going to change some of the new tests (slightly), more LINK tags and such
- # [14:06] <darobin> hallvors: but are you fetching changes from upstream? your fork needs to have those if you want to merge
- # [14:06] <hallvors> so I'm just wondering what the "right" or "best" workflow is
- # [14:07] * Ms2ger resists the urge
- # [14:07] <hallvors> darobin: AFAIK yes
- # [14:07] <Ms2ger> But I think the best workflow is making sure that your patches are reviewed and landed asap
- # [14:08] <hallvors> darobin: "git remote -v" includes origin https://github.com/w3c/web-platform-tests.git (fetch)
- # [14:09] <hallvors> Ms2ger: Critic doesn't allow me to review my own code for some reason :-p
- # [14:09] <darobin> hallvors: it labels that as "origin"? that's weird!
- # [14:09] <hallvors> but feel free to help :)
- # [14:09] <Ms2ger> I'm bad at xhr, sorry :)
- # [14:10] <jgraham> darobin: Well that's origin if that's where it's cloned from
- # [14:10] <hallvors> darobin: it's this: https://github.com/w3c/web-platform-tests/tree/hallvors/XHRtestBugFixes2
- # [14:11] <hallvors> (credit/blame to jgraham for helping me/making me make a branch there ;-))
- # [14:11] <jgraham> Oh, did I?
- # [14:12] <jgraham> Anyway, if you have a branch there, things are easier
- # [14:12] <jgraham> I suggest that you
- # [14:12] <jgraham> git fetch origin
- # [14:12] <jgraham> git checkout master
- # [14:12] <jgraham> git merge --ff-only origin/master
- # [14:13] <hallvors> (and that won't confuse Critic?)
- # [14:13] <jgraham> git rev-parse HEAD (and note the resulting SHA1 somewhere)
- # [14:13] <jgraham> Hmm, I wonder if this will work
- # [14:13] <jgraham> I was *going* to continue
- # [14:14] <jgraham> git checkout hallvors/XHRtestBugFixes2
- # [14:14] <jgraham> git rebase origin/master
- # [14:15] <hallvors> I think I might take Ms2ger's hints and make a new branch for the new stuff instead. So people can finish off that review and merge the big bugfix branch..
- # [14:15] <jgraham> go to critic, select "prepare rebase", "New Upstream / Move" and enter the SHA1 above
- # [14:15] <jgraham> Then force push to github
- # [14:16] <jgraham> But I actuallyu don't know what happens if you rebase a tracked branch
- # [14:17] <hallvors> never mind - I'll just checkout master, do "git pull" there, and then try to set up a new branch - and a new PR
- # [14:31] <darobin> jgraham: yeah, I know that's the origin if that's where it was cloned from, but from the PR I had the impression that hallvors had his own fork
- # [14:31] <darobin> which I'd then expect to be the origin
- # [14:32] <hallvors> I have a fork too
- # [14:32] <jgraham> Well if he cloned his for it would be. But I guess it would also work to clone the original and git remote add myfork
- # [14:32] <Ms2ger> I have a knife?
- # [14:32] * hallvors is configured to confuse darobin
- # [14:32] <Ms2ger> Oh, that kind of fork
- # [14:33] <darobin> hallvors: you and most of the world buddy
- # [14:33] * jgraham would probably actually call one w3c and one jgraham and not have an origin :p
- # [14:40] * Joins: skddc (~anonymous@public.cloak)
- # [14:49] <jgraham> hallvors: After conferring with jl, I have a solution for next time you want to rebase a review
- # [14:49] * Joins: richt (~richt@public.cloak)
- # [14:49] <jgraham> And by that I mean "I asked jl and he told me"
- # [14:55] * Quits: richt_ (~richt@public.cloak) (Ping timeout: 180 seconds)
- # [15:07] * Joins: davidb (~davidb@public.cloak)
- # [15:17] * Quits: smaug (~chatzilla@public.cloak) (Ping timeout: 180 seconds)
- # [15:27] * Joins: marcosc (~marcosc@public.cloak)
- # [15:31] * Joins: richt_ (~richt@public.cloak)
- # [15:31] * Joins: smaug (~chatzilla@public.cloak)
- # [15:37] * Quits: richt (~richt@public.cloak) (Ping timeout: 180 seconds)
- # [16:03] * Joins: davidb_ (~davidb@public.cloak)
- # [16:05] <hallvors> jgraham: did jl consider whether it was "a good idea" in general? :)
- # [16:08] <jgraham> What?
- # [16:08] <jgraham> hallvors: ^
- # [16:09] <hallvors> to rebase a review
- # [16:09] <hallvors> :)
- # [16:09] * Quits: davidb (~davidb@public.cloak) (Ping timeout: 180 seconds)
- # [16:09] * davidb_ is now known as davidb
- # [16:09] <jgraham> Oh well internally we do that all the time
- # [16:09] <jgraham> So it's not generally a problem
- # [16:10] <jgraham> But there is of course more for the author to do
- # [16:10] <jgraham> In general rebasing test reviews should be discouraged, because *in general* tests won't conflict
- # [16:11] <jgraham> In particular writing new tests won't ever conflict, unless two people write the same new tests
- # [16:11] <jgraham> So rebases are only worthwhile if someone changes a common resource, or two people bugfix the same test
- # [16:12] <jgraham> My suggested workflow is at https://github.com/w3c/testtwf-website/blob/gh-pages/github101.md#tips--tricks
- # [16:13] <jgraham> Note step 5
- # [16:13] <jgraham> (yes, I think this is the wrong place for this documentation)
- # [16:14] <jgraham> (and, as I always say, if github wasn't quite so merge-happy I would recommend that the final step be a rebase rather than a merge, so that the history of mainline would be linear and comprehensible)
- # [16:17] <Ms2ger> Mm, comprehensible history
- # [16:19] <darobin> jgraham: how did TestTWF go btw?
- # [16:19] <darobin> (if you were there)
- # [16:20] <jgraham> darobin: I was the kind of "there" which means "on my couch with a laptop at a time on Saturday morning when I would rather have still been asleep"
- # [16:20] <jgraham> But afaict it went pretty well
- # [16:20] <darobin> haha
- # [16:20] <darobin> good to hear
- # [16:20] <jgraham> At least, it was very effective at generating email
- # [16:20] <jgraham> and lots of that email seemed to be tests
- # [16:20] <darobin> that I noticed
- # [16:20] * Ms2ger is still getting email
- # [16:21] <jgraham> and many of those tests seemed to be useful
- # [16:21] <darobin> it was enough email that I mistakenly deleted all of my inbox
- # [16:21] * ArtB hopes tobie et al. is working on that fire hose problem …
- # [16:21] <jgraham> Yeah, I think I got 500ish new mails out of it
- # [16:24] <jgraham> Woah, what's this data-tested-assertions thing?
- # [16:24] <Ms2ger> *assertations
- # [16:24] <Ms2ger> Probably a plinssism
- # [16:25] * Quits: hallvors (~hallvord@public.cloak) (Ping timeout: 180 seconds)
- # [16:26] <jgraham> I hope no spec that uses it ever changes
- # [16:26] <jgraham> Because that will probably mean rewriting every single test
- # [16:27] <jgraham> s/rewriting/updating/
- # [16:35] <darobin> is that from a PR?
- # [16:35] <jgraham> I was looking at http://critic.hoppipolla.co.uk/r/163
- # [16:36] <darobin> mmmm
- # [16:36] <darobin> I wouldn't take that without knowing what it is
- # [16:36] <darobin> it doesn't match anything in the repo
- # [16:38] <Ms2ger> I would support killing all that overly specific stuff
- # [16:38] <Ms2ger> link rel=help, sure, but anything worse...
- # [16:38] <darobin> as you know, I'm in favour of the brutal termination of any form of in-test metadata :)
- # [16:39] <Ms2ger> Hmm
- # [16:39] <Ms2ger> You prefer out-of-test metadata? :)
- # [16:39] <darobin> I prefer vernacular metadata :)
- # [16:39] <darobin> the kind you add without knowing you're adding it
- # [16:40] <darobin> ideally any metadata should have an immediate effect on the test, if at all possible
- # [16:40] <darobin> when you don't see the effect of something you added to your code, you don't notice when it's wrong
- # [16:41] <Ms2ger> Somehow I've heard that argument before ;)
- # [16:42] * darobin shocked
- # [16:42] <Ms2ger> Surprising, eh
- # [16:43] <darobin> well
- # [16:44] <darobin> I'm right now taking some eye drops for a fundus thingie later this afternoon
- # [16:44] <darobin> one of the side effects is "might make you excitable and confused"
- # [16:44] <darobin> so if you bring metadata up again, don't be surprised if you hear the same view anew, except much more excitedly
- # [16:44] <darobin> and less clearly
- # [16:44] <darobin> and probably in a helium-voice of some sort
- # [16:46] <jgraham> They have eye drops that affect the local air density now?
- # [16:46] <jgraham> Awesome
- # [16:47] * Quits: smaug (~chatzilla@public.cloak) (Ping timeout: 180 seconds)
- # [16:47] <darobin> they sure feel like they do
- # [16:49] <jgraham> FWIW I don't want to review it
- # [16:49] <jgraham> Because I can't be bothered to check if it's right
- # [16:49] <jgraham> So that might count as an argument against having it
- # [16:50] <Ms2ger> You don't want to review eye drops?
- # [16:50] <jgraham> (the thing that Philip` had for canvas tests with regexp was a bit better)
- # [16:50] <jgraham> That too
- # [16:50] <Ms2ger> Would be nice if we could get the canvas test generation working again
- # [16:53] <darobin> Ms2ger: do you know what's broken? if you want you can open an issue on that and assign it to me
- # [16:53] * darobin hasn't looked at the problem, but can pick it up
- # [16:54] <Ms2ger> darobin, the script wasn't updated when the tests ended up in all different folders
- # [16:54] <darobin> the script that runs them?
- # [16:54] <Ms2ger> That generates them
- # [16:54] * darobin looks more closely
- # [16:55] <Ms2ger> The actual data is in the .yaml files
- # [16:55] <darobin> ah, gottit
- # [16:56] <darobin> that looks like it's mostly a question of fixing the paths in the script
- # [16:58] <darobin> there, took an issue on it
- # [17:00] <Ms2ger> ta
- # [17:12] * Joins: smaug (~chatzilla@public.cloak)
- # [17:20] * Quits: darobin (rberjon@public.cloak) (Client closed connection)
- # [17:22] * Joins: richt (~richt@public.cloak)
- # [17:29] * Quits: richt_ (~richt@public.cloak) (Ping timeout: 180 seconds)
- # [17:29] * Quits: richt (~richt@public.cloak) (Ping timeout: 180 seconds)
- # [17:39] <timeless> jgraham: hey, does http://critic.hoppipolla.co.uk/r/163 have to require credentials? or is it possible to design it so that readonly is available?
- # [17:48] <jgraham> timeless: It is possible to design it so that readonly is avaliable
- # [17:48] <jgraham> Indeed there is an open review issue somewhere on that
- # [17:49] <jgraham> http://critic-review.org/showcomment?chain=401
- # [18:16] * Joins: jsbell (~jsbell@public.cloak)
- # [18:38] * Quits: ArtB (~abarsto@public.cloak) ("Leaving.")
- # [19:00] * Joins: hober (~ted@public.cloak)
- # [19:01] <timeless> it'd be great if that was on :)
- # [20:05] * Quits: skddc (~anonymous@public.cloak) (Ping timeout: 180 seconds)
- # [20:06] <jgraham> timeless: Well not so much "on" as "implemented" :)
- # [20:06] <jgraham> But yes, I should make the time
- # [20:06] <timeless> :)
- # [20:09] * Joins: sicking (~sicking@public.cloak)
- # [20:14] * Joins: skddc (~anonymous@public.cloak)
- # [20:24] * Joins: abarsto (~abarsto@public.cloak)
- # [20:24] * abarsto is now known as ArtB
- # [20:49] * Joins: dgrogan (~dgrogan@public.cloak)
- # [20:51] * Joins: darobin (rberjon@public.cloak)
- # [21:00] * Quits: Ms2ger (~Ms2ger@public.cloak) ("nn")
- # [21:21] * Quits: skddc (~anonymous@public.cloak) (Ping timeout: 180 seconds)
- # [21:26] * Joins: skddc (~anonymous@public.cloak)
- # [22:04] * Quits: skddc (~anonymous@public.cloak) (skddc)
- # [22:36] * Quits: davidb (~davidb@public.cloak) (davidb)
- # [23:11] * Quits: Lachy (~Lachy@public.cloak) ("Computer has gone to sleep.")
- # [23:20] * Quits: sicking (~sicking@public.cloak) (sicking)
- # [23:21] * Joins: sicking (~sicking@public.cloak)
- # [23:29] * Joins: jeffh (~18047aad@public.cloak)
- # [23:36] * Joins: marcosc_ (~marcosc@public.cloak)
- # Session Close: Tue Jun 11 00:00:00 2013
The end :)