[00:03:53] Is there a certain time without updates before an extension is listed as unmaintained? [00:04:19] nullspoon: no, there is no set duration that triggers the "unmaintained" designation [00:04:24] nullspoon: Probably "until it breaks with the latest MW" or similar [00:04:36] s/latest/latest stable/ [00:05:43] ah [00:05:44] okay [00:06:00] there's an extension I've been using but needed to tweak and noticed it hasn't been updated since mid 2011 [00:06:24] it also doesn't match the current extension architecture suggested in the extension tutorial [00:06:39] i was thinking i might hit up the original author for access to his git repo so I could work on it. [00:06:59] wasn't sure if I should start my own or talk to him first. [00:07:42] nullspoon: The beauty of git is, any work you do can be easily moved back to the original author :) [00:07:58] I'm learning that. [00:08:03] s/git/DVCS systems/, but really, we use git [00:08:26] man I'm glad I learned regex. otherwise I'd have no idea what you're saying. [00:08:26] :) [00:08:33] Agreed [00:08:37] I think you should attempt to do your work and simultaneously talk to the original author [00:08:42] okay [00:08:44] sounds good to me [00:08:59] I'm working on a blog post about using mediawiki for a blogging software and his extension is in there. [00:09:13] but I want to tweak it a bit becasue it doesn't currently support template variables [00:09:16] ie {{{1}}} [00:09:32] nullspoon: btw, you can get a category of your blog put onto http://en.planet.wikimedia.org/ [00:09:37] if you're talking about MediaWiki [00:10:23] interesting [00:10:32] oh yes. definitely talking about mediawiki [00:10:37] i'm an addict now [00:10:38] haha [00:11:35] nullspoon: There are definitely worse things to be addicted to :) [00:13:37] definitely [00:14:04] by the way [00:14:34] while I'm at it, do you guys have any pointers on how to get a friendlier recent changes page? [00:14:49] I've written a script that hits the database and outputs wikitext that I then include in a page [00:15:05] but that's the best I've been able to get so far. IT's a fine solution, but if there's another one that I'm missing, I'd certainly like to hear about it. [00:15:10] Is there a mediawiki template to link to code on gerrit/git ? Extensions either still talk of SVN or insert https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/DidYouKnow.git;a=blob;f=README [00:15:30] makes it hard to move off gerrit ;-) [01:31:00] spage [03:55:31] what did I miss? [03:56:05] * spagewmf back from evening's second MacBook Air hard hang in Ubuntu ;-( [03:59:03] BTW, useful template for your extension on mediawiki.org is {{WikimediaDownload|YOUREXTNAME|server=gerrit}} [04:53:18] RoanKattouw: https://toolserver.org/~krinkle/tmp/test-js-slice/ [04:53:52] RoanKattouw: http://integration.wmflabs.org/testswarm/job/5 [07:47:24] ghello [12:46:45] 12:46:30 [exec] [30;42m[2KOK (32 tests, 1089598 assertions) [12:46:53] 1M assertions for the dump testing system [13:09:41] that is a lot of assertions [13:25:58] New patchset: Hashar; "tweak Ext-Wikibase job to run most tests" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16020 [13:25:59] New patchset: Hashar; "ant: now output directory properties" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16021 [13:25:59] New patchset: Hashar; "make use take-fingerprints ant target" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16022 [13:26:00] New patchset: Hashar; "gitignore jobs prefixed with '-old' or '-back'" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16023 [13:26:48] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16023 [13:26:48] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16022 [13:26:49] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16021 [13:26:50] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16020 [14:29:04] New patchset: Hashar; "less.ly -> gerrit.wikimedia.org" [analytics/gerrit-stats] (master) - https://gerrit.wikimedia.org/r/16029 [14:31:38] New review: Diederik; "Ok." [analytics/gerrit-stats] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/16029 [14:31:40] Change merged: Diederik; [analytics/gerrit-stats] (master) - https://gerrit.wikimedia.org/r/16029 [14:31:57] oh diederik is floating around :-D [14:49:05] New patchset: Hashar; "make sure we delete any LocalSettings.php" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16031 [14:49:27] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16031 [15:00:55] apergos: anyone from features in the office? [15:01:19] I dunno, I'm not in the office :-D [15:01:23] I am sooooo not in the office... [15:04:36] good for you :) [15:04:50] well [15:05:24] http://www.freemaptools.com/tunnel-to-other-side-of-the-earth.htm according to this I'm halfway arouund the world from somewhere off of New Zealand, rather than from SF [15:05:29] but... close enugh :-D [15:05:50] it's a bit early for folks there I think, only 8am [15:14:59] Hey Reedy, got a min? [15:28:22] Reedy: did you manage to raise the AFTv5 lottery for beta yesterday? [15:29:36] hashar: didn't even try [15:29:41] csteipp: sure [15:29:52] Reedy: ok thanks :-] [15:30:36] Reedy: trying to get some info on 38082 [15:30:45] ApiFeedContributions [15:31:13] And why it reads as an anonymous user, instead of the logged in user... if that was intentional or not. [15:32:43] I think it's totally fair if we mark the bug wontfix, but it seems like it might be something we want, unless there's a reason to always do the read as anonymous. [15:36:25] For FeedContributons I pretty much just pulled the code from SpecialContributions https://www.mediawiki.org/wiki/Special:Code/MediaWiki/89566 [15:43:10] directly, neither are using pagesets.. [15:44:22] Yeah, I think I understand what it's doing... [15:47:14] So that puts in for the comment Linker::revComment, which uses $rev->getComment( Revision::FOR_THIS_USER) [15:47:58] This isn't exactly helped by us having 2 different revision deleting systems either [15:48:11] Yeah, no kidding :) [15:48:38] Seems ApiQueryRevisions has likely had the code for a long time [15:50:38] Yeah, that would be another way to do it-- just call userCan before attaching the text. [15:52:27] Alright, thanks for the info. I'll play around with adding in some checks for the user, and see how that goes. I'm guessing changes to the api would go through Roan? [15:54:36] hi csteipp - up early, eh? :-) [16:17:46] marktraceur: hi [16:17:54] drecodeam: Oh hi! :) [16:17:59] How's it going? [16:18:13] (excellent timing, by the way) [16:18:20] its going good! [16:18:31] why were you reviewing my patch or something ?? [16:18:58] drecodeam: I'd done some work in UW, I get emails on new patches, figured I'd look at it :) do you have questions about the review? [16:19:41] no not about a particular review [16:20:21] but I had to ask you that in my new patch I had submitted in the flickr branch, I have fixed the previous issues and coding convention problems [16:20:32] Oh, hm, let me see [16:20:44] Do I also need to submit patches for previous commits I had made earlier ? [16:20:58] Um, I don't think so, hold on [16:21:04] !g I76a5535f | drecodeam [16:21:04] drecodeam: https://gerrit.wikimedia.org/r/#q,I76a5535f,n,z [16:21:08] Is it that one? [16:21:58] If no, I'd appreciate a link to the right one (!g CHANGE_ID should do it) [16:22:01] https://gerrit.wikimedia.org/r/#/c/15547/ [16:22:21] Mmmkay [16:22:24] I'll look at it! [16:22:31] ok thanks. [16:22:47] drecodeam: Is this a _replacement_ patch? As in, you are no longer working on the patch I linked? [16:24:17] marktraceur: no, this is the new functionality I had added for uploading from Photosets, but in the process I have also made the code according to the conventions [16:25:02] Ah, OK [16:25:21] I'm on it! Will ping you in a few minutes with new review, I suspect [16:25:32] Preliminary: Much better. [16:27:16] sumanah: around? [16:27:29] yes! [16:27:40] matanya: I was planning on calling you in a few minutes [16:27:58] whenever you're ready [16:29:14] cool [16:42:47] Oh, drecodeam, I see what's happening [16:42:57] The commit I'm reviewing depends on the old commits? [16:43:19] ya it does [16:43:44] I see now. [16:43:49] Yeah, you're in quite a bind then [16:43:58] Can't update the old commits without merge conflicts [16:45:33] ya.Is this the right way of doing it ?? [16:53:19] drecodeam: I'm not sure. [16:53:27] It makes it difficult for me, to be sure [16:53:52] Because I distinctly remember thinking, "I'll check these commits again when the changes come, at a different level of detail" [16:54:32] marktraceur: In that case I should probably be fixing the previous changes before going onto a new one [16:55:04] *nod* [16:55:11] But I'm not sure the best way to do that [16:57:12] I had a tough time just resolving merge conflicts between two commits [16:57:19] You're sitting on several, IIRC [16:58:40] drecodeam: Had you tried changing the previous commits at all? I can imagine if you had, and found lots of conflicts, you might have done this eventually [17:01:45] marktraceur: no actually when I submit a new change, It asks me "You have more than one commit that you are about to submit. " [17:02:34] so when I submit the new change, It makes it dependent on the previous ones [17:05:45] Hm [17:05:51] I had to ask you this thing only, that since the changes required in the previous commits have been taken care of in this one, should I be changing the previous commits as well ?? [17:06:05] Uhhh [17:06:20] I'd say yes, but if it's too much trouble, I'm not sure it's necessary [17:06:48] There's been a lot of chat on wikitech about "every single commit should be perfect", but I don't think that's worth discussing for UW. [17:08:21] I guess it would not be that much of a problem, if there are not a lot of merge conflicts. [17:08:51] I would also look into why my changes always get dependent on the previous one in this branch. [17:12:13] drecodeam: Well, you're making commits all in a row [17:12:38] If you aren't modifying non-merged commits, you should do git checkout -b new/branch/name [17:12:53] That way, no troubles [17:15:03] matthiasmullie: do you a have a minute or two? [17:15:28] sure [17:15:55] marktraceur: can you explain that a little more [17:16:01] I'd like to show you a feature that was developed on hebrew wiki, and hear what you think [17:16:20] here is the link: https://he.wikipedia.org/wiki/%D7%95%D7%99%D7%9C%D7%99%D7%90%D7%9D_%D7%90%D7%91%D7%A0%D7%A1 [17:16:53] it's a chess board that plays the game described in the article [17:18:02] drecodeam: First, I'd need to know how well you're acquainted with git....do you understand generally how it works, how the branching model works, etc.? [17:18:18] bbl [17:18:42] marktraceur: not a lot, but fairly. [17:19:59] Mmmkay [17:20:43] drecodeam: Basically, gerrit looks at what you're uploading based on the current master branch [17:21:11] It says "I'm at commit hash nth9eu8923c9relu or whatever, and drecodeam is at ht9023peouh099oeu89goegu". [17:21:23] So then it looks at the parents of ht9023peouh099oeu89goegu [17:21:30] (or parent, if not a merge) [17:21:44] If the parent *isn't* nth9eu8923c9relu, there's trouble [17:21:48] lol - that's cool; can't stop watching it ^^ [17:22:21] But it keeps looking, since nth9eu8923c9relu should really be there, or at least, some known commit [17:22:50] Once it finds that known commit (in this case, only a few commits up), it says, "gee, there are three commits that are not currently being directly considered". [17:23:13] So then it gives you that error message, which is good. You should be aware of this. [17:23:47] If you say "do it now!", gerrit tries to recover by finding the new commits it doesn't know about in its database [17:24:03] And just links to them, basically [17:24:27] In order to avoid this, there are a few things you can do [17:25:03] The first, and most obvious, is not to submit review for commits that are based on unmerged commits. But that means not getting review for a long time, which is bad. [17:25:32] The second, and probably best in most cases, is to have separate branches, all based on the current remote's master branch, adding or modifying things. [17:25:45] That approach, however, won't work if you're basing work on previous commits [17:27:05] So whe you added a class to the JavaScript (flickrChooser or something IIRC), and you needed to add more commits that used that class, you couldn't have branched off for that [17:27:06] so matthiasmullie how hard would it be to move it crosswiki? [17:27:50] n00b Q: If I'm reviewing a gerrit patch and I want to propose a multiline improvement, ... ? Presumably there's a way to branch the review changeset, but http://www.mediawiki.org/wiki/Git/Tutorial#How_we_review_code doesn't say [17:27:50] ya thats what I was about to ask [17:29:01] marktraceur: in that case I guess the best way to do it would be by amending changes ? [17:29:21] drecodeam: If that's what you're doing, yes [17:29:46] If you're adding something new, or fixing a different bug, then dependencies are fine [17:29:49] If my new code is dependent on previously unmerged code, I should be ammending it ? [17:30:02] Not necessarily, I'd say [17:30:29] I had one more thing, say I have added the single image upload functionality, like in my case [17:31:10] If your commit message is "add flickrUploadChooser" and your new goal is to fix the way the API key is retrieved (but for some reason you need the flickrUploadChooser changes), you should probably make a new commit [17:31:55] marktraceur: which would be dependent on the previous one ? [17:32:37] Yep [17:32:59] Example: https://gerrit.wikimedia.org/r/#/c/14401/ [17:33:16] I added a tests directory, and a makefile, in a previous commit [17:33:41] But the previous commit was adding API tests, so it didn't make sense to amend it with tests for a class [17:35:21] Ya I get it, so in this case we are bound to phase the problem that if we get back to the previous commit with a new patch, we might have merge conflicts [17:35:39] Yup [17:35:51] I think in general, you've done the right thing [17:36:00] But it makes amending previous commits very difficult [17:36:35] ya I had a similar thing in mind, although you have made it crystal clear now :D [17:36:49] *nod* I'm glad [17:37:31] but since the latest commit takes care of all the previous problems, in the end it would not be a problem. I would try and committing to the previous ones, and lets hope it does not have a lot of merge conflicts [17:39:41] All right, sounds good [17:41:18] marktraceur: looking at your review now [18:55:25] brion: you there? [18:55:35] could you please look at this? https://gerrit.wikimedia.org/r/#/c/15991/ [18:55:52] you are the last person to agree [19:08:55] TrevorParscal: moment [19:09:28] excellent [19:09:32] thanks [19:11:23] ok added a +1 i agree w/ license change [19:13:45] thanks man [19:14:00] sure thing [19:17:04] i'll be in the office later, catch y'all... [21:33:16] running scap now [21:46:38] kaldari are you guys still deploying? [21:46:44] i need to make a config change [21:49:16] bsitu: got a sec? We found something strange in the user table that might be related to Moodbar, can I come and show you when you have a moment? [21:49:44] ori-l ^^ [21:51:02] no we're done [21:51:06] groovy thanks [21:51:14] well, actually... [21:51:20] o [21:51:20] I guess the scap isn't quite finished [21:51:40] or it might just be stuck [21:51:49] i wish scap took less than forever to run [21:52:05] I guess I'll give it a few more minutes [21:52:11] Wait wait [21:52:17] We have a deployment window still [21:52:19] Until 3pm [21:52:22] And I'm about to scap [21:52:35] kaldari ,awjr: ---^^ [21:52:36] this is why i was asking :) [21:52:45] RoanKattou before you scap [21:52:49] can i sneak in a config change? [21:52:50] ? [21:52:52] Sure [21:53:02] thanks it'll take me one minute [21:53:17] but i think kaldari's scap is still running [21:53:20] my scap has been running for 21 minutes [21:53:28] Hmm [21:53:45] sometimes it seems to take as long as 25 [21:53:49] What are you scapping? [21:53:54] Didn't your window end at 1? [21:53:56] WikiLove and PageTriage [21:54:11] maybe :) [21:54:27] we usually so E2 and E3 together, so I don't remember which is first [21:54:30] so=do [21:54:35] waiiit kaldari/ori are listed at 1-3 [21:54:42] oh i see [21:54:43] heh [21:54:50] that is some confusingness [21:54:54] I think I have 2 windows :) [21:55:05] but in this case I think Roan gets my 2nd window [21:55:22] Yeah I'm about to do Ori's deployment :) [21:55:30] Anyway, I'll wait for your scap to finish [21:55:51] well, I'll give it a couple more minutes and then Control-C it [21:55:58] awjr: Push your config change through Gerrit if you wnat [21:56:12] RoanKattouw I just made it on fenari and git-commit'd [21:56:22] OK [21:56:28] Please push it as well [21:56:34] yup done [21:56:52] RoanKattouw: so since scap is still running.. should i jsut let it go out with your scap? [21:57:35] yeah, you can probably ride Roan's scap [21:57:40] mine should be done any minute [21:57:58] is it preferable for me to make config chagnes in my own repo via gerrit rather than on fenari these days? [21:57:59] I'm giving it 20 more seconds [21:58:08] yes [21:58:43] OK, I control-Ced it [21:58:51] awjr: Yes [21:58:51] RoanKattouw: It's all you [21:58:58] Alright thanks [21:59:00] ok i will do that next time [21:59:35] RoanKattouw: Funny question for you....if I wanted my extension to load after WikiEditor, is there a good way to do that in JavaScript? I'm thinking of adding in an event to WikiEditor, but if there's already something you can think of, I'd rather not [21:59:57] Asynchronous question, I'm not in a hurry about it [22:00:11] marktraceur: Best way is to make it depend on ext.wikiEditor but the nasty consequence is that that makes wikiEditor load even if it ordinarily wouldn't [22:00:32] Right--especially since I'm not depending on it, I can fall back on the default editor [22:01:02] The problem is, I guess sometimes WikiEditor doesn't load until after my extension, so mine thinks it needs to fall back [22:01:15] Then its work gets overwritten when WE trashes the old editor bar [22:03:08] Aha, look for the config variables [22:03:16] That might do it [22:05:02] Winner! Thanks for the sounding board, RoanKattouw [22:05:12] RoanKattouw i've got to go afk for a min. can you do me a favor and ping me when that change makes it out? [22:05:48] awjr: Scap starting now [22:05:53] oh awesome [22:06:18] ps im gonna be in the office next week, likely weds - the following tues [22:07:53] Yay [22:08:12] That's like the first time we don't miss each other in three months [22:16:10] :D