[01:24:02] Hey TimStarling / xyzram, I'm looking at OAIRepo, and it looks like the hook that was used to reindex pages on an article save was deprecated by ContentHandler? [01:24:22] So do we only update search now from a cron? [01:28:03] * csteipp starts running for the train [04:00:34] well, I don't know if it's due to lowered expectations or what, but I kinda like Gerrit 2.0 [04:02:46] you mean 2.6? [04:03:02] Yes; the proverbial 2.0 [04:08:47] If I click on my name in the top-right corner, a drop-down menu opens with my name also on it [04:09:05] But I think it's a lot faster [04:09:31] And the diffs seem more precise [04:19:06] i'm a new gerrit fan [04:20:09] a new-gerrit fan or a new gerrit-fan? [04:27:11] the former [04:27:52] actually I liked the onclick handler on the "git fetch" command box better when it threw a fatal error [04:28:21] it allowed me to copy out the ref with a single click and drag [04:28:55] since that's rarely the actual command I want to use [04:38:17] gitblit seems to be missing css [04:38:29] but is still a huge improvement over gitweb [04:47:23] binasher: how do you get to gitblit? [04:48:36] ori-l: plugins -> installed -> gitblit [04:49:17] of course there aren't direct links within changesets [04:49:36] Plugins? I might not have the permission bit. [04:51:10] ori-l: this direct link should work for you https://gerrit.wikimedia.org/r/plugins/gitblit/ [04:51:52] binasher: yup, that worked [04:57:41] fixed it in greasemonkey [05:12:38] TimStarling: share! [05:13:23] http://paste.tstarling.com/p/dVkNPW.html [05:13:55] maybe next time I write one I will try to be smarter about detecting page transitions [05:19:17] thanks [05:30:48] TimStarling: 'window.onpopstate' fires reliably for me on transitions [05:30:57] in Chromium, at least. [09:52:31] uh... gerrit wants me to log in again. can't remember which credentials it wants, havn't done that in months :) [09:52:41] which user name? my commit name? email? labls user? what? [09:52:45] labs user [09:52:49] it got upgraded [09:53:00] labs user and labs password? [09:53:08] doesn't seem to work for me... [09:53:34] I think it's "daniel"? [09:54:10] saper: that's my commit name, but the labs user is "Daniel Kinzler". [09:54:14] and it works now [09:54:20] had to use a space instead of _ [09:55:13] interesting [09:55:18] "daniel" does not work? [09:55:42] nope [09:55:43] ah [09:55:48] * saper bad [09:55:50] that was the first thing i tried [09:55:52] I have cn=uid=sn [09:55:53] yes. [09:56:04] so "saper" works for everything [09:56:22] welcome aboard, anyway :) [09:56:26] yea, i would have preferred that for labs console... [09:56:28] *shrug* [09:57:30] Yes, our gerrit checks for cn ... [10:06:47] maybe it can be swapped [10:06:49] :) [10:07:22] Should we? [10:07:52] I'll ask ^demon when he gets back online. [10:09:42] DanielK_WMDE: I think I already reported the underscore bug [10:09:51] or maybe I just complained without filing it? hmmm [10:10:08] qchris: for DanielK_WMDE only [10:10:20] :-) [10:10:38] why is it a bug? [10:10:47] maybe DanielK_WMDE can edit his own ldap record actually [10:11:02] "bug" as in "not how MediaWiki does it" :D [10:11:10] It is a bug that MediaWiki accepts underscores :) we have spaces in the database [10:11:37] reminds of the famous "Q" problem of IBM AS/4000 [10:11:56] Q is so special there that it even got ignored at the beginning of the password [10:12:07] you can login with QQPASSWORD [10:12:19] AS/400 [10:12:31] MW doesn't "accept" underscores since ages ago, it just normalises them to space [10:12:40] not that this changes your point :) [10:13:13] yeah [10:13:23] * saper never changes his opinion [11:37:34] New review: Krinkle; "Patch Set 1:" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/48450 [13:02:25] New review: Hashar; "Patch Set 1:" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/48450 [13:04:54] * hashar attempt the Chad invocation dance [13:05:04] """ Oh my god Gerrit is down """ [13:11:08] that worked :-] [13:11:58] :-)) [13:15:21] of course, invoking him at roughly 2pm is almost a guaranteed success [13:15:28] ^demon: grats on Gerrit upgrade. [13:15:45] <^demon> Thanks. [13:15:55] Same from me. Looks good! [13:15:58] ^demon: I got two minor issue with the new interface but they are easily solved by tweaking our CSS :-D [13:16:02] <^demon> Zuul had no problems reconnecting, didn't have to kick it. [13:16:03] now grant us the new "Abandon" right!!!!!!!!! ;-] [13:16:10] ah great [13:16:17] <^demon> One thing at a time. First, I'm going to finish breakfast :) [13:16:27] yeah that is important :-] [13:16:41] what would happen if someone edited the Change-Id of a commit msg? [13:16:52] when having my laptop opened at breakfast, I make sure IRC is not connected [13:17:05] Platonides: try it? :-D [13:17:18] I would expect a new change to be created [13:17:56] nope [13:18:06] it still uses the old change-id internally [13:18:18] see https://gerrit.wikimedia.org/r/#/q/Id5c05277f60bed8fd035eb6407da5063dbd06855,n,z [13:18:34] but appears as https://gerrit.wikimedia.org/r/#/q/Id5c05277f60bed8fd035eb6407da5063dbd06857,n,z [13:18:52] the first URL does not give me back any change [13:19:10] yep, but https://gerrit.wikimedia.org/r/#/c/48602/ has that change-id now [13:19:24] ahh [13:19:40] damn [13:19:44] the new diff is really nice [13:19:51] easier to compare between different patchsets [13:19:56] I wonder if it fixed the diff bugs [13:20:04] yeah so you edited the last bit :-D [13:20:28] right [13:21:51] now suppose that I edited that changeset locally [13:21:57] what will happen when pushing? [13:22:18] <^demon> INFO com.google.gerrit.common.ChangeHookRunner : hook[change-merged] exitValue:0 [13:22:24] <^demon> What a completely useless log entry. [13:22:34] <^demon> s/useless/pointless/ [13:22:34] it creates a new change: https://gerrit.wikimedia.org/r/#/c/48616/ [13:23:14] <^demon> New change id -> new change [13:23:40] but only when sending from local [13:23:48] not when changing in the commit msg in the gerrit ui [13:24:52] <^demon> That kind of makes sense, I guess. I could also see it being slightly confusing. [13:25:12] <^demon> If you edit the Change-Id from the UI, you've already got a change_id assigned. [13:25:21] <^demon> When pushing a Change-Id from local, you don't have a change_id yet. [13:25:57] Platonides: I have reverted the change :-D [13:26:00] but if the change id matches the one changed in the commit, it still creates a new one [13:26:21] ^demon: any idea where our Gerrit skin is ? [13:26:29] hashar, I was going to ask where was that button [13:26:36] then realised you did that manually :P [13:26:40] yeah [13:26:49] was looking for something like "revert patchset" or soemthing [13:27:05] I think it's not possible [13:27:17] <^demon> hashar: files/gerrit/skin/* and the theme settings in templates/gerrit/gerrit.config.erb [13:27:18] it is, simply not implemented :-] [13:27:34] ^demon: nice [13:27:46] ^demon: will review the bug opened, maybe open a few and submit css tweaks [13:28:08] what does erb stand for? [13:28:15] Danny_B: templates for ruby [13:28:23] !google define erb [13:28:23] https://www.google.com/search?q=define+erb [13:28:35] (Erbia) Erbium(III) oxide, a pink solid, is a compound of erbium sometimes used as a colouring for glasses and a dopant for optica [13:28:36] hmm [13:28:37] not that one [13:29:07] Extended Ruby Boilerplate ? [13:29:13] thx [13:29:20] I have no idea honestly [13:30:55] an asteroid? :) [[5621 Erb]] [13:31:05] <^demon> Just granted "Edit Topic" to all registered users on all repos. [13:31:38] <^demon> And abandon + remove reviewer to mediawiki/* [13:36:15] <^demon> qchris: So, priority #1 right now is fixing replication. [13:36:39] That is broken? [13:36:42] <^demon> We're getting a quice provisioning error. [13:36:52] <^demon> s/quice/guice/ [13:36:53] Did not notice that :-( [13:36:55] <^demon> http://p.defau.lt/?MPdqAjBhff9xbe7QHWS0ag [13:37:36] Ok. [13:37:53] I'll have a look. [13:38:05] <^demon> Yeah, I started last night but it was like 10:30p :) [13:46:30] for god sake [13:46:40] my coworking place has the worse internet connection ever [13:46:58] sympathies! [13:49:12] ^demon: Did you build the replication plugin against the deployed gerrit version? [13:49:21] (I get compilation errors) [13:49:34] <^demon> Let me see what it was built against, sec. [13:49:57] The problem is not implemented abstract versions. So that may warrant Guice injection errors [13:51:42] <^demon> Hmm, can't find it. But it was an old build. [13:53:02] So I'll work on updating the replication plugin [13:53:14] RequestContext.java also changed meanwhile... :-( [13:53:30] <^demon> https://integration.mediawiki.org/ci/job/gerrit-core/93/ is the build of core we're using. [13:53:55] OK. Thanks. [13:54:16] <^demon> So, shouldn't we deploy https://integration.mediawiki.org/ci/job/gerrit-plugin-replication/92/? [13:56:33] Not sure yet [13:56:40] It builds using 2.5 api [13:56:48] Our gerrit uses 2.6 api I guess [13:56:58] Jenkins says to use 2.4 plugin api? [13:57:08] Trying to make sense of that :-) [13:58:35] <^demon> ...maybe I'm not doing it right in jenkins. [13:58:49] But I am also running into problems locally. [13:59:04] So if it's Jonkins ... it's me and Jenkins doing something wrong :-) [13:59:57] I'll just try updating replication plugin to work against 2.6-SNAPSHOT. Would that work for you? [14:00:12] <^demon> Ah, that'd probably do it. Lemme test locally. [14:00:21] <^demon> Let's both test and see what happens :) [14:00:33] It's not just the pom. That'll be more work. [14:00:55] Gerrit changed a bit since replication hit v1.0 [14:03:35] <^demon> I cleared out my .m2/repository and am doing some fresh builds. I likely have so many versions installed in there that *any* plugin will work :p [14:20:42] ^demon: I've got a version that compiles against 2.6-SNAPSHOT. gerrit-dev does not sport a replication.config. Is there somewhere a replication.config that one can use for testing and that is similar to what we have in production? [14:21:02] <^demon> Not right now, no. [14:21:20] Ok, then I'll have to set it up on my own. [14:24:27] <^demon> It was just removing the second param to that constructor, right? [14:24:46] It was three things [14:24:46] <^demon> Oh, more than that. [14:25:28] Removing AccessPath.GIT, adding changeCache, and fixing dbProvider injection [14:26:45] <^demon> the plugins should probably get branched, if core's going to change in such a breaking way between releases. [14:28:26] Branching hell :-((( [14:28:58] I really hate when there has to be a separate branche for each version. But what can you do? [14:30:19] <^demon> Indeed, we have to do it for mw extensions. [14:30:37] <^demon> I'll clean up jenkins. Right now, it builds each plugin against what's in .m2/repository, which could contain old versions. [14:30:54] <^demon> I'll have it delete the installed gerrit libs before building plugins. [14:33:10] Gerrit got us a server temporarily unavailable :-D [14:33:26] hopefully it is just puppet restarting it [14:34:42] <^demon> Yeah. [14:36:56] ^demon: The updated replication plugin does replicate. [14:37:04] Shall I try to push upstream [14:37:12] Or just send you the jar for a quick fix [14:37:29] <^demon> Both. [14:37:33] Ok :-) [14:37:44] * qchris is lukci I ded not provide more options ;-) [14:37:55] s/lukci/lucky/ [14:38:33] <^demon> Ok, the gerrit-core job now rm's ~/.m2/repository/com/google/gerrit/* before building. [14:38:42] <^demon> Should get more reliable builds now. [14:45:15] ^demon, hi [14:49:13] ^demon, if I type 'A' into the add reviewer box here, I get Internal Server Error: https://gerrit.wikimedia.org/r/#/c/20922/ [14:49:21] <^demon> Yep, working on that. [14:49:53] I was testing https://bugzilla.wikimedia.org/show_bug.cgi?id=39333, but IIRC that bug just meant no results were returned, not an ISE... [14:49:54] okay [14:55:32] ^demon: pushed replication to upstream. Are you working on the "Add reviewer -> ISE" problem or what should I take a look at next? [14:58:13] <^demon> I think I've found it. In GroupMembers.getGroupMembers(), I think we just need to protect against groupCache.get() returning null. [14:58:25] Ok. Great. [14:58:38] So then I'll head back to the "OR search" problem? [14:58:53] <^demon> Sounds good. [14:59:00] <^demon> Oh btw, finished doing a replication start --all, worked great. [14:59:07] \o/ [14:59:07] <^demon> Everything went out. [15:02:48] ^demon: ok so you are really outstanding. [15:02:51] ^demon: Also pushed upstream: https://gerrit-review.googlesource.com/42260 :-D [15:03:10] ^demon: saves me from studying the google license agreement :-] [15:19:11] <^demon> https://gerrit.wikimedia.org/r/plugins/gitblit/ \o/ [15:19:14] <^demon> Whoooohoooooooo [15:20:18] good!! [15:20:29] <^demon> We'll switch over to it full-time soon. [15:20:40] <^demon> Maybe leave it like this for the time being so we can iron out any kinks. [15:20:50] :-) please do! [15:21:05] it is confusing how it shows the parent repositories as standalone [15:21:15] so it shows mediawiki as modified 5 months ago [15:21:25] and clicking there shows you the config changes [15:21:37] you have to browse down to get to mediawiki/core [15:21:59] <^demon> Yeah, gitblit doesn't know the difference. [15:25:49] ^demon: could you grant JenkinsBot submit access on translatewiki repo ? [15:25:50] https://gerrit.wikimedia.org/r/#/admin/projects/translatewiki,access [15:25:52] please ? :-] [15:26:00] that will get you free beer from Siebrand I guess [15:26:07] <^demon> Done. [15:26:10] I tried but failed :/ [15:26:20] no "you are not allowed to do that" message [15:26:20] hehe [15:26:27] <^demon> You couldn't do "Save for review?" [15:26:35] oh [15:26:37] yeah I did that one [15:26:40] that did not show up anything [15:26:43] so I did my change again [15:26:46] and pressed save that time [15:26:48] same thing [15:26:55] <^demon> Hrm, oh well. [15:27:02] <^demon> Will look at that later I guess. [15:27:08] so my changes are probably lost somewhere in Gerrit attic :-] [15:27:11] yeah not a big deal [15:28:24] ^demon: thanks for your quick action! [15:28:27] one less bug! [15:29:32] <^demon> Bug 43807 is also fixed from my side, dunno if there's anything you need to do. [15:31:28] ^demon: well if you feel like https://bugzilla.wikimedia.org/show_bug.cgi?id=43807 is fixed [15:31:33] ^demon: I would be bold and mark it fixed [15:31:49] Dereckson will reopen it if that it is not the case :-] [15:31:54] or we can try submitting a draft and see what happens [15:34:45] Nice, that will be handy. [15:35:38] ^demon: I confirm, the interface to create a new PS with a commit change is easy enough to use to avoid typo in commit bikeshedding anymore. [15:36:28] <^demon> :) [15:36:36] <^demon> Isn't that feature *awesome*? [15:36:54] yay!! [15:40:34] next feature plugin EtherPad so we can collaboratively amend the patchsets [15:41:45] <^demon> There is some idle discussion about allowing editing from the UI. But I think it's pretty far off. [15:41:50] <^demon> It'd be a huge change. [15:42:32] as long at the result is send over `stream-events` for Zuul to trigger Jenkins jobs, I will be fine with it :-] [15:42:35] Hello zeljkof: I see you're willing to mentor a GSoC browser testing project, that would be nice to offer an advanced testing methods project. [15:43:01] Dereckson: in a meeting, will reply in 30 minutes or so :) [15:43:08] Have a nice meeting. [15:44:03] thanks :) [15:44:34] <^demon> hashar: Well, it couldn't do constant saving. Saving would have to recompute the sha1. [15:44:50] <^demon> You wouldn't want to keep making commits and amending every second :p [15:47:32] no, you want to be able to commit at a certain stage [15:47:43] so you've done some collab editing and you like the result in the pad [15:47:59] then there's the save (and the problem with attribution but we'll sweep that under the rug for now) [15:48:42] ^demon: "Patch Set 4: Code-Review+1" -> this is by feature, for more clarity or you still have to configure and put back the old fulltext version? [15:48:53] <^demon> Need to tweak hooks [15:48:56] <^demon> output changed. [15:49:42] https://gerrit.wikimedia.org/r/#/q/project:operations/mediawiki-config+status:open,n,z --> previous col order were also CR V ? [15:52:51] <^demon> Can't remember...doesn't matter much really though. [15:57:32] for attribution, you just add an Author: line for each people editing it [16:00:08] daughter time :-D bbl [16:01:34] off for real now [16:03:03] Dereckson: ok, meeting over [16:03:09] what did you want to ask? [16:16:38] zeljkof: I noticed you enlist as mentor to GSoC, for browser automation. I'm happy of this move: to advocate better test practice, using advanced techniques is nice. [16:17:58] And so, to offer a GSoC project a good opportunity. [16:38:07] Dereckson: I see, you only said you like the idea :) [16:38:15] I thought there was something you wanted to ask [16:40:01] zeljkof: well yes [16:40:23] It would be nice if you could improve the description on the Mentorship_programs/Possible_projects page. [16:41:03] You could first clarify it's related to unit testing, functionnal testing or a mix of the two. [16:42:51] I would have thought it would be valuable to offer a better presentation of the issue, but I asked two developers (one student, one professional, none tester, both familiar with MediaWiki) and the description is clear enough for them. [16:43:58] (My fear were the project entry would only be read, actively acknowledged and understood by Selenium devs/users) [16:44:26] (but apparently I were wrong) [16:55:44] Dans le cadre du ContentHandler, tu as pu repéré les contributeurs experts des formats de fichiers ? [17:19:25] can someone explain "platonides is not allowed to run sudo on webtools-login." ? [17:20:34] You're not allowed to run sudo on webtools-login [17:20:54] well, the question would be why :P [17:21:08] (btw, I thought I was asking on #wikimedia-labs) [17:22:47] ^demon: Reedy: when you get time, would you please look at a change that normalize some maintenance scripts names? --> https://gerrit.wikimedia.org/r/#/c/48430/ [17:23:04] ok, I had to enable myself on Special:NovaSudoer [17:23:20] hashar, wouldn't that break existing scripts? [17:23:31] Platonides: possibly :-D [17:23:45] go ahead and cast your voice on the change :D [17:24:33] Platonides: also when uploading a new patchset, could you comment about what you are changing? ex: https://gerrit.wikimedia.org/r/#/c/47719/ [17:24:47] is that a real email? [17:24:54] waldir@email.com [17:25:21] apparently: $ dig +short MX email.com [17:25:21] 10 mx01.gmx.com. [17:25:22] 10 mx00.gmx.com. [17:25:42] gmx being a well known mail provider [17:26:15] or maybe I am confusing with gmx.de [17:26:49] heh [17:32:32] well, my question was more if he really had an email account in that domain [17:33:06] he does [17:59:18] ^demon: if you are around after all that.... [17:59:59] "Could not load exit nodes." [18:00:04] and guess that torblocklog has: [18:00:13] -rw-r--r-- 1 udp2log udp2log 0 Feb 11 16:51 torblock.log [18:00:18] *what [18:03:22] <^demon> apergos: Why is this cron supposed to run? Is there a comment? [18:03:31] uh [18:04:02] # Puppet Name: tor_exit_node_update [18:04:16] because the nodes change frequently I guess [18:04:18] <^demon> I think it's supposed to be run with --force. [18:04:26] <^demon> --force means "actually fetch them" [18:04:38] <^demon> Without --force, it just falls back to the cache (which has expired, hence, no nodes) [18:04:46] <^demon> We're trying to *update* the cache. [18:05:11] what does it do if it calls back? I mean what does it do after that? [18:05:21] anything useful? [18:05:21] *falls back [18:05:36] <^demon> Nope. [18:05:56] <^demon> I've got a suspicion. Lemme try something. [18:06:50] sure [18:07:38] <^demon> Yup. [18:07:40] <^demon> 2013-02-12 18:07:27 fenari aawiki: Unable to load Tor exit node list: cold load disabled on page-views. [18:07:46] <^demon> That annoying error, I enabled it only for cli. [18:09:03] yep, I see it [18:11:09] <^demon> So, just run it with --force. [18:11:51] every ten mins eh? [18:11:53] meh [18:13:05] <^demon> Hrm, it's supposed to stay in memcached for 30m :\ [18:14:06] well I'm doing it anyways [18:14:47] <^demon> I just ran it with --force, but it's non in memc. [18:14:50] <^demon> wtf. [18:16:43] so this is likely to be overkill [18:16:56] but I'd rather have a bit of overkill than broken [18:17:47] ok it's in, should run like that on next cron [18:18:13] ok it's every 20 mins but it feels like every 10, it's a very chatty script [18:18:27] <^demon> It's very weird that I can't find those keys in memc. [18:18:54] also whoever has bots-bnr1 lab instance seriously needs to clean up their cron script [18:18:55] s [18:19:15] as do the wikidata folks, that's been going on for weeks, several thousand messages a day [18:19:25] [18:19:29] <^demon> wikidata on labs or wikidata on hume? [18:19:58] labs [18:22:43] wikidata-testclient.pmtpa.wmflabs and wikidata-dev-9.pmtpa.wmflabs and wikidata-dev-8.pmtpa.wmflabs [18:31:02] apergos: I worked on tor block a little-- let me know if you need help, but thanks for making sure that runs! [18:31:24] csteipp: thanks, we'll see! [19:00:31] FYI, deploying WikibaseClient to enwiki in a minute or 2 [19:55:21] mlitn: are you deploying? [19:55:37] I am not - feel free if you need to [19:56:14] thanks :) [20:02:32] kaldari, can you fix the e2 list's settings to disallow posts from non-members, like you did with ee? Just got some weird spam from it [20:05:13] <^demon> qchris: I saw your fix for the OR predicate. [20:05:25] Ha! [20:05:31] I hope it gets accepted [20:05:45] The gerrit serach implementation is somewhat nasty. [20:05:47] <^demon> Yeah, feedback would be nice before we deploy. [20:06:45] I started trying to bisect it ... i stopped as I saw it did not work at least half a year ago [20:14:36] ahh [20:14:45] now I understand my confusion on the Gerrit dashboard [20:14:56] <^demon> Hm? [20:14:59] the CR and V columns have been swapped [20:15:05] so we have CR first then Verified [20:15:12] <^demon> I dunno what did that. [20:15:40] not really important :-] [20:15:52] lets just let them like thati [20:16:08] it will only takes a few days for my brain to adapt [20:16:47] The approval_categories has a position column... could that help? [20:16:55] <^demon> Heh, never noticed that. [20:16:59] <^demon> It's no big deal. [20:18:33] * qchris just tested locally. The position column allows to swap them back again [20:18:50] I heard someone else complain about that in some channel before. [20:20:57] kaldari: FYI Thehelpfulone has done what I asked earlier [20:52:53] Krenair: cool. Sorry, I was away at lunch [20:58:19] <^demon> qchris: I have a feeling this won't get accepted :\ http://p.defau.lt/?_j5YfAQq1MmtNRigcFwr4w [20:58:23] <^demon> ;-) [20:59:13] :-) [20:59:40] <^demon> Every time a hook runs, you end up with [20:59:40] <^demon> [2013-02-12 20:57:03,394] INFO com.google.gerrit.common.ChangeHookRunner : hook[comment-added] exitValue:0 [20:59:44] <^demon> So silly. [21:00:11] ^demon: maybe DEBUG level if it exit 0 [21:00:16] else INFO ? [21:00:36] That sounds good. yes. [21:00:51] <^demon> Oh, that could do it. [21:01:22] or even 'warning' [21:01:26] i dont know [21:01:34] <^demon> Nah, there's some code below it also doing info. [21:01:39] <^demon> Would be good to stay consistent. [21:03:02] But nonetheless ... silencing 0 would be good [21:03:22] <^demon> I'll move it off to debug. [21:03:31] <^demon> That way if someone really wants it still, they can turn it on. [21:18:44] marktraceur: https://gerrit.wikimedia.org/r/#/c/48739/ [21:22:38] kaldari: Good plan [22:12:45] Anyone around who might be able to review an interwiki-userrights change? [22:13:59] ^demon|away, how can I rebase a local commit on top of one which is the latest patchset of a gerrit change? [22:14:26] Darn, hotkey "r" doesn't work in Gerrit (when writing a comment in a diff, I can't click Save and then "r" to go to the review form to submit it) [22:14:33] ^demon|away: Ryan_Lane ^ [22:15:00] e.g. double click in https://gerrit.wikimedia.org/r/#/c/48033/1/modules/jquery.footerCollapsibleList.js write a comment and press "r" to submit it. It does pop up the bar "Working.." but doesn't actually end up doing anything [22:15:05] Known or shall I file a bug [22:15:08] git rebase --onto change1 HEAD~1 [22:15:20] Throws "Uncaught TypeError: Cannot read property 'canEditTopicName_0' of null " [22:15:23] when I press "r" [22:15:32] <^demon|away> Krinkle: WFM? [22:15:48] Which is odd since I'm not editing the topic at all [22:15:57] Confirmed, I'm getting the same thing [22:16:10] so you'd git review -d , switch to the change to be rebased on the first, and rebase onto the former [22:16:12] <^demon|away> Hmm, all Registered Users have edit topic rights. [22:16:52] Krinkle: I get "working" hang ups randomly a lot of the time [22:17:00] for various different actions [22:17:44] AaronSchulz, and change1 is the branch name? [22:17:54] the commit id [22:18:03] (the git one, not the gerrit one) [22:18:59] I think cherry-picking would work too [22:19:11] meh [22:21:28] AaronSchulz, great, thanks [22:21:30] "You have more than one commit that you are about to submit." [22:21:32] I guess this is okay.. [22:21:43] you want a dependent change right? [22:21:47] since both are pending [22:21:56] yep [22:22:00] so yeah [22:22:04] git log look fine? [22:22:10] yep [22:25:08] worked as I expected. thanks again [22:28:15] anyone feel like giving a review of a skin? https://github.com/ryan-lane/strapping-mediawiki [22:28:22] I'll pay in liquor [22:29:53] <^demon|away> AaronSchulz: It does "Working..." anytime it's doing a call to the rest api. [22:31:12] Ooh, liquor. [22:31:20] meh [22:31:24] Ryan_Lane: Maybe link to the test of labsconsole with the skin on it? [22:31:24] * AaronSchulz is a cocktail guy [22:31:40] marktraceur: https://labsconsole-test.wmflabs.org/wiki/Main_Page [22:31:53] AaronSchulz: that counts [22:31:56] Ryan_Lane: Sweet. Now, what do you mean by "review" [22:32:04] make sure the code is secure? [22:32:09] Ah right. [22:32:31] Ryan_Lane: did you ask chris? [22:32:33] <^demon|away> What do you know, beer o'clock. [22:33:02] csteipp: mind a security review on a skin? :) https://github.com/ryan-lane/strapping-mediawiki [22:33:47] Ryan_Lane: That's on my list to look at [22:33:59] cool [22:34:05] note that my fork has a lot of changes [22:34:08] Oh, liquor is involved?? [22:34:13] I've upstreamed them via a pull request [22:34:14] That just moved up my priority list.. [22:34:17] hahaha [22:34:19] nice [22:36:46] "Edit LocalSettings.php to add: require_once( "$IP/skins/strapping/strapping.php" );" [22:36:48] * Reedy cries [22:37:02] Reedy: suggest a better way [22:37:06] it's better than using extensions [22:37:17] It's hardly any different [22:37:19] and it's way better than "link the files from the skin directory" [22:37:43] just putting stuff in the skins folder isn't so much "usual" behaviour [22:37:52] we have no proper way to handle skins [22:38:20] I kind of like the idea of keeping them in the skins directory and handling skins somewhat similarly to extensions [22:39:00] I read the last thread about this on wikitech-l [22:39:07] what I came back with was "I hate mediawiki" [22:39:40] no consensus [22:39:54] there's no documentation on how to write skins as extensions [22:40:09] :D [22:40:18] there *is* documentation about how to write skins the way I've done it [22:42:20] For all it matters they can be in the extensions folder like that [22:42:32] not without changing the style paths and such [22:42:59] I'm fine with using any recommended method [22:43:10] alas, there is none ;) [22:43:55] Might want to fix the url too [22:43:55] 'url' => "[...]", [22:44:58] url? [22:45:14] in extension credits [22:45:21] ah [22:45:21] heh [22:50:29] ^demon|away: can you create a repo for me? [22:50:32] heh. away [22:50:33] damn [22:50:38] guess I'll need to make it myself [22:50:47] <^demon|away> Oh darn, how will you ever live :p [22:50:50] \o/ [22:50:52] you're here! [22:50:57] <^demon|away> crap. [22:50:58] mediawiki/skins/mediawiki-strapping [22:51:11] hahaha [22:51:27] <^demon|away4real> You're pulling existing history, right? [22:51:31] yep [22:51:35] <^demon|away4real> k, no initial commit then. [22:52:16] <^demon|away4real> done. [22:52:18] thanks! [22:54:15] now I need to remember how to push in the history [22:55:32] <^demon|away4real> `git clone`, `git pull origin refs/*:refs/*`, `git remote add gerrit ` `git push gerrit refs/*:refs/*` [22:55:48] <^demon|away4real> Might need to adjust acls to grant direct pushing and forge author. [22:55:48] cool [22:55:49] thanks [22:55:56] * Ryan_Lane nods [22:58:58] hooray! imported