[00:02:05] <jeremyb>	 preilly: that demo site uses short URLs with no prefix/namespace
[00:02:08] <jeremyb>	 :(
[00:02:17] <jeremyb>	 (for your github link)
[00:03:10] <jeremyb>	 https://www.mediawiki.org/wiki/Manual:Wiki_in_site_root_directory
[02:53:27] <Emw>	 i've done an initial commit of an extension i'm working on -- https://gerrit.wikimedia.org/r/#/c/46897/ -- could anyone do a high-level review?
[08:30:46] <gerrit-wm>	 New review: Hashar; "and deployed :-)" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/46878
[12:53:32] <^demon>	 qchris: Heya, just got your e-mail (and just sent you one too).
[12:53:44] * qchris  checks mailbox

[12:54:21] <^demon>	 Yeah, gerrit-dev is running 0d625dcb6ac5e6754e96a5c6575a8c5c9fbc0c47 now.
[12:54:45] <^demon>	 I updated it just before you e-mailed me saying it wasn't a working sha1.
[12:54:54] <qchris>	 Ah ... I was too late with my email. Sorry.
[12:55:23] <qchris>	 The culprit seems to be the REST delete commit that you mentioned some days ago.
[12:55:51] <^demon>	 I don't think so. It was having the same problem when it was still using JSON-RPC before it was rewritten to REST.
[12:56:03] <^demon>	 And most of the underlying logic didn't change.
[12:56:03] <qchris>	 I was talking about the regression
[12:56:08] <^demon>	 Ah, ok.
[12:56:18] <^demon>	 Yeah, quite possibly.
[12:56:33] <qchris>	 I am just about to try adding some sleep statements into the code
[12:56:39] <qchris>	 there is some list iterations
[12:56:47] <qchris>	 while also changig the list
[12:56:54] <qchris>	 that looks somewhat fishy
[12:57:21] <qchris>	 btw is there some way to speed up gwtui compilation?
[12:57:31] <qchris>	 Those 6 permutations take ages to compile :-(
[12:57:46] <^demon>	 Faster computer or more patience are the only solutions I've found :\
[12:57:52] <qchris>	 *gg*
[12:58:07] <qchris>	 My wive just told me another one: "dinner's ready"
[12:58:48] * qchris  grabs some food and lets the machine do the work

[12:58:55] <^demon>	 Enjoy dinner.
[12:59:29] <^demon>	 So yeah, I hadn't looked at that list yet, but I was wondering if something like that might be happening. I spent a good part of yesterday delving into gwtorm, and the only reason I could see us tripping this exception would be if we were passing a bogus list of rows.
[13:00:00] <^demon>	 Like if we had a null entry in the list, it would throw off the "count how many records we should update and compare that to what we are going to update" count.
[13:06:01] <^demon>	 hashar: Would it be ok if we installed https://wiki.jenkins-ci.org/display/JENKINS/Bazaar+Plugin?
[13:06:14] <^demon>	 I've got a maven project hosted in a bzr repo I'd like to build.
[13:06:23] <hashar>	 sure go ahead :-]
[13:06:35] <^demon>	 Ok, thanks.
[13:06:59] <^demon>	 "The plugin requires the Bazaar binary (bzr) to be installed on the target machine."
[13:07:00] <^demon>	 Bahhhh
[13:07:42] <hashar>	 ahh
[13:07:45] <hashar>	 we can get it installed
[13:08:25] <hashar>	 let me write the puppet update and do the install manually
[13:09:02] <^demon>	 I don't think it's that important.
[13:09:53] <^demon>	 Don't waste your time :)
[13:10:05] <hashar>	 done
[13:10:27] <hashar>	  https://gerrit.wikimedia.org/r/46931
[13:10:31] <hashar>	 and I have installed it manually
[13:11:26] <hashar>	 it will eventually be merged one day :-]
[13:12:49] <^demon>	 Oh man, I was wrong.
[13:12:54] <^demon>	 I need mercurial, not bzr.
[13:13:07] <^demon>	 I don't think the plugin needs any packages, either.
[13:16:36] <hashar>	 ERR: only one package per day can be processed. Please hold.
[13:17:13] <^demon>	 Huh?
[13:17:18] <hashar>	 ;-D
[13:17:55] <^demon>	 Yeah, don't worry about this. Silly upstream for not using git anyway.
[13:18:15] <^demon>	 If he'd just push 2.1.2 to maven central, I wouldn't even need to build myself.
[13:21:21] <hashar>	 amended patch and installed mercurial manually
[13:24:59] <^demon>	 Seriously, you can't build this in batch mode?
[13:25:03] <^demon>	 What a silly plugin.
[13:25:37] <^demon>	 Haha, it tries to gpg sign the build.
[13:25:38] <^demon>	 That's why.
[13:25:56] <^demon>	 hashar: Forget all of this. I'll *have* to build locally.
[13:28:20] <hashar>	 well you get mercurial
[13:28:36] <hashar>	 what is the problem with gpg?  I guess you can have it skip the signing?
[13:30:59] <^demon>	 Ah, -Dgpg.skip=true might work
[13:31:00] <^demon>	 Lemme try
[13:31:17] <^demon>	 Ah, yep.
[13:32:51] <^demon>	 https://integration.mediawiki.org/ci/view/Java/job/j2bugzilla/ - yay :)
[13:33:57] <hashar>	 goood job
[13:39:58] <qchris>	 ^demon: Should I ping the j2bugzilla author uploading a release, or is it ok for now to build locally?
[13:40:24] <^demon>	 I got it building on jenkins.
[13:41:06] <qchris>	 ok.
[13:43:48] <^demon>	 https://gerrit-dev.wmflabs.org/r/#/admin/plugins/ - wheee
[13:45:06] <qchris>	 "Capability administrateServer is required to access this resource" ;-)
[13:45:29] <^demon>	 Ah, that can be fixed.
[13:45:35] <qchris>	 nono it's ok
[13:45:47] <^demon>	 Put you in admin.
[13:45:50] <^demon>	 It's dev :p
[13:45:50] <qchris>	 I guess it means you got the hooks-bugzilla plugin installed. \o/
[13:46:28] <qchris>	 Now it works.
[13:47:17] <qchris>	 btw. hooks-its did not yet merge all required patches, so hopefully you did not buld hooks-its from master.
[13:48:11] <^demon>	 Jenkins builds it all from master.
[13:48:20] <qchris>	 oh ... rats :-)
[13:48:43] <^demon>	 We can put a custom build on dev if need be.
[13:49:00] * qchris  checks what the missing parte are

[13:51:55] <qchris>	 It should work from master as well. But user names don't show nicely, and you have to give conditions in gerrit ids ('CRVW') instead of treanslated names ('code-review')
[13:53:01] <^demon>	 I haven't configured it yet anyway.
[13:53:08] <^demon>	 We can hold off until the patches go into its.
[13:53:14] <qchris>	 ok.
[14:03:43] <^demon>	 Ah, the stacktrace changed a bit with some of the recent changes in master. It's definitely the writeAudits() that's causing problems.
[14:04:03] <^demon>	 Current stacktrace: http://p.defau.lt/?1Q8OzsPJ4I4_asFs0e9foA
[14:05:15] <^demon>	 The != null should protect against null entries.
[14:06:06] <^demon>	 A-ha!
[14:06:09] <^demon>	 I think I got it.
[14:09:12] <^demon>	 Or not.
[14:13:30] <qchris>	 yop writeAudits it is. And Edwin's also touched that part with his commits.
[14:13:58] <^demon>	 What I need to do is manage to replicate the bug while in debug mode.
[14:14:07] <^demon>	 I've not yet managed to replicate locally. I'm going to do that today.
[14:14:27] <qchris>	 Once we can do that the fix should be easy
[14:15:10] <qchris>	 But even when not in debug mode ... could you just add log statements so we see where it breaks?
[14:15:23] <qchris>	 (and wore importantly "how it breaks")
[14:15:29] <qchris>	 wore -> more
[14:21:58] <^demon>	 I can't replicate when mysql is on my localhost. I'm going to spin up another vm and see what happens when they're on different hosts. Wonder if we're hitting some race condition.
[14:29:12] <qchris>	 I tried that already and did not see any results (even had LDAP on separate machine)
[14:59:58] <^demon>	 qchris: Ok, ldap on one ubuntu vm, gerrit on another, mysql on a third, using a browser on a 4th.
[15:00:02] <^demon>	 Still cannot replicate.
[15:00:10] <qchris>	 :-(((
[15:00:20] <qchris>	 The same for me.
[15:01:25] <qchris>	 Seeing all the "run"s in the code, it somewhat looks like it's a race condition around the looping constructs
[15:05:34] <^demon>	 Maybe we can ask Shawn later and see if he has a clue.
[15:05:41] <^demon>	 I'm out of ideas.
[15:05:45] <qchris>	 Sounds good.
[15:05:58] <^demon>	 Actually, I'll toss it on repo-discuss.
[15:05:59] <qchris>	 It seems I have to leave for now. Be back in a few hours.
[15:06:11] <qchris>	 Yes, sounds good
[15:06:36] <DanielK_WMDE_>	 hi all! i have a few small but important core patches. i'm somewhat blocked on them. anyone up to have a look?
[15:06:41] <DanielK_WMDE_>	 these:
[15:06:43] <DanielK_WMDE_>	 https://gerrit.wikimedia.org/r/#/c/46940/
[15:06:47] <DanielK_WMDE_>	 https://gerrit.wikimedia.org/r/#/c/46876/
[15:06:54] <DanielK_WMDE_>	 i have more, but these are the important ones :)
[15:12:36] <^demon>	 DanielK_WMDE_: +2 on both.
[15:18:34] <DanielK_WMDE_>	 ^demon: thanks!
[15:18:41] <^demon>	 yw
[19:54:14] <kaldari>	 Krinkle: FYI: https://bugzilla.wikimedia.org/show_bug.cgi?id=44525
[20:01:38] <^demon>	 qchris: Edwin seems to have found+fix the "Group Not Found" bit. https://gerrit-review.googlesource.com/#/c/41824/
[20:01:58] <qchris>	 Ah that's good news!
[20:02:31] <qchris>	 I isolated that his commit was the problem but then switched to our more pressing issue.
[20:02:43] <^demon>	 the toString() was encoding the : in ://
[20:02:58] <qchris>	 Ah.
[20:03:02] <qchris>	 That makes sense.
[20:03:49] <qchris>	 I saw your message on repo-discuss
[20:04:34] <qchris>	 Hopefully some good answer comes in from out of nowhere :-)
[20:17:13] <qchris>	 ^demon: I just tried https://gerrit-review.googlesource.com/#/c/41824/
[20:17:25] <qchris>	 I now can delete "Registered Users"
[20:17:37] <qchris>	 But I canno delete ldap groups.
[20:17:49] <qchris>	 Does that work for you as well?
[20:18:00] <^demon>	 https://gerrit-dev.wmflabs.org/r/#/admin/groups/6,members - I can't delete anything from this group.
[20:18:04] <^demon>	 Same orm exception.
[20:18:42] <qchris>	 That's good actually
[20:19:03] <^demon>	 Yeah, it makes me think everything's fixed other than our original bug.
[20:19:04] <qchris>	 Because it means the problem is probably in the code that served as stencil for his code .... or am I wrong here?
[20:19:20] <^demon>	 Yep, my thoughts exactly.
[20:19:42] <qchris>	 I cannot delete ldap/wmf (even locally). Before I could.
[20:19:54] <qchris>	 before = before Edwins two patches
[20:20:02] <^demon>	 What do you get now?
[20:20:14] <^demon>	 (locally)
[20:20:15] <qchris>	 The page you requested was not found, or you do not have permission to view this page.
[20:20:29] <qchris>	 So probably still some redirection problem
[20:20:37] <^demon>	 Ah, probably.
[20:20:41] <qchris>	 Does that work for you locally?
[20:20:45] <^demon>	 Yep.
[20:20:50] <qchris>	 Oh ...
[20:20:57] <qchris>	 Thanks
[20:21:45] <^demon>	 Do you have canonicalWebUrl set?
[20:21:50] <qchris>	 Yes
[20:22:17] <qchris>	 It starts failing with edwins patches.
[20:22:37] <qchris>	 So I'll have to investigate that as well.
[20:22:47] <qchris>	 One gets fixed. Two new open up :-)
[20:29:01] <^demon>	 qchris: Yep, on my local vms master + Edwin's latest patch works just fine.
[20:29:16] <^demon>	 Just confirmed across my 4-vm-sorta-like-labs setup.
[20:29:19] <qchris>	 Ok. Thanks.
[20:29:26] <qchris>	 That's the hell of a bug :-((
[20:37:01] <qchris>	 ^demon: Now I found the reason for the permission problem. It breaks when you try to remove a group that the current user is not part of.
[20:37:38] <qchris>	 I can remove ldap groups that I am part of.
[20:37:59] <^demon>	 Hmm. Worth looking at, but not a blocker really. It's not often that'll happen.
[20:38:34] <qchris>	 Of course.
[20:56:48] <Krenair>	 Ah looks like AaronSchulz beat me, I've been too busy with git log/blame :)
[21:03:21] <Krenair>	 okay, wtf. 16b87692cc7d0542bc225aedcd93949f46662444 (FlaggedRevs), Sat Feb 26 03:04:59 2011 +0000? Really?
[21:03:39] <Krenair>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=44468 ...