[09:19:19] andre__: Bug 31369 - redirects.conf should not hardcode HTTP https://bugzilla.wikimedia.org/show_bug.cgi?id=31369 [09:19:40] andre__: patch is being reviewed under gerrit and waiting for submitter to fix it :-] [09:19:53] (I have updated bug report) [09:20:06] yeah, but that patch already has three -1 [09:20:12] so somebody needs to update it :) [09:20:21] I wouldn't call it "being reviewed". [09:20:28] to me it's reviewed already and considered bad. [09:28:30] yup [09:28:51] andre__: on that topic, that should probably handled by ops but they already have enough stuff to handle :-D [09:29:45] as usual :-/ [13:15:53] andre__: hashar: yes, i should work on it. but would be nice to be able to test it completely in beta. with squid and apache and nginx and everything the same as prod [13:16:14] andre__: hashar: anyway, not happening today... ;( [13:17:05] Does "should" mean "I wish I had time in the next years" (a wishful intention) or more like "I am fairly sure that I have time within the next week" (a plan)? [13:17:14] ;) [13:17:26] somewhere in between [13:17:48] i wish i could get all the code review responses offline [13:17:56] anyway, /me runs away [13:17:57] heh :) [18:05:43] * robla starts poking around for the script [18:05:49] ...to make release notes [18:06:52] ^demon: https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/tools/release.git;a=tree;f=make-release [18:07:15] oops... [18:07:40] https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/tools/release.git;a=tree;f=make-deploy-notes [18:07:42] that one [18:29:48] ^demon: did you get my note earlier about make-deploy-notes? https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/tools/release.git;a=tree;f=make-deploy-notes [18:30:05] <^demon> No, I didn't. Got disconnected. [18:30:22] well, now you have :-) [18:31:09] <^demon> Got it. I have the repo checked out, should've seen that. [18:31:34] <^demon> I made make-wmf-branch much faster this morning. You can now point it to a local on-disk clone rather than having to clone core over ssh. [18:35:18] <^demon> wmf/1.21wmf6 branch created. [18:35:40] <^demon> https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/core.git;a=shortlog;h=refs/heads/wmf/1.21wmf6 [18:39:46] hi ^demon [18:42:10] <^demon> aude: Howdy :) [18:42:53] ^demon: you know which tag / branch we're using for wikibase? [18:44:40] <^demon> branch "mw1.21-wmf6" [18:44:55] yep [18:45:01] <^demon> I've already done the 1.21wmf6 branch, so if anything else needs to go in, we'll need to submodule update. [18:45:05] ok [18:45:11] and for diff and data values tag 2012-12-05 [18:45:19] hi AnjaJ_WMDE [18:45:24] and what AnjaJ_WMDE says [18:45:41] hi katie, was waiting on #wikimedia-wikidata ;) [18:50:09] <^demon> https://www.mediawiki.org/wiki/MediaWiki_1.21/wmf6 updated with all release notes. [18:53:56] aude, AnjaJ_WMDE, ^demon: i'm here [18:54:02] ping me if something comes up [18:54:27] DanielK_WMDE_: ok [18:54:56] grr. [18:55:07] if ^demon wants, i can tag the git submodules as patch against core branch [18:55:08] should have ported Ic8b47e6b to wmf6 not wmf5 [18:55:17] without that, we can't debug the RC stuff [18:55:20] oops? [18:55:20] oh well, perhaps later [18:55:30] AnjaJ_WMDE: not a big deal. can be patched in later [18:55:55] <^demon> I haven't deployed to any wikis yet--if you want to update the submodule you've still got plenty of time. [18:56:07] <^demon> Have to finish a bunch of maintenance tasks and do test/test2 before we do mw.org and wikidata.org [18:56:25] ^demon: ok, doing [18:56:43] test2 has wikibase [18:56:59] <^demon> Ah yes. Well, still a couple of things to do :) [18:57:00] ^demon: so, test2 goes live before any other? in particular, always before wikidata.org? but real wikipedias will get updated ofter wikibase.org, right? [18:57:02] so we probably want the submodules tagged before that [18:57:18] DanielK_WMDE_: i think real wikipedias are a week later [18:57:27] enwiki the following monday and then all the other wikipedias [18:57:42] hm... so the client has to work with an older repo and a newer repos. [18:57:46] that'S going to be a challange [18:57:49] yes [18:57:53] no way around that i think [18:57:59] but gives us a week to test the client [18:58:21] well there is already a change on master that means that the latest repo version will break old clients... [18:58:42] we have to stop doing that [18:58:52] maintain some bw compat [18:58:58] for 2 weeks [18:59:14] aude: well, removing entity data from the change objects is pretty important. and if you do that on the repo and don't update client code... [18:59:18] well, it'll break, no? [18:59:33] we'd have to plan far enough ahead and update the client code in advance [19:00:11] the replacement for entity data has to come first, before we remove the old stuff [19:00:33] ok for now, but from now on [19:00:48] anyway, submodules.... [19:01:24] ^demon: scap time? [19:01:41] <^demon> Almost. Doing the checkout for the new version now. [19:01:44] <^demon> But yeah, we're in motion. [19:02:12] almost.... [19:02:16] xlnt [19:02:42] <^demon> robla: Full release notes are up at [[mw:MediaWiki 1.21/wmf6]]. New hook is in place to link to it too. [19:02:53] \o/ [19:03:01] <^demon> Also, we'll need to manually do release notes for Onboarding and DataValues exts, script was giving an exception. [19:03:52] those are new extensions to 1.21wmf6, so that's probably the problem [19:04:34] well, hmm....DataValues is newly deployed, but not new [19:04:34] datavalues was new in wmf5 [19:04:36] <^demon> Yeah, there's supposedly logic for new extensions, but it must be broken. [19:04:41] <^demon> I just hacked it out [19:05:21] I might be able to figure out what's up [19:06:36] <^demon> Core done, submodules updating. [19:10:46] ok [19:11:12] <^demon> Is the change in the wikibase branch yet? [19:11:22] not yet [19:11:47] 6828c5822a4c6ec77ee4ca806293386a4b510d57 for diff [19:11:54] c403a69d02643aac66b6bd0a0bf5de275524642e for datavalues [19:11:59] and the branch for wikibase [19:12:55] * aude is doing somethign wrong [19:13:03] do i really have to do git submodule update --init for all the extensions? [19:13:47] <^demon> Diff and Datavalues are already correct. [19:15:06] ok [19:15:13] <^demon> If you just update the mw1.21-wmf6 branch on wikibase, I can take care of the wmf branch on core. [19:15:38] ok [19:18:05] <^demon> Ok, ready to run scap. [19:22:48] errr.... [19:23:02] <^demon> Yes? [19:23:30] the submodules was easy last tiem but i'm confused this time [19:23:42] you do it? [19:23:54] <^demon> I said don't worry about the submodule, just get it in the branch. [19:23:54] <^demon> I'll take care of the submodule. [19:24:11] what do you mean? [19:24:18] i think the branch is ready [19:24:36] <^demon> I thought we were saying we wanted a change added to https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/Wikibase.git;a=log;h=refs/heads/mw1.21-wmf6? [19:24:42] ah, ok [19:24:44] i can do that quick [19:28:24] done! [19:28:34] https://gerrit.wikimedia.org/r/#/c/37865/ [19:31:01] <^demon> Got it into the deployment branch. [19:31:13] <^demon> robla: Just started scap. Gonna finish a sudoku while we wait ;-) [19:31:23] :) [19:31:25] yay! [19:35:52] <^demon> AaronSchulz: Pfft. [19:35:54] <^demon> -rw--w---- 1 aaron wikidev 138 2012-12-05 21:49 php-1.21wmf5/maintenance/.mwsql_history [19:36:36] .mwsql? [19:37:02] <^demon> The history file for sql.php if you're in interactive mode, I think. [19:37:08] <^demon> Anyway, scap tries to sync it but I can't read it :p [19:38:44] <^demon> Also, getting [19:38:55] <^demon> "mw11: rsync error: some files/attrs were not transferred (see previous errors) (code 23) at main.c(1536) [generator=3.0.9]" [19:38:55] <^demon> Is this semi-normal with scap these days? [19:38:56] aude: thanks for backporting the debug stuff [19:39:08] ^demon: not that I'm aware of [19:39:53] <^demon> AaronSchulz: Well, they seem to later show done, so maybe. I dunno, I haven't run scap in ages. [19:41:37] <^demon> AaronSchulz: Can you change the permissions on that file in /h/w/c/php-1.21wmf5/ on fenari? I think these errors are permission errors cuz of that. [19:41:56] ok, if i do git submodule update --init for everything, then the submodules work for me [19:41:57] that file is gone [19:42:14] * aude knows for next time to be patience and wait for all the submodules to update [19:42:19] patient [19:42:34] <^demon> thanks. [19:46:28] I can connect to en.wp over http but not https right now [19:47:04] sumanah: oh, you can repro this? [19:47:06] "Cannot connect to destination (en.wikipedia.org)." [19:47:16] works for me [19:47:23] we're talking about this on #wikimedia-operations right now [19:47:26] tell me if there are other steps I should try to repro -- I turn on HTTPS everywhere and .... ok [19:55:19] <^demon> So, anyone seen any good movies lately? [19:55:24] * ^demon is trying to pass the time [19:57:21] heh [19:57:30] ^demon: I saw "Lincoln" & "The Lady Eve" :-) [19:58:06] <^demon> I actually re-watched The Lady Eve last weekend. Great movie. [20:06:40] it is so good! [20:08:04] <^demon> binasher: So, we all agreed last week that today's a bad day to try and sneak an s3 -> s5 move in for wikidatawiki. What does your schedule look like this week for trying to get that done? [20:08:19] <^demon> Also ping aude, DanielK_WMDE_. [20:08:56] hi [20:09:16] any day should work for me [20:11:05] Evening guys - just thought I'd poke my head in and let you know that from what I can see, wikipedia addresses started with https:// are down [20:12:20] I can get to all wikipedia using http://, but using https:// firefox returns an error saying it can't find the server at https://(foo).wikipedia.org [20:13:56] BarkingFish: https://en.wikipedia.org/ is fine for me…. [20:14:36] haven't been able to hit that now for approaching 2 hours, brion [20:14:42] BarkingFish: you in europe by any chance? [20:14:59] oh hi brion [20:14:59] yeah [20:15:44] * brion takes a look at http://nagios.wikimedia.org/ but isn't up on what's what [20:15:53] ^demon: you want to know our preference for moving from s3 to s5? i donh't care much, honestly [20:15:59] mutante or RobH: any issues with SSL in knams? [20:16:02] AnjaJ_WMDE: what do you think --^^ [20:16:14] got a report of trouble reaching https://en.wikipedia.org etc from europe [20:16:27] hrmm [20:16:38] brion: status.wikimedia.org isnt seeing it [20:16:46] but perhaps your getting the start of an issue [20:17:22] me neither, Danielk_WMDE [20:17:35] BarkingFish says has had problems for 2 hrs. could be something weird that's only affecting him, dunno [20:17:39] brion: So status.wikimedia.org is an offsite service provider that hits all our stuff [20:17:53] *nod* where all does it test from? [20:17:56] so it would see something like an EU based outage before most of us US folks will know its happening [20:18:24] oh a bunch of places, it used to be called watchmouse [20:18:28] its another company now, lemme get name [20:18:31] nagios lists ssl3001 as down with 'connection refused', don't know if that's suspicious [20:18:38] * RobH is also grepping about on the servers to see whats up [20:19:00] yea there are two others and its loadbalanced [20:19:09] and the two others are not doing much traffic, but they are doing something [20:19:36] there was an odd dip though [20:19:51] <^demon> RobH: Could you please `chown mwdeploy:wikidev /home/wikipedia/common/wmf-config/ExtensionMessages-1.21wmf6.php` on fenari? [20:21:13] * DanielK_WMDE_ is afk for some minutes [20:21:28] ^demon: yep, done, matches the other extensionwhatevers next to it. [20:21:32] <^demon> Thanks :) [20:21:48] brion: i dont see an issue, but i only have a passing familiarity with this [20:21:53] so if its an odd edge case im more than likely missing it [20:22:03] ^demon: oh, in case you didn't see: we can move to s5 whenever, really. well, thursday evening i'll be on th train for 2 hours, but that's about it. [20:22:08] funky [20:22:18] BarkingFish: any luck with another browser, or closing & restarting firefox? [20:22:21] brion: lets get someone in EU to use https [20:22:45] RobH, brion: i use https per default from germany [20:22:50] wp looks good [20:22:51] what kind of issue you seein? [20:22:54] anything else i should check? [20:23:09] Works fine for me over https from europe [20:23:18] RobH: no issue so far, tell me what to test :) [20:23:33] no clue, brion was just asking if we had issue [20:23:34] RobH: oh - that's logged in [20:23:37] let me try anon [20:23:44] man i hate when problems can't be reproduced :P [20:23:51] Anon works for me too [20:24:05] me too [20:24:12] (Telekom here) [20:24:26] brion: why? just close WFM ;) [20:24:26] brion: see you assume its not fixed. i assume it was broken and the mere idea of me having to fix them scared the servers into compliance. [20:24:36] :) [20:24:59] sorry brion, was busy for a few seconds. [20:25:05] no worries [20:25:11] The ESAMS servers are old servers, they have seen others servers I have seen fit to use as example to how to obey the sysadmins [20:25:16] brion: may i shamelessly abuse the situation to direct you attention towards https://bugzilla.wikimedia.org/show_bug.cgi?id=42825 ? [20:25:16] i will try again from here, but I'm gonna reboot first [20:25:25] since tim is away, i have no idea who to ask about this [20:25:26] k good luck BarkingFish [20:25:26] see you in a bit :) [20:25:48] * aude has karaoke on thursday evening :D [20:25:55] if i'm not busy with more important stuff [20:26:00] * Damianz thinks someone needs to tell BarkingFish that fish go bob not woof [20:26:13] aude: oh, monster ishiban is up on thrusday? maybe i should stay in berlin :P [20:26:33] DanielK_WMDE_: you should! [20:26:48] hrrrmmmm [20:27:12] brion: i'm sure it's a simple issue with ES setup - but i have no idea where to ewven start looking :) [20:27:31] hmm [20:27:55] * aude not sure of the details of how ES works, yet [20:27:59] DanielK_WMDE_: so i have the vaguest impression that ES is separated by databases as well [20:28:09] like there's an 'enwiki' with 'blobs' and a 'dewiki' with 'blobs' etc [20:28:21] mmmmm...... [20:28:34] so just fetching from something like 'DB://cluster25/316188' i think will give you something from the current wiki (or try) [20:28:50] heh, funny, I was just looking at that bug [20:29:27] Where you trying to kill it with a magnifying glass? [20:29:46] brion: that would be pretty bad... nad i'd have be get really mad at Tim then :) [20:30:09] brion: yes there are separate dbs [20:30:13] brion: because he pushed us to "don't copy, just access the other wiki", and i shopwed the spec to him... grrrr. [20:30:20] heh [20:30:23] brion: i replaced the SSL cert earlier, and that required nginx restarts of all nginx boxes.. additionally 2 boxes in esams are broken (hardware) but that isnt new [20:30:30] there should be some way to do it though i just…. don't know what :D [20:30:34] its initialized in addWiki.php for each new wiki [20:30:37] AaronSchulz: separate DBs, yea, but using the same DB:// urls? [20:30:40] and the 2 remaining ones have running services and working config [20:31:06] AaronSchulz: when i access the saame DB:// thingy from different wikis, i should get the same thing, no? [20:31:15] yes, probably the same ones are usd [20:31:21] see the blobs.sql table [20:31:34] AaronSchulz: could you comment on bug 42825 ? [20:31:38] If it is per DB and the id is just an autoincrement column [20:31:38] what section is test2 wiki on? [20:31:48] is the entire idea wrong? or is it just a setup issue? [20:32:00] would be nice to use a proper key/value store :) [20:32:11] * AaronSchulz has a list of recurring things he complains about [20:32:14] brion: 11:55 <+nagios-wm> RECOVERY - HTTPS on ssl1004 is OK: OK - Certificate will expire on 07/19/2016 06:51. [20:32:18] 11:55 < mutante> !log replaced wikipedia SSL cert with new DigiCert (rather than RapidSSL) (RT-3639) [20:32:49] aude: according to the toolserver, on s3 [20:32:53] so there may have been a period of a few minutes where 1004 wasnt running.. one like half an hour where ssl4 was depooled..but no esams issues afaik [20:32:58] so test2 and wikidata are actually on the same shard [20:33:03] that should make things *easier* [20:33:07] still doesn't work :/ [20:33:38] hmmm [20:36:05] Salvete :) Right. I don't know what that did, but I suspect it's a local issue with firefox, possibly my profile is corrupt. I can't hit *any* website which requires an SSL connection, not just wp. This is since I booted about 7.00PM cet [20:36:49] BarkingFish: ok.. thanks for reporting that. we did indeed change the SSL cert we are using on WP earlier, but i did not see any issues with it [20:36:49] I'm gonna reset firefox right back to factory. I can get onto https:// websites with konqueror and chromium, so it's gotta be local to firefox. [20:37:08] fuuuuunky [20:37:53] brion: the difference people would see and that is the only expected thing is that it is now a different cert being served.. by DigiCert, and not by RapidSSL [20:38:18] Well, at least I have my evening meal to plow through while I'm fixing this. Moroccan spiced tuna, tomato and chickpea stew with french bread :) [20:38:23] yeah shouldn't make a big difference [20:38:28] i blame gremlins on BarkingFish's firefox :) [20:38:39] He should just you chrome :D [20:38:44] s/you/use/ [20:38:56] no, he will report how RT and Bugzilla and Etherpad are broken then :) [20:39:19] i blame everything on firefox, brion. It's about as useful as a chocolate sunhat. [20:39:37] mmmm tasty [20:39:39] but it's better than konqueror and chromium for playing bejeweled blitz on facebook :) [20:39:42] heard a dozen reports that dont work with Chrome but with Firefox too :) [20:39:45] <^demon> Does firefox still use more ram than eclipse? [20:40:06] ^demon, not as far as I am aware [20:40:12] ram usage should be way down since last year or so, but web sites may be horribly terrible :) [20:40:36] <^demon> I haven't used firefox in some time, always curious. [20:40:36] ^demon: http://media.bestofmicro.com/9/C/285456/original/wbgp3singletab.jpg ! [20:40:49] I can have firefox and up to about 15 to 20 other apps open on here before my pc starts to slow - 3.2ghz single core processor, 2gb ram, 500mb swap [20:41:51] <^demon> brion: When you got the retina, did you get 8gb or 16gb ram? [20:41:52] Yeah if you want no ram just open NetBeans... [20:42:01] ^demon: 8gb, it's been plenty [20:42:22] <^demon> 16g is sexy. Allows me to run vms with actual sane levels of ram allocated to them :p [20:42:43] <^demon> I can run an ubuntu vm with 8g allocated and still have 8g leftover for osx. [20:42:53] <^demon> I've been spoiled. [20:44:08] 16gb of ram, i7 8 core cpu and a 512 ssd makes your laptop smexy [20:44:17] It's also rather heavy.... damn battery [20:45:18] <^demon> It's lighter than my 13in mbp was. [20:45:34] righty ho - it WAS my firefox profile... I've reset firefox, and started it again, and I can get all https as i normally should. [20:45:50] Next question for me is: what the bejeezus corrupted my profile? [20:47:18] BarkingFish: probably firefox did :P [20:48:34] Maybe the BobingDog did [21:04:58] <^demon> testwiki and test2wiki are both on 1.21wmf6. Can we get some people to give either a whirl before we push to mediawikiwiki and wikidatawiki? [21:05:21] aude: want to look? --^ [21:06:11] looks good to me [21:06:17] looks fine to me at a first glance [21:06:19] * aude missing the lang links, of course [21:06:24] yea [21:06:25] but that's another issue [21:06:28] that still remains to be debugged [21:07:10] ^demon: do you have time tomorrow to do some debugging on test2/wikidata with us? [21:07:13] * aude may update the wikidata item [21:07:14] or maybe Reedy could [21:07:22] <^demon> DanielK_WMDE_: Yes, we can do that. [21:07:27] thanks [21:07:29] cool [21:07:39] any idea when? [21:07:54] i'll try adding some external store to my test wiki to see how it all works [21:07:54] 14:00 UTC or something? [21:08:03] <^demon> We'll meet up in the morning again, when I get online. [21:08:10] aude: you are brave :P [21:08:29] <^demon> DanielK_WMDE_: Yeah, at the latest. I'll shoot for 13:00. [21:08:50] ok thanks! [21:08:56] DanielK_WMDE_: not too much more setup than i already have (e.g. central auth) [21:09:08] it's nice to have one login for all my test wikis :D [21:09:55] ^demon: what's the procedure for getting an extension renamed? [21:10:06] Cf. https://gerrit.wikimedia.org/r/#/c/37610/ [21:10:28] <^demon> Renaming a repo? Can't do it. [21:10:41] Wow. Interesting. [21:10:43] * aude thinks it has to be a new repo [21:10:55] <^demon> You can make a new repository, copy all the history, and then mark the old one as read only (to one day be deleted). [21:10:58] and abandon the old one [21:11:01] <^demon> But you'll lose the review history and such. [21:11:05] I see. [21:11:26] <^demon> StevenW: http://xkcd.com/910/ comes to mind ;-) [21:11:48] Sometimes I wonder if the people who made Gerrit ever actually managed a repository. [21:12:07] <^demon> Indeed, they do. [21:12:21] <^demon> (They use gerrit to develop gerrit, running master usually) [21:12:43] ^demon: that's one repo. we got hundreds :P [21:13:15] <^demon> DanielK_WMDE_: We're a medium sized gerrit install. There's more than a handful of installs with *thousands* of repos and groups and users. [21:13:37] huh, i wonder how they are holding up :) [21:17:07] <^demon> DanielK_WMDE_: I'm going to go ahead and do wikidatawiki now. I'm not seeing any errors from test[2]wiki [21:17:22] ^demon: looks good to me and say go ahead [21:17:25] ok [21:18:36] <^demon> Done. [21:19:07] yay [21:21:46] <^demon> Seems to have been a smooth deploy :) [21:21:54] making new items works on wikidata [21:22:00] everything looks fine to me [21:22:07] <^demon> Spoke too soon. [21:22:08] <^demon> "Fatal error: Call to a member function getEntity() on a non-object at /usr/local/apache/common-local/php-1.21wmf5/extensions/Wikibase/repo/includes/actions/HistoryEntityAction.php on line 38" [21:22:46] it's still calling wmf5? [21:22:56] do we need to kick stuff? [21:23:01] i think we fixed that bug [21:23:14] Excuse me folks, is this the best place to ask about trying to test/setup Parsoid and VisualEditor on a mediawiki instance? [21:23:40] probably not [21:23:51] <^demon> aude: Might've been something still suck in cache. [21:23:58] <^demon> I tried the same url, no error. [21:24:02] ok [21:24:12] i'm 99% sure we saw it and it got fixed [21:24:16] <^demon> Taliaraven: Try #wikimedia-dev, most of the VE folks hang out there. [21:25:31] https://bugzilla.wikimedia.org/show_bug.cgi?id=42768 [21:25:45] https://gerrit.wikimedia.org/r/#/c/37206/ [21:27:23] ^demon: that patch is in the branch [21:27:43] <^demon> Yeah, looks like it :) [21:29:59] *phew* [21:31:13] everything looks uneventful [21:31:45] <^demon> Indeed. I believe I'm going to go afk for a few minutes. [21:32:58] ok [23:07:16] okay, fundrasing deploying ahead of a 5p pst downtime for hw maintenance [23:10:20] !log synchronized payments to 72c1985a2b9b1db [23:10:27] Logged the message, Master [23:27:34] Reedy: is JSHint not liking 1.21wmf6? [23:27:41] No idea [23:27:45] Not been involved today, sorry [23:27:57] ah, sorry, who should I ask? [23:28:17] There's no Hashar... [23:28:22] So Krinkle or ^demon|away [23:28:30] thanks [23:28:38] pgehres: link? [23:28:46] https://gerrit.wikimedia.org/r/#/c/37966/ [23:28:57] http://integration.mediawiki.org/ci/job/mediawiki-core-lint/1355/console : FAILURE [23:29:23] Seems like a generic failure, not mediawiki related [23:29:33] yeah [23:29:38] the jshint process itself failed [23:29:44] so I just need to beat jenkins to merging [23:29:45] no js errors [23:30:17] thanks, just wanted to make sure I wasn't crazy [23:30:18] maybe not [23:31:01] well, it won't let me sumbit that change [23:31:53] I know, but I mean, you may not have to beat it if I can fix it. [23:32:07] aha, now I know what you mean [23:32:19] hashar forgot to test his change, --cwd instead of -cwd [23:35:49] weird [23:38:07] okay, i am going to go ahead and scap 1.21wmf5 since I am running up on a deadline and i don't care about wmf6 until wed [23:38:14] ok [23:41:43] pgehres: I figured it out, it has to do with the broken StartProfiler.php file symlink in the directory [23:41:50] Reedy: Any idea what that's doing there in wmf6? [23:41:57] Yes [23:42:01] There is a broken symlink from StartProfiler.php -> ../StartProfiler.php [23:42:02] It's been broken in every version before [23:42:10] ? [23:42:18] ah [23:42:29] need to be ../wmf-config/StartProfiler.php [23:42:47] that'd still be broken [23:42:51] for jenkins [23:43:18] does that symlink have to be in the repo? can't that be done from commonsettings somewhere? [23:43:31] Oh [23:43:32] lrwxrwxrwx 1 demon wikidev 28 2012-12-10 19:19 StartProfiler.php -> wmf-config/StartProfiler.php [23:43:48] How could you symlink via a settings file? :/ [23:44:03] well, there is no $wgMustHaveSymlink [23:44:10] the symlink is means to an end [23:44:20] can "it" be done from settings? [23:44:23] no [23:44:31] the file has to be in that place? [23:44:43] startprofiler needs to be in the root [23:44:59] and the symlink has to be in the repo? [23:45:17] For the WMF deployment branch [23:45:23] yeah [23:45:33] seems an odd setup, since it isn't pointing to anything [23:45:36] it's in .gitignore... [23:45:40] for a local dev set up [23:45:51] so on my localhost it'd be failing [23:46:01] Looks like ^demon|away made git ignore the gitignore [23:48:40] <^demon|away> Indeed, I did ;-) [23:49:01] <^demon|away> `git add -f ` will add the file even if gitignore says nowai. [23:49:26] Maybe it should be moved to checkoutMediaWiki like AdminSettings [23:49:53] Wait, it's already in there [23:49:53] # Create symlink to wmf-config/StartProfiler.php... [23:49:53] $path = "$destIP/StartProfiler.php"; [23:49:53] $link = "../wmf-config/StartProfiler.php"; [23:49:53] createSymlink( $path, $link, "Created StartProfiler.php symlink." ); [23:50:01] why do we have both.... [23:50:16] <^demon|away> Cuz I didn't know checkoutMediaWiki already did that? ;-) [23:52:55] I've filed a bug against jshint to make it not fatally throw on unresolvable symlinks. [23:53:05] Meanwhile, could it be removed from the repo :) [23:53:17] I'm not sure how to remove it without it dissapearing on deployment. [23:53:55] <^demon|away> Do it in git, pull, re-create on fenari :) [23:54:05] mmm [23:54:05] Right [23:54:11] Might break testwiki [23:54:14] who cares [23:54:43] <^demon|away> For like, 5 seconds while you fix it? [23:54:43] <^demon|away> lol [23:54:54] oh feck [23:55:06] wmf/1.21wmf5 doesn't have it [23:55:13] Yes [23:55:19] Because Chad changed stuff today [23:55:52] <^demon|away> Yeah, we had a deploy ;-) [23:56:04] jfyi in case it matters - I am syncing l10n cache for wmf5 [23:56:09] <^demon|away> testwiki, test2wiki, mediawikiwiki, wikidatawiki -> 1.21wmf6 [23:57:22] Fixed [23:58:25] https://integration.mediawiki.org/ci/job/mediawiki-core-lint/1366/console [23:58:28] indeed [23:58:40] pgehres: "jenkins +1" https://gerrit.wikimedia.org/r/#/c/37966/ [23:58:57] Krinkle: thank you much :-) [23:59:02] pgehres: rebase all the commits [23:59:16] * Krinkle did so already [23:59:26] all 1 commits