[00:25:28] robla: are you guys doing a deployment tomorrow? [00:25:47] asking for fundraising. they are planning a test [00:52:03] Ryan_Lane: no [00:52:25] 16th is the next one [00:58:28] Reedy: thanks [03:08:34] petan: ping [03:31:17] If you or a MediaWiki developer you know is interested in improving watchlists with grouping and usability enhancements, please let me know. I have submitted a clear and practical project proposal for the 2012 Google Summer of Code and I am seeking a mentor. I will, of course, do all of the heavy lifting (coding) to make it happen. All I need from a mentor is knowledgeable guidance and occasional assistance with debugging. [03:31:56] Potential mentors can view the project here: http://www.mediawiki.org/wiki/User:Blackjack48/GSOC_proposal_for_watchlist_improvements [03:34:27] I realize it is very short notice but I've had some trouble finding a mentor. I really think MediaWiki could benefit from my proposed improvements. Feedback is appreciated too - thanks. [05:31:21] blackjack48: Hi. [05:31:39] I was just about to post to the mailing list about you. [05:31:43] But now you're here. [05:31:45] So I don't have to. [05:31:50] You should post to the mailing list. [05:31:52] wikitech-l [05:32:09] https://lists.wikimedia.org/mailman/listinfo/wikitech-l [05:32:17] I did, however I think the post fell a bit back. [05:32:27] So, I'll post again [05:32:45] Did you talk to Sumana? [05:32:50] I think this is what she does all day. [05:33:03] She should be able to find you someone to be a mentor. [05:33:26] Yes, I've been in touch with her and she seems very interested in my proposal. [05:33:35] She can't find you someone? [05:33:54] How married are you to this idea? [05:34:45] I've already submitted the application but I can change details if the community has ideas. [05:35:11] However, I need to find a mentor this week if the idea is to live on. [05:35:19] I was just thinking that if you weren't very attached to it, there aer other projects that Wikimedia would be intersted in. [05:35:29] are [05:35:34] But it's a fine idea. [05:35:39] Better watchlists would be great. [05:35:49] There are plenty of bugs and brainstorming pages on mediawiki.org, I think. [05:35:55] If you haven't found them already. [05:36:09] Watchlist grouping in particular has been endlessly discussed, I think. [05:36:16] Anyway, Sumana should be able to find you someone. [05:36:28] Did you ask her specifically? What was her respose? [05:36:32] response [05:37:22] I'm going to shoot her a quick email, asking if she's found anyone possibly interested. [05:38:48] Okay. [05:39:01] She should be able to find someone. [05:39:11] There are like a million tech staffers now. [05:39:21] Though some are already signed up. [05:39:28] How much hand-holding do you need? [05:39:56] Not much - I learn fast. My work style is outlined briefly in my application. [05:40:09] But basically I check in frequently [05:40:23] and ask for help when I need it. [05:40:31] Frequent check-ins usually mean more work for mentors. ;-) [05:41:50] Sumana is trying to encourage better communication this year... so I don't know whether checking frequently is good or bad :/ [05:42:32] http://www.mediawiki.org/wiki/Summer_of_Code_2012/management [05:51:31] Checking-in is probably good. [05:51:43] Most past GSoC projects haven't amounted to anything. [05:51:51] I'm not sure if they tell you that in the application process. [05:52:40] I don't have exact stats on that, but the general rule is that the larger the project, the more likely it is to fail. People have a tendency to bite off more than they can chew, I think. [05:53:14] No, but I /really/ want the watchlist improvements to amount to something... even if it becomes more of a long term project after GSoC. [05:53:38] (I do realize it'd be a sizable project.) [05:58:18] Anyway thanks for the support, Joan. I'll be logging out shortly. [07:46:09] varnent: . [07:46:13] hi [07:48:48] petan: greetings - how difficult would it be to add support for outreach and some other project wikis to the wm-bot @recentchanges feature? [07:53:30] very easy [07:53:43] it needs to be reported to wikimedia feed [07:53:59] if that wiki has own reports there I can set it up now [07:55:33] where is the easiest place to check that [07:55:44] irc.wikimedia.org [07:56:00] let me check [07:56:19] yeah - it doesn't appear they're in the bot's code or on the IRC page's list [07:57:36] there is a channel for that [07:57:44] however I don't know if works [07:57:58] list these project somewhere and I will check it [08:00:22] I checked in here - http://svn.wikimedia.org/svnroot/mediawiki/trunk/tools/wmib/RClogs.cs - looks like commons, outreach, meta and maybe labsconsole would be helpful additions? we were specifically thinking of a commons and outreach page when this came up [08:03:35] varnent: http://pastebin.com/T5nubaKg [08:03:46] there is config page, links aren't inserted to source code anymore [08:03:50] config file [08:04:19] expand it with wikis you want and I will update it on labs [08:04:29] great - ty :) [08:04:41] however labswiki doesn't have rc feed [08:05:40] it's configured by hand by Ryan and it doesn't use standart setup [08:05:53] makes sense [08:09:08] does it work http://pastebin.com/T5nubaKg [08:09:15] just expand this page [08:19:34] hi [08:26:59] petan: http://pastebin.com/U1TjsAus [08:46:23] varnent: ok [08:47:09] petan: thank you! :) [08:49:24] @recentchanges+ outreach_wiki [08:49:24] Wiki inserted [08:49:27] :o [08:49:28] loading... [08:49:31] ok seems to work [08:49:48] @recentchanges+ commons [08:49:48] Wiki inserted [08:49:58] @recentchanges+ wikitech_wiki [08:49:58] Wiki inserted [08:50:39] fantastic - thank you! :) [08:51:06] you need to set up feed for a list of various pages for it to work [08:51:34] doing that now - working great [08:51:50] ok, let me know if it didn't work [08:56:21] petan: hmm - seems to be having issues on initial tests - adds the items okay - or says it does - but doesn't post changes to the new entries [08:58:21] @RC+ outreach_wiki User:Varnent/Sandbox [08:58:22] Permission denied [08:58:30] oh - ha - right - that's channel specific :) [09:07:16] @RC+ outreach_wiki User:Varnent/Sandbox [09:07:16] Inserted new item to feed of changes [09:08:07] hm... [09:08:24] are you sure there is any activity in that channel on wikimedia irc [09:08:51] @RC+ en_wikipedia User:Petrb/Sandbox1 [09:08:51] Inserted new item to feed of changes [09:09:14] wikimedia-outreach? I believe so - we're looking at using it in wikimedia-lgbt as well [09:09:23] Change on 12en_wikipedia a page User:Petrb/Sandbox1 was modified, changed by Petrb link https://en.wikipedia.org/w/index.php?diff=486973838 edit summary: blah blah [09:09:56] I guess it doesn't work for outreach, eg nothing is reporting changes to that channel on wikimedia server [09:10:39] had the same problem with commons [09:10:41] when you join #outreach.wikipedia on wikimedia you see any changes there? [09:15:34] ok I found it [09:15:35] it's wikipedia, not wikimedia :) [09:15:37] outreach.wikiP [09:15:38] yeah - odd label for that one [09:15:39] Change on 12en_wikipedia a page User:Petrb/Sandbox1 was modified, changed by Petrb link https://en.wikipedia.org/w/index.php?diff=486974480 edit summary: [[WP:AES|←]]Blanked the page [09:15:46] hm, but still doesn't seem to work [09:15:52] @recentchanges+ outreach_wiki [09:15:52] Wiki inserted [09:15:54] @recentchanges+ outreach_wiki [09:15:54] This channel is already watched [09:15:58] ah [09:15:59] that's it [09:16:19] there must be something wrong [09:16:38] it's showing up in the wikimedia irc channels [09:20:26] configuration needed fix [09:20:27] @recentchanges+ commons [09:20:27] This channel is already watched [09:20:28] @recentchanges+ outreach_wiki [09:20:28] This channel is already watched [09:20:30] @RC+ outreach_wiki User:Varnent/Sandbox [09:20:30] Inserted new item to feed of changes [09:20:40] Change on 12outreach_wiki a page User:Varnent/Sandbox was modified, changed by Petrb link https://outreach.wikimedia.org/w/index.php?diff=?diff=33183 edit summary: [09:20:46] works [09:20:50] excellent! [09:21:04] you need to insert it again in your channel [09:21:50] wikitech doesn't exist [09:21:53] so I removed it [09:22:02] ok [09:22:15] it would be nice to have it in feed though [09:22:48] Change on 12outreach_wiki a page User:Varnent/Sandbox was modified, changed by Varnent link https://outreach.wikimedia.org/w/index.php?diff=?diff=33185 edit summary: Blanked the page [09:22:53] yay but link is fucked a bit [09:23:00] https://outreach.wikimedia.org/w/index.php?diff=?diff=33185 [09:25:00] Change on 12outreach_wiki a page User:Varnent/Sandbox was modified, changed by Petrb link https://outreach.wikimedia.org/w/index.php?diff=33189 edit summary: [09:25:27] that's better [09:25:28] there we go :) [09:27:06] works well in the other channel - but not the same luck with commons [09:27:25] @recentchanges- commons [09:27:25] Wiki deleted [09:27:29] @recentchanges+ commons [09:27:29] Wiki inserted [09:27:47] @RC+ commons User:Petrb [09:27:47] Inserted new item to feed of changes [09:28:20] indeed [09:28:42] channel is commons.wikimedia [09:30:27] Change on 12commons a page User:Petrb was modified, changed by Petrb link https://commons.wikimedia.org/w/index.php?diff=69698743 edit summary: [09:30:29] wm2012 works fine [09:30:38] it works on commons too [09:40:59] ty for setting up outreach and others [09:41:08] not sure what's up with commons..but it works somewhat for now :) [09:41:18] I will try to debug it a bit [09:41:26] ty [09:41:32] sorry I gotta dash [09:42:05] right [11:26:59] New patchset: Hashar; "target to take MediaWiki screenshots" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/4752 [11:28:31] New review: Hashar; "> Where is `${builddir}` defined? Is that a ./tmp directory created by ant? Or is that the directory..." [integration/jenkins] (master); V: 0 C: -2; - https://gerrit.wikimedia.org/r/4752 [11:30:40] New review: Hashar; "Patchset 3:" [integration/jenkins] (master); V: 0 C: -2; - https://gerrit.wikimedia.org/r/4752 [12:59:44] vvv: Hey there [12:59:53] vvv: I noticed your nightly toolserver page is breaking a little bit [12:59:57] Krinkle: hi [12:59:59] do you want help in using resoruceloader on it ? [13:00:09] Krinkle: you mean ~vvv/nightly? [13:00:20] I haven't gitified it yet [13:00:25] yeah [13:00:25] or other tools if you use the same css there [13:00:45] I used to fetch MW CSS/JS [13:02:51] yeah, but as of mediawiki 1.17 (so for a while now), mediawiki uses resoruceloader for css/js [13:02:57] not direct .css files anymore [13:03:14] which means they have version numbers in the url and thus change after each version or even more often if you try to link them directly [13:03:35] all you have to do is instead of embedding vector/screen.css embed the startup module and use mw.loader.load [13:03:45] can work on a toolserver page as well, doesn't need mediawiki :) [13:09:11] vvv: Oh, it's not citified yet [13:09:18] gitified* [13:09:22] gitified? [13:09:22] the documentation suggested that it is [13:09:36] Well [13:09:36] I was citing your sentence [13:10:24] mw [[Download]] says somewhere "Download from Git master or [[tools:vvv/nightly|nightly]" [13:10:25] I suppose that was kept from the svn version, without checking [13:11:13] vvv: wow, ./skins-1.5/monobook/main.css [13:11:30] Yeas [13:11:32] Monobook [13:11:59] That's an old and venerable tool [13:12:31] well, as a quick fix you could use this [13:12:32] https://www.mediawiki.org/w/skins-1.20wmf1/monobook/main.css [13:12:40] instead of http://mediawiki.org/skins-1.5/monobook/main.css [13:13:04] even // relative if you like [13:13:15] so that people don't get warnings when using https toolserver [13:16:36] I still have to gitify it [13:22:32] vvv: Do you know how to do it? Or is it just finding the right time [13:22:50] Well, I assume it should be simple [13:36:55] vvv: ok :) [13:37:10] vvv: btw, there is an issue with using git to clone wikimedia repos on toolserver [13:37:19] What issue? [13:37:24] toolserver has their ssl certificates fucked up again [13:37:28] incomplete paths [13:37:40] so it won't let you do it unless you import the CA manually from firefox or something [13:37:47] Ive bugged them about it but no response so far [13:38:11] since gerrit is only available over https it needs that [13:38:26] <^demon> Just clone over ssh? [13:38:30] https (or ssh, but ssh requires private key and that's not recommended and would still need the certificate) [13:39:09] afaik gerrit uses https for anonymous and ssh for authorized, I'm not sure if ssh requires the CA but it would require the ssh key to be on toolserver [13:44:46] <^demon> You could make a separate public/private key pair rather than putting your normal key on TS ;-) [13:45:05] <^demon> Although anonymous clones are nicer and less scary :) [13:45:17] yeah I know, but that's not something I'd like to document on mw.org [13:45:42] "if anonymous checkout isn't working, please sign up for a Gerrit account and create a separate ssh key for it so that attackers can only commit stuff on your name to gerrit" [13:46:03] :P [13:46:06] <^demon> There's also supposedly a way you can clone over https with a user/password for auth, but I haven't tested it really well yet. [13:46:28] which reminds me, I'll bump that bug [13:46:46] still blocking my toolserver tool from automatically doing git stuff [13:47:19] tools like wmfCodeSearch [13:47:33] (aka wikimedia-svn-search but I had to rename it) [14:46:50] https://www.mediawiki.org/w/index.php?title=Extension:Drafts&oldid=524368&debug=true [14:48:58] <^demon> Magic? [14:49:00] <^demon> Heh [14:51:29] re; http://lists.wikimedia.org/pipermail/wikitech-l/2012-April/059989.html [15:04:25] <^demon> Nikerabbit: Can you commit changes to core? I've got some extra debugging things in place and I've already eliminated 2 of the possible points of failure. [15:12:23] New patchset: Hashar; "Install MediaWiki for testswarm usage" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/4752 [15:12:45] New review: Hashar; "(no comment)" [integration/jenkins] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/4752 [15:12:47] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/4752 [15:18:07] hashar: Hm.. 2 questions [15:18:28] hashar: backend uses mediawiki-git fronted says ./mw/master which is correct? [15:18:44] I'll ask in -operations where the convo is [15:24:46] New review: Krinkle; "(no comment)" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/4752 [15:29:10] New patchset: Hashar; "rm MediaWiki jobs that used subversion" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/4788 [15:30:29] New review: Hashar; "(no comment)" [integration/jenkins] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/4788 [15:30:31] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/4788 [15:30:50] ^demon: FYI I have deleted from Jenkins the Mediawiki jobs that used subversion ( change 4788 ) ^^ [15:30:57] <^demon> ok. [15:31:06] Krenair: ^ [15:44:23] (ok) [16:36:27] ^demon: in about 30 minutes [16:37:07] <^demon> Ok, just ping me then. [16:50:07] wikimedia-e3 [16:50:09] whoops [17:03:00] I can't find a create account from the API. Does anyone know if it is there? [17:04:41] i don't believe it's in there [17:08:47] <^demon> Probably because our account creation API is so awful ;-) [17:09:20] * RoanKattouw looks around for a Fabrice [17:11:41] Morning fabriceflorin [17:12:46] Hi RoanKattouw, auroraerose, rsterbin: Sorry for being a bit late, had a SCRUM meeting with the E2 team that just wrapped up. What did I miss? [17:13:26] Not much, just me looking for you [17:14:31] OK, then. So what do you think of the deployment plan I sent you all by email? Does it seem reasonable? Or should we aim to do less for today's deployment? [17:15:25] hi preilly, how are you today? [17:16:08] fabriceflorin: Looks good although I don't know (or maybe don't remember) what Dario's stuff is [17:16:20] preilly: I have a question for you [17:17:55] hi folks [17:18:07] Cool. To sum up our AFT5 plan for the record, here's what's on our plate for today: [17:18:07] • the two revisions from Elizabeth which we couldn't deploy last time: [17:18:09] • the new icons on the feedback page (courtesy of Reha?) [17:18:10] • the upgrade of emails to oversighters to the main mediawiki email system (RT ticket per our emails with Ryan) [17:18:11] • the switch of oversighter emails to their main email account (a simple variable switch, once we confirm speedier delivery) [17:18:12] • Dario's two requests for AFT5 and AFT4, if completed and tested (Dario, can you give us an update on that?) [17:19:26] RoanKattouw, I will dig up the email thread with Dario to figure out what he wanted from us. He said he would be in the office late today, so don't know if we can get him in the loop in time, but will try. [17:19:42] ^demon: ping [17:19:44] hi robla [17:19:53] <^demon> Nikerabbit: Pong. You ready? [17:19:54] robla: looking for Aaron, waiting for preilly [17:19:59] ^demon: export running [17:20:05] takes some minutes [17:20:24] sumanah: how may I help you? [17:20:57] auroraerose and rsterbin, is there anything special that's needed on your end for deployment? Or should we simply wait for RoanKattouw to deploy your new revisions, then test them on production? [17:20:59] hey preilly! so, not sure whether you saw the recent thread on the developers' list about extensions review -- did you? (checking whether you have the background) [17:21:49] Sorry, I meant to say auroraeosrose -- I always forget to add the 'os' part ;o) [17:22:50] * robla pings Aaron irl [17:23:03] sumanah: I didn't read it [17:23:03] fabriceflorin: You can tab-complete nicknames, you know [17:23:11] !tabcompletion | fabriceflorin [17:23:11] fabriceflorin: Instead of manually typing another person's nickname in IRC, you can type the beginning of their name and hit TAB on your keyboard to get it autocompleted, like on the command line. This avoids misspellings. Just make sure it's the right person! [17:23:24] preilly: understood. ok, let me sum up [17:23:32] AaronSchulz: hi! 20% checkin day - already have plans, or need stuff? [17:23:41] hehe [17:23:53] * ^demon adds another ignore rule for mw-bot. [17:23:59] preilly: Tim suggested that it's reasonable & efficient & possibly more pleasant for people to have actual 20% projects [17:24:10] preilly: instead of just doing a bunch of one-off patch reviews of patches by volunteers [17:24:49] preilly: do you think you would prefer to do something like working with a specific community-written extension to review it for possible deployment (which might take multiple 20% days), or continue with 1-off patch review of contributions from volunteers? [17:24:51] small db changed - there is an sql addition (one field) and the script to update the counts fabriceflorin and RoanKattouw [17:24:52] * AaronSchulz can look through extension commits [17:25:03] I've been tailing core all week [17:25:18] AaronSchulz: "tailing"? sorry, don't understand [17:25:21] Yeah [17:25:23] Thanks RoanKattouw and sumanah, what a cool feature! Really appreciate it ;o) [17:25:28] sumanah: I'm not sure [17:25:36] sumanah: I'd need to think about it a bit [17:25:36] sumanah: refreshing and reviewing stuff [17:25:51] preilly: makes sense. ok, so in the interim, here are some patches to consider reviewing today: [17:25:59] preilly: https://gerrit.wikimedia.org/r/#change,4540 [17:26:02] RoanKattouw: I need your help today at 4 pm [17:26:04] preilly: https://gerrit.wikimedia.org/r/#change,4550 [17:26:10] preilly: https://gerrit.wikimedia.org/r/#change,4746 [17:26:27] and preilly are you comfortable with UploadWizard code review? [17:26:29] RoanKattouw: I've got to push a new version of both MF and ZRMA to production [17:26:59] Thanks, auroraeosrose, I appreciate your update about the db change. Is there anything else on your end we need to watch out for? [17:27:00] AaronSchulz: it would also be nice if you could help with reviewing a few patches that have been submitted to BZ [17:27:07] AaronSchulz: that queue is still growing [17:27:15] fabriceflorin: nope, everything else should be good [17:27:17] AaronSchulz: or even just taking a bunch of the most recent ones and sticking them into Gerrit branches [17:28:17] preilly: to help you think about whether you'd prefer a larger 20% project, check out https://www.mediawiki.org/wiki/Review_queue and https://www.mediawiki.org/wiki/Writing_an_extension_for_deployment -- we're trying to be more systematic in doing both user experience and technical/code reviews of community-written projects [17:28:56] RoanKattouw: did https://gerrit.wikimedia.org/r/#change,4601 make into production branch? [17:28:59] preilly: you'd be doing either initial technical reviews of proposed architectures/features, or code reviews of extensions [17:29:07] preilly: let me know your thoughts. [17:29:38] AaronSchulz: does this work for you? https://bugzilla.wikimedia.org/buglist.cgi?cmdtype=runnamed&namedcmd=MW%20extensions%20with%20patches%20that%20need%20review&list_id=106944 [17:29:55] AaronSchulz: Don't know offhand but I can check [17:29:59] AaronSchulz: I shared that search in BZ. 51 patches awaiting review re extensions we deploy [17:30:02] 'There is no saved search named 'MW extensions with patches that need review'. ' [17:30:08] so, I guess no ;) [17:30:29] * AaronSchulz justs uses the link from the sidebar [17:30:30] AaronSchulz: :/ go into your user preferences/shared searches and you can checkbox that one, shared by me [17:30:32] ok [17:31:09] Yoni just emailed me and said that Reha will be available shortly, to look over the deployment of the icons, so we can minimize the risk of binary files (icon images) being mis-deployed (or having the wrong mime-type etc.) [17:31:17] AaronSchulz: 2 for AntiSpoof, 3 for CentralAuth, 4 for ConfirmEdit... [17:32:07] ^demon: okay, export done [17:32:31] <^demon> Ok, fire away. [17:33:08] Thanks auroraeosrose, glad we're all set on your end. RoanKattouw, let us know what you need from us and when you have the code ready for testing. Is anything special needed for moving over the icons for Reha's revision? [17:33:19] AaronSchulz: you see those? [17:33:26] ^demon: done [17:33:35] <^demon> A-ha! I've got it. [17:33:50] AaronSchulz: git tells me it's in there yes [17:33:56] preilly: so maybe next week (when I'm in town) you can let me know what kind of 20% experience you would prefer and I can help get you that :) [17:34:03] preilly: sound good? [17:34:07] sumanah: hi, hope you are doing well. I'm working on a bug, how can I know if there's already a patch for it? [17:34:12] hi davidpardz [17:34:25] sumanah: sure [17:34:27] davidpardz: usually, if you are looking at a bug in Bugzilla and there is already a patch for it, there are a few signals. [17:34:32] thanks preilly, thanks AaronSchulz [17:34:35] RoanKattouw: ok [17:34:38] davidpardz: example: https://bugzilla.wikimedia.org/show_bug.cgi?id=28850 [17:34:43] davidpardz: take a look at that [17:34:47] it would suck if that regressed again [17:35:05] AaronSchulz: For reference I did git checkout wmf/1.20wmf1; git log | grep e82c2cc7eef7d0d58101f43b5fdba289b2c3967e [17:35:16] davidpardz: under all the metadata at the top, you see an attachment, and it's labeled as "(733 bytes, patch) " [17:35:36] davidpardz: also, someone's added the keywords "need-review, patch" to indicate that this is a patch awaiting code review [17:36:05] fabriceflorin: Thanks for reminding me of the icons, I do need to do special things for those [17:36:20] davidpardz: those are accurate like 90-99% of the time, but just to make absolutely sure you are not wasting effort, you scroll down in the bug comments and you read them and see whether anyone says "here is my patch LINK" or something like that [17:37:05] davidpardz: because sometimes, especially recently, the patch is in Gerrit, and not attached to the bug. [17:37:10] davidpardz: does that help & make sense? [17:37:17] davidpardz: so, if you show me a specific bug, I can help you check it [17:37:40] RoanKattouw: My pleasure. When rsterbin gets back, she can provide more info about the icons, but it should be pretty straightforward. If need be, I can send you a Zip file with the PNGs from heatherw. [17:38:09] sumanah: got it, Thank you! yes that makes sense. I think I can still looking at the bug :) - https://bugzilla.wikimedia.org/show_bug.cgi?id=35702 [17:38:33] No worries [17:38:34] I have the files [17:38:42] I just need to explicitly copy them because they're non-text files [17:39:17] davidpardz: there you go. So, you see that there is the bug description + 1 comment, and neither of them mentions nor links to a patch. [17:39:18] <^demon> Nikerabbit: https://gerrit.wikimedia.org/r/#change,4792 should fix it for the last time. [17:39:23] RoanKattouw: can't you do git show --status-only or something [17:39:27] meh, whatever [17:39:40] davidpardz: No patch is attached in the Attachments section, and the Keywords don't indicate any patch is attached, [17:39:51] davidpardz: so I think this one is clear for you to work on. [17:40:03] * AaronSchulz struggles to get gnome3 working properly [17:40:44] sumanah: thank you, yeap, I've been working on the bug but I can't still figure it out [17:40:51] ^demon: need ryan to merge? [17:40:56] davidpardz: UploadWizard doesn't have that many experts who REVIEW patches. As you can see here https://bugzilla.wikimedia.org/buglist.cgi?keywords=patch%2C%20need-review%2C%20&keywords_type=allwords&list_id=106953&resolution=---&query_format=advanced&component=UploadWizard&product=MediaWiki%20extensions there are a few patches that have been awaiting review for a little while. [17:41:10] davidpardz: what's the specific problem you're having, if you can articulate it? [17:41:21] <^demon> Nikerabbit: Or any ops person. I'm going to run out and get some lunch, maybe I'll be back and it'll be successfully merged :) [17:41:43] * Nikerabbit pokes ops [17:42:32] RoanKattouw: Thanks, glad you've got this under control. Give us a shout if you need anything else from us. [17:43:35] Nikerabbit: Ryan just walked in [17:45:00] davidpardz: i am quite familiar with the code, what is the problem you are facing ? [17:45:35] drecodeam: if you have a bit of time to review these patches https://bugzilla.wikimedia.org/buglist.cgi?keywords=patch%2C%20need-review%2C%20&keywords_type=allwords&list_id=106953&resolution=---&query_format=advanced&component=UploadWizard&product=MediaWiki%20extensions and tell the authors what might need fixing, that would be great [17:45:56] drecodeam: https://www.mediawiki.org/wiki/Code_review_guide#Review_checklist is the checklist to follow [17:46:30] sumanah: one of them is mine, i will check the other one. [17:46:48] sumanah: also i have submitted 4 patches on Gerrit, can i only go about reviewing them ? [17:46:58] sumanah: I see your point, and thak you for your help. It has to do with the deeds but It seems like drecodeam can help, thanks again [17:47:08] davidpardz: glad to help [17:47:35] hi [17:47:36] drecodeam: you should ask Ryan Kaldari whether he might be able to review them (probably tomorrow) [17:48:09] sumanah: ya i would do so. [17:48:10] drecodeam: the purpose of code review is to get someone else's eyes checking your code for errors and familiarizing themselves with what you are changing [17:48:30] drecodeam: within Gerrit, when you look at your changeset, you see the "add reviewer" field? [17:48:47] drecodeam: put in Ryan Kaldari's email address, and he'll be notified by email that you want him to review it [17:49:33] sumanah: ya i will do that, although he did tell me that he would look at them as soon as he gets some time. I think he is already aware of it [17:49:33] Also, drecodeam, if you have submitted the patch you mentioned in https://bugzilla.wikimedia.org/show_bug.cgi?id=32247 into Gerrit, please link to the changeset in a comment on that bug [17:50:08] sumanah: no i have not, actually i would remove that patch, because it needs much more work then i imagined when i worked on it. [17:50:10] Hi yoni_omniti ! RoanKattouw is getting things ready now for AFT5 deployment to testing. He knows about the icons and is uploading them separately. auroraeosrose told him about the SQL change to the db. So I think we are all set for the items listed on my email plan -- exept that I'm still checking to see what DarTar wanted us to do this week. Will report back on that once I locate his last email about this. [17:50:25] drecodeam: ok. you can do that by looking at the details for the patch and checking the Obsolete checkbox [17:50:41] thanks, fabriceflorin! [17:50:47] drecodeam: congratulations on learning -- seeing what you did wrong means you're a lot more skilled than you were when you wrote the patch originally :D [17:52:34] fabriceflorin, auroraeosrose, yoni_omniti, rsterbin: OK we should be live on testwiki now [17:52:43] ok [17:52:48] yoni_omniti: My pleasure. Sorry for the misunderstanding about today's deployment. I want us to get in the habit of deploying weekly now, even if it's small revisions, so we don't get further behind on our backlog. When rsterbin joins this chat, we can discuss any open issues which you guys might have, in lieu of our Skype call. [17:53:09] ok [17:53:20] fabriceflorin: i've been here; can you not see me? [17:53:25] sumanah: true that. I had to ask you one more thing. I wanted to start writing documentation for Upload Wizard now that i am getting comfortable with the code, but i am new to it? any starting points or process i should know of ? [17:53:52] drecodeam: a few things, yes [17:54:28] drecodeam: hi, the problem is with the deeds, actually I'm debugging createInputs function [17:54:28] drecodeam: first: you know that the code is commented [17:55:06] drecodeam: for now I gotta go, (school duties) hope see you around [17:55:21] davidpardz: cool, ping me if you are around [17:55:29] drecodeam: | sumanah thank you both [17:55:40] glad to help davidpardz! [17:55:54] drecodeam: the comments also lead to autogenerated documentation [17:56:10] Thanks, RoanKattouw, I am testing all this now. I believe the correct link for the feedback page is: http://test.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow [17:56:11] davidpardz: glad to help ! [17:56:26] sumanah: ya, it does. Helped me a lot in understanding the code. [17:56:42] drecodeam: so, commenting the code in the same fashion is a fantastic way to improve documentation [17:58:28] sumanah:and then i should submit the changes as patches ? [17:58:54] rsterbin: Reha, I'm sorry, I saw you were there, but didn't hear from you, so I thought you were away. Any questions on your end? [17:59:26] drecodeam: sure [17:59:30] sorry, phone call, back in a moment [17:59:42] yoni_omniti rsterbin auroraeosrose : The new features on testing seem to be working for me on Safari. Do they work for you too? [18:00:29] so far so good... [18:01:29] I tested the feedback page both in oversight mode and in logged out mode, and they application works as intended. The permalink is visible to oversighters, and the Go back button from permalink is corectly redirecting to the whole feedback page, for both oversighters and logged-out users. [18:04:18] looks good to me [18:07:34] Ryan_Lane: are you aware of https://gerrit.wikimedia.org/r/#change,4792 yet? [18:07:37] fabriceflorin: dario knows the fix for https://bugzilla.wikimedia.org/show_bug.cgi?id=35590 isn't going up in this release, right? [18:07:45] unless someone has told me, no [18:09:11] Thanks, auroraeosrose -- yoni_omniti and rsterbin, is there any testing issues on your end? Or are we ready to ask RoanKattouw to deploy to production? [18:09:32] i don't see any [18:10:18] Hi, rsterbin, what is causing the hold back of DarTar's ticket? Is it that we didn't have time to test it on prototype? or that RoanKattouw hasn't reviewed the code yet? Or both? https://bugzilla.wikimedia.org/show_bug.cgi?id=35590 [18:10:37] Nikerabbit: I'm confused as to how this helps anything [18:10:51] it removes two log messages and adds an email address to the bot name? [18:10:57] isn't this just going to break? [18:11:13] the username is used for ssh [18:11:19] this is definitely going to break [18:11:34] Ryan_Lane: I guess the email is needed to match the commit properly, but ^demon knows why he did that after he comes back from lucnh [18:12:12] <^demon> Is it also used for ssh? [18:12:14] <^demon> Did I do that? [18:12:16] yes [18:12:20] this change is bad. [18:12:29] * ^demon will fix [18:13:14] you need to fix quite a bit to make this work [18:13:16] RoanKattouw, how are you coming along with implementing the switch to a faster email delivery system on media wiki, per the RT ticket we discussed with Ryan_Lane? https://rt.wikimedia.org/Ticket/Display.html?id=2724 Is this something you think could happen today? [18:13:30] Yes, probably today [18:13:36] <^demon> Ryan_Lane: Waiiittt, no it shouldn't. l10nuser is not used for ssh, that's gerrituser. [18:13:47] <^demon> l10nuser is *only* used to match against project.uploader [18:14:08] you sure? [18:14:19] <^demon> I just grepped, shall I pastebin it for you? [18:14:26] I'll check [18:14:27] RoanKattouw: Cool. Should we first deploy the current revisions from testing to production? Then tackle the email switch? [18:18:50] Yes [18:19:59] RoanKattouw: OK, sounds good. yoni_omniti, rsterbin, auroraeosrose are we good to go to production now, from your standpoint? [18:20:13] yep [18:20:19] yep [18:20:58] Okey dokey. It's all yours, RoanKattouw ! [18:21:36] OK here we go [18:29:27] Not deploying yet, hang on [18:29:35] Only just figured out how to do the submodule update thingy [18:29:59] Reedy: WTF: fatal: reference is not a tree: 87938c1908a553cbbcf0237404c78bed5b2f7513 [18:30:01] Unable to checkout '87938c1908a553cbbcf0237404c78bed5b2f7513' in submodule path 'extensions/AbuseFilter' [18:31:53] ? [18:32:07] Reedy: That's what I get when I run git submodule init && git submodule update [18:32:10] on wmf/1.20wmf1 [18:32:41] I don't [18:32:48] reedy@fenari:/home/wikipedia/common/php-1.20wmf1$ git submodule init && git submodule update [18:32:48] reedy@fenari:/home/wikipedia/common/php-1.20wmf1$ [18:34:31] <^demon> You shouldn't need to re-run init unless you're adding new submodules. [18:34:58] ^demon: Initial checkout [18:35:50] <^demon> Ah [18:36:00] <^demon> Well, that shouldn't happen ;-) [18:36:26] <^demon> http://stackoverflow.com/questions/2155887/git-submodule-head-reference-is-not-a-tree-error [18:36:37] Hey rsterbin, any idea why RoanKattouw is having problems with this? Unable to checkout '87938c1908a553cbbcf0237404c78bed5b2f7513' in submodule path 'extensions/AbuseFilter' yoni_omniti auroraeosrose [18:36:50] OK deploying now [18:36:52] fabriceflorin: Not their problem [18:37:05] fabriceflorin: This is ^demon 's or Reedy 's fault [18:37:11] i have no idea whatsoever. [18:37:11] <^demon> Not my fault. [18:37:15] I blame git [18:37:16] LOL [18:37:23] git - gits fault [18:37:25] Gits giving the error, not me ;) [18:37:31] Thanks, RoanKattouw, glad this is nothing on our end. We'll check production when you tell us it's ready for testing. [18:39:00] Oh God, it's gonna do TWO localization cache rebuilds now [18:39:05] This will take a while [18:39:28] * yoni_omniti sleeps [18:39:44] <^demon> Ryan_Lane: How's that patch look? [18:40:06] Presumably, depending on what is actually up with Git, the only thing I can presume is it didn't like how I pulled the rev into AF [18:40:39] * Reedy shrugs [18:40:41] OK, thanks. On the plus side, it's better to learn all these Git-related issues now, with a small deployment, rather than in the middle of a major push ;o) [18:40:45] I don't proclaim to know anything about git :p [18:41:17] <^demon> The first step is making sure 87938c1 is actually in AF, which it should be. [18:41:22] <^demon> Unless Reedy made a typo. [18:42:12] OH [18:42:15] ^demon: it's not merged [18:42:19] I wonder if that's fucking it over [18:42:25] Probably [18:42:26] <^demon> Well duh. [18:42:34] * ^demon whacks Reedy with a Clue-by-4 [18:42:35] If you git review -d 'ed it locally it'd work [18:42:38] https://gerrit.wikimedia.org/r/#change,4708 [18:43:45] Getting review is hard when no one is around [18:51:19] ^demon: which patch? [18:51:34] <^demon> gerrit hooks you said looked wrong but you'd look at them again. [18:51:42] <^demon> https://gerrit.wikimedia.org/r/#change,4792 [18:53:06] gimme a sec, then [18:53:50] ah [18:53:51] I see [18:55:19] ^demon: merged [18:55:23] you want a force puppet run? [18:57:24] <^demon> Ryan_Lane: If you wanna pull https://gerrit.wikimedia.org/r/#q,4796,n,z in too, we'll go ahead and force a puppet run. [18:57:39] <^demon> 4796 makes "bug 123" and "I54bb..." clickable. [19:00:31] fabriceflorin, rsterbin, auroraeosrose, yoni_omniti: Deployment done [19:00:40] (sigh, that took forever) [19:00:56] thx [19:01:44] thanks [19:01:50] cache? [19:02:55] RoanKattouw: all the messaging is redone now? [19:03:07] weeee [19:03:09] "messaging redone"? [19:03:20] new translation messages available ;) [19:03:28] Oh heh [19:03:33] Yeah they should be I think [19:03:38] Let me know if you see anything weird [19:04:40] <^demon> Ryan_Lane: I'm guessing gerrit's starting up? [19:06:39] probably, I forced a puppet run [19:07:28] it's back up [19:08:28] Thanks, RoanKattouw. The feedback page on production looks good to me: http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow - I tested it as an oversighter, monitor and logged-out user. [19:09:08] <^demon> Ryan_Lane: Thanks, everything looks great. [19:09:13] sweet [19:09:15] yoni_omniti, auroraeosrose and rsterbin : does this work for you on production as well? [19:09:21] yep [19:09:23] works for me [19:09:47] looks good, checked the log as well, messages are correct there [19:10:02] yay [19:10:36] Cool. auroraeosrose, thanks for checking the log. RoanKattouw, nice work! [19:11:45] clicktracking? [19:14:35] yoni_omniti: I don't know if there is any click tracking involved in today's deployment. I think we were not able to release DarTar's request today, though I'm still waiting to hear from rsterbin whether that was because we didn't get a chance to test it on prototype or because RoanKattouw didn't have time to review the code. [19:14:56] ok [19:15:13] I saw some related code come in yesterday [19:15:16] But I didn't review it [19:15:17] i checked that in yesterday — it would have been unreasonable to expect Roan to get to it [19:15:33] OK, looks like this went quite well [19:15:37] woot [19:15:38] I'm gonna get lunch now [19:15:43] enjoy [19:15:49] Thanks, rsterbin, that's what I thought, but wanted to make sure we were clear on this. [19:16:27] Yes, RoanKattouw_away, I'll go grab lunch too. Maybe I'll see you at Working Girls? [19:16:45] If there are no further issues with today's deployment, I think that OmniTI developers can be released now, because I RoanKattouw and I can handle the switch to a faster email system and subsequent switch of email address from aft@wikimedia.org to the oversighter email. [19:18:02] Nice work, you guys! I'm glad we did this push, because it means we can start getting oversighters in the loop, which we couldn't do until we got auroraeosrose's fixes in -- and having the icons on production will also make everyone feel happier. ;o) [19:19:33] yoni_omniti: do we need to schedule a call or chat later today, or are you guys OK until Monday's weekly call? (I am on vacation Friday). [19:19:52] fabriceflorin: i think we're ok until monday [19:20:29] Cool. I'm off to lunch now, and can continue this conversation via email. Over and out for now! [19:58:45] I wonder how we have non utf-8 compliant crap in the imagelinks tables [19:58:55] example: [19:59:12] http://de.wikipedia.org/wiki/Benutzer:Brickbat this bad file link turns into [19:59:26] | 1791721 | Http://www.jorgemata.info/jap�n/Edwin-Valero.jpg | [19:59:31] none too good [19:59:33] ^demon: so suddenly we have links in commit messages :-D [19:59:48] ^demon: https://gerrit.wikimedia.org/r/#change,4797 has link to bug and change-id \O/ [19:59:59] <^demon> :) [20:00:05] <^demon> You're welcome. [20:02:24] ^demon: so was the patch okay? [20:02:41] <^demon> I believe so. We'll find out when you try again :) [20:02:53] ^demon: should I try right now? [20:03:21] <^demon> Go for it. [20:03:27] core or extension? [20:03:35] <^demon> Try core again. [20:03:54] let's see if there are any new translations yet [20:11:57] ^demon: hmmm looks like the removal of automatic rebasing broke my scripts [20:12:35] <^demon> Arrrggghhh. [20:13:27] and looks like there isn't command line option to force it [20:14:03] <^demon> Nikerabbit: I mean we could put a `git rebase origin/master` before the `git review` [20:14:12] ^demon: yeah I can do that [20:15:22] ^demon: does rebase have to be after or before git commit? [20:15:59] <^demon> After commit, before review. [20:22:51] ^demon: that didn't work [20:23:00] git review still wants to submit to patches [20:23:21] I have: git commit -a -v -e -m "$COMMITMSG" && git rebase origin/master && git review -t l10n || : [20:23:36] s/to/two/ [20:24:23] <^demon> Lemme think. [20:26:34] 7ee06f2 (HEAD, master) Localisation updates from http://translatewiki.net. [20:26:38] c37aa67 (origin/master, origin/HEAD) Localisation updates from http://translatewiki.net. [20:27:18] c37 is merged already [20:28:14] <^demon> I'm confused why it would want to try pushing that again if it's already in the tree :\ [20:28:30] OK WTF [20:28:32] same here [20:28:39] https://gerrit.wikimedia.org/r/#change,4803 [20:28:40] no idea what is going on [20:28:40] Nikerabbit, checkout origin/master and cherry-pick it? [20:28:48] Why the hell are the tests broken *on 1.20wmf1* ?!? [20:28:58] Platonides: it=? [20:29:03] <^demon> RoanKattouw: All the tests are broken on non-master [20:29:05] Nikerabbit, ^demon: Run git fetch gerrit and that message should go away [20:29:08] I'm at master tracking origin/master [20:29:11] ^demon: Why? [20:29:18] <^demon> Ask Antoine. [20:29:27] * hashar hides [20:29:52] yeah 1 second timeout again [20:29:52] https://integration.mediawiki.org/ci/job/MediaWiki-Tests-Misc/378/console [20:30:15] RoanKattouw: do I need to always run that command? [20:30:32] what was auroraeosrose doing here? [20:30:45] Nikerabbit: It's not important, its only real function is to make git-review not complain. You can bypass that message with the -y flag too [20:30:55] Platonides: She works for OmniTI, on the ArticleFeedbackv5 extension [20:31:07] wow, I didn't know [20:31:18] I know about her as a php developer [20:31:20] RoanKattouw: have a look at git show d664ffd | https://gerrit.wikimedia.org/r/#q,I2f5613b0c0215023c413dec8e84804175d8c8d6e,n,z [20:31:28] RoanKattouw: we need to raise the timeout a bit [20:31:34] RoanKattouw: but wont that resubmit commits already submitted? [20:31:40] or find a nice way to disable the timeout :-)) [20:31:47] (and merged!) [20:31:49] Nikerabbit: No it won't [20:32:02] I guess I should fix this in git-review [20:32:15] <^demon> hashar: 1s is just too low. I think 2s for low will eliminate most of the false positives. [20:32:34] indeed, that is why I have kept it and put it to 2s [20:32:38] Is there a reason this happens for non-master branches specifically [20:32:41] or is this just a fluke? [20:32:54] also this afternoon (EU time) I have had to delete millions and millions of files on gallium [20:32:59] using 16 threads of rm -fR [20:33:17] so that surely caused tests to timeout randomly when accessing disk [20:33:38] OK [20:33:53] should probably have sent an email to wikitech-l =/ [20:34:07] How can I make it rerun the tests for that rev? [20:35:10] you have to log in Jenkins [20:35:26] then browse to that build page. On the left is a "Retrigger" link [20:35:50] cherry pick of phpunit timeout for 1.20wmf1 : https://gerrit.wikimedia.org/r/#q,4805,n,z [20:35:52] hashar, have you tried moving the sqlite db to /dev/shm ? [20:36:13] that should make them faster, as the fsync() would no longer be going to disk [20:36:17] Thanks [20:37:03] Platonides: maybe we could turn off fsync() in sqlite :-D [20:39:17] Platonides: anyway, I will probably have the test to run with MySQL by default [20:39:29] sqlite / postgre being run asynchronously as a second pass job [20:40:26] bahh stil failing https://integration.mediawiki.org/ci/job/MediaWiki-Tests-Parser/599/console [20:40:40] 1) ParserTests::testParserTest with data set #0 ('Blank input', '', '', '', '') [20:40:41] [exec] PHP_Invoker_TimeoutException: Execution aborted after 2 seconds [20:40:44] preilly: OK I approved your commit, overriding Jenkins [20:40:48] Sorry about the delay there [20:41:04] preilly: Now, to deploy, you would run svn up php-1.19/extensions/{your,stuff} as usual [20:41:26] preilly: and then for the git part, it's cd php-1.20wmf1; git submodule update extensions/MobileFrontend I believe [20:41:37] I am heading bed now. Take care [20:41:59] hashar: good night [20:42:27] RoanKattouw: okay [20:45:42] preilly: (and please do verify it worked correctly; you're kind of my guinea pig) [20:46:13] RoanKattouw: okay [20:46:37] ^demon: do we maintain a fork on github ? [20:46:49] <^demon> Not yet, no. We need to setup replication. [20:46:59] thanks [20:48:04] while true; do sleep 60; git pull gerrit/master && git push github --mirror; done; [20:49:15] hashar, it's really easy to remove fsync() in sqlite [20:49:19] but you need to recompile it [20:49:28] which then would lead to recompile the php module [20:49:37] I don't think you want to go that way [20:50:02] in-memory partitions teh rock [20:51:32] I actually tried making SQLite tests run on an in-meory DB (new DatabaseSqliteStandalone( ':memory:' )), but it actually worked a bit slower than from HDD [20:54:42] Platonides: /dev/shm could probably work [20:54:50] not sure how much RAM will be needed though [20:55:11] anyway, I will have to find out a way to disable that timeout, or maybe raise the limit yet again [20:55:21] or make all tests @medium (aka 10sec timeout) [20:55:49] I am sleeping now [20:55:51] bye bye [21:23:59] Who just broke: PHP Parse error: syntax error, unexpected T_STRING, expecting ')' in /home/wikipedia/common/wmf-config/InitialiseSettings.php on line 2173 [21:25:08] <^demon> Works for me? sync-file shouldn't have sync'd it if there's a syntax error. [21:25:18] Reedy: Was that you? [21:25:37] ^demon: it was only on the local f [21:25:37] i [21:25:38] le [21:25:42] system [21:25:57] sorry for the multiline spam [21:26:16] <^demon> Oh, the clone from gerrit you mean? [21:27:05] ^demon: on fenari [21:27:33] <^demon> I just did php -l and it said no syntax errors. [21:27:59] Yeah, cause I fixed them [21:28:03] working with weird charsets [21:28:47] Reedy: thanks, for fixing it [23:08:51] Tim-away: ping [23:12:19] hello [23:15:43] TimStarling: I just ran the scap command [23:15:46] sorry for missing the checkin earlier...was anyone here for it? [23:15:54] TimStarling: and I get a good deal of errors [23:16:10] TimStarling: see http://pastebin.mozilla.org/1566435 [23:17:10] yeah, mergeMessagesFiles.php gives a lot of notices [23:17:17] we should probably disable notices while it runs, at least [23:17:48] the way it works is fairly hackish, but for some reason it worked just fine in 1.18 and throws tons of warnings in 1.19 [23:18:36] the include() warnings from 1.20 probably need to be fixed [23:20:07] the PrefStats extension seems to be missing [23:20:28] TimStarling: so, do I need to be worried about this scap run [23:20:32] that is ^demon's responsibility I guess [23:20:46] <^demon> I thought we pulled PrefStats in? [23:21:16] preilly: at worst, only mediawiki.org will be broken [23:21:22] so it's not a big deal [23:21:28] <^demon> Ah, it wasn't. [23:21:31] <^demon> I'll fix that. [23:21:32] and nothing newly broken, it would have been broken before [23:25:38] <^demon> Ugh, I can't checkout 1.20wmf1 [23:25:50] <^demon> checkout -b wmf/1.20wmf1 --track origin/wmf/1.20wmf1 [23:25:50] <^demon> fatal: Not a git repository: /opt/local/apache2/htdocs/mediawiki/core/.git/modules/extensions/AbuseFilter [23:26:00] <^demon> Reedy: ^ [23:26:37] TimStarling: scap is still running for me [23:26:56] TimStarling: it's still outputting srv243: MediaWiki 1.19: Compiling texvc...ok [23:26:57] srv243: MediaWiki 1.20wmf1: Compiling texvc...ok [23:26:58] srv243: Done [23:26:59] srv248: Copying to srv248...ok [23:27:02] etc, etc.... [23:27:42] mark's idea was to replace scap's use of rsync with git [23:27:49] maybe we will do that soon, that would be much faster [23:28:08] TimStarling: okay, well it's been running for awhile [23:28:10] it is kind of amusing how close the feature list we wrote for a scap replacement is to the feature list of git [23:28:29] What uses PrefStats? [23:28:55] I pulled in PrefSwitch as SimpleSurvey blindly includes it... [23:29:51] It should just be dropped from ExtensionMessages-1.20wmf1 [23:35:37] ExtensionMessages-1.20wmf1.php is automatically generated, you can't just copy it and edit it [23:36:03] it even says so in the header [23:37:49] nobody run scap for a bit [23:38:45] I am running it with ExtensionMessages-1.20wmf1.php deleted [23:42:15] scap doesn't have set -e, so it just continues even when mergeMessageFileList.php fails [23:58:51] it's strange, I don't know where the PrefStats.i18n.php could have come from