[00:03:32] <TimStarling>	 looks good
[00:03:42] <TimStarling>	 I updated the scribunto extension page, btw
[00:05:44] <sumanah>	 "Scribunto" sounds like a command that Hermione Granger might use :) (it's just because it's Latin I think)
[00:05:56] <sumanah>	 Yay dox! Thanks.
[00:08:08] <Krinkle>	 Or a new distro of Ubuntu (that was my first thought, but after a quick google I found the Latin definition too)
[00:08:35] <sumanah>	 Krinkle: Tim added the "and this is why it's called that" explanation to the extension page
[00:09:09] <Krinkle>	 Ubuntu, Edubuntu, Kubuntu, Lubuntu, Xubuntu, .. Scribunto ?
[00:09:36] <TimStarling>	 if you pronounced it correctly it would sound like screebuntaw
[00:10:05] <TimStarling>	 which sounds a bit different to ubuntu
[00:10:18] <sumanah>	 It's a cool name
[00:10:46] <TimStarling>	 it's nice that it's a short name and a mission statement in one
[00:10:58] <TimStarling>	 I guess that's why latin is so popular for mottos, brevity
[00:11:48] <sumanah>	 and fancy-sounding!
[00:12:16] <vvv>	 And it also translates nices into English, because English does not have verb forms, and "they will write" sounds cool
[00:12:29] <vvv>	 (it does not translates that nicely into Russian)
[00:12:30] <sumanah>	 is it imperative or descriptive?
[00:12:41] <sumanah>	 Oni pyesyatcya! or something
[00:12:49] * sumanah  does not remember her Russian

[00:13:05] <vvv>	 [04:12:36] <sumanah> is it imperative or descriptive? — well, that depends upon how you pronounce it
[00:13:54] <sumanah>	 vvv: remind me what the situation is wrt the Berlin hackathon? there are 19 tickets/seats left
[00:14:01] <sumanah>	 vvv: I mean, regarding whether you're coming
[00:14:31] <vvv>	 Well, I will have to travel there with my old passport, since my new one will not be ready till mid-May
[00:14:36] <vvv>	 I will most probably come
[00:14:46] <vvv>	 By the way, I wanted to ask you about a venue
[00:15:23] <sumanah>	 go ahead and ask
[00:15:45] <sumanah>	 vvv: also, if you will need financial assistance to come to the Berlin hackathon, please register ASAP, preferably by 1 May
[00:16:04] <vvv>	 sumanah: where's the venue located?
[00:16:26] <sumanah>	 vvv: It's nearly settled but not certain; can I tell you early next week?
[00:16:33] <vvv>	 Yes, please
[00:16:49] <vvv>	 I need to book accomodation ASAP
[00:16:51] <sumanah>	 vvv: I'll put it on the hackathon page and in other relevant locations as soon as I know.
[00:17:01] <sumanah>	 vvv: something vaguely near the s-bahn will do, right?
[00:17:16] <vvv>	 Well, yes
[00:17:22] <vvv>	 Wait
[00:17:40] <vvv>	 Do you mean near any S-Bahn, or near any specific S-Bahn station?
[00:17:45] <RoanKattouw>	 Doesn't WMDE provide arranged accommodation this year?
[00:18:28] <sumanah>	 I presume that you want to arrange your own hotel for some reason, vvv?
[00:18:42] <vvv>	 Yes
[00:18:46] <sumanah>	 vvv: or would you prefer to have us book a space for you in the same hostel as everyone else?
[00:18:47] <sumanah>	 ok
[00:21:08] <sumanah>	 vvv: it is likely that we will be near Potsdamer Platz.
[00:21:11] <sumanah>	 back in a bit
[00:21:18] <RoanKattouw>	 TrevorParscal: Ha, getLengthDifference() is also unused, and, unsurprisingly, also broken
[00:21:22] <RoanKattouw>	 (replace doesn't update it)
[00:21:46] <TrevorParscal>	 lol
[00:47:16] <jcrowgey>	 Hi there, we recently upgraded from wikimedia 1.12.0 to 1.15.5-8 and i'm having some trouble getting the right output for <style> elements, can anyone advise?
[00:47:30] <jcrowgey>	 in the old version, pages get output with a <style> element that's being generated by index.php, but in the new version no such style element is output.  can anyone point me to documentation for index.php?action=raw&gen=css
[00:47:56] <jcrowgey>	 the style declaration im missing is output dynamically by index.php with some particular arguments.
[00:51:19] <RoanKattouw>	 Maybe the function you were using (addStyleSomething) was deprecated or removed or something?
[00:59:25] <TimStarling>	 the 1.15 branch point was 3 years ago
[01:00:21] <TimStarling>	 it's difficult to remember details about it
[01:04:22] <preilly>	 RoanKattouw: have you talked to awjr recently
[01:04:29] <RoanKattouw>	 I have not
[01:04:51] <preilly>	 RoanKattouw: how can we cherry pick a change to the wmf-1.20wmf1 branch?
[01:05:04] <RoanKattouw>	 [[wikitech:How to deploy code]] documents all of this
[01:05:26] <RoanKattouw>	 I think the relevant section is https://wikitech.wikimedia.org/view/How_to_deploy_code#Case_1b:_extension_changes
[01:05:46] <preilly>	 RoanKattouw: The site's security certificate is not trusted!
[01:05:52] <RoanKattouw>	 Yeah, known problem
[01:05:56] <RoanKattouw>	 Sorry for the https link
[01:06:02] <RoanKattouw>	 http://wikitech.wikimedia.org/view/How_to_deploy_code#Case_1b:_extension_changes
[01:09:49] <preilly>	 RoanKattouw: do we need to set-up a special deployment branch?
[01:10:16] <preilly>	 RoanKattouw: can we just cherrypick the change we want in to the existing submodule?
[01:12:11] <awjr>	 hello
[01:12:15] <RoanKattouw>	 Yes, you should set up a new deployment branch
[01:12:26] <preilly>	 awjr: see that reply?
[01:12:30] <awjr>	 yep
[01:12:43] <preilly>	 RoanKattouw: but, we don't have rights to git push origin wmf/1.20wmf1
[01:12:49] <RoanKattouw>	 Use git review instead
[01:13:06] <awjr>	 RoanKattouw: the proper tracking branch would be wmf/1.20wmf1, yes?
[01:13:08] <RoanKattouw>	 The docs are a bit wrong still because I haven't fixed reality to match the docs completely
[01:13:10] <RoanKattouw>	 Yes
[01:13:22] <awjr>	 RoanKattouw: also are there wikis still running 1.19?
[01:13:26] <RoanKattouw>	 Yes there are
[01:13:28] <RoanKattouw>	 Lots of them
[01:13:29] <awjr>	 k
[01:13:33] <RoanKattouw>	 See /h/w/common/wikiversions.dat
[01:13:34] <preilly>	 awjr: most of them
[01:13:41] <RoanKattouw>	 OK I'm heading out for drinks with Ironholds et al
[01:13:49] <RoanKattouw>	 If you need me, I have gtalk on my phone
[01:13:53] <RoanKattouw>	 firstname dot lastname at gmail
[01:13:55] <preilly>	 RoanKattouw: argh
[01:14:21] <RoanKattouw>	 (Sorry, they're waiting for me)
[01:14:35] <awjr>	 RoanKattouw before you go!
[01:14:35] <awjr>	 awjrichards@virtualbot:~/Dev/wikimedia_git/extensions/MobileFrontend$ [master]$ git branch --track wmf/1.20wmf1 origin/wmf/1.20wmf1
[01:14:36] <awjr>	 fatal: Not a valid object name: 'origin/wmf/1.20wmf1'.
[01:14:46] <RoanKattouw>	 Yeah you didn't create it yet
[01:14:54] <RoanKattouw>	 Wait
[01:14:58] <awjr>	 ah need to create it in gerrit?
[01:15:01] <RoanKattouw>	 Did you cd to your *core* checkout?
[01:15:13] <awjr>	 RoanKattouw no
[01:15:19] <awjr>	 but the docs say to do it in the extension checkout
[01:15:21] <RoanKattouw>	 Which instructions are you following
[01:15:27] <RoanKattouw>	 Are you following [edit] Creating an extension deployment branch  ?
[01:15:34] <awjr>	 oh shit i was reading the wrong section
[01:15:51] <RoanKattouw>	 The "[edit] Now that your extension has a deployment branch " section?
[01:15:55] <RoanKattouw>	 I named it that for a reason :)
[01:16:03] <awjr>	 :p
[01:17:04] <RoanKattouw>	 OK, heading out now
[01:17:13] <RoanKattouw>	 If you need more help, hit me up on gtalk (details above)))
[01:56:54] <TimStarling>	 where is the gerrit user list?
[01:59:52] <Krinkle>	 TimStarling: There isn't any afaik, its in LDAP
[02:00:13] <TimStarling>	 that's no excuse
[02:00:35] <Krinkle>	 There is https://gerrit.wikimedia.org/r/#admin,group,11
[02:00:40] <Krinkle>	 and such
[02:00:52] <TimStarling>	 there was no user list for subversion either, until I made one
[02:01:21] <Krinkle>	 well, we no longer have USERINFO so..
[02:01:37] <TimStarling>	 I tried https://gerrit.wikimedia.org/r/#admin,group,3 but there's no list there
[02:01:59] <Krinkle>	 yeah, it's a magic one, just like ops https://gerrit.wikimedia.org/r/#admin,group,7
[02:02:08] <Krinkle>	 external group defined in LDAP
[02:02:16] <TimStarling>	 actually users.php works by parsing /etc/passwd
[02:02:40] <TimStarling>	 USERINFO is just for links and real names
[02:02:56] <Krinkle>	 TimStarling: Still? I thought that was changed, but either would work (/etc/passwd would be  more omelet of course)
[02:02:57] <Krinkle>	 complete*
[02:03:19] <Krinkle>	 I suppose any script hosted inside there could get the data from ldap and make a public listing.
[02:03:24] <TimStarling>	 sure, it might have been changed
[02:03:34] <TimStarling>	 my point is that the lack of USERINFO is not really an excuse either
[02:04:05] <Krinkle>	 I did a similar thing on the Toolserver once to get a nice public view of who is in one group, can be easily fetched from php. labs console might have it already
[02:04:07] * Krinkle  checking

[02:06:00] <Krinkle>	 nope, labs console only has a full user list, https://labsconsole.wikimedia.org/wiki/Special:ListUsers, which is in a way complete since all accounts there are made through ldap,
[02:06:25] <Krinkle>	 would be nice to have a way (like Github has) to generate lists of all users with access, and all contributors (the latter including authors that don't have access, but who's commit was merged in)
[02:06:30] <Krinkle>	 per repo
[02:07:44] <Krinkle>	 the latter wouldn't even require having access to ldap
[02:07:47] <TimStarling>	 no grafzahl
[02:08:07] <Krinkle>	 http://www.commandlinefu.com/commands/view/4519/list-all-authors-of-a-particular-git-project http://stackoverflow.com/questions/6482436/list-of-authors-in-git-since-a-given-commit no idea how that performs on medawiki/core.git
[02:08:40] <Krinkle>	 too bad all svn commits were imported with fake author addresses, would be nice if that was mapped to the superseding git account where possible
[02:09:08] <TimStarling>	 there is an Alexander K. but I thought gerrit usernames were linked to LDAP cn
[02:09:11] <TimStarling>	 which is grafzahl
[02:09:21] <Krinkle>	 grafzahl?
[02:09:52] <Krinkle>	 I suppose you don't mean the german character Grafzahl of the german Sesame Street ?
[02:10:13] <TimStarling>	 possibly a distant relative
[02:10:25] <Krinkle>	 Oh, its a user name :D
[02:10:36] <TimStarling>	 [[Git/Conversion/Extensions_queue]] asks for the maintainer's gerrit username
[02:11:09] <TimStarling>	 so I did an ldapsearch on formey and he has both uid and cn set to grafzahl
[02:11:10] <Krinkle>	 What is a gerrit username anyway, emailaddress? or git author "name"
[02:11:30] <TimStarling>	 but there's no grafzahl on that labsconsole user list
[02:11:43] <TimStarling>	 like I say, I think it is cn, i.e. common name
[02:12:16] <Krinkle>	 right, labsconsolewiki accounts are named after the shell account names, not the "real names", which is what Gerrit mostly uses
[02:12:26] <TimStarling>	 yours is "Krinkle", i.e. same as your uid except with an initial capital letter
[02:12:32] <Krinkle>	 yeah
[02:12:42] <Krinkle>	 but my Git author name is Timo Tijhof
[02:13:00] <Krinkle>	 which is also my gerrit username (although I always log in with my emailaddress)
[02:13:19] <TimStarling>	 I think the git author name is arbitrary
[02:13:27] <TimStarling>	 some of my commits have "Tim Starling" and some have "tstarling"
[02:13:36] <Krinkle>	 Hm.. strange
[02:13:40] <Krinkle>	 I can't commit otherwise
[02:13:41] <TimStarling>	 I probably forgot to set some environment variable
[02:13:48] <Krinkle>	 my default git config is set to  Krinkle
[02:14:10] <Krinkle>	 but in wm-git repo clones I have to set it locally to my full name, otherwise it won't let me commit
[02:14:17] <Krinkle>	 ah, its the emailaddress, not the name
[02:14:30] <Krinkle>	 Krinkle <krinklemail@gmail.com> and Timo Tijhof <ttijhof@wikimedia.org> are a set
[02:14:38] <Krinkle>	 its the mail that does it, of course
[02:14:50] <TimStarling>	 your gerrit username is Timo Tijhof?
[02:15:15] <Krinkle>	 all I know is my shell name is Krinkle and for Git I use Timo Tijhof <ttijhof@wikimedia.org>
[02:15:17] <TimStarling>	 so I guess I have to hack the gerrit database to work out the username, in that case
[02:15:30] <TimStarling>	 because there's nothing in LDAP for you that says Timo Tijhof
[02:15:37] <Krinkle>	 So gerrit has its own username thingy?
[02:15:41] <TimStarling>	 guess so
[02:16:00] <Krinkle>	 https://gerrit.wikimedia.org/r/#settings,contact
[02:16:07] <Krinkle>	 Hm.. it says krinkle Krinkle, nowhere Timo Tijhof
[02:16:07] <TimStarling>	 I mean, yes it does, I hacked it once before
[02:16:20] <TimStarling>	 to fix my own account
[02:16:23] <Krinkle>	 hehe
[02:16:47] <Krinkle>	 using a central LDAP, only to have Gerrit keep its own little thing
[02:19:23] <TimStarling>	 your gerrit account: http://paste.tstarling.com/p/lFxrFZ.html
[02:20:15] <TimStarling>	 mysql> select * from account_external_ids where external_id like '%grafzahl%';
[02:20:16] <TimStarling>	 Empty set (0.00 sec)
[02:20:43] <TimStarling>	 anyway this seems excessively complicated just to migrate an extension
[02:21:13] <Krinkle>	 so, just so we're clear. why would "grafzahl" exist? Are you sure he has an account? I can't find grafzahl or Alexander K. around git related stuff
[02:21:27] <TimStarling>	 he doesn't have an account, hence the lack of results above
[02:21:31] <Krinkle>	 right
[02:21:53] <Krinkle>	 http://svn.wikimedia.org/viewvc/mediawiki/USERINFO/grafzahl?view=co
[02:56:49] <IL|afk>	 hello! i'm trying to fiddle with a usabilty/accessabiliy gadget at enWS. basically, I want to make an image on the page greyscale, but i can't access the image data since it is a cross-domain request. is there any way to use something like CORS on WMF projects to get the image data over?
[03:01:39] <TimStarling>	 no
[03:02:59] <inductiveload>	 TimStarling: good to know, thanks
[03:03:18] <inductiveload>	 at least i can stop banging my head on the desk, it has a dent already
[03:03:21] <TimStarling>	 there's a bug for it if you're interested
[03:04:03] <inductiveload>	 sure :-)
[03:04:31] <TimStarling>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=28700
[03:06:22] <inductiveload>	 thanks
[05:54:56] <davidpardz>	 sumanah hi, hope you are fine. it's quite late probably you are busy. But I'd like to know if someone got interested on my proposal ? =/
[05:55:25] <sumanah>	 davidpardz: it is late and I am indeed busy.  I infer from what you are saying that no mentor has contacted you with interest. Is that correct?
[05:55:51] <davidpardz>	 sumanah It is
[05:58:24] <sumanah>	 davidpardz: I have heard nothing from any mentor who would be interested in mentoring your proposal.
[05:58:48] <davidpardz>	 sumanah thank you
[05:58:56] <sumanah>	 davidpardz: Did you see all my emails to wikitech-l (the mailing list) where I talked about how unlikely it is that mentors would be interested in 2 proposals covering the exact same topic?
[05:59:29] <sumanah>	 davidpardz: I'm a little confused as to why you chose to concentrate on the topic you did, given that you saw that at least one other student had already gotten started on it
[06:01:38] <sumanah>	 davidpardz: I'm sorry, but I have to go -- it is 11pm here and I need to have a meeting with someone else before I go to sleep
[06:01:40] <davidpardz>	 sumanah yes, I did. I emailed about my interest on that idea and no one had taken it. But my proposal was ready days before I showed my interest.
[06:01:50] <davidpardz>	 sumanah thank you
[06:02:12] <sumanah>	 davidpardz: that timeline is still a little confusing to me, but that's ok -- I know we have talked about this stuff before
[06:02:53] <sumanah>	 in any case I wish you the best -- and thank you for contributing to MediaWiki!  look forward to seeing you become one of our UploadWizard wizards :)
[06:02:55] <davidpardz>	 sumanah thank you again, I'll try harder next time
[06:03:13] <davidpardz>	 sumanah yeap see you around
[06:03:16] <sumanah>	 davidpardz: :)
[06:03:53] <sumanah>	 davidpardz: did you ever read http://webchick.net/embrace-the-chaos ?
[06:03:56] <sumanah>	 might be interesting
[06:04:06] <sumanah>	 ok, talk to you later
[08:56:58] <gerrit-wm>	 New patchset: Hashar; "testswarm: parse error in submit script" [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5406
[08:57:19] <gerrit-wm>	 New review: Hashar; "(no comment)" [integration/testswarm] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/5406
[08:57:21] <gerrit-wm>	 Change merged: Hashar; [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5406
[09:03:13] <gerrit-wm>	 New patchset: Hashar; "rename .ini file to ini-sample to avoid collision" [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5408
[09:03:30] <gerrit-wm>	 New review: Hashar; "(no comment)" [integration/testswarm] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/5408
[09:03:33] <gerrit-wm>	 Change merged: Hashar; [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5408
[10:33:31] <gerrit-wm>	 New patchset: Hashar; "gitignore testswarm-submit.ini" [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5412
[10:33:32] <gerrit-wm>	 New patchset: Hashar; "need to submit to vhost integration.wikimedia.org" [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5413
[10:33:33] <gerrit-wm>	 New patchset: Hashar; "gitweb URL had duplicate sha1 as parameter" [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5414
[10:33:33] <gerrit-wm>	 New patchset: Hashar; "emit an error on non HTTP 200 error code" [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5415
[10:33:34] <gerrit-wm>	 New patchset: Hashar; "invalid test URL, should be Special:JavaScriptTest" [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5416
[10:34:03] <gerrit-wm>	 New review: Hashar; "(no comment)" [integration/testswarm] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/5416
[10:34:11] <gerrit-wm>	 New review: Hashar; "(no comment)" [integration/testswarm] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/5415
[10:34:20] <gerrit-wm>	 New review: Hashar; "(no comment)" [integration/testswarm] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/5414
[10:34:30] <gerrit-wm>	 New review: Hashar; "(no comment)" [integration/testswarm] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/5413
[10:34:41] <gerrit-wm>	 New review: Hashar; "(no comment)" [integration/testswarm] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/5412
[10:34:43] <gerrit-wm>	 Change merged: Hashar; [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5416
[10:34:44] <gerrit-wm>	 Change merged: Hashar; [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5415
[10:34:45] <gerrit-wm>	 Change merged: Hashar; [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5414
[10:34:46] <gerrit-wm>	 Change merged: Hashar; [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5413
[10:34:47] <gerrit-wm>	 Change merged: Hashar; [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/5412
[10:34:58] <hashar>	 should be up to date now
[10:35:02] <hashar>	 lunch time!!!
[11:04:21] <mutante>	 what's a good way to check if refreshLinks.php  recently ran on a wiki and did what it is expected to?  anything better than "look for a wiki with empty categories, run it, and see if they are gone"?
[11:05:31] <petan|wk>	 mutante: logs?
[11:05:53] <petan|wk>	 and if it doesn't make any, we should fix it
[11:07:09] <mutante>	 right, also just looking for an example wiki to see cleanup from user side
[11:07:43] <mutante>	 or if the wiki users can see that action themselves even
[11:08:02] <petan|wk>	 I don't understand what action you mean now
[11:08:31] <petan|wk>	 refreshLinks is shell tool, users can't start it
[11:08:39] <mutante>	 an action like fixes to these "Categories are empty or don't show all pages, if "What links here?" doesn't work well or in some other link-related trouble"
[11:09:19] <mutante>	 petan|wk: i realize that, but it could have been that using the shell script still creates log on the wiki
[11:09:47] <petan|wk>	 I meant debug logs (text file)
[11:09:58] <petan|wk>	 at some point these jobs do create logs on wiki too
[11:10:26] <petan|wk>	 because these jobs rebuild the recentchanges and such histories
[11:11:19] <petan|wk>	 but only in case it fixes some "action" which should have been logged and for some reason wasn't or it's logged as done by Default user
[11:11:30] <petan|wk>	 Mediawiki default
[11:12:03] <petan|wk>	 but I don't believe that refreshLinks.php does log anything on wiki
[11:12:57] <mutante>	 ok, so maybe that would be a little enhancement to suggest, make it add a "refreshLinks ran and cleaned up some broken stuff for your convenience" line to RC
[11:13:30] <petan|wk>	 hm... I think it would be better to keep this technical info away from regular users who would be just confused about that
[11:13:47] <petan|wk>	 having a system log even publicly accessible would be better
[11:13:50] <mutante>	 hrmm.. then filter for admins and 'crats
[11:14:47] <petan|wk>	 no it's ok to make it public just it shouldn't be mixed with regular logs, and in the end admins and crats on wikipedia are mostly non technical users as well
[11:15:19] <petan|wk>	 actually I like that idea, but I would still preferer to have a nice easily browseable logfile somewhere for sysadmins
[11:15:48] <petan|wk>	 so that regular users aren't bothered with stuff they don't need to see, and sysadmin can easily check what is going on
[11:15:54] <mutante>	 btw logfiles, you saw that non-roots can read syslog on labs now?
[11:16:18] <petan|wk>	 which syslog? /var/log/messages
[11:16:30] <petan|wk>	 most of them used to be only for root dunno why
[11:16:38] <mutante>	 syslog and messages currently
[11:16:57] <mutante>	 other files in /var/log can easily be added to an array if requested
[11:16:59] <petan|wk>	 ah, that's cool, actually it would be cool if apache logs would be readable too
[11:17:21] <petan|wk>	 squid as well
[11:18:27] <mutante>	 https://gerrit.wikimedia.org/r/#patch,sidebyside,4157,3,manifests/base.pp
[11:18:43] <mutante>	 $common_logs =
[11:18:47] <petan|wk>	 nice
[11:19:16] <mutante>	 just limited to /var/log at this point
[11:19:30] <petan|wk>	 ok, that's actually where most of that is
[11:19:55] <mutante>	 but this is not the ideal way yet, supposed to change it and let the logfiles be written with right permissions in the first place, rather than changing them later and constantly
[11:20:30] <petan|wk>	 I am wondering why it was restricted in past
[11:20:35] <mutante>	 well, now we are on -labs topic and should probably switch channels
[11:20:43] <petan|wk>	 standard ubuntu has readable logs
[11:21:27] <petan|wk>	 anyway that logging idea is actually not bad, but maybe we should discuss it on wikitech-l
[11:21:48] <petan|wk>	 current system of logging is really bad
[11:21:51] <mutante>	 sure, why not
[11:22:14] <petan|wk>	 I mean the debug logs, not on wiki logs
[11:22:38] <petan|wk>	 how you on production search anything in debug logs? do you even have them?
[11:23:09] <petan|wk>	 I enabled debug logging for some common devs on labs so that they can find out what's going on, and logfile is a mess
[11:23:16] <mutante>	 there is a central logserver and udp2logs
[11:23:24] <petan|wk>	 that must be ton of logs
[11:23:41] <mutante>	 not every single service writes logs
[11:23:44] <petan|wk>	 like imagine you get error page when you open wiki
[11:24:03] <petan|wk>	 how would you search the error in logs for that specific session
[11:25:03] <petan|wk>	 or something is just broken, but not everytime it just randomly happen, you open a page and you need to find corresponding logs to debug why it's broken
[11:25:32] <petan|wk>	 there is million of lines in that logfile and they aren't even sorted or prefixed
[11:25:51] <petan|wk>	 the logs we have on labs were of no use until I improved them with unique prefixes
[11:26:17] <petan|wk>	 we did it yesterday with hashar to debug memcached
[11:26:20] * mutante  nods.. yeah, that sounds like something for the list now. i don't have answers to all of that

[13:00:46] <qchris>	 ^demon: May I disturb you for a second?
[13:00:56] <^demon>	 What's up?
[13:01:07] <qchris>	 Would you mind having a look at:
[13:01:14] <qchris>	 https://gerrit.wikimedia.org/r/#change,5421
[13:01:33] <qchris>	 As there are some problems with lengths in the db, we'd need this patch to get accepted
[13:01:50] <qchris>	 Jenkins fails, but that is because a test now takse >2 seconds
[13:02:03] <qchris>	 I'll see to that afterwards
[13:02:25] <^demon>	 Is this something that needs to make its way into master too?
[13:02:38] <qchris>	 We are not sure yet.
[13:03:04] <qchris>	 Some things need to be checked and then we'll see how to overcome the main issue
[13:03:38] <^demon>	 Mmk. We don't use this outside of the dumps, so it'll be fine for now probably. We'll need a better solution before we branch 1.20wmf2 though.
[13:03:39] <^demon>	 Sound ok?
[13:03:52] <qchris>	 Yes.
[13:05:12] <hashar>	 qchris: you can  raise the timeout for a test by  using   @medium
[13:05:40] <qchris>	 hashar: Yes, that was my plan. The test passes locally for me.
[13:05:46] <hashar>	 \o/
[13:05:57] <^demon>	 hashar: It's that silly empty wikitext parser case that's failing.
[13:06:19] <^demon>	 https://integration.mediawiki.org/ci/job/MediaWiki-Tests-Parser/773/testReport/junit/(root)/ParserTests__testParserTest/testParserTest_with_data_set__0/
[13:06:26] <hashar>	 \O/
[13:06:40] <apergos>	 ^demon: the short summary is that revision text returned now (maybe) has \r stripped when in the past it didn't always (this is a best guess but it fits the numbers); the lengths in the db for many revisions appear to count the \r s
[13:06:44] <hashar>	 that is the first test
[13:07:02] <hashar>	 I am pretty sure that timeout is caused by some parser igniter that load a ton of stuff
[13:07:03] <apergos>	 this means we are likely going to have to repopulate that column on all wikis
[13:07:22] <hashar>	 for example rebuilding l10n caches
[13:07:23] <^demon>	 hashar: That's what I'm assuming.
[13:07:32] <^demon>	 apergos: Gotcha. Sounds fine, I'll go ahead and approve it
[13:07:32] <apergos>	 this fix means the dumps don't die the death while we schedule a run of those things
[13:07:33] <Platonides>	 apergos, I'm quite sure \r has never been in the revisions
[13:07:36] <apergos>	 thanks
[13:08:17] <apergos>	 I wouldn't be so sure. these are revisions from 2003
[13:08:21] <^demon>	 Merged. Hopefully we can get a real fix in master before we branch again :)
[13:08:46] <qchris>	 Platonides: There are some guards against \r in revisions in dump's text pass. So I assume they at least have been there at some point
[13:08:50] <apergos>	 I would love it.  although I don't know what the fix should be, maybe just "repopulate with what seems to be the correct length"
[13:09:02] <qchris>	 ^demon: Thanks
[13:09:40] <^demon>	 Does this need to be scapped out, or do you guys deploy this on your own?
[13:10:46] <apergos>	 Platonides: here's an example.  rev_len listed as 399 for an 8 line text.  actual length when looking at mysql output form the command line or what mw thinks the text ought to be: 391.
[13:11:02] <apergos>	 it's core so generally I don't much with it
[13:11:12] <apergos>	 if it were my pile of scripts that would be us
[13:11:16] <apergos>	 *muck
[13:11:28] <^demon>	 apergos: 5 lines up ^
[13:11:48] <apergos>	 ^demon: 3 lines up ^
[13:11:49] <apergos>	 :-P
[13:12:16] <^demon>	 So is that a yes I need to sync-file?
[13:13:20] <apergos>	 yes, that would be awesome (I dunno if you want to clear it with someone there or not)
[13:14:32] <^demon>	 Clear with whom?
[13:14:34] <^demon>	 ;-)
[13:14:49] <apergos>	 I dunno, whoever the buck gets passed to for deployments :-P :-D
[13:15:18] <^demon>	 Well it's sync'd :)
[13:15:27] <apergos>	 when is the next branch likely to happen? (gives me a timetable for how long our grace period to repopulate rev_length is
[13:15:28] <apergos>	 )
[13:15:30] <apergos>	 thanks
[13:16:12] <^demon>	 I don't know what date we're pencilling in for the next deployment yet. Reedy might know.
[13:16:25] <^demon>	 I know we were probably going to take 3 weeks for the first cycle instead of 2.
[13:16:38] <apergos>	 that took care of the issue for now, yay for that
[13:16:52] <apergos>	 ok, so that should be plenty of time
[13:17:24] <apergos>	 lemme look at that maintenance script
[13:17:39] <apergos>	 everything's on 1.20 now right?
[13:17:48] <^demon>	 Don't know :)
[13:17:51] <^demon>	 Maybe?
[13:18:27] <apergos>	 hahaha
[13:18:43] <apergos>	 I think the length issue predates 1.20 so it's probably ok
[13:19:23] <Reedy>	 Nope
[13:19:27] <Reedy>	 Wikipedias aren't...
[13:19:28] <Reedy>	 IIRC
[13:19:30] <apergos>	 ok
[13:19:43] <Reedy>	 Monday, April 23, 18:00-20:00 UTC (11am-1pm PDT): MediaWiki 1.20wmf01 deployment - Deploy to English Wikipedia
[13:19:43] <Reedy>	 Wednesday, April 25, 18:00-20:00 UTC (11am-1pm PDT): MediaWiki 1.20wmf01 deployment - Deploy to other Wikipedias
[13:19:53] <apergos>	 All *wiki wikis to 1.20wmf1
[13:20:11] <apergos>	 19:51 apr 19
[13:20:16] <apergos>	 that's not the pedias?
[13:20:39] <^demon>	 Maybe he meant *wiki.+
[13:20:52] <^demon>	 :)
[13:21:04] <apergos>	 http://wikitech.wikimedia.org/view/Server_admin_log  :-P
[13:21:15] <apergos>	 apr 18, sorry
[13:23:04] <apergos>	 yeah they seem to still be 1.19 regardless of the log entry
[14:28:00] <hexmode>	 I abandoned and fixed my commit the minute after it was up, but ^demon had already -1'd it :P
[14:28:30] <^demon|away>	 I'm fast :)
[15:42:21] <hexmode>	 "A non well formed numeric value encountered in /home/mah/domains/winkyfrown.com/public_html/mediawiki-1.18.0/includes/CryptRand.php on line 379"
[15:42:24] <hexmode>	 hrm
[15:42:57] <hexmode>	 and that is 1.18.2 fwiw
[15:43:19] <Reedy>	 I think that's been fixed in git
[15:43:22] <Reedy>	 Maybe.
[16:08:44] <Krinkle>	 hashar: Is the snapshotting of mediawiki live?
[16:08:54] <Krinkle>	 Or is it pending review?
[16:09:03] <Krinkle>	 (for ci / testswarm)
[16:09:03] <hashar>	 live
[16:09:12] <hashar>	 was waiting for ya :-D
[16:09:25] <Krinkle>	 Ah, I see
[16:09:26] <Krinkle>	 http://integration.mediawiki.org/clone/mw/master/
[16:09:42] <Krinkle>	 http://integration.mediawiki.org/clone/mw/master/5430/1/
[16:09:44] <Krinkle>	 readonly db
[16:10:06] <Krinkle>	 http://integration.mediawiki.org/clone/mw/master/5427/1 is working though
[16:10:12] <hashar>	 https://integration.mediawiki.org/testswarm/user/MediaWiki/
[16:10:19] <hashar>	 5430/1 is building right now
[16:10:23] <Krinkle>	 :O
[16:10:27] <hashar>	 oh no that is 5438
[16:10:31] <Krinkle>	 I didn't look at testswarm yet
[16:10:33] <hashar>	 well there is an issue of some sort then
[16:10:33] <Krinkle>	 nice
[16:10:51] <hashar>	 one issue is that once the test block has finished, it does not close
[16:10:51] <Krinkle>	 hashar: You need to add some config
[16:10:58] <Krinkle>	 https://integration.mediawiki.org/testswarm/job/1909/
[16:10:58] <hashar>	 and all tests are reporting as timeout
[16:10:59] <Krinkle>	 inject.js
[16:11:13] <Krinkle>	 testswarm hooks are not registered, you need to tell mw where testswarm is
[16:11:55] <Krinkle>	 hashar: https://gerrit.wikimedia.org/r/#change,3702
[16:12:24] <Krinkle>	 $wgJavaScriptTestConfig['testswarm-injectjs']
[16:12:28] <Krinkle>	 @example '//integration.mediawiki.org/testswarm/js/inject.js'
[16:13:03] <hashar>	 why isn't it injected by default?
[16:13:41] <Krinkle>	 hashar: because it is version specific
[16:13:47] <Krinkle>	 testswarm version specific
[16:14:04] <Krinkle>	 and not everybody wants to use testswarm, so loading by default is pointless too
[16:14:57] <hashar>	 k
[16:15:07] <hashar>	 will have to add that to ExtraSettings.php
[16:15:45] <Krinkle>	 Is ExtraSettings.php included or substituted?
[16:16:27] <hashar>	 it is copied to  /mw/core/something and then required()
[16:16:38] <hashar>	 so would have to update the existing one manually :-]
[16:16:51] <Krinkle>	 okay, good
[16:17:17] <Krinkle>	 should be as static as can be, the manual updating is not a problem. better than changes reflecting retro-actively on asynchronous tests
[16:17:55] <Krinkle>	 hashar: If you add it to the existing ExtraSettings files I'll reset the job via testswarm interface
[16:18:13] <Krinkle>	 login succesful
[16:19:43] <Krinkle>	 hashar: yeah /clone/mw/master/5438/1 , I don't know whether it is building or not, but it has been submitted to testswarm and clients crash over it
[16:23:29] <Krinkle>	 hashar: several load.php requests are also failing
[16:23:30] <Krinkle>	 exception 'DBQueryError' with message 'A database error has occurred.  Did you forget to run maintenance/update.php after upgrading?  See: https://www.mediawiki.org/wiki/Manual:Upgrading#Run_the_update_script
[16:23:30] <Krinkle>	 Query: REPLACE INTO module_deps (md_module,md_skin,md_deps) VALUES
[16:23:32] <Krinkle>	 ('mediawiki.legacy.shared','vector','["\/var\/lib\/testswarm\/mediawiki-git\/clone\/5427\/1\/skins\/common\/images\/Checker-16x16.png","\/var\/lib\/testswarm\/mediawiki-git\/clone\/5427\/1\/skins\/common\/images\/feed-icon.png","\/var\/lib\/testswarm\/mediawiki-git\/clone\/5427\/1\/skins\/common\/images\/ajax-loader.gif","\/var\/lib\/testswarm\/mediawiki-git\/clone\/5427\/1\/skins\/common\/images\/spinner.gif","\/var\/lib\/testswarm\/mediawiki-gi
[16:23:33] <Krinkle>	 lone\/5427\/1\/skins\/common\/images\/help-question.gif","\/var\/lib\/testswarm\/mediawiki-git\/clone\/5427\/1\/skins\/common\/images\/help-question-hover.gif","\/var\/lib\/testswarm\/mediawiki-git\/clone\/5427\/1\/skins\/common\/images\/tipsy-arrow.gif"]')
[16:23:34] <Krinkle>	 Function: DatabaseSqlite::replace/single-row
[16:23:35] <Krinkle>	 Error: 8 attempt to write a readonly database
[16:23:36] <Krinkle>	 ' in /var/lib/testswarm/mediawiki-git/clone/5427/1/includes/db/Database.php:937
[16:23:54] <Krinkle>	 resourceloader cache for module stuff
[16:24:26] <hashar>	 I think apache can't write to the sqlite file
[16:24:31] <hashar>	 already thought about that bug
[16:25:24] <Krinkle>	 the first ones are working fine, the last 2 are failing with db
[16:25:30] <Krinkle>	 hashar: the module_deps cache bug?
[16:28:13] <hashar>	 Krinkle: ExtraSettings.php updated for all snapshots
[16:28:21] <hashar>	 $wgEnableJavaScriptTest = true;
[16:28:22] <hashar>	 $wgJavaScriptTestConfig['testswarm-injectjs'];
[16:28:54] <Krinkle>	 I assume you did set that array to something?
[16:29:00] <Krinkle>	 I've reset the jobs in testswarm
[16:29:11] <hashar>	 no just copy pasted ahah
[16:29:41] <hashar>	 should I set a boolean 'true' value
[16:29:42] <hashar>	 ?
[16:30:00] <Krinkle>	 No
[16:30:08] <Krinkle>	 @example '//integration.mediawiki.org/testswarm/js/inject.js'
[16:31:05] <hashar>	 $wgJavaScriptTestConfig['testswarm-injectjs'] = '//integration.mediawiki.org/testswarm/js/inject.js' ;
[16:31:09] <Krinkle>	 Right :)
[16:31:17] <Emw>	 	i'm developing a mediawiki extension for webgl-enabled interactive models. the extension will require a way to create static snapshots of a given model for clients that don't support webgl. i've got scripts to do this, but they're written in python and not php. would that be a problem if i aimed to eventually get the extension included in wikipedia's deployment of mediawiki?
[16:31:55] <hashar>	 find . -name ExtraSettings.php -exec cp /var/lib/jenkins/jobs/_shared/ExtraSettings.php {} \;
[16:31:56] <hashar>	 \o/
[16:32:19] <hashar>	 Emw: you want to talk about it on wikitech-l
[16:32:24] <hashar>	 Emw: our mailing list ;-D
[16:32:37] * Krinkle  resets jobs again

[16:32:39] <hashar>	 Emw: python has a few fan in WMF but I guess we prefer php
[16:34:06] <Krinkle>	 hashar: not working yet
[16:34:12] <Krinkle>	 mw.config({"QUnitTestSwarmInjectJSPath":false});
[16:34:31] <Emw>	 hashar: ok.  also, the static image generator uses scripts in GIMP and a FOSS version of PyMOL (http://en.wikipedia.org/wiki/PyMOL).  what are the changes that those dependencies would be an issue?
[16:34:33] <Krinkle>	 hashar: It is not set or included some how
[16:35:08] <hashar>	 Emw: GIMP  PyMOL are entirely unknown to me sorry :/
[16:35:28] <hashar>	 I don't really have time to look that up right now sorry :-((
[16:35:40] <Emw>	 np, thanks though
[16:36:03] <hashar>	 Emw: you really want to ask wikitech-l  , will ensure almost all developers will have a look at it :-D
[16:36:14] <hashar>	 Emw: or wait for the US dudes to wake up ;-))
[16:36:29] <Emw>	 sounds good
[16:36:40] <hashar>	 Emw: maybe brion would be interested in someone rendering molecules in webgl
[16:36:55] <hashar>	 he likes new and fresh technologies ;-]
[16:37:09] <brion>	 that'd be spiffy. :) but would need a 2d canvas fallback probably :)
[16:37:15] <Emw>	 i've got a prototype working locally, and it looks very cool
[16:38:53] <Krinkle>	 hashar: Looks like all dbs are read only, not only the last 2 snapshots. The earlier ones are working because the needed cache is already in the db for them, but the older ones also have load.php errors where module_deps can't be updated
[16:39:16] <Krinkle>	 Query: INSERT OR IGNORE INTO msg_resource (mr_lang,mr_resource,mr_blob,mr_timestamp) VALUES ('en','user.options','{}','20120420163745') Function: MessageBlobStore::insertMessageBlob/single-row
[16:39:17] <Krinkle>	 Error: 8 attempt to write a readonly database
[16:39:23] <Emw>	 no 2d canvas fallback yet though.  that isn't yet supported by the library i'm using yet.  a non-mediawiki demo is available at http://webglmol.sourceforge.jp/glmol/viewer.html.  i've optimized my prototype to load significantly faster.
[16:42:07] <hashar>	 Krinkle: I made a change to let jenkins change the sqlite database to be in www-data group
[16:42:13] <hashar>	 Krinkle: but can't find it anymore :-(
[16:42:13] <Krinkle>	 ok
[16:44:39] <hashar>	 ohhh
[16:44:48] <hashar>	 there is a g+s flag on the directory
[16:44:57] <Krinkle>	 hashar: http://etherpad.wikimedia.org/ContinuousIntegration
[16:44:58] <hashar>	 so any file created there are in the www-data group
[16:45:01] <hashar>	 I just need to g+w
[16:45:25] <hashar>	 and update the build script
[16:48:27] <hashar>	 Krinkle: all sqlite files are now writable
[16:49:43] <Krinkle>	 hashar: http://etherpad.wikimedia.org/ContinuousIntegration
[16:50:03] <Krinkle>	 hashar: added @fixme: Fix testswarm useragent local hack
[16:50:13] <Krinkle>	 I assume you added Firefox 11?
[16:50:27] <hashar>	 I have no idea
[16:51:08] <Krinkle>	 db is working now :)
[16:51:52] <hashar>	 yeah I manually g+w the files
[16:52:11] <Krinkle>	 the extra settings isn't working though
[16:52:18] <Krinkle>	 strange
[16:55:15] <gerrit-wm>	 New patchset: Hashar; "enable testswarm submissions" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/5440
[16:55:38] <hashar>	 here are the various changes that were sitting on gallium
[16:56:07] <Krinkle>	 k, but the extra settings, module_deps and user agents need to be fixed before submitting more jobs to testswarm
[16:56:51] <csteipp>	 ls
[16:59:00] <hashar>	 Krinkle: I need to leave the coworking place
[16:59:06] <Krinkle>	 ok
[16:59:20] <hashar>	 would it be possible to copy paste the ether pad issues as bug reports ?
[16:59:24] <hashar>	 I think it will be easier to handle
[16:59:39] <hashar>	 well no
[16:59:43] <hashar>	 that is just trivial stuff
[16:59:45] <hashar>	 lets keep it there
[16:59:52] <Krinkle>	 yeah, the other bugs aren't marked fixed yet
[16:59:59] <Krinkle>	 this is just intermediary stuff
[17:02:10] <hashar>	 so I am out now Krinkle
[17:02:13] <hashar>	 space closed
[17:02:15] <hashar>	 cya!!!
[17:02:17] <Krinkle>	 cya
[17:02:23] <Krinkle>	 thanks for the progress, getting really close!
[17:24:39] <awjr>	 RoanKattouw: ping
[17:24:45] <RoanKattouw>	 Pong
[17:24:49] <RoanKattouw>	 Was looking for you
[17:24:54] <awjr>	 how's it going?
[17:25:01] <awjr>	 sorry had some trouble getting going this moring
[17:25:05] <RoanKattouw>	 No worries
[17:25:34] <RoanKattouw>	 So I heard you got stuck somewhere in my documentation
[17:25:34] <awjr>	 yeah i was setting up the remote branch for mobile frontend
[17:25:54] <awjr>	 at the end of this section: http://wikitech.wikimedia.org/view/How_to_deploy_code#Creating_an_extension_deployment_branch
[17:26:16] <awjr>	 the git push part was where i got stuck. i was trying to do what i thought was the equivalent with git review but was not working
[17:26:24] <RoanKattouw>	 Ooooh
[17:26:29] <RoanKattouw>	 You can't git push to your own extension?
[17:26:36] <RoanKattouw>	 Not even to create new branches? That's strange
[17:26:39] <awjr>	 lemme try again
[17:27:00] <RoanKattouw>	 OK you know what, I am just going to fix the Gerrit perms right now
[17:27:11] <awjr>	 actually… it just worked
[17:27:14] <RoanKattouw>	 Then reality should match the docs, and we can try it with that
[17:27:15] <RoanKattouw>	 aha
[17:27:18] <awjr>	 it's possible that i assumed it would not have worked
[17:27:21] <awjr>	 and just went with git review instead
[17:27:28] <RoanKattouw>	 Assume nothing! ;)
[17:27:38] <awjr>	 :p
[17:27:47] <awjr>	 ok that all worked
[17:28:21] <RoanKattouw>	 Sweet, I can see your wmf branch in gitweb now
[17:28:29] <awjr>	 the following steps i do in my dev checkout of MobileFrontend, or from core?
[17:28:51] <RoanKattouw>	 awjr: Each of the blocks has a cd command at the beginning
[17:29:17] <awjr>	 ok got it
[17:29:27] <awjr>	 actually i dont
[17:29:36] <awjr>	 http://wikitech.wikimedia.org/view/How_to_deploy_code#Now_that_your_extension_has_a_deployment_branch
[17:30:05] <awjr>	 the first cd is ambiguous to me - is that for my extension in the core checkout, like, the submodule?
[17:30:57] <RoanKattouw>	 Oh, sorry
[17:31:02] <RoanKattouw>	 No, the "real" extension checkout
[17:31:04] <RoanKattouw>	 I'll fix that
[17:31:09] <awjr>	 from the instructs that follow, i assume it should NOT be the submodule
[17:31:09] <awjr>	 ok yeah
[17:31:18] <awjr>	 but yeah, didn't want to assume anything :)
[17:31:20] <RoanKattouw>	 Or clarify, rather
[17:31:37] <RoanKattouw>	 It is technically not ambiguous because the examples that use core use mediawiki/core/extensions/Blah
[17:31:45] <RoanKattouw>	 And the ones that use real extension checkouts use mediawiki/extensions/Blah
[17:31:49] <RoanKattouw>	 But that should be clarified somewhere
[17:31:50] <awjr>	 ah yeah you're right
[17:32:04] <awjr>	 hmm
[17:32:12] <awjr>	 $ [master]$ git branch --track wmf/1.20wmf1 origin/wmf/1.20wmf1
[17:32:12] <awjr>	 fatal: Not a valid object name: 'origin/wmf/1.20wmf1'.
[17:32:12] <awjr>	 awjrichards@virtualbot:~/Dev/wikimedia_git/extensions/MobileFrontend
[17:32:24] <awjr>	 do i need to git pull first?
[17:32:46] <awjr>	 yes
[17:32:49] <awjr>	 git pull first helped
[17:33:15] <RoanKattouw>	 Aha
[17:33:20] <RoanKattouw>	 Yeah I need to add that too, thanks
[17:33:31] <awjr>	 np, thanks for even having documentation :p
[17:33:45] <RoanKattouw>	 Oh, I have a git pull further down
[17:34:24] <awjr>	 for the cherry pick -
[17:34:29] <awjr>	 which hash do i use?
[17:34:37] <awjr>	 sorry, i feel like i should know this by now
[17:34:47] <awjr>	 this is the change: https://gerrit.wikimedia.org/r/#change,5400
[17:35:09] <awjr>	 would it be the commit hash? https://gerrit.wikimedia.org/r/gitweb?p=mediawiki%2Fextensions%2FMobileFrontend.git;a=commit;h=ba9e5dbeb25be279d9c90e581152e91e5325fea1
[17:35:39] <RoanKattouw>	 Yes, the 40-character SHA-1 hash
[17:35:45] <RoanKattouw>	 of the change that ended up in master
[17:36:04] <RoanKattouw>	 So yeah that's ba9e5... in this case
[17:36:28] <awjr>	 awjrichards@virtualbot:~/Dev/wikimedia_git/extensions/MobileFrontend$ [wmf/1.20wmf1]$ git push origin wmf/1.20wmf1
[17:36:28] <awjr>	 Counting objects: 5, done.
[17:36:28] <awjr>	 Compressing objects: 100% (3/3), done.
[17:36:28] <awjr>	 Writing objects: 100% (3/3), 383 bytes, done.
[17:36:28] <awjr>	 Total 3 (delta 2), reused 0 (delta 0)
[17:36:28] <awjr>	 remote: Resolving deltas:   0% (0/2)
[17:36:28] <awjr>	 To ssh://review/mediawiki/extensions/MobileFrontend.git
[17:36:29] <awjr>	  ! [remote rejected] wmf/1.20wmf1 -> wmf/1.20wmf1 (prohibited by Gerrit)
[17:36:31] <RoanKattouw>	 (If you have the Ubuntu git package, you should be able to tab-complete commit hashes in git commands)
[17:36:33] <RoanKattouw>	 Ah yes
[17:36:35] <awjr>	 this is where i need to use git-review?
[17:36:36] <RoanKattouw>	 That I have yet to fix
[17:36:41] <RoanKattouw>	 Well
[17:36:45] <RoanKattouw>	 If you have a few minutes I can fix the perms
[17:36:49] <RoanKattouw>	 And you can be my guinea pig
[17:36:58] <awjr>	 sure thang
[17:43:02] <RoanKattouw>	 awjr: What is the e-mail address or username you use for Gerrit?
[17:43:09] <awjr>	 arichards@wm.o
[17:43:38] <RoanKattouw>	 Ah yes
[17:43:41] <RoanKattouw>	 OK, try pushing now
[17:43:53] <awjr>	 that appears to have worked
[17:44:31] <RoanKattouw>	 Yup that worked
[17:44:48] <RoanKattouw>	 You should also have push rights on the wmf/1.20wmf1 branch in core
[17:44:58] <RoanKattouw>	 So you should be able to do all the submodule update stuff too
[17:45:14] <awjr>	 woohoo thanks
[17:45:31] <RoanKattouw>	 (I added you to the wmf-deployment group just now, for some reason only deployers from the platform group plus yours truly were added to that)
[17:48:39] <awjr>	 yay i think that worked
[17:49:09] <awjr>	 so RoanKattouw: now that i have my fix checked in, do i need to do anything special on fenari besides git pull?
[17:49:47] <RoanKattouw>	 Yes, it's in the documentation
[17:49:53] <awjr>	 of course
[17:49:56] * awjr  reads more

[17:49:58] <RoanKattouw>	 Under step 2
[17:50:57] <awjr>	 RoanKattouw: does this look right?
[17:50:59] <awjr>	 awjrichards@fenari:/home/wikipedia/common/php-1.20wmf1/extensions/MobileFrontend$ git branch
[17:50:59] <awjr>	 * (no branch)
[17:51:10] <RoanKattouw>	 Yes
[17:51:14] <awjr>	 ok
[17:51:19] <RoanKattouw>	 That happens with submodules
[17:51:39] <awjr>	 ok cool
[17:51:57] <awjr>	 i dont see any conflicts on the calendar, but is there any reason i should NOT run scap? this is an i18n change
[17:52:04] <RoanKattouw>	 No go ahead
[17:52:08] <RoanKattouw>	 Just make sure it runs to completion
[17:52:25] <awjr>	 aye
[17:52:42] <RoanKattouw>	 When I run it then for some reason it doesn't run the final steps (the steps that deploy the actual code and log "sync done" to the channel)
[17:53:12] <awjr>	 that is weird
[17:53:30] <RoanKattouw>	 Yeah, no idea why that happened
[17:53:43] <awjr>	 solar flares? gremlins?
[17:53:57] <awjr>	 all seems very plausible.
[17:56:02] <brion>	 anybody care to do some quick gerrit reviews for me?
[17:56:08] <brion>	 https://gerrit.wikimedia.org/r/#dashboard,25 <- under 'started by Brion VIBBER'
[18:08:45] <awjr>	 RoanKattouw: scap appears to ahve finished although no log showed up in wiimedia-tech
[18:09:06] <RoanKattouw>	 Did it take like 15 minutes to sync stuff over, or no?
[18:09:12] <RoanKattouw>	 Excluding the cache rebuilds I mean
[18:09:13] <awjr>	 surprisingly no
[18:09:17] <RoanKattouw>	 Aha
[18:09:32] <awjr>	 last informative thing i saw was Copying style sheets to apaches...
[18:09:37] <RoanKattouw>	 Yeah thought so
[18:09:40] <RoanKattouw>	 That's exactly what happens to me
[18:09:44] <RoanKattouw>	 And for the life of me I do not know why
[18:09:45] <awjr>	 :|
[18:09:46] <RoanKattouw>	 Anyway
[18:09:48] <RoanKattouw>	 You need to run ddsh -F30 -cM -g mediawiki-installation -o -oSetupTimeout=10 '/usr/bin/scap-1'
[18:10:14] <awjr>	 k running
[18:12:14] <awjr>	 RoanKattouw is that just what actually syncs the changes?
[18:12:20] <RoanKattouw>	 Yes
[18:12:23] <awjr>	 cool
[18:12:28] <RoanKattouw>	 The other stuff is all preparatory or peripheral
[18:12:36] <RoanKattouw>	 Cache rebuilds, syncing CSS/JS stuff, etc
[18:31:31] <RoanKattouw>	 Rawr
[18:31:41] <RoanKattouw>	 I am not a Gerrit administrator (anymore?) so I can't fix the gerrit perms properly
[18:39:03] <RoanKattouw>	 Oh hey ^demon
[18:39:11] <RoanKattouw>	 I *just* sent you an e-mail about the wmf/* perms thing
[18:40:03] <^demon>	 I just read it.
[18:40:33] <RoanKattouw>	 I think we need to discuss how this will work in a bit more detail
[18:41:08] <RoanKattouw>	 Also, while trying to deny rights from project owners on mediawiki/*:wmf/* (which it doesn't let me do, probably because I'm not an admin), I realized that it may not work
[18:41:24] <^demon>	 Why not?
[18:41:26] <RoanKattouw>	 Because project owners can set perms on their project and probably find a way around the system and into the wmf/* branches for their extensions
[18:41:50] <^demon>	 We can adjust it so owners can't change their own config.
[18:41:55] <^demon>	 Config changes should be requested, anyway
[18:41:58] <RoanKattouw>	 Unless Gerrit's permission system has magic mind-reading capabilities, or works in a strange way that ends up doing exactly what we want
[18:42:01] <RoanKattouw>	 Oh, hah, yeah that works
[18:43:42] <^demon>	 Just deny permissions on refs/meta/config or w/e it is.
[18:44:45] <RoanKattouw>	 BTW could I be put in the Administrators group?
[18:44:46] <^demon>	 I also added you back to admin group, not sure why you weren't there.
[18:44:50] <RoanKattouw>	 Ah thansk
[18:44:52] <^demon>	 Yeah I already did that :)
[18:44:56] <RoanKattouw>	 Yeah I think I was there at some point
[18:45:05] <RoanKattouw>	 And maybe I was removed, or I just misremember
[20:20:23] <csteipp>	 Hey RoanKattouw, do you know if wmf is using the Interwiki extension on any of our sites?
[20:35:05] <RoanKattouw>	 No I don't believe we use it
[20:35:09] <RoanKattouw>	 ( csteipp  )
[20:35:24] <csteipp>	 cool. thanks!
[21:50:13] <sumanah>	 csteipp: http://kernsec.org/wiki/index.php/Linux_Security_Summit_2012 reminded me to talk to you (linked in LWN)
[22:21:54] <brion>	 kaldari: hey can you review a couple of my commits on UploadWizard if you hvae the chance?
[22:22:12] <kaldari>	 I'm looking at the first one now
[22:22:28] <brion>	 https://gerrit.wikimedia.org/r/#change,4935 & https://gerrit.wikimedia.org/r/#change,4926
[22:22:29] <brion>	 thx
[22:22:41] <brion>	 also have a related one on core: https://gerrit.wikimedia.org/r/#change,4924
[22:52:56] <Amir1>	 Hello! I've commit access but i can't commit :(
[22:53:24] <Amir1>	 I think problem is my public key
[22:57:35] <sumanah>	 hi Amir1
[22:57:38] <sumanah>	 what error message?
[22:57:52] <Amir1>	 just a moment
[22:59:22] <Amir1>	 Commit failed (details follow):Access to '/svnroot/pywikipedia/!svn/act/9af6b0c1-3066-6b4f-a71a-658b53dc2cb1' forbidden
[23:00:44] <sumanah>	 Amir1: pywikipedia, hm.
[23:00:57] <Amir1>	 yes
[23:01:02] <sumanah>	 Amir1: that is weird. when was the last time you tried to commit?
[23:01:26] <sumanah>	 (and did so successfully)
[23:01:28] <Amir1>	 I don't remember
[23:01:40] <Amir1>	 but I did so many times
[23:02:00] <sumanah>	 I'm trying to figure out where the prob happened
[23:02:08] <Amir1>	 http://www.mediawiki.org/wiki/Special:Code/pywikipedia/author/amir
[23:02:27] <sumanah>	 so, 31 july 2011
[23:02:39] <sumanah>	 Amir1: I have gotten someone who will be dropping in to help you
[23:02:50] <RoanKattouw>	 ewwww
[23:02:54] <RoanKattouw>	 That is a SCARY error message
[23:03:06] <Amir1>	 sumanah: Thanks
[23:03:21] <sumanah>	 hi tparveen! how is it going?
[23:03:31] <sumanah>	 btw amir:
[23:03:35] <sumanah>	 !events | amir
[23:03:35] <mw-bot>	 amir: We run face-to-face events for MediaWiki developers and people who want to learn: https://www.mediawiki.org/wiki/MediaWiki_developer_meetings We also have online meetings in IRC to triage bugs: https://www.mediawiki.org/wiki/Bug_management/Triage
[23:03:40] <sumanah>	 !berlin | amir
[23:03:40] <mw-bot>	 amir: Registration is now open for the Berlin hackathon, 1-3 June 2012. See https://www.mediawiki.org/wiki/Berlin_Hackathon_2012
[23:03:45] <Amir1>	 I think problem is my public key because my toolserver account expired and my public key was there
[23:04:19] <sumanah>	 Amir1: https://www.mediawiki.org/wiki/Developer_access/Subversion#Updating.2Freplacing_your_SSH_key
[23:04:25] <tparveen>	 hello Sumanah...I am ok..how are you...trying to figure out interwikilink...
[23:04:55] <sumanah>	 tparveen: I'm not bad!  You're trying to figure out this thing? https://www.mediawiki.org/wiki/Interwiki_link
[23:05:33] <tparveen>	 yes
[23:05:46] <RoanKattouw>	 Amir1: No, that won't affect the validity of your key
[23:05:56] <RoanKattouw>	 Amir1: I know the error message for 'your key is wrong' and it looks different
[23:06:14] <Amir1>	 RoanKattouw: OK
[23:06:15] <RoanKattouw>	 I think this is the "you don't have permission to commit to this specific thing" error
[23:06:18] * RoanKattouw  wants Chad to be here

[23:06:34] <sumanah>	 well that's strange, though, because https://www.mediawiki.org/wiki/Special:Code/pywikipedia
[23:06:45] <RoanKattouw>	 Yes, he has prior commits there
[23:06:50] <sumanah>	 other people can commit to pywikipedia
[23:06:52] <RoanKattouw>	 And other people committed to it today
[23:07:07] <sumanah>	 indeed
[23:08:08] <tparveen>	 sumanah: yes to test commons - trying to make some oracles using commons so I can test labs.common
[23:08:32] <RoanKattouw>	 Ryan_Lane: Maybe you can debug this issue?
[23:08:36] <sumanah>	 tparveen: I don't actually know what "make some oracles" means here - is "oracles" a QA term, or are you talking about the RDBMS?
[23:08:44] <RoanKattouw>	 I have been looking around on formey a bit but I can't find the pywikipedia access controls
[23:08:48] <sumanah>	 there are none
[23:08:55] <RoanKattouw>	 :O
[23:08:57] <RoanKattouw>	 Right
[23:09:00] <RoanKattouw>	 That's why the file is empty :D
[23:09:07] <sumanah>	 RoanKattouw: http://wikitech.wikimedia.org/view/Svn.wikimedia.org#Access_Control
[23:09:11] <sumanah>	 All SVN users have access to the pywikipediabot project, even those who aren't in /svnroot/mediawiki/conf/authz.
[23:09:13] <tparveen>	 Sumanah: :) testing terms - it means sort of baseline
[23:09:21] <RoanKattouw>	 Amir1: Is this problem reproducible?
[23:09:22] <sumanah>	 interesting!
[23:09:34] <RoanKattouw>	 Does it occur every time, even if you try to commit and try again?
[23:10:01] <Amir1>	 RoanKattouw: Yes
[23:10:12] <RoanKattouw>	 What are you committing exaclty?
[23:10:29] <Amir1>	 cosmetic_changes.py
[23:10:35] <RoanKattouw>	 Also, when you run svn info , what is the URL that it gives you?
[23:10:36] <Amir1>	 in pywikipedia folder
[23:12:07] <Amir1>	 RoanKattouw: I can't find info
[23:12:20] <RoanKattouw>	 Amir1: Run the 'svn info' command
[23:12:50] <Amir1>	 I use "Tortoise SVN"
[23:12:55] <Amir1>	 I windows
[23:12:58] <Amir1>	 in windows
[23:14:10] <RoanKattouw>	 Oh, hm
[23:14:18] <Amir1>	 but i can tell where i checked out
[23:14:22] <RoanKattouw>	 There must be some way that you can ask it what URL you checked out
[23:14:34] <Amir1>	 http://svn.wikimedia.org/svnroot/pywikipedia/trunk/pywikipedia/
[23:15:55] <Amir1>	 commit to:http://svn.wikimedia.org/svnroot/pywikipedia/trunk/pywikipedia/cosmetic_changes.py
[23:16:28] <Ryan_Lane>	 Roawhich issue?
[23:16:34] <Ryan_Lane>	 hahaha
[23:16:36] <Ryan_Lane>	 RoanKattouw: which issue
[23:21:50] <sumanah>	 tfinc: https://thestrangeloop.com/about/coverage - no recordings from last year that I can see, but you might want to skim a few blog entries about the event
[23:22:07] <sumanah>	 hi bianca!
[23:22:10] <sumanah>	 got a moment to chat?
[23:29:30] <sumanah>	 bianca: ping
[23:39:48] <sumanah>	 bianca: you have 2 emails from me in your inbox - it would be great if you had a moment to chat with Ryan_Lane or andrewbogott
[23:39:51] <sumanah>	 I think Ryan is afk.
[23:41:56] <sumanah>	 And soon I will be too.
[23:45:05] <andrewbogott>	 bianca:  I'll be in transit for most of the weekend, but I will try to catch up with you soon!  Feel free to pester me here on Tuesday if I haven't gotten back to you by then.