[00:00:39] New review: Krinkle; "(1 comment)" [mediawiki/extensions/SemanticMediaWiki] (master) C: -1; - https://gerrit.wikimedia.org/r/63397 [00:12:11] New patchset: Mwjames; "Fix some JS related docs in smw-core (JSDuck)" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/63397 [00:12:57] Change merged: jenkins-bot; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/63397 [00:19:19] New patchset: MarkAHershberger; "Remove remaining JSON release notes" [mediawiki/core] (REL1_21) - https://gerrit.wikimedia.org/r/63368 [00:19:19] New patchset: MarkAHershberger; "bump to 1.21.0rc5" [mediawiki/core] (REL1_21) - https://gerrit.wikimedia.org/r/63398 [00:30:07] New review: Krinkle; "fixme: various issues that suggest this has never been tested with jsduck, thus making the "fixes" r..." [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/63397 [00:30:28] New review: Krinkle; "Looks like it was self-merged while I was reviewing." [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/63397 [00:31:42] New review: Krinkle; "Indeed, we'll need to keep one of the exceptions so that projects can still have it V+2 in a pipelin..." [integration/zuul-config] (master) C: -1; - https://gerrit.wikimedia.org/r/61720 [00:39:35] Change merged: jenkins-bot; [mediawiki/core] (REL1_21) - https://gerrit.wikimedia.org/r/63368 [00:39:50] New review: Krinkle; "Also needs a @mixin block and to actually be documented (see the others and jsduck/config.json for m..." [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/63382 [00:41:01] New review: Krinkle; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63382 [00:46:38] Change merged: MarkAHershberger; [mediawiki/core] (REL1_21) - https://gerrit.wikimedia.org/r/63398 [00:57:42] Change merged: jenkins-bot; [mediawiki/core] (REL1_21) - https://gerrit.wikimedia.org/r/63365 [01:04:43] Change merged: jenkins-bot; [mediawiki/core] (REL1_21) - https://gerrit.wikimedia.org/r/63367 [01:04:43] Change merged: jenkins-bot; [mediawiki/core] (REL1_21) - https://gerrit.wikimedia.org/r/63366 [01:05:52] New review: Mwjames; "Thanks for the comments, I'll have a crack on this when I have more time on hand." [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/63397 [01:11:30] New patchset: Mwjames; "Use QUnit.start()" [mediawiki/extensions/SemanticResultFormats] (master) - https://gerrit.wikimedia.org/r/63402 [01:12:30] Change merged: jenkins-bot; [mediawiki/extensions/SemanticResultFormats] (master) - https://gerrit.wikimedia.org/r/63402 [01:14:40] heh, html5 added
[02:08:18] New patchset: Mwjames; "Add smw.dataItem.wikiPage.getText()" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/62536 [02:11:07] Change merged: jenkins-bot; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/62536 [02:12:47] New patchset: Mwjames; "srf.formats.datatables() use getNamespaceId()" [mediawiki/extensions/SemanticResultFormats] (master) - https://gerrit.wikimedia.org/r/63405 [02:13:33] New review: Mwjames; "Thanks QUnit ..." [mediawiki/extensions/SemanticResultFormats] (master) C: 2; - https://gerrit.wikimedia.org/r/63405 [02:13:42] Change merged: jenkins-bot; [mediawiki/extensions/SemanticResultFormats] (master) - https://gerrit.wikimedia.org/r/63405 [02:41:32] New patchset: Daniel Friesen; "Whitelist a bunch of url protocols." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63406 [02:45:17] New patchset: Daniel Friesen; "Improve language handling of page titles." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63378 [02:45:55] New review: Daniel Friesen; "Adding updates to hooks.txt." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63378 [03:07:52] New review: Peachey88; "looks sane" [mediawiki/core] (master) C: 1; - https://gerrit.wikimedia.org/r/63406 [03:14:36] New patchset: Arrbee; "Adding more test scenarios for ULS user settings" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/63140 [03:21:59] New patchset: Daniel Friesen; "Shorten our in-comment urls to the whatwg HTML spec." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63407 [03:24:24] New patchset: Daniel Friesen; "Whitelist the element." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63408 [04:34:21] mooeypoo: I just read your email. That's quite an ambitious first project [04:34:50] * bawolff hasnt read yoir code yet but is impressed [04:36:05] New review: Parent5446; "Looks good. Will give it a little while in case there are objections, and then merge." [mediawiki/core] (master) C: 1; - https://gerrit.wikimedia.org/r/63406 [04:38:03] hey bawolff :) [04:38:20] well to be honest, I wanted to do something "small" to help me since I'm not as well versed in git [04:38:33] ironically, this was started because i was lazy.... :D [04:39:12] but I think I might be in the right direction. It works well considering -- but it's a bit slow to run because of the multiple requests to external sites. [04:39:22] and thank you very much for the comment :) [04:39:26] Hey there are a lot of proverbs about laziness = good programmers [04:40:02] yeah, work hard on a project to save you repeated redundant work :) [04:40:57] New patchset: Rfaulk; "fix. abort method does not attempt reset to a start tag when one does not exist." [sartoris] (master) - https://gerrit.wikimedia.org/r/63409 [04:40:58] New patchset: Rfaulk; "fix. dulwich tagging." [sartoris] (master) - https://gerrit.wikimedia.org/r/63410 [04:41:41] I found a php library that enables use of git without having git installed. Might be a good way to replace it with the current screen-scraping of gerrit. [04:41:59] And there are also notes on how programmers hate many of the supposed rules of writing English you're taught. [04:42:18] Stupid english [04:42:41] Horrible language. Super ambigious [04:42:55] ^_^ Of course when you actually do some research those rules on writing are complete FUD [04:43:31] * mooeypoo has a good excuse; she's a foreigner [04:43:35] New patchset: Rfaulk; "fix. remove debugger statement + pep8." [sartoris] (master) - https://gerrit.wikimedia.org/r/63411 [04:43:36] :D [04:43:44] * bawolff does not [04:44:53] * DanielFriesen proudly puts a comma before and instead of omitting it in 3+ lists and puts his punctuation outside of a quote. [04:44:58] When I was a gsoc student in 2010, pretty much every code review I got said something about my spelling [04:45:13] as long as you don't misuse "their/they're/their", I think that makes you better than many other native english speakers. [04:45:16] s/and/'and' [04:46:16] DanielFriesen: I have "Strunk and White" and even they disagree on commas. [04:46:35] a, b, and c FTW! [04:46:56] I write stories, so I use commas for flow control, mostly. [04:47:06] ^^ see what I did, there.. [04:47:18] :p [04:49:12] Lol [04:50:06] Every language has weirdness in it, but English idioms can be really weird to non-native English speakers. [04:50:15] Even something as small and wide-spread as "Break a Twenty". [04:50:41] * DanielFriesen rips mooeypoo's bill in half. [04:50:46] First time someone asked me if I could break a twenty, I thought he was pulling my leg (which is another funny idiom, but at least graphically understandable) [04:51:02] Ah, see, even that isn't really breaking it. You'd need to freeze it first. [04:51:05] * mooeypoo thought a lot about this [04:51:10] lol [04:52:44] "To drink like a fish" is another weird idiom. I get the meaning, but if you think about it, it's really quite silly. [04:53:20] New review: Tim Starling; "Looks good to me, apart from minor issues inline." [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/58924 [04:56:52] I always assumed the break a 20 thing went back to old days where people used a $20 lump of gold that could be broken in two [04:57:17] I think it does, yeah. But as a newly-minted English speaker, it was a bit weird to hear for the first time. [04:58:30] When I get tired, some of the rules get jumbled in my head. That results in some amusing English errors (usually spoken, I catch them when I type) -- like, for instance, saying "I fat into it" instead of "fit" [04:58:38] I mean really, English. Sit/Sat. Fit/Fat. [04:58:58] Drink Drank... [04:59:31] Lol [04:59:46] Also, this one I get right, but it confused me at first -- One Louse, Many Lice. One Mouse, Many Mice. One House... [05:00:17] One mouse, many mouse... [05:01:48] Another confusing inconsistency: Rough, Tough, Cough --> Dough [05:08:16] New review: Tim Starling; "Why is auto-login still needed? Why not just wait until the user actually visits another wiki?" [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/62194 [05:44:27] New review: Daniel Friesen; "Good point Tim. Though when on the other wiki we'll need some way to know whether the user is logged..." [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/62194 [06:05:12] hm, anyone knows where I can find the mediawiki developer's styleguide? [06:05:16] is there any such thing? [06:05:37] mooeypoo: I'm sure there is... [06:05:48] yeah I thought so too, but I can't find it atm. [06:05:59] Huh, let me try (once Firefox works again) [06:06:10] oh! [06:06:13] http://www.mediawiki.org/wiki/Style_guide [06:06:22] Oh, is that it? [06:06:24] ... that was easy O.o haha [06:06:34] * Matthew_ high5s mooeypoo [06:06:36] Nice :) [06:06:45] :D hehe i was looking for it the wrong way. bah hehe [06:07:17] Heh, OK. That's quite all right, no harm done. [06:07:19] * Matthew_ reads [06:07:46] ah, spaces, spaces. I need to remember spaces. [06:23:07] If in doubt add spaces [06:23:24] yeah there's a bunch I need to change, style-wise. [06:23:46] oh! shoot, I forgot to take off the error reporting. bah. Good catch on that email. [06:24:36] I don't understand the first comment. What does it mean "Your RL module should have a remote path in addition to a local one" ? What's "RL module" ? [06:40:08] Mooeypoo: rl = resource loader [06:41:22] mooeypoo: http://www.mediawiki.org/wiki/Manual:Coding_convention is what you want [06:41:30] not Style_Guide [06:42:39] Mooeypoo: style wise your extension is pretty good [06:43:18] Mooeypoo: the only thing that stuck out to me was the scCommit class [06:45:23] Mooeypoo: since usually in mw class names start with capital letters, and are rarely abbreviated (however given its an ext and not core, people are somewhat less picky on the nitpicky things) [07:04:26] bawolff & p858snake|l_ : I looked at the styleguide http://www.mediawiki.org/wiki/Style_guide and it seems I'm not quite there... but do I go by that, or by the coding convention? [07:05:23] you want the coding convention page [07:05:28] bawolff: also, I was worried of using "Commit" as is in fear it may be already existing somewhere (I know there's a Git class used by MW) but I can change it, or I can rename it to uppercase SECommit [07:06:00] the style guide is for for the visual stuff, but don't really worry about that till you have something decent and functional [07:06:10] Ah, okay. I'll go over that, then. [07:06:13] Thanks :) [07:06:15] I would reccomend SECommit [07:06:44] Wow I didnt even know we had a ui style guide... [07:06:54] oh, wait, this: http://www.mediawiki.org/wiki/Manual:Coding_conventions/PHP <-- is what I was looking at, it's linked from the style guide [07:07:50] bawolff: we don't, the foundation does [07:08:09] :p [07:08:14] that randomly strung up without community involvement [07:08:17] *sprung [07:08:26] ROTFL... Comment spam has been getting stranger and more unique recently but this one is the nuttiest one I've ever seen. [07:08:33] okay, hm, I'm wondering about the language comment. The Special:Version page uses $wgLang->timeanddate( $datething ) and I used the same. The comment in the email suggests I use $this->getLanguage() but then it won't be consistent. [07:08:38] Should I be inconsistent? [07:08:51] mooeypoo: ->getLanguage is consistent. [07:08:52] as in, should I do "better than" Special:Version ? [07:08:58] Anything using $wgLang is wrong. [07:09:01] DanielFriesen: hi! thanks for the comments :D [07:09:18] Die wgLang [07:09:28] oh. Okay, Special:Version used it hehe :) that's why I did too. I'm looking for the ->getLanguage that converts date in the documentation now [07:09:45] The only reason some parts of Special:Version are still using $wgLang is probably because there are a few spots that are static when they shouldn't be and hence have no context. [07:09:48] Its the same [07:10:47] $this->getLanguage () usually points to the same object as $ wgLang [07:11:02] 99% of the time [07:11:18] Or more specifically $wgLang is basically RequestContext::getMain()->getLanguage(); [07:11:27] ohh [07:11:29] It's a deprecated global. [07:11:37] I didn't see that in the documentation, thanks for that! [07:13:03] Mooeypoo: btw some other things - you should have a qqq language with documentation about messages [07:15:13] Ooh, okay. [07:15:51] Well I guess it doesnt matter if your ext is not at translatewiki [07:16:28] Any reason for using git hub instead of wikimedia's git repo [07:18:27] oh, I don't know how to open a repo on gerrit? [07:18:40] You also use {{Special:Version}} in one of the messages. You probably meant to use [[ instead of {{ [07:18:43] And I wans't sure if this was a wayward attempt at an extension or if it was worthy of work [07:19:18] Opening a repo on gerrit is an annoyingly beurocratic process [07:20:03] Basically you ask chad. There is a request page somewhere on mediawiki.org [07:20:05] yeah I figured that I'll have GitHub first, and if it's good enough I can transport it into gerrit [07:20:27] that's also why I didn't put the external link yet btw.. i figured I'll wait with making it an Extension: page in mw [07:20:36] 'till I know if it's a worthwhile extension. [07:20:45] I definitely think its worthwhile [07:21:11] :) Thanks! [07:21:37] hi, anyone knows where the HTTP_X_SUBDOMAIN value comes from? Who sets it? [07:21:43] New review: preilly; "04:53:43 + tee pep8.txt" [sartoris] (master); V: -1 C: -1; - https://gerrit.wikimedia.org/r/63410 [07:21:52] grepping is not helping [07:22:42] btw, what do you think about the "fallback" of checking file modification time? I was really struggling with this decision, because it could be misleading if someone copied their folders or something like that, but I am not sure I have any other solution if git's not available. [07:23:26] Change merged: preilly; [sartoris] (master) - https://gerrit.wikimedia.org/r/63409 [07:25:22] Im not sure if you have to many other options [07:26:13] Change merged: preilly; [sartoris] (master) - https://gerrit.wikimedia.org/r/63410 [07:26:33] Yurik: i have never heard of that before - but if its a http header, probably varnish [07:26:51] Change merged: preilly; [sartoris] (master) - https://gerrit.wikimedia.org/r/63411 [07:28:38] bawolff, yes, but the closest i found in varnish is "set req.http.x-subdomain = ..." -- very close... experimenting... [07:29:51] I believe php turns x-foo headers into array keys like HTTP_X_FOO [07:30:20] For thinks like $_SERVER globals [07:30:27] *things [07:33:28] Mooeypoo: btw the version numbs on extension pages are usually horrible outdated [07:33:40] yea that was also my fear [07:34:08] that's why I went for the commits, I figured the version number itself is probably not reliable, but if we know how many times people pushed changes, we can guesstimate if the extension is outdated or not [07:35:01] but that means there'll be a problem with extensions that don't use git repos, or that use external git repos. I think maybe adding some option to add an external repo path in the settings could be the best solution for the moment. [07:35:05] Still looking into that, though. [07:35:14] I'm a little meh about the fact i'm basically screen-scraping gerrit. [07:35:38] gerrit web, I mean. I'm pretty sure there should be a better way to do it. [07:36:29] hm, another question, guys.. I just fixed DanielFriesen's suggestion regarding the ResourceLoader external path, but then I noticed that there's also an option for 'messages' array you can have in there too [07:36:49] It's optional [07:36:55] what's the difference with that message array and the i18n file? [07:37:19] mooeypoo: That array is entirely for RL, the system that delivers css and js. [07:37:46] I see, so if I want to have messages in my js file, that's where I put them? [07:37:53] No [07:37:59] oh :) [07:38:16] mooeypoo: Messages in RL refers to a list of keys of i18n messages. This basically tells RL to take the i18n messages for those keys and then serve them to the browser so that scripts can use them. [07:38:33] You define all the i18n in the i18n file. Only a list of keys that you intend to use go in the RL module. [07:38:40] ohh. Gotcha. [07:38:42] thanks! [07:39:09] bawolff, yes, just checked, it does this weird magic. Thx. stupid varnish [07:39:20] thx :) [07:40:14] Of course im hoping that mw isnt dependant on weird varnish magic... [07:40:55] New patchset: Amire80; "Update from upstream" [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/63416 [07:41:35] hm, I have to say, the MW docs leave some to be desired, they tend to be confusing at times. [07:42:10] but, granted, I'm a newbie who's used to super-noobish-friendly systems like wordpress extension documentation [07:42:21] That's putting it diplomaticly [07:42:45] hehe [07:42:46] Mooeypoo: its the number 1 bug! [07:42:58] Quite literally [07:43:11] is it? well.. hm. That's encouraging actually. I though I was being an idiot. [07:43:18] !bug 1 [07:43:34] Stipid bugs [07:43:43] *bots [07:44:11] Http://bugzilla.Wikimedia.org/1 [07:44:36] (Alas it is not #1 in priority) [07:45:47] maybe you guys should add it to next year's GSoC proposals.. I don't think I saw it there this year, could be something a student can do over the summer. [07:46:14] It's a little hard to do when your contributors are employees of a non-profit focused on using the software for their own sites and volunteers who contribute for their own personal reasons. [07:46:20] I can try and recommend some changes as I encounter them, but I definitely don't feel confident enough to *actually* change them. Maybe I sould post comments in the talk pages? [07:46:27] Mooeypoo: per the standard {{sofixit}} guilt trip, if you see anything easy to fix up, feel free to edit it ;) [07:46:41] * mooeypoo beat ya to it [07:46:55] vs. A for-profit company who's employees build the software for general use. [07:47:30] Don't try to win guilt-trip over me, I have a lifetime of experience by simply being Jewish. My mom after dinners and before trips? Bootcamp in guilt trip. [07:47:47] Lol [07:48:31] Remember when wmf hired someone to write docs? [07:48:36] DanielFriesen: Well, speaking of comparisons, here's a fair one: I've dabbled in some Wordpress extensions, and drupal and joomla, and I say this with utmost seriousness here -- you guys are the most newbie-welcoming so far. [07:49:05] Not much came of that if I recall [07:49:06] Wordpress, Joomla and Drupal communities (drupal a tad less so) are a bit snobbish and curt. It never felt comfortable to go and chat and ask questions like I can here. [07:49:29] Mooeypoo: that's excellent to hear [07:49:46] there's this attitude of "oh great, a newbie, why don't you read the damn docs in that page no one else found either before oyu come and ask this stupid question" [07:50:01] New patchset: J; "(bug 48300) escape title in edit url" [mediawiki/extensions/TimedMediaHandler] (master) - https://gerrit.wikimedia.org/r/63417 [07:50:12] See, we can't fall back to that [07:50:25] Because the docs suck too mich [07:50:29] So I think newbie-welcoming and true cooperation totally beats having good docs. [07:50:45] hehe i'll try not to take it as a reason not to improve them :P [07:50:53] New patchset: J; "(bug 46650) increase memory limit for frame extraction" [mediawiki/extensions/TimedMediaHandler] (master) - https://gerrit.wikimedia.org/r/63418 [07:50:58] Ideally we would have both [07:51:05] bawolff: Should I comment on the talk pages of those documents? I don't know if I should edit any of them directly (or if I even can) [07:51:48] You should be able to edit them [07:52:18] Some might require approval before the edit goes live [07:52:41] 'be able to' perhaps. "Should" not sure.. I'm still newbie, not sure I trust myself to write proper corrections. But I can point out things that might be confusing.. [07:53:17] The 'getting started' pages are pretty good [07:53:27] Feel free to do whatever you feel most comfortable with [07:53:49] but then the drill-down documentations are lacking. I resorted to looking at the raw code in https://doc.wikimedia.org/mediawiki-core/master/php/html/index.html [07:54:33] Remember that if you say something wrong its easy to undo yoir edit [07:54:48] Hm, true, but people may be confused until the error is caught [07:54:54] unless the fixes require review like code? [07:55:24] Some but not all pages are configure to require review [07:55:57] * bawolff mutters about flagged revs being ineffective on mw.org [07:56:34] Yeah the code comments are quite helpful imo [07:56:55] They are, but then you usually have to backtrack a lot to see where something comes from [07:57:07] Although I say that as a non newbie, I imagine rtfs is a bit intimidating [07:57:15] There's something about an XML reader as part of MW core I think? I was thinking of using it at first for the DOMDocument parsing [07:57:27] then gave up completely and resorted to the built in domdocument and xparse [07:58:02] We have an xml class, but its only for generating xml [07:58:08] Or really html [07:58:11] bawolff: they're excellent when you want to understand the code in the core, but as documentation to what you should use in your own extensions, it's messy to follow. [07:58:41] The xml vs html class divide is weird [07:58:58] Yeah, that's what I saw. [07:59:55] .... omg [08:00:07] I was about to say "you know what will be awesome? gerrit REST api" [08:00:15] Lol [08:00:15] and then I found that.. [08:00:34] https://gerrit-review.googlesource.com/Documentation/rest-api.html [08:01:00] hm. Seems you still need to authenticate, though. Not helping my extension. Could've saved the screen scraping and multiple html file reads. [08:01:02] I have a script actually that does stats from it [08:02:00] So does Susan (his is much better than mine) [08:03:39] hm. The way I do it works, but it takes a bit to load because I ask for several gerrit repo pages and parse their html for details. [08:04:13] if I can work with JSON replies, that'll make everything faster. I'll check into that. As far as I can see, the gerrit people have been talking about it in the groups but it wasn't yet fully implemented. [08:05:21] I do believe that we rin somewhat of a bleeding edge version of gerrit [08:05:29] *run [08:13:51] hmm. [08:20:52] oh! someone already made an extension like mine with external repos, but git-only. Collab-time? woot. [08:23:59] Mooeypoo: btw in regards to api, all wikimedia repos are also duplicated on gerrit, so you could potentially use their api [08:24:19] *duplicated on github [08:24:31] oh.. really? [08:24:36] * bawolff curses virtual keyboards [08:24:44] regularly? [08:25:00] Near instantly afaik [08:25:13] oh that's great [08:25:15] New patchset: Amire80; "Put the cog icon in the middle of the section title" [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/63419 [08:25:18] the only problem is finding where's what [08:25:48] gerrit is convenient in that aspect, all extensions have the same url with the difference of name only. Might have to find a creative solution to get them off GitHub [08:26:35] https://github.com/wikimedia [08:27:03] I think there is reasonble naming conventions [08:27:08] you know, thinking about it, I should also use the summary page instead of the shortlog page when i call the repo. I call the shortlog and then use the link of each commit to call its details..It was easier to parse but it is prbably redundant. [08:31:57] hm bawolff there's another problem with using the release convention -- Cite extension for instance, wasn't updated for a while, but I submitted a (tiny) fix to it a week ago for CSS. That doesn't really warrant a separate release [08:32:30] but it *was* responsible for a bug that VE caught, and was solved after upgrading Cite. [08:32:57] People backport changes to old release branchs sometimes [08:33:00] Not sure the releases, even if they were by convention, would've helped here, especially when you don't have a single "owner" for a lot of the extensions [08:33:34] yeah, that too. I think it may be safer to stick to reading actual changes [08:34:35] For example there was recently a fix for intersection extension that I backported to REL1_19 branch [08:35:22] Generally its only safe to assume the version from the release branch of the ext works with that version of mediawiki [08:35:31] New patchset: Spage; "Reposition nested FancyCaptcha, conditional help." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/62542 [08:36:06] New patchset: Spage; "Reposition nested FancyCaptcha, conditional help." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/62542 [08:36:18] New patchset: Spage; "Reposition nested FancyCaptcha, conditional help." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/62542 [08:36:29] And we encourage people to downliad the version corresponding to their version of mediawiki [08:36:57] Hm, good point. [08:38:00] oi, it's 5am again. That hour keeps coming every day, despite my attempts to ignore it. [08:38:04] Otoh its rare for an extension change to get backported for most extensions [08:38:11] Lol [08:38:27] Wait, are you in my timezone? [08:38:35] and I have exams this week.. I do studying during the day, and relaxing programming by evening, then before I know it, eveing turns into morning. [08:38:52] bawolff: NYC. Otherwise known as "The Center of Everything[TM]" [08:38:52] (Atlantic canada) [08:39:44] Canada! I never been, but I hear you guys have beauuuutiful scenery. [08:39:53] 1 hour behind me I think [08:40:03] You're at 3:40 now? [08:40:15] 540 [08:40:24] actually "behind you" woudl mean you're at-- yes. [08:40:29] No, it's 4:40 here. [08:40:50] so *almost* 5am. That's better than yesterday, I noticed 5 was coming when it was already 6am. [08:41:07] Canada scenery verries where you go [08:41:36] We also have a lot of boring scenery no one mentions in the travel guide [08:41:59] I'm sure. It's like that everywhere. [08:42:27] Other than New York, if you believe The Media, of course. It's all awesome and cookies here. [08:42:29] Most of the really pretty stuff is in the west [08:42:34] And Times Square. [08:42:56] I hear america is the land of unicorns [08:43:13] oh is it? Well, the Niagara Falls are in the east [08:43:29] It is. The run rampant. [08:43:38] Well unicorns armed with machine guns depending on which news station you eatch [08:44:12] Also, everything's huge here. [08:44:29] Of course my reference is puny in comparison, I'm originally from Israel [08:45:03] first time I saw a Wallmart I thought the entire state of Israel can be housed in it. [08:45:19] Yeah there is Niagara, but a large part of tourism is banf and jasper [08:45:24] Lol [08:45:56] And yet one can never actually find what you are looking for in a wallmart [08:47:11] I hear you need a map. [08:48:59] Gps [08:49:25] And if you're there too long, they call search and rescue. [08:52:50] You're getting your spaceman back soon, bawolff [08:52:59] Did you see his *awesomely win* music video? [08:53:13] I don't think so [08:53:34] https://www.youtube.com/watch?feature=player_embedded&v=KaOC9danxNo#! [08:53:39] pure space win. [08:55:42] That's awesome [08:55:44] David Bowie tweeted him back "Hallo Spaceboy" [09:20:47] ok, I'm off to bed, finally... [09:20:56] night guys, thanks for the help! [09:25:03] Night [09:25:52] New patchset: Siebrand; "Update formatting." [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63386 [09:30:56] New patchset: Spage; "Blank some new Create account messages by default" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/61395 [09:32:16] New review: Spage; "Review feedback." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/61395 [10:00:34] New patchset: Zfilipin; "Updated parallel_tests Ruby gem" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/63427 [10:08:54] New review: Peachey88; "> Though when on the other wiki we'll need some way to know whether the user is logged" [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/62194 [10:13:03] New review: Daniel Friesen; "When the user visits the wiki." [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/62194 [10:30:05] hi all [10:30:21] i've got a simple PHP style question [10:30:30] which isnt answered in the PHP code conventions [10:30:48] in skins/Vector.php and skins/MonoBook.php [10:30:54] there are pieces of code that go: [10:31:02] if ( $foo == $bar ) { ?> [10:31:07]
some content
[10:31:12] [10:31:15] instead of: [10:31:26] if ( $foo == $bar ): ?> [10:31:32]
some content
[10:31:38] [10:31:46] i fancy that the second way is clearer [10:31:53] is there a reason for using the first ? [10:38:35] question answered #mediawiki [10:54:44] New review: Denny Vrandecic; "We need to unstuck this. I do not follow Krinkle's argument given at https://bugzilla.wikimedia.org/..." [mediawiki/core] (master) C: 1; - https://gerrit.wikimedia.org/r/62785 [10:58:47] New review: Aklapper; "There shouldn't be an issue: longdescs.thetext does not store linebreaks when adding a new comment, ..." [wikimedia/bugzilla/modifications] (master) - https://gerrit.wikimedia.org/r/62218 [11:00:17] gerrit spam! [11:00:22] New patchset: Ori.livneh; "Add Redis module" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63438 [11:00:22] New patchset: Ori.livneh; "Enable uploads and InstantCommons by default" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63439 [11:00:22] New patchset: Ori.livneh; "Add mediawiki::parsoid class" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63440 [11:00:23] New patchset: Ori.livneh; "Vagrantfile shared folder config: f/dmode to 755" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63441 [11:00:23] New patchset: Ori.livneh; "Add redis::php class & 'needs_update' param for extensions" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63442 [11:00:23] New patchset: Ori.livneh; "Add roles for VisualEditor, EventLogging, MobileFrontend, Echo and GettingStarted" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63443 [11:03:11] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63438 [11:03:28] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63439 [11:03:43] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63440 [11:04:02] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63441 [11:04:16] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63443 [11:04:16] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63442 [11:04:27] ..sorry. [11:11:54] New patchset: Ori.livneh; "Add explanatory comment re: run stages" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63445 [11:12:16] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63445 [11:12:41] New patchset: Siebrand; "Remove unused local variables" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63446 [11:15:27] New patchset: Siebrand; "Use localised count" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63447 [11:26:49] New patchset: Siebrand; "Make TranslateSandboxReminderJob::newJob() static" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63448 [11:29:04] New patchset: Siebrand; "JSON standard does not allow comments." [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63449 [11:40:12] New patchset: Siebrand; "Update to jQuery UI Position 1.10.3" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63453 [11:40:12] New patchset: Siebrand; "Update jQuery Autosize to master" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63454 [11:47:18] New patchset: Ori.livneh; "Add a Rakefile for running puppet-lint against repo" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63455 [12:00:14] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63455 [12:08:37] New patchset: Matthias Mullie; "Claim anon feedback after sign up" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/33383 [12:48:12] New patchset: Amire80; "Add a MediaWiki preference for enabling ULS IME" [mediawiki/extensions/UniversalLanguageSelector] (master) - https://gerrit.wikimedia.org/r/61770 [12:52:12] @User:Jeremyb [12:52:14] New patchset: Matthias Mullie; "Quickly enable/disable AFTv5" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/61797 [12:52:40] hi I want to join Lab bot project [12:52:52] would you please add my name to this project? [12:54:07] @Jeremyb [12:58:09] New patchset: Ori.livneh; "Improve in-line comments for MFE and VE roles" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63457 [12:58:26] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63457 [13:02:33] New patchset: Luis Felipe Schenone; "(Bug 5798) Prevent overflow in the footer when there are no links" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63458 [13:04:19] New review: Zfilipin; "Runa, what changed in patch set 3? You have uploaded a few new feature files?" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/63140 [13:05:37] New patchset: Matthias Mullie; "Feedback link on article pages" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/35168 [13:05:38] New patchset: Matthias Mullie; "Quickly enable/disable AFTv5" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/61797 [13:19:14] New patchset: Siebrand; "Fix incorrect file name" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63459 [13:33:10] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63458 [13:44:00] ugh, why do people /part ? [13:47:08] hrmmm, reza registered with NickServ on August 15, the same hour as NickServ last saw them [13:50:17] Change merged: Zfilipin; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/63277 [13:51:28] New review: Zfilipin; "I have merged the commit into master since Karim gave +1 to my changes." [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/63277 [13:56:47] New patchset: Siebrand; "Tweak error message" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63462 [13:56:47] New patchset: Siebrand; "Message tweak." [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63463 [13:56:48] New patchset: Siebrand; "Remove unused message" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63464 [13:56:50] New patchset: Siebrand; "Make proofread terminology more consistent" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63465 [13:56:51] New patchset: Siebrand; "Remove key not used in resource" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63466 [14:17:46] New patchset: Matthias Mullie; "Bug 43190 - AFTv5 API results should be wrapped in a {'query':results} object (for JSON)" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/39038 [14:24:04] New patchset: Reedy; "Update Wikidata related branch points" [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/63470 [14:24:20] Change merged: Reedy; [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/63470 [14:26:47] New patchset: Reedy; "Remove final commit as no changes are there" [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/63471 [14:27:02] Change merged: Reedy; [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/63471 [14:28:09] New patchset: Matthias Mullie; "Bug 43190 - AFTv5 API results should be wrapped in a {'query':results} object (for JSON)" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/39038 [15:34:52] Change merged: jenkins-bot; [mediawiki/skins/erudite] (master) - https://gerrit.wikimedia.org/r/32035 [15:48:39] New review: Krinkle; "Aside from whether this change needs tests, please try harder. There is no such thing as "not having..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63282 [15:48:55] New patchset: Reedy; "Update VisualEditor to master" [mediawiki/core] (wmf/1.22wmf4) - https://gerrit.wikimedia.org/r/63479 [15:54:42] New review: Anomie; "@Tim Starling:" [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/62194 [15:56:40] hashar: Krinkle heya, this "continuous integration" meeting that is on the WMF Engineering calendar that starts in 5 minutes, is it meant to be public? [15:56:43] I just noticed it, and well, it is relavant to andre__ and I's interests :) [15:56:46] greg-g: I think it is just timo and I [15:57:11] greg-g: And it is meant to be public yes [15:57:16] greg-g: What do you mean by public exactly? [15:57:30] well, joinable by people from WMF, I suppose [15:57:36] I just didn't see an announcement about it [15:57:53] It is our bi-weekly standup [15:57:54] like many teams have [15:57:54] oh, interesting! [15:58:01] I didn't realize [15:58:05] so, sort of private (not for non-staff) [15:58:06] and that is the very first one :-] [15:58:08] andre__: ^^ [15:58:11] however since we're both remote... [15:58:15] * greg-g nods [15:58:16] we do it via hangout [15:58:39] I didn't make it private, I mean, people can walk into office standups as well. You're welcome to peek :) [15:58:45] andre__ and I were wondering if we should join, but I'll leave you to it for now ;) [15:58:51] hashar: shall we? [15:58:54] I din't realize either, but I've copied it to my calendar now [15:58:59] :) [15:59:06] It's definitely great to know that it exists, so I'll try to be around from time to time for it [15:59:15] (plus CI in Bugzilla has quite some unprioritized tickets) [15:59:18] andre__: You should be able to add yourself to it (instead of copying it, thus disconnecting from the event loop) [15:59:35] edit the engineering calendar event and invite yourself if you like [15:59:42] definitely join [16:00:08] Krinkle: we can get a one / one later on if you want :-] [16:00:08] will take a minute, session lost with google [16:00:13] Krinkle, hashar: Any chance of integrating Selenium/cross-browser automated UI testing? That's my number 1 ask (after stability). ;-) [16:00:44] James_F: zeljkof is taking care of Selenium stuff with chrismcmahon leading it. [16:01:02] James_F: given the length of selenium tests, it is unlikely we will ever have them on patch submission :-] [16:01:25] hashar: Krinkle and-re has a office hours right now in -office, so he'll be multitasking for this one at least ;) [16:01:38] greg-g: andre__ I have added you to the even (at least for the one right now) [16:01:47] awesome, thanks much [16:01:54] hashar: join the hangout [16:02:01] hashar: :-( [16:02:09] hashar: CI without CI sucks. [16:02:11] Krinkle: yeah whenever google let me in [16:02:25] James_F: that takes ages to run :-] [16:02:42] hashar: Sure, but that's normal. Certainly a fail of Selenium should block merge. [16:02:58] then each merge will have to wait X hours :-] [16:03:00] hashar: And you can always throw a Selenium-runner box at it if hardware's an issue. [16:03:12] are the selenium tests ~10mins? [16:03:17] s/are/aren't/ [16:03:33] hashar: So be it. We've been completely broken in VE master for some browsers five times in the past month because of the lack of Selenium. [16:03:45] James_F: Selenium tests right now are too slow and too expensive to run on each patch set [16:03:56] James_F: Q/A (not CI) is currently hangling these asynchronously [16:05:03] Krinkle: Yes, I know, but with no reply channel it might as well not exist. It should only merge if the tests run. [16:05:14] Krinkle: And "too expensive" is a value judgement for each team. :-) [16:06:52] James_F: Contact Željko, he has a bot that runs them daily (if not more) [16:07:01] twice a day iirc [16:08:32] Krinkle: Per our stand-ups and your comments, we've been waiting for the Jenkins integration of that for 6 months now, if you remember. Now you tell me it's not going to happen? :-( [16:14:50] New patchset: Matthias Mullie; "Bug 43167 - list=articlefeedbackv5-view* should accept page titles as an option" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/63480 [16:15:51] New patchset: Matthias Mullie; "(bug 42722) Show Protect Feedback (action=protect) for pages in whitelisted namespaces only" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/63120 [16:24:27] 16 [16:24:54] * Nemo_bis stares at self [17:54:37] RoanKattouw: https://gerrit.wikimedia.org/r/#/c/63489/ [17:55:36] AaronSchulz: +2ed [17:56:08] New patchset: Matthias Mullie; "(bug 42722) Show Protect Feedback (action=protect) for pages in whitelisted namespaces only" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/63120 [17:56:14] New patchset: Aaron Schulz; "Added wiki ID to backend copy script errors." [mediawiki/core] (wmf/1.22wmf3) - https://gerrit.wikimedia.org/r/63492 [17:56:41] okay. fixed the flow server. [17:57:50] Change merged: Reedy; [mediawiki/core] (wmf/1.22wmf4) - https://gerrit.wikimedia.org/r/63490 [17:57:57] Change merged: jenkins-bot; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63446 [18:06:30] Reedy: Thanks man [18:08:10] Change merged: jenkins-bot; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63447 [18:10:03] New patchset: Mwjames; "SMW\ApiResultPrinter check for is_integer/is_bool as well" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/63494 [18:11:15] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63489 [18:11:36] * MaxSem pokes ori-l [18:11:53] New patchset: Mwjames; "SRF\EventCalendar eliminate parameters bool hack" [mediawiki/extensions/SemanticResultFormats] (master) - https://gerrit.wikimedia.org/r/63495 [18:18:29] Change merged: jenkins-bot; [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/63494 [18:19:01] Change merged: jenkins-bot; [mediawiki/core] (wmf/1.22wmf3) - https://gerrit.wikimedia.org/r/63492 [18:19:21] New review: Mwjames; "Dependency merged [1]" [mediawiki/extensions/SemanticResultFormats] (master) C: 2; - https://gerrit.wikimedia.org/r/63495 [18:20:15] New patchset: Aaron Schulz; "Strengthened validation for HashRing a bit." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63497 [18:21:06] New patchset: Umherirrender; "jquery.byteLimit: Fix infinite loop if text is longer than max" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63282 [18:21:08] Change merged: Demon; [mediawiki/extensions/timeline] (master) - https://gerrit.wikimedia.org/r/47319 [18:21:30] Change merged: jenkins-bot; [mediawiki/extensions/SemanticResultFormats] (master) - https://gerrit.wikimedia.org/r/63495 [18:22:29] Change merged: jenkins-bot; [mediawiki/extensions/MWSearch] (master) - https://gerrit.wikimedia.org/r/62443 [18:22:42] Change merged: Demon; [mediawiki/tools/release] (master) - https://gerrit.wikimedia.org/r/56194 [18:23:27] New patchset: Aaron Schulz; "Clear stat cache before determining local copy error message." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63486 [18:23:41] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63406 [18:25:36] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63491 [18:30:23] New review: Alex Monk; "(1 comment)" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/63120 [18:32:10] New review: Umherirrender; "Another way is to remove the methods, but that should be done later, if at all." [mediawiki/core] (master) C: 1; - https://gerrit.wikimedia.org/r/63338 [18:33:30] New patchset: Matthias Mullie; "Post without comment" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/37229 [18:35:17] New patchset: Umherirrender; "jquery.byteLimit: Fix infinite loop if text is longer than max" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63282 [18:36:47] New patchset: Krinkle; "jquery.tablesorter.test: Reset transform tables in mw.config" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63482 [18:37:55] ^demon: Ta for the merge [18:38:04] <^demon> yw [18:40:10] New patchset: Pgehres; "Unsetting the email address for a user when the email address is invalidated." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/61907 [18:42:58] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63482 [18:43:24] New patchset: Matthias Mullie; "(bug 42722) Show Protect Feedback (action=protect) for pages in whitelisted namespaces only" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/63120 [18:45:16] greg-g: I'd like to use the lightning deploy window today please. I also suggest that we ceremonially name every 2nd Monday's LD window as the Roan Kattouw ZOMG VisualEditor Is Terribly Broken Must Fix It Nao Window [18:46:29] RoanKattouw_away: yes indeed, you may, and feel free to name it such ;) [18:47:39] Change merged: jenkins-bot; [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/63120 [18:50:21] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63338 [18:51:17] greg-g: BTW, are we moving to the one-week deployment window model? 'Cos I'm about to have to create new milestones for the VE product, and... :-) [18:52:03] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/62827 [18:53:14] James_F: we are. what it looks like and when is tbd. That is the email I'm thinking about/writing in my head during lumch now ;) [18:53:20] mmm, lumch [18:53:41] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63497 [18:59:17] New patchset: Aaron Schulz; "Clear stat cache before determining local copy error message." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63486 [19:02:40] New patchset: Demon; "Allow setting of connection timeouts for HTTP requests using cURL" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/58247 [19:15:50] greg-g: OK. Will create a single other two-week one and leave it for now. :-) [19:22:09] New patchset: Wikinaut; "(bug 47264) Special:Version: add last local update timestamp for core and extensions" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/59373 [19:24:18] New patchset: Yurik; "(Bug 44790) API bug with sqlite returns duplicate categories" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/48098 [19:28:29] MaxSem: oops, missed your ping. What's up? [19:29:16] New patchset: Adamw; "WIP: banner DonationForm backend support" [mediawiki/extensions/DonationInterface] (master) - https://gerrit.wikimedia.org/r/61929 [19:29:26] ori-l, now that labs is fixed, I'm continuing with EL. basically, what's needed to check from the web how it works? [19:30:50] not totally sure what you mean. you want to log events and see them come in? [19:30:57] yes [19:31:21] what data store (if any) do you want them going into? mysql? there isn't a canonical one per se; eventlogging is a pipe. [19:31:53] there's good support for just streaming and not writing at all, dumping to a file, writing to mongo, or writing to a mysql db [19:32:27] 'just streaming' may be perfectly adequate if you just want the ability to see things as they come in in a separate terminal window, say [19:33:01] if you actually want to run sanity checks against the data you log in dev, you probably want to set up a mysql database [19:33:13] let me look at your manifest; just a sec [19:34:30] currently, i have only a vanadium analog [19:35:02] do you have an actual URL endpoint set up? [19:35:30] no [19:35:33] meaning: varnish instance that 204s a specific request pattern, and a varnishncsa instance that follows the shm log and sends those out by udp [19:35:53] ah, this I have [19:36:03] (need to check how it works) [19:36:08] New review: Alex Monk; "(1 comment)" [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/53979 [19:36:16] New review: Alex Monk; "Needs rebase." [mediawiki/extensions/ArticleFeedbackv5] (master) C: -1; - https://gerrit.wikimedia.org/r/53979 [19:37:37] MaxSem: there's a sample vcl file in /examples in the EL repo, and there's also this: https://wikitech.wikimedia.org/w/index.php?title=EventLogging&oldid=63180#Varnish [19:37:55] lo [19:38:03] hi hashar [19:39:45] MaxSem: another options (if you have the time & interest) is to write a simple MediaWiki API endpoint that behaves in the same way -- like ClickTracking used to have. It's a perfectly viable approach for smaller wikis, and it would make EL usable as a standalone extension. [19:40:06] ori-l, the former should be taken care of in https://gerrit.wikimedia.org/r/#/c/60923/3/manifests/role/cache.pp [19:41:05] I don't need an API endpoint, what's needed is a way to check on beta if EL is still working (with mobile) [19:42:14] ok, so if you're on 10.4.1.14 / deployment-eventlogging and run nc -ul 8422, and then curl the event.gif endpoint, do you see it show up? [19:42:38] meh, that instance was recreated, need to update the manifest:) [19:43:14] New review: Nikerabbit; "I43ace2120812 indicates comments shouldn't be a problem." [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63449 [19:46:47] MaxSem: ori-l : are you setting up EventLogging on labs? :-D [19:46:48] yes [19:46:49] MaxSem: if you get that up, the rest of the EL module would have been reusable as-is in labs, except i did not parametrize the database settings, which I'll do now [19:47:09] hashar: MaxSem is doing it and i'm mostly deflecting his questions with more questions :) [19:47:42] "why are you asking?" [19:47:52] are you doing that on 'beta' ? :D [19:47:55] no, why are you asking?:P [19:47:58] Change merged: jenkins-bot; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63448 [19:48:04] haha [19:48:16] New review: Nikerabbit; "Really? Only comment changed?" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63453 [19:48:21] hashar, yes - on deployment-prep [19:49:32] oy. this was my first puppet module and it shows. [19:49:46] i'll make it parametrized so that it's usable for labs. [19:50:16] you guys are awesome [19:50:25] * hashar logs an entry [19:51:31] and here you are https://www.mediawiki.org/wiki/Beta_cluster/status#2013-05-13 [19:51:55] I actually started it like 2 weeks ago [19:52:11] then new instance creation broke [19:52:23] New review: Nikerabbit; "Ooops." [mediawiki/extensions/Translate] (master) C: 2; - https://gerrit.wikimedia.org/r/63459 [19:52:48] humm err: /Stage[main]/Rsync::Server/Service[rsync]: Failed to call refresh: Could not start Service[rsync]: Execution of '/etc/init.d/rsync start' returned 1: at /etc/puppet/modules/rsync/manifests/server.pp:56 [19:53:39] New review: Pgehres; "Or, maybe they are just broken. This change should be covered by them as all it does is set the ema..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/61907 [19:53:44] Change merged: jenkins-bot; [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/62982 [19:54:06] MaxSem: yeah that is really dumb from puppet. It should log stderr [19:54:34] MaxSem: try starting it manually, maybe that dump some useful error [19:54:46] I'm doing it manually [19:55:04] ah, rsyncd, not puppet [19:55:27] Change merged: jenkins-bot; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63459 [19:56:05] MaxSem: is that instance going to use mediawiki? [19:56:11] no [19:56:16] good [19:56:18] only an EL endpoint [19:56:41] cause I still have to migrate /data/project, by default it uses the glusterFS shared volume [19:56:50] I have made the apaches to use a NFS server instead [19:58:11] MaxSem: Address already in use, rsyncd[18563]: unable to bind any inbound sockets on port 873 [19:58:25] pfft:) [19:58:33] from /var/log/syslog [19:59:04] crazy [19:59:20] New review: Nikerabbit; "recheck" [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63462 [19:59:40] just re-ran and it works [20:00:51] some dependency issue maybe [20:02:46] mmm slooooooow [20:02:55] Queue lengths: 170 events, 1 results. [20:05:16] Build has been executing for 11 min on master [20:05:22] https://integration.wikimedia.org/ci/job/mwext-Translate-phpcs-HEAD/1283/ [20:08:13] New review: Demon; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63486 [20:08:36] New review: Kaldari; "(1 comment)" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63007 [20:12:18] New review: Aaron Schulz; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63486 [20:14:17] New patchset: Aaron Schulz; "Clear stat cache before determining local copy error message." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63486 [20:15:23] hashar: isn't 17 minutes for phpcs excessive? [20:15:40] Nikerabbit: depends on what you lint [20:15:52] see https://integration.wikimedia.org/ci/job/mwext-Translate-phpcs-HEAD/1283/ [20:15:53] Nikerabbit: bug report it please, we can take a look at it later on [20:15:55] <^demon> It was doing i18n files. [20:15:58] i am on conf calls tonight [20:16:02] ah i18n :( [20:16:06] supposed to be ignored [20:16:07] as far as I know i18n files are supposed to be ignored [20:16:26] New review: Alex Monk; "(1 comment)" [mediawiki/extensions/ArticleFeedbackv5] (master) C: -1; - https://gerrit.wikimedia.org/r/56877 [20:16:40] the alias file is only 1000 lines long [20:16:42] <^demon> eg: 20:15:12 Processing Translate.i18n.php [247996 tokens in 35026 lines]... DONE in 1144 seconds (2738 errors, 2314 warnings) [20:17:34] oh it finally finished [20:17:43] <^demon> Yeah, queue's going down now [20:19:41] New review: Alex Monk; "(1 comment)" [mediawiki/extensions/ArticleFeedbackv5] (master) C: -1; - https://gerrit.wikimedia.org/r/42759 [20:22:18] MaxSem: "Because the problem with EventLogging's puppet code languages is it's got all this implicit environment that it carries around with it. You wanted a banana but what you got was a gorilla holding the banana and the entire jungle." [20:22:30] that was joe armstrong on OOP, but it applies http://www.johndcook.com/blog/2011/07/19/you-wanted-banana/ [20:22:43] heh [20:22:50] anyways, i see what you mean and am fixing it [20:23:14] thanks:) [20:26:14] New patchset: EBernhardson; "Database updates for respecting oversight within Echo" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63076 [20:27:41] New patchset: Anomie; "Fix uncaught exception in LuaStandalone" [mediawiki/extensions/Scribunto] (master) - https://gerrit.wikimedia.org/r/63564 [20:27:47] Nikerabbit: sorry back in [20:28:34] Skipping 13,534 KB.. Full Log [20:28:53] New patchset: Anomie; "Invalid data should cause an error, not a silent placeholder" [mediawiki/php/luasandbox] (master) - https://gerrit.wikimedia.org/r/63565 [20:29:02] hmm. No [20:31:02] New review: Demon; "(1 comment)" [mediawiki/extensions/Echo] (master) C: -1; - https://gerrit.wikimedia.org/r/63076 [20:31:34] Change merged: jenkins-bot; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63463 [20:31:34] Change merged: jenkins-bot; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63464 [20:31:34] Change merged: jenkins-bot; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63462 [20:31:39] New patchset: EBernhardson; "Database updates for respecting oversight within Echo" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63076 [20:33:21] Nikerabbit: filling a bug, will take care of it over the week if possible [20:33:31] hashar: ;)! [20:33:49] Nikerabbit: I probably need to overhaul the phpcs shell scripts we are using [20:33:54] New patchset: EBernhardson; "Database updates for respecting oversight within Echo" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63076 [20:38:36] New review: Alex Monk; "Fixes some things broken in I7ee9fc31 PS3." [mediawiki/extensions/ArticleFeedbackv5] (master) C: 2; - https://gerrit.wikimedia.org/r/60678 [20:38:49] Change merged: jenkins-bot; [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/60678 [20:43:04] New patchset: Anomie; "Add JavaScript login check against the central wiki" [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/58924 [20:43:33] New patchset: Hashar; "phpcs-HEAD now ignores i18n files" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/63571 [20:43:46] New review: Anomie; "(1 comment)" [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/58924 [20:43:48] New review: Hashar; "For phpcs-HEAD : https://gerrit.wikimedia.org/r/63571" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/58638 [20:45:12] New review: Hashar; "Deployed for mwext-Translate-phpcs-HEAD . Still have to update all the other jobs though." [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/63571 [20:45:38] Nikerabbit: that will be tracked under https://bugzilla.wikimedia.org/show_bug.cgi?id=48420 . I have updated the Translate job and retriggering a build [20:48:52] New patchset: EBernhardson; "Database updates for respecting oversight within Echo" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63076 [20:48:52] New patchset: EBernhardson; "Respect revision suppression within Echo" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63572 [20:51:25] hey sumanah, how are you? [20:51:39] hi hoo, I am all right, how are you? [20:51:56] Same here ;) [20:52:38] hoo: I am a little :( because of some of the conversation in https://en.wikipedia.org/wiki/Wikipedia_talk:Notifications recently [20:53:32] sumanah: Wow... that's a lot of discussion [20:53:47] e.g. "seemed to me that the developers viewed the bar as an old and low tech way of doing something that could be done with new flashy java code and took action." (I believe the person means "JavaScript" and not Java) [20:54:18] New review: Kaldari; "(1 comment)" [mediawiki/extensions/Echo] (master) C: -1; - https://gerrit.wikimedia.org/r/63070 [20:54:33] That's a common mistake... once somebody said he was against using JavaScript because of Oracle's security flaws -.- [20:55:11] If you felt like dropping some knowledge in there I would appreciate it; I also understand if you don't want to [20:55:43] sumanah: Maybe I take a deeper look later ;) [20:55:48] OK :) [20:55:54] sumanah: do you have any idea about sightseeing around the Hackathon? I plan to arrive a bit earlier on the 23rd to get a quick look at the town, but of course I don't want to do taht alone ;) [20:56:27] Also this weekend my husband and I baked brownies with chocolate, marshmallows, bourbon, and cacao nibs. He usually bakes in our household and he had to sort of talk me down from adding NINE things [20:56:36] hoo: I have never been to Amsterdam! [20:56:49] and I actually enjoy sightseeing alone but I guess you are different [20:57:06] sumanah: Neither have I ;) [20:57:20] :D I love brownies [20:57:29] So I think you are asking "are there plans for group tourism" right? [20:57:52] (I figure https://en.wikivoyage.org/wiki/Amsterdam has some info about general things to see) [20:58:04] sumanah: Indeed, just a few people with the same idea would be nice [20:58:50] There is going to be a city cruise one of the nights of the event, but it would also be reasonable for you to post to https://www.mediawiki.org/wiki/Talk:Amsterdam_Hackathon_2013 and mention the specific things you're most interested in seeing -- the Rijksmuseum, for instance -- and then people could coordinate [20:59:53] New patchset: EBernhardson; "Respect revision suppression within Echo" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63572 [21:01:24] sumanah: I'll post here... I really need to book my train tickets today [21:01:29] * there [21:01:52] thanks hoo! Looking forward to a bit of sightseeing [21:02:34] MaxSem: I'm going to modularize the MySQL dependency too, and then it should be quite straightforward to pick and choose an adequate setup for labs. [21:02:42] I'll get to it sometime this week. [21:04:24] hoo: I have now fallen into Wikivoyage :-) https://en.wikivoyage.org/wiki/Amsterdam/Canal_Ring There is a pancake restaurant called "Pancakes! Amsterdam" [21:05:17] sumanah: :D Do you know of https://meta.wikimedia.org/wiki/Association_of_Stroopwafel_Addicts ? [21:05:44] yes! [21:05:53] ori-l, thanks - meanwhile i'm going to try https://gerrit.wikimedia.org/r/63566 [21:07:27] Nikerabbit: fixed for Translate :-] https://integration.wikimedia.org/ci/job/mwext-Translate-phpcs-HEAD/1286/console [21:08:14] it's unfortunately going to include 'eventlogging::supervisor' which hard-codes a client to write to db1047.. and yes, I know, there will be a price to pay in the afterlife for such things [21:09:18] hoo: are you going to stay an extra day before and/or after? [21:09:24] New patchset: Anomie; "Add JavaScript login check against the central wiki" [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/58924 [21:09:51] New review: Anomie; "PS15: Work around jshint stupidity." [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/58924 [21:10:12] sumanah: From Thursday the 23rd until Sunday (going to leave on Sunday to be back normal on Monday) [21:10:46] New patchset: Anomie; "Add JavaScript login check against the central wiki" [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/58924 [21:11:25] New patchset: Anomie; "Replace image-based AutoLogin with iframes" [mediawiki/extensions/CentralAuth] (master) - https://gerrit.wikimedia.org/r/62194 [21:11:56] hoo: well I bet there will be people who want to see stuff together in Amsterdam on Thursday :-) [21:12:33] New review: Hashar; "deployed" [integration/jenkins-job-builder-config] (master) C: 2; - https://gerrit.wikimedia.org/r/63571 [21:12:53] I guess so... sadly all the people I know and yet have asked are going to arrive later [21:12:56] New review: EBernhardson; "We could probably un-block this now?" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/61163 [21:13:03] (WMDE staff) [21:13:53] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/63571 [21:17:09] hoo: well, I'm fairly sure some people will arrive on Thursday and be hanging out in the hotel lobby - I don't know who, but statistically it's probable [21:18:12] heh, I guess I'm just going to lurk for Wikimedians in the lobby then :D [21:18:53] That's what I do! [21:19:19] hoo: I'll be around on Thurs. I'll find you in the lobby at some point :) [21:19:25] \o/ [21:21:22] csteipp: In case you want to do a fast and easy merge: https://gerrit.wikimedia.org/r/58042 [21:25:24] greg-g: confirmed with mwlaker and fundraising that they are no longer using test NFS style [21:27:48] hoo: /me is sad that he doesn't get to go to the hackathon [21:28:47] pgehres: Oh :/ What about Wikimania, will you be there? [21:29:08] pgehres: awesome, thanks [21:29:50] hoo: sadly, wasn't approved for that either. and, I am leaving the WMF at the end of this week. [21:30:10] hopefully though I will still be able to contibute as a volunteer and come to wikimania in London [21:30:26] I hope so too [21:30:52] * pgehres has never been to a hackathon (except for the SF one, but that was almost entirely staff anyway) [21:30:54] pgehres: if you find yourself in India or Israel there are often events in those places as well, in case you wanna go [21:30:58] pgehres: Oh... would be great to see you around from time to time... admin tool related extension always need love (reviews and stuff) [21:31:06] pgehres: well, no, it wasn't almost entirely staff, do you want the numbers? [21:31:17] nah, I was exaggerating [21:32:14] but I knew very few of the staff then anyway, so I spent a lot of time getting to know them [21:32:24] New review: Kaldari; "I'm taking off the -2, but I don't really understand why we need an extra UI element for this. If su..." [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/37229 [21:35:50] New patchset: Mattflaschen; "Fix Qqq link for userlogin-resetpassword-link" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63577 [21:38:54] csteipp: do you happen to know if anyone besides fundraising uses Twig? [21:39:11] I'm trying to figure out if I can update the vendors repo to Twig 1.13 (from 1.7) [21:39:44] New review: Mwalker; "yes -- in progress :)" [wikimedia/fundraising/civicrm] (master) - https://gerrit.wikimedia.org/r/53653 [21:41:11] mwalker: Not that I know of [21:44:04] Reedy: do you know where the sidebar donate link comes from in new WMF wikis? [21:44:12] Err [21:44:14] Link? [21:44:33] Oh, sitesupport-url|sitesupport [21:44:47] I guess the default message for that … hmm [21:47:06] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63577 [21:47:20] New review: Anomie; "(1 comment)" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/30775 [21:47:35] csteipp: yay! do you know if we have a new location (in git) for svn.wikimedia.org/svnroot/wikimedia/vendors ? [21:48:20] mwalker: ask k4, that was pretty much a fundraising repo [21:48:22] https://gerrit.wikimedia.org/r/#/admin/projects/ [21:48:50] Reedy: where is the new wiki creation script? [21:49:05] Why? What wiki are you creating? :p [21:49:32] just curious, was trying to see if I could add a small piece to set the sitesupporturl [21:49:47] Ah [21:49:50] since it should have a {{subst:}} in it [21:49:55] extensions/WikimediaMaintenance.git [21:49:59] addWiki.php [21:50:08] :-) ty [21:53:06] New review: Nemo bis; "Thank you for this. I've not yet checked if this is all of them and I've no particular opinion on pr..." [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/63465 [21:54:56] Any idea what could be causing a wiki with file cache enabled to be very slow still? [21:55:14] The hits are cache hits, (the html contains a where the date is many days ago [21:55:27] the requests are giving max=age=0, expires 1970 [21:55:33] http://docs.jquery.com/Main_Page [21:55:47] request takes like 10 seconds [21:56:29] https://github.com/Krinkle/docs.jquery.com/blob/master/www/InitialiseSettings.php#L169-L195 [21:58:05] This started happening about 2 hours ago with: [21:58:06] "JobQueueDB::doAck". Database returned error "1205: Lock wait timeout exceeded; try restarting transaction (localhost)". [21:58:15] New review: Emufarmers; "If questions and answers aren't stored in the database, it won't be possible to collect statistics o..." [mediawiki/extensions/ConfirmEdit] (master) - https://gerrit.wikimedia.org/r/63342 [21:58:29] I upgraded MediaWiki to latest master and ran update.php but no improvement, though the error is gone, it is still very slow. No idea what could be causing this. [22:04:12] mwalker: I'm the wrong person for that sorry. maybe ^demon? [22:04:36] It takes 6 seconds to serve a 304 [22:05:06] <^demon> mwalker: I just finished importing the history of that svn repo to git like you asked on wiki :) [22:05:31] ^demon: thanks :) [22:05:43] <^demon> yw. [22:06:17] What's your job run rate? Is it done as part of other requests? [22:07:09] New patchset: EBernhardson; "Remove dead code" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63583 [22:07:13] Krinkle: Job run rate? [22:08:34] Reedy: Here's more details http://cl.ly/image/3L3u34171j27 [22:08:51] Reedy: Job run rate is the default I think. let me check [22:09:36] Reedy: Indeed, wgJobRunRate = default (1) [22:10:06] The cache-control private and max-age of 0 don't make sense [22:10:17] Especially since it does become 304 apparently [22:10:32] and then it says served in 0.210 secs [22:10:47] New review: MarkAHershberger; "Statistics! nice. I think I can get your version tested if you could share the source." [mediawiki/extensions/ConfirmEdit] (master) - https://gerrit.wikimedia.org/r/63342 [22:12:13] Reedy: Both the live and the stage wiki are slow, so it can't be job rate afaik. Since the stage wiki is in readonly mode (and has been for months) [22:12:36] http://stage.docs.jquery.com/ [22:13:03] Change merged: jenkins-bot; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63486 [22:14:38] New patchset: Jdlrobson; "Move addDefaultModules from OutputPage to the skin" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63584 [22:15:27] New patchset: Jdlrobson; "Move addDefaultModules from OutputPage to the skin" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63584 [22:21:58] New review: Bsitu; "(1 comment)" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/62624 [22:24:34] New patchset: Bsitu; "(bug 48302) Trailing dot in link is ignored in plain text email" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63007 [22:24:59] New patchset: Mwalker; "Making a symlink to 1.7.0" [wikimedia/fundraising/twig] (master) - https://gerrit.wikimedia.org/r/63586 [22:25:47] New patchset: Mwalker; "Adding Twig 1.13.0" [wikimedia/fundraising/twig] (master) - https://gerrit.wikimedia.org/r/63587 [22:30:12] what's the best way to amend a commit that's some way back in history? [22:30:17] New review: Krinkle; "Not all of these are skin related. Allowing skins to override these will cause situations where exce..." [mediawiki/core] (master) C: -1; - https://gerrit.wikimedia.org/r/63584 [22:30:25] TimStarling: interactive rebase [22:30:41] At least, that's what I usually do. [22:31:04] git rebase -i; then pick->edit the one you want to change. [22:32:21] interesting, thanks [22:32:25] essentially it does this (which would be the manual alternative); git reset --hard ; git cherry-pick ; git commit --amend; git cherry-pick [22:32:44] New review: Kaldari; "(1 comment)" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/62210 [22:33:47] yeah, I've done that sequence before I think [22:34:12] The alternative, which works nicely if you've already got the changes in your working copy, is to commit - reorder - squash [22:39:09] New patchset: Bsitu; "Make Echo talk page notification rely on enotifusertalkpages" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63070 [22:39:18] New review: Jdlrobson; "It seems perfectly reasonable that I should be able to write a skin which adds no javascript to the ..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63584 [22:45:28] RoanKattouw: define reorder? [22:46:09] Krinkle: Make changes ; commit them ; interactive rebase ; move change commit up to be just below the target commit ; pick->fixup [22:46:12] Moving the change is the "reorder" bit [22:46:33] rebase --interactive allows you to reorder the lines and that will move the commits around [22:46:34] Righ [22:47:08] Thats for when you want to amend the change itself [22:47:20] If you just need to amend the commit message you could do it right away in git rebase [22:47:59] RoanKattouw: btw, thanks for that one. I didn't think of using git rebase to "add" a fixup commit in the history [22:48:21] I still used the manual reset/amend/cherry-pick path for that scenario [22:48:36] Yeah for the commit msg you just use reword [22:48:37] Right [22:49:07] Yeah I usually use the workflow you described (rebase interactive, edit) when I'm thinking ahead, but when I've already made the changes locally I just commit them then use an interactive rebase to clean it up [22:49:15] indeed, reword, not edit. [22:49:29] RoanKattouw: Oh, interesting. [22:49:36] RoanKattouw: Actually, In that case I use git stash [22:49:45] and then pop it when at the past commit [22:49:48] New review: Bsitu; "(1 comment)" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/62210 [22:49:51] Yeah I sometimes do that as well [22:49:52] saves me one step of merge conflicts [22:50:18] Right [22:50:21] hm.. not merge conflicts, but less staging/committing [22:50:28] yeah [22:50:39] I've showed you git add -p, right? [22:50:47] Yeah, I use that all the time [22:51:12] I used to use git diff and git add, but now I use git add -p exclusively before any commit. [22:51:16] New review: Kaldari; "I support the idea of being able to override the default modules from the skin. This will give us mu..." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63584 [22:51:39] https://help.github.com/articles/interactive-rebase#example [22:52:56] RoanKattouw: TimStarling: btw, any idea what could be causing a wiki with file cache enabled and CACHE_ANYTHING/CACHE_ACCEL enabled to still take 10 seconds for a 304 response, logged-out? [22:53:08] more details http://cl.ly/image/3L3u34171j27 [22:53:29] Oooooh I didn't know about --autosquash, that's hot [22:53:46] Krinkle: do you have profiling enabled? [22:53:57] AaronSchulz: No [22:54:07] AaronSchulz: You mean as a cause or to find the cause? [22:54:20] the later [22:54:31] even $wgProfiler['class'] = 'ProfilerSimpleText'; in StartProfiler.php might be enough [22:54:37] gotcha, will enable in staging. [22:54:50] what wiki is this? [22:55:10] no idea [22:55:26] profiling would work to identify this, or strace [22:55:40] AaronSchulz: http://stage.docs.jquery.com/Special:BlankPage [22:55:46] the non-stage is also affected [22:55:56] stage is a readonly copy [22:57:23] New review: Bsitu; "We also need some comments for new object and functions, it would help the reviewers get a brief ide..." [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/62624 [22:57:55] scap is moving to tin now [22:59:22] Change merged: jenkins-bot; [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/61565 [22:59:27] TimStarling: so we are not waiting a week? [22:59:47] no [23:00:02] you mean a week, oh no, that's amsterdam, make it a month? [23:00:02] meh, fine by me :) [23:00:17] it was scheduled for yesterday but I forgot [23:01:07] Oh, hah [23:01:10] meeting notes from an hour ago say "Tim will complete the migration today." [23:01:31] I guess this means my lightning deploy window at 4pm is gonna be postponed, huh? :) [23:01:39] but my dotfiles on fenari!!! [23:01:45] i have colorized enhancements of sync-* [23:02:12] New review: Alex Monk; "Needs rebase." [mediawiki/extensions/ArticleFeedbackv5] (master) C: -1; - https://gerrit.wikimedia.org/r/59387 [23:02:36] Enabling profiling http://stage.docs.jquery.com/mw/profileinfo.php [23:02:41] greg-g: Who all is signed up for LD today? Tim is doing the fenari->tin migration so I'm not so sure we can do LD today [23:03:04] Looks like it stays empty [23:03:21] http://dpaste.de/SGhVK/raw/ [23:03:22] * AaronSchulz was thinking of using ProfilerSimpleText [23:03:52] $wgProfiler['class'] = 'Profiler'; [23:03:52] $wgProfileToDatabase = true; [23:03:52] $wgEnableProfileInfo = true; [23:03:53] RoanKattouw: you can still do your lightning deployments [23:03:55] I've set those [23:04:02] you just will need to manually copy the files with a carrier pigeon or something [23:04:06] haha [23:04:56] or maybe the patchset will work and you'll be able to deploy out of tin in 5 minutes [23:05:22] TimStarling: So is sync-file/sync-dir functional on either fenari or tin right now? Or should I wait? [23:05:23] Yippie, build fixed! [23:05:23] Project browsertests-en.wikipedia.org-linux-chrome build #57: FIXED in 3 min 57 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-linux-chrome/57/ [23:05:40] Fingers crossed :) [23:06:01] AaronSchulz: Also enabled wgDebugComments [23:06:19] I have to leave in about half an hour so if it's not up by then I can do it tomorrow, not a big deal [23:06:21] krow lliw ti ebyam [23:07:10] RoanKattouw: I don't think it would be functional [23:07:34] puppet will be progressively changing the scap-1 scripts on the apaches to use tin as an rsyncd server [23:07:59] so either you have to override that rsyncd choice (which is quite possible) or you have to wait until rsyncd on tin is truly functional [23:08:12] Krinkle: where did you set $wgProfiler['class'] = 'Profiler'; ? [23:08:39] New patchset: Demon; "Clean up sublime ignores" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/63591 [23:08:43] TimStarling: Ooooh I see. So it's gonna take a couple hours before the change has fully propagated? In that case I'll just do it tomorrow [23:08:47] if you just want sync-file, note that it is like one line of code [23:09:02] you can just write a slightly different one line of code directly into your terminal [23:09:11] Wait, scap-1 [23:09:15] AaronSchulz: LocalSettings.php [23:09:16] Yeah I'm not doing anything related to scap, just sync-file/sync-dir [23:09:23] AaronSchulz: I guess that won't work [23:09:36] I don't think sync-dir uses the two-stage sync process, does it? [23:09:56] RoanKattouw: you mean 1) sync, 2) pray? [23:10:17] hahaha [23:10:18] Krinkle: did you create StartProfiler.php in the same directory [23:10:25] sync-dir is just: [23:10:31] it should be a PHP file with the $wgProfilter['class'] config [23:10:46] sudo -u mwdeploy rsync -a --delete --exclude=.svn --exclude=.git --exclude=cache/l10n --no-perms $MW_RSYNC_HOST::common/$DIR/ $DESTDIR [23:11:08] so you can run that on fenari as long as $MW_RSYNC_HOST is still correct there [23:11:10] AaronSchulz: No, I figured it would work without a StartProfiler.php file when in LocalSettings. Why does it have to be in a separate file? - ah because LocalSettings.php is included later [23:11:16] or if not, you can run: [23:11:23] Right [23:11:23] Where "correct" means? [23:11:30] sudo -u mwdeploy rsync -a --delete --exclude=.svn --exclude=.git --exclude=cache/l10n --no-perms 10.0.5.8::common/$DIR/ $DESTDIR [23:11:33] 10.0.5.8? [23:11:35] yes [23:11:35] Ah yes [23:12:01] OK, well let's give that a shot then [23:12:06] Yippie, build fixed! [23:12:07] Project browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_9 build #46: FIXED in 10 min: https://wmf.ci.cloudbees.com/job/browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_9/46/ [23:12:20] New patchset: Catrope; "Update VE to fix category bug" [mediawiki/core] (wmf/1.22wmf4) - https://gerrit.wikimedia.org/r/63592 [23:14:25] AaronSchulz: Done, still nothing being logged it appears. [23:14:47] Change merged: Catrope; [mediawiki/core] (wmf/1.22wmf4) - https://gerrit.wikimedia.org/r/63592 [23:15:38] Switching to ProfilerSimpleText [23:17:04] Krinkle: you said it is read only? [23:17:06] TimStarling: Hmm MW_RSYNC_HOST doesn't seem to be defined anywhere on fenari, it's not in /usr/lib/mw-deployment-vars.sh [23:17:10] then the DB one won't work [23:17:12] AaronSchulz: Yes [23:17:25] AaronSchulz: wgReadonly, not mysql read only, though [23:17:33] Change merged: jenkins-bot; [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/61559 [23:18:03] http://stage.docs.jquery.com/Special:BlankPage now shows it on the bottom (view-source for better version) [23:19:04] TimStarling: Alright, I did export MW_RSYNC_HOST=10.0.5.8 then ran sync-dir, and that seems to have worked [23:19:07] Krinkle: doesn't matter [23:19:16] that makes MW not write to the profiling table [23:19:27] * RoanKattouw is done with LD [23:19:37] AaronSchulz: I figured, which is why I switched to SimpleText instead [23:22:04] AaronSchulz: 100.00% 0.346467 [23:22:11] so that's not where it is spending the 10 seconds. [23:22:47] use strace [23:23:07] Change merged: jenkins-bot; [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/62210 [23:23:26] Krinkle: yeah, I noticed the "served by" time was also low [23:23:40] try: ps -C apache2 -o pid,wchan [23:23:44] while the request is running [23:23:56] maybe in a shell script loop that runs it every second [23:24:04] Krinkle: that means it is either in restInPeace() or the network [23:24:07] while true; do date; ps -C apache2 -o pid,wchan ; done [23:24:37] that will confirm that apache is running while you are waiting, and also tell you what wchan it is waiting in [23:24:41] TimStarling: Most are at flock_lock_file_wait, some are at "?" or "stext" [23:24:56] excellent [23:25:09] total of like 20-30 instances of apache2 -k start [23:25:17] Change merged: jenkins-bot; [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63070 [23:25:34] flock_lock_file_wait is unusual, it is probably the problem [23:25:41] so pick one of the PIDs in that state, and do: [23:25:50] lsof -p $pid ; strace -p $pid [23:26:04] TimStarling: Thanks, I hadn't used strace before to debug apache. Never had to, either. Usually PHP side is to blame. [23:26:10] strace should tell you what filehandle it is waiting on [23:26:14] OK [23:26:22] then lsof should let you correlate that with a path [23:26:31] checking what those commands are first. just a moment. [23:26:54] Change merged: jenkins-bot; [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63007 [23:28:24] oh, I missed a sleep in my shell script loop earlier [23:28:59] if you ran it and your terminal was flooded with 10 Mbps of messages, you probably realised that already [23:29:02] There were plenty of pending requests, so I didn't use the loop one [23:29:18] New review: Bsitu; "(1 comment)" [mediawiki/extensions/Echo] (master) - https://gerrit.wikimedia.org/r/63076 [23:30:57] New patchset: Pgehres; "Adding function to set the MW:Sitesupport-url" [mediawiki/extensions/WikimediaMaintenance] (master) - https://gerrit.wikimedia.org/r/63594 [23:32:08] TimStarling: Yeah [23:32:10] :) [23:32:15] TimStarling: Snippet of output: https://gist.github.com/Krinkle/5572420 [23:32:50] I've cut out a few sections, but this is most of it [23:32:53] strange that it doesn't show up in profiling [23:33:05] mmm, I guess not so strange [23:33:22] it probably is in the shutdown function, whereas profiling stops before shutdown [23:33:45] anyway, yes, PHP sucks and its session functions are unusable by default [23:34:09] you can use memcached for sessions [23:35:12] Right [23:35:27] TimStarling: It is being slow even for anon requests though [23:35:32] which is what is so puzzling about it [23:35:41] those requests are served from static file cache [23:35:47] well, maybe anon requests are starting sessions [23:36:01] have it write a backtrace to a debug log when a session is started [23:36:03] that would hurt [23:36:34] Hm.. could be. I wonder how that could change overnight, but lets see [23:37:32] Krinkle: the file cache code still goes through normal setup/restInPeace()/shutdown [23:37:40] Sure [23:37:40] New patchset: Pgehres; "Updating default sitesupport-url" [mediawiki/extensions/WikimediaMessages] (master) - https://gerrit.wikimedia.org/r/63595 [23:37:41] so any session handler can still slow it down [23:38:20] sessions are closed in a register_shutdown_function() callback [23:38:37] * AaronSchulz recalls having some index.php code to bypass that or something ages ago, meh [23:39:09] TimStarling: AaronSchulz: I checked the session directory this instances php uses, it only has about half a dozen files in it [23:39:27] which roughly matches the number of users I know have write access to the wiki and are using it right now. [23:39:49] so maybe you have two problems [23:40:03] maybe sessions are making it slow for logged-in users, causing the requests you normally see active [23:40:14] and maybe additionally it is slow for anons for some other reason [23:40:16] New review: Alex Monk; "The scroll bar this produces on Chrome is weird and doesn't work properly." [mediawiki/extensions/ArticleFeedbackv5] (master) - https://gerrit.wikimedia.org/r/61568 [23:40:30] Apache is jumping to 24% for 3 of the 30 processes right after "sudo /etc/init.d/apache2 restart" [23:41:03] you're assuming there is 10s of CPU usage for a file cache hit? [23:41:09] that would be a lot of CPU usage [23:41:29] As far as I know, at the moment there are near 0 http requests. [23:41:29] I assumed it would just be blocking on something for 10s, hence strace as preferred tool [23:43:01] I think you should strace a request that you generate, rather than a random apache child [23:43:28] this is on labs? [23:43:42] is it a public IP or a proxy? [23:43:53] can you reproduce with curl or does it need a browser? [23:44:00] New review: Alex Monk; "In Chrome part of the scroll is not properly selectable, and in Firefox it overlaps with the helpful..." [mediawiki/extensions/ArticleFeedbackv5] (master) C: -1; - https://gerrit.wikimedia.org/r/61568 [23:44:03] No, on an old machine at jquery which is being migrated. John Resig set it up in 2006, last month I updated it to MediaWiki 1.22 and all was well. [23:44:05] it's slow with curl [23:44:24] AaronSchulz: curl is slow indeed, I noticed that doing the same curl from on the server itself is not slow at all [23:44:37] New review: Kaldari; "Yeah, looks like that was left over from before we were using payload. I support removing the unused..." [mediawiki/extensions/Echo] (master) C: -1; - https://gerrit.wikimedia.org/r/63583 [23:44:57] try: lsof -i4 -n | grep [23:45:18] I guess... [23:45:25] lsof -i4 -n | grep .*:80 [23:45:30] to filter out your SSH connection [23:45:52] with the right criteria, you can pull out the PID with awk, like [23:46:12] lsof -i4 -n | grep whatever | awk '{print $2}' [23:46:53] pid=`lsof -i4 -n | grep whatever | awk '{print $2}'` ; lsof -p $pid; strace -ttt -p $pid [23:47:27] and since I know you like to know how everything works... [23:47:33] -i4 = IPv4 sockets [23:47:41] -n = numerical IP address rather than reverse DNS [23:47:57] strace -ttt = timestamp each log line so that it makes sense when you pastebin it [23:48:08] so that waits are obvious [23:48:14] lsof is more awesome than I thought [23:48:15] very nice [23:48:34] :D [23:50:37] TimStarling: It seems it won't match anything. My IP is showing up for the ssh one, but nothing else. I've tried a few times with multiple tabs and this in a loop. [23:51:03] They are all taking 5+ seconds [23:51:27] ok, so maybe the apache request is very fast, too fast to capture this way [23:52:10] TimStarling: The apache process dies before the request closes? [23:52:14] try: tcpdump src host and dst port 80 -s1500 -A [23:52:25] Does it grant php authority to send directly to the client? /me doesn't know [23:52:49] New patchset: Ori.livneh; "Remove unused "force" parameter" [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63597 [23:53:10] no it doesn't [23:53:11] Change merged: jenkins-bot; [mediawiki/vagrant] (master) - https://gerrit.wikimedia.org/r/63597 [23:53:13] k [23:53:28] TimStarling: tcpdump isn't installed [23:54:10] so install it [23:54:16] hehe [23:54:16] ;) [23:54:34] so sayeth Tim [23:54:39] This is an end-of-life machine without any puppetisation (the other servers are puppetized) [23:54:46] I guess one more unpuppetized package won't hurt [23:55:13] New patchset: Adamw; "WIP: banner DonationForm backend support" [mediawiki/extensions/DonationInterface] (master) - https://gerrit.wikimedia.org/r/61929 [23:55:36] TimStarling: Ah, it was installed. PATH is just messed up (not containing sbin) [23:56:05] tcpdump lets you confirm that the problem is indeed with apache and not some other place in the network stack [23:56:20] maybe you want a bidirectional trace... [23:56:38] this should work: tcpdump host and port 80 -s1500 -A [23:58:02] TimStarling: I can't find any command 'dst'. Typo? [23:58:35] "and" was an actual tcpdump parameter, not separating two commands [23:58:44] Oh, I see :) [23:59:38] but like I say, maybe you don't want to use dst/src anyway, since that only gives you a flow in one direction