[00:00:04] "All these timing scripts rely on microtime which relies on gettimebyday(2). This can be inaccurate on servers that run ntp to syncronise the servers time. For timing, you should really use clock_gettime(2) with the CLOCK_MONOTONIC flag set." [00:00:11] time() can be set back in windows though [00:00:33] he means gettimeofday not gettimebyday [00:01:05] anyway, what will happen if the clock is set back? the timeout will be extended by the amount it is set back by? [00:01:18] it doesn't sound like a disaster to me [00:02:03] I suppose if the timeout was being routinely hit, and then the clock was set back by 5 minutes or something, then you could get an overload [00:02:32] but timeouts probably shouldn't be routinely hit [00:03:05] and it's really unlikely that the clock will ever be set back by more than a few milliseconds on a WMF apache since we monitor them for working ntpd [00:03:21] you have to learn to be more pessimistic :) [00:03:39] anyway, it's easy enough to check for well enough [00:04:32] suit yourself [00:04:47] * AaronSchulz is amending [00:04:49] maybe we should have a microtime() wrapper that's guaranteed to be monotonic [00:04:58] stores the last value in a static variable [00:06:33] you calling usleep() until catches up (dying if it takes to long)? heh [00:08:12] *you mean [00:08:35] my hands are like a CD writer that randomly skips some byes not and then... [00:08:52] *bytes, *ahem* [00:09:11] $end = wfMonotonicTime() + $timeout; [00:09:30] while ( wfMonotonicTime() < $end ) { ... } [00:09:49] or some variation of that [00:10:18] then we could write a C implementation of it that calls clock_gettime() [00:13:39] TimStarling: have you thought about queue migration a bit more? [00:15:51] progressively migrate wikis to redis, have job runners doing both mysql and redis until all mysql queues are zero length, seems reasonable to me [00:16:12] having a lot of mysql job runners doing nothing won't cost us much, with aggregate caching [00:17:01] you mean dedicated mysql job runners? [00:17:12] I mean dual purpose with redis [00:17:14] or just each runner using both? [00:17:17] oh ok [00:17:41] if memory usage is excessive, we can reduce the number of processes per server slightly during the migration [00:18:21] I think redis will have like 54gb of the 72gb for use [00:18:31] I mean PHP memory usage [00:18:56] the number of job runners is constrained partly by PHP memory usage, right? [00:19:21] probably not constrained very much, we can probably just double the number of job runners [00:19:40] the pop rate will be higher during migration due to there being more processes, but once migration is complete, it will settle down pretty quickly [00:20:52] I don't think it is constrained that much, no :) [00:22:48] TimStarling: have you had a chance to see https://gerrit.wikimedia.org/r/#/c/53718/6 ? [00:24:04] yeah, quite a big diff [00:24:26] the class is actually smaller I think (in spite of adding a bunch of comments) [00:25:43] it will be fairly complex to review, I will have to spend a bit of time with the redis manual [00:35:08] anomie, still around? [00:35:30] MartijnH- Yes [00:35:36] TimStarling: I'm sure you'll enjoy more Lua [00:36:23] anomie: cool. (Un?)fortunately, my testcase proved rather simple, though I managed to mess up pretty well. Could you give some feedback on http://en.wikipedia.org/wiki/Module:Revision ? [00:36:40] MartijnH- Sure, just a minute [00:37:21] there is also the slightly more complicated http://en.wikipedia.org/wiki/Module:AfC which has way more room for me to mess up [00:39:16] TimStarling: I remember the whole "are these two edits the same" difficulty and suggested the self::identicalPageObjects() style check [00:39:38] interesting that the summary was also thrown in for that reason [00:39:52] * AaronSchulz feels like inline docs would help [00:41:46] I suppose one could made a UUID in EditPage, pass it to the hook, and pass that to doEditContent() and to the hook [00:41:54] *could make [00:41:58] but meh... [00:42:57] MartijnH- Module:Revision looks decent, although it's so simple I wouldn't be too surprised if the parser function version is as fast. [00:45:25] anomie, that one is mainly because it's called a lot in other templates, and I find frame:expandTemplate{ title = 'diff', args = { 'even', 'four', 'is', 'much' } } to be an annoyance [00:45:55] MartijnH- In that case, make it take the args directly instead of going through the frame object. [00:47:06] well, it should at least be callable with frame IMO. the lua template is trivial, and - though I admit being bad at reading mediawiki templates - is practically unreadable to me. [00:48:02] though on the other hand, the invocation with frame.newChild() is just as cumbersome as the expandTemplate call [00:48:09] that is what would be needed right? [00:49:23] MartijnH- frame.newChild() has a low limit, to prevent people from abusing it too much. The usual thing to do is to write the version that takes args directly, and then write a wrapper that extracts args from the frame object and calls the first function. [00:49:30] maybe a convenience alias would make sense [00:49:42] heh [00:49:50] yeah, that :) [00:50:17] are there any naming convensions for that in place yet? [00:50:51] MartijnH- Looking at Module:AfC: if the functions are not going to be called from outside the module, there's little point in putting them in the table for export. Your p.shorttitle might cut things off in the middle of a multi-byte Unicode character. Use mw.ustring instead to avoid that. Also, did you know that when the module is invoked from inside a template (which is the normal use case), frame:getParent().args has the arguments passed to the [00:50:51] *template* even if they weren't specifically passed into the #invoke? [00:50:54] I'd imagine nobody wants to invoke {{#invoke|foowrapper|args}} [00:51:30] The #invoke is usually inside a template. Like [[Template:Lua hello world]] calls [[Module:Bananas]]. [00:52:41] ah, that makes sense [00:52:44] sort of [00:53:00] in a backwards compatible sort of way [00:55:08] so if you {{#invoke:Foo|bar}} from template:baz, frame.getParent().args holds the parameters passed to baz? [00:57:39] that's both convenient and horrible [00:57:53] MartijnH- Exactly. If you look at the work going on at [[Module:Citation/CS1]], you can see they use this to great advantage: frame:getParent().args has all the arguments passed to the citation template, while frame.args has a bunch of configuration info to change behavior between say {{cite web}} and {{cite news}} and so on [00:58:01] Why horrible? [00:58:24] because that would mean the callee must have knowledge of the arguments passed to the caller [00:58:53] so it only works if its called from the way you expect it to be called [00:59:48] Most should be, anwyay. It's much like how you didn't often call Template:str len/core directly, before that was Luaized. [01:00:46] well, I can certainly see the convenience [01:01:19] Another template that uses frame:getParent() to good effect, IMO, is [[Template:On this day]]. See how it's used on a page like [[Talk:Teachers' Day]]. [01:06:17] I probably need some more experience with this to have a *valid* opinion, rather than an opinion based on my gut [01:09:11] Well, I'm out for the night (in my timezone). [01:09:50] ... [01:09:56] well, thanks for the help :D [09:37:23] New patchset: Hashar; "(bug 45084) update beta database automatically" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/54635 [09:37:44] New review: Hashar; "deployed" [integration/jenkins-job-builder-config] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54635 [09:37:44] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/54635 [12:37:03] Susan: https://www.wikimedia.org/w/extract2.php?template=Secure.wikimedia.org_template&title=Secure.wikimedia.org_portal& [12:37:16] Wee [12:37:22] _portal still works when triggered manually [12:37:31] $1 replacement in _template with rendered wikitext of _portal. [12:37:41] See also https://meta.wikimedia.org/wiki/Talk:Project_portals/links [12:37:58] (don't worry, they don't take arbitrary page names of meta) [13:15:41] YuviPanda: I need to know Rahul_21's email address in Bugzilla to find out if Rahul is able to change bug statuses :) [13:23:06] New patchset: Zfilipin; "Updated multi_json Ruby gem" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54648 [13:37:42] New review: Cmcmahon; "maintenance" [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54204 [13:37:43] Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54204 [13:38:27] New review: Cmcmahon; "maintenance" [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54469 [13:38:28] Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54469 [13:39:34] New review: Cmcmahon; "Mobile docs to proper location" [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54518 [13:39:34] Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54518 [13:40:23] New review: Cmcmahon; "maintenance" [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54648 [13:40:23] Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54648 [14:13:18] zeljkof: hey there :-] [14:13:33] hashar: what's up? :) [14:13:34] zeljkof: was wondering which of the beta wiki you are using to run the browser tests :-] [14:13:49] zeljkof: I got a jenkins job to update the beta enwiki database once per hour [14:13:58] hashar: let me check [14:14:32] the db updating job is at http://integration.wikimedia.org/ci/job/beta-updatedb-enwiki/ =] [14:17:57] hashar: so by default tests run at test2 [14:18:19] hashar: but some tests run at en.wikipedia.beta.wmflabs.org [14:19:06] aftv5 tests run on beta right now fwiw [14:19:12] good morning :) [14:19:26] chrismcmahon: good morning :) [14:19:28] zeljkof: nice. So en.wikipedia.beta.wmflabs.org has its database updated automatically once per jour [14:19:30] hour [14:19:44] hashar: thanks for letting me know [14:19:48] thanks hashar! [14:19:51] nice! [14:20:01] hashar: the updates should not break anything, right? :) [14:20:19] that should make it possible to port the page_triage test there immediately. [14:20:52] chrismcmahon: I am working on https://bugzilla.wikimedia.org/show_bug.cgi?id=46223 at the moment, I have installed android setup on my machine [14:53:05] zeljkof: database updates might break something : -] [14:53:15] zeljkof: but I believe it is way better than NOT having any DB update [14:55:00] chrismcmahon: while you are around, I am wondering if we really need all the wiki we have in beta. There is ~500 of them right now [14:55:15] chrismcmahon: I thought we could drop some of them since they are mostly empty and probably not used at all [14:55:56] hashar: I think we can drop some, especially if we can add them later if we need them. [14:57:32] chrismcmahon: that is what I though. Could you cast your voice on https://bugzilla.wikimedia.org/show_bug.cgi?id=46104 ? [14:57:34] and cc [14:57:56] I will follow up with a list of wiki to eliminate =) [15:01:34] Welcome everybody to this BUG DAY about open LiquidThreads bug reports which will take place in this channel for the next couple of hours. [15:01:38] Please see https://www.mediawiki.org/wiki/Bug_management/Triage/20130318 for more information. [15:01:41] As usual, the Etherpad will links to Bugzilla is located at http://etherpad.wmflabs.org/pad/p/BugTriage [15:03:32] siebrand died out of happiness [15:03:57] hehe [15:07:13] chrismcmahon, do you have an available Firefox 22 somewhere? [15:10:24] so for this LQT bugday, the list of bug reports is linked from http://etherpad.wmflabs.org/pad/p/BugTriage [15:11:05] New patchset: Hashar; "triggers for mw/ext/DoubleWiki" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54656 [15:11:20] ...and there are about 220 open reports to choose from :) [15:11:24] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54656 [15:12:09] hmm, does anybody still run IE6 on an ancient machine, to test bug 24772 ? :) [15:12:52] andre__: Hah, I'll check that out. [15:13:24] IE6 = Windows XP? I wonder which Windows versions ditched which IE versions... [15:13:47] andre__, yes, WinXP came with ie 6 [15:14:28] valeriej: Somebody (actually no idea who) added "Page to test LQT on test2: https://test2.wikipedia.org/wiki/Talk:Xpd " to the Etherpad :) [15:14:31] hi Platonides! [15:14:36] in for some triaging too? :) [15:14:36] * Platonides sees http://test.wikipedia.org has Liquid Threads installed, good [15:15:04] andre__: Nevermind. You can test pages using different browser modes, but it only goes to IE7. [15:15:11] andre__, I have to leave soon [15:15:12] andre__: I added it. :P [15:15:13] ah! [15:15:21] Platonides, heh, pity :) [15:15:21] but if I can help a bit in the meantime... :) [15:15:48] Pick your favorite LQT bug and update it :P [15:15:52] As written, the list of bug reports is linked from http://etherpad.wmflabs.org/pad/p/BugTriage [15:16:11] hey, i am new to mediawiki [15:16:17] hi dell [15:16:31] i would like to get involved with mediawiki dev [15:16:53] hello dell, welcome [15:17:41] dell: a good starting point is http://www.mediawiki.org/wiki/How_to_contribute [15:17:49] dell: do you have anything specific in mind? [15:18:24] (also note that there's a bugday going on here right now, testing and updating open bug reports about the LiquidThreads MediaWiki extension) [15:20:41] i was working on etherpad... [15:21:06] * andre__ takes a look at https://bugzilla.wikimedia.org/show_bug.cgi?id=22800 as part of the bugday [15:22:04] dell: not sure if I understood you correctly - do you refer to the Bugday Etherpad, or in general? [15:22:50] etherpad in general [15:23:24] * valeriej is looking at https://bugzilla.wikimedia.org/show_bug.cgi?id=26383 [15:23:29] hashar: got a minute for a small change that would help with debugging? https://gerrit.wikimedia.org/r/#/c/54348/ [15:23:48] dell, ah. It's one tool we use in Wikimedia and MediaWiki development for meetings, yeah. [15:23:58] dell: Does http://www.mediawiki.org/wiki/How_to_contribute help you? [15:24:16] i had already taken a look at it.. [15:24:20] thanks a lot [15:24:32] Krinkle|detached: Sorry, just saw your note here. That wiki page isn't protected any longer. [15:26:23] DanielK_WMDE_: I have seen it, not sure what to say about that patch :-]  I need to actually review it and think about it [15:26:51] So the testcases in https://bugzilla.wikimedia.org/show_bug.cgi?id=22800 aren't working anymore, so I asked for updates... [15:27:14] ...and might close this in a few weeks if there's no answer, because there would be no clear goal in what to fix. [15:28:42] I'm also adding this to http://etherpad.wmflabs.org/pad/p/BugTriage to keep track of the bug reports that we've taken a look at here. [15:29:19] dell: if you have any specific questions in mind, don't hesitate to ask. For example it might be good to know which programming languages you know [15:29:33] plus qgil might be able too to help you find your way into contributing to MediaWiki [15:29:50] dell ^ [15:30:18] New patchset: Hashar; "job mwext-PoolCounter-pep8" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/54663 [15:30:55] dell, are you a student? Just wondering because of our mentorship programs e.g. Google Summer of Code [15:32:07] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/54663 [15:32:23] dell, e.g. https://www.mediawiki.org/wiki/Summer_of_Code_2013 [15:32:55] New patchset: Hashar; "triggers for mw/ext/PoolCounter" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54664 [15:33:12] dell, in general what is useful is to fill your user page at http://mediawiki.org with some details about you, your skills and interests. This way it is easier to help you finding an interesting first task [15:33:17] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54664 [15:34:00] andre, i am quite comfortable with php, c, python and java [15:34:12] yeah i am a student [15:35:10] could you give me some clue how to go abt bug 22800 [15:41:45] dell: I wouldn't recommend bug 22800 when it comes to coding. As I wrote there, it's not clear enough. [15:42:12] dell: We're right now triaging older bug reports here a bit, see the IRC topic of this channel, that's why some LiquidThread bugs get mentioned here. [15:43:31] Hm, so I tried to create a thread on my user talk page on twn to test bug 26383, but it's not saving the thread. [15:44:57] valeriej, user talk page on which wiki? [15:45:34] andre__: translatewiki.net [15:45:40] ah. [15:46:27] New patchset: Hashar; "remove mwext-PoolCounter-pep8" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54665 [15:46:31] ...and I see in the etherpad that qgil is triaging LQT bugs too. Welcome! :) [15:47:13] heya, I'll start with enhancement requests + High (it's becoming a personal tradition) [15:47:26] valeriej: I'd either ask on #mediawiki-i18n (because it's translatewiki.net) if that setting is still correct. Or maybe in this channel, Nikerabbit (who filed that report) knows? [15:47:31] Change merged: Hashar; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/54665 [15:47:32] qgil, thanks man! [15:47:49] andre__: Thanks. [15:47:55] andre__: i18n team is in meeting now [15:48:59] andre__: valeriej - can you suggest a specific bug for me to try to reproduce? [15:49:45] oh, I see [15:50:14] re that bug: we disabled all anon editings a while ago [15:50:23] ah, thanks. valeriej ^ [15:50:49] so I have no idea whether the bug still exists [15:51:05] I had already noted this on the buday page [15:51:09] *bugday [15:51:28] Nikerabbit: Do you need special permissions to start a discussion? I tried to start a discussion on my talk page, but it won't save. [15:51:40] sumanah: Perhaps this one: https://bugzilla.wikimedia.org/show_bug.cgi?id=23382 [15:52:09] New patchset: Hashar; "Revert "job mwext-PoolCounter-pep8"" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/54667 [15:52:10] Nemo_bis: noted what? [15:52:17] Change merged: Hashar; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/54667 [15:54:36] I have to go; another meeting in a few minutes -- but if someone else comes in who is asking for suggestions, they can now look in the etherpad and there are about 4 starter bugs there. [15:54:56] Thanks, sumanah! [15:55:01] Nikerabbit: that anons and non-translators don't have rights for everything [15:56:20] ok [15:56:25] valeriej: try now [15:56:43] Nemo_bis: not that it helps for this particular bug [15:56:48] Nemo_bis, oh cool, thanks for the quick help [15:56:58] We are currently holding a bug day! Anyone interested can help us triage LiquidThreads reports. More info here: https://www.mediawiki.org/wiki/Bug_management/Triage/20130318 and etherpad here: http://etherpad.wmflabs.org/pad/p/BugTriage [15:57:26] unless we are on an another bug already [15:57:28] valeriej: you don't have a babel anywhere? [15:58:12] I assume that's https://www.mediawiki.org/wiki/Extension:Babel in context. [15:59:35] hallo! [15:59:53] Nemo_bis: It worked. Thanks! [16:00:01] lizzard: Hello! [16:00:25] Nemo_bis: No, I don't. [16:01:34] valeriej: https://translatewiki.net/wiki/Special:FirstSteps would be appreciated, it takes a minute [16:01:49] lizzard: Welcome! Here's the bug day etherpad, if you're interested: http://etherpad.wmflabs.org/pad/p/BugTriage [16:01:57] Nemo_bis: Will do! [16:02:20] hashar: fair enough, thanks for looking [16:03:06] I'm trying to reproduce https://bugzilla.wikimedia.org/show_bug.cgi?id=24646 now on test2.wikipedia.org as part of the bugday... [16:05:28] oh, i have to test your bot [16:05:34] Hello Bug Squad! [16:05:53] darn I thought i would trigger something :D [16:05:56] lizzard: "Welcome to the bugday, lizzard! How can I help you?" [16:06:02] LOLLL [16:06:03] * andre__ dresses as a bot [16:07:15] *tries to resist pushing the self-destruct button* [16:07:17] but a bot wouldn't be the worst idea, as long as we're hosting bugdays in a busy development IRC channel. We'll likely try a separate channel (#wikimedia-office) next time though. Still unsure what's better. [16:07:36] hmm, for bug 24646 I cannot reproduce, plus the testcase does not exist anymore. So I'll ask the bug reporter how I could check if this is still a problem and leave it like that. [16:08:45] on your etherpad, you have the search lnk and then suggested bugs to look at. are they suggested specially because they are easier? or becasue they are difficult? [16:09:37] Good question. I delegate that one - Wasn't me. :) [16:09:45] or some other reason like priority? or kind of random? :D [16:09:57] old? [16:10:25] nearly are all old, they are LQT bugs! :P [16:14:13] lizzard: I think sumanah suggested because they are easier. [16:14:31] "sumanah: I have to go; another meeting in a few minutes -- but if someone else comes in who is asking for suggestions, they can now look in the etherpad and there are about 4 starter bugs there." [16:15:16] andre__: not true, LQT has new bugs constantly, even without being modified [16:15:28] Nemo_bis, I know, I wasn't serious. :) [16:15:43] :P [16:15:47] hi bawolff [16:15:59] Hi [16:16:38] Commenting on bug 26383 with IRC comments and adding testme keyword to see if someone can confirm it. [16:20:50] * andre__ tries to reproduce https://bugzilla.wikimedia.org/show_bug.cgi?id=34271 as part of the bugday [16:25:35] The sig length in mediawiki is configurable. Presumably mark meant the default length [16:26:37] * valeriej thinks https://bugzilla.wikimedia.org/show_bug.cgi?id=22678 can be closed [16:27:17] * andre__ cannot reproduce https://bugzilla.wikimedia.org/show_bug.cgi?id=34271 - adding a comment about that [16:27:27] bawolff, oh, I see [16:28:14] Mw default is appearantly 255 (characters? Bytes? The docs don't say) [16:28:55] Which would also be a natural limit if lqt was storing it in the db [16:29:03] I only tried with ascii so no idea either - 255 is default. [16:29:44] If I was on a real computer instead of a cell phone id look up what lqts limit is [16:31:26] valeriej: Re: https://bugzilla.wikimedia.org/show_bug.cgi?id=22678 - interface has only changed a little bit ("Search terms" instead of "Search discussion") on https://test2.wikipedia.org/wiki/Talk:Xpd [16:31:47] search taking way more space than "Start a new discussion". [16:31:53] so might be still valid. [16:32:32] andre__: Ah, I see. [16:32:34] That bug sounds familar ;) [16:32:41] hehe [16:38:50] * andre__ looking at https://bugzilla.wikimedia.org/show_bug.cgi?id=45618 [16:39:30] https://bugzilla.wikimedia.org/show_bug.cgi?id=25456 and https://bugzilla.wikimedia.org/show_bug.cgi?id=25866 seem to be related but I'm not quite sure. Is it better to just post a comment mentioning the other bug or should I use the "See also" field? [16:39:52] ...which has a patch, which depends on another patch, which has a bug number. Adding that bug as dependency. [16:39:55] Somewhere I read that "See also" is for external URLs, that's why I'm asking [16:40:32] Krinkle: is it possible to find out if some feature for example "Event Logging" has tests running under Jenkins at https://integration.wikimedia.org/ci/ ? [16:40:50] chrismcmahon: Use search? [16:40:56] qgil, probably both. People love to not see the "See also" field [16:41:07] andre__, heh - ok [16:41:33] yeah, maybe the See Also field should be somewhere else in the Bugzilla UI. More like "in your face". [16:41:39] Krinkle: OK. Search turns out to be case sensitive, thanks [16:41:48] Ah, I see [16:41:52] Didn't know [16:41:58] I wish see also just took normal bug numbers [16:45:37] Krinkle: another question, is there a way to know if a build is actually empty without going to look, e.g. https://integration.wikimedia.org/ci/job/mwext-GuidedTour-testextensions-master/ [16:47:29] plop [16:47:38] chrismcmahon: What do you mean by "a build that is empty [16:47:40] ? [16:48:17] Krinkle: when you click "Recent Changes" on that link above, you see "no builds" [16:48:23] https://integration.wikimedia.org/ci/job/mwext-GuidedTour-testextensions-master/changes [16:49:27] Krinkle: "Project mwext-GuidedTour-testextensions-master" looks to be just an empty placeholder right now, unless I'm reading that wrong. [16:49:42] It is a job that has an empty build history. [16:49:59] It isn't part of any pipeline or the pipeline it belongs to has not been triggered yet [16:50:14] Krinkle: so it is a build that has no tests either, is that correct? [16:50:20] No that is not correct. [16:50:25] ah, OK [16:50:30] It may have plenty of tests. That's something else. [16:50:34] It hasn't been run yet ever. [16:50:56] You can determine it by looking at the job page. I'm not sure how else you'd determine it. What do you expect? [16:51:54] Krinkle: it seems strange to have a "job with no build history". [16:52:16] chrismcmahon: It means someone set up the configuration and shell commands that run on a certain event [16:52:26] But the event itself has not occurred yet [16:52:29] nothing strange [16:52:51] Most likely it was created very recently and no commit has been pushed yet since it was created. [16:52:57] Meaning it hasn't ever run yet. [16:53:52] Krinkle: is there any way from Jenkins to know what unit tests would be called by that build? [16:55:43] chrismcmahon: The "mwext-{ext-name}-testextensions-{branch}" pattern of jobs just install MediaWiki, and run phpunit.php [16:55:49] Look at the extension repo. [17:01:21] Site note: Parallel to the LQT bugday in this channel, an IRC Office Hour on Wikimedia's issue/bug tracker ( https://bugzilla.wikimedia.org ) and Bug management in general ( https://www.mediawiki.org/wiki/Bug_management ) starts NOW in #wikimedia-office . If you'd like to know how to better find information in Bugzilla that interests you, if you have ideas and criticism how to make Bugzilla better, this is your chance! ;) [17:21:56] * andre__ tries to reproduce https://bugzilla.wikimedia.org/show_bug.cgi?id=26490 as part of the LQT bugday [17:28:30] andre__: uselang=ksh ;) [17:29:05] Nikerabbit, any semantic mediawiki developer here? [17:29:25] Rahul_21: how hardcode needed? [17:29:34] Rahul_21: there is also #semantic-mediawiki [17:30:13] Nikerabbit, good point. Done. :) [17:30:14] Nikerabbit, no one 's replying there!i was looking jereon de dauw,but he seems away [17:31:03] Nikerabbit, i read about smw and i liked so i thought ill dive into it! [17:31:36] Rahul_21: it's cool stuff (though not always polished) [17:31:55] did you have something specific in mind? [17:34:02] mlitn, fabriceflorin, denis_o: hello together. Sorry my internet is currently slow and not stable, so maybe irc-messages to/from my can get lost, please ping here as this channel is logged [17:34:37] thanks se4598 - good to know [17:40:24] Hi denis_o and se4598 : No worries, we will also send email updates about today's AFT5 deployment on German Wikipedia. Per my last email, we're waiting on one final code review from lwelling1 and a final approval from tewwy. Once we have go-ahead, mlitn estimates that it would take about an hour or so to deploy to German Wikipedia, most likely after 20:00 your time, possibly later. We will keep you posted. In the meantime, this checklist can [17:40:25] provide more details on what we are attempting to do: http://etherpad.wikimedia.org/AFT5-release [17:42:16] Continuing with the LQT bugday in this channel: I cannot reproduce https://bugzilla.wikimedia.org/show_bug.cgi?id=26490 anymore - attached a screenshot and closed as WORKSFORME. [17:43:55] * valeriej is checking out https://bugzilla.wikimedia.org/show_bug.cgi?id=21703 [17:44:21] fabriceflorin I'm good on the last patch I was to look at. I think your only blocker now is making tewwy happy about that slave sync patch [17:45:46] Thanks, lwelling1, much appreciated! mlitn, can you give us your view on tewwy's question about the slave sync patch? [17:48:26] I'm not happy :-) [17:49:07] But I'm okay with it at the moment. I just want to make sure the risk is much lower. As far as I can tell this only makes sure data is consistent. DB load is tiny so I don't think this patch is a problem. [17:49:53] THis only affects the feedback database so who cares if we go to master at the moment. [17:50:05] emphasis on: at the moment. [17:50:38] Are we sure that rolling AFTv5 out again (note: not this patch) will not cause a breakage like Friday? [17:52:58] By "at the moment" I just want to explain. [17:53:25] I defer to mlitn about this, but my understanding from my conversation with him is that what caused the outage on Friday was the fact that he had forgotten to disable the hook to My Contributions, which I believe has been patched now. [17:53:35] I'm okay with this patch going out because it doesn't affect any of our dbs that matter. But in general I'm worried about the patch in the long term. [17:54:07] fabriceflorin: Yes, but there could be other problems that this crash papered over. I want to make sure those aren't there. (or as sure as possible) [17:54:55] fabriceflorin: Think of it this way: what is the benefit of AFTv5 going out now? Let's weight against risk. Are we sure there aren't other tables that aren't going to be affected other than contribs? [17:55:29] qgil, hello is the bug still not fixed https://bugzilla.wikimedia.org/show_bug.cgi?id=40062? [17:55:31] tewwy: I agree with you, and defer to you, mlitn and Asher to confirm that the risk is low enough to proceed. Sadly, I don't have the technical expertise to provide more recommendations at this time. :( [17:55:48] fabriceflorin: I'm okay with that. [17:56:32] fabriceflorin: I'm also okay with https://gerrit.wikimedia.org/r/#/c/53982/ going out "at the moment" though let's make sure we don't have self-approval in the future. Things aren't that important that we need to clutter our code with that "habit" [17:57:27] tewwy: I've looked at the log Asher had saved after Friday - those issues should be taken care of (although, since AFTv5 was mostly inactive, there's always a risk that more will be uncovered) [17:57:31] mlitn: I know Asher approved it offline, but it looks like self-approval and it sets a bad precedent. If he is uncomfortable with a +2 then lwelling1 should +2. [17:57:33] hi Rahul_21 [17:57:54] tewwy: I believe that Matthias checked with Asher before giving the self-approval, and only did it after Asher confirmed that it was OK for him to deploy. But again, we should let mlitn comment himself, so that we hear it from the primary source ;o) [17:58:03] qgil, hello is the bug still not fixed https://bugzilla.wikimedia.org/show_bug.cgi?id=40062? [17:58:20] Rahul_21, all I know is what you see in that report. I don't think anybody is working on it. [17:58:38] Rahul_21, are you interested? [17:59:09] fabriceflorin: That's not how it looks in gerrit. And someone else may see this and think it's okay to self-review, especially of they know the importance of this patch relative to impact on the wiki (small). This is a precedent I want to avoid for anything other than the most essential patches (hot fixes to site where nobody is available) [17:59:16] qgil, yes [17:59:41] Rahul_21, the best you can do is comment your intentions in the bug report. [17:59:48] fabriceflorin: Right now there is a spate of people self-reviewing changes and this has to stop. If it continues, people will lose access. And I don't want this happening on features AT ALL unless it's a hot fix. [17:59:55] Rahul_21, what solution you propose [17:59:58] tewwy: I definitely agree, I kind of rushed into that one, shouldn't have [18:00:54] mlitn: I'm okay with the dewiki deploy. But the "at the moment" I want to discuss. [18:01:04] Couldn't reproduce bug 21703 using the steps in comment 6 or on MediaWIki talk pages. Asked for retest. [18:01:59] mlitn: Basically the problem I have with the patch (in the long term) is that the general philosophy of sharding/memcache is to make it so read loads are against memcache and write loads are against the db. IN order to ensure this is the case, we can't ever delete from cache and expect the db to update it with the canonical. [18:02:49] qgil, i guess a simple css class should suffice!i do agree with helder's comment there [18:02:51] mlitn: While the db **is** canonical, we should be write-through the cache as well as read-through. By this I mean, in general we shouldn't have to read from the db unless that entry has been LRU'd out. [18:03:33] valeriej: that toolbar is WikiEditor so might make sense to check for existing bug reports under "MediaWiki Extensions > WikiEditor" [18:03:52] andre__: Oh, thanks for the tip. Checking now. [18:03:53] mlitn: So when we do anything like clear the cache entry and expect the db to update on the next request, we have a problem in design. RIght now this isn't an issue because the only thing consuming the cluster is AFTv5, but this will be an issue for Flow down the read if we keep along this path. [18:04:39] Rahul_21, comment in the bug report. The CSS class exists and is being used in Wikipedia. Another thing is whether MediaWiki should support that by default. But I'm just someone commenting, what matters are the opinions of the maintainers. [18:04:43] valeriej: but I cannot find a ticket when searching for "dup" and "twice" in the bug summary of WikiEditor bug reports, so I guess we're fine. [18:05:04] mlitn: So while I'm okay with the patch as is for now, I want you and others to consider if there is a way to avoid it for later because that later needs to set a pattern of interaction for access to flow, etc. IN general a sharded cluster should be so efficient that there isn't a need for read db replicas and the only db slave should be the failover hot spare. [18:05:35] tewwy: mlitn: fabriceflorin: please excuse me to intervene: to me as a non-tech-person it seems, that all critical topics *for the moment* have been solved and might be discussed in a moment less filled with expectations of the de-community and our promises to them? [18:05:38] mlitn: But for this release I'm okay with it, if that makes sense. At least I'm okay with https://gerrit.wikimedia.org/r/#/c/53982/ [18:06:00] That's not going to hurt anyone, and even if it never went out, all that would happen was data in AFTv5 would be inconsistent. [18:06:22] andre__: I couldn't find anything either. [18:06:22] denis_o: You are correct. [18:06:36] thank you very much tewwy [18:06:53] denis_o: Just be aware that because of the failed deploy in Friday, we're not 100% sure that this won't go without a hitch. :-) [18:07:26] denis_o: But you are correct, they can proceed with the dewiki deploy and I'm oaky with the last gerrit patch for the release [18:07:41] that's great news tewwy [18:07:56] denis_o: In fact, I'm okay with this patch being there for fabriceflorin's schedule for enwiki and frwiki AFTv5 stuff. [18:08:32] denis_o: It's just that I was sick yesterday, come in, see a huge thread on AFTv5 and got all kind of pissy. :-) [18:08:54] tewwy: that makes sense; do you mind me picking the discussion up again with you later this week - the original plan was to rebuild _all_ caches rather than purging, but there was a reason I changed something; I have to think it through again first, can't remember exactly what made me do it [18:08:59] tewwy: Thank you very much for your approval to deploy -- and for your reasonable guidance for improving our processes, which is much appreciated ;o) [18:12:34] denis_o and se4598 : Thanks you very much for your patience with all this -- and I really appreciate that you posted the URLs to the new help pages on our checklist: http://etherpad.wikimedia.org/AFT5-release -- this will enable mlitn to update the local configuration files to point to the correct help pages. [18:13:30] fabriceflorin: is there any delay to expect now or will we be up and running in 45 min.? [18:13:39] mlitn: Please let us know what we can do to support you in deploying to German Wiki. We are at your disposal for anything you need. [18:14:37] denis_o: I think mlitn can answer your question about timing better than I could. [18:15:00] *turningheadtomlitn* [18:16:15] denis_o: there is definitely some delay; I hope to have things up in around 1:15 [18:17:01] thx mlitn [18:22:10] mlitn: Not at all. Proceed. [18:34:46] denis_o and se4598 : Thanks for adding the link to the feedback guidelines in the Etherpad, which I had omitted to include. I just added a spot for the link to 'Terms of use' as well, which is a link in the feedback form. On the English Wikipedia, I believe we link to this version: http://wikimediafoundation.org/wiki/Feedback_privacy_statement - Did WMF legal translate this into German as well? [18:35:23] fabriceflorin: http://wikimediafoundation.org/wiki/Feedback_privacy_statement/de [18:36:30] PeterSymonds did this, I am unsure about his status to do this as a part of WMF legal and the need to be a part of that [18:37:20] valeriej, got an IE7 somewhere? https://bugzilla.wikimedia.org/show_bug.cgi?id=24378 would be another IE ticket if you have some time :) [18:38:00] Not sure if https://bugzilla.wikimedia.org/show_bug.cgi?id=31251 is a dup of bug 24292. Asked if commenters are still seeing a notification of new messages when they have none. [18:38:04] andre__: Checking now. [18:38:41] denis_o, fabriceflorin: It's a translation created by a normal wiki user: http://meta.wikimedia.org/w/index.php?title=Feedback_privacy_statement/de&action=history [18:38:47] valeriej, oh nice catch! [18:39:22] se4598: is http://wikimediafoundation.org an area for normal wiki-users? [18:40:11] denis_o: I will ping Michelle and Peter right now and Cc: you. In the meantime, should we link to the version Sebastian just posted? (Also, FYI, I have not yet received confirmation from mlitn that all these links will be enabled tonight, since he is working on getting the basic functionality deployed first -- it may be a while until the links are enabled.) [18:40:12] andre__: It's in the comments as possibly related but was never marked a dup, so I figured I'd ask explicitly if it's been fixed for them. [18:40:37] Ok, I'm done with LQT triaging for today. That was interesting. Will try to keep hashing out LQT enhancement requests during the rest of the week. [18:40:44] denis_o: No, I (and someone else earlier) asked PeterSymonds to transfer the message from meta to wmf [18:41:03] valeriej, good point [18:41:09] qgil, thanks so much for your help! [18:41:28] qgil: Thanks! [18:41:51] fabriceflorin: thx - michelle should know about this case as we mailed on it already in november/december [18:45:48] andre__: Can confirm Bug 24378 [18:49:10] valeriej, ah, pity. thanks :) [18:50:44] valeriej: by the way, your blogpost frequency becomes impressive. :) Again a great post, really thanks for summarizing the latest bugdays! [18:52:39] andre__: Ha, thanks! I've learned I have to write down the first drafts of my posts, so I don't get distracted. It's tedious, but it works. [18:55:13] valeriej, I see. Nice. [18:55:49] hoo|busy: Sorry to bother you, but could you explain the comment here: https://gerrit.wikimedia.org/r/#/c/54524/4/modules/ext.thanks.thank.js [18:56:25] Are you saying that results with errors get handled by fail()? [19:00:52] * andre__ away from keyword for a while [19:02:21] Bug 30698 is still valid. Commenting on report. [19:02:50] chrismcmahon: Hi Chris! I think I'm about ready to work on the automation part. I have my environment set up properly to do Ruby and run tests. The only thing is I don't have a local wiki, but bundle exec rake does work and generate a report. I'm not sure how accurate it is. [19:06:59] denis_o: I just emailed Michelle and Cc:d you and se4598 -- will keep you posted. For now, my recommendation is to use this link on a temporary basis, if it works for you: http://de.wikipedia.org/wiki/Wikipedia:Artikel-Feedback/Notizen (unless Michelle prefers we link to the official English version for now) [19:08:20] fabriceflorin: thx - but why Notizen as a link? thats nothing legal, but the aftv5-bugreport-page? [19:13:11] chrismcmahon: Hi Chris! Are you available now to talk with me? [19:13:19] Hi denis_o, This is for the privacy policy link that is shown at the bottom of all feedback forms: 'You agree to transparency '. It makes it clear to users how their feedback will be made public, even in a limited pilot. I defer to you on which link to use for the German Wikipedia. On the English Wikipedia, we now link to this privacy policy: http://wikimediafoundation.org/wiki/Feedback_privacy_statement -- I don't think [19:13:20] this is a big issue for today, but we do want to take care of it in coming days. [19:16:39] denis_o and se4598 : I'm just using the time productively to take care of minor loose ends, while we wait for mlitn to give us an update on the deploy :) [19:18:26] * valeriej if trying to confirm this bug: https://bugzilla.wikimedia.org/show_bug.cgi?id=23567 [19:19:24] fabriceflorin: the import script is currently running for dewiki data; I can not yet estimate how long it'll take to complete [19:19:42] meanwhile, I'll start looking if entries appear in other filters [19:29:57] hi rachel99, thanks for doing the Search thing with us last week [19:30:18] chrismcmahon: Sure thing. [19:30:57] chrismcmahon: I think my environment is all set up now. I'm using cygwin now and I've installed all the right gems, etc. [19:31:14] rachel99: great, cygwin is a good way to go I think [19:32:06] rachel99: so have you cloned the latest qa-browsertests repo from gerrit from earlier today? [19:32:18] fabriceflorin: se4598: denis_o: I expect the merge script to run for about another 45 minutes [19:32:20] chrismcmahon: No, I haven't done it today [19:32:42] judging from DB though, entries are now correctly marked as resolved etc [19:32:52] rachel99: or you could do "git pull" in your master branch to get the latest changes I think [19:33:00] mlitn: thx [19:33:04] zeljkof: are you around? [19:33:08] chrismcmahon: I can do it now. What about local wiki? Is that necessary to run tests? [19:33:13] mlitn: Thanks for the update. Glad that the data looks clean, and I've got the whole afternoons blocked off for testing this. [19:33:44] rachel99: no local wiki needed. one "design principle" (so far) is that the browser tests all point to public shared test environments [19:34:06] fabriceflorin: mlitn : se4598 : possibly i have to leave around this time [19:34:22] chrismcmahon: Great. Then I think it's working, as it does generate a cucumber.html file under reports. [19:34:26] mlitn: Would you be able to update the help page links that are now at the bottom of the AFT checklist? http://etherpad.wikimedia.org/AFT5-release [19:34:36] se4598: would you mind to make the announcement just in case? the text is ready [19:35:31] rachel99: if you do for example "$ bundle exec cucumber features/guided_tour.feature" do you see Firefox start up and jump around? [19:35:53] rachel99: (that would be from within the /browsertests directory) [19:35:56] chrismcmahon: Yes, I do see that. [19:36:03] great! [19:36:09] denis_o: it's ok for me [19:36:35] denis_o: If you want to make an announcement, I would definitely let people know that there may still be some issues. It's usually prudent to wait to make a broad announcement until we've confirmed in a small team that the application is working. But it's your call entirely, just wanted to recommend caution ;o) [19:36:50] mlitn: how about the negative numbers for the filters we've encountered friday? [19:36:57] rachel99: updating my own repo now, one sec... [19:37:38] thx fabriceflorin, will think twice [19:39:18] se4598: Matthias told us yesterday that the negative number issue should be solved -- it was apparently due to an incorrect article path name that was not matching the name on the German Wikipedia. [19:40:01] fabriceflorin: Will be this deployment now a test on full article set or just over a small subset? [19:40:39] rachel99: Would you like to add the scenario we wrote last week about searching with accents http://www.mediawiki.org/wiki/Qa/test_backlog to the real code base and start writing the code to automate it? [19:40:59] se4598: on testwiki; the numbers appear to be correct now [19:41:01] rachel99: there is one last piece of gerrit config we might have to do along the way [19:41:17] chrismcmahon: Yes, definitely. I'm trying to do git pull now. [19:41:28] se4598: caches are now still messed up though; after the data-import I'll clear them and hopefully have the same correct results on dewiki now [19:41:33] se4598: I believe this deployment will affect all 13k articles, as we had discussed on Friday. It may be possible to limit it somehow, but it might take more time to do this. So my current inclination is to try all 13k articles now, and to revert if we encounter serious issues -- unless mlitn has a better idea. [19:42:19] rachel99: ok, great. When your repo is up to date (or if you just want to clone from scratch), open /features/search.feature in your favorite editor and let me know... [19:42:23] chrismcmahon: git pull seems to have worked and updated the files [19:42:38] I think taking away the category on a lot of articles would be a lot of work [19:43:00] rachel99: you probably need to do "bundle install" to get the latest gems then [19:43:28] chrismcmahon: I did the bundle install last night. Should that be good enought? [19:44:20] rachel99: try running the test with "$ bundle exec cucumber features/guided_tour.feature" to see if you have errors now that need bundle install [19:44:27] mlitn: Is is at all easy for you to only deploy AFT5 for this article at first? http://de.wikipedia.org/wiki/Studentendoppelgrab (if that works on that article, we would deploy on all 13k articles). If this is not easy and will require more time, we can stick to the current plan, but wanted to ask, just in case. [19:48:02] chrismcmahon: you rang? :) [19:48:33] fabriceflorin: I'd rather not; there is a way to test the Special page prior to full re-activation though, I'll email more details [19:48:33] hi zeljkof rachel99 is just about to be in a position to start automating the Scenario we wrote last week [19:49:09] chrismcmahon: Would you have any time to help us test on German Wikipedia, when we deploy in about 30 mins.? [19:49:10] chrismcmahon, rachel99: great news! :) [19:49:18] and problems? [19:49:20] fabriceflorin: I should have time [19:49:39] chrismcmahon: cool, thanks. I will keep you posted. [19:50:09] zeljkof: well, I'm just running the guided_test feature now. Firefox is popping up and running the tests. [19:50:28] zeljkof: rachel99 is doing 'bundle install' now and I was just going to have her edit search.feature to add the new scenarios about searching with accent characters, and then try to have 'git review' work, and then watch what Cucumber does with a new .feature file [19:50:49] rachel99, chrismcmahon: sounds good to me [19:51:14] I will probably go to sleep soon, but fell free to send me mail if there are any problems [19:51:41] mwalker: cluster config, the script is mediawiki-config/favicon.php [19:51:56] zeljkof: I'll probably be called away in about 30 minutes, if you happen to still be around [19:52:44] Someone in the room? :) [19:53:05] zeljkof, chrismcmahon : the test finished, but there was one failure. Should I dpaste that to you? [19:53:10] rachel99: so I say open search.feature in your favorite editor and add the Scenarios we created from http://www.mediawiki.org/wiki/Qa/test_backlog to the search.feature file [19:53:22] rachel99: nope, I just wanted to make sure it would run [19:53:35] rachel99: known failure, being fixed right now [19:54:15] https://bugzilla.wikimedia.org/show_bug.cgi?id=35070 I was able to delete a post just by blanking the message and subject. [19:54:30] chrismcmahon: Ok, I will add scenarios to search.feature file. Just paste them into the search.feature file? [19:55:41] rachel99: yes, but follow the conventions in the existing Scenarios. Also, I think we should put in the two Scenarios that will pass right now and leave the one that will fail in the backlog page [19:55:54] rachel99: let me know when you do that and save the file [19:56:07] chrismcmahon: ok [19:56:12] How is the wikimedia/Wikipedia strategy regarding JavaScript? [19:59:03] I was wondering all these 'world' graphics here http://en.wikipedia.org/wiki/GDP [19:59:05] mich2: more like *what* is the WMF strategy re: JS [19:59:33] Would be such a nice use case for: http://jqvmap.com/ [20:00:43] I mean these graphics are static, and to make them a map which you could feed in data as you want (which year you like) or which colors you like, would be nice, not? [20:01:42] not in my opinion, no. [20:01:59] And why you think that? [20:02:01] But that's a content question which would be addressed by the community of the project. [20:02:22] Regarding the content ,yes [20:02:53] But is it technically possible to use/implement something similar today in Wikipedia ? [20:02:56] I generally use a 1.2GHz netbook; I'd rather not have every site I visit offloading more and more of their content processing onto my browser, thank you very much. [20:03:22] I see that point [20:03:29] Especially since that processing is not being cached for the tens of thousands of visitors who may see exactly the same content. [20:03:54] chrismcmahon: I did it and saved file [20:04:15] Krenair: Do you know what hoo-man is referring to in this comment: https://gerrit.wikimedia.org/r/#/c/54524/4/modules/ext.thanks.thank.js [20:04:20] rachel99: you have edited search.feature file? [20:04:22] fabriceflorin: mlitn: se4598: gotta leave now, sorry, didn't expect this to last so long [20:04:28] zeljkof: yes [20:04:37] zeljkof: on my system [20:04:42] se4598: you have the announcement-text [20:04:59] rachel99: now run it with this: bundle exec cucumber features/search.feature [20:05:00] place it when everything is ok. [20:05:05] Amgine: Yes but that's the point , you force them into seeing the exactly same content, interactivity should be at least an option , not? [20:05:17] rachel99: cucumber will let you know that some scenarios are not implemented [20:05:37] fabriceflorin: mlitn: se4598: i am reading e-mails [20:05:44] cu [20:05:48] denis_o: cya [20:05:55] denis_o: Sorry this took longer than expected. Sadly, it seems that major deployments like that almost always go overtime. :( se4598, will you be available to stay longer? [20:06:13] mich2: Some interactivity, yes. As simple and low-cost as possible. [20:06:57] Amgine: Good, what are the currently available technology's to do something like this today in Wikipedia? [20:06:59] (also keep in mind the range of platforms, especially the rapidly growing mobile) [20:07:05] hi rachel99 [20:07:23] rachel99: if you do 'git status' now, it should show the file you changed [20:07:25] Amgine: Oh I'd love to use this on my iPad [20:07:53] chrismcmahon: Hi. I'm running the bundle exec cucumber command now. will do that next. [20:08:12] wikitester2501: but would someone in Egypt want it loaded on their Nokia handset? [20:08:22] https://www.mediawiki.org/wiki/Resource_loader <- have you looked at this? [20:08:29] rachel99: OK. I was going there next for the new Search tests, but I want to make sure git is all ready to go [20:08:33] rachel99: what OS are you on? [20:08:50] zeljkof: rachel99 is using cygwin on Windows [20:10:16] Amgine: Thanks I didn't know the resource loader, I will have a look [20:10:22] yw. [20:11:16] fabriceflorin: yes, at least 2 hours [20:11:19] Amgine: I'm not trying to sell JS or something , I know you can do cool stuff without JS very well: http://queue.acm.org/detail.cfm?id=2436698 [20:11:41] chrismcmahon: git status looks ok. The cuke test failed at line 16- I have another feature line there. Maybe that shouldn't be there? [20:12:19] Amgine: I also really like efforts making Wikipedia available to low end hardware, like Wikipedia Zero [20:12:54] * zeljkof is done for today [20:13:09] rachel99: when you do 'git status' do see a line that says something like "# modified: features/search.feature"? [20:13:22] Amgine: That said, my Egypt friend in Cairo now already has an android mobile so :) [20:13:31] rachel99: feel free to send me mail if you have any questions and I am offline, I have sent you my e-mail in private message [20:13:32] se4598: Wonderful. Thanks so much for making the time to help us deploy! I will also be available for the next 3 hours, as I blocked most of the day for this project. [20:13:45] zeljkof: ok, thanks. [20:14:26] chrismcmahon: no. it says changes not staged for commits. [20:15:34] rachel99: does the whole message look something like http://dpaste.com/1028196/ [20:15:38] ? [20:17:52] Amgine: It's not like I'm trying to turn Wikipedia into wolfram alpha, maybe you could generate fixed images from data and use the image then as resource in articles [20:17:55] chrismcmahon: yes. it is here: paste.com/1028198/ My vi crashed, and during the recovery I made a new file and then moved it to the original search.feature name. [20:19:28] chrismcmahon: I'm not sure if that messed up things? [20:19:30] rachel99: OK, I think we're good. Next, take a look at http://www.mediawiki.org/wiki/Gerrit/Getting_started just for reference. What we need to do next is to make a branch... [20:20:53] rachel99: so the next step would be to "$ git checkout -b MEANINGFUL_BRANCH_NAME" so we know it is your branch that we'll be working on. Right now you're in the master branch, and we rarely if ever want to work directly in the master branch. [20:21:44] chrismcmahon: ok. so I should do that command now? [20:22:03] rachel99: yes "$ git checkout -b MEANINGFUL_BRANCH_NAME" (pick whatever branch name you like) [20:24:54] chrismcmahon: ok. done. branch is rachel99_mods [20:24:58] wikitester: no worries, I'm just a curmudgeon. Far, far too much js, imo, is implemented without even considering if it needs to be. [20:25:53] * marktraceur hands Amgine a copy of the Lynx browser - problem solved :) [20:25:59] rachel99: OK, now you want to add the file you changed, so you want to do "git add features/search.feature", exactly like "git status" showed [20:26:29] marktraceur: I use that nearly daily. [20:27:41] chrismcmahon: ok. done. how do I see if that worked? [20:28:08] rachel99: now if you do "git status" you'll see a msg about "Changes to be committed..." [20:28:23] chrismcmahon: I do see that. [20:29:01] rachel99: so the next step is to commit your changes so you can do "git commit -m "MY_MSG"" [20:30:31] chrismcmahon: Done. [20:31:10] Krenair: Do you know what hoo-man is referring to in this comment: https://gerrit.wikimedia.org/r/#/c/54524/4/modules/ext.thanks.thank.js [20:31:12] Hi [20:31:14] rachel99: OK, now for the last and trickiest step. What happens when you type "git review" right now? [20:31:23] I think he's saying that if you get an error, it won't use the .done callback? [20:32:03] * marktraceur high-fives Amgine [20:32:05] Krenair: That was what it sounded like, but that seem weird. [20:32:15] Um not really [20:32:17] New patchset: Rachel99; "modified search.feature to add test for accents search" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54719 [20:32:28] Pretty sure you're supposed to use .fail not .done to catch errors... I might be wrong there though [20:33:09] what data counts as errors though? [20:33:16] chrismcmahon:I think that worked- gerrit reported a new patchset. [20:33:30] The Lynx engine is used by many audio browsers, marktraceur. [20:33:38] I guess any top level item named 'error'? [20:33:53] kaldari, have you read the source code for this? [20:33:55] rachel99: Awesome! If you go to https://gerrit.wikimedia.org/r/#/ do you see your commit under Changes/My? [20:34:14] Krenair: no, I was hoping to avoid that :) [20:34:23] rachel99: I mean My/Changes. (you might have to log in to gerrit) [20:34:39] } else if ( result.error ) { var code = result.error.code === undefined ? 'unknown' : result.error.code; apiDeferred.reject( code, result ); [20:34:46] Amgine: What is an audio browser? [20:35:06] I imagine it's something like "Browser for audio files" but I may be wrong. [20:35:26] kaldari, https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/core.git;a=blob;f=resources/mediawiki.api/mediawiki.api.js;h=HEAD;hb=refs/heads/master#l150 [20:36:16] oops that downloads. try https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/core.git;a=blob;f=resources/mediawiki.api/mediawiki.api.js;h=cf7443f3b3bfa35e6c52a7043e76a2f17c52adf8;hb=refs/heads/master#l150 [20:36:57] chrismcmahon: yes. it appears to be there. I don't see the actual search.feature file though. You download it from that screen? [20:37:14] yeah, looks like you're right. [20:38:01] rachel99: no, that's fine. You should see a button on the gerrit page for your commit saying "Add reviewer". Could you please add cmcmahon and zfilipin as reviewers using that button? [20:40:12] chrismcmahon: did that [20:40:20] is terry chay on irc? [20:41:04] marktraceur: http://www.pcmag.com/encyclopedia_term/0,1233,t=audio+browser&i=57712,00.asp [20:41:55] rachel99: this is super. I was worried that 'git review' might be trouble. The reason we go through those steps is that we always do code review for any code that is not a pure experiment. [20:41:55] 404'd [20:42:30] chrismcmahon: yes, I understand. I am happy it all worked! [20:43:05] http://www.pcmag.com/encyclopedia/term/57712/audio-browser [20:43:17] Project browsertests-windows-internet_explorer_9 build #241: UNSTABLE in 12 min: https://wmf.ci.cloudbees.com/job/browsertests-windows-internet_explorer_9/241/ [20:43:18] MatmaRex, yes, he's signed in as tewwy [20:43:18] * zeljko.filipin: Updated multi_json Ruby gem [20:43:18] * zeljko.filipin: Updated Ruby gems [20:43:19] * zeljko.filipin: Moved MobileFrontend documentation to MobileFrontend repository [20:43:19] * zeljko.filipin: Updated multi_json Ruby gem [20:43:28] Ah, like a screen reader. Awesome. [20:43:40] rachel99: so after you make whatever changes or additions you a) put them in a branch that is not master b) do git-review to tell gerrit that there is code ready to be reviewed and c) get that code reviewed by others on the project. Zeljko always reviews mine and me his (even though he has more experience than me, he still gets things reviewed) [20:44:14] rachel99: also, once code is in gerrit we get automatic emails about status. it is very helpful. [20:44:58] chrismcmahon: I got it. Good system. [20:45:56] rachel99: we don't always pair-program, but we do always do code review. in fact, approving your own code changes in some projects can get the approver sanctioned in unpleasant ways. :) [20:46:27] chrismcmahon: Ok. got it. [20:51:11] rachel99: what does Cucumber show now when you run "bundle exec cucumber features/search.feature"? [20:51:38] MatmaRex: Yes, you're messing with my lunch btw. :-) [20:52:32] sorry, tewwy [20:52:53] MatmaRex: I haven't eaten since last night so I'm irritable. In any case, you got your revert +2'd, I'll make sure S has your suggetstions before he puts it up for review again. We'll also add you to the bug. It won't be for a couple weeks though [20:52:59] MatmaRex: because it's not exactly urgent. [20:53:07] but there's no hurry here, finish your lunch leisurely :) [20:53:09] MatmaRex: :-) [20:53:33] tewwy: i'll hopefully reply to your points on the revert commit when i'm able to [20:54:40] MatmaRex: I understand where you are. I'll have Howie force the design team do something about making a system for handling Agora. Right now if it appears design changes are sub rosa it's not because it's intended it's because Agora is very disorganized right now and doesn't have a proper process. [20:54:43] chrismcmahon: I get this: features/search.feature: Parse error at :16. Found feature when expecting one of: comment, doc_string, row, scenario, scenario_outline, step, tag. (Current state: step) [20:54:44] it's already 10 PM here, and i've got real-life work to do still :/ [20:55:05] rachel99: what do you think is causing that? [20:55:07] also, on an unrelated note. [20:55:08] MatmaRex: Understood. And you can send an e-mail to tchay@wikimedia anytime. I do read them. :-) [20:55:18] i've been having this conversation on three different irc channels. [20:55:26] there's something very wrong about communication fragmentation. [20:55:36] eh. [20:55:57] MatmaRex: That's not unrelated. ;-) [20:56:29] chrismcmahon: an extra feature command, or perhaps an extra space after word Feature? [20:57:25] MatmaRex: What happened here is small potatoes compared to https://bugzilla.wikimedia.org/show_bug.cgi?id=44394#c38 , so I'm not that put out by the revert ;-) [20:58:13] rachel99: yes, the extra Feature line [20:58:21] rachel99: I'll do a thing, one moment [20:58:32] tewwy: yeah, i just got the bugmail, reading it [20:59:28] New review: Cmcmahon; "There should only be one "Feature" line in any .feature file. The two Scenarios look OK to me, but ..." [qa/browsertests] (master); V: -1 C: -1; - https://gerrit.wikimedia.org/r/54719 [20:59:29] MatmaRex: I especially enjoy ori-l's patch [21:01:05] off to lunch [21:01:08] rachel99: gerrit should be sending you email now that your code was reviewed and got a comment. (hopefully the message isn't too negative) [21:01:28] rachel99: gerrit should also show that in the UI under My/Changes [21:02:52] Today's bug day is now over. Thanks to everyone who participated! Feel free to triage LQT reports throughout the week. Info. can be found here: https://www.mediawiki.org/wiki/Bug_management/Triage/20130318 [21:03:07] chrismcmahon: Ok, I see message in my mailbox. Should I just vi the file and make the change? [21:04:26] valeriej: just wondering - do you have themes ready for further bug days? :) [21:04:32] rachel99: we are getting into an area where I lack expertise, but to the best of my knowledge the bottom of this page is correct: http://www.mediawiki.org/wiki/Gerrit/Getting_started about amending commits. [21:05:36] Yeah! Thanks everybody for joining this bugday on LQT, and big thanks to valeriej! [21:05:56] MatmaRex: Good questions! We do not. I'm making a thread now looking for suggestions. Give me a minute... [21:06:19] MatmaRex, ideas are always welcome! [21:07:25] best place to propose might be https://www.mediawiki.org/wiki/Project_talk:WikiProject_Bug_Squad [21:07:30] chrismcmahon: I will follow that. Do I need to do a git pull first as it says? [21:08:08] rachel99: also, check what "bundle exec cucumber features/search.feature" shows you before you amend your commit. [21:08:42] valeriej: https://bugzilla.wikimedia.org/buglist.cgi?quicksearch=%3Ainterface ? ;) [21:08:44] andre__: ^ [21:09:12] (sorry about not joining today, i was busy ;), but i WORKSFORME'd two interface bugs instead.) [21:09:32] MatmaRex, hehe, thanks [21:09:38] yeah, let's try that [21:09:46] rachel99: you shouldn't need to do git pull I don't think [21:09:56] MatmaRex: https://www.mediawiki.org/wiki/Thread:Project_talk:WikiProject_Bug_Squad/Bug_Day_Topics_for_April :) [21:12:24] valeriej: thanks, i'll comment [21:12:41] MatmaRex: Thanks! [21:13:28] bluh, my attempts at being smart are failing. Scribunto question: If there is a Wikipedia module:Revision, I should be able to rettable = require("Revision") from a lua script right? [21:15:11] Hi guys. [21:16:22] chrismcmahon: the Test ran ok, thought it came back with: 4 scenarios (2 undefined, 2 passed) . Do you want me to dpaste the output> [21:18:35] Is this LQT bugday still on? [21:19:31] Krenair: The bug day finished, but you're welcome to triage reports. We'll be doing that throughout this week. [21:19:37] Krenair: andre__ was thanking everyone about 15 min ago. [21:19:44] Ah okay. [21:19:48] https://www.mediawiki.org/wiki/Bug_management/Triage/20130318 [21:19:58] I had my sixth form interview earlier so wasn't really able to take part. [21:20:00] Oh, should change the topic. [21:20:02] valeriej: I was thinking I might come try to do a bug per day this week :) [21:20:05] I've still not caught up with all my emails. [21:20:16] Krenair: No worries! [21:20:22] lizzard: That would be great! [21:20:30] I want to be more familiar with your bugzilla and its quirks! [21:20:37] rachel99: Cucumber should have said something like "You can implement step definitions..." [21:21:04] lizzard: Feel free to ask any questions! I'll have to join your next bug day, as well! [21:21:27] valeriej: awesome - the next focused one will be on Bugzilla itself so useful to many of us [21:21:35] chrismcmahon: yes, it did say that on the first line. [21:21:36] i will ping y'all when i set up the page for it [21:21:43] lizzard: Awesome, thanks! [21:23:06] Anyone know the usual topic for this channel? [21:23:07] rachel99: awesome, check in the latest to gerrit as an amendment to that change OR you could just abandon the current commit and commit the latest correct changes as a fresh commit [21:30:14] New patchset: Rachel99; "modify search.feature again to delete extra feature line" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/54724 [21:31:44] mlitn: http://de.wikipedia.org/wiki/Spezial:Artikelr%C3%BCckmeldungen_v5 is "nicht aktiviert", richtig? [21:32:07] chrismcmahon: at this point, it is [21:32:59] New review: Cmcmahon; "ready for next steps" [qa/browsertests] (master); V: 2 C: 2; - https://gerrit.wikimedia.org/r/54724 [21:33:45] rachel99: did you notice that not only do we have email msgs. from gerrit, we also have an IRC bot? ^^ [21:33:59] chrismcmahon: yes, I do see it. [21:34:33] chrismcmahon: I will have to go soon. Can we continue this tomorrow or what do you suggest? [21:34:44] rachel99: congratulations! what we did in gerrit just then was very sophisticated [21:35:06] chrismcmahon: Thanks! Glad it worked :) [21:35:22] rachel99: for the next move, look over some of the step definitions, that's what we'll do next https://github.com/wikimedia/qa-browsertests/tree/master/features/step_definitions [21:35:59] TimStarling: just updated https://gerrit.wikimedia.org/r/#/c/53718 [21:36:50] rachel99: tomorrow is great, I'll have some time around mid-day, or ask zeljkof about writing step definitions if you're around in the morning. (Z is better than me by a long way, too) [21:37:07] chrismcmahon: will do. Thanks for all! [21:37:28] rachel99: thank you! you're doing really great picking this all up from scratch [21:38:23] chrismcmahon: thanks. I have worked on a lot of related things, such as subversion, and scripting languages, so that is helping. [21:39:11] chrismcmahon: Bye for now. [22:15:30] AaronSchulz: https://www.mediawiki.org/wiki/User:Bsitu/Echo_Database_Discussion [22:48:59] ^demon: around? [22:49:06] help with a repo creation? [22:49:12] (and import)? [22:49:14] <^demon> sup? [22:49:32] hey ^demon [22:49:45] can you create analytics/limn-mobile-data and import github.com/wikimedia/limn-mobile-data [22:49:45] ? [22:50:12] this thing is going to get cloned by puppet and 'deployed' in the cluster (stat1), so needs to be in gerrit [22:53:21] <^demon> Done. [22:53:33] ^demon: thank you! [22:54:08] <^demon> yw [22:54:21] Hi ^demon [22:54:29] <^demon> howdy [22:54:30] Is something funny with Gerrit searching at the moment? [22:54:33] https://gerrit.wikimedia.org/r/#/q/status:open+project:mediawiki/core+-Verified-1+-CodeReview-1+-CodeReview-2,n,z [22:54:38] New patchset: Krinkle; "Add mwext-VisualEditor-docgen job" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53995 [22:54:52] It's including results with CodeReview-1 and -2 [22:55:10] New patchset: Krinkle; "Enable mwext-VisualEditor-docgen job" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/53997 [22:55:32] something funny going on with Gerrit* [22:56:22] mlitn and se4598: How is the deployment coming along on your end? Would you like to continue or postpone? (I am prepared to continue, but I am conscious of the fact that it's really late on your end). [22:56:56] I swear searches like this used to exclude such results, so the query I'm giving it should be valid... [22:57:38] <^demon> Weird. [22:57:42] <^demon> I dunno, offhand. [22:58:09] fabriceflorin: the totals issue appears to be resolved, so we're pretty much ready to re-enable for testing purposes, if that's still ok [22:59:42] mlitn: I am prepared to go ahead with the testing, but only if you are comfortable to do this at this late hour, and if you are not worried about possible issues. [23:02:15] fabriceflorin: we can go on, I will stay till 1am [23:02:51] fabriceflorin: one hour ;) [23:03:55] tewwy: is Echo made so that it can go on the extension1 cluster? [23:07:21] Aaron|laptop: I'm not certain, I got the impression it could be but isn't right now. [23:07:40] What is the extension1 cluster? [23:07:42] Aaron|laptop: You or lwelling should ping bsitu about it. [23:07:49] or binasher :) [23:08:10] Krenair: a db cluster independent of core mediawiki, it's being used by AFTv5. [23:08:36] Krenair: it's a holding are for things that can be sharded in the future (because joins can't be done against mediawiki's core tables) [23:09:06] Aaron|laptop: tewwy: i think i'll try to organize an echo db meeting for later this week [23:09:13] Krenair: IN this case, it'd be because the fact that the mailbox queue is not supposed to be in the db in the long term (it should be stored in redis or something, but there is no time) [23:09:44] binasher: Thanks! Add lwelling and required, you can put me as optional, but I'll be happy with whatever outcome you guys agree to. [23:09:54] and of course bsitu [23:12:07] tewwy: great! is kaldari working on echo too? [23:12:23] ya [23:12:50] although more front-end, less back-end [23:13:12] ok.. kaldari i'll add you as optional to the echo db review, feel free to ignore if you'd like :) [23:13:22] thanks [23:16:24] mlitn se4598: As discussed, I will be available to test on the German Wikipedia for the next 90 minutes or so, then available again to test some more this evening, while you are sleeping ;o) Please let us know as soon as possible when the application is up and running -- and ready to test. Also, chrismcmahon may be able to test. [23:17:17] fabriceflorin, chrismcmahon: it just got enabled [23:17:17] fabriceflorin: you can start testing ;) [23:17:39] fabriceflorin: http://de.wikipedia.org/wiki/Spezial:Artikelr%C3%BCckmeldungen_v5 is up [23:18:17] mlitn: Aaron|laptop: notpeter: https://ishmael.wikimedia.org/sample/?hours=120&host=db1030&sort=ratio [23:18:39] aftv5 is looking much better this time, at least so far [23:19:10] binasher: thanks; I'll keep an eye on that chart too [23:19:11] let's hope it keeps behaving [23:19:56] fabriceflorin: can you confirm that feedback form is not shown to logged out users? https://de.wikipedia.org/wiki/Gro%C3%9Fer_Ameisenb%C3%A4r [23:20:00] mlitn: so AFTv5 does not appear on random articles, only a select list [23:20:07] indeed [23:20:52] chrismcmahon: https://de.wikipedia.org/wiki/Kategorie:Wikipedia:Artikel-Feedback_Grundstock [23:21:01] se4598: feedback form *is* shown to logged-out users [23:21:30] se4598: example: http://de.wikipedia.org/wiki/Gro%C3%9Fer_Ameisenb%C3%A4r [23:23:16] chrismcmahon: ok, seems *several* action=purge helps [23:23:22] New patchset: Krinkle; "Add mwext-VisualEditor-docgen job" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53995 [23:25:27] mlitn: the link to hide own feedback seems not to work for me [23:25:35] indeed, just noticed that too [23:25:46] i know what the problem is, but fixing it today won't work anymore [23:26:20] getting *many* not-logged-in feedback entries right now on the central page [23:27:51] mlitn: fabriceflorin on the Central page, sorting by Neueste etc. doesn't work, reverts to Relevant regardless of what is chosen in the selectbox [23:28:56] Change merged: Krinkle; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53995 [23:29:03] New patchset: Krinkle; "Enable mwext-VisualEditor-docgen job" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/53997 [23:30:02] chrismcmahon: seems to be only an UI-issue, sorting works for me for "Nicht relevant" and others [23:30:17] chrismcmahon: it does seem to sort as relevant, but the value in the dropdown still shows "relevant" [23:30:33] se4598: OK. this is Chrome/OSX, checking FF [23:31:54] mlitn: se4598 I don't think those sorts are correct. Alteste shows Relevant and the top entry is vor 3 Monate [23:32:19] * chrismcmahon can't type A-umlaut easily, sorry [23:32:40] chrismcmahon: if I click alteste, I do get the oldest posts (3 months) [23:32:54] but the value it displays in the dropdown is still displayed as "Relevant" [23:33:15] the sort seems to work, but the value in the dropdown shows "Relevant" regardless [23:33:18] or so it seems to me [23:35:38] mlitn: OK, but sorting by Neueste shows vo 5 Tage, and that can't be right. [23:36:01] chrismcmahon: are you in another filter? [23:36:45] chrismcmahon: filter "Hervorgehoben" (="featured")? [23:36:51] se4598: that's what it was, thank you, I wasn in Ungepruft [23:36:58] :) [23:37:02] mlitn: the feedback (#17) isn't 3 months old but was posted on 4. Dezember 2012 [23:37:32] se4598: do you have a link to that one? [23:37:35] * chrismcmahon used to be able to converse in German, this is fun :) [23:37:42] mlitn https://de.wikipedia.org/wiki/Spezial:Artikelr%C3%BCckmeldungen_v5/Schach/17 [23:37:47] thank you [23:38:36] se4598: 4 december to 20 march is not much over 3 months, right? [23:39:15] on april 4th, it should switch to saying "4 monate" - or am I missing something? [23:39:26] mlitn: confused mind, my fault [23:39:40] it's ok, it's late :) [23:41:19] I may have messed up something on https://de.wikipedia.org/wiki/Spezial:Artikelr%C3%BCckmeldungen_v5/Schach/17 I lack context to figure it out though [23:41:59] what do you think you've done? [23:42:12] (btw, chris, if the german is hard, add ?uselang=en to the url ;) ) [23:43:27] mlitn: checking, thanks. [23:44:37] mlitn: want to check that the history under https://de.wikipedia.org/wiki/Spezial:Artikelr%C3%BCckmeldungen_v5/Schach/17 is sane? [23:45:48] chrismcmahon: makes sense: se marked as useful, you undid, you marked as useful, you undid, you marked as non-actionable, you undid, you marked as useful [23:45:55] seems legit [23:46:09] mlitn: sounds reasonable, thanks [23:49:12] any concerns against publishing the announcements? fabriceflorin [23:49:46] se4598: mlitn other than the selectbox display not being sticky? [23:50:15] "hide my post" doesn't work; selectbox value is sticky; those are the only 2 I'm aware of [23:50:16] (always shows "Relevant") [23:52:11] Hi mlitn, se4598 and chrismcmahon, I am sorry that I could not do any testing until now. I will give you my answer on announcing this in about 30 minutes. Is that OK with you? In the meantime, could each of you email me a list of all the known errors so far, along with your recommendations as to whether or not we should announce this deployment? Thanks. [23:53:43] fabriceflorin: see email [23:53:58] we've uncovered only 2 so far, listed in the email [23:54:03] Yippie, build fixed! [23:54:03] Project browsertests-windows-internet_explorer_9 build #242: FIXED in 14 min: https://wmf.ci.cloudbees.com/job/browsertests-windows-internet_explorer_9/242/ [23:54:32] fabriceflorin: I'll stick around for another 0:30 until you finish testing [23:54:47] se4598: will you still be around in 0:30, for the announcement? [23:55:08] mlitn: fabriceflorin I'll be signing off in about 5min [23:55:25] chrismcmahon: okay, thanks for the assistance! [23:55:39] mlitn: good job making this happen! [23:55:56] mlitn, fabriceflorin: yes, ok; but I will now prepare for sleep, so afk ;) [23:56:59] mlitn: Thanks, I just read your email, and on that basis, I would go ahead with the announcement, as long as we clearly disclose to the German users that the feature still has issues, and list those issues precisely. I am assuming that se4598 is comfortable with making that announcement. [23:57:54] binasher: now that it's been re-enabled in full on dewiki for abour an hour, do you see any warnings on your end?