[14:47:14] evening [15:26:54] Reedy: ping ? ;- D [15:27:08] Reedy: the images directory in /home/wikipedia/common is ignored in .gitignore [15:27:29] I guess we could put them in git as well [15:31:36] heh [15:31:49] i'll have a loo [15:36:25] @google define:loo [15:36:57] I should take notes of all the english words I am seeing :-D [15:38:08] that would be a small list! [15:38:19] arhh [15:38:32] I should also take note of all the time I am forgetting in my sentences [15:38:48] I meant "all the NEW TO ME english words" ;-] [15:40:59] hashar: could you request two features for wiktionary mobile: word review list, and "new to me" button (which adds current word to word review list) ? [15:41:28] amgine: ? Cant you open enhancement requests in bugzilla? :-] [15:41:42] yes, can, just not in a time/space to do so... [15:42:05] well do that when you are available to think about the feature :D [15:42:18] kk\ [15:42:33] there is probably no point in having someone just copy paste "word review list button" and "new to me button" [15:42:38] that is not going to help the developers [15:42:48] you will want to write a few paragraphs ;-] [15:43:23] You are evil. [15:43:34] na just pragmatic hehe [15:53:11] hashar: surely loads of those images are old logos, not used now? [15:53:28] the sul ones might be.. [15:53:31] at least the images/sul/ should be used ffor single logging [15:53:37] so any news on what happened with jenkins-bot yesterday? [15:53:47] nullspoon: no idea :-( [15:53:57] nullspoon: I investigated a bit, found nothing and gave up. [15:54:13] One off isn't so bad [15:54:19] Its when it happens regularily/repeatedly [15:54:19] the gallium box crashed for some unknown reason. I have just manually retriggered the jobs [15:54:21] hashar: Hopefully it wasn't my submitting to patch sets to close together for the same change :) [15:54:23] Reedy: I will add the images/sul/ [15:54:25] two* [15:54:31] ook [16:00:34] Reedy: https://gerrit.wikimedia.org/r/13030 and https://gerrit.wikimedia.org/r/13031 [16:00:44] Reedy: could you review them please ? [16:00:47] did the change on fenari [16:00:57] so master there is a fork currently :-] [16:01:23] ah I forgot the .gif [16:03:16] thanks! [16:19:32] arrahhh putenv( "GDFONTPATH=/usr/local/apache/common/fonts" ); [16:19:36] I love our conf [16:41:30] Does anyone here have a lot of experience with database work in MW extensions? I'm in the middle of writing one, it's not cooperating [16:41:34] https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/EtherEditor.git;a=blob;f=includes/EtherEditorPad.php;h=6d31b554c817ed15244766dc238681cf603447ca;hb=7c7339395b160d16f2a8108789b0bac87814c609 [16:42:25] Essentially the problem is this: I call authenticateUser, then separately I call getContribs(). If I do it in that order, getContribs() should return the name of the current user, at least [16:42:58] And on a local server with sqlite, it does. On the remote server running MySQL, it returns null. [16:56:22] I am out see you tomorrow [17:04:59] marktraceur, that looks very odd [17:05:20] Platonids: I can't disagree with you, to be sure [17:05:31] (though I have this sinking feeling that you mean my code) [17:05:38] when you test using MySQL is it a single server or are slaves present? [17:05:54] maybe you're just hitting a lagged slave [17:06:13] Platonides: Uhhh, I did nothing special to set it up, I'm guessing it's just a single server [17:06:50] is it a local mysql install? [17:07:24] Platonides: Yes [17:07:49] than I don't think you have inadvertedly setup mysql replication :) [17:10:03] Platonides: I sure hope not, sounds painful [17:11:17] Hello, AaronSchulz matthiasmullie yoni_omniti chrismcmahon: We're getting ready to deploy Article Feedback 5 more widely today. Before we can deploy, we have a half dozen Gerrit tickets that need to be reviewed, and bsitu and kaldari are kindly helping us with this. AaronSchulz , would you be able to help as well, in RoanKattouw's absence? [17:12:29] Here are some of the gerrit tickets that still need to be approved. Aaron, bsitu and kaldari, can you please let us know ASAP which tickets you can review our have questions about? [17:14:39] https://gerrit.wikimedia.org/r/#/c/12973/ - Change Iec9a6199: update lottery checks in JS (I am listing these AFT5 revisions in order of priority) [17:15:22] https://gerrit.wikimedia.org/r/#/c/13002/ - Change I5e6938c3: Bug 37950 - Article Feedback - 'View Activity' does not appear in Editor's view [17:15:41] * AaronSchulz finishes some hiring stuff [17:15:50] https://gerrit.wikimedia.org/r/#/c/13006/ - Bug 37953 - Article Feedback Page - New call to action below page title [17:16:20] https://gerrit.wikimedia.org/r/#/c/12711/ - Bug 37847 - Add profiling to AFTv5 functionality [17:16:43] https://gerrit.wikimedia.org/r/#/c/12217/ - Bug 37580 - Article Feedback - Special Display for IPv6 Addresses [17:17:12] https://gerrit.wikimedia.org/r/#/c/13019/ - lower AFTv4 odds to kickstart AFTv5 at 1% (inverse odds) [17:17:24] * AaronSchulz looks at https://gerrit.wikimedia.org/r/#/c/12711/ [17:17:40] https://gerrit.wikimedia.org/r/#/c/10394/ - AbuseFilter: Make it possible to have differing default warning messages for differing filter groups. [17:18:51] Thanks, AaronSchulz, much appreciated! [17:19:22] I'll look at 12973 [17:19:41] Thanks, Kaldari, that's wonderful! [17:20:04] hi kaldari [17:20:15] Did you get my email about the jury tools? [17:20:28] oh yeah, sorry I haven't replied to it... [17:20:31] Platonides: Did you have any other ideas? /me is pretty much blocked on this [17:21:17] kaldari: Nah, don't reply, just build it ;-) [17:22:12] multichill: I think it would be a huge help. I was wondering if any of the POTY folks would be interested in adapting their code, as it seems like it serves a similar function. [17:22:31] I think mono worked on some of that [17:23:05] multichill: anyway, I'll have to circle back around on this, as we're preparing for a deployment right now [17:23:23] * AaronSchulz looks at https://gerrit.wikimedia.org/r/#/c/12217/ [17:23:25] Maybe you can respond to the thread later [17:24:45] What are you deploying today kaldari? The version numbers are increasing rapidly [17:25:35] multichill: We're getting ready to deploy some updates to ArticleFeedback [17:27:44] matthiasmullie: what am I looking at anyway? (dashboard.js). The file comment just says "Script for Article Feedback Extension" [17:30:04] marktraceur, I can only suggest you to try create a reduced testcase [17:30:16] Hrm [17:31:24] Oh, I guess the other interesting bit is getOtherPads--also returns null, though other pads exist [17:31:48] Maybe my database is totally shot [17:31:54] I know it was buggy before [17:33:39] * marktraceur sighs [17:33:47] The tables are there, they have stuff in them [17:33:55] Hmm, it looks like I should get the editor-engagement group to own these extensions. :-( [17:36:08] bsitu: do you have +2 on aft? [17:36:24] AaronSchulz: no [17:36:33] ah, that explains [17:36:41] :) [17:36:45] Kaldari: Thanks for reviewing Gerrit #12973 - Here is the Bugzilla ticket which explains how our lottery works: https://bugzilla.wikimedia.org/show_bug.cgi?id=37616 (the explanation for how article IDs are parsed is explained at the end of this thread) [17:41:23] Platonides: If I understand this correctly, the only thing that's not working is the select query [17:43:22] AUGH [17:43:25] Got it [17:43:33] matthiasmullie: why was the code checking can_delete before https://gerrit.wikimedia.org/r/#/c/12711/? [17:43:40] was that just a copy/paste error? [17:46:14] likely c/p error, yes; didn't match requirements with these too strict permissions [17:54:56] Hi matthiasmullie : for the record, I find that the 'BETA' label at the top of the feedback page is showing a garbled message on prototype (<articleFeedbackv5-beta-label>): http://prototype.wikimedia.org/release-en/Special:ArticleFeedbackv5/Golden-crowned_Sparrow (I assume that this is because we did not update the messages, but wanted to point that out, so you are aware of it) [18:01:55] fabriceflorin: yes thanks, can't clear the message over there indeed; works fine locally though, but let's not forget to look at it on test as well [18:05:21] are you guys keeping wmgArticleFeedbackLotteryOdds at 100 for en.wiki or changing that? [18:07:02] Matthias: Cool. Let's definitely remember to check this on test and production. On a separate note, how easy would it be to only show the alternate user name method for anonymous users to only show when they have a new IPv6 code, not a standard IP code, as originally requested? (I realize this would be for next deployment, but wanted to ask anyway). https://bugzilla.wikimedia.org/show_bug.cgi?id=37580 [18:07:11] they should be going to 99% (https://gerrit.wikimedia.org/r/#/c/13019/) at which point aftv5 will kick in for the other 2% [18:07:14] 1%* [18:08:44] fabriceflorin: <30min [18:12:00] Matthias: Thanks! I added this request on that Bugzilla ticket for the next deployment. [18:13:20] matthiasmullie: Can you let us know which critical Gerrit tickets are still awaiting review? Please include their Gerrit URL, to make it easier for our reviewers ;o) [18:16:28] AFTv5: https://gerrit.wikimedia.org/r/#/c/13006/ - Bug 37953 - Article Feedback Page - New call to action below page title [18:16:30] AFTv5: https://gerrit.wikimedia.org/r/#/c/12973/ - update lottery checks in JS (Kaldari is looking into that one) [18:17:52] Thanks, Matthias, we just discussed 12973 and I think we clarified Kaldari's questions about it. [18:18:16] Config: https://gerrit.wikimedia.org/r/#/c/13019/ - lower AFTv4 odds to kickstart AFTv5 at 1% (inverse odds) [18:18:34] The first ticket you listed, 13006 should be very easy to approve, since it's primarily a text copy change. [18:18:42] fabriceflorin: I'm sorry - kaldari approved already [18:19:41] AbuseFilter: https://gerrit.wikimedia.org/r/#/c/10394/ - AbuseFilter: Make it possible to have differing default warning messages for differing filter groups. [18:19:43] [18:20:45] those are the 3 open critical ones [18:20:48] Does anyone know anything about rest of the Berlin videos? It seems after the first sprint of 4 videos last month it stopped? [18:21:10] matthiasmullie: Cool. AaronSchulz, bsitu and Kaldari, are any of you reviewing this Config revision? https://gerrit.wikimedia.org/r/#/c/13019/ [18:21:29] oh actually, I did have one remaining question about 12973... is there any mechanism for making sure that AFT doesn't show up on the Main Page? [18:22:21] is that in the PHP blacklist code? [18:22:44] matthiasmullie: ^ [18:22:58] kaldari: Good point. I do not know the answer. Is the Main Page considered to be in the article namespace? [18:23:05] yes [18:23:36] kaldari: it's in PHP indeed; JS won't be included if blacklisted [18:23:50] thanks [18:24:03] matthiasmullie: Great news! Glad you're on top of it ;o) [18:24:49] and for the article page, the check is in JS once more (to counter page cache issues from when the chech was not yet in php) [18:26:27] AaronSchulz, bsitu and Kaldari, are any of you reviewing this AbuseFilter revision? https://gerrit.wikimedia.org/r/#/c/10394/ This revision was originally made by Werdna (Andrew Garrett), but RoanKattouw requested a couple minor tweaks. Since Werdna is too busy with school, matthiasmullie made these minor tweaks on his behalf. [18:27:32] probably not [18:28:32] reviewed 13019 [18:28:36] what's left? [18:28:43] AaronSchulz: Can you please clarify what you mean by 'probably not'? [18:28:59] heh, I could look at 10394 if no one else is [18:29:27] Kaldari, thanks! That would be great if you could take 10394. Is anyone covering 13019? [18:29:46] 13019 is done [18:29:58] Oops, sorry, spoke too soon ;o) [18:32:19] matthiasmullie: Anything else that you think we need for you to deploy? Can you merge all these approved revisions on your end? Also are any of these 'non-critical' tickets more important than we thought? [18:32:20] • 12849: Bug 37616 - Article Feedback - Increase Test Sample to 10% of English Wikipedia [18:32:21] • 13027: fix typo's [18:32:22] • 13025: feedback on 11283 [18:34:43] 12849 - abandoned; doesn't need to get in; 13027 - really just some unimportant comment typo's; 13025 - merged by aaron already [18:34:56] so none that need to get in :) [18:35:04] just fixing up 1 last commit [18:37:32] matthiasmullie: Wonderful! Many thanks to AaronSchulz, bsitu and kaldari for jumping in and helping get this code reviewed! Much, much appreciated. [18:39:11] matthiasmullie: Did you hear back from Asher about today's DB schema change? Any questions on his end? Shouldn't be, since we did exactly what he proposed ... [18:39:41] A couple minor issues in 10394 that need to be fixed [18:39:57] AaronSchulz: Can you still be on call to advise matthiasmullie as we start pushing to test and en-wiki in the next couple hours? [18:43:38] I guess so, I'm eating some left over chinese food ;) [18:43:46] * AaronSchulz isn't going anywhere [18:49:43] matthiasmullie: One more question... The comments say "Rule out MSIE 6/7/8", but it looks like only IE 6 and 7 are actually excluded. Is that correct? [18:50:24] i.e. the feedback dashboard is now supported on IE8? [18:50:34] mmh; one of this weeks items is IE8 fixes; so guess we should remove that in the comments [18:50:58] * matthiasmullie is looking at fabrice for confirmation that IE8 will now be supported [18:51:19] fabriceflorin: ^ [18:52:58] Hi guys, sorry for not responding sooner. Once we confirm that IE8 is now working after deployment, we can remove the exception from the comments. But we have only tested it on prototype, not production. So if you are feeling bold, you can remove it now, or remove it later ;o) [18:54:00] kaldari: https://gerrit.wikimedia.org/r/#/c/10394 - updated [18:55:12] the exception has already been removed, so unless anyone says otherwise, it will be activated for IE8 users (despite the comment to the contrary) [18:56:34] fabriceflorin: everything was fine on prototype? then I'd go for pushing it to prototype (we'll have to do it some day, might as well do it sooner rather than later, since we'll only be upping the lottery from now on) [18:58:05] rereviewing 10394 [18:58:41] matthiasmullie: Yes, everything was fine on prototype from my standpoint. I am not sure if I understand your question above, can you please clarify? [18:59:21] 10394 merged [18:59:34] if everything was fine on prototype, I see no reason to not yet push it to test & production [19:00:12] or did I misunderstand you? :) [19:01:40] * AaronSchulz misses out on ice cream [19:04:08] matthiasmullie: Yes, I think we are ready to push to testing now. Thank you all for your help! (AaronSchulz: there is most likely time for you to quickly get some ice cream, as it will take a bit of time for matthiasmullie to push the code to testing. ) [19:04:21] one rev still needs fixing [19:04:50] AaronSchulz: Oops. Spoke too soon. Which rev still needs work? [19:05:01] https://gerrit.wikimedia.org/r/#/c/13006/ [19:05:34] I fixed up some of the comments in the dashboard script: https://gerrit.wikimedia.org/r/#/c/13050/ [19:05:42] Thanks! matthiasmullie: does this seem doable on your end? [19:06:40] kaldari: thanks for making the comments easier to understand - matthiasmullie, you may want to be a bit more verbose in your comments in the future, to make it easier for new reviewers to understand the code ;o) [19:12:07] will keep that in mind ;) [19:16:10] https://gerrit.wikimedia.org/r/#/c/11282/6/includes/specials/SpecialContributions.php will break FlaggedRevs [19:17:08] maybe I should revert that from master unless it gets fixed [19:17:54] AaronSchulz: Thanks for the heads-up. matthiasmullie: It seems we should revert this for this deployment, so we can push everything else in the short amount of time we have left. [19:21:51] AaronSchulz: it's just about the

not being included in $ret? [19:22:49] matthiasmullie: it's about the

yes [19:23:48] AaronSchulz: Cool. That's probably just a one-minute fix, according to Matthias (I just spoke to him on Skype). [19:24:16] I'd suggest adding $classes to the hook, so FR can just use a class instead of tag wrapping [19:24:46] then FR wouldn't have to care if the

s is there or not [19:24:49] *are there or not [19:29:52] AaronSchulz: https://gerrit.wikimedia.org/r/#/c/13051/ (core) & https://gerrit.wikimedia.org/r/#/c/13052/ (aftv5) - just a quick update that should revert the lineendings hook to the previous situation and makes aftv5 work with it [19:29:56] matthiasmullie: also, $ret is not always defined [19:30:10] oh - you're right [19:30:13] hmm, what's the $row->rev_id check for? [19:30:44] ok; might want to revert it and I'll take a closer look at it next week [19:31:16] it's there to make sure that the data being processed over there is actually a revision (and not *other* data, like an AFTv5 line) [19:33:55] ok [19:34:18] matthiasmullie: revert the whole commit? [19:35:09] that seems to be the safest choice atm [19:35:39] Yes, I agree. In the interest of time, it might make sense to revert now, and try again next week. :( [19:37:01] matthiasmullie and AaronSchulz : are you guys comfortable pushing to testing now? then to production? I realize that this will add a couple hours to our schedule, but this is an important deployment. I am fully prepared to do it if you guys can do it as well. [19:38:20] i'm ok [19:40:35] AaronSchulz: OK with you to push to testing now? [19:44:49] matthiasmullie: Let [19:45:23] matthiasmullie: Let's start pushing to testing now, in the interest of time. I'm assuming that AaronSchulz may have stepped away from his desk. [19:46:06] fabriceflorin: for testing, yes [19:47:38] AaronSchulz: Thanks! -- assuming all works out on testing, are you OK with going to production today? I realize this will take more time, but I am ready to do it if you are, so we can stay on schedule. [19:52:58] fabriceflorin: is ryan k around? [19:56:07] AaronSchulz: Yes, Kaldari is around. But Terry Chay had recommended that we involve you for any tricky back-end issues related to this deployment. For example, we are deploying the new DB schema approved by Asher, as well as the lottery, and Matthias may need to pick your brain if any issues come up once we're on production. [19:57:23] I'm not too familiar with the schema though, so I'd be a little weary [20:01:44] AaronSchulz: Terry's instructions were to go ahead with this deployment, provided you can provide some oversight. If nothing breaks on testing, we should be in good shape on production. If not, we can revert back. Are you OK with this plan? What about you, matthiasmullie ? [20:02:10] fabriceflorin: how simple is the rollback strategy? [20:05:51] AaronSchulz: see DM [20:06:37] AaronSchulz: I defer to matthiasmullie on your question about our rollback strategy. I assume that we would push back the old code. However, the comments DB would be split into two files, which may not be easy to revert. [20:24:24] matthiasmullie: is it on test? [20:25:22] AaronSchulz: not yet; i'm ready to pull the code on test; but awaiting confirmation on how to go about the schema change (can I just execute it?) [20:25:49] matthiasmullie: just for test? [20:26:08] right now; yes [20:26:34] if you're comfortable with pushing to production today - ok as well [20:26:39] if you'd rather wait, ok as well [20:26:58] what's the schema change? Did asher ok it "just to be done"? [20:27:56] I doubt there are many rows on testwiki [20:28:13] still [20:28:20] adding a bazillion indexes wouldn't be cheap :p [20:28:47] AaronSchulz/Reedy: https://bugzilla.wikimedia.org/show_bug.cgi?id=37707 [20:38:22] AaronSchulz: We just had a call to review our options with matthiasmullie, kaldari and bsitu. We determined that it is too risky for us to attempt deployment to testing or production today. We would like to wait another week until RoanKattouw returns, and deploy with his help next Tuesday, July 3. We will also coordinate with Asher to make sure that we make the DB change at the same time. I assume that you will support that decision, since [20:38:23] you seemed weary about going all the way to production today, for very good reasons. I will write our notes and send an email later. In the meantime, I would like to thank you for all your help today, it was really appreciated! [20:40:13] ok