[01:23:03] New review: Diederik; "Ok." [analytics/udp-filters] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/29469 [01:23:03] Change merged: Diederik; [analytics/udp-filters] (master) - https://gerrit.wikimedia.org/r/29469 [14:02:09] New review: Diederik; "Ok." [analytics/gerrit-stats] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/28367 [14:02:09] Change merged: Diederik; [analytics/gerrit-stats] (master) - https://gerrit.wikimedia.org/r/28367 [14:06:36] New review: Diederik; "Feedback has been incorporated." [analytics/wikistats] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/28506 [14:39:27] New patchset: Stefan.petrea; "Added module to parse iPad user agents and tests." [analytics/wikistats] (master) - https://gerrit.wikimedia.org/r/28506 [14:40:36] New review: Diederik; "Fixed merge conflict." [analytics/wikistats] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/28506 [14:51:31] New review: Diederik; "Ok." [analytics/wikistats] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/28506 [16:26:58] * marktraceur points the world at https://www.mediawiki.org/wiki/Google_Code-In [16:45:51] I am getting the following error while running UploadWizard extension : Call to undefined method Language::fetchLanguageNames(). Can someone guide me through it ? [16:51:14] hi drecodeam! [16:51:16] marktraceur: ^ [16:51:23] got a moment? [16:52:20] hey sumanah, i was actually writing you a mail [16:52:24] :) [16:52:39] I had to discuss about the bangalore meetup [16:52:49] drecodeam: too old mw core I guess [16:53:07] drecodeam: I'm not going to be at the Bangalore meetup in November [16:53:18] drecodeam: Tomasz Finc and Alolita Sharma are the people to discuss that with mostly [16:53:51] sumanah: Aha, thanks for that. I had to discuss about the sponsorhip, I just hope I am not pretty late for that :D [16:53:56] alolita: hey [16:54:09] drecodeam: financial sponsorship to attend? [16:54:51] sumanah: ya just the traveling cost, I was actually waiting for my sessional exams datesheet before I could plan out anything [16:55:22] I have arranged my staying bit, but needed help with the traveling [16:55:29] drecodeam: understood. OK, yeah, email me + Tomasz + Alolita with a cost estimate so we can approve it, and let us know whether you can be reimbursed afterwards or whether you need the money in advance [16:55:59] sumanah: Alright, I was writing that mail only. I guess I am fine with both the options:D [16:56:11] Great! :) [16:56:12] nice to see you again online, drecodeam [16:56:53] Ya, it has been a long time. I had started on a new job, so was waiting for things to stabilize before I could get back to contributing. [16:57:12] Its time I finish off whatever is left in my code :) [17:03:12] drecodeam: YAY [17:33:23] drecodeam: Welcome back! Did updating MW fix it, by any chance? [17:46:36] yuvipanda: "pops a cap" <-- this may have multiple meanings [17:47:18] I'm culturally aware of a single one - I've a bottle of Rum with a cap, and I did, indeed, pop it. (it's cheap rum) [17:47:57] marktraceur: I just googled. Damn [17:47:57] *nod* [17:49:04] Hi mlitn, spagewmf, bsitu : Look forward to a small deployment for Article Feedback and Page Curation today! Here's the main thing we're deploying for AFT5: https://trello.com/card/aft5-signup-cta-conversions-using-current-data/5069e3859c41bc8f712411a3/6 [17:49:07] marktraceur: I keep bumping into those things. Both IRL and on email. [17:49:17] marktraceur: thanks for pointing that out! [17:49:26] No problem! [17:51:59] fabriceflorin: changes are on test, could you notify Dario so he can verify everything's ok? [17:52:23] as well as kaldari/bsitu regarding the PT change [17:53:32] mlitn, are you going to do the E3 deploy? It's HEAD of master, '58f36f0 Merge "track campaign (for AFT Signup CTA)"' [17:53:41] hi fabriceflorin, ping me if I can help with anything [17:54:23] spagewmf: sure [17:55:32] mlitn: it looks good, thx [17:56:06] mlitn, spagewmf: Thanks! I just told Dartar that the AFT5 changes are up on test. He's going to check on this right now. [17:57:22] spagewmf: e3 master is now on test [17:58:13] chrismcmahon: Thanks for the offer. This is mostly a metrics deployment, so there may not be much to test on the AFT5 front. And the Page Curation bug fix is also something we can test on our end. But we'll let you know if we need anything on your end. Cheers. [18:00:08] mlitn, fabriceflorin, so how do I get Called To Action!? http://test.wikipedia.org/wiki/Sandbox seems to have the old AFT v 4 feedback. [18:01:18] spagewmf: Try this perennial article on test: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow [18:02:55] spage: When I click on the CTA4 call to action, I get directed to this version of Account Creation: http://test.wikipedia.org/w/index.php?title=Special%3AUserLogin&returnto=Golden-crowned_Sparrow&type=signup&campaign=aftv5_cta4&c=250 [18:04:19] spagewmf: FYI, the URL above has one more parameter besides the campaign code, that one is for the number of the feedback post (c=250). [18:05:05] fabriceflorin, looks good. Please try guessing the CAPTCHA wrong [18:05:10] fabriceflorin, spagewmf, mlitn: ready to test and check log data we're collecting [18:05:17] you should still get the blue button [18:07:21] spagewmf: message should be "Login error [18:07:21] Incorrect or missing confirmation code." upon guessing CAPTCHA wrong? [18:07:41] spagewmf: just tested, I remain bucketed in ACUX when a server error is triggered [18:09:00] spagewmf: I tested the wrong captcha as you requested and got this error message, as you had expected: "Login error [18:09:05] Incorrect or missing confirmation code." http://test.wikipedia.org/w/index.php?title=Special:UserLogin&action=submitlogin&type=signup&returnto=Golden-crowned_Sparrow&campaign=aftv5_cta4 [18:09:39] chrismcmahon that sounds right. ACUX just takes the output of the current Userlogin?type=signup form and munges it, so interactions like that should be unaffected. [18:10:20] fabriceflorin good, on redisplay the campaign is in the URL and Dario should continue to get the campaign in events [18:10:49] funnel completed, pulling data from emery [18:10:52] spagewmf: OK. that is slightly funky because the CAPTCHA is actually labeled "Security Check" not "confirmation code", but small potatoes I guess. [18:11:04] spagewmf: And if I answer the captcha correctly, I get a "login successful" message, with this URL: http://test.wikipedia.org/w/index.php?title=Special:UserLogin&action=submitlogin&type=signup&returnto=Golden-crowned_Sparrow&campaign=aftv5_cta4 [18:11:21] spagewmf: sorry for late answer, I think you already gathered that http://test.wikipedia.org/wiki/Golden-crowned_Sparrow should get you AFTv5 [18:12:04] So it looks like we're good to go on the AFT5 front, unless I am missing something. mlitn , is there anything else we need to test for AFT5? Have we disabled the click tracking for CTA1? [18:14:07] fabriceflorin: yes, cta1 is disabled (you can't even manually trigger it with ?aftv5_cta=1 anymore) [18:14:33] so, DarTar, you're getting everything correctly, right? [18:14:40] not quite, I am mailing you a report [18:14:48] bear with me [18:14:56] alright [18:15:02] BTW, there's a prize if you can figure out why "Real name:" appears on signup on test wikis but not on en. I think it's a hidden configuration to hide the hidden realname pref. 8-) [18:17:30] spagewmf, fabriceflorin, mlitn: mail for you [18:20:17] spagewmf: "Real name:" does appear on enwiki [18:20:46] DarTar: "the CTA signup pitch event is missing" [18:20:50] what exactly do you mean with this? [18:21:01] the CTA impression [18:21:49] oh ok [18:21:53] as well as all other AFT-related events [18:21:57] check http://meta.wikimedia.org/wiki/Research:Article_feedback/Clicktracking#CTA1_update_.28October_2012.29 for comparison [18:22:20] AFT impression, AFT submit, CTA impressions: they are missing (if I look up that token) [18:24:50] bleh, sounds a lot like last week, before the problem was suddenly, mysteriously, solved [18:33:06] mlitn: could this be because of the weird setup of testwiki? [18:34:18] DarTar: that's what I'm guessing [18:34:26] point 2 of your mail should be fixed [18:34:47] point 1 is what we also had last week, but suddenly worked [18:35:04] k great - point 1 also includes all the upstream AFT events (not reported in my mail) [18:35:13] was that the same issue we had last week? [18:35:15] I forgot [18:35:27] and again, I'm getting this result: {"warnings":{"main":{"*":"Unrecognized parameter: 'time'"}}} [18:35:59] shite [18:37:16] I don't think anyone wrote a patch for that, the problem fixed itself when the code was pushed to enwiki, do I remember correctly? [18:39:21] the problem had fixed itself on test [18:42:02] * chrismcmahon twitches with impatience to get beta labs working for stuff like this [18:49:12] mlitn: what do you recommend at this stage? [18:50:21] DarTar: can you test again? [18:50:28] DarTar: let it fix itself [18:50:31] :D [18:50:31] :-D [18:50:58] bsitu: yeah that's always my preferred strategy [18:50:58] mlitn: sure [18:53:13] w00t [18:53:28] mlitn: I get this error message when trying to post feedback [18:53:40] "The anontoken is not 32 characters" [18:53:55] believe me it is 32 characters [18:54:17] 0G2GU6a3tYKDnl8Zht5IKKoKJigv85yP [19:02:51] weird thing, that error, can't reproduce it [19:04:35] k let me try again [19:05:02] oh that's interesting [19:06:46] I just noticed that despite having logged out and cleared all my cookies, Chrome was serving me a cached version of the page, which presumably means AFT was trying to read a mw.user.id() instead of a non-existent cookie [19:06:47] could that be the reason why this happens? [19:08:07] the sooner we fix the confusion created by mw.user.id() the better [19:08:23] anyway, starting with a clean browser again [19:09:38] mlitn: error occurred again, same pattern as above "The anontoken is not 32 characters" [19:09:48] how to reproduce on Chrome/Mac OS: [19:09:59] 1) clear all cookies [19:10:44] 2) hit a page with AFT enabled (Golden-Crowned Sparrow) and check if that sets the mediaWiki.user.id cookie with a token [19:11:45] 3) check if you're being served a fresh or cached version of the page, if it's the latter (indicating that I am still logged in), try posting feedback [19:12:30] you will get the above error message [19:12:42] oh, so you're logged out, but chrome servers you a page of you - supposedly - logged in, right? [19:12:57] and indeed mw.user.id() returns the logged in username, thereby failing to pass a valid token [19:13:01] indeed [19:13:02] yep [19:13:06] ok, i see the problem now [19:13:21] bah [19:14:04] I suggest removing the additional check for this in AFT [19:14:30] if we get served the username rather than an anon token, that's ok as well, I guess [19:17:35] marktraceur: Hey, I am back, and I am currently on MW 1.19.2, and its still not working [19:18:29] drecodeam: Could you try with latest git? :) [19:18:51] marktraceur: on it ! [19:19:00] Woooooo [19:19:59] mlitn: hmm, no, that's going to mess up with the log data [19:20:27] ACUX only relies on the cookie [19:20:29] so we have two options: [19:20:34] spagewmf: Dunno if you got the earlier ping--are you still not seeing "Real name:" on enwiki? [19:21:26] either we think that this issue with caching will only hit a small number of users (and since the CTA4 experiment is targeted to anons that's probably a good assumption) [19:21:42] or we rely on the cookie as opposed to calling user.id() [19:21:51] but that probably requires more work, doesn't it? [19:22:19] marktraceur were you logged when you saw Real name? We only munge the form to the new look for anonymous users. [19:22:46] well, user.id() fetches from cookie (if present) [19:24:00] spagewmf: I'm not sure it was the new form, but I saw real name [19:24:02] only, it appears that somehow chrome is still sending cached data [19:24:16] spagewmf: If there's a parameter I need to pass, what is it? [19:24:36] if would suggest option 1 because we need something really fast, and because it has been like this since forever and a day [19:25:01] and noone ever complained about not being able to submit feedback, so I'm guessing almost noone encounters the problem :) [19:25:47] are you getting all data already, btw? [19:26:23] well, I cannot submit feedback ;) [19:26:41] emptying the cache and restarting [19:28:16] ok, funnel completed, looking up the log [19:29:01] marktraceur, if you want to see the new appearance even though you're logged in, http://en.wikipedia.org/wiki/Special:UserLogin?type=signup&acux_bucket=acux_2&acux_eligible=1 [19:29:54] mlitn: *sigh* same problems as before, sending you a snippet from the log [19:31:05] check your inbox [19:32:17] mlitn: unless you need me in the next 10 mins I'm going to grab some food, brb [19:33:31] bon appetit [19:41:59] spagewmf: Confirmed, I see Real name: [19:43:15] marktraceur but the experiment is only for anonymous (logged out) users, and I believe on enwiki they don't ever see "Real name" on the account creation form. [19:43:35] spagewmf: I was logged out just now when I saw it [19:44:57] mlitn: What is your recommendation about what to do about today's AFT5 deployment? It seems like the main options are: 1) Fix the problem and deploy; 2) Deploy without a fix, assuming the problem will go away on production; 3) abort the deployment. Which do you recommend at this point? (If we select 2, on the assumption that the problem is with the test environment, we may need to revert if the problem doesn't go away on production, so that [19:44:57] may add a couple more hours to your schedule. ) [19:48:44] back [19:50:00] mlitn: My understanding from what DarTar told me is that this issue would only affect a small number of users in some very peripheral edge cases (logged in users who log out and click on the create account button, correct me if this is inaccurate). So it may be worth trying it out on production, if we feel confident that 1) the revision is very easy to revert if problems occur and 2) it's unlikely to break things for large numbers of users. [19:50:59] that applies to the token problem only, we still have the issue with the missing/ill-formed events [19:51:05] fabriceflorin: yes, the problen however is that DarTar is not receiving all tracking events [19:51:19] btw fabriceflorin just went out for a few mins [19:51:20] DarTar: issue #2 should be fixed: i had not yet pulled it on test [19:51:40] oh ok, I assumed it was already live [19:52:27] shall I try now? [19:53:50] please do :) [19:54:05] although no progress has been made on #1 [19:54:53] k [19:57:50] DarTar: could you try one more time? :) [19:57:54] mlitn: duplicate bucket fixed indeed, thanks [20:00:19] DarTar: any chance you could give it (yet) another spin? [20:00:34] done, same pattern [20:00:37] #1 still not fixed? [20:00:54] nope [20:01:06] funnel starts from cta click [20:01:19] well ehm, I guess we can push to production, hoping the problem is solved there, but if it isn't, we're not worse off than we are now.. [20:01:33] let me look up a few events before the submit, 1 sec [20:03:32] ok interesting, assuming I am the only one making tests, I see the missing events but with a different token [20:03:54] possible explanation, some events might still be using the CT token instead of the anon token? [20:04:01] if the token is OUgGqijjbPROSGucqX9cHRzbEXY2bHkxR, it's me [20:04:58] no, there's another token which presumably is me (we should test on different pages ;) ) [20:05:01] sending you a dump from the log [20:09:55] right... [20:10:16] CT's $.trackActionWithOptions() still uses $.cookie( 'clicktracking-session' ) [20:10:25] ahaaa [20:10:33] is that the version on testwiki only? [20:10:43] should be the same on enwiki, right? [20:11:32] yes [20:11:33] so either we change that to mw.user.id(), or revert AFT back to $.cookie(...) [20:11:57] I am confused, we didn't have this issue with the CTA1 test afaik [20:11:59] indeed [20:12:03] was about to say that :p [20:12:48] I'm going to talk to halfak in a minute about this data, but my QA suggested that everything was coming through as expected [20:14:03] we definitely want to use the token, not the CT session [20:14:10] as that's what ACUX is logging [20:14:45] indeed [20:15:08] give me a second though, doublechecking my statement as I'm not too familiar with CT ;) [20:16:00] sure [20:25:41] ok DarTar [20:25:47] any change you can review https://gerrit.wikimedia.org/r/#/c/29687/ [20:26:02] spagewmf: I see it without logging in. [20:26:07] I didn't want to touch the code in clicktracking, since who knows what else depends on it [20:26:25] so I'm just doing the clicktracking call from AFT, with the correct token [20:27:01] mlitn: bsitu is looking at your revision now. Do we need a second reviewer? [20:27:10] that said, I'm pretty sure last week's data will exhibit the same problem... [20:27:21] It seems CT's jquery.clickTracking.js trackAction() adds time to the tracked event, but CT's own API does not do anything with this parameter (it logs wfTimestampNow() ) , and thus some wikis warn "Unrecognized parameter: 'time'". The left hand does not know what the right hand is doing! [20:27:48] thx for the fix, the problem affecting last week is worrying - I'll have to review this with aaron [20:28:01] spagewmf: that's indeed another issue, though it seems that's not the one blocking us :) [20:28:14] should still be fixed though [20:28:44] fabriceflorin: if bsitu feels confident merging it, we don't really need some1 else [20:32:37] mlitn: aside from today's deployment, we will need to do a post-mortem for CTA1 to understand if we're missing a number of events which may affect the analysis (I really hope that's not the case) [20:33:24] well, all events should be there, but - looking at the current code - they will have a different token [20:34:27] DarTar: [20:34:28] if that's the case there should be no CTA impressions at all that we could match with CTA submits [20:34:57] mlitn: Cool. Bsitu thinks your revision looks good, and he asked kaldari to look at it as well. [20:34:58] I'd be surprised if that happened because halfak was looking at clickthorugh rates and had to rely on tokens [20:35:20] but I can check with him later today [20:35:34] I hope I'm wrong ;) [20:36:32] oh [20:36:32] just a second [20:37:02] dropping a line to aaron now [20:37:34] DarTar: page revision id is also saved in the logs [20:37:41] yes [20:37:48] that's as expected [20:38:18] what are the buggy events using the CT session? Anything up to the CTA impression? [20:38:47] yes, and that's good in case last week's tokens are screwed up [20:38:48] just a sec [20:39:02] sure [20:42:32] init, noedit-init, impression, submit_attempt, submit_success, submit_error_, ctaXYZ-impression [20:43:06] these use CT javascript, so will probably have another token [20:43:36] mlitn patch comment sent [20:43:39] ok, thanks - let me report this to aaron [20:44:25] now, about the revision id, given that we also know this value, it is semi-possible to match both token values, assuming that it's relatively unlikely that multiple people will be submitting feedback at the same page at the same timeframe [20:45:20] only those leaving feedback on the same revision at the same timeframe can't be matched with a 100% certainty, but if that happens, it does't really matter (i think) which start (CT cookie) we can tie to which end (mw.user.id) [20:49:10] spagewmf: re-commented ^^ [20:50:07] mlitn: good point, do you mind posting this in a reply to Aaron? I just sent a short report [20:50:14] scrap that, I'll do that [20:50:37] okay [20:52:22] done [20:52:25] mlitn yes the token in CT logs is its clicktrackingSession. Dario often ignores it. yes, put the actual anonymous user token (that mw.user.id() should return if the user is logged out) in the info along with your other stuff. [20:52:39] Hi [20:52:55] I think i may have discovered a bug with mw.loader.load [20:53:09] Edokter: yes? [20:53:34] It doesn't work :) [20:53:34] spagewmf: you're suggesting to still add it to "additional" rather than submit it as "token", right? [20:53:41] that's not very useful [20:53:55] at least not here: http://en.wikipedia.org/wiki/User:Edokter/MenuTabsToggle.js [20:53:58] Yes, that's what ACUX does, another | separated value. [20:54:03] spagewmf: no I don't ignore that ;) I'm aware of this, we add to use this solution in ACUX because of the problem with matching client and server events [20:54:10] I had to replace it with importScriptURI [20:54:24] but in the previous round of CTA1 that should not have been the case [20:54:47] otherwise, it will not work when loaded from *other* wikis [20:54:47] meaning that either using CT or the token is fine as long as they are consistent [20:55:16] in which case, the CSS fails to load, or loads too late [20:55:34] Edokter: The fade.png background with on ".. li" is causing each menu item in the dropdown menu to have a fade [20:55:36] ok, in that case, DarTar is it ok if I submit _all_ events with the CT cookie as token, and add mw.user.id into the |-seperated additional values, for _all_ events? [20:55:38] good grief, hope all these typos are due to Mountain Lion's automatic spellchecking [20:55:44] Edokter: that looks broken, was that intentional? [20:56:13] mlitn: sure, if that's the fastest way to get this, it's fine with me [20:56:13] KRinkle, not... that is quite intentional [20:56:17] Edokter: k.. [20:56:18] I like grandients :) [20:56:57] Krinkle, the problem is the CSS not loading or loading too late when using mw.loader.load [20:57:12] Edokter: anyway, it is working fine here [20:57:12] making the arrows invisible [20:57:12] let me show you [20:57:31] Edokter: http://cl.ly/image/1f1C1m3z452E [20:57:47] (notice the gradients) [20:57:58] It's working probably because I reverted to using importStylesheetURI [20:58:01] No [20:58:06] Look at the screenshot [20:58:21] mlitn, possible code fix added [20:58:27] I'm not using that gadget [20:58:32] Oops... left out some info... fails in IE8 only [20:59:10] windows xp? [20:59:24] yes [21:00:02] But as mw.loader.load and importStylesheetURI do practically the same, why does the first fail? [21:00:17] lets see if it actually fails first [21:00:42] how can I determin that? [21:01:17] well, you say you already did [21:01:26] right? [21:01:26] now I'm trying to reproduce it [21:01:42] mlitn: talking to aaron [21:02:38] mlitn possible code fix in the review, I tested new $.aft.additional() and it seems to work. [21:03:02] ok, i put the importStylesheetURI back in [21:03:09] wait for caching [21:03:40] No need to [21:03:50] hey mlitn, I just talked to fabriceflorin, we're thinking it makes more sense to defer until the next deployment window on Thursday [21:03:50] I'm just executing the mw.loader.load line from the console [21:04:12] if your script is relevant, I'm not even starting to debug, cause that would mean a whole different kind of problem that I can't go into now. [21:04:26] spagewmf: indeed, i'm going to add that, but also change real 'token' value to the one currently set by CT, just to make sure we don't get too confused [21:04:34] it doens't seem to work indeed, though I do get a few signals [21:04:36] let me try something [21:04:49] I am reviewing the data with Aaron over the phone and if we come to the conclusion that CTA1 was affected too we may want to run that again instead of CTA4 [21:04:55] DarTar, do you want AFTv5 to always include an anon user token "Golden-crowned_Sparrow|147947|zJm7M8JgHzUtmsZWXA7EZKZb0cKFEwba" , or can mlitn leave it off or blank when the user isn't anonymous? [21:06:17] DarTar: okay, keep me posted [21:06:37] spagewmf, mlitn: at this stage we're calling it off [21:07:02] I'll send a report after talking to Aaron [21:07:03] fabriceflorin: how about the pagetriage fix, should I still deploy that one, or can it be delayed to thursday? [21:07:07] mlitn I don't understand. If you call trackActionWithIfo(), it'll supply CT's token for you. [21:08:22] the stylesheet *is* added as last child to [21:08:31] spagewmf: indeed, but not everything uses trackActionWithInfo(), there's some that uses this too: [21:08:32] $.trackActionURL = function( url, id ) { [21:08:33] return mw.config.get( 'wgScriptPath' ) + '/api.php?' + $.param( { [21:08:34] (according to IE8 console) [21:08:34] 'action': 'clicktracking', [21:08:36] 'format' : 'json', [21:08:37] 'eventid': id, [21:08:42] 'namespacenumber': mw.config.get( 'wgNamespaceNumber' ), [21:08:42] 'token': $.cookie( 'clicktracking-session' ), [21:08:42] 'redirectto': url [21:08:43] } ); [21:08:43] }; [21:08:56] (which is a real link that registered with CT and then redirects to the actual link) [21:09:20] (paste paste http://dpaste.org/ paste :) ) [21:09:29] Edokter: I actually don't even get that [21:09:35] Edokter: its as if nothing happens [21:09:41] mlitn: We should deploy the PageTriage fix, if you don't mind -- or Kaldari and Benny can do this on this end, if you are getting tired. Thanks for your good work today! [21:10:18] Krinkle: you don't get my comment, or why it's not working? [21:10:31] Edokter: I don't even get a in the [21:10:38] (in WinXP/IE8) [21:11:14] ah, it only appears after refreshing the DOM (not the page) [21:11:41] Hm.. it seems the debugger in IE is kinda misleading [21:12:05] I just did $('head').empty() and I see it happening (all stylesheets gone), but the dom looks the same still [21:12:40] okay, so now I see it in the dom [21:12:48] fabriceflorin: it's ok, I'll push PageTriage out, will need to undo staged AFT stuff first though [21:12:49] it doesn't always update properly after manipulating the DOM [21:13:12] could this be a timing issue> [21:13:50] ori-l: regarding postedit, I'm about to submit a change for you, includes mw.msg fix [21:14:22] Krinkle: yay! should I abandon 29699 [21:14:24] ? [21:32:00] Krinkle: shall I make a bug report? [21:32:23] Edokter: yes, please do. Feel free to assign to me. I think I have a fix, but a reminder doesn't hurt :) [21:32:33] OK [21:40:23] ^demon: is the wmf-deployment group in gerrit meant to match the mortals group in admins.pp? [21:40:53] <^demon> More or less. wmf-deployment is "People who deploy mediawiki stuff to the cluster" [21:41:18] Krinkle: bug filed: Bug:41331 [21:41:51] ^demon: what other things would people deploy? [21:42:34] <^demon> Um, nothing really. I didn't know if mortals was more inclusive than "fenari deployers" or not. [21:42:51] andre__: Can you make some updates to the Echo Bugzilla component for me? [21:43:06] Haha, okay. It just seemed a little strange that they differed. [21:43:28] kaldari: What do you need? :) [21:43:43] <^demon> pgehres: If anyone in mortals isn't in wmf-deployment and needs to be, we can easily add them :) [21:44:25] We probably want to take brion off and add bsitu@wikimedia.org as default recipients for the component (unless brion says he wants to keep getting notices for it) [21:44:40] ^demon: I am not sure. The missing people I found were me, K 4-713, c steipp, and Rob La [21:44:42] kaldari, if you drop "probably" I can do that, sure [21:44:55] well, definitely add bsitu [21:44:59] nah i don't need to be on it constantly [21:45:09] but i get lots of bug mail so i won't notice if i stay ;) [21:45:16] okay, fixed. [21:45:19] <^demon> brion: I saw you got your shell access back, btw ;-) [21:45:24] remove brion, added bsitu. [21:45:25] ssssh :) [21:45:29] <^demon> How long did you manage to avoid that? [21:45:31] andre_: thx [21:45:38] np [21:45:38] and can you confirm that I'm already listed? [21:45:43] ^demon: I guess add brion to that list too :-) [21:45:54] as a default bug recipient for Echo? [21:46:07] so far i've only used it to upload mobile app files :) I need to be retrained on deployment at some point i suspect ;) [21:46:21] * pgehres sees a wall of :-) [21:46:35] :P [21:46:41] <^demon> brion: git pull, scap. Hasn't changed too much. [21:46:49] <^demon> Other than het deploy. [21:47:08] ah het deploy. fun! [21:48:06] andre__: Can you make sure I'm listed as a default recipient as well, I'm not sure if I am currently or not. [21:48:36] <^demon> pgehres: Shoot me an e-mail with all those names and I'll poke it later tonight. [21:48:45] <^demon> It's dinnertime for me :) [21:48:51] ^demon: enjoy your food :-) [21:48:59] off to sleep, g'night [21:50:12] kaldari: you were not in default CC for Echo, now you are. If we ever upgrade to Bugzilla 4.4 in a few years you will be able to find out yourself :) [21:50:33] can't wait [21:50:47] well, actually I guess I can [21:51:00] but it will be hard [21:51:30] tell me! [22:26:29] Haven't seen that error before: [22:26:37] Notice: Did not find alias for special page 'Notifications'. Perhaps no aliases are defined for it? [Called from SpecialPageFactory::getLocalNameFor in /var/www/MediaWiki/Git/core/includes/SpecialPageFactory.php at line 576] in /var/www/MediaWiki/Git/core/includes/debug/Debug.php on line 283 [22:29:13] Krenair: That's new to me, too! What did you do? :) [22:29:14] echo? [22:29:21] Reedy, yes [22:30:00] $specialPageAliases['en'] = array( [22:30:00] 'Notifications' => array( 'Notifications' ), [22:30:00] ); [22:30:09] create an alias file and add that [22:31:30] and load it ;) [22:36:36] Reedy, shall I upload that to Gerrit then? [22:36:49] Yeah, I would ; [22:36:51] ;)