[00:00:19] <preilly>	 TimStarling: this is so annoying https://github.com/stefanesser/suhosin/issues/14
[00:01:29] <TimStarling>	 yeah, it doesn't seem to be a high priority for him
[00:03:15] <preilly>	 TimStarling: plus he isn't the easiest person to work with
[00:03:49] <TimStarling>	 I've heard that
[00:07:37] <TimStarling>	 I may need a small Zend engine patch: a macro which disables inline assembly
[00:07:43] <TimStarling>	 do you have commit access to it, preilly?
[00:08:38] <preilly>	 TimStarling: I should
[00:08:56] <preilly>	 TimStarling: if I don't I can add myself
[00:09:58] <TimStarling>	 LLVM has a new JIT called MCJIT which supports inline assembly, and I wasted a lot of time trying to switch to it before discovering that so many features are unimplemented that it's not possible
[00:10:40] <TimStarling>	 so it has to use the plain C versions of addition etc. somehow
[00:17:20] <preilly>	 lame
[00:18:06] <TimStarling>	 woot
[00:18:20] <TimStarling>	 http://paste.tstarling.com/p/IasJwI.html
[00:18:52] <TimStarling>	 just got it working
[00:20:36] <TimStarling>	 and with an 8% time reduction before I even look at the optimisation settings or anything
[00:20:43] <preilly>	 TimStarling: why does just one test  use no php.ini file?
[00:21:01] <preilly>	 TimStarling: just to avoid optimisation settings?
[00:21:02] <TimStarling>	 because I had the default php.ini set up to use LLVM to save typing
[00:21:13] <TimStarling>	 with -n it shows you plain Zend
[00:21:47] <TimStarling>	 I didn't use the "time" shell command because it would have been unkind
[00:21:53] <TimStarling>	 it takes a while to start up the JIT
[00:22:12] <TimStarling>	 I just used microtime(true)
[00:24:33] <preilly>	 TimStarling: and this is with no caching correct?
[00:25:07] <TimStarling>	 right, but like I say, I omitted the part that would be cached from the benchmark, by waiting until the oparray is compiled before starting the timer
[00:25:42] <preilly>	 TimStarling: got it
[07:34:12] <hashar>	 hello
[07:39:36] <hashar>	 TimStarling: good evening :-]
[07:39:56] <hashar>	 TimStarling: is that PHP llvm you talked about something which is actively developed ?
[07:40:25] <TimStarling>	 no
[07:40:31] <TimStarling>	 it is actively developed since last friday
[07:40:36] <TimStarling>	 which is when I started actively developing it
[07:43:12] <TimStarling>	 but it's so small
[07:43:58] <TimStarling>	 it's not something that needs active development, just a day of maintenance a year
[07:44:11] <TimStarling>	 and it gives a 12% performance improvement on a tight loop benchmark
[07:44:22] <hashar>	 compared to PHP 5.4 ?
[07:45:28] <hashar>	 apparently brion took a peak at it too :-) http://gitorious.org/php-llvm
[07:46:45] <hashar>	 TimStarling: just wondering but do you get any contacts with the Facebook developers of hiphop
[07:47:01] <hashar>	 they were supposedly writing some kind of virtual machine ala Java
[07:49:45] <TimStarling>	 yes, hhvm was an inspiration
[07:53:29] <hashar>	 there are some recent push https://github.com/facebook/hiphop-php/commits/vm
[07:53:30] <hashar>	 nothing fancy though :D
[07:58:19] <Platonides>	 oh, so you started doing it
[07:58:39] <Platonides>	 I may need to change my last-night appreciation
[08:00:28] <Platonides>	 hehe, you're taking over
[08:00:40] <Platonides>	 one goes when the other arrives
[08:40:24] <hashar>	 we have soooo many changes opened
[08:48:34] <Nikerabbit>	 hashar: oh really?
[08:53:28] <hashar>	 ahhh
[08:53:38] <hashar>	 someone send some code to have groups in what lists!!!!!
[08:59:44] <Nikerabbit>	 ?
[09:22:32] <hashar>	 sent
[09:22:39] <hashar>	 Nikerabbit: that is a google summer of code project :-]
[09:22:46] <hashar>	 Nikerabbit:  https://gerrit.wikimedia.org/r/16419
[09:22:59] <hashar>	 Nikerabbit: will apparently let you manage several whatlists
[09:23:04] <hashar>	 by grouping articles
[09:37:40] <TimStarling>	 Platonides: I was doing it before last night
[10:03:50] <gerrit-wm>	 Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/16113
[14:45:58] <marktraceur>	 hashar: You're awesome! Thanks so much for the review, I hadn't done much testing before. I'll check them again in an hour or so, make sure things are working somewhat
[14:46:14] <hashar>	 marktraceur: :)))
[14:46:29] <hashar>	 marktraceur: it can still be greatly improved but they are overall fine
[14:46:42] <marktraceur>	 And yeah, I realized this morning that I probably broke things with the rebase
[14:47:00] <hashar>	 marktraceur: I thought it would be better for you to just send follow up instead of amending the existing one :-D
[14:47:11] <hashar>	 I did a rebase this morning too
[14:47:15] <hashar>	 might have broken yet more stuff
[14:47:21] <hashar>	 so you will need to run the test suite again :-(
[14:47:39] <marktraceur>	 It's fine, this one only takes a few seconds in all
[17:07:58] <sumanah>	 hi rmoen, how's your day going?
[17:08:28] <sumanah>	 (I promised my partner that we'd go out to lunch together, so I'm hoping to check in with 20%-related stuff a little early if possible)
[17:08:58] <rmoen>	 sumanah: good.  That's fine by me :)
[17:09:12] <sumanah>	 rmoen: if you already have 20%-related plans, I won't interfere
[17:09:45] <rmoen>	 sumanah: I was planning on wrapping up some ve commits as I was out sick yesterday and then jumping into some code review
[17:09:56] <rmoen>	 sumanah: question, do ve bugfixes count ?
[17:09:58] <sumanah>	 rmoen: makes sense to me. hope you're feeling better
[17:10:04] <rmoen>	 sumanah: lol
[17:10:29] <rmoen>	 sumanah: thanks, I am doing much better.
[17:10:44] <sumanah>	 rmoen: looking at https://www.mediawiki.org/wiki/Wikimedia_engineering_20%25_policy#Scope ... unfortunately, no, I don't think VE bugfixes count yet; they will once it's rolled out to more places, though
[17:11:06] <sumanah>	 robla can correct me re the scope of the community service stuff.
[17:11:40] <rmoen>	 sumanah: i figured.  That's ok by me :)
[17:11:48] <sumanah>	 :)
[17:12:13] <sumanah>	 rmoen: got some code review already queued up? you are probably already set on what is in Gerrit awaiting review
[17:13:00] <rmoen>	 sumanah: I'm all set.  Thanks
[17:13:06] <sumanah>	 I know Trevor is also set
[17:13:11] <sumanah>	 not seeing Brion online yet
[17:13:22] <sumanah>	 jdlrobson: hi, can I check in with you briefly re your 20% plans for the day?
[17:13:38] <jdlrobson>	 sumanah: my 20% day is friday so haven't thought about it yet!
[17:13:54] <sumanah>	 jdlrobson: it moved, argh, changing the wiki page now
[17:14:08] <jdlrobson>	 thanks!
[17:14:46] <sumanah>	 subbu: remind me, did you pick a day?  https://www.mediawiki.org/wiki/Wikimedia_engineering_20%25_policy#IRC_checkins
[17:15:14] <sumanah>	 rmoen: is Roan around?
[17:15:37] <subbu>	 sumanah, i picked thursday.
[17:16:02] <subbu>	 i thought i added it to the page.  let me check
[17:16:17] <sumanah>	 thanks subbu - I added it to an additional part of the page :)
[17:16:55] <subbu>	 ah, i only added it to 'scheduling' and not to irc checkin.
[17:17:28] <sumanah>	 At some point someone should overhaul that page, it's out of date. I don't have time right now and ought to do that + updating the [[Developers]] page at some point
[17:17:29] <subbu>	 thanks, but, maybe these are two different things .. are they?
[17:19:40] <sumanah>	 subbu: they ought to be in sync and sadly often are not
[17:19:48] <subbu>	 ok,
[17:20:00] <sumanah>	 thanks
[17:22:31] <sumanah>	 hi TrevorParscal, happy codereviewing today
[17:23:06] <sumanah>	 rmoen: thought you might be interested to know that Ryan Faulkner & bawolff are progressing re SubpageSortKey https://www.mediawiki.org/wiki/Special:Code/MediaWiki/95382%26path%3D#c32923
[17:23:39] <sumanah>	 TrevorParscal: thanks for the review on OnlineStatusBar; looks like we're waiting for petan to fix a few issues you & Timo found
[17:24:37] <TrevorParscal>	 nice
[17:27:10] <sumanah>	 kaldari: do you happen to know what's going on with SignupAPI?  (Leaving a note in the bug as well)
[17:27:23] <sumanah>	 TrevorParscal: am I right in presuming your code review queue is all full for today?
[17:28:01] <TrevorParscal>	 i have code review plus some documentation and gsoc stuff to catch up on
[17:28:09] <kaldari>	 E3 is supposed to take it over one of these days
[17:28:10] <TrevorParscal>	 what did you have in mind?
[17:28:38] <sumanah>	 Yay GSoC & doc, TrevorParscal - sounds good to me :-) I'm sure I've requested your eyes on a changeset or two in the last week but I can't recall.
[17:28:52] <TrevorParscal>	 ok
[17:28:57] <sumanah>	 TrevorParscal: that's all, just checking whether you wanted any input
[17:29:19] <thekaryn>	 TrevorParscal - yes, e3 is going to take over signup api. i think it's going to go to S
[17:30:27] <sumanah>	 Ah, as swalling just said, "still in the work queue for E3, but since we're in the middle of
[17:30:27] <sumanah>	 testing/deploying a different experiment it's on the backburner."
[17:30:27] <sumanah>	 ok, thanks all!
[17:30:28] <sumanah>	 and I presume brion and Roan have their hands full today as well
[17:30:29] <sumanah>	 off to lunch I go
[17:31:41] <sumanah>	 ah, thanks thekaryn - just updated https://www.mediawiki.org/wiki/Review_queue#Extensions with that.
[18:31:28] <matthiasmullie_>	 fabriceflorin: chrismcmahon: everything's been pushed to testwiki, let's start testing
[18:32:21] <chrismcmahon>	 hi matthiasmullie_ thanks
[18:33:21] <fabriceflorin>	 Thanks, matthiasmullie_ ! I'm on it. For the record, here are the new revisions we plan to deploy today (sorry to other IRC participants for this long list, but it is really helpful for us to have these links handy in this conversation).
[18:33:26] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/15521/ - Bug 37822 - Text size of legal text from feedback form
[18:33:27] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/15796/ - Article Feedback Tool - Internet Explorer 7 Display Issues
[18:33:28] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/15798/ - allow posting feedback without a comment for both positive & negative vote
[18:33:30] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/15708/ - fix patchwork messages
[18:33:31] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/16133/ - Replace x with × for close icon
[18:33:33] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/16425/ - Bug 38499 - Problem with hidden feedback overlay
[18:33:34] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/16277/ - Only show the link to the central feedback page to editors
[18:33:36] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/15877/ - don't display talk page link to anons
[18:33:37] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/15875/ - up version number
[18:33:39] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/16032/ - don't display CTA2 to editors
[18:33:40] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/16410/ - Add a 'Refresh page' button on the feedback page
[18:33:42] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/16521/ - lower AFTv4 odds to display AFTv5 at 10% (inverse odds)
[18:33:43] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/16301/ - Fix issue where it was possible to flag feedback multiple times
[18:33:45] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/16033/ - second link on contribs should go to special page with post on #top
[18:33:46] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/16030/ - add class to AFTv5 entries in mycontribs
[18:33:47] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/15876/ - Show 'All Comments' filter by default for editors
[18:33:48] <fabriceflorin>	 https://gerrit.wikimedia.org/r/#/c/15718/ - Bug 37631 - Avoid modal behavior in "Improve this page" link at the toolbar
[18:35:59] <marktraceur>	 fabriceflorin: Pastebin or etherpad?
[18:36:29] <marktraceur>	 Like http://etherpad.wikimedia.org/NewDeploysToday for example
[18:37:18] <shepazu>	 how much control do we have over the page output? the body content is not really what we'd like, and uses tables for layout too much… could we do something with a skin, or an extension?
[18:37:39] <fabriceflorin>	 markraceur: Good idea. I will do this next time. I figured I would do this now, since the IRC chat room seemed completely quiet. But your point is well taken, and your suggestion much appreciated.
[18:38:12] <kaldari>	 bsitu: https://www.mediawiki.org/wiki/Page_Triage/Templates#Multiple_issues
[18:38:35] <bsitu>	 cool, thx
[18:38:47] <kaldari>	 bsitu: although someone still needs to make them into templates
[18:40:20] <chrismcmahon>	 matthiasmullie_ fabriceflorin IE7 can now see the Central page (and it looks fine) but AFTv5 feedback form is not appearing for IE7 at the bottom of the Sparrow page
[18:41:15] <fabriceflorin>	 chrismcmahon: Thanks! Is this on Windows 7 as well? or only Vista or XP?
[18:42:10] <chrismcmahon>	 fabriceflorin: this is on Vista on my personal VM
[18:42:44] <chrismcmahon>	 I can look on crossbrowser...
[18:44:42] <fabriceflorin>	 matthiasmullie_: I am no longer seeing CTA2 for editors, but I am seeing CTA1 as an auto-confirmed editor, which shouldn't be happening. Not a show-stopper for this deployment, but wanted to point this out. Editors should only see CTA5. So perhaps we need to revisit the CTA bucketing to make it based on user rights, rather than percentages?
[18:47:13] <matthiasmullie_>	 fabriceflorin: i'll add in that check for CTA1 as well by next deployment
[18:47:56] <matthiasmullie_>	 don't think we need to drift away from percentages though, as I can exclude CTA's based on certain conditions (like the permissions)
[18:48:53] <matthiasmullie_>	 so we can still use %-based bucketing, though correcting false bucketing (like based on permissions, or on whether the page can be edited)
[18:49:05] <chrismcmahon>	 fabriceflorin: matthiasmullie_ somewhat bizarre, IE7s on crossbrowsertesting are OK, but the IE7 on my VMWare VM does not see AFTv5 on the Sparrow page.  I'm willing to call it a local fluke issue on my VM
[18:50:18] <fabriceflorin>	 matthiasmullie_: Thanks! I also note that if you hit helpful/unhelpful a whole bunch of times, you can game your score up by a couple points, as shown for this post: http://test.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow/180 -- it should only have a relevance of +1, but I managed to give myself a +3 by clicking on helpful/unhelpful very fast a dozen times. Not a huge deal, but worth noting for next time.
[18:53:03] <matthiasmullie_>	 chrismcmahon: that's odd, as I'm not seeing it on my VM either
[18:53:31] <matthiasmullie_>	 if you got to see it though crossbrowsertesting, I'm gone blame some caching though
[18:53:44] <matthiasmullie_>	 (not sure)
[18:54:05] <chrismcmahon>	 matthiasmullie_: that sounds reasonable
[18:54:13] <matthiasmullie_>	 fabriceflorin: I'll look into the numbers by next deployment as well
[18:57:06] <chrismcmahon>	 ha. matthiasmullie_ AFTv5 just started appearing for my IE7, so server caching was the culprit
[18:57:37] <fabriceflorin>	 chrismcmahon: I just tested IE7 on XP using crossbrowsertesting.com and everything is looking good for the feedback form and feedback page. No link appears on the talk page, but I believe that is because I am anonymous, as intended. Would you like me to test on Vista as well, to see if I can reproduce your issue with crossbrowsertesting?
[18:57:42] <matthiasmullie_>	 ok great :)
[18:58:07] <chrismcmahon>	 fabriceflorin: I think we're good ^^
[18:58:32] <matthiasmullie_>	 so do I - didn't encounter any problem with the new code
[19:02:11] <fabriceflorin>	 matthiasmullie_: No problems on my end either. I just checked that we are correctly displaying articles on 'Parameters' and 'Planets_of_doom' for the 10% lottery as well, as recommended by you in this Bugzilla ticket: https://bugzilla.wikimedia.org/show_bug.cgi?id=37616 -- so I think we are good to go to production, unless Chrismcmahon has any issues.
[19:10:22] <chrismcmahon>	 fabriceflorin matthiasmullie_ could you do a sanity check?  the Refresh button at bottom of central page is OK in Chrome and IE7 but seems not to be refreshing in Firefox
[19:12:17] <chrismcmahon>	 fabriceflorin matthiasmullie_  I may be mistaken on how Refresh should work though
[19:12:29] <marktraceur>	 hashar: Thanks again for the help with the test suite, it's getting better all the time. Adding the new API calls and everything, so it should be complete once more in an hour or so. I'm confident enough in the code that I can +2 it, methinks.
[19:14:01] <chrismcmahon>	 actually fabriceflorin matthiasmullie_ scratch that, no refresh happens if no new feedback exists, my mistake
[19:15:21] <matthiasmullie_>	 chrismcmahon: it doesn't refresh the page, it just re-fetches the data in the list (so when new feedback has been added, it will be visible in the list after hitting refresh list, without having to refresh the page)
[19:16:28] <fabriceflorin>	 chrismcmahon: I could not reproduce the issue either. It works fine for me on FF, after posting feedback in a separate tab. The one thing that can be confusing is that your previous post appears on top, even if it wasn't the last one, but I think this is a small use case that mostly affects testers, not real users.
[19:20:03] <hashar>	 marktraceur: good to know :-]]
[19:20:15] <hashar>	 marktraceur: you did a nice job there, testing is far from being an easy task :D
[19:23:47] <marktraceur>	 hashar: Agreed, but something that's very helpful
[19:24:02] <marktraceur>	 I imagine it will make stupid mistakes a lot less common :)
[19:24:17] <hashar>	 marktraceur: definitely!!
[19:24:33] <hashar>	 marktraceur: meanwhile, do you have anyone to review all the changes you made to UploadWizard? (I am not volunteering btw)
[19:24:45] * werdna  reviews hashar 

[19:24:55] <werdna>	 speaking of reviews, hashar, you need to revisit my pretty timestamps rev
[19:25:13] <marktraceur>	 hashar: kaldari and drecodeam (GSoC student) are the only two who have worked on it in a sustained fashion, but it's definitely been a while
[19:25:22] <hashar>	 werdna: the human friendly time formatting ? ;)  Have you looked at DateInterval ?
[19:25:30] <werdna>	 hashar: did you read my comments
[19:25:36] <werdna>	 you have completely misunderstood how it works :)
[19:25:42] <hashar>	 marktraceur: I am confident you will find out someone to review all those changes!
[19:25:50] <hashar>	 werdna: haven't read them yet no
[19:25:57] <marktraceur>	 hashar: It's almost all JavaScript, so my choices for potential reviewers are very limited. Krinkle did some review on EtherEditor's frontend, maybe he could help....else, I don't know
[19:26:08] <werdna>	 hashar: basically, you seem to think it is entirely about relative timestamps
[19:26:17] <werdna>	 but relative timestamps are only used for the last 90 minutes
[19:26:18] <marktraceur>	 I've been trying to catch frontend changes by others and review them where possible, trying to lighten the load
[19:26:51] <hashar>	 marktraceur: aren't you in SF? if so maybe you can organize a code review sprint in the office :-D
[19:27:36] <werdna>	 after that it does the time (if it is the same day as now, according to the user's timezone), then "Yesterday at [time]", if it was yesterday, according to the user's time, then "[weekday] at [time]" if it was in the last five days, then "[day] [month]", then "[full date]
[19:28:53] <chrismcmahon>	 fabriceflorin: matthiasmullie_ I think the only thing I haven't looked at is the hidden overlay, I don't have the right permissions on testwiki, but Matthias and I checked it a bit on prototype
[19:29:29] <matthiasmullie_>	 chrismcmahon: as far as I can tell, the overlay is functioning just fine
[19:30:00] <matthiasmullie_>	 I'm just not sure if it'll fix the problem for those it was broken
[19:32:08] <marktraceur>	 hashar: That would be pretty awesome. I think Eloquence had mentioned something similar to that in an email
[19:32:39] <fabriceflorin>	 matthiasmullie_: Yes, I think this is primarily an issue for fluffernutter, who I'm told has an older system. I can test this on production, where I have temporary oversighter rights. Are we pushing to production now?
[19:33:08] <matthiasmullie_>	 fabriceflorin: yes we are
[19:33:25] <fabriceflorin>	 Cool, thanks!
[19:33:35] <chrismcmahon>	 fabriceflorin: there is at least one and possibly two other people with fluffernutter's issue
[19:33:57] <chrismcmahon>	 I've had an email exchange with one of them.
[19:34:40] <fabriceflorin>	 chrismcmahon: Do you know which exact configurations seem to be causing the problem? I can test these as oversighter on production, using crossbrowsertesting.com
[19:37:05] <chrismcmahon>	 fabriceflorin: I have been unable to reproduce the issue even with identical js/gadgets to fluff, so it doesn't seem to be a configuration thing.  I have isolated it to that one bit of js that is not showing up.  Matthias changed how that blue overlay is generated, we're hoping that the change either a) fixes the problem or b) provides more information for us
[19:39:07] <fabriceflorin>	 chrismcmahon: Thanks for the detailed clarification. I will test with my current browsers and let you know if I can reproduce the issue. Is the 'blue overlay' the same thing as the 'gray mask' that obscures the contents of a hidden post? Or is it something different?
[19:39:54] <chrismcmahon>	 fabriceflorin: yes, it's the mask with the button "View content" that does not appear for fluff
[19:40:27] <fabriceflorin>	 Thanks. Got it.
[19:47:12] <matthiasmullie_>	 fabriceflorin: chrismcmahon: the changes have been pushed to production - let's test
[19:51:09] <fabriceflorin>	 Sounds good. I'm on it!
[19:51:13] <chrismcmahon>	 matthiasmullie_ fabriceflorin I have an email draft for the fluff and the other oversighter, I'll get that out too
[19:52:54] <fabriceflorin>	 chrismcmahon: Thanks for contacting fluff and the other user directly, good idea!
[19:59:17] <fabriceflorin>	 I was able to reproduce Fluff's issue on production, using Safari. I hid something, then clicked on Details, and what appears to have happened is that I got logged out in the process, so that I could no longer access any of my user rights. All it said was "This post #183,520 was hidden by Fabrice Florin 3 minutes ago. http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5/Golden-crowned_Sparrow/183520
[20:01:00] <fabriceflorin>	 I will say, for the record, that the Wikipedia site is constantly logging me out at the strangest times, so I do not know if this is related to AFT or to the site itself. Still, it is really strange that as soon as I log in again, then try to go to the details page linked above, it immediately logs me out again.
[20:02:10] <matthiasmullie_>	 that's weird indeed
[20:02:20] <chrismcmahon>	 fabriceflorin:  fluff actually sees something slightly different, the text of the hidden posts are visible for her,  but can't be manipulated
[20:03:23] <matthiasmullie_>	 chrismcmahon: just a new reply
[20:03:36] <matthiasmullie_>	 doesn't appear to have fixed the problem
[20:04:45] <matthiasmullie_>	 now narrows it down to pages though - central feedback page seems fine
[20:06:21] <fabriceflorin>	 One little nitpick: the Help icon on the Activity log should go to the Help page, linking to the FAQ called 'How can I restore feedback that was wrongly hidden?' Right now it does nothing for me.
[20:07:22] <fabriceflorin>	 By Activitly log, I am referring to the flyout panel that pops up when you click on 'View activity' on the feedback page (not the central activity log).
[20:08:11] <chrismcmahon>	 yeah, I'm stumped, the page elements are exactly the same in both views
[20:09:28] <chrismcmahon>	 at least they're not blocked, just inconvenienced
[20:12:46] <fabriceflorin>	 chrismcmahon: That's good. As long as they have a recourse, that makes this less critical.
[20:14:28] <preilly>	 ^demon: gerrit is broken in chrome for viewing diffs on *.i18n.php files
[20:14:31] <matthiasmullie_>	 fabriceflorin: ok - got that; should the icon always link to the "how can I restore feedback that was wrongly hidden" faq, even if the feedback isn4t hidden?
[20:14:54] <preilly>	 ^demon: I get a, "Uncaught Class$eRc: Array index 82 out of range 1A8267964652EF1ADC8DB786BC4B3F40.cache.html:2319
[20:14:54] <preilly>	 (anonymous function)" error
[20:15:05] <^demon>	 preilly: Has it been reported upstream?
[20:15:43] <preilly>	 ^demon: there is a similar issue http://code.google.com/p/gerrit/issues/detail?id=1155
[20:15:51] <preilly>	 ^demon: but that is with CSS files
[20:16:23] <^demon>	 Ugh, another issue with code-prettifier.
[20:16:29] <^demon>	 This is the second time today I've cursed it.
[20:16:51] <^demon>	 preilly: Best advice I can give is complain upstream. I have zero clue how to fix this.
[20:18:28] <fabriceflorin>	 matthiasmullie_: Good point. If we are looking for a general case, it's probably better to take them to this question instead: 'How can I get more information about a post?' http://en.wikipedia.org/wiki/Wikipedia:Article_Feedback/Help/Monitors#How_can_I_get_more_information_about_a_post.3F
[20:20:40] <matthiasmullie_>	 ok, got it - will fix that for next release - ok?
[20:23:12] <fabriceflorin>	 matthiasmullie_: Yes, sounds good. I'm just noting all this on IRC, so I don't forget. Let me know if it's a problem to note all these small issues in this chat.
[20:24:07] <matthiasmullie_>	 it's not - I quicky mail them to myself and I'll sort them out tomorrow ;)
[20:24:24] <matthiasmullie_>	 if you ever catch me failing though, do let me know :p
[20:26:46] <fabriceflorin>	 Splendid. More importantly, I can confirm that we now have AFT5 correctly deployed on 10% of en-wiki, starting with the 2 articles you told us to check first: 1976_NFL_season and 1974_in_art. I also get about 1 in 12 articles when I keep clicking on the 'Random article' link.
[20:27:30] <matthiasmullie_>	 sounds about right :)
[20:45:10] <fabriceflorin>	 matthiasmullie_ chrismcmahon : I thoroughly tested this on IE7 and it all looks great, in both reader and editor mode. Nicely done!
[20:45:32] <chrismcmahon>	 fabriceflorin:  \o/ me too :)
[20:57:12] <fabriceflorin>	 matthiasmullie_ chrismcmahon: I have tested most of the features on production, and I think we are good to go for this deployment. Thanks again for a job well done!
[20:57:31] <chrismcmahon>	 I think so too, thanks fabriceflorin
[20:57:50] <matthiasmullie_>	 they appear to be working fine here as well :)
[20:57:51] <chrismcmahon>	 nothing on the back end seems to be blowing up either
[20:59:26] <TrevorParscal>	 Krinkle: dude, I don't have merge rights on WikiEditor?
[20:59:28] <fabriceflorin>	 Yes, so far, so good. Now that we are at 10%, we will ask Asher to do an analysis of our overall performance for AFT5, so we can determine whether or not we will need the sharding servers before we can deploy this tool to 100% in September.
[20:59:40] <TrevorParscal>	 given that I'm one of the primary authors, that seems like a mistake
[20:59:51] <Krinkle>	 TrevorParscal: ?
[20:59:53] <RoanKattouw>	 That's weird
[20:59:58] <RoanKattouw>	 Let me look at that in a minute
[20:59:59] <Krinkle>	 checking
[21:00:13] <TrevorParscal>	 I +2'd https://gerrit.wikimedia.org/r/#/c/9053/ and it still says review in progress
[21:00:38] <Krinkle>	 TrevorParscal: RoanKattouw It doesn't have its own access group. Only inherits from mediawiki/extensions/* admins
[21:00:52] <TrevorParscal>	 hmmm
[21:01:04] <TrevorParscal>	 I'm allowed to merge to core, but not any extension?
[21:01:09] <RoanKattouw>	 Wait you +2ed it
[21:01:12] <RoanKattouw>	 So you do have *some* rights
[21:01:29] <RoanKattouw>	 "Your change could not be merged due to a path conflict."
[21:02:00] <^demon>	 I really want to enable the feature that will tell you first if you can't merge.
[21:02:05] <^demon>	 Stupid freaking memory leak
[21:02:13] <TrevorParscal>	 lol - didn't see that
[21:02:18] <TrevorParscal>	 I never seem to see that
[21:02:26] <RoanKattouw>	 Well there's a bug in this version of gerrit
[21:02:36] <^demon>	 hrm? bug?
[21:02:39] <Krinkle>	 never mind the access group thing, that doesn't matter since we're both in the admins group
[21:02:49] <RoanKattouw>	 Previously, the comment saying "Trevor Parscal gave this a +2" came before the comment saying "Successfully merged" or "Couldn't merge"
[21:02:53] <RoanKattouw>	 But now, those are reversed
[21:03:03] <Krinkle>	 Either way, that are however both expanded by default
[21:03:08] <RoanKattouw>	 ^demon: Look at the order of the last two comments on https://gerrit.wikimedia.org/r/#/c/9053/ . This happens everywhere
[21:03:13] <RoanKattouw>	 Krinkle: Not for me
[21:03:26] <^demon>	 RoanKattouw: Yuck yuck. Reported upstream?
[21:04:13] <RoanKattouw>	 Not yet
[21:04:21] <RoanKattouw>	 Will file
[21:04:21] <^demon>	 Wonder if it's related to http://code.google.com/p/gerrit/issues/detail?id=1443.
[21:05:00] <^demon>	 Na, unlikely.
[21:05:12] <RoanKattouw>	 What's our current version of Gerrit? 2.4?
[21:05:24] <^demon>	 2.4.2
[21:10:33] <RoanKattouw>	 ^demon: https://code.google.com/p/gerrit/issues/detail?id=1490
[22:32:27] <marktraceur>	 drecodeam: New patchset? :)
[22:32:40] <drecodeam>	 ya
[22:33:23] <drecodeam>	 marktraceur: the messages API being changed, does that also change how we call it through Javascript ?
[22:40:52] <marktraceur>	 drecodeam: Messages API on Flickr's side?
[22:41:26] <drecodeam>	 no, the MediaWiki side
[22:51:07] <marktraceur>	 drecodeam: Oh, huh, I hadn't heard about that--is there a link explaining?
[22:56:26] <drecodeam>	 marktraceur: its the same thing you pointed me to : http://www.mediawiki.org/wiki/WfMessage#Comparison_with_the_deprecated_wfMsg.2A_functions
[22:56:42] <marktraceur>	 drecodeam: OHHHH
[22:56:51] <marktraceur>	 I'm sorry, I'm apparently a little slow today
[22:56:56] <marktraceur>	 Trying to write unit tests
[22:57:12] <marktraceur>	 drecodeam: No, IIRC, that only applies to PHP
[22:57:21] <drecodeam>	 marktraceur: hehe, thats fine. If you are
[22:57:41] <marktraceur>	 We could glance at the latest JS documentation and make sure UW is using proper l10n methods
[22:57:51] <drecodeam>	 sorry about that .. that was for the previous one
[22:58:06] <drecodeam>	 marktraceur: can you point me to the latest documentation ?
[22:58:27] <marktraceur>	 drecodeam: http://www.mediawiki.org/wiki/L10n#Using_messages_in_JavaScript
[22:58:39] <marktraceur>	 AFAIK it's all right the way it is
[22:58:56] <marktraceur>	 The one thing that might be messed up (my fault) is .html vs. .text
[23:00:18] <drecodeam>	 ya I guess everything else seems to be fine
[23:10:03] <marktraceur>	 drecodeam: Have you been working on untangling your commits, or is your current stance "eh, it's too much work, the status quo is fine" (which is perfectly valid)
[23:11:06] <drecodeam>	 marktraceur: what do you mean by untangling your commits ?
[23:14:33] <marktraceur>	 drecodeam: You had a bunch of dependencies, and you were modifying the last one rather than fixing the original problems
[23:15:37] <drecodeam>	 no marktraceur, I was working on the geo branch this weekend, I would be resolving the problems in the flickr branch now
[23:15:54] <marktraceur>	 Oh mmkay
[23:16:33] <marktraceur>	 drecodeam: I'm more wondering if you're intending to put the changes to your older commits into the older commits, or if you intended to continue changing the old one.
[23:17:30] <drecodeam>	 marktraceur: I was thinking of changing the older commits only, but what do you suggest I do ?
[23:23:12] <marktraceur>	 drecodeam: I would suggest untangling them, and changing the older commits, definitely
[23:23:21] <marktraceur>	 But it can be tough, sometimes
[23:23:38] <drecodeam>	 how do i go about untangling them ??
[23:24:02] <marktraceur>	 drecodeam: Basically you'll want to go to each patchset, start at the first one
[23:24:19] <drecodeam>	 marktraceur:  ya and then ?
[23:24:31] <marktraceur>	 Do `git review -d CHANGE-ID` where CHANGE-ID is the hash you see at the top of the page
[23:25:04] <marktraceur>	 Then make the changes kaldari and I suggested in various reviews, and do `git commit --amend`
[23:25:17] <RoanKattouw>	 Instead of the change-id you can also use the five-digit number in the URL
[23:25:20] <marktraceur>	 Once that's done, you can run `git review -R` to push the change again
[23:25:45] <marktraceur>	 RoanKattouw: ++, but I always found the change ID faster to copy and paste
[23:26:39] <drecodeam>	 marktraceur: ya thats what I was gonna do, but I thought you had something else in mind when you meant Untangling them
[23:28:33] <marktraceur>	 drecodeam: No, just a different way of saying it :)
[23:28:47] <drecodeam>	 marktraceur: cool, on it then :D
[23:50:47] <preilly>	 werdna: do we even need this file: https://gerrit.wikimedia.org/r/#/c/16580/2/xmpp/XMPPHP/XMPP_Old.php
[23:51:07] <werdna>	 preilly: probably not, it just came with the XMPPHP dist
[23:51:32] <werdna>	 I'm not convinced that either of the libs I'm using are the correct choice
[23:51:38] <werdna>	 but they seem to work for now
[23:52:14] <preilly>	 werdna: okay
[23:58:07] <preilly>	 werdna: I added another comment to https://gerrit.wikimedia.org/r/#/c/16580/2/xmpp/XMPP.php