[02:50:21] ori-l: got a moment? [02:59:38] sumanah: hey [03:00:15] ori-l: just wanted to let you know that, as a MW core person, you have the power to help approve or to disapprove requests at https://www.mediawiki.org/wiki/Git/Gerrit_project_ownership [03:04:39] yes, i know, and i saw the new requests, but it seemed a bit rude to step in so early. i figured i should lurk a while and get a sense for the considerations. what do you think? [03:06:08] ^ sumanah [03:07:40] ori-l: I can understand that urge [03:11:21] ori-l: no prob [03:11:39] ori-l: maybe next month [03:11:53] ori-l: you can also nominate people to be co-maintainers of particular repositories [03:15:10] sumanah: "guess we could be more liberal about giving people access and taking it away from them if they break certain yet-to-be-determined rules, but right now we don't have such a process and we don't have those rules spelled out very well." (Roan, https://www.mediawiki.org/wiki/Git/Gerrit_project_ownership/Archive) [03:15:16] yeah [03:15:32] nomination != final approval, though [03:15:46] you could also simply notice people and tell them they should self-nominate [03:16:12] happy to do that, obviously :) [03:16:30] great! thanks [03:18:47] ok, sleep [09:14:42] hashar: Have you met werdna? [09:35:49] Reedy: hi [09:35:54] Reedy: not sure I have met him [09:38:50] ahh found a pic of him on twitter : https://si0.twimg.com/profile_images/299952477/IMG_0917.jpg [09:39:02] Reedy: so no, I have never met him face to face [09:39:20] hahaa [10:37:46] I think I will write a RFC to rewrite mediawiki in python :-D [10:40:14] python is slow [10:40:28] let's rewrite it in forth [10:40:40] you know it has near-assembly performance [10:49:41] TimStarling: never heard of that language :-] [10:53:21] I've been reading about it over the past week, I thought it might be useful as a microcontroller language [10:53:32] but it turns out to be more suited to 16-bit than 8-bit [10:54:30] it was a kind of hard core competitor to BASIC in the 80s [10:55:15] reading some forth code, it really looks like some high level assembly [10:55:30] have you looked at golang ? [10:55:42] no [10:55:49] supported by Google IIRC, that really looks like an interpreted language [10:55:53] but it is actually compiled [10:56:19] http://en.wikipedia.org/wiki/Go_(programming_language) [10:56:43] forth is quite unique [10:56:56] it's stack-based, written in RPN form [10:57:13] 1 2 + 3 = [10:57:23] that sort of thing [10:57:39] RPN is how my father introduced me to computer arithmetic [10:58:01] and its "words", which are equivalent to functions, can be customised in both the compiler and at execution time [10:58:26] i.e. you can run arbitrary forth code while other forth code is being compiled [10:59:36] like a C macro? [11:00:46] no... the code actually runs during compilation, it doesn't just expand [11:01:04] like a C macro if the C preprocessor let you run arbitrary C code [11:01:14] which had access to the compiler state [11:03:44] types are unsafe, the words just expect a certain type as input, and if you give them some other type, they will just interpret the memory to be the wrong type, bitwise [11:04:07] like C if every argument were void* [11:04:43] you can see that this kind of language would appeal to a different crowd than BASIC, right? [11:04:55] not really [11:05:03] I am really a noob when it comes to all that language stuff :( [11:05:16] I barely understand the need to execute forth code during compilation [11:05:42] I don't know any other language that does that [11:06:22] as you said earlier, that seems to be a language for real hackers :] [11:06:41] yeah [11:06:45] not for the child I was in the 80's enjoying basic :-] [11:07:07] yes, I was too young for it as well [11:07:33] I actually read about forth when I was about 16, in a magazine article [11:07:38] two things I regret are: not learning C cause turbo pascal + assembly filled my needs [11:07:43] assembly looked much easier and more interesting at the time [11:07:46] and not learning java cause it is everywhere nowadays ;-] [11:08:44] I loved assembly, it makes you feel like you are actually moving electrons from a transistor to another one [11:09:07] that surely taught me that reading the doc is important [11:09:26] yep :) [11:09:34] who know what INT 13,1 does? :-] [11:11:00] anyway got the new jenkins/gerrit gateway to finally work properly [11:11:12] it is time for me to grab a lunch downtown. [11:11:16] have a good night! ;-] [11:11:22] well done, bye [11:31:25] I'm fixing some bugs in Phabricator to enforce everywhere the differential.allow-self-accept setting. From Arcanist (arc diff command) or from the add reviewer form, it weren't possible to add an author as reviewer. [11:59:18] hi, I'd appreciate a comment about a part of the code inside SpecialUserlogin.php [12:00:32] hi r4ph431 - go ahead and ask your question, I shall see if I can find someone to help [12:01:31] authenticateUserData() defines $this->mExtUser if an ExternalUser is setup (what I did) [12:01:56] then call $this->mExtUser->authenticate() (which does the real job of authentication) [12:02:26] but soon after it call $u->checkPassword() unconditionnally [12:02:39] that's the 1st thing I don't understand [12:03:34] (BTW, just to check, is this the latest version of MediaWiki you're running?) [12:03:53] REL1_19 [12:17:00] r4ph431: normally, you should have a class extending AuthPlugin with methods to perform this check, it's what $u->checkPassword() should call. [12:17:28] What authentication extension do you use? [12:18:29] I working on enhancing AuthDrupal [12:19:02] so the answer is: I should have 2 classes, one extending ExternalUser, the other AuthPlugin, is that right ? [12:19:53] you should already have one as I see in the plugin page: [12:19:54] Auth_Shibboleth (ShibAuth) http://shibboleth.internet2.edu, Extension:Shibboleth_Authentication [12:19:58] and this is the way Auth_Shibboleth does [12:20:28] ok, I actually do have an active AuthDrupal extension [12:20:33] yes: [12:20:34] AuthDrupal.php:class AuthDrupal extends AuthPlugin { [12:20:39] then comes my second question [12:21:03] how can I call ExternalUser->authenticate() from AuthDrupal->authenticate() [12:21:28] since they both do the very same work and I already implemented this inside my ExternalUser object. [12:22:26] seems like I've no access to LoginForm->mExtUser from AuthDrupal->authenticate() [12:22:50] (otherwise I would just do $XXX->mExtUser->authenticate() [12:24:12] on a side note I don't know how extauth/MediaWiki.php would deal with AuthPlugin and this call to checkPassword() [12:24:44] (thanks for helping, Dereckson) [12:24:52] I'm sorry I can't straight help you on this matter. I'm not familiar with ExternalAuth, according https://www.mediawiki.org/wiki/ExternalAuth it's planned as a remplacement of AuthPlugin. I wrote my auth extensiuon against the old AuthPlugin. [12:25:10] I read on the wiki page "A crude version of ExternalAuth was checked in as of r53497. It should be considered experimental and not suitable for production use unless you're desperate. Possibly the most notable omission is that nothing whatsoever is currently logged, so local accounts will magically appear without any record of when." [12:25:46] Maybe r4ph431 should you wait the code is more mature before migrate AuthDrupal to use it. [12:25:51] (hello sumanah) [12:27:43] When you say "I working on enhancing AuthDrupal", do you have a goal to achieve, something not working you would want it works now? [12:28:06] I would wait (or code), but the issue is that it's not compatible with the API [12:28:22] so I can't login using Drupal via the MW API [12:28:40] (so I can't use getCurrentDrupalUsername :) ) [12:28:54] oups, should have pasted git-remote-mediawiki) [12:29:50] I already did some work on this but upstream is far from reactive on both Drupal and MW side so I worked on gitorious [12:30:55] r4ph431: I try to help people who want to make changes in MW upstream, so if you point me to the talk pages/bugs/changesets/mailing list discussions that made you feel like you were not making headway, I will see what I can do for the future [12:38:39] sumanah: do you know if someone is still working on ExternalAuth? It's a code submitted as a draft by Simetrical, who doesn't seem active anymore. [12:39:16] Dereckson: Shall we look at its Gerrit repo? [12:39:43] * sumanah looks at https://www.mediawiki.org/wiki/ExternalAuth [12:40:23] Oh, it does not have one https://gerrit.wikimedia.org/r/#/admin/projects/ [12:41:30] * sumanah looks at the includes dir [12:42:50] investigation takes me to https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/core.git;a=history;f=includes/ExternalUser.php;h=9a01deb70b8c2e75a2c202d5df8c58d4a9907c36;hb=refs/heads/master [12:43:35] * sumanah waits for gitweb [12:44:48] Dereckson: from https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/core.git;a=history;f=includes/ExternalUser.php;h=9a01deb70b8c2e75a2c202d5df8c58d4a9907c36;hb=refs/heads/master and https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/core.git;a=history;f=includes/extauth;hb=refs/heads/master it looks like Reedy & iAlex are the last to fiddle with ExternalAuth stuff [12:45:11] if at all [12:45:23] and Aryeh has not worked on it since 2009 [12:52:51] Sep 26 00:49:32 2009 reworked the whole (0d1d984313) [12:55:28] but the ExternalUser::newFromName() was already inside includes/Login.php (before the rename) [12:55:52] btw r4ph431 the mediawiki-l list might be able to help you more thoroughly [12:55:54] !lists | r4ph431 [12:55:54] r4ph431: mediawiki-l and wikitech-l are the primary mailing lists for MediaWiki-related issues. See https://www.mediawiki.org/wiki/Mailing_lists for details. [13:14:07] r4ph431: if you wish to improve MediaWiki external authentication support, your contributions would be welcome. We've the CentralAuth infrastructure for WMF wikis, and it (clumsily) works for other authentication ways with AuthPlugin, so this is not a current development priority. [13:15:25] r4ph431: you should find relevant documentation on https://www.mediawiki.org/wiki/How_to_become_a_MediaWiki_hacker (especially coding conventions, and how to setup a working environment in the section 'Submit your changes') [13:17:25] I think Dereckson is 98% right [13:17:43] "this is not a current development priority" merely means that most of the existing developers aren't prioritizing working on it directly [13:18:09] but that does not mean anyone is opposed to improving it of course [13:18:35] I only want to stress that because I think there are other projects where "not a priority" translates to "we don't want this" so I want to emphasize that here, we don't mean that :) [16:18:44] chrismcmahon: can you point me to zeljkof's and Michelle's BZ accounts so I can make sure they have all the privs they need? [16:25:49] sumanah: I'm not sure if they have BZ accounts set up. I have admin powerz on BZ too, so if it's an issue I can handle that on the spot. [16:25:58] okay chrismcmahon [16:26:06] I encourage them, of course, to get BZ accounts [16:26:56] sumanah: of course! :-) I'll be talking to them both soon, will cover that. [16:27:01] OK! Thanks [16:46:38] chrismcmahon: you may find https://gerrit.wikimedia.org/r/#/q/project:test/mediawiki/extensions/examples,n,z useful in your training - that's what we use in the https://www.mediawiki.org/wiki/Git/Tutorial to train new people in both submitting and accepting patchsets [16:51:04] sumanah: I did that one myself. For Z and M, I actually made changes to operations/wmf-config, because hopefully RSN all three of us will be making real changes in there e.g. https://gerrit.wikimedia.org/r/#/c/26611/ [16:51:45] sumanah: and of course discovered a bug that I'm going to have Z work on with Antoine. [16:52:01] chrismcmahon: I saw that you were working in there, yeah -- one nice thing about the training\examples stuff is that we can easily give these new folks +2 access in that repo so they can see what it's like to merge or abandon [17:18:16] sumanah: hello [17:18:34] srikanthlogic: I wanted to introduce you to andre__ in case you had not virtually met before :) [17:18:56] srikanthlogic: andre__ is the new Bug Wrangler & will probably be feeding you bugs or asking you questions about language engineering stuff [17:19:02] we just had Bugzilla contact I think :) [17:19:06] I have been seeing andre__'s bug updates. /me waves [17:19:10] hi srikanthlogic! [17:19:18] hello andre__ [17:19:21] that's all :) [17:19:52] hehe [18:20:29] hey mlitn, I'm here when you need me [18:22:02] DarTar: ok cool - commits still being reviewed though; I'll mention you when it's on test ;) [18:22:59] sounds good [18:32:39] lwelling1: There are two of you!? :) [18:34:11] no marktraceur, there's only 1, see the number 1 at the end of her name? Duh. :p [18:35:36] Reedy: Right--lwelling[1] is the second element in the array! [18:35:42] Reedy: Also, it's Luke Welling :) [18:44:52] im having troubles running phpunit tests - they die with a php parse error on a nonexistent line, eg: [18:44:52] awjrichards@ubuntu:/mnt/hgfs/testing/core/tests/phpunit$ make phpunit [18:44:53] php phpunit.php --configuration /mnt/hgfs/testing/core/tests/phpunit/suite.xml [18:44:53] PHP Warning: Unterminated comment starting line 308 in /mnt/hgfs/testing/core/includes/Defines.php on line 308 [18:44:53] PHP Parse error: syntax error, unexpected '*' in /mnt/hgfs/testing/core/includes/Defines.php on line 309 [18:44:53] make: *** [phpunit] Error 255 [18:45:06] anyone have any idea what might be going on? [18:46:00] that's from a checkout of a core changeset; if i switch to master, similar problem: [18:46:00] awjrichards@ubuntu:/mnt/hgfs/testing/core/tests/phpunit$ make phpunit [18:46:01] php phpunit.php --configuration /mnt/hgfs/testing/core/tests/phpunit/suite.xml [18:46:01] PHP Parse error: syntax error, unexpected '*' in /mnt/hgfs/testing/core/maintenance/Maintenance.php on line 1304 [18:46:01] make: *** [phpunit] Error 255 [18:47:26] wtf [18:47:26] awjrichards@ubuntu:/mnt/hgfs/testing/core/tests/phpunit$ php -l ../../maintenance/Maintenance.php [18:47:27] PHP Parse error: syntax error, unexpected '*' in ../../maintenance/Maintenance.php on line 1304 [18:47:27] Errors parsing ../../maintenance/Maintenance.php [18:47:42] awjr: Where's the change? [18:47:56] marktraceur: i am currently trying this on master [18:48:02] *nod* [18:48:18] marktraceur: but the change that i first was trying this with is https://gerrit.wikimedia.org/r/#/c/24115/16 [18:54:41] huh, i think it was something resulting from the remote mount in my VM, i fiddled in the host OS and now i think it's ok. weird. [20:18:44] mlitn: https://bugzilla.wikimedia.org/show_bug.cgi?id=40775#c9 [20:18:56] testing section edit / tab edits next [20:19:06] thats's fast :p [20:24:42] mlitn: updated the ticket, nothing else to report [20:25:07] on it [20:25:26] those missing impressions (particularly CTA impressions) are the only blocker [20:31:15] DarTar: any idea if "$.trackActionWithInfo" of clicktracking extension still works fine? [20:31:47] or: do you have any of these events: submit_attempt, init, ... [20:32:32] hold on let me check [20:33:33] you're right: I am getting no submit_attempt events (I was mostly looking at CTAs) [20:34:04] but these are not in the specs (mind you, it would be good to have them) [20:34:24] as for inits, they are sampled so I am not surprised they are not showing up [20:34:34] indeed [20:35:14] it appears that all JS tracking (which uses $.trackActionWithInfo) fails [20:40:21] aarg [20:42:21] wondering if that could be related to some changes that we introduced in e3 [20:42:40] spagewmf, ori-l ^^ [20:44:35] DarTar: the "error" appears to be in clicktracking extension [20:44:49] just pushed a patch for what I think is causing the problem [20:44:54] https://gerrit.wikimedia.org/r/#/c/27621/ [20:45:03] ori-l: could you take a look at it? [20:45:10] DarTar: we haven't deployed anything today or yesterday [20:45:14] i'll take a look [20:45:27] don;t know too much about clicktracking, so might overlook something :p [20:46:12] ori-l: I was thinking of the recent fixes we made to CT, not necessarily in the last couple of days [20:46:32] we made one fix that isn't conceivably related to this [20:47:09] specifically: have a function that initiated an AJAX request return the jqXHR, whereas previously it returned undefined. [20:47:20] well, didn't return anything [20:47:43] mlitn: are you adding an additional parameter to CT API calls? [20:48:21] couldn't you simply encode it into the "extra" field? [20:48:56] sorry - you kind of lost me - what extra field? [20:49:08] the "time" var appears to be added by CT itself [20:49:35] but some additional data is being pushed, yes (through $.trackActionWithInfo) [20:52:42] ok, just had a quick chat with ori-l [20:53:07] making these changes to the parameters, even if time is optional, is potentially risky [20:53:36] as we haven't tested how this may affect other clicktracking data we're currently collecting [20:54:35] so if passing this parameter is all we need to fix the issue, ori-l suggests that you append it to the additional field [20:55:11] with the standard separator (|) [20:55:20] we can just discard this field when processing the data [20:55:46] but I am with ori-l on not making changes to the API itself [20:55:53] mlitn, CT has always logged the time value, so it's weird that the API is complaining about it. Hmm. [20:56:15] spagewmf: it's added by the back-end, no? [20:56:28] it may still complain if you're sending it as part of the payload from the client [20:56:34] i haven't investigated [20:57:29] mlitn: does the fix that ori-l suggested work for you? [20:57:53] yes, trackActionXxx call trackAction() which always adds time and does the API call. [20:58:12] DarTar: no, the time parameter is not being added by AFT, but by CT [20:58:43] I could change the commit to just remove the JS that adds the time param though [20:59:49] which does indeed mean it's crazy that all this time CT has been sending a parameter that isn't in its API. Did MediaWiki's API tighten up to reject this? [21:01:18] spagewmf if that's true, I am now wondering if this may affect other events in the log [21:01:40] that we might be missing [21:02:20] mlitn: I am running out to get some food, I'll be back in 10 [21:05:46] mlitn, do you have the call to trackActionXxx and/or the api post? Maybe something else going wrong, but we don't think it's the missing time param. [21:05:47] food [21:06:40] sure, just a second [21:06:51] http://test.wikipedia.org/wiki/Golden-crowned_Sparrow [21:06:54] at the bottom [21:07:03] click "Yes" or "No" [21:07:07] fill out something [21:07:09] and submit [21:07:28] in console, you'll find a couple of ajax calls to CT [21:07:51] the response has now changed though [21:07:57] response I'm getting now is: [] [21:08:06] no longer the warning about time-var [21:08:15] DarTar: still not getting any "impression" entries? [21:16:11] back [21:16:38] * DarTar looking up the log [21:17:16] yay, I have them [21:17:31] -click_yes [21:18:02] -submit_attempt [21:18:06] -submit_success [21:18:24] -cta_edit-impression [21:18:45] mlitn: ^^ [21:19:14] ok cool, looks like it's fixed [21:19:31] the other 2 minor issues should have been fixed as well [21:19:44] Nemo_bis: Sorry about my disappearance re: TranslateWiki [21:20:00] let me check [21:20:18] Nemo_bis: I'm sort of caught up in another project, so UW won't get my attention for a while (or until I decide to do it in my free time) [21:20:42] mlitn: yes, both underscore and suffix are now removed [21:20:48] thanks, we're good to go [21:20:58] what was the issue btw? [21:21:05] marktraceur: I see, is there someone else who is/could be taking care of it in the meanwhile? [21:21:41] Nemo_bis: Errrr....thedj has been known to work on UW, but other than that I can't think of anyone [21:21:48] DarTar: no idea, but since the patch to clicktracking got merged in and afterward it seems fixed, I'm assuming that did the trick [21:21:58] Maybe someone on the l10n team could hit it? [21:21:58] Maybe see with matmarex [21:22:06] he likes JS/CSS [21:22:09] not sure what magic got cooked by ori-l & spagewmf though - they might know better ^^ [21:22:55] mlitn: i didn't intend to approve it, and the merge a minute or so after [21:22:58] ok, that's likely the last data collection ever we run with clicktracking so I am not too worried [21:23:01] marktraceur: we have literally hundreds of twn and i18n bugs/requests, they need to be taken care of by each "owner" or caring dev [21:23:02] *and reverted [21:23:29] then i have no clue what fixed the problem, but something did :) [21:23:31] Dereckson: I'll try to ping him [21:24:14] Nemo_bis: He's in #mediawiki methinks [21:37:37] mlitn, as I recall every response I've seen to clicktracking api requests is []. But now you've got me wondering why/ [21:40:21] quite odd, yes [21:40:21] ' [21:40:36] nothing changed on AFT & nothing did for CT [21:41:21] pretty much the only thing I can think of would be main api code suddenly being stricter on parameters for a short while, but that doesn't really make sense i guess [21:43:09] CT's excute() doesn't getResult()->addValue(), that's why the posts to api.php return is []. Maybe that's to minimize overhead, but it makes it hard to tell what's going on. [21:55:32] K4-713: pgehres - do you have any objection if I create a quick stub on mediawiki.org for the ContactPageFundraiser extension? [22:05:27] hi vvv [22:05:36] hi sumanah [22:05:48] K4-713: pgehres - never mind, figured out that it's nearly identical to https://www.mediawiki.org/wiki/Extension:ContactPage and put in a redirect [22:22:55] mlitn test.mediawiki.org's behavior can jump around, that's on fenari and people try stuff out there all the time. [22:37:36] kaldari: on an anon enwiki pageview that loads a hidden jimmy banner - i don't see an en.wikipedia request that looks like it would be the banner list [22:43:54] kaldari: nm, http://en.wikipedia.org/w/index.php?cache=%2Fcn.js&title=Special%3ABannerListLoader&language=en&project=wikipedia&country=US [22:50:36] binasher: http://wikitech.wikimedia.org/view/CentralNotice#Banner_delivery [22:50:38] binasher: http://wikitech.wikimedia.org/view/CentralNotice/Optimizing_banner_loading [22:51:13] thanks [22:51:34] Ooo. [22:52:22] This is what we changed since last year: http://wikitech.wikimedia.org/view/CentralNotice/Optimizing_banner_loading#2012_update [22:52:38] but I'll write up an email with more info [22:57:49] binasher: https://bugzilla.wikimedia.org/show_bug.cgi?id=40532 [23:06:32] kaldari, you are with the fundraising banners? [23:06:57] I got one the other day and my first impression was that it arrived at an advertising domain instead of wikipedia [23:07:43] I'm not related to fundraising, but I do some coding work for them on Fridays :) [23:08:09] isn't still Thursday for you? :) [23:08:25] yeah, you caught me :P [23:08:30] xD [23:08:34] it's Friday here [23:08:44] I remembered that bit of feedback when reading you talking about banners [23:09:05] I don't really know who is working on that [23:09:49] Officially, I don't work on fundraising at all, but I used to in past years, so I pitch in a bit in the fall :) [23:10:34] I actually don't know who designing the banners these days [23:10:39] it might even be outsourced [23:11:25] [23:22:08] kaldari: here's the rule from last year [23:22:09] acl wpstatic urlpath_regex ^/wiki/Special\:Banner(Controller|ListLoader) [23:22:18] I can just make that work again [23:23:15] that would work. [23:23:33] as long as it doesn't matter if the colon is escaped [23:24:26] Although I'm not sure if that works in other languages or not... [23:25:10] binasher: yeah, looks like we're using the localized Special page name now, so that won't work [23:25:32] we changed it to avoid a redirect [23:25:56] or extra cache hit or something :) [23:26:21] Reedy: I have good news for you [23:27:32] kaldari: is ListLoader a constant? [23:27:57] no, varies by language [23:28:07] for example: http://fr.wikipedia.org/w/index.php?cache=%2Fcn.js&title=Sp%C3%A9cial%3ABannerListLoader&language=fr&project=wikipedia&country=US [23:28:35] You could chop off the Special part [23:29:08] and we don't have to worry about the Controller any more, that's ResourceLoaderified [23:30:21] but listloader isn't [23:30:32] why isn't that an API module or RL module? [23:30:44] Or just not yet? [23:31:09] I has to have custom caching [23:31:24] so? [23:31:31] that shouldn't be an issue [23:31:31] but it's in the process of being converted to an API module [23:31:37] awesome [23:31:53] actually, it's already been converted, but we have to do more profiling on it [23:32:46] and make sure there's no performance hit [23:33:14] we're still using the old version in the meantime [23:35:41] binasher: looks like we can't rely on the page name due to aliasing. I would suggest something like cache=custom or some other query string param. [23:41:24] sumanah: oh? [23:41:56] kaldari: we were previously not rewriting cache-control on api requests but are now [23:42:01] Reedy: I'm going to be submitting a patchset but basically I see no reason why we need to keep the reference to PrefSwitch in default.conf [23:42:07] i'm not sure if that's actually intentional [23:42:07] it's not referred to anywhere else [23:42:35] i can do something like cache=custom, but if it's moving to an api request, i could also just make that browser cacheable [23:42:55] that might work [23:45:17] binasher: One issue with that though is that we would be changing 2 variables at once - switching to an API request and reimplementing browser caching at the same time, which might complicate troubleshooting [23:45:46] that's relatively minor though [23:52:05] binasher: just talked to Matthew. Since we only have 2 weeks left, he thinks switching to an API call might be a bit risky, so we're going to go with the regex solution for now. [23:52:10] sumanah: AFT used to use it, via simple survey... [23:52:28] nod [23:52:30] spagewmf tidied it up, so as long as AFTv5 doesn't need it, it can go. Same for simplesurvey [23:52:45] binasher: or however you would want to detect for the query param [23:55:51] kaldari: how about bcache=1 [23:56:02] sounds good to me [23:56:12] what's "bcache" from? [23:56:24] i was just thinking browser cache :) [23:56:27] browser cache? [23:56:29] yeah :) [23:56:38] Sumanah, how is your patchset different from https://gerrit.wikimedia.org/r/#/c/26653/ ? [23:57:12] spagewmf: the answer is that it is not and that I had neglected to check whether anyone had already done this! [23:57:32] Thank you for stopping me from duplicating more work [23:58:07] sumanah, the more code killers the better [23:59:37] The new http://docs.webplatform.org/wiki/Special:Version loads PrefSwitch and SimpleSurvey, thus the nonsensical http://docs.webplatform.org/wiki/Special:PrefSwitch