[10:41:52] can't find any doc for mediawiki.action can somebody point me to the right place? thx [10:43:25] the javascript part ? [10:45:15] well i see in one script [10:45:17] mw.loader.using( 'mediawiki.action.edit', function() { [10:45:28] so i wonder about its behavior [10:45:48] the code is under resources/mediawiki.action/ [10:45:54] but I am not familiar at all with that part of code [10:45:58] not even sure what it is meant to do [10:46:14] it is probably the javascript counterpart of includes/actions/*.php [10:46:53] might be JS code loaded only when &action= is used [10:47:00] no more idea sorry :( [10:47:23] sounds reasonable [10:47:44] hmm, no roan, no krinkle to ask... :-( [10:47:50] yeah Roan is in SF [10:47:59] Krinkle in Europe but usually there during SF timezone [10:48:04] (aka past 7pm europe time) [10:48:09] Trevor in US too :-( [10:48:25] so your best luck is to send krinkle / roan an email [10:48:31] and wait for an answer from them :-( [10:48:43] or try your luck on wikitech-l but you might not be any answer before this evening [10:48:53] anyway [10:48:56] I am off for some sleep [10:49:00] be back later [15:48:40] Reedy: I have did a bunch of mediawiki-config this afternoon [15:48:55] Reedy: also helped nike to deploy some l10n extension this morning [15:49:10] also deployed like 6 minor configuration changes [15:51:31] Medication still seems to be screwing my sleep schedule even more [15:52:13] :-( [16:23:33] Hi, anyone know why my ssh connection to gerrit is timing out? "ssh -p 29418 cneubauer@gerrit.wikimedia.org". I can hit gerrit.wikimedia.org on port 22 just fine (although it gives me a pubkey error) but the gerrit instance times out.. [16:26:13] Do you not get a placeholder text? [16:27:06] on port 29418 I just get a timeout. On 22 I get "Permission denied (publickey)." and that's it [16:27:08] What does -vvv give you? [16:28:10] The last line before the timeout is "debug1: Connecting to gerrit.wikimedia.org [208.80.154.152] port 29418". Do you want the whole output? [16:28:37] then "debug1: connect to address 208.80.154.152 port 29418: Connection timed out" [16:30:52] alex@alex:~$ ssh -p 29418 krenair@gerrit.wikimedia.org [16:30:52] **** Welcome to Gerrit Code Review **** [16:30:54] works for me [16:32:49] so it's something on my end... I can SSH to other sites just fine... [16:34:07] well, in this case, port is the issue [16:34:10] you can reach the host fine [16:34:20] and I guess it's fine in a web browser? [16:34:45] yeah, I can log in and get to my settings page in gerrit and such [16:34:58] Are you sure your keys are in Gerrit and Labsconsole? [16:35:11] it's in gerrit, I'll check labsconsole [16:35:47] Um. okay. [16:35:53] hmm, should it be listed under the "manage your ssh keys" option on the left? [16:35:53] So apparently I don't have any in labsconsole anymore [16:36:34] log out, log in again [16:37:12] And they're back. What. [16:37:24] So that's a known bug then? [16:37:35] yeah [16:38:04] mw cookie auth bug [16:38:19] "There were no Nova credentials found for your user account. Please ask a Nova administrator to create credentials for you." [16:38:51] yeah [16:38:53] logout, login [16:39:04] "Have you tried turning it off and on again?" [16:39:12] :D [16:39:56] ok, my key is in there now and in gerrit [16:40:28] is there any chance a corporate firewall would be blocking that port somehow, even though its the destination port? that doesn't make sense right? [16:40:45] it's possible [16:40:47] quite possible [16:41:08] they probably only allow outbound requests on specific ports [16:41:14] ie whitelist rather than blacklist [16:41:19] but isn't the outbound request on 22? [16:41:23] no [16:41:32] -p 29418 [16:41:33] yeah it's not is it [16:41:35] damn [16:41:51] On this note, it was discussed about working around this [16:42:02] and tying port 22 on a different ip to allow people to not need the port [16:42:15] we can then route that to 29418 on the other ip and have it all hunky dory [16:42:35] I assume that's in the pipe for some point down the road? [16:43:14] as a workaround, is it possible to setup an SSH tunnel for an SSH connection? [16:43:16] indeed [16:43:26] yeah, you can do that [16:43:45] really! hmm, okay I'll mess with that. thanks very much for the help [16:43:47] you can clone via HTTP too [16:43:59] wait, I can? that would solve the problem [16:44:04] indeeeed [16:44:15] you should be able to push via HTTPS, but that doesn't seem to qite work [16:44:28] I'm going by these instructions: http://www.mediawiki.org/wiki/Git/Workflow#Clone_the_repository [16:45:01] https://labsconsole.wikimedia.org/wiki/Git#Restrictions_and_Anonymous_access [16:45:17] My git remotes are set up to HTTPS [16:45:27] And I haven't had any trouble pushing stuff to Gerrit for review [16:45:49] oh, fine then :p [16:46:00] Or.. Hmm [16:46:25] victory! [16:46:28] It looks like I set them to HTTPS, then git-review changes them to SSH [16:46:40] but I'll have to tunnel to push [16:46:41] oh well [17:21:10] Hello matthiasmullie, rsterbin, yoni_omniti, RoanKattouw, chrismcmahon : Look forward to today's deployment of article feedback! Are we all set to go? Did we get the code reviews for what we are deploying today? [17:21:23] apart from today's new stuff, we did [17:21:27] I CRed a bunch of things last night [17:21:30] Like at 1am [17:21:49] so I'll have RoanKattouw review the brand new stuff first [17:22:38] matthiasmullie, RoanKattouw: Thanks for getting these final code reviews in! Please let me know if there are any questions for me. [17:23:56] all ready [17:24:58] Hmm [17:25:28] $out->addJsConfigVars( 'wgArticleFeedbackLotteryOdds', .... ) --> that'll add it for that page only [17:25:39] The wgAFLotteryOdds should be in makeGlobalVariablesScript instaed [17:25:57] Otherwise there's still a caching problem [17:26:46] Ahm [17:26:51] I meant GetConfigVars of course [17:31:32] matthiasmullie, rsterbin : Any thoughts on RoanKattouw's comment about GetConfigVars, wgAFLotteryOdds and makeGlobalVariablesScript ? [17:31:46] I'm leaving some comments on Gerrit in a minute [17:32:02] i'll wait for all the code review comments, then fix it [17:32:08] Done, commented on https://gerrit.wikimedia.org/r/#/c/12024/ [17:32:20] Ok done reviewing [17:33:51] Thanks, RoanKattouw ! rsterbin, let us know if you have any questions. (And thanks for burning the midnight oil last night to make this happen, much appreciated ;o) [17:34:57] should be a simple fix for these two issues from CR. RoanKattouw — ok if I refactor after launch? I don't want to hold things up. [17:35:22] Sure [17:35:46] Fix it first, make it pretty late [17:35:48] r [17:37:56] am heading out now. will look at other bugs once i am back. heading out shortly. [17:39:23] RoanKattouw, rsterbin, matthiasmullie: Just so I understand, are we going to fix https://gerrit.wikimedia.org/r/#/c/12024/ now, before we deploy? or are we going to launch with the current version, and refactor it later? [17:40:02] we need to get it fixed now or it won't be working yet [17:40:34] oh nvm [17:41:10] matthiasmullie: OK, thanks. Keep us posted. We only have a couple hours left, so want to make sure we can fit it all in. [17:41:18] RoanKattouw: so it can be merged? [17:41:48] ? [17:41:53] No, 12024 needs to be fixed [17:41:58] I thought Reha was doing that [17:42:01] nvm my rambling; i got it right the first time :p [17:42:15] yes, she probably is; got confused on the "refactor later" part [17:42:50] Cool. While we wait for Reha, would you like to give us an overview of what we plan to deploy today? [17:44:17] sure [17:44:19] RoanKattouw: done. [17:44:25] rsterbin: Merged [17:44:31] thx [17:45:20] Whew I'm finally logged back in. :-) [17:46:04] Hidden filters were missing a where clause in filter count refresh script [17:46:06] Refactored the flagging class [17:46:08] fix problem with mood not registering for form 6 votes [17:46:10] Bug 37630 - Feedback Form simplification [17:46:12] Show the just-posted comment at the top of the list [17:46:14] bug 36771 - CTA 4: Sign up or login [17:46:16] Bug 37434 - "Learn more" links on actions need to link to the help materials [17:46:18] Make the AFTv5 lottery the inverse of the AFTv4 lottery [17:46:20] those will be getting deployed today [17:47:16] RoanKattouw, rsterbin: thanks for the prompt turn-around on fixing and merging 12024! matthiasmullie: thanks for the overview of what we are deploying today! Are we ready to start deployment now? [17:47:45] fabriceflorin: do you want CTA4 to show up already and at what percentage? [17:49:05] matthiasmullie: Can we deploy CTA4 at 0% for now, and check that it works with the URL you gave us yesterday? If not, then you can set it at 10%, and give 90% to CTA5 for the rest of the week. [17:50:29] we sure can [17:51:49] matthias: OK, sounds good. We want to give CTA5 the lion's share of the traffic this week, so we collect more usage data on the feedback page. Starting next week, we will distribute the traffic evenly between CTA1, CTA2 and CTA4. [18:01:29] Hello jpostlethwaite: FYI, we are just now starting deployment of AFT5 to testing, waiting for matthiasmullie to give us the go ahead to start testing. (The first hour was spent doing final code reviews and critical fixes). [18:20:25] hi fabriceflorin [18:20:38] Hi gammamatrix ! [18:21:11] oh, i did not realize i am using my nockname [18:21:31] nickname [18:21:40] nockname :) [18:21:50] *knock knock* [18:21:58] i am getting AFT running on my laptop [18:22:04] :) [18:26:29] jpostlethwaite: Cool. The new version will be up on testing shortly, once Matthias is done deploying there. [18:27:38] Here is the URL for the feedback form, which we can test once matthiasmullie tells us it's ready: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_form=6 [18:28:21] And here is the URL for testing the new CTA4, when it has been pushed to testing: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_cta=4 [18:28:36] fabriceflorin: you shouldnt need ?aftv5_form=6 since it'll be the default one [18:28:48] you'll need the ?aftv5_cta=4 though to test that cta [18:31:24] Thanks, matthiasmullie! We will also want to test that the CTA5 correctly shows your last post on top of the feedback page when you click on 'See all comments'. I'm guessing we should be able to trigger that with this URL: test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_cta=5 Is this a correct assumption? [18:32:15] you could trigger it with that url, although omitting it should do as well( since that one's on 100%, it should show by default) [18:32:25] fabriceflorin: it's on test, take her for a spin :) [18:33:32] Cool! I can already confirm that the CTA5 is showing the last post at the top of the page, at this URL: http://test.wikipedia.org/w/index.php?title=Special:ArticleFeedbackv5/Golden-crowned_Sparrow&ref=cta#125 [18:34:46] fabriceflorin: is the placeholder text in the feedback-textarea fine for you? [18:35:04] matthiasmullie: Did you run the script to clear the messages? I am getting some gibberish in the gray text box prompt for step 2 of Option 6: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_form=6 [18:35:18] looks like it's messed up for you too [18:36:30] Yes, it says <articlefeedbacv5-bucket6-question=-placeholder-no> in the text box. Perhaps it's because we shortened that message? [18:36:58] yes same thing here [18:37:04] working locally, but not on test [18:37:06] l10n cache update? [18:37:38] i'd assume cache, but i've ran the script twice and still persists :) [18:38:25] Also, I am not able to see the CTA4 at all, even when I use this URL: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_cta=4 [18:38:30] There's a typo in the message key [18:38:37] "articlefeedbacv5" is missing a 'k' [18:38:50] RoanKattouw: +1 [18:39:01] Also, the equals sign is probably a typo [18:39:17] +2 [18:39:50] RoanKattouw: the real message has the k though, and not the = [18:40:02] i'm assuming fabriceflorin just typed it over with a typo [18:40:07] RoanKattouw: Sorry for the typos on my end, I had to retype this manually, because I cannot copy and paste this message, since it disappears when I click on it (as it should). So these errors do not exist in the code. [18:40:08] OK [18:40:22] Well then see if the message key in the i18n file is exactly the same [18:40:29] Including capitalization [18:40:43] Also, did you rebuild the l10n cache after pushing out your most recent changes? [18:40:50] oh; apparantly, there's more messages missing (http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_cta=4, all the way to the end) [18:41:52] RoanKattouw: I did, twice [18:42:46] What's the command that you used? [18:43:08] mwscript rebuildLocalisationCache.php --wiki=testwiki --outdir=/home/wikipedia/common/php-1.20wmf5/cache/l10n [18:43:30] (the message key is there btw, the i18n file is up to date) [18:46:30] OK [18:46:34] Try running this [18:46:39] Hmm wait [18:47:04] First, try going to http://test.wikipedia.org/wiki/MediaWiki:articlefeedbackv5-bucket6-question-placeholder-no [18:47:20] That seems to work [18:47:26] So I think it's the RL message cache that's the problem [18:47:38] Run mwscript extensions/WikimediaMaintenance/clearMessageBlobls.php --wiki=testwiki [18:47:42] * RoanKattouw needs to document this stuff [18:48:47] Blobls? [18:49:09] Blobs, I assumed ^^ [18:49:12] heh [18:49:34] RoanKattouw: thank you, that fixed it [18:49:51] Yay [18:49:54] Yeah Blobs [18:55:08] Cool, very nice! I can confirm that the garbled text is fixed in Form 6, that the last user post appears at top of the feedback page when clicking on CTA5, and that the CTA4 works as intended when logged out (but not when logged in, as expected): http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_cta=4 [18:55:42] The one thing I don't know how to test is the new AFT5 lottery as a reverse of the AFT4 lottery. Any ideas on how we could test that? [18:56:43] I outlined that in one of my e-mails [18:56:52] Basically, you find articles with page IDs ending in the right numbers [18:57:46] rsterbin, matthiasmullie : I am still seeing filter counter errors on testing: for example, it says there are 2 unhelpful comments, but it is showing me 5 comments on this feedback page: http://test.wikipedia.org/w/index.php?title=Special:ArticleFeedbackv5/Golden-crowned_Sparrow [18:57:47] you can test any article without the aft5 category [18:57:59] that fits the numbers set in the config [18:59:38] rsterbin: I tried testing an article that doesn't have the aft5 category, and it still shows the Option 6 form: http://test.wikipedia.org/wiki/Page426 [18:59:53] probably because AFT4 is still set to 100% [19:00:35] matthiasmullie: what's the AFT4 lottery set to? [19:00:56] also, did you re-run the filter count script? [19:02:55] rsterbin: didn't re-run filter count script yet [19:03:01] rsterbin, matthiasmullie, RoanKattouw : How can I tell if an article that is not in the aft5 category won the lottery anyway? Or is that feature not going to be deployed until next week? I am trying to figure out why Page426 is showing the feedback form, the talk page link and the feedback page: http://test.wikipedia.org/wiki/Special:ArticleFeedbackv5/Page426?ref=talk [19:04:04] What is the lottery set to? [19:04:26] Right now, every random article I click on has been showing AFT5 feedback form -- that's dozens of articles so far, not a single one has NOT included AFT5. [19:04:50] but no aft4, right? [19:05:02] I have yet to see AFT4 anywhere on testing. [19:05:08] if aft4's lottery is set to zero, that would cause it [19:05:33] that hasn't been changed; might've been 0 though, lemme check [19:05:43] default => 0 [19:05:45] enwiki => 100 [19:05:52] testwiki is not listed so it gets the default [19:07:26] indeed [19:08:31] RoanKattouw: can we easily and rapidly change that, or how does one go about config change? [19:08:40] It's easy, yeah [19:08:59] Clone ssh://gerrit.wikimedia.org:29418/operations/mediawiki-config.git [19:09:08] got it already [19:09:10] The setting is in wmf-config/InitilaliseSettings.php [19:09:21] added it already [19:09:25] git review? [19:09:30] Change it, commit it, submit for review, approve (or ask me to), then pull it down on fenari [19:09:32] Yes [19:09:52] matthiasmullie, RoanKattouw : Glad to hear it seems so easy. Will we have to do the same on en-wiki too? [19:10:06] no, en-wiki is already set to 100% [19:10:09] No, enwiki has it alerady [19:10:12] AFTv4 is at full 100 already there, so not neccessary [19:10:30] Splendid. Thanks so much for explaining this to me. ;o) [19:10:46] RoanKattouw: https://gerrit.wikimedia.org/r/#/c/12042/ [19:10:51] ^^ [19:11:08] OK, approved [19:11:17] Go to /h/w/common and run git pull, that should fetch it [19:11:28] Then run sync-file wmf-config/InitialiseSettings.php 'Enable AFT4 on testwiki' [19:14:47] ok so that's done [19:14:52] AFTv4 is at full 100% [19:15:11] meaning all pages, except those whitelisting AFTv5 through the category, should be displaying AFTv4 [19:15:45] rsterbin: If you have a moment, could you please update the Bugzilla ticket about the wider 10% deployment with instructions on how to test it? Be sure to include the numbers from the config file, so we can check that these numbers do indeed work: https://bugzilla.wikimedia.org/show_bug.cgi?id=37616 [19:16:54] i believe RoanKattouw's suggestion was to go live with 100% AFTv4 today [19:17:07] then slowly ramp up [19:17:57] fabriceflorin: right now, it should be displaying only AFTv4, AFTv5 has no lottery yet (we first have to make sure some pages are being edited and their cache is rebuilt to expose the article ID to the javascript before we turn on the lottery) [19:18:29] yeah [19:18:57] when we start enabling it next week, i'll be saying which articles are in & out of the lottery (since it depends on exactly which percentage we're showing it at) [19:19:42] matthiasmullie: Cool, that did the trick! I can confirm that AFT4 is showing on all pages that are not in the AFT5 category, on a variety of browsers, including Chrome incognito. (And thanks for the explanation about the lottery not being active yet.) So from my standpoint, I think we are good to go to production on en-wiki now. Any reason we should not? [19:19:56] or I might already start gathering the articles for the increments Roan proposed, will save us the work when we deploy later [19:20:24] I can't see any reason - looks good here [19:20:39] you managed to see CTA4 btw? since you reported not seeing it first? [19:21:45] Yes, I am seeing the CTA4 when logged out, as intended. Ok, let's deploy to production, then! [19:22:27] okay [19:23:14] Thanks! I will be away from my desk for 10-15 mins, while you push to production. Good luck! [19:47:45] fabriceflorin: rsterbin: jpostlethwaite: yoni_omniti: chrismcmahon: it's on enwiki, can you all test? [19:47:56] yes [19:48:00] looking [19:48:01] sure thing [19:50:12] Cool. I'm on it. jpostlethwaite, you can adapt all the links in the thread above, and replace 'test' with 'en'. We typically post feedback on this page, to keep the rest of the encyclopedia free of testing comments; http://en.wikipedia.org/wiki/Golden-crowned_Sparrow [19:52:00] on http://en.wikipedia.org/wiki/Special:ArticleFeedbackv5 I see "5 feedback posts since Dec. 2011"? [19:57:33] chrismcmahon: Nice bug! matthiasmullie did you run the filter counter script on en-wiki? [19:57:57] it resulted in an error - looking into it with rsterbin as we speak [19:59:38] Also, AFT4 is showing up for all non-AFT5 articles as intended, however the talk page link still appears for these articles, as on this page: http://en.wikipedia.org/wiki/Talk:Diving_at_the_Summer_Olympics [20:01:10] My expectation was that we would not show the talk page link if the article is not in the AFT5 group. Fortunately, when you click on it, we get the correct message that feedback is not enabled for this article, but it is adding a link that doesn't work to 100% of en-wiki, which is not good. [20:03:59] Actually, it appears that it's not 100% of en-wiki, as I just found an AFT4 page that does not have the link: http://en.wikipedia.org/wiki/Talk:Stones_Brewery [20:05:40] i see the feedback-link on there as well [20:06:39] i don't [20:07:22] hm. it's possible that some talk pages have very old html — from before the last push [20:07:59] Interesting: I don't see the Stones Brewery talk page link on Firefox, but I see it on Chrome and Safari. Either way, the vast majority of the random articles I have checked so far have the incorrect talk page link. Any ideas of what we can do about this? [20:08:20] fabriceflorin: i can do that [20:09:22] fabriceflorin: we could swap the js on the feedback page to default to false [20:10:00] fabriceflorin: just to make sure: are you talking about this page: http://test.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_cta=4 [20:13:10] the trouble is that we have two old versions of the html for the feedback page, but no way to tell them apart. [20:13:42] so we either make both result in a link, or both result in no link [20:15:11] it looks like it was already done [20:20:07] rsterbin matthiasmullie: I think it would be OK to swap the JS on the feedback page to default to false, if I understand you correctly. Does this mean that the talk page link would only show up if the HTML was updated? If so, that seems like a better solution to me than putting a useless link on 100% of the encyclopedia. [20:20:23] yeah [20:21:18] OK, so can we implement that solution instead? I would rather have a lot fewer talk page links than have way too many, as we do now. [20:25:15] hello @chrismcmahon [20:25:21] just saw your tweet [20:25:22] fabriceflorin: we're looking into a temp solution to get rid of it [20:25:39] (to be revised this week, but it should get rid of this problem) [20:25:54] is roan around though? we need him to approve it :) [20:26:12] hi cheezy! [20:26:26] jpostlethwaite: We're done with the deployment on testing, and have moved on to en-wiki. Here are some of the links you could test on production: http://en.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_cta=4 [20:26:35] looking [20:26:57] http://en.wikipedia.org/wiki/Golden-crowned_Sparrow?aftv5_form=6 [20:27:11] i am not sure what i need to test :P [20:27:20] they are all en.* [20:27:56] jpostlethwaite: if you need to get caught up fast, http://www.mediawiki.org/wiki/QA/Article_Feedback_Test_Plan#Pages_to_test might be helpful [20:27:58] jpostlethwaite: Yes, we are now testing on en-wiki. We are done testing on the test site. [20:28:08] ok [20:28:34] jpostlethwaite: some of that doc is not relevant right now, but has some good fast info [20:28:42] thanks [20:29:13] chrismcmahon: so i should give inappropriate feedback? [20:29:21] to see it go through [20:29:29] jpostlethwaite: inappropriate is OK on the Sparrow page [20:29:37] ok [20:29:45] probably better not to mess with other pages [20:30:29] cheezy: not too much automation going on here, but thanks for dropping in :-) [20:32:16] matthiasmullie rsterbin : I just went down on the 3rd floor and asked RoanKattouw if he could approve that last fix you just submitted for review. He's on it. [20:32:56] he's done it already - getting it ready [20:34:03] Thank you RoanKattouw, much appreciated! Hopefully this will the the last time we have to bug you -- I know you have a big deadline on the Visual Editor launch on Thursday, and wish you all the best with it! [20:35:32] chrismcmahon: Would it be possible for you to give a full briefing on AFT5 to jpostlethwaite ? I think it would be helpful, so he doesn't have to guess what's going on ;o) [20:35:50] that would be helpful :) [20:36:51] Maybe you guys could have a chat on Skype, with screensharing set up, or something like that. Alternatively, jpostlethwaite I could give you an overview in person, if you plan to be in the WMF office later this week. Thanks! [20:37:32] in person would probably be more efficient (and fabriceflorin is the resident expert) [20:37:45] fabriceflorin: rsterbin: chrismcmahon: yoni_omniti: jpostlethwaite: the talk-page link should be gone now [20:37:52] I may be able to come up tomorrow afternoon [20:38:22] at first sight, seems to be working [20:38:26] jpostlethwaite: I haven't been on the project 100% so I probably have blind spots [20:38:53] no problem - I am still working on getting it running on my laptop. [20:40:19] jpostlethwaite: also, there is an *enormous* amount of documentation for AFTv5. my test plan was an effort to boil it all down for people seeing it for the first time [20:40:23] fabriceflorin: just FYI: we'll have to look into it further tomorrow, and it might be possible that this jeopardizes next week's lottery start - just so you know [20:40:37] will be able to tell more once we've looked deeper into what exactly the issue was [20:41:05] i have been going through that, it definitely is a lot :) [20:41:08] matthiasmullie: I'm afraid I am still seeing the talk page link on random pages that should not have them. That happens after I clear my caches, in both Firefox and Chrome incognito. [20:41:33] do you have an example url? [20:46:35] I just visited a dozen random pages and it was right on all of them [20:47:39] anyone else testing this? [20:48:01] matthiasmullie: The talk page links went away soon after I reported the problem, so I think we are fine for now. It appears that it is now also missing from some of the AFT5-enabled pages on Firefox, but that is not as big a deal as having the link everywhere, and I can live with it. Here is an AFT5 page that doesn't have a talk page link on Firefox, but has it on Chrome: http://en.wikipedia.org/wiki/Talk:Oxygen [20:48:33] fabriceflorin: I'm seeing it [20:48:47] but it might be an unintended side-effect [20:49:47] there's cases where we're not certain when the javascript is not getting the values because it's some cached html (in which case it may or my not be allowed) [20:49:58] at first, we let too many pass [20:50:20] it's possible that it's a little too tight now, combined with the caches [20:50:33] but it should hold up this week :) [20:51:00] matthiasmullie: Yes, this could be related to caches, either on the server or on my end. Either way, I think the main issue has been solved. Though I understand this could screw up our plans for next week. Let's discuss tomorrow. It's getting a bit late for our proposed call today, so I recommend we talk at 11am PT tomorrow, if that works for you, rsterbin, yoni_omniti and chrismcmahon. [20:51:39] not certain about screwing up next weeks plans, but it could - just want to give you advanced heads-up; we'll look into it more tomorrow and this week [20:51:48] tomorrow works for me [20:51:49] ok [20:51:56] thats 2pm ET? [20:52:03] yes [20:52:13] works for me [20:52:43] Yes, yoni_omniti that would be 2pm ET. I think we need to let matthiasmullie go home now, it's 11pm his time :( [20:53:03] ouch [20:53:03] lol [20:53:12] fabriceflorin: that's noon for me, should be fine [20:53:22] luckily, I'm not morning person :) [20:53:28] OK, sounds good. I am moving our appointment to Wed. at 2pm PT then. [20:55:01] thanks! [20:55:37] Many thanks to matthiasmullie for all his great work today, as well as to rsterbin for staying up late last night to get this first part of the scaling/lottery code started. And kudos to RoanKattouw for guiding our steps despite the fact that he has a huge deployment for Visual Editor on Thursday. Thanks as well to yoni_omniti chrismcmahon and jpostlethwaite for all your help with the testing today. Speak to you guys tomorrow ... [20:55:57] later [20:55:59] :) [20:56:18] cya [20:56:49] Get some rest, you guys! Thanks again ... [21:04:02] New patchset: Krinkle; "Pass --branch to testswarm-submit" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/12108 [21:04:28] New review: Krinkle; "(no comment)" [integration/testswarm] (master); V: 0 C: 0; - https://gerrit.wikimedia.org/r/11830 [21:41:07] Krinkle: any thoughts on possibly exposing the methods in mediawiki.action.watch.ajax.js ? Reason being, we are integrating VisualEditor into mw.org and would like to be able to call updateWatchLink from our save dialog. [21:41:38] rmoen: Yes/No [21:41:50] rmoen: Move them to a mediawiki.page* module, and call them from action.watch.action [21:42:01] or actually, I'm not sure it is even relevant to action.watch [21:42:30] let me check [21:42:31] it can be made public, for sure [21:43:06] Krinkle: ok cool [21:43:10] arch, I hate that Mac OS X considers ./resources/mediawiki.action to be an Automator script (the .action extension) [21:43:27] means I can't double-click to open and files in it are not searchabe/findable in Spotlight [21:43:31] ew yeah [21:43:48] Anyone who knows an acceptable solution for that, I owe you a bear if you tell me [21:43:55] beer* [21:43:59] I don't have any bears [21:44:34] rmoen: oh that file looks like a mess, a bit of everything [21:44:42] mwUriGetAction [21:46:14] rmoen: so yeah, it looks like none of this is related to viewing an action=watch page. this is about ajax-ifing links to action=watch on any page [21:46:23] Right [21:46:32] feel free to move to mediawiki.page.watch.ajax and make public what you want to make public [21:47:00] Krinkle: OK thanks ;) [21:47:07] rmoen: if you want to, perhaps also dedicate a commit to move mwUriGetAction into mediawiki.Uri module and use that in there. [21:47:13] there's an open bug for that iirc [21:47:42] Sounds good [21:48:17] (if you do that, make sure that aside form using wgActionPaths, you also make sure that the given url is indeed on wgServer and wgScriptPath, since mediawiki.Uri is used for other urls as well. [21:48:35] not wgScriptPath, just wgServer [21:48:56] e.g. so that google.com/view/Foo does not yield "param action = view" [21:49:12] good luck (with either or both), and thanks you too [21:49:30] Krinkle, gotcha. [21:49:34] I may ping you again [21:49:45] is my assuption correct that mw.loader.using( 'mediawiki.action.edit', function() { ... }); can be turned to [ResourceLoader|dependencies=mediawiki.action.edit] in gadget definition? [21:49:46] I'll be here for a couple more hours [21:50:00] Danny_B|backup: Yes, that is correct [21:50:00] Krinkle: i thought you're in europe ;-) [21:50:10] Danny_B|backup: Yep ;-) [21:50:24] but on US time ;-) [21:50:29] almost [21:54:26] the undefined in copypaste example on cc/js should be removed i guess, because then it does not pass jslint/hint with only globals mw $ as suggested [21:56:33] Krinkle would editing information template on commons break the site? [21:56:47] its used on over 10 million pages [21:56:49] No, but it will create a big queue [21:56:58] test thoroughly before doing that [21:57:02] right [21:57:07] consensus, talk page, etc. [21:57:09] but edit wouldnt break the site. [21:57:33] If the edit is needed and thoroughly tested, no it won't break the site [21:57:52] could you state the technical issues at http://commons.wikimedia.org/wiki/Commons:Village_pump#Merge_extensions_of_.7B.7BInformation.7D.7D_to_it ? [21:58:21] assuming the new version is throughly tested in a sandbox [21:58:41]