[00:00:28] we'll need to mod the script to actually pass through a tag to git, and also tag the extensions in git, presumably [00:00:50] doesn't seem too hard to do, but not done yet [00:01:25] or use a fixed release date and find the revision preceding it in each extension [00:03:32] maybe check if a tag exists, and use it if it does [00:03:38] if not, use the date [00:03:48] then if an extension does need to be updated, we can just add a tag [00:03:55] No tags on ConfirmEdit... [00:04:06] What date should I use? The original release date? [00:04:28] Or date of the version Reedy last included? [00:04:46] probably the date of the version Reedy last included [00:04:55] yeah [00:05:05] Cool. I think I can track that down. [00:05:08] I think it may just be easier to tag the version we think is the correct one, and not make the script conditional [00:05:41] sure, whichever way seems easier [00:05:54] I agree-- I was thinking I would pass the version into exportEtension, and then export the tag... after I've taged them all [00:06:37] Oh, or were you saying one tag for both 1.18 and 1.19? [00:09:26] tags for each [00:09:39] whatever makes it easiest for the script, really [00:10:09] you don't want to be forced to use the same extension version for every MW version [00:10:40] csteipp: Tim and I have a phone call starting now-ish. are you good to go? if you need anything, go ahead and ping [00:10:55] I'm good. Thanks! [00:18:20] kaldari, marktraceur - are we doing a uw deploy tomorrow? [00:18:45] Yes [00:18:50] awesome [00:19:44] We have pretty much all of the important changes merged in now [00:20:02] and I've fixed all of the issues you identified with the Flickr uploader... [00:20:32] although I think there was one that I couldn't reproduce... [00:20:40] time to pull up a new rev on mwreview? [00:21:21] 'extension error ("we only accept the following extensions .."' [00:21:32] I haven't seen that one [00:21:38] re: key issues, it would be great if we could get https://bugzilla.wikimedia.org/show_bug.cgi?id=37108 fixed before the deploy .. it's basically causing dupe checks to happen at the final step. mark has an old changeset here: https://gerrit.wikimedia.org/r/#/c/8937/ [00:21:41] yeah, would be a good time to update it [00:21:47] Anyone here have contact with a google developer who could fix this? https://www.mediawiki.org/wiki/Thread:Project:Support_desk/Google_spreadhseets%27_ImportXML_not_allowed._Is_Google%27s_user-agent_blacklisted%3F [00:21:57] ok, I'll do an mwrev update now and poke at it [00:23:18] I'll take a look at the timestamp change [00:24:26] there are still a couple missing pieces in the Flickr uploader, like pulling the image description, but those should be done soon. [00:27:16] woo! [00:27:25] http://mwreview.wmflabs.org/wiki/index.php/File:Fabola-7577403396.jpg [00:28:27] kaldari, want me to make some noise on commons to test it? [00:28:48] yes, please :) [00:29:27] We might need to set up some template on that wiki though [00:44:07] kaldari: It looks like you guys have broken stuff? Error: Error: Syntax error, unrecognized expression: .[name="file"] [00:44:16] (Looks like a syntax error in the jQuery selector engine) [00:44:20] *thrown by [00:44:44] Where at? [00:44:56] enwiki Special:Upload [00:45:03] Ah, strike that [00:45:10] enwiki, Upload file link in the sidebar [00:45:16] So I figure that might be UploadWizard or it might not be [00:46:06] kaldari: https://bugzilla.wikimedia.org/show_bug.cgi?id=39779 [00:46:21] Ooooh, it's not UW it seems [00:46:35] Yeah, that's en.wiki's homegrown uploadwizrd [00:47:00] MediaWiki:FileUploadWizard.js [00:48:31] gotta love templates .. importing Template:Information + dependencies from Commons was all that was required to bring my little labs instance to its knees (including exceeding $wgMaxArticleSize post-expansion, apparently) [00:48:58] alright, let's do a shallow template:information clone instead [00:54:29] en.wiki uploadwizard should be fixed now [00:55:03] Oh, looks like kaldari edit-conflicted with me [00:55:11] heh [00:56:09] that 'uploadwizard' is the most unusable interface I've ever seen [00:57:19] looks like my edit only fixes 1 of many errors though [00:58:15] it should be renamed to uploadconjurer until it gets more experience points. [00:59:01] might be totally fixed now [00:59:06] lol [01:01:07] kaldari, small note - the flickr batches don't trigger the copymetadata code for applying metadata across a batch [01:01:35] look like en.wiki is working now [01:01:52] ah, I'll make a note of that [08:02:45] hello [08:09:00] hi [08:09:10] hashar is there a way to update git? [08:09:18] what do you mean? [08:09:19] on beta [08:09:22] git pull [08:09:24] for everything [08:09:26] updating the git software version? [08:09:27] ahh [08:09:30] well hmm yeah [08:09:40] I don't want to do that I just want to know it [08:09:40] cd /home/wikipedia/common/php && git pull [08:09:53] mm [08:09:55] then cd /home/wikipedia/common/php/extension && git submodule foreach git pull [08:10:02] does it update extensions and so [08:10:04] aha [08:10:04] maybe we should add that in some FAQ on labsconsole [08:10:15] ok [08:10:16] if you ever do it [08:10:23] please add a log :-) [08:10:27] sure [08:10:36] I am going to make that automatic anyway [08:10:39] ok [08:11:17] if you want to do it right now, go ahead :-D [08:11:49] fine [08:52:10] hashar we got some broken icons on beta [08:52:13] you know that? [08:52:31] logo's on bottom [09:12:06] petan: I think we got a bug about itt [12:09:00] JeroenDeDauw: hey. What is your AbstractTestCase class for https://gerrit.wikimedia.org/r/#/c/16833/ [12:09:09] JeroenDeDauw: looks like a way to test an abstract class isn't it ? [16:32:30] Hashar: "git submodule foreach git pull" <- that's not the method at [[mw:Download from git]] ? [16:35:10] Amgine: that would update all submodules yes [16:37:08] <^demon> Amgine: That's what the submodule subscription thing does for us automatically in mediawiki/extensions.git [16:37:17] <^demon> So we don't have to run a cron all the time doing that to keep them up to date. [16:42:43] hashar, ^demon: MW suggests "git submodule update --init" for the pull; is the foreach git pull a 'better' method? [16:43:12] (because the former has occasionally crapped out on the farm, like with BlameMaps.) [16:43:17] <^demon> No, git submodule update means "update my submodules to the version the parent repo says they should be" [16:43:34] <^demon> foreach git pull would update each of the submodules to the latest HEAD. [16:43:40] <^demon> Which would still need to be committed in the parent [16:44:04] k. [16:44:59] oh, apergos: we have another request for Wiktionary->DICT dumps. [16:45:14] great [16:45:27] heh. [17:02:56] When I do a push-for-review I get the error (change 16946 closed). I googled around and noticed that this tends to occur during a merge when a rebase should have been done. However, I did a rebase. My log also doesn't show a merge. [17:03:09] Can someone give me a pointer or two? [17:03:52] Also, I did a git show and all of the changes look right. [17:06:07] nullspoon: "I did a rebase" <-- how? [17:06:42] Man marktraceur. You always take care of my git issues. lol. Thanks for helping. [17:06:50] I did a git fetch origin master [17:06:55] then git rebase master [17:07:03] for my branch that is [17:07:10] *nod* K [17:07:20] Oh, can't forget the git clean and git reset --hard HEAD on master [17:07:53] nullspoon: I'd like to try a slightly different method....this is how I do it, maybe it will help [17:08:11] nullspoon: git checkout master && git pull --rebase origin master [17:08:24] * sumanah thought marktraceur might pop up to help with this :D [17:08:35] nullspoon: git checkout && git rebase -i master [17:08:55] awwwright. Give me a few. [17:08:58] nullspoon: Then see if the commit(s) you're trying to push are the only things in the text editor [17:09:10] sumanah: You know I'm always up for a good git puzzle in the morning! [17:10:05] marktraceur: You know, I'm becoming addicted to git. I'm using it for everything I can now. [17:10:28] marktraceur: No Gerrit since I'm the only one using the repos, but still. Git is an amazing product. [17:10:40] :) agreed [17:11:14] nullspoon: Look up "etckeeper", it will version your /etc with git [17:11:24] I haven't used it, but I've seen it in use and it seems pretty awesome [17:11:25] zomgeeeees!!! [17:11:48] marktraceur: so, instead of a real journaling filesystem? ;-) [17:11:58] I've got sparkleshare set up on my wife's laptop and I just run git from command line to transfer files to her. [17:12:04] sumanah: Or rather, on top of one! [17:12:15] I'm also using git to backup my website [17:12:23] and all my database dumps [17:12:23] nullspoon: a real sparkleshare user, in the wild! ;-) more seriously, I don't think I've run into someone using ss [17:12:32] sumanah: Blame paulproteus for this suggestion, he's the one I saw using it [17:12:48] sumanah: I'm not a user of it. It's just nice to have a client that seamlessly pulls from git repos. [17:13:02] sumanah: I looked at it and saw that it required mono and promptly did NOT install it. [17:13:12] My wife runs windows. :) [17:13:56] active development, https://github.com/hbons/SparkleShare/commits/master , good to see. maybe I should try it [17:14:08] It's pretty cool. Has some problems on linux though [17:14:14] Okay - I lied. I tried it. [17:14:15] ;) [17:14:50] It does a pretty good job but it stores lots of info in the mono 'registry' which causes problems if you want to completely wipe it out and start from scratch. [17:14:54] took me a while to figure that one out. [17:15:23] boo to that [17:15:39] yeah [17:15:43] I'm not a fan. [17:15:51] They have a config file you can edit [17:15:55] but it doesn't store everything. [17:16:24] I kept wiping the file out and finding that when I booted ss again it still had a bunch of my settings. It was a while before it hit me that Mono had a registry. [17:16:43] marktraceur: my change was the only thing in the interactive rebase editor [17:17:28] marktraceur: however, your way of pulling and rebasing worked. I just pushed for review. [17:17:36] nullspoon: Sweet. [17:17:44] I couldn't tell you what was different [17:17:49] But I do know that way works for me [17:17:57] * sumanah is a fan! BBBRRRRRRR BBBRRRRRR BBRRRRRR (sweep right)  [17:18:14] * sumanah is also silly and imitating electrical appliances [17:18:16] marktraceur: I like that one command pull and rebase. I'm definitely going to use that. [17:18:30] :) it's in the git docs on mw.org! It's very useful. [17:18:41] It's okay sumanah. I do that exact imitation daily. [17:18:43] :) [17:19:12] "I ran the backup and the server was all "brrrrrrrrrrrrrr....rr...rr...bleh.." [17:19:38] At first I thought "what appliance are you imitating, and why is it related to you being a fan of something?" but then I realized. [17:19:52] oh gosh. That just hit me too. [17:19:55] nullspoon: you and I should do a show that's like Click and Clack [17:19:55] And then the sun rose. [17:19:59] ..wow....just...wow [17:20:08] haha [17:21:01] oh by the way [17:21:08] on the topic of using git for everything [17:21:13] I'm using git for backups now as well [17:21:22] to back up a mediawiki instance and its database dumps no less [17:21:22] Yayyyyy git for everything [17:21:36] nullspoon: you should write up a little case study of that setup and send it to mediawiki-l [17:21:43] I bet other people would like to copy you [17:21:44] It's awesome because it only checks in changes. Run git gc and the database dump one shrinks super small. [17:21:58] sumanah: mediawiki-l? [17:22:02] What's that [17:22:13] nullspoon: it's a mailing list for MediaWiki sysadmins [17:22:17] & extensions developers [17:22:27] nullspoon: https://www.mediawiki.org/wiki/Mailing_lists [17:22:57] Ha, using a bot in one channel to support someone in another. [17:22:59] !lists [17:22:59] mediawiki-l and wikitech-l are the primary mailing lists for MediaWiki-related issues. See https://www.mediawiki.org/wiki/Mailing_lists for details. [17:23:11] marktraceur: I was actually being much more obscure than that [17:23:32] wow. it never occured to me to sign up for that. [17:23:35] marktraceur: once, Chad said that it seemed like I used !lists in #mediawiki at least once a day [17:23:47] It's becoming very clear I have never contributed to an open source project before [17:23:59] nullspoon: we are delighted to be your gateway drug to FLOSS ;-) [17:24:11] haha [17:24:40] nullspoon: http://lists.wikimedia.org/pipermail/mediawiki-l/2011-August/037870.html has a case study from another wiki on how they set stuff up.... I like it when people send those in, because sysadmins so often find themselves reinventing stuff from scratch [17:24:46] and I like it when people show off cool stuff they've done [17:25:03] sumanah: I'd be happy to write it up and send it in. [17:25:10] \o/ [17:25:11] I like helping people out. THat's why my website is around. [17:25:51] <^demon> sumanah: Another one to star ;-) http://code.google.com/p/gerrit/issues/detail?id=1544 [17:26:26] ^demon: done. :-) are your chorohoe & innocent... Google IDs some kind of Jekyll & Hyde thing? [17:26:50] <^demon> innocent... is my personal account. chorohoe is wmf. [17:27:02] so, yes ;-) [17:27:03] <^demon> I always forget which one I'm logged in to [17:27:10] * sumanah is being silly, never mind me [17:28:04] ^demon: do the Gerrit devs frown upon open canvassing to star issues? and do stars have an effect on what they fix sooner? [17:28:40] <^demon> Not sure about the first, oh well :p [17:28:56] <^demon> I don't really know how much of an effect it has, but I figure it can't hurt to raise attention. [17:32:16] marktraceur: I'm getting an build fail on my defaultsettings.php [17:32:16] nullspoon: Possibly useful: http://www.mediawiki.org/wiki/Talk:Git/Workflow#Helpful_script.2C_switch_between_branches_with_intermediate_rebase_18062 [17:32:29] nullspoon: Oh, want to link me to the change? [17:32:53] https://gerrit.wikimedia.org/r/#/c/11137/ [17:32:57] ah [17:32:58] I' [17:33:05] I've actually gotta head out here [17:33:11] K, enjoy [17:33:16] but I checked it and can't seem to find why it's breaking [17:33:17] I'll look at it and review [17:33:28] the syntax matches almost exactly for wgWhitelistRead [17:33:39] hokay. Thanks so much marktraceur! [17:33:42] be back in a bit [17:36:07] Oof, example regex in the comments was "/Special.*/" and it closed the comment. Nasty way to go. [17:55:22] RoanKattouw: I was just thinking the other day about something (not really important right now). about how to store internal wiki links in the canonical html dom for wiki pages (e.g. where templates are not expanded and internal links not expanded with i18n messages and wgArticlePath) [17:55:28] RoanKattouw: a custom protocol :) [17:55:50] better than href="Page name" or href="/Page name", which is indeed ambiguous and ultimately suboptimal [17:55:58] mw-link://Page namw [17:56:01]  [17:56:07] what do you think? [17:56:45] That could work, I guess [19:13:52] <^demon> K4-713: I was going to make your 2 new repos. Does wikimedia/fundraising/* seem like a good home for them? [19:14:29] ^demon: Sounds perfect. [19:14:29] Thank you! [19:18:46] <^demon> K4-713: Ok, two new repos created, wikimedia/fundraising/DonationEmailUnsubscribe and wikimedia/fundraising/PaymentsListeners - "fundraising" group in gerrit owns them, has review permissions, etc. [19:23:14] ^demon: Great! I'll let my team know. [19:44:52] kaldari: Going well so far I take it? [20:09:20] kaldari, ping? [20:09:37] deploying... gimme a bit [20:10:15] no rush. [20:23:27] hey mlitn [20:24:36] mlitn: ping me if you're around [20:52:14] kaldari: jquery.badge is merged (why are you not in #mediawiki) [20:53:02] Krinkle: thanks! [20:53:52] kaldari: and https://gerrit.wikimedia.org/r/#/c/20224/10..11/resources/jquery/jquery.badge.js some minor changes I made [21:01:20] Dantman: can you rebase https://gerrit.wikimedia.org/r/#/c/19008/ ? [21:01:27] (probably just release notes) [21:01:45] ...we need to fix that [21:13:05] * nullspoon feels dumb that he was bested by an elusive comment close. [21:14:23] nullspoon: Don't, it happens :) [21:15:21] marktraceur: Sure, but...but...but. why me??? [21:15:38] [21:16:11] I'm kinda surprised that my syntax highlighting didn't catch that. [21:26:59] nullspoon: Give it another shot? [21:27:21] yep [21:27:23] Fixed that. [21:27:39] Realized that in my last checkin I removed two variables that were neededin the scope of my if statement [21:27:41] oops [21:27:56] Mmmkay, I anxiously await the next one :) [21:28:49] haha [21:28:58] oddly the code runs on my server [21:29:02] but the unit tests fail [21:29:10] due to the scope issue that is [21:29:26] Right [21:31:11] nullspoon: Passed! [21:31:22] yay! [21:32:05] I was thinking by the way [21:32:42] marktraceur: Did you see my comment about the elseif statement causing my check for wgWhitelistReadRegexp to be skipped entirely? [21:33:15] No, I didn't [21:35:53] OK, I see it now [21:36:45] Dereckson: Saw your comment about commenting the dataTestIsValidMoveOperation function. I didn't write that one. Should I still write up the comment for it? [21:36:47] Sorta makes sense [21:37:07] marktraceur: do you think we should perhaps split those elseifs out into individual if statements? [21:37:43] marktraceur: seems to me that since we're setting whitelisted to true, it doesn't matter if they're elseif statements or not plus it will help future modifications to avoid the variable declaration order issue. [21:37:44] nullspoon: Which elseifs, exactly? [21:37:49] one sec... [21:38:56] marktraceur: it starts at line 2020 [21:39:10] Ah [21:39:29] marktraceur: There is the initial use shortcut. Then the user isAllowed, then the isSpecial, then the wgWhitelistRead, then mine [21:39:40] Right [21:39:44] nullspoon: Why does it matter? [21:39:59] If it's whitelisted in one way, not much use checking if it's whitelisted another way [21:40:39] I mean, you could have it *more whitelisted* and put some kind of halo on the page, but functionally it makes no difference [21:40:45] I could be misunderstanding the problem [21:42:23] marktraceur, nullspoon > what's the comment policy for unit tests? [21:42:35] Dereckson: From what I've seen "who cares" [21:42:40] But I could be wrong [21:43:00] On a personal basis, I would prefer if each function could be documented. But now, I'd understand documentation there isn't a priority. [21:43:49] Dereckson: No mention of comments in the unit testing docs, I'd say it's pretty low priority [21:45:35] Dereckson: For example, http://svn.wikimedia.org/doc/classApiEditPageTest.html [21:45:41] Almost no comments [21:46:06] (though holy crap, amazing inheritance diagrams) [21:46:37] Krinkle, Yup... RELEASE-NOTES [21:49:14] kaldari: There was an i18n bug that got fixed in core yesterday that I was hoping would make it into production, I don't suppose we have time [21:49:30] marktraceur, Dereckson: sorry. Had to step away for a second. [21:49:36] where's the fix? [21:49:40] kaldari: https://gerrit.wikimedia.org/r/21954 [21:49:59] kaldari: There was also https://gerrit.wikimedia.org/r/21960 for WLM [21:50:24] Dereckson: I don't mind documenting it at all. Not trying to shirk the system, just wasn't sure if I should be documenting someone else's work. [21:50:34] the 2nd one was deployed [21:50:38] Oh good [21:50:40] but not the first one [21:50:55] nullspoon: it seems the answer is "it would be nice if you add comment, as it would improve the code readibility but it's not mandatory in tests" [21:51:03] I didn't know about the 1st one :( [21:51:09] Dereckson: Sure thing. I'll take care of it. [21:51:36] I'll see if I can deploy it as well... [21:51:43] kaldari: I don't see the page_id commit in the commit log for the revision on commons [21:51:57] Oh, hold [21:52:04] marktraceur: Back to the elseif statement thing we were talking about. Here's the issue - if you add another variable for more whitelisting awesomeness, say wgAwwYeah for instance, then it won't be reached if the person has even used the wgWhitelistRead array at all. [21:52:39] nullspoon: That doesn't make any sense. /me re-reads code [21:52:54] the elseif statement will reach wgWhitelistRead and end the if statement regardless of following elseifs because wgWhitelistRead exists as an array and count is > 0 [21:52:56] Ohhhhh [21:52:58] where is the commit log for the revision on commons? [21:52:58] Oh oh oh [21:53:00] :) [21:53:31] kaldari: Special:Version points to https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/core.git;h=f5298434f7a17359256de7ba0cb74a20a89f85fb [21:53:36] Which gives a 400 error [21:54:18] well, the deployment to Commons is still running [21:54:29] *nod* that might do it [21:54:36] might take another 10 minutes [21:55:09] to 20.. [22:02:38] marktraceur: I was thinking of just doing multiple consecutive if statements. It's not necessary but might help with future possible bugs that won't get caught like mine was. [22:02:50] marktraceur: Not sure how that will impact performance though. [22:19:28] nullspoon: Why not make the check for $wgWhitelistRead another function, have that function return whether that method yields a whitelist result, and then if it doesn't, do the $wgWhitelistReadRegexp bit [22:19:38] (in fact, you could make both of them functions [22:19:54] ) [22:23:32] hmmm [22:23:43] should I shoot for that in this patch set? [22:24:11] You could, it might make people more likely to merge it....I dunno, maybe not. Up to you. [22:29:21] Hmmm [22:37:06] kaldari: Do you have an image with EXIF date in it by chance? [22:37:19] sure, but I see what the problem is... [22:37:37] looks like you need to update line 850 [22:37:49] to add the time values there [22:38:39] er sorry line 849 [22:38:52] or maybe 848 :) [22:39:02] That's the one [23:26:37] "Click OK to confirm..." [23:26:42] [Yes] [No] [23:32:26] Reedy: "Are you sure you want to cancel?" [OK] [Cancel]