[00:11:45] ^demon|away: no .gitreview' file found in the MobileFrontend repository. [00:13:17] preilly: do you know how to clone the wmf extensions? [00:13:36] AaronSchulz: individually yes [00:13:56] * AaronSchulz was starring at update-submodules.sh [00:14:15] staring [00:14:29] <^demon|away> preilly: http://p.defau.lt/?HCNdLo7C1AyNzaCemDLpWw - create a .gitreview with that as the contents. [00:14:32] <^demon|away> Then you can push that. [00:14:50] <^demon|away> AaronSchulz: Yeah that's not 100% yet :\ [00:14:58] no script to clone them all? [00:15:50] <^demon|away> Well you could do something similar, the ls-projects would work fine. [00:16:58] what are we doing on /home? [00:17:01] <^demon|away> https://gerrit.wikimedia.org/r/Documentation/cmd-ls-projects.html has an example. Although I wouldn't use --bare [00:17:24] <^demon|away> The plan there is to be a 'wmf' branch on core, pulling in the deployed extensions as submodules. [00:17:31] <^demon|away> Have that checked out to home. [00:17:59] hmm, maybe in the meantime you can fix +verified :D [00:18:08] I can't merge anything by anyone [00:18:17] <^demon|away> I can give us the ability to +verify for now. [00:18:17] ^demon|away: done [00:18:26] <^demon|away> But hashar needs to fix the hooks [00:18:33] yeah [00:19:08] * AaronSchulz hugs Git Bash [00:19:14] <^demon|away> Try now [00:20:06] ^demon|away: did you switch OpenStackManager over as well? [00:20:23] <^demon|away> Oh, nope. [00:20:27] heh [00:20:31] <^demon|away> If it wasn't in make-wmf-branch, I probably missed it :) [00:20:36] ^demon|away: yeah, it probably should be... [00:20:48] makes sense, though it's still a wmf extension :) [00:20:58] <^demon|away> Other than the TMH stuff I was asked to pull. [00:21:08] can you also move LDAP as well? [00:21:30] <^demon|away> I suppose I could be bribed to do those this evening :) [00:21:46] heh [00:21:57] hmm — Your branch is ahead of 'origin/master' by 41853 commits. [00:21:59] you want openstackmanager to ever get better? :) [00:22:08] because I planned on working on it on the plane this friday [00:22:24] <^demon|away> preilly: I rewrote some history for it today, was this a fresh clone or from earlier? [00:22:59] ^demon|away: okay, after a, "git reset --hard origin/master" it's clean [00:24:33] should LocalSettings.php be in the our .gitignore [00:27:36] <^demon|away> Probably. [00:27:59] yeah, just like it's in the svn ignore [00:28:16] <^demon|away> Dumping OpenStackManager and LDAPAuthentication [00:31:26] ^demon|away: maybe the CachedSpecialPage stuff should be rejected and redone as a single patch [00:31:32] it's like 1/3 of the commits ;) [00:31:53] <^demon|away> You can do that. [00:32:41] preilly: status:open project:mediawiki/extensions/MobileFrontend [00:32:45] ^demon|away: just -2 everything? [00:32:59] Ryan_Lane: okay [00:33:01] does that actually push it out of the 'open' queue or is abandon needed? [00:33:28] <^demon|away> AaronSchulz: "Abandon change" [00:33:29] abandon is needed to remove it [00:33:59] * AaronSchulz doesn't want to abandon all the little kitties [00:34:25] * preilly — thinks wow [00:34:40] preilly: maybe you can do it then? [00:34:52] AaronSchulz: nope [00:35:24] AaronSchulz: Animal euthanasia really isn't my thing [00:39:18] preilly: https://gerrit.wikimedia.org/mediawiki-extensions.txt [00:40:40] Does anyone know offhand how to create/delete an article in a unit test? I need it to actually get saved to the database since the code I'm testing creates a bunch of dependent records based on the page_id. [00:40:59] (ie. just making an Article object isn't enough) [00:42:20] preilly: /bin/ls -1d /var/lib/gerrit2/review_site/git/mediawiki/extensions/*.git | sed 's#.*/##' | sed 's/\.git//' > /var/www/mediawiki-extensions.txt [00:42:37] Ryan_Lane: thanks! [00:42:47] yw [13:12:48] * Krinkle rewrite 120% of TestSwarm this week, and about to add another 80% by the end of this week :) [13:13:12] It's becoming a little bit like MediaWiki [13:17:05] ha [13:17:55] sumanah: where do I get an access to the extensions I wrote? [13:18:20] vvv: everyone has access to comment or to give merge requests. [13:18:27] vvv: do you mean Gerrit project ownership? [13:18:35] I think that's what I mean [13:18:44] I am unfamiliar with gerrit terminology [13:19:05] vvv: https://gerrit.wikimedia.org/mediawiki-extensions.txt are your extensions in there? [13:19:16] vvv: yeah, we are all learning some new jargon [13:19:32] sumanah: well, I am an "initial author" of TitleBlacklist [13:19:33] vvv: I need to update https://www.mediawiki.org/wiki/Git/Workflow#Other_MediaWiki_extensions [13:20:07] Plus I have contributed to some extensions to the extent I am capable of reviewing code there well [13:20:53] vvv: for TitleBlacklist and other extensions that have already moved to Git/Gerrit, I'm setting up a procedure right now for people to request Gerrit project ownership. For extensions that haven't moved yet, use https://www.mediawiki.org/wiki/Git/New_repositories [13:20:53] vvv: Anyone can push things to gerrit (well anything with a Wikimedia LDAP account that is). [13:21:11] vvv: But you mean you want to get master-merge permission to review things in that extension? [13:21:25] vvv: https://www.mediawiki.org/wiki/Git/Workflow#How_to_comment_on.2C_review.2C_and_merge_code_in_Gerrit has directions on how you can comment on and review code [13:21:37] Krinkle: well, what's equivalent to code review permission? [13:21:40] vvv: and https://www.mediawiki.org/wiki/Git/Workflow#How_to_submit_a_patch has directions on submitting code. [13:21:45] vvv: There is two [13:21:59] vvv: Gerrit project ownership lets you approve or veto code for merger into trunk (master). [13:22:01] vvv: there is the ability to mark things OK and the ability to perform the merge into the master [13:22:22] vvv: The merge-master is a bit like merging to a release branch/deployment, so I think that is more restricted [13:22:30] vvv: anyone with an LDAP account can give OKs I believe [13:22:38] OKs in gerrit are "+1" [13:22:38] I see [13:22:48] under the Review button [13:23:42] anyone with commit access previously had code review access, and now anyone with an LDAP account can give comments and +1's in gerrit. [13:23:58] except that we no longer have to maintain a user group separately, because it's one thing now :) [13:24:55] I think I used to "ok" revisions on CentralAuth and TitleBlacklist back when there was some pre-release backlog [13:25:22] vvv: did you read http://thread.gmane.org/gmane.science.linguistics.wikipedia.technical/59681 ? [13:27:59] <^demon> Roan needs to finish cleaning up his mess :) [13:28:00] <^demon> https://gerrit.wikimedia.org/r/#q,status:open+project:mediawiki/core+owner:Catrope,n,z [13:28:30] sumanah: well, right now I do [13:30:30] sumanah: sounds like I am eligible for title blacklist [13:30:43] yes, the conf is stored in instantiation, it has no lazy-init https://github.com/Krinkle/testswarm/blob/8e263b1907bb79239ba0cee72ed6ffbf5e1109be/inc/TestSwarm.php#L41 [13:30:54] eh, wrong chan [13:30:56] vvv: I'm setting up a place for requests now [13:31:10] I'm putting you in as one of the first in the queue :-) [13:31:33] Oh cool [13:32:11] * vvv is lucky he is used to that [13:32:43] ^demon: Do you know anything about how authorship is supposed to be linked? From what I understand the svn-import was supposed to use the new emailaddresses where available, but it looks like that didn't happen at all. my svn commits as attributed to Krinkle [13:33:05] same for other svn users [13:33:28] <^demon> All svn users were attributed to @users.mediawiki.org [13:33:42] Why? [13:33:51] I thought the mail from USERINFO was supposed to be used [13:34:13] so that git viewers consider them the same author, based on mail address [13:34:25] and so that the mail address actually works,t oo [13:36:42] <^demon> We're going to set up those redirects. [13:37:03] <^demon> It was mainly for ease and consistency. [13:37:23] <^demon> And if you associate that e-mail address in gerrit, it'll show up as you :) [13:37:32] <^demon> Regardless of what e-mail. [16:31:03] sumanah: I've noticed we're low on information to tell people how to migrate from SVN Tags to using [16:31:09] * to a using a git clone [16:32:55] Reedy: this sounds accurate. [16:32:59] Reedy: I don't know what to tell them. [16:34:12] Reedy: if you have time and inclination to imagine out a few steps, that would be nice [16:34:37] I'm putting some generic information into these release announcement emails [16:34:44] ok [16:35:55] hashar: if you have time later today or tomorrow, check out https://www.mediawiki.org/w/index.php?title=Talk:Git/Workflow&curid=80233&diff=514244&oldid=514189 and see if you can respond? [16:36:10] or I should say https://www.mediawiki.org/wiki/Talk:Git/Workflow [16:36:11] in general [16:39:31] back in 30 min or so [16:54:45] coffee break sorry [17:16:51] hey hexmode [17:17:06] sumanah: hey [17:17:12] hexmode: it's 1:15! [17:17:24] preilly: hi, how are you today? [17:17:27] so it is :) [17:18:53] hexmode: Gabriel's already gone for the day I think, and Aaron is sick, and Tim is busy with fairly urgent stuff [17:19:20] right... and preilly isn't around? [17:19:30] hexmode: I don't know! [17:19:33] * robla doesn't see preilly [17:19:33] https://gerrit.wikimedia.org/r/#q,reviewer:preilly%2540wikimedia.org,n,z [17:20:05] sumanah: just to verify, you got my email about gadget authors, right? [17:20:33] hexmode: oh wow! yes I did! I'm sorry, I saw it in my inbox but missed that it was new [17:21:02] sumanah: k, just let me know if you need the data munged somehow [17:21:20] hexmode: thank you for this! wow! [17:21:53] this is gold [17:22:07] * hexmode agrees [17:22:14] are you the goldsmith? :D [17:22:34] I shall have to reach out to them this week re Berlin hackathon [17:23:00] https://en.wikipedia.org/wiki/Wikipedia_talk:Gadget#Gadget_creators_invited_to_Berlin_hackathon_1-3_June_2012 [17:24:41] ok, so, when preilly comes by, we can talk about 20% time [17:25:12] other than that, I think there's nothing in particular we need to say/do here. [17:25:19] checkin can be over 5 min early :-) [17:25:23] thanks hexmode, robla [17:25:42] sumanah: The top person on the list was a complete surprise for me :) [17:26:07] hexmode: I had just run across their name a few days previous for the first time [17:26:58] kevin_brown: ping. [17:27:51] hexmode: not sure if you saw this, but in Gerrit you can search by reviewer like so: https://gerrit.wikimedia.org/r/#q,reviewer:preilly%2540wikimedia.org,n,z [17:28:25] yeah, saw your link above [17:30:04] nod. [17:30:21] sumanah: I'm not going to be able to do 20% time today [17:30:44] sumanah: we are behind on a Zero launch and I need to focus on that today [17:30:53] preilly: got it. rescheduling for another day then? [17:31:01] good luck! hope this new Foy guy helps :D [17:31:02] sumanah: tomorrow [17:31:28] preilly: got it. Thanks. and if you have any questions re Gerrit you know we're around :) [17:50:37] ^demon: how do we do deploys now? [17:56:58] do we not have a lint check and bot for the verified status in gerrit? [17:57:24] hashar: did you look for me? [17:57:50] aharoni: yeah [17:58:14] aharoni: was answering some of your gerrit question on https://www.mediawiki.org/wiki/Talk:Git/Workflow#Branches_and_patch_sets [17:58:28] about branches and patchset :-D [17:58:45] robla: gah, the openstack thing is today [18:00:26] hey, devs, how do we deploy now? [18:00:43] Ryan_Lane: yeah, I'd love to know [18:00:49] do we have any docs for this [18:00:49] ? [18:01:12] hashar: so it's even weirder than i thought. If `git review' sends only one commit, why bother with branches? it sounds like i have to create a branch for every single commit. [18:01:14] Ryan_Lane: well, http://wikitech.wikimedia.org/view/How_to_deploy_code#Basic_common_sense doesn't look right [18:01:21] hi Ryan_Lane [18:01:25] robla, sumanah: how do we deploy now? [18:01:41] Ryan_Lane: it just has a, "THIS NEEDS TO BE UPDATED TO REFLECT THE NEW GIT/GERRIT WORKFLOW." message at the top [18:01:43] if we have site issues, we're going to have some major problems if we don't know how [18:01:52] http://thread.gmane.org/gmane.science.linguistics.wikipedia.technical/59832 has the most current info I think [18:02:01] some guy with git-svn will merge git changes to deployment? [18:02:08] * Aaron|home isn't quite sure [18:02:12] Ryan_Lane: what sumanah said [18:02:14] aharoni: yeah this way you can work on a change. If you have to interrupt, just switch to a new fresh branch [18:02:18] Aaron|home: what openstack thing? [18:02:30] 6th floor collab [18:02:46] aharoni: also using branches locally means your submitted commits will always be depending upon a change that was already merged [18:03:30] hashar: thanks. i started a meeting, TTYL. [18:03:32] aharoni: you can think about local branch as giving a name to a commit sha1 :-) [18:03:42] I'm not seeing any docs on how to deploy.... [18:03:45] aharoni: ok come to me tomorrow morning we can talk about it plenty [18:03:59] we're deploying from svn? [18:04:02] Ryan_Lane: use 1.19wmf1 branch IIRC [18:04:19] so, we need to git to svn into the branch, then deploy from there? [18:04:30] Ryan_Lane: that's what I think I heard [18:04:35] heh [18:04:39] that's not going to be fun [18:04:43] don't quote me on that but I think that is the transitional plan [18:04:53] ok. we need to document how to do this now [18:04:54] I was thinking "I don't have git-svn :/" in my mind [18:05:05] and by now, I mean right now [18:05:15] Ryan_Lane: Chad's in the middle of a test right now I think [18:05:17] if we have an outage we're *fucked* [18:05:19] right now, right now, right now! [music] [18:05:42] Reedy: you had the skeleton of a plan? [18:05:58] Ryan_Lane: I think the bug is https://bugzilla.wikimedia.org/show_bug.cgi?id=34138 [18:06:01] sumanah: cepacol pills seem to work (at least so far for the few min)...let's see how long they go [18:06:09] Ryan_Lane: which have some dependencies [18:06:11] good luck Aaron|home :-) [18:06:13] robla: do you not see openstack on the calender [18:06:19] it's at like 6pm [18:06:42] you guys know we have no way to switch masters right now, right? [18:07:32] ok…. so, for now we just push into svn [18:07:45] how do you switch masters ? [18:07:52] the same way we did before [18:07:54] for now [18:07:56] isn't it a change in some load balancing php file on fenari ? [18:08:00] yes [18:08:09] on a local svn right? [18:08:11] yep [18:08:13] that should be ok [18:08:21] so, for now, at least for that, we continue as we did before [18:08:30] that seems unrelated to the MediaWiki svn move to git [18:08:39] no. it's related [18:08:43] we did not migrates everything to git overnight [18:08:45] the configuration repo is now git [18:08:49] ohhh [18:09:00] * sumanah needs to go, meeting [18:09:01] so is core, and all of our extensions [18:10:11] /home/w/c/php-1.19 is still svn.wikimedia.org/svnroot/mediawiki/branches/wmf/1.19wmf1 [18:10:15] extensions probably the same [18:10:39] wmf-config has a .svn and now .git [18:11:41] grr and no .git [18:13:09] wait is it or is it not git? [18:13:15] * Aaron|home is dreadfully confused [18:14:32] I am too [18:14:43] to me the good old paths are not git [18:14:48] might want to ask Reedy [18:14:57] I am heading out now [18:15:00] will be back tomorrow [18:16:31] call on my mobile phone if needed, it is on office wiki and in contact file on fenari [18:16:37] ++ [18:33:49] is there supposed to be a separate gerrit for each extension? i'm getting this for aftv5: http://pastebin.com/raw.php?i=ZBnDGtL0 [18:35:12] Aaron|home: nothing on fenari should have changeds [18:40:05] rsterbin: I think you've just gotta push a gitreview file [18:40:13] But you need to speak to someone who actually knows [18:42:58] rsterbin, yes, but not all extensions are migrated yet [18:43:27] rsterbin, it needs to be configured [18:43:35] is this something i should do myself, or do i need someone at wmf to do it? [18:43:51] Has AFTv5 only been marked r/o in SVN this afternoon? [18:44:41] i was able to commit in svn this morning [18:44:50] Yeah, that's what I noticed [18:44:51] (before i remembered i was supposed to be using git...) [18:45:02] last commit at 14:48, 22 March 2012 [18:45:13] [/trunk/extensions/ArticleFeedbackv5] [18:45:13] * = r [18:45:21] Try committing again? [18:45:30] Just out of interest more than anything [18:45:49] ok [18:46:17] If it lets you, the SVN config must be broken [18:46:38] sweet, commit failed [18:46:55] which is something i never expected to say ;) [18:47:05] I wonder if Chad "fixed" it for AFTv5 earlier then [18:47:16] * Reedy goes to look for more blind people to follow [18:51:22] Platonides: is garrit configuration something i can do, or do i need someone at wmf to handle it? [18:52:22] rsterbin: i've got a feeling you can copy the .gitreview file from another repo and commit it [18:52:27] Yea [18:52:28] in some way o ranother [18:52:38] I'll do it for AFTv5 [18:54:09] rsterbin, Reedy: done [18:54:44] thanks [18:55:48] ouch, fail [18:56:05] http://pastebin.com/raw.php?i=rxMcNwsk [18:56:40] https://www.mediawiki.org/wiki/Git/Workflow#missing_Change-Id [18:57:12] thx [18:58:32] yeah, mine's very different [18:58:58] lol [18:59:06] It's probably not for that reason [18:59:14] It's probably just that you made those commits before you installed the hook [18:59:39] Run git commit --amend and save the commit message without changing it [18:59:48] Then run git log and see if that magically added the Change-Id [19:00:08] If it did, you can use git rebase --interactive master to reword all of the commit summaries [19:11:31] RoanKattouw: that was it, thanks -- all good now [19:11:54] Sweet [19:19:02] RoanKattouw_away -- (for when you get back) is prototype set up for git yet? [19:37:44] blah blah blah back [19:38:03] Ryan_Lane: sorry the hacker place I was in was closing [19:38:30] Ryan_Lane: is there still an issue about deploying live hack / switching masters ? [19:45:16] Ryan_Lane: ^demon - heads-up re the email just now from robla to wikitech-l - [Wikitech-l] MediaWiki core deployments starting in April, and how that might work [19:47:15] <^demon> Got it, will look. [19:47:42] * sumanah makes Gerrit accounts for people [19:49:48] rsterbin: No it is not -- but I need to move it to labs anyway [19:50:09] ok [19:50:23] fabrice will be sad about that ;) [19:51:09] Hmm, right [19:51:56] I'll work on this later today, after I finish other misc git migration tasks [19:58:56] Ok, so how do I commit to the 1.17/1.18/1.19 branches, and then make the git equivalent of svn tags? [20:00:29] sumanah: do you know the timeline for when .gitreview files will be added to all of the extensions? [20:00:48] pgehres: When you say all the extensions, you mean all the extensions that are already in Gerrit, right? [20:00:54] correct [20:01:07] and resubmit https://www.mediawiki.org/wiki/Special:Code/MediaWiki/114374 [20:01:22] pgehres: ok, just wanted to make sure. ^demon knows more about that than I do. If there's a specific one that you need to be updated, if you tell him he can probably frontload it [20:01:52] <^demon> pgehres: Just do it as you see it. [20:02:05] ^demon: I can't commit to master [20:02:19] <^demon> On what project? [20:02:21] ah, pgehres is not in the Gerrit project owners group for this extension [20:02:40] almost all the ones I've look at only have the Admins as owners [20:02:44] yes [20:03:00] not even the "mere mortals" on the cluster could commit [20:03:18] pgehres: hold on, you can commit, you just can't merge, right? [20:03:21] <^demon> pgehres: Just push it for review and someone will approve it. [20:04:09] I didn't think I could push for review with a .gitreview, is that not correct? [20:04:52] <^demon> Yeah just make the .gitreview file, then add and push [20:05:03] <^demon> It doesn't have to be submitted to master to work :) [20:05:23] ah, okay, I thought I tried that and it didn't work. let me try again [20:05:48] pgehres: it's a bit confusing :-/ [20:11:22] Hmm, being told I am missing a Change-Id, double checked per https://www.mediawiki.org/wiki/Git/Workflow#missing_Change-Id and that seems to be correct. Any ideas? [20:13:37] pgehres: I know RoanKattouw was missing a ChangeID a bunch today :-) [20:14:42] pgehres: You might have made those commits before that hook was installed [20:15:00] Check git log master..HEAD for any commit messages that don't have change-ids [20:15:21] If it's just one and it's the tip of your branch, you can run git commit --amend and save without changing anything, that'll add the change-id [20:15:39] If there's multiple, you'll have to use an interactive rebase, but I'm not gonna explain that unless you need it :) [20:22:59] hey it looks like it worked, thanks RoanKattouw and ^demon [20:27:07] yay [20:27:48] <^demon> pgehres: Where's the change? I don't see it :) [20:28:02] <^demon> Ah, DonationInterface. Got it [20:28:04] https://gerrit.wikimedia.org/r/#patch,unified,3454,1,.gitreview [20:28:25] yes, one of our little babies in Fundraising :-) [20:28:40] hashar: I was just confused about how deployment worked [20:28:51] <^demon> pgehres: Approved + merged. [20:28:52] hashar: the answer is apparently: "For now, keep doing it the same way: [20:29:17] \o/ thanks ^demon [20:29:55] Ryan_Lane: good !!!!! [20:30:08] Ryan_Lane: at one point I was like "Oh my god we screwed Ryan :-(" [20:30:23] heh [20:30:50] we are supposed to create a wmf branch with submodule [20:30:57] and migrate the deployement tools from svn to git [20:31:10] might take a bit of tim ethough [20:31:54] * Ryan_Lane nods [20:32:02] no worries, as long as it won't affect current operations [20:32:05] <^demon> The deployment tools are pretty much scm-agnostic. [20:32:28] <^demon> It'll just be a matter of `git pull && git submodule update` rather than `svn up` [20:32:34] we could probably drop rsync in favor of git pull :-] [20:33:04] <^demon> Let's not change deployment process just yet. If it ain't broke ;-) [20:33:12] <^demon> But long-term, yeah that'd be awesome. [20:33:50] well it does not change the process [20:33:53] just the tool :-] [20:34:05] use the build in git rsync system 8-] [20:34:11] <^demon> Well, one step at a time. [20:34:16] sure sure [20:34:48] ^demon: speaking of it . I finally abandoned the huge Jeroen changes [20:35:00] Roan is squashing them in a new change as we write [20:35:19] <^demon> *nod* [20:35:44] <^demon> Should we abandon the stuff to test/mediawiki/core? [20:36:28] ^demon: I think i talked about them with saper [20:36:50] I think he already squashed those as a new changes in mediawiki/core [20:36:59] anyway, he was aware of the test status [20:37:33] ^demon: also I am wondering if we should enforce a commit message [20:37:37] format [20:37:54] aka: 1 line of 60 chars at max, 1 empty line, rest wrapped at X chars [20:38:08] <^demon> Gerrit yells at you if you disobey. [20:38:45] well we have several changes which do not respects them [20:38:54] so I am not sure Gerrit is loud enough [20:39:20] <^demon> Meh, bigger fish to fry right now. [20:39:25] <^demon> Put it on the "Unscheduled" list [20:39:49] if there is anything I can do, please drop me a list of tasks [20:40:05] i18n seems fine for now NikeRabbit seems to be an expert now :-] [20:42:44] hah [20:42:55] I'm planning to try to solve it fully tomorrow [20:43:03] I am working tomorrow [20:43:08] so we can hack together if you want [20:43:14] one Q: is the l10n user accound yet created? [20:43:23] oh that one [20:43:24] is git pull equivalent to svn update? [20:43:28] completely forgot about requesting it [20:43:36] bsitu1: Pretty much yeah [20:43:39] bsitu1: roughly yes [20:43:58] I got a bunch of conflict after git pull [20:44:05] bsitu1: svn update get the diff and apply them. git pull fetch objects (git fetch) and merge them (git merge) [20:44:11] but I did not make any change to core though [20:44:37] bsitu1: Let me come over to your desk and help [20:44:53] ^demon: are you able to create user on gerrit ? Translatewiki would need a dedicated user to submit translation [20:45:32] bsitu1: hug Roan for me when ever he arrives at your desk :-]]] [20:47:55] <^demon> hashar: Yeah, is there a public key I can use for it? [20:48:13] <^demon> Stash the key somewhere on formey and I'll get to it in ~2 hours when I'm back [20:48:20] fine [20:49:04] Nikerabbit: should we commit with betawiki user key ? [20:49:26] Heh, turned out bsitu1 had a clone of one of the older incarnations of core.git [20:49:42] He cloned it before Chad rewrote some of the history, and that confused the crap out of his git client [20:49:53] dohh [20:49:58] poor bsitu1 [20:49:59] yeah, I cloned it on last friday [20:50:15] that the price for being an early adopter :-] [20:50:29] :-D [20:51:00] ^demon: key on formey:/svnroot/translatewiki.pub [20:51:22] ^demon: if you have no idea for the name, I guess translatewiki will be fine [20:51:23] Reedy: may I poke you? [20:51:31] Maybe [20:51:40] <^demon> hashar: Was gonna do l10n-bot [20:52:00] Reedy: I need someone to tell me how to get this shell request ( https://bugzilla.wikimedia.org/show_bug.cgi?id=34005 ) fixed [20:52:01] Nikerabbit not around :/ [20:52:07] It's there for 2 months [20:52:19] So I guess there may be some issues with it [20:52:41] ^demon: so, I want to upgrade labsconsole [20:52:46] to 1.18.2 [20:52:51] how do I do this in git? [20:52:55] <^demon> Sam hasn't put the tags in git yet. [20:52:59] I do a clone and switch to 1.18.2? [20:53:00] <^demon> He's working on that :) [20:53:07] you've got to be kiddin gme [20:53:09] <^demon> But clone then checkout, yeah [20:53:22] <^demon> He made a release without tagging it? Oh yeah, it happened ;-) [20:53:26] so, we just released a security release with no reasonablr way to apply it? [20:53:56] that has a way for people to reset arbitrary passwords [20:54:00] ^demon: whenever you create the user, can you poke a mail to Nikerabbit and I ? [20:54:03] or to take over people's accounts [20:54:23] http://download.wikimedia.org/mediawiki/1.18/mediawiki-1.18.2.patch.gz [20:54:33] Reedy: I pull from svn... [20:54:38] Indeed [20:54:40] I'd prefer to be able to pull from git [20:54:50] why didn't we do one more svn release for this? [20:54:58] It's too late now [20:55:03] We've already moved to gi [20:55:08] t [20:55:13] * Ryan_Lane sighs [20:55:16] so putting stuff in svn gets in a mess [20:55:17] I know [20:55:34] we should have pushed the git migration till after the security release [20:55:38] this is a bad position to be in [20:56:25] bed time for now :) [20:58:08] <^demon> Reedy: We could *attempt* at doing the tags in svn, then importing them to git. [20:58:11] <^demon> I make no promises. [21:03:08] heh [21:06:31] hashar: was? [21:12:27] Nikerabbit: ^demon was asking for the username to submit translation to gerrit [21:12:42] l10n-bot translator translatewiki [21:15:07] bed time [21:15:12] fuzzybot! but really, l10n-bot seems to be clearest [21:20:15] reedy@ubuntu64-web-esxi:~/phase3$ git review [21:20:15] Errors running git rebase -i remotes/gerrit/master [21:20:15] error: could not apply 2a4d580... 1.17: Back out r77893 (beginning of skin system refactor, committed the day before the branch point). Doing something this major right before a release that already has so many major changes sounds like a bad idea. [21:20:15] hint: after resolving the conflicts, mark the corrected paths [21:20:16] hint: with 'git add ' and run 'git rebase --continue' [21:20:18] Could not apply 2a4d580... 1.17: Back out r77893 (beginning of skin system refactor, committed the day before the branch point). Doing something this major right before a release that already has so many major changes sounds like a bad idea. [21:20:22] stupid software [21:20:54] Oh, yeah, that might mess up sometimes [21:20:57] Try git review -R [21:21:15] Also, what branch are you submitting to? [21:21:18] What are you even doing [21:22:32] submit the damn security fixes to 1.17 [21:22:44] To the REL1_17 branch? [21:22:52] git-review is probably trying to submit to master instead [21:22:55] though, after it gives that error, it seems to have reset itself back to head [21:23:11] Reedy: git rebase --abort gets you back to where you were [21:23:56] Reedy: Also, do you mean to push this for review or just to push it? [21:24:03] I couldn't care less [21:24:05] Presumably it's already reviewed and you can just push it [21:24:07] OK [21:24:12] so, run git rebase --abort if you've not already done so [21:24:34] That should put you back where you were, with your security stuff queued up (check with git log), and on the REL1_17 branch (check with git branch) [21:24:55] If that all checks out, run git push origin REL1_17 [21:26:37] To ssh://reedy@gerrit.wikimedia.org:29418/mediawiki/core.git [21:26:37] ! [remote rejected] REL1_17 -> REL1_17 (prohibited by Gerrit) [21:26:37] error: failed to push some refs to 'ssh://reedy@gerrit.wikimedia.org:29418/mediawiki/core.git' [21:26:44] I'll give you perms, hang on [21:29:59] Reedy: Try now? [21:30:17] That looks to have worked... [21:30:26] Ok [21:30:33]