[00:18:02] hi kaldari [00:43:56] drecodeaM: howdy [00:44:14] er... drecodeam [00:44:28] my pinky got trigger-happy there [01:07:47] i got through GSoC [01:07:56] thanks a lot, i am extremely excited for it :D [01:08:29] kaldari: ^^ [03:57:58] Change abandoned: Catrope; "No activity in 2.5 months" [analytics/reportcard] (master) - https://gerrit.wikimedia.org/r/2322 [09:11:24] hi hashar [09:12:05] hellloo Nikerabbit [09:14:42] hashar: it looks like I'm not in wmf-deployment group [09:16:48] on gerrit? [09:16:59] hashar: yep [09:17:23] Nikerabbit so you arent deploying? :p [09:22:44] ToAruShiroiNeko: that makes it a bit hard [09:23:23] aw [09:23:32] Nikerabbit on an unrelated note maybe you can explain me something [09:23:35] Nikerabbit: I think you have to formally ask on https://www.mediawiki.org/wiki/Git/Gerrit_project_ownership [09:23:42] Who would be the right person to contact for https://bugzilla.wikimedia.org/show_bug.cgi?id=22521 ? [09:24:22] hashar: really? I have had SVN deployment access for a while [09:24:43] I have been explicitly asked to not modify any user rights on Gerrit [09:25:14] sorry :-( [09:25:16] hashar what about user lefts? :p [09:25:38] have to poke Roan and sumana then [09:25:50] as I understand it, we want only ops and a few additional people to be allowed to alter the cluster [09:25:58] Nikerabbit just challenge them to a pokemon battle [09:25:59] so the list has been set to a very restrictive subset for now [09:26:00] :p [09:26:24] Nikerabbit: you should most probably be granted the deployment rights though [09:26:31] but that is not something I am allowed to do :-] [09:27:01] if you get any access the list is https://gerrit.wikimedia.org/r/#admin,group,21 [09:27:52] hashar: that's how I found out I don't have access [09:27:55] which is basically ops + techs from Platform Engineering (Aaron Schulz, ^demon, reedy and me) + preilly & awjrichards for the mobile frontend [09:29:05] ToAruShiroiNeko: hmm some core stuff... perhaps ialex or aaron? [09:31:24] but people are just mostly working on their own projects, so it's hard to get anyone to work on other stuff [09:31:46] I am curious who worked with cascade proteciton [09:31:55] if you can find this out maybe [09:32:00] I dont even know where to look [09:32:13] I merely want to prod that person whom might fix the issue with ease as it is their code [09:34:29] you could fine some piece of code which is related and run git annotate on it [13:39:04] hashar: ping [13:39:07] swarm is down [13:39:09] Not connected: Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2) [13:39:10] http://integration.mediawiki.org/testswarm/user/mediawiki/ [13:39:11] Happened within the last 10-20 minutes [13:39:12] Just got back from the supermarket, no idea whats going on yet [13:39:45] btw, nice work on the swarm, running very smoothly from jenkins, SpecialJavaScriptTest is also working nicely in all browsers [13:40:23] Krinkle: mutante and I are upgrading mysql there [13:40:26] to log slow queries [13:40:31] (well mutante does everything) [13:40:51] oh, okay, np :) [13:43:02] Krinkle: so somehow I reenabled testwarm yesterday [13:43:24] there was an issue with the injector and another with the way we let apache write to sqlite [13:43:24] what do you mean, "somehow"? You did that last week. [13:43:34] set-group-id just does not work when using a java app [13:43:52] oh, the sql issue [13:44:02] I have logged that somewhere, can't remember [13:44:06] yeah I was wondering how you fixed it [13:44:22] the $wgJavaScriptTesting something array was missing a layer [13:44:30] namely the unit testing framework ['qunit'] [13:44:33] aha [13:45:13] https://gerrit.wikimedia.org/r/#patch,unified,5440,1,jobs/_shared/ExtraSettings.php [13:47:53] New patchset: Hashar; "Correct wgJavaScriptTestConfig value" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/5794 [13:48:23] that is the fix I did yesterday [13:49:04] New review: Hashar; "Already deployed in production yesterday. That did fix JS tests to timeout." [integration/jenkins] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5794 [13:49:07] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/5794 [13:49:33] Krinkle: and the other one (sqlite files not being writable) is fixed by https://gerrit.wikimedia.org/r/#change,5737 [13:49:38] basically, I chmod o+w [13:49:41] which is VERY LAME [13:49:43] but works :-] [13:49:52] ok [13:51:40] <^demon> hashar: When you get a chance, could you take a peek at Ic4d67fe8? I started rewriting the gerrit hooks but I'm a python noob :) [13:51:54] !g Ic4d67fe8 [13:51:54] https://gerrit.wikimedia.org/r/Ic4d67fe8 [13:51:59] 404 :-( [13:52:14] someone screwed up !g [13:52:17] !gerritsearch [13:52:17] https://gerrit.wikimedia.org/r/#q,`e1,n,z [13:52:24] !gerritsearch Ic4d67fe8 [13:52:24] https://gerrit.wikimedia.org/r/#q,Ic4d67fe8,n,z [13:52:29] !g del [13:52:30] Successfully removed keyword: g [13:52:30] danke [13:52:32] noooo [13:52:34] !g is fine [13:52:34] Successfully added keyword: g [13:52:35] !g alias gerritsearch [13:52:35] Alias "g" already exists. Use "!g del" to remove it. [13:52:51] hashar!!! [13:52:53] !g [13:52:53] fine [13:52:55] !g del [13:52:55] Successfully removed keyword: g [13:52:59] !g is for looking a gerrit number such as change 1234. You could : !g 1234 and it gives https://gerrit.wikimedia.org/r/1234 [13:52:59] !g alias gerritsearch [13:52:59] Successfully added keyword: g [13:53:00] Alias "g" already exists. Use "!g del" to remove it. [13:53:08] stop defining !g :P [13:53:10] !g del [13:53:11] Successfully removed keyword: g [13:53:19] !g alias gerritsearch [13:53:19] Successfully added alias: g [13:53:32] I know what !g does, but /r/ only works for change numbers [13:53:41] gerritsearch works for everything, including change numbers [13:53:45] but also sha1 [13:54:16] !g Ic4d67fe8 [13:54:16] https://gerrit.wikimedia.org/r/#q,Ic4d67fe8,n,z [13:54:35] lol [13:55:37] hashar: Could you garbage all mw-clones before 5724 ? [13:55:48] http://integration.mediawiki.org/testswarm/job/1969/ [13:55:52] http://integration.mediawiki.org/testswarm/job/1968/ [13:56:10] 1969 = e6836d4 = 5725/1 [13:56:18] 1968 and lower is removed by me last week [13:56:27] (which were broken) [13:56:35] they're no longer in testswarm [13:57:17] after TestSwarm 1.0.0 is ready and Jenkins is using it to aggregate results, we can clean up old installs from Jenkins after a run is completed, but for now we'll have to periodically clean up [13:57:58] at least this once to clean up unused installs, we don't have to clean up regularly if we focus on getting the results into Jenkins soon-ish [13:59:21] We should probably work together with jQuery on that (since I don't know enough about Jenkins or Java, better for you to do it). The jQuery team is already using Jenkins and TestSwarm on the old version, but version 1.0.0 provides a much better way to extract the results (JSON API instead of html scraping), but the Jenkins plugin will have [13:59:28] ...have to be rewritten for that [13:59:53] ^demon: good to know you like python :) [13:59:57] To avoid duplicate efforts, we can settle on who will do it (or work on it together) [14:00:10] ^demon: I am busy with labs and gallium today then I am out of town till wednesday [14:00:38] <^demon> Ah, maybe I'll pester Ryan later then :) [14:00:40] ^demon: then preparing the trip to SF + more labs. So I will probably not be able to review the Gerrit hook system any time soon :-( [14:05:46] * hashar opens mail clients [14:06:07] * hashar notices how much mails he received today. Close mail clients. :-D [14:07:01] hashar: did you get my message above? [14:14:00] To avoid duplicate efforts, we can settle on who will do it (or work on it together) ? [14:14:01] :D [14:14:08] Krinkle: ohh the lonnng stuff above [14:14:16] :) [14:15:19] so as I understand it [14:15:29] we need to upgrade Testswarm so we get the new JSON API [14:15:40] then install the jenkins plugin [14:16:30] 1) Wait until TestSwarm 1.0.0 is tagged (will be about 1-2 weeks now), 2) Remember it is a rewrite, all pre-1.0 is incompatible so we need to nuke our install and re-install, 3) Start working on making 3 things work again: [14:16:38] * node-browserstack [14:16:45] * testswarm-browserstack [14:16:49] * testswarm-jenkins [14:16:51] those three :) [14:17:00] the first one already works, and requires no changes [14:17:22] the second one jquery has already started on working (they originally wrote it and will maintain it) [14:17:56] the third one was originally written by github.com/appendto but needs updating either jquery or we will do t hat [14:18:53] since you know a lot of about jenkins, I think you're the right one to do that (to have the jenkins job poll the testswarm api until it the job is ready in testswarm, and then aggregated the results from the testswarm api into Jenkins so that the build results can be viewed in Jenkins. [14:19:19] jzaefferer from jQuery also contributed to the that Jenkins plugin and will be able to help as well [14:25:51] well I think Jenkins plugins use java [14:25:56] well I am pretty sure they do [14:26:08] so I am unlikely to be ever able to rewrite testswarm-jenkins :-∏ [14:26:42] Hm.. I thought you were writing in Java for the things you did so far in Jenkins? [14:27:04] Or using existing plugins that take xml configuration manifests and instructions [14:29:24] Krinkle: Jenkins is a java apps with lot of plugins [14:29:30] so I just cherry picked some plugins [14:29:38] and clicked / filled some fields [14:29:42] I know, but I thought you were writing in Java [14:29:50] at least a little bit [14:29:52] the stuff I wrote is the build.xml ant script [14:29:54] but that is about it [14:29:56] okay [14:30:21] anyway, we will want to get testswarm 1.0 deployed [14:30:26] so you wouldn't know how to make those build results in a Jenkins plugin (those tables with stats, and seemingly endless subpages with more details about each individual test) [14:30:32] which mean updating the debian package, and have it installed on gallium [14:30:42] yeah, regardless of jenkins, we can upgrade to testswarm 1.0 anyway [14:30:46] that would be a great first step. Then we can talk about jenkins/testswarm plugin :-D [14:30:55] since we don't use the plugin yet, so there is no breakage in upgrading [14:31:38] hashar: do we need something in the debian package to tell that the existing install can not be upgraded? [14:31:44] i.e. nuke the db and www files? [14:32:09] maybe [14:32:11] obviously we need it, I meant to ask, is it possible? [14:32:12] I have no idea honestly [14:32:26] we could do it manually by just using dpkg --purge testswarm [14:32:30] then install the new package [14:32:41] k [14:34:58] Krinkle: do you still use the testswarm instances on labs? Can I nuke them all as well as the project? [14:35:06] I no longer use htem [14:35:11] nuke at will :) [14:35:17] good [14:35:26] cause I have not been able to login them for the last 5 months or so [14:36:32] oh ? [14:37:14] GitHub "delete repo" is pretty cool [14:37:15] http://cl.ly/2F1E1e2l2r3O1h1p2P1c [14:37:21] that warning is just epic [14:37:30] "Unexpected bad things will happen if you don't read this [14:37:32] XD [14:39:38] Krinkle: https://bugzilla.wikimedia.org/36241 [14:39:43] got it [14:39:44] bye bye testswarm project :D [14:46:09] <^demon> !unreviewedcode is Here's a list of everything that hasn't had a human reviewer in Gerrit yet: [14:46:09] Successfully added keyword: unreviewedcode [15:04:50] ^demon: !unreviewedcode is a lovely query [15:05:01] <^demon> Isn't it? :) [15:05:25] <^demon> You can get some *really* awesome queries when you start combining things like age: and reviewer: and the current status. [15:05:58] <^demon> The UI may be absolute shit, but the Google guys certainly know how to make an advanced search. [15:30:05] Krinkle: Good morning [15:30:14] Hey [15:34:51] <^demon> RoanKattouw: Did you know of http://gerrit-dev.wmflabs.org/r/Documentation/config-gerrit.html#_a_id_theme_a_section_theme ? [15:35:17] <^demon> Seems like most of the UI can be re-colored without CSS. [15:35:23] Ooooh nice [15:35:33] Krinkle might be interested in that [15:35:58] <^demon> I've been reading through a lot of the gerrit config this morning and came across that :) [15:37:25] Yeah, would help the basic thing. Although it is limited, but I didn't expect it to be possible to them Gerrit anyway, so this is better and it is documented/supported [15:37:33] RoanKattouw: Gonna check out your change 5290 now [15:37:54] I have the skins frontend mostly ready [15:37:56] !g 5290 [15:37:57] https://gerrit.wikimedia.org/r/#q,5290,n,z [15:38:09] Thanks [15:38:11] https://www.mediawiki.org/wiki/ResourceLoader/Version_2_Design_Specification/Task_management [15:38:15] It's still broken, I have some improvements locally [15:38:49] RoanKattouw: btw, there is something else we haven't though of yet [15:38:56] there is a scheduled task for normalization before saving [15:39:04] but we also need it before 'getting' it from the API [15:39:05] Yeah [15:39:10] e.g. when 'skins' is added [15:39:21] Oh, you mean for when stuff is added later [15:39:25] yep [15:39:39] I would like to avoid that [15:39:40] I'd rather not hardcode defaults on the client side [15:39:58] because they are already in the backend somewhere [15:40:17] The default set is in the backend and can easily be exported to the client side [15:40:20] Gadget::getPropertiesBase [15:40:22] Yes [15:40:35] Ideally we wouldn't add new mandatory props when this goes stable [15:40:46] well, future proof :) [15:40:52] There's a few options, like normalize on every read, or renormalize everything with a script [15:41:08] But I'd really just like new props to be optional in that the client can deal with them [15:41:10] Or ... wait [15:41:19] We don't really need to re*normalize* [15:41:24] We just need to fill in the blanks (missing fieldS) [15:41:38] Yeah [15:42:15] a maintenance script would be fine too, that just reads the gadgets blobs from the rows in chunks of N, create gadget objet and re-encode and save back [15:42:17] or something likek that [15:42:57] I'm not sure how that'll go though, it would have to be ran before the front-end is updated. [15:43:07] so the front-end and backend need to be able to deal with unknown properties [15:43:14] which is against the normalization [15:43:24] maybe normalize on read is better? [15:43:29] read and save [15:44:01] when in deployment, we'd have both versions in the database [15:44:25] with normalize on read, there is no issue afaik [15:51:01] Normalize on read is fine, as long as "normalize" means "add missing properties" [15:51:25] As opposed to the more elaborate song and dance we do (or are supposed to be doing) when you save something [15:54:31] yeah, probably just an array merge of base props and real props [15:54:37] Yeah [15:54:40] or something like that, filling in gaps [15:54:40] That wouldn't be hard to do [15:54:51] * RoanKattouw adds to todo list [15:56:43] Oh that would unbreak my localhost too probably [16:00:14] hi RoanKattouw [16:00:24] Hi Nike [16:00:37] RoanKattouw: shouldn't I be in the wmf-deployment gerrit group? [16:01:29] Yes you should [16:01:38] I'll add you right away, sorry for the oversight [16:02:07] Nikerabbit: Added [16:02:34] RoanKattouw: I also submitted the cron stuff to gerrit, it's certainly wrong but I don't know how to improve: https://gerrit.wikimedia.org/r/#change,5783 [16:03:53] <^demon> Nikerabbit: I think I found out how to enable blame on gitweb. [16:04:18] <^demon> https://gerrit.wikimedia.org/r/#change,5810 [16:04:41] ^demon: 5810 doesn't lint [16:04:49] <^demon> I saw. I don't know why. [16:05:08] Nikerabbit: Looks good mostly apart from the user thing, you'll need to ask for ops input on that. I commented on two other things [16:05:38] Your logrotate conf looks good, and you're creating the directory you're storing logs in (this is a common pitfall) [16:06:39] ^demon: Found it [16:07:37] <^demon> Ah ok, missed that. [16:09:57] RoanKattouw: I'm currently a bit blocked by the backend changes by you (skins/position need backend support, and error handling needs the backend to throw errors in certain cases). I've stash-popped my working copy on skins/position front-end in another RL2 branch, will work a bit now on RL QUnit testing, that's on the todo list as well [16:10:09] OK [16:10:14] I'll get this done as fast as I can [16:10:23] no rush, there is enough to work on [16:10:38] i'm blocked on this, but there is enough other stuff to do ): [16:12:46] RoanKattouw: can you look at https://gerrit.wikimedia.org/r/#change,5771 [16:19:19] preilly: done [16:33:36] <^demon> Nikerabbit: I think I just shut up L10n bot e-mails with a 1-click config change. [16:34:03] How so? [16:34:27] <^demon> On the group options page, check "Send email notifications about comments and actions by users in this group only to" [16:34:30] <^demon> [x] Authors [16:34:48] Ooooooh [16:34:49] Sweet [16:35:00] <^demon> That's what spearce claims :) [16:35:48] <^demon> I know it'll shut them up for reviewers, he didn't answer on whether it shuts it up for watchers. I hope so. [16:35:54] * ^demon is waiting for an answer on that [16:36:44] <^demon> hashar: Is there a reason gerrit2 is in the L10n-bot group? [16:39:51] ^demon: I have no idea :-( [16:39:55] I don't think I did it [16:40:00] maybe RoanKattouw ? [16:40:30] Not me [16:40:53] <^demon> Removing it. [16:58:37] <^demon> yes. it should then only email the owner of a change and/or the commit author if the author is a registered user and isn't the same as the owner [16:58:38] <^demon> but users following by stars or watches won't be notified [16:58:38] <^demon> and neither will reviewers [16:58:46] <^demon> RoanKattouw ^ [16:59:02] Coo [16:59:04] l [17:11:50] rsterbin: did you see my last mail? Hoping for a miracle... [17:12:00] hang on... [17:12:17] yes, those events are present [17:12:26] are they missing in your logs? [17:24:50] hi 20% people, sorry I am late [17:25:03] Niklas and who else? [17:25:05] * sumanah checks [17:25:45] used to be Ian but he's Friday now [17:31:30] <^demon> !unreviewedcode | sumanah [17:31:31] sumanah: All commits in Gerrit that haven't had a human review yet: https://gerrit.wikimedia.org/r/#q,is:open+project:mediawiki/core+-CodeReview%252B2+-CodeReview%252B1+-CodeReview-1+-CodeReview-2,n,z [17:31:36] <^demon> Thought you might like that :) [17:31:51] Thank you, ^demon! [17:31:54] <^demon> Oh, you can drop the project:mediawiki/core part if you'd like [17:31:58] and it's not even my birthday! [17:33:57] hi sumanah, what is going on with eventbrite? [17:34:09] hi multichill - I don't know, and I will simply register you manually [17:34:26] We could try one last time? Maybe it was because the link had gotten old? [17:36:34] multichill: nah, I will just register you [17:36:37] it'll take a second [17:40:21] I'm in a bad mood today. Is there a patches-needs-hate list? [17:40:47] Only if you promise to be nice-ish to the authors [17:41:10] I'm always nice to authors :) [17:41:36] vvv: there's a shared query in BZ [17:41:42] vvv: of patches awaiting review against MW [17:41:52] and another shared search of patches awaiting review against MW extensions that WMF deploys [17:46:48] Found it [17:46:49] Thanks [17:47:02] *evil laughter* [17:47:07] Welcome! thanks for the patch review... hmmm...... [17:49:14] plop [17:49:20] hi there Nikerabbit [17:50:19] Krinkle: OK got the skins&top stuff to work [17:50:22] Pushing in a minute [17:50:51] sumanah: is there something that doesn't involve code review in gerrit? It drives me nuts [17:50:55] Nikerabbit: sure [17:51:03] Nikerabbit: so, would you like to help review patches in BZ? [17:51:18] Nikerabbit: or perhaps work on extension review re Special:Interwiki? [17:51:59] sumanah: iw review is planned for next sprint [17:52:00] I know that is already in your backlog https://bugzilla.wikimedia.org/show_bug.cgi?id=22043 but I figure it is ok to double-count it [17:52:37] Nikerabbit: so if you prefer to not do that now, that's ok [17:52:45] * sumanah looks at https://www.mediawiki.org/wiki/Wikimedia_engineering_20%25_policy#Scope [17:53:09] Nikerabbit: there are about 180 patches awaiting review in BZ, and you can do that mostly without Gerrit [17:53:58] hmm [17:54:11] BZ isn't really better than gerrit for review, but I can try [17:54:57] Krinkle: OK, https://gerrit.wikimedia.org/r/5812 and https://gerrit.wikimedia.org/r/5290 now ready for review [17:55:24] Nikerabbit: how about, instead, putting in some GSoC time? [17:55:49] sumanah: naturally [17:56:03] you could spend some of today working with Robin re his plans [17:56:41] not sure if I will catch him today, but later this week definitely [17:57:43] ok, and if nothing else, you could start prepping the exercises you want him to do :D [17:58:05] I also did more translatewiki.net [[support]] work last week, I hope that it counts :) [17:58:12] Sure! [17:58:13] :) [17:58:32] that sounds like it is in scope https://www.mediawiki.org/wiki/Wikimedia_engineering_20%25_policy#Scope [17:58:40] anyway, you have ideas; go forth :D [17:58:53] sumanah: yup [18:57:09] RoanKattouw, hey can you merge & push a fix on Math for me? [18:57:13] https://gerrit.wikimedia.org/r/#change,5243 [18:57:19] On it [18:57:22] disables a couple XSS vectors in non-default setting [18:57:23] thanks! [18:58:31] shall i do a review push to wmf branch or leave that to you? [18:59:40] Go ahead [18:59:53] ok i'll cherry pick it [19:00:16] If nothing else has changed in Math (modulo i18n) you could just update it to master [19:00:39] most parser test cases and i18n, that's safe [19:00:53] <^demon> Yeah, just update to master then :) [19:00:56] hmm [19:01:07] ok i'm not sure where the wmf branches are for this [19:01:19] <^demon> mediawiki/core.git -b wmf/1.20wmf1 [19:01:24] <^demon> submodules in extensions/ [19:01:28] ahhhhh right [19:01:43] Ah yes [19:01:46] <^demon> submodules are awesome :) [19:01:48] Just updating to master will be fine [19:02:10] brion: "How to deploy code" on wikitech has elaborate docs on this should you need them [19:02:25] urgh, it's conflicting with my local extension checkouts [19:02:52] Yeah it does that for me too [19:03:19] One day I might give up and use a separate clone for wmf/* stuff [19:04:38] wheeeee cloning extensions [19:05:05] <^demon> I officially think adding a reviewer is the single best feature we never had in CodeReview. [19:05:26] <^demon> Now it's so easy for me to say "Go look at this Brion" and if I don't get a response in a couple of days I can tell I'm being ignored :) [19:05:44] haha [19:07:01] <^demon> That, and for totally non-urgent things it's nicer than pinging someone on IRC and distracting them. [19:07:25] the danger is there's no context, so it's like "wtf is this code?" :D sometimes [19:08:10] <^demon> Yeah like Aaron always adding me to review his FileBackend stuff ;-) [19:08:19] yeah i gotta brush up on that code :) [19:08:38] what keep you busy nowadays brion ? [19:09:01] code review, mobile, and anxiety mostly :P :) [19:09:26] alllllmost done checking out these submodules [19:09:39] <^demon> Yeah the first `git submodule update --init` takes forever. [19:09:55] anxiety could be solved by bicycling a bit and taking some nice vacations :-D [19:10:18] <^demon> It's pretty much entirely CPU-bound, figuring out ~100 or so repos worth of deltas. [19:10:39] yay vacations :D [19:10:49] riding a bicycle in sanf rancisco would just make me more anxious, i'd die ;) [19:10:54] ok let's try this update [19:10:58] that is what Rob told me [19:11:20] <^demon> I need a vacation. 57 days \o/ [19:11:24] he said that bicycling just after landing at SF and being jet lagged is surely a bad idea [19:11:45] so I will skip bicycling entirely :-((( [19:12:20] Sunday biking in GG park is awesome, though, and there's bike rental stations all over [19:12:26] ^demon: I tried to do submodule init/update on core master, and nothing happened. What am I doing wrong? [19:12:44] <^demon> vvv: Wrong branch, submodules are on wmf/1.20wmf1 [19:12:56] No submodules on master? [19:12:59] RoanKattouw, ok 047543b6805a268c8d689a7a1ce12ec545ef79a9 [19:13:01] er [19:13:02] <^demon> No, not for extensions. [19:13:04] robla: is GG to SF what Central Park is to NY ? [19:13:07] https://gerrit.wikimedia.org/r/#change,5823 [19:13:28] hashar, pretty exactly. city planners in the 19th century were like "what does NY have that we don't? A BIG RECTANGULAR PARK" [19:13:34] but it's not in the middle, it's off to the side :) [19:13:35] <^demon> vvv: Most people cloning core won't want to clone all extensions too. Once we update to 2.3 (next week!), we'll have mediawiki/extensions.git setup properly to clone all. [19:13:36] ahah [19:13:49] my wife is a city planner :-]]] [19:14:00] but being in Europe, you have to be much more creative! [19:14:09] fun fact: golden gate park was originally sand dunes. all the plants were imported, even those that are native california species [19:14:09] <^demon> Although, putting the extensions we're "bundling" in the tarball as submodules might make sense. [19:14:12] hashar: They tried to get the guy that did Central Park [19:14:15] He said it couldn't be done [19:14:20] Oh yeah, European cities are creative [19:14:52] anyway we will see how I survive my first day in SF [19:15:10] <^demon> Hopefully less eye-opening than my first day in SF. [19:15:24] ? [19:15:31] ^demon: do tell [19:15:40] <^demon> I showed up the weekend of "Up Your Alley" and some friends took me to it. [19:15:45] <^demon> It's a smaller Folsom St. Fair [19:15:52] hah [19:16:22] given I land at 1pm or 5am (france time). I think the first afternoon will be an awful experience [19:16:25] <^demon> This was...2-3 hours after landing in SFO. [19:18:07] well.... that's the san francisco experience ^demon [19:18:48] <^demon> I can't wait to move :) [19:18:56] toward or away ;) [19:19:35] r0an looks busy w/ something else, i'll ping him in a few minutes :) [19:19:49] Guess what is that the first floor of the hotel? « Café de la Presse », so French [19:20:04] might be able to have coffee and croissant there [19:20:07] Yeah you're close to some French restaurant [19:20:56] I wanted to taste american food and then someone told me there was none per see :-( [19:21:01] <^demon> brion: Toward. I need to get the hell off this coast. [19:21:24] everyone seems to want out of virginia these days [19:22:02] i was sure i had some question for mdale but i have forgotten it. d'oh! [19:22:18] I will be around if you remember ;) [19:22:23] whee [19:22:55] hashar: On the other side of the hotel there's Rouge & Blanc, which is also a breakfast place (annex wine bar) [19:23:36] * robla points out that that's not American food [19:24:00] hashar: Mo's burgers in North Beach is awesome [19:24:11] <^demon> I usually end up at Rouge & Blanc :p [19:24:37] i never make it up to north beach, they need a train line up there [19:24:53] willt he chinatown line eventually get nearish there? [19:25:05] also close by is a pizza by the slice place that's pretty great .... [19:25:10] * robla tries to remember the name [19:25:20] <^demon> For breakfast/brunch, I highly recommend Cafe Greco. I walk up there from the Triton for breakfast before heading to the office some days. [19:25:23] I will use the bicycle to get there :-D [19:25:36] and you're not too far from chinatown either, there's some really good food if you know where to hit [19:25:36] <^demon> It's on Columbus I think. [19:25:46] which i don't cause i hardly ever go there ;) [19:25:59] not american food though ;) [19:26:04] but at least it's different eh [19:26:10] \o/ [19:26:39] robla: Oh you mean that pizza place on Kearny near Bush [19:26:41] ? [19:27:10] guess I will ask for recommendations each days [19:27:19] * robla looks it up [19:28:14] it's on Green near Grant. Golden Boy Pizza [19:28:38] http://www.goldenboypizza.com/ [19:28:39] RoanKattouw, when you have a chance: https://gerrit.wikimedia.org/r/#change,5823 ready for merge/deploy [19:29:10] I will skip pizza entirely unless they are really made of fresh food [19:29:10] brion: Merged it just now, deploying [19:29:10] ok...now I'm hungry [19:29:22] \o/ thanks RoanKattouw [19:29:30]