[00:25:46] gn8 folks [04:10:34] !log on hume: running purgeList.php to purge all pages on uzwiki from squid [04:10:45] Logged the message, Master [04:12:14] Susan: ^ [16:11:17] dumps.wikimedia.org is slow, only 40 Kbytes/sec :-( [16:21:27] Dispenser: from where? if TS, maybe someone else is downloading? [16:22:47] home connection (and TS has insane bandwidth often over 40 MByte/sec), http://dumps.wikimedia.org/other/poty/2011/ [16:53:24] Dispenser: I'll have a look at that in a little while [17:29:37] Dispenser: still there? [17:29:44] yes [17:30:04] I'm pulling 44 KB/s at the moment [17:30:13] what are you using to do it? [17:30:21] Firefox [17:30:30] hi guys [17:30:34] is there something or somebody i should poke about fixed bugs that really need a point-release? [17:30:44] right now i'm thinking about what is described in second paragraph of https://bugzilla.wikimedia.org/show_bug.cgi?id=42452#c41 [17:31:02] I'd ask in the dev channel (or wait for somene more dev-ish to pop up) [17:31:05] (in this one) [17:31:52] asked -dev already [17:32:09] they are your best bet I guess [17:32:18] nobody replied, and there are more people here :P [17:32:30] i'm hoping to get an answer today ;) [17:32:41] Reedy: ^ ? [17:32:50] Krinkle|detached: ^ (well, krinkle not here) [17:33:23] Dispenser: we rate limit but not that low by any means [17:33:53] giv me a url? I wanna test sometjhing [17:34:43] Dispenser: [17:34:50] http://dumps.wikimedia.org/other/poty/2011/poty-2011-media.tar.gz [17:34:55] thanks [17:36:18] hm bad from within the network, atrocious in fact [17:47:52] It's improving ~90 KB/s [17:53:59] Dispenser: https://bugzilla.wikimedia.org/show_bug.cgi?id=43647 [17:56:27] I downloaded poty 2008 and 2009 just fine on Dec 22, so it's recent [18:01:24] so what's interesting is that some people still have decent speeds when I look at traffic on various connections there [18:01:34] can you give me a traceroute to dataset2? [18:01:36] pastebin [18:04:02] Dispenser: [18:08:20] 40 KB/s here, let's do one [18:08:58] apergos: http://p.defau.lt/?oEVYSgRZXWFYIWiLRwBJfA [18:09:10] thanks [18:09:24] apergos: http://pastebin.com/swD7sqLZ [18:09:48] apergos: same on TS http://p.defau.lt/?3wIWRchB1wroXaewEbOBGA [18:09:57] apergos: 30-40K/s here too [18:09:58] er not 1001 [18:10:02] I want to dataset2 [18:10:03] please [18:10:35] I included both [18:10:48] ah, off my screen [18:11:47] apergos: same even from labs http://p.defau.lt/?ZrZTsVn4DZPD_o650iPKAA [18:12:01] apergos: not a network issue. [18:12:13] apergos: no lost packets or retransmits [18:12:16] apergos: there was a spike of ZIP downloads probably in the last few days, but it should be over as es.wiki removed the sitenotice [18:13:33] apergos: lvm has 100% i/o utilization [18:13:53] the disk util has been uniformly hight for the past week [18:13:57] checked atop [18:14:08] for 14 days ago [18:14:13] (before the issue cropped up) [18:14:19] it's pretty comparable [18:16:24] es.wiki sitenotice was from 2012-12-28 to 2013-01-04 [18:16:34] dunno if it matters, just saying [18:17:02] apergos: where do the dumps come from? lvm? [18:17:21] dataset2 [18:18:01] there are people downloading over there with 4 5 Mb/sec [18:19:12] I told you, it's not a network issue [18:19:25] hm [18:19:31] the file is being read fine though. [18:21:48] apergos: anyway, I'll leave you up to it. it's reproducible here though, so if you need help I'd be happy to be your guinea pig [18:21:55] thanks [18:22:25] but I didn't see any retransmissions, so I don't think it's network-related [18:24:17] MatmaRex: Yes, "soon" ish [18:25:24] alright, thanks [18:25:53] hashar: thanks for enabling pep8 on eventlogging [18:25:56] ! [18:25:56] ok, you just typed an exclamation mark with no meaning in the channel, good job. If you want to see a list of all keys, check !botbrain [18:26:21] ori-l: you are welcome :-] I can make the pep8 a blocker (aka vote verified -1 on lint failure) [18:26:27] heh. a little aggressive of wm-bot. [18:26:32] someone _please_ shutup wm-bot [18:26:33] hashar: yes, that would be good! [18:28:05] ! | Nemo_bis [18:28:06] Nemo_bis: ok, you just typed an exclamation mark with no meaning in the channel, good job. If you want to see a list of all keys, check !botbrain [18:29:28] ! | wm-bot [18:29:28] wm-bot: ok, you just typed an exclamation mark with no meaning in the channel, good job. If you want to see a list of all keys, check !botbrain [18:29:42] ori-l: https://gerrit.wikimedia.org/r/42583 will make pep8 to vote -1. Will deploy that later tonight [18:29:46] * hashar waves [18:33:24] Reedy: mid-air collision, see https://www.mediawiki.org/w/index.php?title=Version_lifecycle&diff=625191&oldid=611383 [18:33:39] «Since Debian (the Linux distribution with the longest release cycle) has a two year cycle between each freeze and we've gotten MediaWiki 1.19 into Wheezy, we'll support MW 1.19 for the next two years.» [18:34:38] "support" here means very serious bugs and security updates [18:34:57] yes, what else could it mean? [18:35:15] normal bugs? [18:35:27] anyone thinking that? [18:35:31] it was about https://bugzilla.wikimedia.org/show_bug.cgi?id=43715 btw [19:05:59] * robla assumes deployment is proceeding apace. [19:06:06] I'll be off in a meeting [19:06:50] Yeah, I'm in no rush [19:06:53] just finishing eating [19:26:15] aude: what? [19:27:00] 17:30 < MatmaRex> is there something or somebody i should poke about fixed bugs that really need a point-release? [19:27:12] 17:30 < MatmaRex> right now i'm thinking about what is described in second paragraph of https://bugzilla.wikimedia.org/show_bug.cgi?id=42452#c41 [19:27:23] I'm busy right now, but what is the problem? The bug is fixed. [19:27:25] you might be able to advise or help MatmaRex ? [19:27:40] maybe have that backported or something? [19:27:44] * aude not sure [19:28:08] right [19:28:21] I'm backporting now (if not already). I don't do releases though [19:28:29] ok [19:28:29] Vector extension doesn't need release, just a branch update. [19:28:35] MediaWiki core could use a release for that indeed. [19:28:40] ok [19:29:38] Krinkle: okay, so i can assume this is being taken care of? :) [19:29:45] Yes [19:29:50] awesome. thanks [19:29:55] working on it now [19:38:29] MatmaRex: Hm.. [19:38:36] MatmaRex: The fix doesn't apply cleanly [19:38:45] MatmaRex: Because the change it is fixing isn't in REL1_20 [19:39:21] MatmaRex: Vector was branched at the same time as core [19:39:38] MatmaRex: Are you sure this is an issue on third-party wikis running 1.20? [19:41:40] Krinkle: well, this guy reported that it was happening, and reported that the patch fixed it for him... [19:42:00] Krinkle: maybe the tarball contains something weird? [19:42:50] * MatmaRex goes to check [19:42:57] https://gerrit.wikimedia.org/r/#/c/30361/ is the core change [19:43:01] it was not backported to 1.20 [19:43:11] (unless someone did not follow protocol) [19:43:32] ah, no [19:43:38] the core change isn;t in 1.20 [19:43:47] but apparently the half-broken vector changes are included [19:44:04] there were three or four patchsets for vector [19:44:15] first to make it match the new code in core master [19:44:17] In the extension yes, but only one in core. [19:44:21] MatmaRex: No [19:44:24] then to make it also work with 1.20 and older [19:44:26] MatmaRex: The core change isn't broken, it is a feature [19:44:42] MatmaRex: https://gerrit.wikimedia.org/r/#/c/30361/ is the change that changes stuff from h5 to h3 [19:44:59] Krinkle: yes, and it's only on master, not on 1.20, and this is good [19:45:00] that isn't in 1.20, and neither are the bugs that core change caused in the Vector extension. [19:45:10] so there is no problem [19:45:25] MatmaRex: I'm reading the bug but I don't see anyone say there is an issue in the 1.20 tarball? [19:45:38] D: [19:45:41] wait, let me recheck things [19:46:25] i'm downloading the 1.20.2 release right now to see what's inside [19:46:31] ok [19:46:41] Krinkle: https://bugzilla.wikimedia.org/show_bug.cgi?id=43227 states that it broke after the 1.20.2 upgrade [19:46:58] Yeah, I just got there to [19:47:02] and that it was fixed by my latest patch, https://gerrit.wikimedia.org/r/#q,I0a05ed39,n,z [19:47:24] | * 4cbb313 - (bug 43227) fix noncollapsible sidebar due to HTML changes (2 weeks ago) [19:47:25] | * 685027d - Making previous fix compat with CSS style convention and adding comment (6 weeks ago) [19:47:25] * 6b9c6ff - temporary fix for bug 42452 - adding forward compatibility (6 weeks ago) [19:47:26] | * 987d51a - (bug 42354) Make collapsibleNav work with h3 headings introduced in I9a2ebd50 (7 weeks ago) [19:47:30] i only want this patch backported to the vector extension version that will be included with 1.20.3 release. [19:47:32] Those are the patches for the Vector extension. [19:47:40] Etiher all of them or none. [19:48:02] both ways will do, probably. [19:48:14] MatmaRex: But https://gerrit.wikimedia.org/r/#/c/30361/ (core change) isn't in REL1_20, so this doesn't make sense. [19:48:28] I'm not backporting fixes that fix a bug introduced in a change that doesn't exit. [19:48:29] argh, why is my connection so slow [19:48:29] exist* [19:48:42] i'm at 40% of the tarball download... [19:49:14] MatmaRex: Might not be your connection [19:49:15] Krinkle: some of the "fixes" for that core change made to Vector were incompatible with pre-that-core-change core [19:49:18] Just git-checkout REL1_20 ? [19:49:24] Reedy: no, it is my connection. [19:49:47] 1mbps shared on three computers, yay. [19:50:01] MatmaRex: Maybe, but if so then it is even more important that we don't backport them unless the h5-to-h3 change made it in 1.20 [19:50:16] MatmaRex: I'm getting less than 50 KB/s from download.wm.o [19:50:17] we want backwards compatibility in vector? [19:50:26] ^ apergos Did you see that bug about it being reported as slow? [19:50:39] aude: If you use 1.20 there is no point in having backwards compatibility fixes for 1.21alpha [19:50:47] yes, though I am stumped at the moment [19:51:15] Krinkle: so there's a 1.20 release for vector? [19:51:22] ( Reedy ) [19:51:27] sorry, mostly listening to meeting [19:51:28] * aude only uses extensions from git [19:51:37] Krinkle: both pre-https://gerrit.wikimedia.org/r/#/c/35817/ Vector and current master Vector should be good for 1.20 [19:51:44] aude: There is no "release" there is a branch, because 1.20 has bundled extensions. [19:51:58] Krinkle: if you have a sec, can you take a peek at https://gerrit.wikimedia.org/r/#/c/42329/ ? If I were able to get that merged and backported to the deployment branch I'd be able to avoid copy/pasting code in two places. [19:52:44] Krinkle: re `git checkout`, i want to make sure i'm looking at what's actually in the tarball [19:52:44] Krinkle: ok [19:53:07] Krinkle: and there are way too many branches and tags on all those repos [19:53:21] MatmaRex: what do you mean? [19:53:22] ah Reedy while you are in here, are you the one who knows the guts of the interwiki cdb file? [19:53:27] MatmaRex: (re: number of branches)_ [19:53:41] and to make matters worse, i currently have both origin and gerrit remotes [19:53:41] cause I have docs that someone who knows it should read and kick me about where they are wrong [19:53:48] MatmaRex: I can confirm https://gerrit.wikimedia.org/r/#/c/30361/ is not in core REL1_20 [19:53:56] I saw that bug.. [19:53:59] REL1_20 does not have the new headings [19:54:04] so there shouldn't be a problem [19:54:09] I"ll wait for you to confirm the tarball as well [19:54:19] 90%... [19:54:32] um not the bug, actual docs [19:54:41] although yes the bug too :-P [19:56:21] http://www.mediawiki.org/wiki/Interwiki_cache/Setup_for_your_own_wiki [19:56:46] it is so not urgent. I just want to get the right person's eyeballs on it [19:56:58] I saw it and wtf'd at the dupe keys [19:57:05] heh [19:57:20] Krinkle: eh, it looks like 1.20.2 has only h5s everywhere, in core and in E:Vector [19:57:26] see it doesn't pay to look too close at anything [19:57:29] like sausage... [19:57:36] Krinkle: so i have no idea what bug https://bugzilla.wikimedia.org/show_bug.cgi?id=43227 was about [19:57:59] Krinkle: the person might have upgraded to newest E:Vector [19:58:05] MatmaRex: It had become about 1.21 (ie. wmf wikis), which was a genuine bug [19:58:19] but the reporter's situation on his 1.18>1.20.2 wiki, makes no sense. [19:58:19] ah, yes [19:58:25] "our copy of the Vector extension is from git-Head" [19:58:35] so MW core 1.20.2 release, E:Vector git master [19:58:46] i should have read it more carefully first [20:00:01] Krinkle: so in fact it seems like there is nothing left to fix here right now [20:00:28] * MatmaRex self-flagellates [20:00:34] note to self: less writing, more reading. [20:01:31] Krinkle: i'll summarize this on both bugs. is this okay? [20:01:45] MatmaRex: I'm almost done writing that muself [20:02:19] one [20:02:23] alright, i'll get out of the way then :) [20:02:24] myself* done* [20:06:25] Krinkle: i expanded on your comment on https://bugzilla.wikimedia.org/show_bug.cgi?id=43227#c16 [20:07:13] ok [20:20:52] did we ever fix SSL certs for some of the mobile sites? I notice that https://commons.m.wikimedia.org/ spits out a warning that it's only good for *.m.wikipedia.org [20:27:56] brion: no [20:28:04] nnnnnnooooooooooooooooo [20:28:30] well we're probably going to send more people to commons on mobile in the future so we might want to work that out [20:28:30] brion: ply RobH with beer now that he's in the office [20:28:31] we don't have enough IP addresses to do so [20:28:33] Then make him buy it ;) [20:28:34] will do [20:28:48] naw i don't think he's got a company card anymore ;) [20:30:50] brion: thanks for your blog post about FirefoxOS :-] [20:31:13] brion: I didn't even know about that phone OS before reading your blog. [20:31:39] yeah it's got some potential, but it's still very early [20:31:47] they're working on it like *crazy* to hit their v1 production deadlines :) [20:32:51] brion: I emailed a link to your FirefoxOS post to my friend Christie who works on FirefoxOS dev relations [20:33:01] in case she talked to you later or something [20:33:11] not yet but i'll keep an eye out :) [20:33:25] that reminds me -- i have to check for the firefox market credentials so i can update our app [20:33:41] (runs on firefox aurora or nightly for android too) [20:34:44] anyone know about this ubuntu phone os thing they just announced? [20:34:58] i'm curious but they've got no hardware partners which is… not impressive yet [20:35:02] yeah [20:35:12] All I heard is people talking about the toolkit - Qt or GTK, I forget [20:35:27] Qt, which smells very non-Canonical to me :) [20:35:44] but probably wise, as BlackBerry 10 and the remains of Meego are both Qt [20:46:27] o/ brion [20:47:17] howdy