[00:16:34] Krinkle, tgr: so what's the consensus? Revert or keep or ...? [00:18:21] matching call_user_func_array() doesn't make me sad (which is what the patch does) since that is what I would have used to make the feature if it worked for constructors [00:18:51] but I get the possible confusion argument that Krinkle is putting forth [00:19:50] I'd say revert. We're not using it (the example case was actually associative), and using it would lead to fragility/confusion. And it seems PHP 5.6+ omitted the tolerance for this in at least one feature. [00:20:48] ah tgr already has a patch up that reverts and adds a check for non-numeric keys -- https://gerrit.wikimedia.org/r/#/c/255052 [00:33:27] Did you have a role in selecting Mr. Scheier? From the wording, in your email, I'm inclined to think that you did not, but I want to get clarity on that. [00:34:01] duh, middle-click copy strikes again [00:34:12] I should disable that somehow [05:35:22] Thanks bd808, tgr|away :) [06:04:31] TIL: the POSIX standard provides guidance on how command-line utilities ought to name and document their parameters: http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap12.html [11:34:37] bd808: I updated the demo server for scholarships but I don't have permissions to access and update the database. I tinkered with it for a while but didn't work. https://dpaste.de/uR7u These are the files pending migration on that server. [13:39:51] Thanks for the work Niharika :) [13:39:57] I may submit some patch later btw [13:40:17] Nemo_bis: Thanks....but for what? [13:41:23] I have yet to decide! [13:42:14] Nemo_bis: Wait, which work are you thanking me for in the first place? :P [13:43:41] Anything scholarship-related [13:43:54] I know the schedule is tight [13:45:12] Nemo_bis: You need to thank bd808 for the awesome work he did, twice as fast and efficiently as me! :) [16:11:29] Niharika: looking at wikimania-scholarships.wikimania-support.eqiad.wmflabs now [16:19:28] Niharika: *grumble* this server is really out of date [16:19:37] not your fault obviously [16:20:10] bd808: But I did update it earlier today. Are we looking at the same server? [16:20:10] oh it's still a precise host... [16:20:28] Niharika: wikimania-scholarships.wikimania-support right? [16:21:04] I sshed to wikimania-scholarships.eqiad.wmflabs [16:21:14] *nod* [16:22:02] I'm messing with the labs-vagrant bits and then realized why what I did was wrong [16:23:05] * bd808 steps back slowly and starts over [16:33:00] Niharika: I got the db updated and I think I put back all the things I had kind of messed up in the labs-vagrant checkout [16:33:38] we should probably build a new test vm of this soon [16:34:00] the one we have is very very old and works a bit weird because of that [16:34:21] bd808: we could destroy and re build this one, right? [16:34:48] yeah. when we do we will lose the db unless we back it up first (which we should do) [16:35:22] It might be easiest to build a new vm in the project, copy over the db and then when everything is working kill the old one [16:35:34] Yep. Although this db is mostly empty except for the few test applications. [16:36:03] I'll put it on my todo list. [16:36:07] yeah and some accounts for you and me and sati [16:36:22] and kushal apparently too [16:37:52] Niharika: the .my.cnf to get into the db easily belongs to the vagrant user (sudo -- su vagrant). [16:40:10] I did that. Without the -- and still got access denied. [16:41:03] hmm.. I just used my knowledge of the horrible password for the mysql db and did it all on the mysql command line [16:41:24] mysql -u root --password=vagrant scholarships < whatever [16:44:23] Niharika: when you were trying to review https://gerrit.wikimedia.org/r/#/c/254784/ did you notice that it is in the wikimedia/slimapp project? It downloads ok for me. [16:45:12] bd808: I was in the master folder, so that shouldn't have been a problem? [16:45:38] I'll try it again by being that project folder to see if that does anything magical. [16:46:32] you need to be in a clone of slimapp to download it I think [18:25:40] AaronSchulz: is it possible to backport your fix for https://phabricator.wikimedia.org/T117332#1828826 (links tables not being updated) to wmf.7? [19:07:48] s [19:07:52] ah [19:49:02] bd808: anomie: looks like my browser decided this was enough video for today [19:49:24] *nod* we are going to wrap it up. [21:37:45] AaronSchulz: (and cc urandom) does UIDGenerator generate UUIDv1s? It looks like newTimestampedUID128 could be it, but I can't tell if it is exactly a UUIDv1 or something very similar. [21:45:29] no, there was a patch for that at point, though it didn't have a use [21:47:30] oh yeah, here we go [21:47:31] https://gerrit.wikimedia.org/r/#/c/56567/ [21:47:34] urandom: ^ [21:48:28] ah ha! [21:55:58] \O/ [21:56:13] urandom: sorry for suggesting to use UUIDGenerator instead of 3rd party lib [21:56:40] hashar: no no, i had concerns about pulling that in, in the first place [21:56:49] we can also go the over way around: use the 3rd party lib and phase out UUIDGeneator [21:57:01] but our class has some unique features I belibe [21:57:09] in the end i decided to ask for forgiveness rather than permission [21:57:21] at least that let you develop a prototype easily [21:57:52] maybe one day our UUIDGenerator will be made a standalone lib [21:58:13] https://git.wikimedia.org/tree/operations%2Fpuppet/5b7895dcd5b49b385f97e99438acf837f6a1a1d8/files%2Fmisc%2Fscripts [21:58:18] https://github.com/wikimedia/operations-puppet/tree/production/files/misc/scripts [21:58:23] gitblit is seriously out of date [21:58:23] anyway sleep time for me, quite happy to see one more person will dig in UUIDGeneator [21:58:38] AaronSchulz's gerrit here looks promising, but it looks like it adds dependencies too (gmp and bcmath)) [21:58:49] Reedy: varnish cache maybe? [21:59:03] Reedy: or the Gerrit replication is broken [21:59:23] hashar: thank you for taking the time to look at that extension, and for the helpful suggestions [21:59:36] https://git.wikimedia.org/log/operations%2Fpuppet.git [21:59:39] It seems up to date [22:00:14] urandom: thanks. You will need someone else to give a serious review about the actual code -:} [22:00:23] https://git.wikimedia.org/tree/operations%2Fpuppet/0ac1d89/files%2Fmisc%2Fscripts#foo [22:00:33] Yeah... Looks like it's something cached and it's pointing at an old url :) [22:00:49] Reedy: well your URL has commit 5b7895dcd5b49b385f97e99438acf837f6a1a1d8 which is from 2014 ... [22:00:54] Yeah [22:01:02] I've no idea where kaldari got that from [22:01:12] https://git.wikimedia.org/tree/operations%2Fpuppet/production/files%2Fmisc%2Fscripts [22:01:17] that one is fine :-} [22:01:21] Yeah [22:01:23] thanks [22:01:33] I think the default of gitblit is to expand refs to the sha1 [22:01:37] to give you permalinks [22:01:50] or kaldari got it from some old post/email [22:02:06] well [22:02:10] if you go the main page https://git.wikimedia.org/summary/operations%2Fpuppet.git [22:02:19] there is green label [production] [22:02:25] which uses the sha1 [22:02:27] urandom: bcmath is usually bundled on windows and in the debian/ubuntu distros. I think there other other distros were you have to install it though. [22:02:36] so yeah most probably some old link floating around [22:02:56] and yeah, that really should made into a library [22:03:49] one day we will want to build a nice website showing all our reusable libs [22:03:53] and start marketing them [22:20:26] UIDGenerator uses a ton of wf*() functions. Not completely trivial to convert to a lib today [22:30:40] wfBaseConvert is now a library, I just didn't replace the usages [22:31:11] gj legoktm [22:31:11] :D [22:32:41] I don't think that global function can go anywhere anytime soon [22:32:50] 5 WMF extensions use it [22:33:06] + 2 others [22:34:27] Do we need to add the composer dependancy to all the extensions to replace it? [22:34:41] It seems a bit of a grey area if core depends on it [22:35:50] if core depends on it, you can assume its there [22:37:26] We should be able to migrate most of core [22:37:34] Just leave the global wf and deprecate [22:38:58] should be an easy search and replace [22:39:29] Yeah, exactly [22:39:38] GCI task? [22:39:44] Or not worth the hassle? ;) [22:42:28] * Reedy jfdi [22:47:50] legoktm: the benchmarks are still in core [22:48:19] Reedy: yeah, that wasn't easily librarizable. [22:48:35] cdb.php is still in core too [23:08:53] $this->mDescription = "Benchmark for Wikimedia\base_convert."; [23:09:07] Should I have double backslashed? [23:41:33] Will PHP let you subtract skins? [23:41:38] $end = Wikimedia\base_convert( $block->getRangeEnd(), 16, 10 ); [23:41:38] $start = Wikimedia\base_convert( $block->getRangeStart(), 16, 10 ); [23:41:38] $size = log( $end - $start + 1, 2 ); [23:41:43] string - string +1?