[01:31:22] TimStarling: I don't know if we ended up backporting PHP > 7.2 things yet (considering master/wmf still php 7.2, we plan to break compat later in the 35.x cycle though). [01:31:35] so might work to upgrade on php 7.2 first [01:31:47] and then do the php upgrade later, to avoid too many changes at once [01:32:09] "unofficiall" [01:32:15] I tried suppressing MW's complaints about PHP 7.2, the next problem was that it was using a version of DBAL that depended on 7.3 [01:32:33] maybe that doesn't matter for WMF if we're not running the updater [01:36:20] that's been of a PITA on branches, as they rename classes in various ways, making it not so easy to support multiple versions of DBAL [01:36:52] ah, didn't realize dbal was 7.3+ [01:36:53] interesting. [01:37:11] I guess the gate has been opened then [01:37:33] There's a 7.2 version [01:38:08] well, a ^7.1 version [01:43:46] or let me try again... in master we support multiple versions of dbal (because of 7.2 in wmf production) [01:43:59] in rel1_35/1_36 we dropped that support for the older version to reduce complexity [17:57:00] Reedy, James_F, TimStarling: did any of you help produce a list of developers eligible to vote in the 2017 board election? Folks are asking questions on how to do that for the upcoming board election and andre_ and I have questions. :) [17:57:21] I figured that it would have been one of the three of you or Chad that did it before [17:57:54] I didn't. [17:58:00] Probably it was Chad. [17:58:42] https://gerrit.wikimedia.org/r/c/mediawiki/extensions/SecurePoll/+/350371 looks like it was James A [17:59:17] oh, list of developers [18:00:08] legoktm: yeah, it's the "developers" bit here -- https://meta.wikimedia.org/wiki/Wikimedia_Foundation_elections/2017/Board_of_Trustees#Requirements [18:00:45] which feels like a set of criteria from the svn days [18:01:07] mapping Developer accounts to SUL accounts is one non-trivial part of the problem [18:02:16] "commit access" is definitely SVN [18:02:21] how is "Wikimedia Foundation utilized repo" defined? :P [18:02:37] these are the questions we are asking Majavah :) [18:03:38] We are also pointing out that this has a particular bias for MediaWiki deve work over all the other kinds of off-wiki tech work that our technical volunteers do [18:06:46] if it's strictly "a repository deployed on the production cluster" I don't think I meet that criteria, all of my commit access is on wmcs things and I'm just a normal contributor on mediawiki and related [18:07:00] The task at https://phabricator.wikimedia.org/T163854 makes me wonder if they even bothered with developer rules last time [18:07:59] https://phabricator.wikimedia.org/T271884#6798568 suggested people were manually added on request [18:08:02] Majavah: yup. I used Magnus as the canonical example in the meeting we just had, but you are likely to fall into the same sort of thing. Lots of tech work to support the movement, but not by adding to MW [18:08:31] bd808: I do contribute to mw fairly often, I just don't have commit access [18:09:03] if "commit access" == +2 rights, but that's an assumption [18:09:10] yeah, meant +2 [18:09:14] "commit access" means nothing tangible to me [20:30:30] > changing $http->runMulti( $reqs ); to $val = $http->runMulti( $reqs ); sends the purge request to varnish else nothing gets sent. [20:30:36] well that's weird. https://phabricator.wikimedia.org/T279948 [20:30:56] doesn't look like there's a scoped callback or destructor involved there afaik? [20:31:18] maybe some unintelligent optimizer at work?