[00:05:08] Krinkle: https://gerrit.wikimedia.org/r/#/c/mediawiki/core/+/555601/4/includes/specialpage/LoginSignupSpecialPage.php [00:05:28] the extra code "used to create empty line to separate help msg" definitely makes a difference when testing [00:05:50] But surely that should be doable using CSS alone? [00:10:43] Reedy: indeed. Or not even CSS, presumably htere's an option you can set on the previous or next item to give it a block if needed. I'd recommend asking for screenshots of current, proposed without this "empty line" and with, and go from there. maybe CC ppl who worked on this before as well. Not sure who. maybe Volker or Bartosz. [00:10:54] I'm not very up to date on HTML/vform/ooui stuff. [00:11:01] HTMLForm* [00:11:14] I did test it, and it definitely looks better with the extra linebreak thing [00:11:19] It just definitely feels the wrong way to do it [00:11:35] MatmaRex: About? ;P [00:11:52] * Reedy makes some screenshots [00:12:06] kind of [00:13:58] Reedy: Krinkle: there aren't many options for changing the layout of HTMLForm, you'll probably have to write a line or two of CSS for this [00:14:11] (assuming the empty line there is needed at all) [00:14:40] https://phabricator.wikimedia.org/T122124#5734227 [00:14:52] I think it looks better with the extra spacing [00:14:57] It's too close together otherwise [00:15:51] yeah [00:16:19] oh also this is probably not needed in OOUIHTMLForm. but the signup form still uses the old stuff [00:16:36] Yeah, we really should finish that. [00:16:46] And by "we" I mean MatmaRex doing all the work, as always. ;-) [00:16:51] something something change something something [00:16:56] James_F: It takes two to merge a patch ;D [00:17:15] i wrote a checklist about this [00:17:16] in 2017 [00:17:23] my work here is done [00:17:24] Reedy: Yes, the main thing is there's lots of fiddly crap added to CreateAccount that no-one knows who owns. [00:17:53] Does it matter if we throw away X? Does it matter if Y is three pixels narrower? Who knows!? [00:19:44] The Wikimedians will know you change something [00:20:13] Yeah, them I'm not worried about. [00:20:15] it's mostly captcha and mobile stuff that requires some more thought [00:20:20] Yeah. [00:20:27] last comment on https://gerrit.wikimedia.org/r/c/mediawiki/core/+/284902 has the current state [00:20:29] as of 2016 [00:20:41] lol [00:21:07] anyway i'm not picking this up right now. feel free to amend the patch [00:28:33] * James_F grins [02:10:04] TimStarling: (re: php-memcached) I suppose you've ruled out "unserialize", igbinary and/or __wakeup()? [02:10:51] no worries if not, I'll look into that tomorrow if not. [05:03:27] good point, no I haven't ruled them out [06:38:12] <_joe_> igbinary is supposed to be the way to go for php-memcached, but I've never tested performance [13:58:46] MaxSem: im missing something, but where is assertIsArray ? [13:58:53] relating to https://gerrit.wikimedia.org/r/#/c/mediawiki/extensions/Wikibase/+/556666/ ? [14:07:46] https://github.com/wikimedia/mediawiki/search?q=assertIsArray&unscoped_q=assertIsArray [14:07:51] Somewhere in phpunit I guess [15:30:38] I'm just confused because that patch got merged and CI said yes etc, but as far as I could tell in the phpunit docs for the version we are using in CI that method didnt exist yet :D [15:30:49] probably some bad docs or magic somewhere though [15:34:40] It's in my 7.5.17 phar of phpunit [15:48:58] re T239724, has anyone wondered why such an apparently cheap request got a timeout? [15:48:58] T239724: Fatal error: "Object does not support method calls" (from MemcachedPeclBagOStuff) - https://phabricator.wikimedia.org/T239724 [16:02:35] if unserialize failed due to an exception, there should have been a warning logged