[14:30:12] tgr|away: did you mean 'this morning' instead of 'tomorrow' in T379589 ? [14:30:13] T379589: scap backport fails at purgeMessageBlobStore.php with getaddrinfo failed - https://phabricator.wikimedia.org/T379589 [14:43:06] Krinkle: thanks, I meant yesterday [14:44:22] https://sal.toolforge.org/production?p=0&q=tgr+scap&d=2024-11-11 - two successful scaps, then got this the third time [14:45:21] even today, it seems most backports went fine [14:46:29] right. I guess it coincidences with gadget definitinos being in memcache or not [14:46:48] when not, it'll try to fetch 'em from the database [14:47:01] in order to know which ext.gadget.* modules exist and thus which modules to purge [14:47:26] It's not meant to depend on knowing each module, though. [14:47:47] There's a singular global shared key for touching the cache [14:48:26] but through the bless of dependency injection, that 1 line method now depends on everything from MBS and in turn everything from RL in order to discover that service. [14:49:03] I fixed this for the installer by making clearGlobalCacheEntry a static function. [14:50:05] But this maintenance script isn't using that. That has the benefit of not hardcoding which WANObjectCache instance to pass but probably worth trading that for depending on services->getResourceLoader()->getMessageBlobStore() because that will, in prod with Gadgets installed, depend on the database to register additional modules. [14:50:20] * Krinkle adds comment to task [20:20:01] Hi, I'm wondering if someone could review/merge https://gerrit.wikimedia.org/r/c/mediawiki/extensions/TemplateStyles/+/1088709 and https://gerrit.wikimedia.org/r/c/mediawiki/extensions/CSS/+/1088710 please? [20:42:52] so HTMLFormFfield has a getLabel() and a getLabelHtml() method, where getLabelHtml is supposed to return raw HTML, while getLabel is... also supposed to return raw HTML?