[00:52:57] <AaronSchulz>	 TimStarling: I amended https://gerrit.wikimedia.org/r/c/mediawiki/core/+/696631/11 a bit and left a comment
[01:07:16] <TimStarling>	 is this a postgres/sqlite support patch? 
[01:08:08] <TimStarling>	 I mean, the point of it is to avoid edge cases in the standard SQL implementation of upsert in Database::doUpsert(), right?
[02:54:25] <AaronSchulz>	 TimStarling: yes (aside from being easier to reason about generally)
[15:11:55] <Krinkle>	 > Amir1 T288639> AbstractContent::getParserOutput creates a new PO, renders it and send it back every time
[15:11:55] <Krinkle>	 I'm pretty sure this didn't used to be the case, or if it was there must've been a wrapper that was typically called instead? Maybe the recent refactor that moved stuff in ContentHandler caused this to no longer be cached.
[15:11:55] <stashbot>	 T288639: SpamBlacklistHooks::onEditFilterMergedContent causes every edit to be rendered twice - https://phabricator.wikimedia.org/T288639
[15:22:29] <Amir1>	 Krinkle: it can be, I habe no clue unfortunately (afk today btw)
[21:37:32] <AaronSchulz>	 Krinkle: maybe similar to flaggedrevs relying on the preparedEdit field of WikiPage that was removed
[22:41:55] <Krinkle>	 I'm in the call now.