[15:53:10] ebernhardson: hey, can you help me out with a failing flow test? [15:53:12] https://integration.wikimedia.org/ci/job/mediawiki-extensions-hhvm-jessie/40450/console [15:53:21] it's failing for https://gerrit.wikimedia.org/r/c/405015/ [15:54:02] i don't have the foggiest idea what the offending code is even trying to do, nor in what way my patch changes the behavior [15:55:48] DanielK_WMDE_: maybe, lemme look [15:57:04] DanielK_WMDE_: that's a huge patch :P will take a bit [15:57:34] ebernhardson: assume that the path Does Nothing (tm). [15:57:59] as it should change nothing. it's "just refactoring" that gives me the freedom to ad MCR stuff. [15:58:21] ebernhardson: what I'm trying to understand is what change in behavior causes the test to fail. [15:58:45] what behavior in core is it relying on? [15:59:05] If you can tell me that, I can figure our how and why I my patch changes it :) [15:59:38] DanielK_WMDE_: no clue :P it's been about 3 or 4 years since i looked at this [16:00:07] right... also, the stack trace is... crazy. [16:00:58] git says Stephane Bisson write the test... is he on irc? [16:01:08] DanielK_WMDE_: i could dig you up some 500 function stack traces in java if it makes you feel better ;)_ [16:01:23] DanielK_WMDE_: or even better, babel cross-compiled javascript with all kinds of junk functions between the real function calls in the stack :P [16:01:30] nope, been there, done that... [16:02:04] DanielK_WMDE_: should be sbisson in #wikimedia-collaboration [16:02:21] but maybe not online right now? [16:02:25] thanks. do you know what time zone he's in? [16:02:52] DanielK_WMDE_: east coast canada [16:03:42] DanielK_WMDE_: at a high level, the test is explicitly saying 'dont set new values', but new values are being set. I'd probably have to set some breakpoints and step through before/after the patch to figure out whats going on [16:06:20] ebernhardson: yea, never mind, i was hoping you'd just know. [16:06:29] i'm poking stephane now [17:03:58] I'll look at the patch this afternoon