mirror of
https://github.com/Bubberstation/Bubberstation.git
synced 2026-01-14 02:43:16 +00:00
3536d0df62
## About The Pull Request Fixes #75251 Fixes #75376 Fixes #75377 Fixes #75394 uh... I think? Here's the thing, is that golem arms are not actually particularly different to other kinds of arms, so I have been struggling to find a reason that this test is flaking _only_ for golem arms, in relation to a list that I didn't touch. My best guess is that the reason they're sometimes sticking to the "available hands" list is related to golems not being dismemberable and thereby skipping the code path which would remove a removed arm from the hand list, which sounds plausible except that nightmares are also NODISMEMBER and don't have the same problem. Regardless, I made some boilerplate for when you give someone an arm which registers to the destruction and dismember signals and removes them from your available hand list when either of those things happen under any circumstances, whereas before it was reliant on passing through a much more specific pipeline. Notably the text which set this also remarked `We only want to do this if the limb being removed is the active hand part.` which certainly does not sound true to me (although it does for the other line scoped inside that if check) and now also isn't, but this didn't seem to be widely affecting arms in general anyway. I couldn't replicate the test failing after making these changes but as a flaky result that doesn't really tell us very much because to some extent it might just have been down to luck. ## Why It's Good For The Game Memory management, hopefully will stop randomly triggering CI failures. ## Changelog Not player facing