Closed cfis closed 1 year ago
Yeah, this all looks good, thanks for your work here! The renames I'm sure are fine and if there is someone who asks why the methods got renamed, we can re-add the older names as fallbacks. But yeah I doubt that is going to be an issue. And yeah 4.1 is a good version bump for this amount of changes.
I do now see random failure on windows without gc.stress for the instance tracking feature (not really sure its related to gc.stress but it wasn't happening before). Will take a look.
Ok, took a look at this and can sporadically replicate it. I've been wondering if instance tracking is a good idea or not, and this makes me think it is not. Going to play with this a bit more.
@jasonroelofs I think I'm at a good stopping point for a 4.1 release if that works for you. Tests pass, docs updated, changelog updated.
Note I called this 4.1, but there are a couple breaking changes so I could see some logic in calling it 5.0. Breaking changes are:
The fix is simple, and I would guess these methods are not used much.