Chris Fallin

To summarize what I think is a good approach with the help of a small Cranelift extension, based on discussion in the Cranelift meeting this morning: - We should have...

I agree, this would be ideal to check. I don't have time to add it currently; if you want to make an attempt, I'm happy to review though.

The same label can be attached to different vregs across different ranges of instruction indices, though these shouldn't overlap (i.e. the label should be in only one place at a...

That indeed seems invalid! It looks like [this loop](https://github.com/bytecodealliance/regalloc2/blob/3c84b6fe1efc56a1a4c9a2a51e8a6daf0aa845e1/src/fuzzing/func.rs#L362-L375) ensures that over a single instance of the loop, it generates non-contiguous ranges, but it picks an arbitrary label and that...

Unfortunately there's no way for Cranelift to know what regalloc-inserted moves/spills/reloads are "for"; even at the RA2 level, we have an abstraction of liveranges, and we know that the liverange...

@gkgoat1 if you're interested in working on this, I'd be happy to review the work! The main design question is how this manifests in the IR; we're addressing the parallel...

I don't think the core issue is actually resolved at all; and in the linked PR it mentions that it resolved a subpart of the problem only. This could still...

HI @whitequark -- thank you so much for starting to look into this! Wasm EH will be an important feature to support for a bunch of reasons and I'm happy...

> we don't intend to ever unwind native frames, right? I think that's probably a good starting point (though this along with all the other details is up for discussion!)....

Exception handling support has now been merged in #11326 -- closing as completed.