Chris Fallin

For what it's worth, for previous large Cranelift features where we wrote RFCs (e.g., ISLE, or the mid-end egraphs framework) the RFC was more of a snapshot of proposed design...

Hi @Iizerd -- I'd still like to see the fuzzing for this before we merge it, to be sure we've got it right -- any update on this? > As...

@sparker-arm Is there any feature flag or other way to detect whether the underlying hardware does a partial write or not (other than, say, the runtime performing one and catching...

@sparker-arm one other thought: is it sufficient, architecturally, to do a load (and throw away the result) to a stored-to address before every store? We would then take the fault...

Thinking about the load-before-store idea a bit more: a page fault is a precise exception, in the sense that all side-effects of earlier instructions should occur, and none of later...

> Even if this does work, I don't really think we should be coming up with hacks to get WebAssembly to run well on a very prominent architecture. It seems...

Strong +1 to this -- this is the right way to represent idempotent side-effects IMHO. (Doing it the simpler way was a matter of expediency on my part as I...

@SoniEx2 could you clarify which part you'd like to be improved? The type mismatch itself is roughly the same in both ("expected i32" vs. "expected [i32]", "nothing on stack" vs....

I think we still want to use the domtree traversal for other reasons (or more precisely: my head hurts trying to think about how a computed-on-demand traversal order would interact...

Ah, I had recalled you having thoughts about integrating constant-phis with the egraph pass, and how that might work with a fixpoint (or in the cheap version, no fixpoint, only...