Where do I go to have a productive conversation about the state of the rust trait solver?

116 points Zistack

It is deeply broken. I have to dodge compiler bugs (actually, I suspect it's one bug or a small number of closely related bugs) within my codebase on a regular basis. From what I have seen around, almost everybody who tries to implement generic types that support some level of generic arithmetic runs into these bugs.

I have two MWEs for compiler bugs now. I reported the first one a while ago. It appears to have almost completely ignored. I would report the new one, but I strongly suspect that it will be ignored like my first one (and all the other similar reports that seem to happen on a semi-regular basis (see my report - I link a few of them)).

I know that there's been some movement on re-implementing that subsystem of rustc, but I can't tell what the current state of all of that work is, or if anything has or ever will come of it. What happened to chalk? Was the re-implementation project I briefly saw on zulip/github at all related? Did anything ever come of either project?

I would actually have some useful things to say about the situation with trait coherence rules being a bit too restrictive (also an obstacle for me and anyone doing generic arithmetic on generic types, but a much smaller one) and what might be done about that if I weren't concerned that there isn't really anybody working on even the basic correctness of the system as it is defined currently.

I've looked at the zulip, but it mostly appears to be internal discussions about process and stuff. I can't tell where someone like me would go about asking about stuff like this there.