Daniel Thornburgh

I think we need to fix #443 before doing this. We also need to have a strategy for either distributing both x86 and ARM versions of the SDK for Mac,...

> Ideally, I want my ISR to only save what is transitively calculated from other function calls that I make use of, and the compiler has that information. At least...

Yep ^ this. I'll be looking into adding a proper error message (or more optimistically, making it so that it works correctly), after I finish this week's SDK release.

> Modern C uses a macro that calls a function for errno. If we did this, all the error checking in the standard library can be skipped and apps that...

This will regress with the latest pull from upstream, but I'm not particularly inclined to keep messing with LLVM's register allocator, since I'm in the middle of writing a new...

Since this relanded, I've been getting pre-commit breakages on a libc change: https://buildkite.com/llvm-project/github-pull-requests/builds/120029#01933169-a0ae-443c-9de3-0827c512a223 I can't find anything in the logs that indicates that a build or test failure occurred. Does...

Thanks for the quick fix; my PR is green now!

This should be a fairly straightforward change to the legalizer; should be a really nice first project. I'm not likely to get to it myself soon, but I can definitely...

Generally agree w @johnwbyrd. If we had detailed steps about how to set up a build environment on e.g. Debian, then from what I've seen it's customary to include `build-essentials`....

> I agree. One small side note - perhaps the more appropriate behavior for a C= program would be to jump to the reset vector, as opposed to infinitely looping?...