Daniel Thornburgh

It looks like this was built without the optimization profile used in the SDK. What do you get at `-Os`?

Unfortunately, getting this to work requires the detailed memory layout of the process to be emitted as markup. I don't think it's likely that you'd get N utility authors to...

Sorry, I had meant to imply (and should have stated) that I'd like to see a change to `backtrace_print` for this. Is the format output by that function part of...

This is probably one of those lovely instances of Apple triggering `if (DARWIN)` code that doesn't bother checking anything else. I don't know anything more than that though, like why...

I have no idea about any of this. LLVM's CMake is a wild and wooly beast, and it's difficult to model its behavior accurately in one's head. Unfortunately, that's about...

> Hmm, I‘ve also just noticed. In contrast to CD, none of the CI jobs build `llvm-mos` using `clang/cmake/caches/MOS.cmake`. Instead, they build it using a stripped down CMake config which...

> As of this writing, I don't see any indication that upstream builds on M1. Welllll, this kind of thing is usually somewhat implicit. If you break Apple's downstream M1...

While the compiler shouldn't ever crash, in this case it seems reasonable for it to emit an error, since the argument isn't a constant value, as required by `i`.

Hey retrac0, thanks for the suggestion! Without having tried it, it seems like the simplest way to do something like this would be to add an "address space" for cartridge...

@asiekierka , if you get around to this, take a look at isAddressSpaceSupersetOf in clang/AST/Type.h. I think that logic determines whether an implicit conversion between address spaces is attempted. EDIT:...