Chandler Carruth

> We should consider accepting such cases as integer literals once Carbon is more established and the divergence from C and C++ is a less significant concern. Curious -- why...

Ah, I hadn't found that page when doing my upgrade. I instead mostly read the standing documentation. =/ Anyways, this all makes sense (and matches that I had to use...

@BlkPingu and @yanzhang0219 - Maybe take this to a separate issue or discussion? I'd like to keep this focused on improving the discoverability of the change in workflow here.

> > Is there a way to print this information from Homebrew? It might already be printed and I missed it while updating everything, which is entirely on me. >...

I tried to build a minimal reproduction, but while it produces a similar assembly sequence, the link step doesn't do the same relaxation I observe in the real world test...

As pointed out on Twitter, LD64 doesn't seem to assert that the `adrp` destination and `add` destination match, but it also doesn't update them to match, just folds the offset...

> @chandlerc Are you sure that the bug reproduces when linking with ld64.lld? It did with -fuse-ld=lld, but I can check later to see if this was actually still using...

Several folks on Twitter helped me learn about the LOH hints that need to be present for the linker behavior I saw to fire, and so took another shot at...

> > @chandlerc Are you sure that the bug reproduces when linking with ld64.lld? > > It did with -fuse-ld=lld, but I can check later to see if this was...

> > For me, the most convincing reason to change the spelling of `me` to `self` would be to support shorthand syntax `[self]` and `[addr self]`, which don't really make...