Open henryiii opened 3 weeks ago
Thank you for this contribution, @henryiii.
Someone with a machine on macOS 13 should dive into this problem.
Note that we are only maintaining the distributions of mamba and micromamba on conda-forge, but we aren't against someone maintaining the distributions on Homebrew and upstreaming patches so that mamba and micromamba are portable or more systems.
Any suggestions or thoughts on https://github.com/mamba-org/mamba/issues/3495 ? The problem is that the GHA runs here use the latest LLVM compilers, and not the AppleClang compilers that come with macOS, and systems like homebrew use the AppleClang compilers. I think there are a couple of uses of C++ features that aren't supported by the older compilers. So far it looks pretty easily fixable, I think? Maybe just adding some explicit types instead of using the deduction guidelines.
There's also a linux issue there.
So far it looks pretty easily fixable, I think? Maybe just adding some explicit types instead of using the deduction guidelines.
Yes, definitely.
New GHA workflows can be set to test Homebrew's toolchains.
macOS 13 is a target for homebrew, and still an important target for GHA, as it's the last Intel build they provide. I think on macOS 13 there will be a missing deduction guide error, but will have to see.
FYI, mamba-org/setup-micromamba@v2 is broken, probably see #3576.
(Likewise, I don't think GCC 11 is supported, only 13+)
Versions don't match, this is reporting Clang 17, while the homebrew job is using AppleClang 15.