sykhro
sykhro
Nope, I'll remove them. I don't see why these are still failing. I can test on an intel machine in a couple days, it works fine on M1
> @sykhro, I took the liberty of rebasing this onto master and removing the now unused llvm@17 and ninja dependencies. > > I leave it in your hands to decide...
In theory, AppleClang + ninja should yield the maximum speed in terms of build time
> Ah, but llvm@17 is _not_ AppleClang, right? Correct, it's clang 17 from llvm.org. AppleClang (Xcode) is Apple's fork of clang, which has some tweaks for M1, Swift, and it's...
> Ok, it's a deal. We could also try installing ninja but not llvm@17 to see if the increase in build-speed offsets the dependency installation time. On second thought, not...
Yes, I have a fix - will send soon; opened the issue to avoid forgetting :)
I'm terribly sorry! I thought I sent this ages ago, I have rebased it on top of the `SDL2` branch; see #10909
The old behaviour can be restored by passing `-fcaret-diagnostics-max-lines=1 -fdiagnostics-show-line-numbers=0`
I'm happy to contribute but the project is a bit hard to nagivate. Could you point me to where options are passed to clang-tidy? For now I can at least...
> I think that according to https://blog.conan.io/2022/10/13/Different-flavors-Clang-compiler-Windows.html, the runtime can be compiler.runtime=dynamic/static, but it cannot use simultaneously compiler.libcxx=c++_shared definition, the libcxx for Windows-clang is used for msys2 backend, not a...