klausler wrote:

> > A Fortran compiler and its runtime are tightly coupled; the compiler can 
> > only work with this runtime, this runtime can only be used by this 
> > compiler, and they share common API definitions and data structures.
> 
> So is c C/C++ compiler and its runtime(s). There is still a need to support 
> cross-compilation and other build modes like 
> `FLANG_EXPERIMENTAL_CUDA_RUNTIME`.
> 
> > I don't mind if you move the sources and headers that are used _only_ by 
> > the runtime into a new top-level directory, if you must. But moving common 
> > headers and sources used by _both_ the compiler and the runtime builds out 
> > of the compiler tree makes no sense to me.
> 
> As said, this was the result of the RFC. I can update the PR and leave them 
> in `flang/`.

I am not in disagreement with your objectives.  But I think that you will be 
able to achieve them without scrambling the source locations so much that I 
won't be able to continue maintaining them upstream.

https://github.com/llvm/llvm-project/pull/110298
_______________________________________________
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
  • [... Tom Eccles via llvm-branch-commits
  • [... Michael Kruse via llvm-branch-commits
  • [... Tom Eccles via llvm-branch-commits
  • [... Peter Klausler via llvm-branch-commits
  • [... Valentin Clement バレンタイン クレメン via llvm-branch-commits
  • [... Michael Kruse via llvm-branch-commits
  • [... Michael Kruse via llvm-branch-commits
  • [... Valentin Clement バレンタイン クレメン via llvm-branch-commits
  • [... Peter Klausler via llvm-branch-commits
  • [... Michael Kruse via llvm-branch-commits
  • [... Peter Klausler via llvm-branch-commits

Reply via email to