naveen-seth wrote:

I wouldn’t mind this either 🙂. Admittedly, my knowledge of module 
internals/serialization is still shallow, but if that is no issue with the 
other modules maintainers, I'd be happy to.

> > and this line of PRs to add dependency scanning to the clang driver was the 
> > main blocker for CMake C++ module support.
> 
> I didn't follow this. CMake already supports C++ modules. And I lack an 
> interesting for this since I feel the feature is majorly used for using 
> modules without a build system. But this is only common tutorial in my mind. 
> @naveen-seth could you clarify?

I also don’t remember this being a main blocker; my understanding is that the 
CMake-team and other stakeholders mainly wanted the `clang-scan-deps` 
functionality to be available directly from `clang` itself ([[Modules] 
Bi-Weekly Meetup Notes, Feb 17, 
2026](https://discourse.llvm.org/t/modules-bi-weekly-meetup-notes-feb-17-2026/89851)).
 

I believe this was the reasoning: 
https://github.com/llvm/llvm-project/issues/76726#issuecomment-1874571241
(cc: @mathstuf, @Bigcheese).

Moving `clang-scan-deps` into `clang` probably also relates to the [earlier 
comments](https://github.com/llvm/llvm-project/pull/152770#issuecomment-3344634967)
 about unifying more of the modules tooling.
I’d like to also follow up on this in the next modules meeting and can draft an 
RFC afterward if there’s still interest.

https://github.com/llvm/llvm-project/pull/152770
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to