[PATCH] D134267: [C++] [Modules] Support one phase compilation model for named modules

[Chuanqi Xu via Phabricator via cfe-commits]

ChuanqiXu added a comment.

In D134267#3864416 <https://reviews.llvm.org/D134267#3864416>, @ruoso wrote:

> Currently, no. I think we do need a paper to discuss the requirements of the remote execution protocol and how they relate to the implementation of C++ modules.
>
> The simple summary is that one way that remote execution is implemented is by just wrapping the compiler execution, creating a Merkle tree with all the inputs, identify all outputs, ship that to a remote worker, and return another Merkle tree with the outputs.

Yeah, it is always better to have standard protocols. So the current state about the suffixes of modules (or `a differently named argument for output files` in your terms) is still need to be discussed. Personally, I agree with the special suffix for the sake of readability. @iains I think your draft patches may be better to be suspended until the SG15 get consensus. From my understanding, the problem may not be related with the client/server modes, right? So I guess this may not block your future works.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D134267/new/

https://reviews.llvm.org/D134267