[PATCH] D113676: WIP: [clang][lex] Fix search path usage remark with modules

Wed Jan 5 06:27:00 PST 2022

jansvoboda11 added a comment.

In D113676#3168219 <https://reviews.llvm.org/D113676#3168219>, @Bigcheese wrote:

> Can we not consider a modulemap used when we load an AST that depends on that modulemap? It's possible this breaks if the module only includes textual headers though.

I don't think that would be 100% correct. I think it's possible for `Module` instances (created by parsing a modulemap file) to have semantic meaning even without their AST being loaded. One example (that I just made up but sounds plausible enough to me) is that we could have a fixit for an `@import` of non-existent module that suggests importing another available module (discovered through header search) with the most similar name. We don't need the AST file for that, but removing search paths could change the diagnostic output.

> It really feels like we should have a single place where we actually know if a module is used or not. Long term I would really like to separate modulemap parsing from `Module` creation, which would be a great place to actually do this.

Agreed. But I'm afraid that's a big undertaking. Do you have any other ideas on how to detect module usage with the current state of things?

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D113676/new/

https://reviews.llvm.org/D113676