hokein added inline comments.
================
Comment at: clang-tools-extra/include-cleaner/lib/WalkAST.cpp:86
report(UD->getLocation(), TD,
IsUsed ? RefType::Explicit : RefType::Ambiguous);
}
----------------
VitaNuo wrote:
> kadircet wrote:
> > hokein wrote:
> > > we should report all references as explicit.
> > i think having `Ambiguous` here for unused symbols is fine. we'd like to
> > consider such symbols for the purposes of saying "yeah this include is
> > probably used" but we shouldn't be inserting headers for the unused ones.
> >
> > do we have an example for the contrary?
> @hokein so what would be the final conclusion then? Should I re-introduce the
> "isUsed" check?
oh, right. `Ambiguous` is better, this is similar to OverloadExpr, we can't
prove that the symbol is used. (sorry, I somewhat had an impression this should
be explicit).
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D138821/new/
https://reviews.llvm.org/D138821
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
