[PATCH] D94017: [flang] Fix bogus message on internal subprogram with alternate return

Peter Klausler via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Jan 4 10:34:49 PST 2021


klausler added a comment.

In D94017#2477615 <https://reviews.llvm.org/D94017#2477615>, @PeteSteinfeld wrote:

> In D94017#2477542 <https://reviews.llvm.org/D94017#2477542>, @klausler wrote:
>
>> I'm confused by this change.  The call checking code that's been modified pertains to absent actual arguments; but alternate returns can't be optional and must have corresponding positional actual arguments.
>
> The code is checking for how to handle the situation where the actual argument is missing.  That's potentially the case for optional arguments, but it also happens for alternate returns.  So new code is saying that it's OK for the actual argument to be missing if the dummy argument is an alternate return.  Formerly, we emitted an error message when calling an internal subprogram with alternate returns.

But how can it be okay for the actual argument to be missing when the corresponding dummy argument is an alternate return?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D94017/new/

https://reviews.llvm.org/D94017



More information about the llvm-commits mailing list