[Patch, fortran] PR39295 - [4.2/4.3/4.4 Regression] Too strict interface conformance check

[Tobias Burnus]

Paul Richard Thomas wrote:
> This one is straightforward - please look at the comments in the
> testcase and the patch to interface.c
>
> Bootstrapped and regtested on FC9/x86_64 - OK for trunk and 4.3?
>   
Looks OK.

Can you correct the typo here:

+  /* At this point, both symbols are procedures.  It can happen that
+     a external procedures are compared where one is identified by usage

(remove the "a ")

Tobias

> 2009-02-26  Paul Thomas  <pault@gcc.gnu.org>
>
> 	PR fortran/39292
> 	* interface.c (compare_type_rank_if): Return 1 if the symbols
> 	are the same and deal with external procedures where one is
> 	identified to be a function or subroutine by usage but the
> 	other is not.
>
> 2009-02-26  Paul Thomas  <pault@gcc.gnu.org>
>
> 	PR fortran/39292
> 	* gfortran.dg/interface_25.f90: New test.
> 	* gfortran.dg/interface_26.f90: New test.
>