[patch, fortran] Fix PR 26039, lack of conformance checking in some intrinsics

[Paul Thomas]

Thomas

:REVIEWMAIL:

>Because I didn't know that this exists :-) Thanks for the hint.
>  
>
You are entirely welcome!  I am just using it for checking elemental 
subroutine arguments.

>What about this?  OK for 4.2 and (after some days) for 4.1?
>
>	Thomas
>
>2006-01-31  Thomas Koenig  <Thomas.Koenig@online.de>
>
>	PR fortran/26039
>	expr.c (gfc_check_conformance):  Reorder error message
>	to avoid plural.
>	check.c(gfc_check_minloc_maxloc):  Call gfc_check_conformance
>	for checking arguments array and mask.
>	(check_reduction):  Likewise.
>  
>
I am stunned that changing the error message did not bring the testsuite 
down in ruins.  Ah well, "chapeau" as they say around here.

>2006-01-31  Thomas Koenig  <Thomas.Koenig@online.de>
>
>	PR fortran/26039
>	maxval_maxloc_conformance_1.f90:  New test.
>  
>
>------------------------------------------------------------------------
>
>Index: expr.c
>===================================================================
>--- expr.c	(revision 110306)
>+++ expr.c	(working copy)
>@@ -1821,7 +1821,7 @@ gfc_check_conformance (const char *optyp
> 
>       if (op1_flag && op2_flag && mpz_cmp (op1_size, op2_size) != 0)
> 	{
>-	  gfc_error ("%s at %L has different shape on dimension %d (%d/%d)",
>+	  gfc_error ("different shape for %s at %L on dimension %d (%d/%d)",
>  
>
Yes, that's good.

The rest is fine; OK for trunk and for 4.1, after a few days as you say.

Thanks

Paul