[PATCH] PR fortran/56667 -- Issue a sane error message

[PATCH] PR fortran/56667 -- Issue a sane error message

[Steve Kargl]

[-- Attachment #1: Type: text/plain, Size: 676 bytes --]

For the mangled code in the new testcase, gfortran issues
a somewhat obtuse error message.  The problem is the matcher
for a complex entity runs prior to the matcher for an implied
do-loop.  Errors are queued in that order so an error messagei
for a mangled complex constant is emitted.

Regression tested on x86_64-*-freebsd.

2018-03-07  Steven G. Kargl  <kargl@gcc.gnu.org>

	PR fortran/56667
	* primary.c (match_sym_complex_part): Give the matcher for an implied
	do-loop a chance to run.

2018-03-07  Steven G. Kargl  <kargl@gcc.gnu.org>

	PR fortran/56667
	* gfortran.dg/implied_do_2.f90: New test.
	* gfortran.dg/coarray_8.f90: Update for new error message.

-- 
Steve

[-- Attachment #2: pr56667.diff --]
[-- Type: text/x-diff, Size: 2264 bytes --]

Index: gcc/fortran/primary.c
===================================================================
--- gcc/fortran/primary.c	(revision 258227)
+++ gcc/fortran/primary.c	(working copy)
@@ -1248,8 +1248,22 @@ match_sym_complex_part (gfc_expr **result)
 
   if (sym->attr.flavor != FL_PARAMETER)
     {
-      gfc_error ("Expected PARAMETER symbol in complex constant at %C");
-      return MATCH_ERROR;
+      /* Give the matcher for implied do-loops a chance to run.  This yields
+	 a much saner error message for "write(*,*) (i, i=1, 6" where the 
+	 right parenthesis is missing.  */
+      char c;
+      gfc_gobble_whitespace ();
+      c = gfc_peek_ascii_char ();
+      if (c == '=' || c == ',')
+	{
+	  m = MATCH_NO;
+	}
+      else
+	{
+	  gfc_error ("Expected PARAMETER symbol in complex constant at %C");
+	  m = MATCH_ERROR;
+	}
+      return m;
     }
 
   if (!sym->value)
Index: gcc/testsuite/gfortran.dg/implied_do_2.f90
===================================================================
--- gcc/testsuite/gfortran.dg/implied_do_2.f90	(nonexistent)
+++ gcc/testsuite/gfortran.dg/implied_do_2.f90	(working copy)
@@ -0,0 +1,7 @@
+! { dg-do compile }
+! PR fortran/56667
+program error_message
+   implicit none
+   integer :: ir
+   write(*,*) ( ir, ir = 1,10    ! { dg-error "Expected a right parenthesis" }
+end program error_message 
Index: gcc/testsuite/gfortran.dg/coarray_8.f90
===================================================================
--- gcc/testsuite/gfortran.dg/coarray_8.f90	(revision 258227)
+++ gcc/testsuite/gfortran.dg/coarray_8.f90	(working copy)
@@ -145,7 +145,7 @@ end module mmm4
 
 subroutine tfgh()
   integer :: i(2)
-  DATA i/(i, i=1,2)/ ! { dg-error "Expected PARAMETER symbol" }
+  DATA i/(i, i=1,2)/ ! { dg-error "Syntax error in DATA" }
   do i = 1, 5 ! { dg-error "cannot be an array" }
   end do ! { dg-error "Expecting END SUBROUTINE" }
 end subroutine tfgh
@@ -153,7 +153,7 @@ end subroutine tfgh
 subroutine tfgh2()
   integer, save :: x[*]
   integer :: i(2)
-  DATA i/(x, x=1,2)/ ! { dg-error "Expected PARAMETER symbol" }
+  DATA i/(x, x=1,2)/ ! { dg-error "Syntax error in DATA" }
   do x = 1, 5 ! { dg-error "cannot be a coarray" }
   end do ! { dg-error "Expecting END SUBROUTINE" }
 end subroutine tfgh2

Re: [PATCH] PR fortran/56667 -- Issue a sane error message

[Thomas Koenig]

Hi Steve,

> For the mangled code in the new testcase, gfortran issues
> a somewhat obtuse error message.  The problem is the matcher
> for a complex entity runs prior to the matcher for an implied
> do-loop.  Errors are queued in that order so an error messagei
> for a mangled complex constant is emitted.
> 
> Regression tested on x86_64-*-freebsd.

OK for trunk.

Thanks for the patch!

Regards

	Thomas