Re: [PATCH] PR fortran 88116,88467 -- Catch array constructor errors

[Segher Boessenkool]

On Sun, Dec 16, 2018 at 07:46:45AM -0800, Steve Kargl wrote:
> On Sun, Dec 16, 2018 at 11:45:11AM +0100, Dominique d'Humières wrote:
> > Hi Steve,
> > 
> > The patch works as expected.
> > 
> > Is "Can\’t " instead of "Can’t " really necessary?
> > 
> 
> I don't remember how dejagnu handles quotes in the
> regex expression.  "Can\'t" seems to work (for me).
> If you want to change it, go ahead.  Personally, I
> would prefer "Cannot". 

A backslash in a double-quoted string is handled by backslash-substitution,
before it is ever seen as a regexp.  "\'" is exactly the same as "'".

In a braced string it does not get backslash-substitution and then it _is_
handled by the regular expression parsing, but again, as RE {\'} is exactly
the same as {'}.


Segher

Re: [PATCH] PR fortran 88116,88467 -- Catch array constructor errors

[Steve Kargl]

On Sun, Dec 16, 2018 at 11:45:11AM +0100, Dominique d'Humières wrote:
> Hi Steve,
> 
> The patch works as expected.
> 
> Is "Can\’t " instead of "Can’t " really necessary?
> 

I don't remember how dejagnu handles quotes in the
regex expression.  "Can\'t" seems to work (for me).
If you want to change it, go ahead.  Personally, I
would prefer "Cannot". 

-- 
Steve

Re: [PATCH] PR fortran 88116,88467 -- Catch array constructor errors

[Dominique d'Humières]

Hi Steve,

The patch works as expected.

Is "Can\’t " instead of "Can’t " really necessary?

Thanks for the patch,

Dominique

[PATCH] PR fortran 88116,88467 -- Catch array constructor errors

[Steve Kargl]

The attached patch has been regression tested on
x86_64-*-freebsd.  It fixes to PR.  Both PR are related
to errors in array constructors.  PR fortran/88116 was not 
reporting an error on type mismatches with nested array
constructors.  PR fortran/88467 was silently accepting
a CHARACTER entity within nested array constructors.

OK to commit?

2018-12-13  Steven G. Kargl  

PR fortran/88116
PR fortran/88467
* array.c (gfc_match_array_constructor): Check return value of
gfc_convert_type().  Skip constructor elements with BT_UNKNOWN,
which need to go through resolution.
* intrinsic.c (gfc_convert_type_warn): Return early if the types
martch (i.e., no conversion is required).
* simplify.c (gfc_convert_constant): Remove a gfc_internal_error,
and return gfc_bad_expr.

2018-12-13  Steven G. Kargl  

PR fortran/88116
* gfortran.dg/pr88116_1.f90: New test.
* gfortran.dg/pr88116_2.f90: Ditto.

PR fortran/88467
* gfortran.dg/pr88467.f90: New test.

-- 
Steve
Index: gcc/fortran/array.c
===
--- gcc/fortran/array.c	(revision 267063)
+++ gcc/fortran/array.c	(working copy)
@@ -1246,7 +1246,9 @@ done:
 	{
 	  c = gfc_constructor_first (head);
 	  for (; c; c = gfc_constructor_next (c))
-	gfc_convert_type (c->expr, , 1);
+	if (!gfc_convert_type (c->expr, , 1)
+		&& c->expr->ts.type != BT_UNKNOWN)
+	  return MATCH_ERROR;
 	}
 }
   else
Index: gcc/fortran/intrinsic.c
===
--- gcc/fortran/intrinsic.c	(revision 267063)
+++ gcc/fortran/intrinsic.c	(working copy)
@@ -5030,6 +5030,13 @@ gfc_convert_type_warn (gfc_expr *expr, gfc_typespec *t
   if (expr->ts.type == BT_UNKNOWN)
 goto bad;
 
+  /* In building an array constructor, gfortran can end up here when no
+ conversion is required for an intrinsic type.  We need to let derive
+ types drop through.  */
+  if (from_ts.type != BT_DERIVED
+  && (from_ts.type == ts->type && from_ts.kind == ts->kind))
+return true;
+
   if (expr->ts.type == BT_DERIVED && ts->type == BT_DERIVED
   && gfc_compare_types (>ts, ts))
 return true;
Index: gcc/fortran/simplify.c
===
--- gcc/fortran/simplify.c	(revision 267063)
+++ gcc/fortran/simplify.c	(working copy)
@@ -8360,7 +8360,7 @@ gfc_convert_constant (gfc_expr *e, bt type, int kind)
 
 default:
 oops:
-  gfc_internal_error ("gfc_convert_constant(): Unexpected type");
+  return _bad_expr;
 }
 
   result = NULL;
Index: gcc/testsuite/gfortran.dg/pr88116_1.f90
===
--- gcc/testsuite/gfortran.dg/pr88116_1.f90	(nonexistent)
+++ gcc/testsuite/gfortran.dg/pr88116_1.f90	(working copy)
@@ -0,0 +1,4 @@
+! { dg-do compile }
+program p
+   print *, [integer :: 1, [integer(8) :: 2, ['3']]] ! { dg-error "Can't convert" }
+end
Index: gcc/testsuite/gfortran.dg/pr88116_2.f90
===
--- gcc/testsuite/gfortran.dg/pr88116_2.f90	(nonexistent)
+++ gcc/testsuite/gfortran.dg/pr88116_2.f90	(working copy)
@@ -0,0 +1,7 @@
+! { dg-do run }
+program p
+  real :: a(2) = [real :: 1, [integer :: (real(k), k=2,1), 2]]
+  real :: b(1) = [real :: [integer :: (dble(k), k=1,0), 2]]
+  if (a(1) /= 1. .or. a(2) /= 2. .or. b(1) /= 2.) stop 1
+end
+
Index: gcc/testsuite/gfortran.dg/pr88467.f90
===
--- gcc/testsuite/gfortran.dg/pr88467.f90	(nonexistent)
+++ gcc/testsuite/gfortran.dg/pr88467.f90	(working copy)
@@ -0,0 +1,4 @@
+! { dg-do compile }
+program foo
+   print *, [integer :: 1, [integer(8) :: 2, '3']] ! { dg-error "Can\'t convert" }
+end program foo