[Fortran, patch, pr71936, v1] [6/7 Regression] ICE in wide_int_to_tree, at tree.c:1487

[Fortran, patch, pr71936, v1] [6/7 Regression] ICE in wide_int_to_tree, at tree.c:1487

[Andre Vehreschild]

[-- Attachment #1: Type: text/plain, Size: 393 bytes --]

Hi all,

attached patch fixes the ICE by ensuring that when the SOURCE=/MOLD=
expression is an array-valued function call with no ref, the ref of
object to allocate is taken. The array properties nevertheless are
taken from the function's result.

Bootstraps and regtests ok on x86_64-linux-gnu/F23. Ok for trunk and
gcc-6?

Regards,
	Andre
-- 
Andre Vehreschild * Email: vehre ad gmx dot de 

[-- Attachment #2: pr71936_1.clog --]
[-- Type: text/plain, Size: 360 bytes --]

gcc/fortran/ChangeLog:

2016-08-07  Andre Vehreschild  <vehre@gcc.gnu.org>

	PR fortran/71936
	* trans-array.c (gfc_array_allocate): When SOURCE= is a function
	stick with the ref of the object to allocate.

gcc/testsuite/ChangeLog:

2016-08-07  Andre Vehreschild  <vehre@gcc.gnu.org>

	PR fortran/71936
	* gfortran.dg/allocate_with_source_21.f03: New test.



[-- Attachment #3: pr71936_1.patch --]
[-- Type: text/x-patch, Size: 1939 bytes --]

diff --git a/gcc/fortran/trans-array.c b/gcc/fortran/trans-array.c
index 7572755..6d5a0ef 100644
--- a/gcc/fortran/trans-array.c
+++ b/gcc/fortran/trans-array.c
@@ -5431,12 +5431,19 @@ gfc_array_allocate (gfc_se * se, gfc_expr * expr, tree status, tree errmsg,
 
   if (ref->u.ar.type == AR_FULL && expr3 != NULL)
     {
+      gfc_ref *old_ref = ref;
       /* F08:C633: Array shape from expr3.  */
       ref = expr3->ref;
 
       /* Find the last reference in the chain.  */
       if (!retrieve_last_ref (&ref, &prev_ref))
-	return false;
+	{
+	  if (expr3->expr_type == EXPR_FUNCTION
+	      && gfc_expr_attr (expr3).dimension)
+	    ref = old_ref;
+	  else
+	    return false;
+	}
       alloc_w_e3_arr_spec = true;
     }
 
diff --git a/gcc/testsuite/gfortran.dg/allocate_with_source_21.f03 b/gcc/testsuite/gfortran.dg/allocate_with_source_21.f03
new file mode 100644
index 0000000..fbf3159
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/allocate_with_source_21.f03
@@ -0,0 +1,52 @@
+! { dg-do compile }
+
+! Check fix for pr71936.
+! Contributed by Gerhard Steinmetz
+
+program p
+  type t
+  end type
+
+  call test2()
+  call test4()
+  call test1()
+  call test3()
+contains
+  function f_p()
+    class(t), pointer :: f_p(:)
+    nullify(f_p)
+  end
+
+  function f_a()
+    class(t), allocatable :: f_a(:)
+  end
+
+  subroutine test1()
+    class(t), allocatable :: x(:)
+    allocate (x, mold=f_a())
+    deallocate (x)
+    allocate (x, source=f_a())
+  end subroutine
+
+  subroutine test2()
+    class(t), pointer :: x(:)
+    allocate (x, mold=f_p())
+    deallocate (x)
+    allocate (x, source=f_p())
+  end
+
+  subroutine test3()
+    class(t), pointer :: x(:)
+    allocate (x, mold=f_a())
+    deallocate (x)
+    allocate (x, source=f_a())
+  end
+
+  subroutine test4()
+    class(t), allocatable :: x(:)
+    allocate (x, mold=f_p())
+    deallocate (x)
+    allocate (x, source=f_p())
+  end subroutine
+end
+

Re: [Fortran, patch, pr71936, v1] [6/7 Regression] ICE in wide_int_to_tree, at tree.c:1487

[Paul Richard Thomas]

Hi Andre,

That looks fine to me. OK for 6-branch and trunk.

Thanks for the patch.

Paul

On 7 August 2016 at 15:32, Andre Vehreschild <vehre@gmx.de> wrote:
> Hi all,
>
> attached patch fixes the ICE by ensuring that when the SOURCE=/MOLD=
> expression is an array-valued function call with no ref, the ref of
> object to allocate is taken. The array properties nevertheless are
> taken from the function's result.
>
> Bootstraps and regtests ok on x86_64-linux-gnu/F23. Ok for trunk and
> gcc-6?
>
> Regards,
>         Andre
> --
> Andre Vehreschild * Email: vehre ad gmx dot de



-- 
The difference between genius and stupidity is; genius has its limits.

Albert Einstein

Re: [Fortran, patch, pr71936, v1] [6/7 Regression] ICE in wide_int_to_tree, at tree.c:1487

[Andre Vehreschild]

[-- Attachment #1: Type: text/plain, Size: 905 bytes --]

Hi Paul,

thanks for the review. Committed to trunk as r239237.

I will commit to gcc-6 in one week.

Regards,
	Andre

On Mon, 8 Aug 2016 11:33:06 +0200
Paul Richard Thomas <paul.richard.thomas@gmail.com> wrote:

> Hi Andre,
> 
> That looks fine to me. OK for 6-branch and trunk.
> 
> Thanks for the patch.
> 
> Paul
> 
> On 7 August 2016 at 15:32, Andre Vehreschild <vehre@gmx.de> wrote:
> > Hi all,
> >
> > attached patch fixes the ICE by ensuring that when the SOURCE=/MOLD=
> > expression is an array-valued function call with no ref, the ref of
> > object to allocate is taken. The array properties nevertheless are
> > taken from the function's result.
> >
> > Bootstraps and regtests ok on x86_64-linux-gnu/F23. Ok for trunk and
> > gcc-6?
> >
> > Regards,
> >         Andre
> > --
> > Andre Vehreschild * Email: vehre ad gmx dot de  
> 
> 
> 


-- 
Andre Vehreschild * Email: vehre ad gmx dot de 

[-- Attachment #2: submit.diff --]
[-- Type: text/x-patch, Size: 2974 bytes --]

Index: gcc/fortran/ChangeLog
===================================================================
--- gcc/fortran/ChangeLog	(Revision 239236)
+++ gcc/fortran/ChangeLog	(Arbeitskopie)
@@ -1,5 +1,11 @@
 2016-08-08  Andre Vehreschild  <vehre@gcc.gnu.org>
 
+	PR fortran/71936
+	* trans-array.c (gfc_array_allocate): When SOURCE= is a function
+	stick with the ref of the object to allocate.
+
+2016-08-08  Andre Vehreschild  <vehre@gcc.gnu.org>
+
 	PR fortran/72698
 	* trans-stmt.c (gfc_trans_allocate): Prevent generating code for
 	copy of zero sized string and with it an ICE.
Index: gcc/fortran/trans-array.c
===================================================================
--- gcc/fortran/trans-array.c	(Revision 239235)
+++ gcc/fortran/trans-array.c	(Arbeitskopie)
@@ -5431,12 +5431,19 @@
 
   if (ref->u.ar.type == AR_FULL && expr3 != NULL)
     {
+      gfc_ref *old_ref = ref;
       /* F08:C633: Array shape from expr3.  */
       ref = expr3->ref;
 
       /* Find the last reference in the chain.  */
       if (!retrieve_last_ref (&ref, &prev_ref))
-	return false;
+	{
+	  if (expr3->expr_type == EXPR_FUNCTION
+	      && gfc_expr_attr (expr3).dimension)
+	    ref = old_ref;
+	  else
+	    return false;
+	}
       alloc_w_e3_arr_spec = true;
     }
 
Index: gcc/testsuite/ChangeLog
===================================================================
--- gcc/testsuite/ChangeLog	(Revision 239236)
+++ gcc/testsuite/ChangeLog	(Arbeitskopie)
@@ -1,5 +1,10 @@
 2016-08-08  Andre Vehreschild  <vehre@gcc.gnu.org>
 
+	PR fortran/71936
+	* gfortran.dg/allocate_with_source_21.f03: New test.
+
+2016-08-08  Andre Vehreschild  <vehre@gcc.gnu.org>
+
 	PR fortran/72698
 	* gfortran.dg/allocate_with_source_20.f03: New test.
 
Index: gcc/testsuite/gfortran.dg/allocate_with_source_21.f03
===================================================================
--- gcc/testsuite/gfortran.dg/allocate_with_source_21.f03	(nicht existent)
+++ gcc/testsuite/gfortran.dg/allocate_with_source_21.f03	(Arbeitskopie)
@@ -0,0 +1,52 @@
+! { dg-do compile }
+
+! Check fix for pr71936.
+! Contributed by Gerhard Steinmetz
+
+program p
+  type t
+  end type
+
+  call test2()
+  call test4()
+  call test1()
+  call test3()
+contains
+  function f_p()
+    class(t), pointer :: f_p(:)
+    nullify(f_p)
+  end
+
+  function f_a()
+    class(t), allocatable :: f_a(:)
+  end
+
+  subroutine test1()
+    class(t), allocatable :: x(:)
+    allocate (x, mold=f_a())
+    deallocate (x)
+    allocate (x, source=f_a())
+  end subroutine
+
+  subroutine test2()
+    class(t), pointer :: x(:)
+    allocate (x, mold=f_p())
+    deallocate (x)
+    allocate (x, source=f_p())
+  end
+
+  subroutine test3()
+    class(t), pointer :: x(:)
+    allocate (x, mold=f_a())
+    deallocate (x)
+    allocate (x, source=f_a())
+  end
+
+  subroutine test4()
+    class(t), allocatable :: x(:)
+    allocate (x, mold=f_p())
+    deallocate (x)
+    allocate (x, source=f_p())
+  end subroutine
+end
+

Re: [Fortran, patch, pr71936, v1] [6/7 Regression] ICE in wide_int_to_tree, at tree.c:1487

[Richard Biener]

On Mon, Aug 8, 2016 at 11:58 AM, Andre Vehreschild <vehre@gmx.de> wrote:
> Hi Paul,
>
> thanks for the review. Committed to trunk as r239237.
>
> I will commit to gcc-6 in one week.

If you want to catch the 6.2 release then please make sure to have it on the
branch this Friday.

Richard.

> Regards,
>         Andre
>
> On Mon, 8 Aug 2016 11:33:06 +0200
> Paul Richard Thomas <paul.richard.thomas@gmail.com> wrote:
>
>> Hi Andre,
>>
>> That looks fine to me. OK for 6-branch and trunk.
>>
>> Thanks for the patch.
>>
>> Paul
>>
>> On 7 August 2016 at 15:32, Andre Vehreschild <vehre@gmx.de> wrote:
>> > Hi all,
>> >
>> > attached patch fixes the ICE by ensuring that when the SOURCE=/MOLD=
>> > expression is an array-valued function call with no ref, the ref of
>> > object to allocate is taken. The array properties nevertheless are
>> > taken from the function's result.
>> >
>> > Bootstraps and regtests ok on x86_64-linux-gnu/F23. Ok for trunk and
>> > gcc-6?
>> >
>> > Regards,
>> >         Andre
>> > --
>> > Andre Vehreschild * Email: vehre ad gmx dot de
>>
>>
>>
>
>
> --
> Andre Vehreschild * Email: vehre ad gmx dot de

Re: [Fortran, patch, pr71936, v1] [6/7 Regression] ICE in wide_int_to_tree, at tree.c:1487

[Andre Vehreschild]

[-- Attachment #1: Type: text/plain, Size: 1480 bytes --]

Hi all,

to catch the gcc-6-release I have committed the patch as r239354.

Regards,
	Andre

On Tue, 9 Aug 2016 11:09:39 +0200
Richard Biener <richard.guenther@gmail.com> wrote:

> On Mon, Aug 8, 2016 at 11:58 AM, Andre Vehreschild <vehre@gmx.de>
> wrote:
> > Hi Paul,
> >
> > thanks for the review. Committed to trunk as r239237.
> >
> > I will commit to gcc-6 in one week.  
> 
> If you want to catch the 6.2 release then please make sure to have it
> on the branch this Friday.
> 
> Richard.
> 
> > Regards,
> >         Andre
> >
> > On Mon, 8 Aug 2016 11:33:06 +0200
> > Paul Richard Thomas <paul.richard.thomas@gmail.com> wrote:
> >  
> >> Hi Andre,
> >>
> >> That looks fine to me. OK for 6-branch and trunk.
> >>
> >> Thanks for the patch.
> >>
> >> Paul
> >>
> >> On 7 August 2016 at 15:32, Andre Vehreschild <vehre@gmx.de>
> >> wrote:  
> >> > Hi all,
> >> >
> >> > attached patch fixes the ICE by ensuring that when the
> >> > SOURCE=/MOLD= expression is an array-valued function call with
> >> > no ref, the ref of object to allocate is taken. The array
> >> > properties nevertheless are taken from the function's result.
> >> >
> >> > Bootstraps and regtests ok on x86_64-linux-gnu/F23. Ok for trunk
> >> > and gcc-6?
> >> >
> >> > Regards,
> >> >         Andre
> >> > --
> >> > Andre Vehreschild * Email: vehre ad gmx dot de  
> >>
> >>
> >>  
> >
> >
> > --
> > Andre Vehreschild * Email: vehre ad gmx dot de  


-- 
Andre Vehreschild * Email: vehre ad gmx dot de 

[-- Attachment #2: submit_gcc6.diff --]
[-- Type: text/x-patch, Size: 3017 bytes --]

Index: gcc/fortran/ChangeLog
===================================================================
--- gcc/fortran/ChangeLog	(Revision 239353)
+++ gcc/fortran/ChangeLog	(Arbeitskopie)
@@ -1,5 +1,12 @@
 2016-08-11  Andre Vehreschild  <vehre@gcc.gnu.org>
 
+	Backport from trunk:
+	PR fortran/71936
+	* trans-array.c (gfc_array_allocate): When SOURCE= is a function
+	stick with the ref of the object to allocate.
+
+2016-08-11  Andre Vehreschild  <vehre@gcc.gnu.org>
+
 	Backport from trunk
 	PR fortran/72698
 	* trans-stmt.c (gfc_trans_allocate): Prevent generating code for
Index: gcc/fortran/trans-array.c
===================================================================
--- gcc/fortran/trans-array.c	(Revision 239353)
+++ gcc/fortran/trans-array.c	(Arbeitskopie)
@@ -5421,12 +5421,19 @@
 
   if (ref->u.ar.type == AR_FULL && expr3 != NULL)
     {
+      gfc_ref *old_ref = ref;
       /* F08:C633: Array shape from expr3.  */
       ref = expr3->ref;
 
       /* Find the last reference in the chain.  */
       if (!retrieve_last_ref (&ref, &prev_ref))
-	return false;
+	{
+	  if (expr3->expr_type == EXPR_FUNCTION
+	      && gfc_expr_attr (expr3).dimension)
+	    ref = old_ref;
+	  else
+	    return false;
+	}
       alloc_w_e3_arr_spec = true;
     }
 
Index: gcc/testsuite/ChangeLog
===================================================================
--- gcc/testsuite/ChangeLog	(Revision 239353)
+++ gcc/testsuite/ChangeLog	(Arbeitskopie)
@@ -1,6 +1,12 @@
 2016-08-11  Andre Vehreschild  <vehre@gcc.gnu.org>
 
 	Backport from trunk:
+	PR fortran/71936
+	* gfortran.dg/allocate_with_source_21.f03: New test.
+
+2016-08-11  Andre Vehreschild  <vehre@gcc.gnu.org>
+
+	Backport from trunk:
 	PR fortran/72698
 	* gfortran.dg/allocate_with_source_20.f03: New test.
 
Index: gcc/testsuite/gfortran.dg/allocate_with_source_21.f03
===================================================================
--- gcc/testsuite/gfortran.dg/allocate_with_source_21.f03	(nicht existent)
+++ gcc/testsuite/gfortran.dg/allocate_with_source_21.f03	(Arbeitskopie)
@@ -0,0 +1,52 @@
+! { dg-do compile }
+
+! Check fix for pr71936.
+! Contributed by Gerhard Steinmetz
+
+program p
+  type t
+  end type
+
+  call test2()
+  call test4()
+  call test1()
+  call test3()
+contains
+  function f_p()
+    class(t), pointer :: f_p(:)
+    nullify(f_p)
+  end
+
+  function f_a()
+    class(t), allocatable :: f_a(:)
+  end
+
+  subroutine test1()
+    class(t), allocatable :: x(:)
+    allocate (x, mold=f_a())
+    deallocate (x)
+    allocate (x, source=f_a())
+  end subroutine
+
+  subroutine test2()
+    class(t), pointer :: x(:)
+    allocate (x, mold=f_p())
+    deallocate (x)
+    allocate (x, source=f_p())
+  end
+
+  subroutine test3()
+    class(t), pointer :: x(:)
+    allocate (x, mold=f_a())
+    deallocate (x)
+    allocate (x, source=f_a())
+  end
+
+  subroutine test4()
+    class(t), allocatable :: x(:)
+    allocate (x, mold=f_p())
+    deallocate (x)
+    allocate (x, source=f_p())
+  end subroutine
+end
+