This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: [Fortran, Patch, PR72698, v1] [5/6/7 Regression] ICE in lhd_incomplete_type_error, at langhooks.c:205
- From: Andre Vehreschild <vehre at gmx dot de>
- To: Thomas Koenig <tkoenig at netcologne dot de>
- Cc: GCC-Patches-ML <gcc-patches at gcc dot gnu dot org>, GCC-Fortran-ML <fortran at gcc dot gnu dot org>
- Date: Mon, 8 Aug 2016 11:49:54 +0200
- Subject: Re: [Fortran, Patch, PR72698, v1] [5/6/7 Regression] ICE in lhd_incomplete_type_error, at langhooks.c:205
- Authentication-results: sourceware.org; auth=none
- References: <20160807135234.7dd4c7fc@vepi2> <b434cfb0-2fc7-4e52-14d0-5032c2ce870c@netcologne.de>
Hi Thomas,
thanks for the review. I have changed the test to run and result check
as requested. Committed to trunk as r239236.
Will commit to gcc-6 and -5-branch in one week.
Regards,
Andre
On Mon, 8 Aug 2016 08:42:08 +0200
Thomas Koenig <tkoenig@netcologne.de> wrote:
> Am 07.08.2016 um 13:52 schrieb Andre Vehreschild:
> Hi Andre,
>
> > attached patch fixes the ICE caused by a zero-sized string.
> > Assigning that string to a temporary variable obviously did not
> > work out. The patch fixes this by checking for zero-sized strings
> > in SOURCE= and not producing the code to assign "nothing" to the
> > temporary variable and later to the allocated memory. The version
> > for gcc-5 had to be adapted slightly, because the version of the
> > ALLOCATE() implementation is way behind.
> >
> > Bootstrapped and regtested on x86_64-linux-gnu/F23. Ok for trunk,
> > gcc-6 and gcc-5?
>
> Looks good.
>
> With the test case, you might consider changing that into a runtime
> test to make sure that the correct result is obtained.
>
> So, OK with that change.
>
> Regards
>
> Thomas
>
--
Andre Vehreschild * Email: vehre ad gmx dot de
Index: gcc/fortran/ChangeLog
===================================================================
--- gcc/fortran/ChangeLog (Revision 239235)
+++ gcc/fortran/ChangeLog (Arbeitskopie)
@@ -1,5 +1,11 @@
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.
+
+2016-08-08 Andre Vehreschild <vehre@gcc.gnu.org>
+
PR fortran/70524
* trans-array.c (gfc_trans_dummy_array_bias): Ensure that the
location information is correctly set.
Index: gcc/fortran/trans-stmt.c
===================================================================
--- gcc/fortran/trans-stmt.c (Revision 239235)
+++ gcc/fortran/trans-stmt.c (Arbeitskopie)
@@ -5448,9 +5448,19 @@
}
gfc_add_block_to_block (&block, &se.pre);
gfc_add_block_to_block (&post, &se.post);
+
+ /* Special case when string in expr3 is zero. */
+ if (code->expr3->ts.type == BT_CHARACTER
+ && integer_zerop (se.string_length))
+ {
+ gfc_init_se (&se, NULL);
+ temp_var_needed = false;
+ expr3_len = integer_zero_node;
+ e3_is = E3_MOLD;
+ }
/* Prevent aliasing, i.e., se.expr may be already a
variable declaration. */
- if (se.expr != NULL_TREE && temp_var_needed)
+ else if (se.expr != NULL_TREE && temp_var_needed)
{
tree var, desc;
tmp = GFC_DESCRIPTOR_TYPE_P (TREE_TYPE (se.expr)) || is_coarray ?
@@ -5679,11 +5689,8 @@
gcc_assert (expr3_esize);
expr3_esize = fold_convert (sizetype, expr3_esize);
if (e3_is == E3_MOLD)
- {
- /* The expr3 is no longer valid after this point. */
- expr3 = NULL_TREE;
- e3_is = E3_UNSET;
- }
+ /* The expr3 is no longer valid after this point. */
+ expr3 = NULL_TREE;
}
else if (code->ext.alloc.ts.type != BT_UNKNOWN)
{
@@ -6012,7 +6019,7 @@
fold_convert (TREE_TYPE (al_len),
integer_zero_node));
}
- if (code->expr3 && !code->expr3->mold)
+ if (code->expr3 && !code->expr3->mold && e3_is != E3_MOLD)
{
/* Initialization via SOURCE block (or static default initializer).
Classes need some special handling, so catch them first. */
Index: gcc/testsuite/ChangeLog
===================================================================
--- gcc/testsuite/ChangeLog (Revision 239235)
+++ gcc/testsuite/ChangeLog (Arbeitskopie)
@@ -1,3 +1,8 @@
+2016-08-08 Andre Vehreschild <vehre@gcc.gnu.org>
+
+ PR fortran/72698
+ * gfortran.dg/allocate_with_source_20.f03: New test.
+
2016-08-08 Alan Modra <amodra@gmail.com>
* gcc.c-torture/compile/pr72802.c: New.
Index: gcc/testsuite/gfortran.dg/allocate_with_source_20.f03
===================================================================
--- gcc/testsuite/gfortran.dg/allocate_with_source_20.f03 (nicht existent)
+++ gcc/testsuite/gfortran.dg/allocate_with_source_20.f03 (Arbeitskopie)
@@ -0,0 +1,21 @@
+! { dg-do run }
+
+! Check that PR72698 is fixed.
+! Contributed by Gerhard Steinmetz
+
+module m
+contains
+ integer function f()
+ f = 4
+ end
+end
+program p
+ use m
+ character(3), parameter :: c = 'abc'
+ character(:), allocatable :: z
+ allocate (z, source=repeat(c(2:1), f()))
+ if (len(z) /= 0) call abort()
+ if (z /= "") call abort()
+end
+
+