public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r14-3633] Fortran: runtime bounds-checking in presence of array constructors [PR31059]
@ 2023-09-01 20:42 Harald Anlauf
  0 siblings, 0 replies; only message in thread
From: Harald Anlauf @ 2023-09-01 20:42 UTC (permalink / raw)
  To: gcc-cvs

https://gcc.gnu.org/g:6f06152541d62ae7c8579b7d7bf552be19e15b05

commit r14-3633-g6f06152541d62ae7c8579b7d7bf552be19e15b05
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Aug 31 22:19:58 2023 +0200

    Fortran: runtime bounds-checking in presence of array constructors [PR31059]
    
    gcc/fortran/ChangeLog:
    
            PR fortran/31059
            * trans-array.cc (gfc_conv_ss_startstride): For array bounds checking,
            consider also array constructors in expressions, and use their shape.
    
    gcc/testsuite/ChangeLog:
    
            PR fortran/31059
            * gfortran.dg/bounds_check_fail_5.f90: New test.

Diff:
---
 gcc/fortran/trans-array.cc                        | 23 ++++++++++++++++++++
 gcc/testsuite/gfortran.dg/bounds_check_fail_5.f90 | 26 +++++++++++++++++++++++
 2 files changed, 49 insertions(+)

diff --git a/gcc/fortran/trans-array.cc b/gcc/fortran/trans-array.cc
index 90a7d4e9aef3..6ca58e98547a 100644
--- a/gcc/fortran/trans-array.cc
+++ b/gcc/fortran/trans-array.cc
@@ -4740,6 +4740,29 @@ done:
       for (n = 0; n < loop->dimen; n++)
 	size[n] = NULL_TREE;
 
+      /* If there is a constructor involved, derive size[] from its shape.  */
+      for (ss = loop->ss; ss != gfc_ss_terminator; ss = ss->loop_chain)
+	{
+	  gfc_ss_info *ss_info;
+
+	  ss_info = ss->info;
+	  info = &ss_info->data.array;
+
+	  if (ss_info->type == GFC_SS_CONSTRUCTOR && info->shape)
+	    {
+	      for (n = 0; n < loop->dimen; n++)
+		{
+		  if (size[n] == NULL)
+		    {
+		      gcc_assert (info->shape[n]);
+		      size[n] = gfc_conv_mpz_to_tree (info->shape[n],
+						      gfc_index_integer_kind);
+		    }
+		}
+	      break;
+	    }
+	}
+
       for (ss = loop->ss; ss != gfc_ss_terminator; ss = ss->loop_chain)
 	{
 	  stmtblock_t inner;
diff --git a/gcc/testsuite/gfortran.dg/bounds_check_fail_5.f90 b/gcc/testsuite/gfortran.dg/bounds_check_fail_5.f90
new file mode 100644
index 000000000000..436cc96621dc
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/bounds_check_fail_5.f90
@@ -0,0 +1,26 @@
+! { dg-do run }
+! { dg-additional-options "-fcheck=bounds -g -fdump-tree-original" }
+! { dg-output "At line 13 .*" }
+! { dg-shouldfail "Array bound mismatch for dimension 1 of array 'ivec' (2/3)" }
+!
+! PR fortran/31059 - runtime bounds-checking in presence of array constructors
+
+program p
+  integer              :: jvec(3) = [1,2,3]
+  integer, allocatable :: ivec(:), kvec(:), lvec(:), mvec(:), nvec(:)
+  ivec    = [1,2]   ! (re)allocation
+  kvec    = [4,5,6] ! (re)allocation
+  ivec(:) = [4,5,6] ! runtime error (->dump)
+  ! not reached ...
+  print *, jvec + [1,2,3] ! OK & no check generated
+  print *, [4,5,6] + jvec ! OK & no check generated
+  print *, lvec + [1,2,3] ! check generated (->dump)
+  print *, [4,5,6] + mvec ! check generated (->dump)
+  nvec(:) = jvec          ! check generated (->dump)
+end
+
+! { dg-final { scan-tree-dump-times "Array bound mismatch " 4 "original" } }
+! { dg-final { scan-tree-dump-times "Array bound mismatch .*ivec" 1 "original" } }
+! { dg-final { scan-tree-dump-times "Array bound mismatch .*lvec" 1 "original" } }
+! { dg-final { scan-tree-dump-times "Array bound mismatch .*mvec" 1 "original" } }
+! { dg-final { scan-tree-dump-times "Array bound mismatch .*nvec" 1 "original" } }

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2023-09-01 20:42 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-09-01 20:42 [gcc r14-3633] Fortran: runtime bounds-checking in presence of array constructors [PR31059] Harald Anlauf

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).