This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: My unreviewed patches
Tobias SchlÃter wrote:
> Steven Bosscher wrote:
>
>>On Feb 23, 2005 04:00 PM, Tobias SchlÃter <tobias.schlueter@physik.uni-muenchen.de> wrote:
>>
>>>2005/02/07 [gfortran] use string pool where appropriate
>>>http://gcc.gnu.org/ml/fortran/2005-02/msg00047.html
>>
>>This looks OK.
>
>
> Unfortunately this caused regressions on Steve Kargl's amd64 box, so I backed
> the patch out again. We're trying to find out what happened.
The reason for this was a slight oversight which allowed certain code to pass
NULL as format string to the C library. Once Steve and I were awake at the
same time we were able to quickly identify the issue. It didn't affect Linux
(or AIX for that matter) because their C libraries gracefully ignore this
error. I've audited the patch for similar problems and didn't find any, so
I've reinstalled the patch, and added this patch on top of it under the
obviously correct rule after we both tested it.
Sorry for the inconvenience this might have caused,
- Tobi
2005-02-24 Tobias Schl"uter <tobias.schlueter@physik.uni-muenchen.de>
Unrevert previously reversed patch, adding this patch:
* module.c (find_true_name): Deal with NULL module.
Index: module.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/fortran/module.c,v
retrieving revision 1.29
diff -u -p -r1.29 module.c
--- module.c 23 Feb 2005 22:14:11 -0000 1.29
+++ module.c 24 Feb 2005 18:17:06 -0000
@@ -675,7 +675,10 @@ find_true_name (const char *name, const
int c;
sym.name = gfc_get_string (name);
- sym.module = gfc_get_string (module);
+ if (module != NULL)
+ sym.module = gfc_get_string (module);
+ else
+ sym.module = NULL;
t.sym = &sym;
p = true_name_root;