[Bug fortran/52370] [4.7/4.8/4.9 Regression] Spurious may be used uninitialized warning for check of optional argument

2014-02-11 Thread jakub at gcc dot gnu.org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52370

Jakub Jelinek jakub at gcc dot gnu.org changed:

   What|Removed |Added

 CC||jakub at gcc dot gnu.org

--- Comment #5 from Jakub Jelinek jakub at gcc dot gnu.org ---
Indeed, this regressed with r182211.


[Bug fortran/52370] [4.7/4.8/4.9 Regression] Spurious may be used uninitialized warning for check of optional argument

2014-02-11 Thread jakub at gcc dot gnu.org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52370

--- Comment #6 from Jakub Jelinek jakub at gcc dot gnu.org ---
Until we get more fine-grained warning control, I believe the only way to fix
this is make sure TREE_NO_WARNING is set on Fortran optional argument related
variables (not just the data pointer (b.0 here), but I think the others too
(bounds, size, ...).


[Bug fortran/52370] [4.7/4.8/4.9 Regression] Spurious may be used uninitialized warning for check of optional argument

2014-02-11 Thread jakub at gcc dot gnu.org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52370

--- Comment #7 from Jakub Jelinek jakub at gcc dot gnu.org ---
Created attachment 32107
  -- http://gcc.gnu.org/bugzilla/attachment.cgi?id=32107action=edit
gcc49-pr52370.patch

So, how about this patch?


[Bug fortran/52370] [4.7/4.8/4.9 Regression] Spurious may be used uninitialized warning for check of optional argument

2014-02-11 Thread jakub at gcc dot gnu.org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52370

--- Comment #8 from Jakub Jelinek jakub at gcc dot gnu.org ---
Author: jakub
Date: Tue Feb 11 20:48:26 2014
New Revision: 207698

URL: http://gcc.gnu.org/viewcvs?rev=207698root=gccview=rev
Log:
PR fortran/52370
* trans-decl.c (gfc_build_dummy_array_decl): Set TREE_NO_WARNING
on decl if sym-attr.optional.

* gfortran.dg/pr52370.f90: New test.

Added:
trunk/gcc/testsuite/gfortran.dg/pr52370.f90
Modified:
trunk/gcc/fortran/ChangeLog
trunk/gcc/fortran/trans-decl.c
trunk/gcc/testsuite/ChangeLog


[Bug fortran/52370] [4.7/4.8/4.9 Regression] Spurious may be used uninitialized warning for check of optional argument

2013-12-19 Thread rguenth at gcc dot gnu.org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52370

Richard Biener rguenth at gcc dot gnu.org changed:

   What|Removed |Added

   Target Milestone|--- |4.7.4


[Bug fortran/52370] [4.7/4.8/4.9 Regression] Spurious may be used uninitialized warning for check of optional argument

2013-12-14 Thread dominiq at lps dot ens.fr
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52370

Dominique d'Humieres dominiq at lps dot ens.fr changed:

   What|Removed |Added

 Status|UNCONFIRMED |NEW
   Last reconfirmed||2013-12-14
Summary|Spurious may be used   |[4.7/4.8/4.9 Regression]
   |uninitialized warning for  |Spurious may be used
   |check of optional argument  |uninitialized warning for
   ||check of optional argument
 Ever confirmed|0   |1

--- Comment #4 from Dominique d'Humieres dominiq at lps dot ens.fr ---
I get the warning with 4.6.4 up to trunk (r205944), but not for 4.5.4. The
change occurred between revisions r182107 (2011-12-08) and r182980
(2012-01-07).

 That the warning only shows up in 4.7 but not in 4.7.2 is probably due 
 to the patch for PR 50923. (The warning is only printed for -O1 but not 
 for -O2 or -O0.)

The fix for PR 50923 is in the above range.