Uros Bizjak wrote:
> Attached patch initializes problematic array to zero instead of
> uninitialized value.
>
> 2014-06-17 Uros Bizjak
>
> * gfortran.dg/pr61335.f90 (cp_unit_create): Initialize
> unit_id and kind_id to zero.
>
> Tested on alphaev68-linux-gnu and x86_64-linux-gnu.
> OK fo
On Fri, Jun 6, 2014 at 10:07 AM, Uros Bizjak wrote:
> On Fri, Jun 6, 2014 at 9:47 AM, Uros Bizjak wrote:
>
>>> 2014-05-28 Richard Biener
>>>
>>> PR tree-optimization/61335
>>> * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
>>> new range fails, drop to varying.
>>>
>>> * gfortran.
On Fri, Jun 6, 2014 at 9:47 AM, Uros Bizjak wrote:
>> 2014-05-28 Richard Biener
>>
>> PR tree-optimization/61335
>> * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
>> new range fails, drop to varying.
>>
>> * gfortran.dg/pr61335.f90: New testcase.
>
> This testcase triggers SIGFPE
Hello!
> 2014-05-28 Richard Biener
>
> PR tree-optimization/61335
> * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
> new range fails, drop to varying.
>
> * gfortran.dg/pr61335.f90: New testcase.
This testcase triggers SIGFPE on alpha due to the use of denormal
operand. Maybe uni
This fixes PR61335 - symbolic range propagation in VRP is so weak
that we didn't notice this very old serious bug ... compare_values
can return "don't know" which is -2, but PHI node visiting handles
it the same as -1 (less than...). Oops.
Bootstrap and regtest running on x86_64-unknown-linux-gn
