Re: [PATCH] Fix IVOPTs introducing undefined signed overflow

2013-10-15 Thread Richard Biener
On Tue, 15 Oct 2013, Richard Biener wrote: > > This makes sure that we do not record a signed GIV that we do not > know whether it overflows or not. For the testcase IVOPTs else > can end up replacing an unsigned computation with a signed one. > > Note that alternatively we may decide that it i

[PATCH] Fix IVOPTs introducing undefined signed overflow

2013-10-15 Thread Richard Biener
This makes sure that we do not record a signed GIV that we do not know whether it overflows or not. For the testcase IVOPTs else can end up replacing an unsigned computation with a signed one. Note that alternatively we may decide that it is not desirable for SCEV to return signed { 2147483643,