On Wed, Apr 12, 2023 at 06:35:29PM +0200, Bernhard Reutner-Fischer wrote:
> >+relation_kind ret = le_op1_op2_relation (lhs);
> >+if (ret == VREL_GT
> >+&& (op1.known_isnan ()
> >+|| op1.maybe_isnan ()
> >+|| op2.known_isnan ()
> >+|| op2.maybe_isnan ()))
> >+
On 12 April 2023 16:21:24 CEST, Jakub Jelinek via Gcc-patches
wrote:
>--- gcc/range-op-float.cc.jj 2023-04-12 12:17:44.784962757 +0200
>+++ gcc/range-op-float.cc 2023-04-12 16:07:54.948759355 +0200
>@@ -835,10 +835,17 @@ public:
> bool fold_range (irange &r, tree type,
>
On Wed, Apr 12, 2023 at 12:33:39PM +0200, Jakub Jelinek via Gcc-patches wrote:
> On Tue, Apr 11, 2023 at 04:58:19PM -0400, Andrew MacLeod wrote:
> > This bootstraps and has no regressions, and is fine by me if you want to use
> > it.
>
> Thanks, looks nice.
> My incremental patch on top of that wo
On Tue, Apr 11, 2023 at 04:58:19PM -0400, Andrew MacLeod wrote:
> This bootstraps and has no regressions, and is fine by me if you want to use
> it.
Thanks, looks nice.
My incremental patch on top of that would then be below.
Though,
FAIL: gcc.dg/tree-ssa/vrp-float-6.c scan-tree-dump-times evrp "
On 4/11/23 04:21, Jakub Jelinek wrote:
Hi!
This patch was what I've tried first before the currently committed
PR109386 fix. Still, I think it is the right thing until we have proper
full set of VREL_* relations for NANs (though it would be really nice
if op1_op2_relation could be passed eithe
On 4/11/23 10:21, Jakub Jelinek wrote:
Hi!
This patch was what I've tried first before the currently committed
PR109386 fix. Still, I think it is the right thing until we have proper
full set of VREL_* relations for NANs (though it would be really nice
if op1_op2_relation could be passed eit
Hi!
This patch was what I've tried first before the currently committed
PR109386 fix. Still, I think it is the right thing until we have proper
full set of VREL_* relations for NANs (though it would be really nice
if op1_op2_relation could be passed either type of the comparison
operands, or even
