Re: [Patch] PR 51938: extend ifcombine

2012-07-24 Thread Richard Guenther
On Mon, Jul 23, 2012 at 10:58 PM, Marc Glisse marc.gli...@inria.fr wrote: On Wed, 20 Jun 2012, Richard Guenther wrote: On Sun, Jun 10, 2012 at 4:16 PM, Marc Glisse marc.gli...@inria.fr wrote: Hello, currently, tree-ssa-ifcombine handles pairs of imbricated ifs that share the same then

Re: [Patch] PR 51938: extend ifcombine

2012-07-24 Thread Marc Glisse
On Tue, 24 Jul 2012, Richard Guenther wrote: On Mon, Jul 23, 2012 at 10:58 PM, Marc Glisse marc.gli...@inria.fr wrote: On Wed, 20 Jun 2012, Richard Guenther wrote: On Sun, Jun 10, 2012 at 4:16 PM, Marc Glisse marc.gli...@inria.fr wrote: Hello, currently, tree-ssa-ifcombine handles pairs

Re: [Patch] PR 51938: extend ifcombine

2012-07-23 Thread Marc Glisse
On Wed, 20 Jun 2012, Richard Guenther wrote: On Sun, Jun 10, 2012 at 4:16 PM, Marc Glisse marc.gli...@inria.fr wrote: Hello, currently, tree-ssa-ifcombine handles pairs of imbricated ifs that share the same then branch, or the same else branch. There is no particular reason why it couldn't

Re: [Patch] PR 51938: extend ifcombine

2012-06-26 Thread Richard Guenther
On Sat, Jun 23, 2012 at 7:18 PM, Marc Glisse marc.gli...@inria.fr wrote: Hello, thanks for looking into the patch. A couple more details now that I am back from a conference: On Wed, 20 Jun 2012, Marc Glisse wrote: On Wed, 20 Jun 2012, Richard Guenther wrote: On Sun, Jun 10, 2012 at

Re: [Patch] PR 51938: extend ifcombine

2012-06-26 Thread Marc Glisse
Thanks for the comments, I'll look into it later in the summer. On Tue, 26 Jun 2012, Richard Guenther wrote: On Sat, Jun 23, 2012 at 7:18 PM, Marc Glisse marc.gli...@inria.fr wrote: Hello, thanks for looking into the patch. A couple more details now that I am back from a conference: On

Re: [Patch] PR 51938: extend ifcombine

2012-06-24 Thread Joseph S. Myers
On Sat, 23 Jun 2012, Marc Glisse wrote: Actually, what would you think of s/flag_trapping_math/flag_signaling_nans/ in invert_tree_comparison? Are there other possible ways ab can trap than having a sNaN for a or b? Ordered comparisons raise exceptions for both quiet and signaling NaNs.

Re: [Patch] PR 51938: extend ifcombine

2012-06-24 Thread Marc Glisse
On Sun, 24 Jun 2012, Joseph S. Myers wrote: On Sat, 23 Jun 2012, Marc Glisse wrote: Actually, what would you think of s/flag_trapping_math/flag_signaling_nans/ in invert_tree_comparison? Are there other possible ways ab can trap than having a sNaN for a or b? Ordered comparisons raise

Re: [Patch] PR 51938: extend ifcombine

2012-06-23 Thread Marc Glisse
Hello, thanks for looking into the patch. A couple more details now that I am back from a conference: On Wed, 20 Jun 2012, Marc Glisse wrote: On Wed, 20 Jun 2012, Richard Guenther wrote: On Sun, Jun 10, 2012 at 4:16 PM, Marc Glisse marc.gli...@inria.fr wrote: Hello, currently,

Re: [Patch] PR 51938: extend ifcombine

2012-06-23 Thread Marc Glisse
On Wed, 20 Jun 2012, Marc Glisse wrote: On Wed, 20 Jun 2012, Richard Guenther wrote: On Sun, Jun 10, 2012 at 4:16 PM, Marc Glisse marc.gli...@inria.fr wrote: Hello, currently, tree-ssa-ifcombine handles pairs of imbricated ifs that share the same then branch, or the same else branch. There

Re: [Patch] PR 51938: extend ifcombine

2012-06-20 Thread Richard Guenther
On Sun, Jun 10, 2012 at 4:16 PM, Marc Glisse marc.gli...@inria.fr wrote: Hello, currently, tree-ssa-ifcombine handles pairs of imbricated ifs that share the same then branch, or the same else branch. There is no particular reason why it couldn't also handle the case where the then branch of

Re: [Patch] PR 51938: extend ifcombine

2012-06-20 Thread Marc Glisse
On Wed, 20 Jun 2012, Richard Guenther wrote: On Sun, Jun 10, 2012 at 4:16 PM, Marc Glisse marc.gli...@inria.fr wrote: Hello, currently, tree-ssa-ifcombine handles pairs of imbricated ifs that share the same then branch, or the same else branch. There is no particular reason why it couldn't

[Patch] PR 51938: extend ifcombine

2012-06-10 Thread Marc Glisse
Hello, currently, tree-ssa-ifcombine handles pairs of imbricated ifs that share the same then branch, or the same else branch. There is no particular reason why it couldn't also handle the case where the then branch of one is the else branch of the other, which is what I do here. Any