About Float nan comparison patch

 
Hi all,
I dare pushing this http://bugs.squeak.org/view.php?id=7260 subject.
A VM fix has been provided, it is simple, and got support from Andreas
if I remember.
Ah yes, http://lists.squeakfoundation.org/pipermail/squeak-dev/2009-January/133443.html
...well, at least the VM fix :)
Any taker ?

Nicolas

 
On Thu, Jun 11, 2009 at 09:53:20PM +0200, Nicolas Cellier wrote:
>  
> Hi all,
> I dare pushing this http://bugs.squeak.org/view.php?id=7260 subject.
> A VM fix has been provided, it is simple, and got support from Andreas
> if I remember.
> Ah yes, http://lists.squeakfoundation.org/pipermail/squeak-dev/2009-January/133443.html
> ...well, at least the VM fix :)
> Any taker ?

Nicolas,
Thanks for the reminder.

Andreas,
Can you comment as to whether the Mantis 7260 patch for handling Nan
in the Float comparison primitives should be included in the VMMaker
package? I think that the answer is yes, but I'm not certain if there
might be side effects for existing code.

I will wait one week and then add the changes to VMMaker on SqueakSource
if no objections have been raised.

Dave

 
David T. Lewis wrote:
> Can you comment as to whether the Mantis 7260 patch for handling Nan
> in the Float comparison primitives should be included in the VMMaker
> package? I think that the answer is yes, but I'm not certain if there
> might be side effects for existing code.

The changes should be included. They won't have an effect on any code
unless that code is using NaNs in comparisons. And wherever these are
used the behavior is currently very much undefined and not IEEE754
compliant.

Cheers,
   - Andreas

 
On Fri, Jun 12, 2009 at 12:49:55AM -0700, Andreas Raab wrote:
Thanks, the changes are now in VMMaker-dtl.123 on SqueakSource.

Dave