RE: [gnugo-devel] Floating point arithmetics

[Portela Fernando]

Hi,

I don't have much time to reply in detail, sorry. In short :

There seems to be a consensus about the patch not going into 3.6.
And I fully agree.

Since we aren't rushed, I will rework the patch :

- integrate Gunnar's macros idea. I already had some, but missed the
  FP() idea. It works perfectly on VC++, so I doubt there will be any
  problem with other compilers.
- try to reduce the performance impact. I already tried ideas along the
  line of what Arend suggested, unfortunately with no success at all.
  I want to try a couple other things over the week-end before
  submitting a patch.

-- nando