2010/11/16 Roberto Bagnara <bagnara@cs.unipr.it>

Fabio,

I saw that we use ldexpl() twice in the floating_point branch.
However:

1) that function is not available on several platforms;
2) in both cases we pass an argument of type double
  as the first argument;
3) it seems we use that function in a way that is not
  conditional on the floating-point format(s)
  involved.

When you have time, can you please look at this issue?
Thanks,

  Roberto


Sure. I should be able to have a look at this on Thursday.
(Roberto, sorry if you receive a duplicate of this mail: I had forgotten to use "Reply to all")