[Ecls-list] [ANN] ECL v9.10.2

[Andy Hefner]

On Wed, Nov 18, 2009 at 6:01 PM, Juan Jose Garcia-Ripoll
<juanjose.garciaripoll at googlemail.com> wrote:

> It is disabled by default, I would say. To disable all floating point
> exceptions, use (EXT:TRAP-FPE T NIL), as discussed in a different thread.

Sorry, wasn't following the other thread. On my setup (Linux/x86, ECL
from Git circa 30 minutes ago, threads disabled), underflow conditions
are signalled from both the ECL repl and in compiled executables (e.g.
(* 1.2f-38 0.1f0) causes an error).

> To selectively activate / deactivate exceptions, use the condition names:
> division-by-zero, floating-point-overflow, floating-point-underflow,
> floating-point-invalid-operation and floating-point-inexact as first
> argument

That works great, thanks.