[Openmcl-devel] more bugs on Linux/AMD64

[Gary Byers]

On Thu, 13 Jul 2006, Gary Byers wrote:

>
>
> On Thu, 13 Jul 2006, Eric Marsden wrote:
>> * I notice that floating point overflow and underflow detection
>>  default to off, at least on my system. Should stuff be expected to
>>  break if (set-fpu-mode :underflow t) is used?
>
> Overflow detection should be on by default, and it is in the 060705
> image that I have here.
>
> The best answer that I can give ("maybe something will break, but I
> don't remember what") isn't a very good answer.  My recollection is
> that whatever that something is, it could probably disable/restore
> the overflow-enable bit on a local basis.
>

I'm not sure if this is what I was (vaguely) remembering, but:

? (ccl::set-fpu-mode :underflow t)  ; enable FLOATING-POINT-UNDERFLOW
                                     ; exceptions
? 1e-50
-> IMPOSSIBLE-NUMBER reader-error

I think that the READER-ERROR is allowed and probably correct there
(see 2.3.1.1); with underflow detection off, 0.0 is returned.

2.3.1.1 just says that a READER-ERROR might be signaled if a potential
number's exponent is out of bounds; I suppose that this is a case where
underflow detection could be locally disabled under control of some
preference for people who'd prefer getting a 0.0 back in cases like this.