Sign of tanh ???

Sign of tanh ???

[Bobby D. Bryant]

Isn't tanh supposed to preserve the sign of its argument?

---8<-------
with
  Ada.Float_Text_IO,
  Ada.Numerics.Elementary_Functions;

use
  Ada.Float_Text_IO,
  Ada.Numerics.Elementary_Functions;

procedure Test is
begin
   Put( Tanh(  0.5 ), Fore => 3, Exp => 0 );
   Put( Tanh( -0.5 ), Fore => 3, Exp => 0 );
end Test;
---8<-------

% gnatmake test
gnatgcc -c test.adb
gnatbind -x test.ali
gnatlink test.ali
gnatlink: warning: executable name "test" may conflict with existing
command

% ./test
  0.46212  0.46212

% rpm -q gnat
gnat-3.13p-5


Thanks,

Bobby Bryant
Austin, Texas

Re: Sign of tanh ???

[Bobby D. Bryant]

Wilhelm Spickermann wrote:

> On 21-Apr-01 Bobby D. Bryant wrote:
> > Isn't tanh supposed to preserve the sign of its argument?
>
> Sure. But I get the result
>   0.46212 -0.46212
> and I use the gnat-3.13p-5 rpm (from ALT) too.
>
> Perhaps it´s the C-library. Gnat uses the C-Library to compute the tanh (if
> the result cannot be easily found to be -1.0 1.0 or the argument). Does the
> problem occur in C too?

No, it works OK in C.

Odd...

Bobby Bryant
Austin, Texas

RE: Sign of tanh ???

[Wilhelm Spickermann]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=us-ascii, Size: 383 bytes --]


On 21-Apr-01 Bobby D. Bryant wrote:
> Isn't tanh supposed to preserve the sign of its argument?

Sure. But I get the result
  0.46212 -0.46212
and I use the gnat-3.13p-5 rpm (from ALT) too.

Perhaps it�s the C-library. Gnat uses the C-Library to compute the tanh (if
the result cannot be easily found to be -1.0 1.0 or the argument). Does the
problem occur in C too?

Wilhelm