Hi,
am Montag, 24. Februar 2025 um 15:59 schrieben Sie:
> I am releasing onto an unsuspecting world a new version of libm, the math
> library of awesomeness.
> What I would like to know is, how close is this to 1.0?
I haven't seen any discussion of it on freedos-devel@lists.sourceforge.net, so
the status
is probably "untested" or "untested by anybody except Gregory".
Looking at the source makes me wonder: did you implement all this by yourself,
or is it 'simply'
some linux code ported to DOS?
Not doubting you, but some of the alghorithms are fairly advanced. It would be
clearly helpful to
have the alghoritm source indicated.
I also didn't find *immediately* test code, where the library is tested against
known outcomes similar to
if (abs( sin(0.12345678) - 0.12065265261521762716862781)) ) 0.00000000001)
throw_error(__LINE__);
or similar. How is the correctness assured?
>- Testing, testing, more testing, and debugging of the functions that
> are already written.
That's ridiculous without indication what has already debugged, and what not.
So certainly not "1.0".
I have no plan to "test" sin() for all values 0..10^6.
Tom
_______________________________________________
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel
