[RFC] libm replacement for Busybox

[Roberto A. Foglietta]

2008/10/10 Loïc Grenié <loic.grenie at gmail.com>:

>    They are neither very fast nor very precise (the
>  trigonometric functions are awful, up to 16 bits are
>  false) but libm is rather large and usually a user of
>  Busybox on an initrd or embedded platform does not
>  need neither speed nor excellent precision. Otherwise
>  it would always be possible to disable them anyway.

 How not very precise? You claim they are false after 16 bits but a
test of 100 runs of your "quick and dirty way" to check the precision
says:

 for i in $(seq 100); do ./libmtest; for f in err*; do tail -1 $f |
sed "s/^/$f: /"|cut -f2 -d\: | cut -f1; done; done | tee pippo.txt

 roberto at rafbook:/tmp$ sort -g pippo.txt | head -n1
 -52.034560116350050407

 roberto at rafbook:/tmp$ sort -g pippo.txt | tail -n1
 -4

 precision varies between 52 (very good) and 4 (very scarse) bits. Are
the these the correct meaning of log2 you had said before?

 Cheers,
-- 
/roberto