Re: A missing (?) function: exponentfp(x)

[Ilya Zakharevich <nospam-abuse@ilyaz.org>]

On Thu, Sep 19, 2024 at 10:39:51AM +0200, Loïc Grenié wrote:
> > Unfortunately, thinking of this more: no go.  You cannot pass through
> > a double without a correction of large exponents…

>     The proposed patch of Aurel passes only the first word of the mantissa
>   to log2.

Ah!  Sorry!  Did not pay enough attention (to rootpol.c, of all the places!)

Unfortunately, the complex branch of dbllog2() seems to be majorly
wrong…

Using flog2() wrapper defined earlier in this thread:

      	       flog2delta(x)=flog2(x)-log(abs(x))/log(2);
  (03:01) gp > flog2delta(1+0.0008*I)
  %284 = -4.6166226535255913560320062458846938056 E-7

It seems that somebody wanted to write a Taylor series of order
BYTES_IN_LONG/4 (in 4**(-abs(x-y))/2?), but stopped at writing the 0th
term only…

Yours,
Ilya

  P.S.  But I think that a better solution is to bite the bullet and
  	permit "d" in the import signature, allowing returning
  	double.¹⁾ Maybe even allow "d" for convert-double-to-1-word,
  	and "dd" for convert-double-to-64-bits.

	Of course, even then one would still need to allow all the
	other types in dbllog2()…

	   ¹⁾ Allowing double arguments would lead to a combinatorial
	      explosion, but double return is just one more 1-line
	      branch in addition to 4–5 already present.