This is the mail archive of the
guile@sourceware.cygnus.com
mailing list for the Guile project.
Re: more cvs guile stuff :)
- To: gjb at cs dot washington dot edu (Greg J. Badros)
- Subject: Re: more cvs guile stuff :)
- From: "Dale P. Smith" <dpsm at bigbird dot en dot com>
- Date: Wed, 8 Mar 100 23:39:06 -0500 (EST)
- Cc: guile at sourceware dot cygnus dot com
>
> "Dale P. Smith" <dpsm@en.com> writes:
>
> >
> > Has anyone looked at the patches I sent in that update numbers.c with the
> > latest logxxx routines from scm?
>
> To where did you send them? I'm seeing a message from you about that
> code, but no patch... I applied Eric's patch, which may have duplicated
> and/or broken yours. (And yours may have had stuff done differently or
> fixed more problems than his, so I definitely don't want it to get lost).
I sent one patch to the list. You can get it at
http://sourceware.cygnus.com/ml/guile/2000-02/msg00142.html
I also sent in the same patch with a ChangeLog entry to bug-guile or guile-bug
@gnu.org. (wherever the guile page says to send bug reports).
> Please re-send, and if you can, comment on Eric's changes (now already
> in CVS) relative to yours.
My patch stomps all over his. The current code in cvs has only a few lines
in each of the logxxx functions. The scm code is about 20 lines for each,
with about 6 new scm_big_xxx functions.
My understanding is that the current cvs code uses immediate numbers or just
enough of a bignum to fit in an int (or a long with Eric's changes). The
acutal operation is done on the int (or long).
The new code operates on immediate numbers directly, only using the bignum
routines for bignums.
The code should be faster for smaller numbers, no conversion to bignums
and back to longs for everything.
> And I apologize if I missed your earlier patch.
No problem. ;)
--
Dale P. Smith
dpsm@en.com home
dsmith@altustech.com work