Ilya Zakharevich on Mon, 7 Apr 2003 16:52:49 -0700


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: [PATCH CVS] broken proto


On Mon, Apr 07, 2003 at 11:47:19PM +0200, Bill Allombert wrote:
> On Mon, Apr 07, 2003 at 02:16:46PM -0700, Ilya Zakharevich wrote:
> > On Mon, Apr 07, 2003 at 12:17:26PM +0200, Bill Allombert wrote:
> > > I think we should remove completly valence code from GP. They are not
> > > used anymore, not documented and wrong.
> > 
> > They are used (by Math::Pari), documented, and are correct (at least
> > with my patches).  Moreover, now they are very easy to check for
> > correctness.
> 
> That's what I cannot understand: could not you generate them yourself 
> at build time instead of requiring them to be in the PARI source ?

a) Why generate them when they are already there?

b) *Where* to store the values?

c) If you can answer (a) and (b), who is going to implement this?
   (Although with Math::PariBuild, this may be not that complicated)

> Anyway if you insist to keep them in PARI, then please send the
> documentation and the perl script that make the check.

The documentation is there.  The perl script is Makefile.PL of Math::Pari.

 * Valence 0 reserved for functions without mandatory args.
 * Valence 99 reserved for codes which do not correspond 1-to-1 to valences.
 * Any other valence (what to do with 0?) should correspond to exactly
 * one code.

Basically, if you change the code, change the valence; if the new code
coincides with one of an existing function, set the valence to the
valence of this function.  If not, set the valence to 99.

Hope this helps,
Ilya