Karim BELABAS on Fri, 19 May 2000 09:27:16 +0200 (MET DST)

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

Re: problem with polredabs and temporary variables

[Carl Witty:]
> I ran across a problem using polredabs and temporary variables.  The
> polredabs0 function (actually the nilord function, which is called
> indirectly by polredabs0) creates a new temporary variable and assumes
> that it is lower priority than the variable of the input polynomial.
> Ordinarily this works fine, but it fails if you're using the library
> interface and call the function with your own temporary-variable
> polynomial.
> It's probably not worth fixing this problem...it looks like that would
> be a major rewrite.  Is it worth adding a warning about this type of
> problem to the "temporary variables" section of the manual?

There should be a common interface to access (or create) a variable
with the requested relative priorities, e.g fetch a variable with lower
priority than vx, and higher than vy. Since variable numbers are restricted
(from 0 to MAXVARN) a general solution appears out of reach.

OTH, any particular instance should be easy to fix in the library:
1) check variables at the start of the relevant function (the one which is
going to introduce a new variable number and requires that it be of lower
priority than anything in sight) [here it would be nilord or factmod9, a priori
the former]
2) if trouble is foreseen, change main variable in relevant objects.
3) change it back in final result, if needed.

It's already done in a few functions; a few more are left... I'll try to
automate the previous solution once a stable release is out (it's an annoying
design flaw but it only concerns library programming).


Karim Belabas                    email: Karim.Belabas@math.u-psud.fr
Dep. de Mathematiques, Bat. 425
Universite Paris-Sud             Tel: (00 33) 1 69 15 57 48
F-91405 Orsay (France)           Fax: (00 33) 1 69 15 60 19
PARI/GP Home Page: http://www.parigp-home.de/