Another nfdisc bug

[Igor Schein <igor@txc.com>]

Hi again,
nfinit() performes irreducibilty check before going any further,
but nfdisc() doesn't.  Here's how it can be *exploited*:

% gp -q -f
? f(n)=for(j=-n,n,for(k=-n,n,for(l=-n,n,nfdisc(x^3+j*x^2+k*x+l))))
? f(1)
  ***   negative or zero argument in an arithmetic function
? f(1)
  ***   negative or zero argument in an arithmetic function
? \q
  ***   bus error: bug in GP (please report).   

This and the previous bug report are based on 2.0.12.alpha as of Nov 6 1998,
with no later patches applied.  The exact behavior may vary depending
on the patch level.  Speaking of patches, I think it's a good idea
to assign a unique ID number to every patch released, so that
later it's easy to refer someone to a specific patch.  The system
of ID assignment doesn't have to be overly sophisticated, as long as
the ID is unique and is self-explanatory as to what sources it's to
be applied.

Thanks

Igor

P.S. Karim, is there a tentative schedule for a beta release?