nfields test

[Karim BELABAS <Karim.Belabas@math.u-psud.fr>]

[Ilya wrote:]
> Looking into pari-2.0.11.beta\src\test\32\nfields I can see
> 
> ? p2=x^5+3021*x^4-786303*x^3-6826636057*x^2-546603588746*x+385389051407205
> 7
> x^5 + 3021*x^4 - 786303*x^3 - 6826636057*x^2 - 546603588746*x + 385389
> 0514072057
> 
> Note extra 7s at the start and the end.  ?!!!

It's not junk. In test mode (gp -test, mentionned in the man since 2.0.10
(I think)) all lines are wrapped at 74 chars. This makes the output of diff
human readable (without ediff). And has the added benefit of by-passing a
bug a many versions of sed (don't accept lines longer than 4096 chars. 4 or
5 lines had this property in the old bench).

> If the testdata contains junk, how one can trust it?  Is not it
> manually checked?

It is manually checked.

   Karim.

P.S: Our mail server crashed and was down for the whole of August so I
could not release the stable version. It turned out that _many_
bugs/inefficiencies were reported during this delay (e.g Igor's message on
factoring polynomials). So I'm going to release another beta shortly
(featuring extensive kernel changes...). 

Ilya, I won't include your last two patches in this one (valence +
highlevel). (Besides: I want to keep file names DOS compatible, please
restrict to 8.3 format for files which are not platform specific).
--
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://pari.home.ml.org