imap from reals to rationals seems to loose signs of coefficients

[Madalina (User Forum)]

reported by Madalina (User Forum, 28 Aug 2009)

here's a code sample to illustrate the effect:

ring r=real, (x,y), dp;

poly f= -1+2x2+3y3-2x-3y; poly g= -1-2x2-3y3+x+3y; ideal m=f,g; ring t=0,(x,y),dp; ideal n=imap(r,m); n[1];

3y3+2x2+2x+3y+1

n[2];

3y3+2x2+x+3y+1

poly newf=imap(r,f); poly newg=imap(r,g); newf;

3y3+2x2+2x+3y+1

newg;

3y3+2x2+x+3y+1

[hannes]

fixed sign bug in kernel/longrat.cc: nlMapR

[hannes]

fixed sign bug in kernel/longrat.cc: nlMapR