File: [local] / OpenXM_contrib2 / asir2000 / engine / H.c (download)
Revision 1.10, Sat Aug 8 14:19:41 2015 UTC (9 years, 1 month ago) by fujimoto
Branch: MAIN
Changes since 1.9: +4 -1
lines
Added fflush(stderr) after fprintf(stderr, ) for mingw32/mingw64.
|
/*
* Copyright (c) 1994-2000 FUJITSU LABORATORIES LIMITED
* All rights reserved.
*
* FUJITSU LABORATORIES LIMITED ("FLL") hereby grants you a limited,
* non-exclusive and royalty-free license to use, copy, modify and
* redistribute, solely for non-commercial and non-profit purposes, the
* computer program, "Risa/Asir" ("SOFTWARE"), subject to the terms and
* conditions of this Agreement. For the avoidance of doubt, you acquire
* only a limited right to use the SOFTWARE hereunder, and FLL or any
* third party developer retains all rights, including but not limited to
* copyrights, in and to the SOFTWARE.
*
* (1) FLL does not grant you a license in any way for commercial
* purposes. You may use the SOFTWARE only for non-commercial and
* non-profit purposes only, such as academic, research and internal
* business use.
* (2) The SOFTWARE is protected by the Copyright Law of Japan and
* international copyright treaties. If you make copies of the SOFTWARE,
* with or without modification, as permitted hereunder, you shall affix
* to all such copies of the SOFTWARE the above copyright notice.
* (3) An explicit reference to this SOFTWARE and its copyright owner
* shall be made on your publication or presentation in any form of the
* results obtained by use of the SOFTWARE.
* (4) In the event that you modify the SOFTWARE, you shall notify FLL by
* e-mail at risa-admin@sec.flab.fujitsu.co.jp of the detailed specification
* for such modification or the source code of the modified part of the
* SOFTWARE.
*
* THE SOFTWARE IS PROVIDED AS IS WITHOUT ANY WARRANTY OF ANY KIND. FLL
* MAKES ABSOLUTELY NO WARRANTIES, EXPRESSED, IMPLIED OR STATUTORY, AND
* EXPRESSLY DISCLAIMS ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS
* FOR A PARTICULAR PURPOSE OR NONINFRINGEMENT OF THIRD PARTIES'
* RIGHTS. NO FLL DEALER, AGENT, EMPLOYEES IS AUTHORIZED TO MAKE ANY
* MODIFICATIONS, EXTENSIONS, OR ADDITIONS TO THIS WARRANTY.
* UNDER NO CIRCUMSTANCES AND UNDER NO LEGAL THEORY, TORT, CONTRACT,
* OR OTHERWISE, SHALL FLL BE LIABLE TO YOU OR ANY OTHER PERSON FOR ANY
* DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE OR CONSEQUENTIAL
* DAMAGES OF ANY CHARACTER, INCLUDING, WITHOUT LIMITATION, DAMAGES
* ARISING OUT OF OR RELATING TO THE SOFTWARE OR THIS AGREEMENT, DAMAGES
* FOR LOSS OF GOODWILL, WORK STOPPAGE, OR LOSS OF DATA, OR FOR ANY
* DAMAGES, EVEN IF FLL SHALL HAVE BEEN INFORMED OF THE POSSIBILITY OF
* SUCH DAMAGES, OR FOR ANY CLAIM BY ANY OTHER PARTY. EVEN IF A PART
* OF THE SOFTWARE HAS BEEN DEVELOPED BY A THIRD PARTY, THE THIRD PARTY
* DEVELOPER SHALL HAVE NO LIABILITY IN CONNECTION WITH THE USE,
* PERFORMANCE OR NON-PERFORMANCE OF THE SOFTWARE.
*
* $OpenXM: OpenXM_contrib2/asir2000/engine/H.c,v 1.10 2015/08/08 14:19:41 fujimoto Exp $
*/
#include "ca.h"
#include "inline.h"
#include "base.h"
#include <math.h>
#if 1
#define Mulum mulum
#define Divum divum
#define Mulsum mulsum
#define Gcdum gcdum
#endif
#define FCTR 0
#define SQFR 1
#define DDD 2
#define NEWDDD 3
LUM LUMALLOC();
void berle(int index,int count,P f,ML *listp)
{
UM wf,wf1,wf2,wfs,gcd;
ML flist;
int fn,fn1,fm,m,n,fhd;
register int i,j,inv,hd,*ptr,*ptr1;
n = UDEG(f);
wf = W_UMALLOC(n); wf1 = W_UMALLOC(n); wf2 = W_UMALLOC(n);
wfs = W_UMALLOC(n); gcd = W_UMALLOC(n);
for ( j = 0, fn = n + 1; (j < count) && (fn > 1); ) {
m = sprime[index++];
if ( !rem(NM((Q)UCOEF(f)),m) )
continue;
ptoum(m,f,wf); cpyum(wf,wf1);
diffum(m,wf1,wf2); gcdum(m,wf1,wf2,gcd);
if ( DEG(gcd) > 0 )
continue;
hd = COEF(wf)[n]; inv = invm(hd,m);
for ( i = n, ptr = COEF(wf); i >= 0; i-- )
ptr[i] = ( ptr[i] * inv ) % m;
fn1 = berlecnt(m,wf);
if ( fn1 < fn ) {
fn = fn1; fm = m; fhd = hd;
for ( i = n, ptr = COEF(wf), ptr1 = COEF(wfs); i >= 0; i-- )
ptr1[i] = ptr[i];
}
j++;
}
DEG(wfs) = n;
*listp = flist = MLALLOC(fn); flist->n = fn; flist->mod = fm;
/* berlemain(fm,wfs,(UM *)flist->c); */
if ( fm == 2 )
berlemain(fm,wfs,(UM *)flist->c);
else
newddd(fm,wfs,(UM *)flist->c);
for ( i = DEG((UM)(flist->c[0])),
ptr = COEF((UM)(flist->c[0])),
hd = fhd, m = fm; i >= 0; i-- )
ptr[i] = ( ptr[i] * hd ) % m;
}
int berlecnt(int mod,UM f)
{
register int i,j,**c;
int d,dr,n;
UM w,q;
int **almat();
n = DEG(f); c = almat(n,n);
w = W_UMALLOC(mod + n); q = W_UMALLOC(mod + n);
for ( i = 1; ( d = ( mod * i ) ) < n; i++ )
c[d][i - 1] = 1;
DEG(w) = d; COEF(w)[d] = 1;
for ( j = d - 1; j >= 0; j-- )
COEF(w)[j] = 0;
for ( ; ( i < n ) && ( ( dr = divum(mod,w,f,q) ) != -1 ); i++ ) {
for ( j = dr; j >= 0; j-- )
COEF(w)[j + mod] = c[j][i - 1] = COEF(w)[j];
for ( j = mod - 1; j >= 0; j-- )
COEF(w)[j] = 0;
DEG(w) = dr + mod;
}
for ( i = 1; i < n; i++ )
c[i][i - 1] = ( c[i][i - 1] + mod - 1 ) % mod;
return berlecntmain(mod,n,n-1,c);
}
/* XXX berlecntmain should not be used for large mod */
int berlecntmain(int mod,int n,int m,int **c)
{
register int *p1,*p2,i,j,k,l,a;
int *tmp,inv;
int cfs;
for ( cfs = 1, j = k = 0; j < m; j++ ) {
for ( i = k; ( n > i ) && ( c[i][j] == 0 ); i++ );
if ( i == n ) {
cfs++; continue;
}
if ( i != k ) {
tmp = c[i]; c[i] = c[k]; c[k] = tmp;
}
p1 = c[k]; inv = invm((p1[j] + mod) % mod,mod);
for ( l = j; l < m; l++ )
p1[l] = ( p1[l] * inv ) % mod;
for ( i = k + 1; i < n; c[i][j] = 0, i++ )
if ( i != k && ( a = -c[i][j] ) )
for ( l = j + 1, p2 = c[i]; l < m; l++ )
p2[l] = (a*p1[l] + p2[l]) % mod;
k++;
}
return ( cfs );
}
UM *berlemain(int mod,UM f,UM *fp)
{
UM wg,ws,wf,f0,gcd,q;
int n;
register int i;
n = DEG(f); wg = W_UMALLOC(n); mini(mod,f,wg);
if ( DEG(wg) <= 0 ) {
f0 = UMALLOC(n); cpyum(f,f0); *fp++ = f0;
return ( fp );
}
f0 = W_UMALLOC(n); cpyum(f,f0);
ws = W_UMALLOC(n); wf = W_UMALLOC(n);
q = W_UMALLOC(n); gcd = W_UMALLOC(n);
for ( i = 0; i < mod; i++ ) {
cpyum(f0,wf); cpyum(wg,ws);
COEF(ws)[0] = ( COEF(ws)[0] + mod - i ) % mod;
gcdum(mod,wf,ws,gcd);
if ( DEG(gcd) > 0 ) {
if ( DEG(gcd) < n ) {
divum(mod,f0,gcd,q); f0 = q; fp = berlemain(mod,gcd,fp);
}
break;
}
}
fp = berlemain(mod,f0,fp);
return ( fp );
}
void hensel(int index,int count,P f,ML *listp)
{
register int i,j;
int q,n,bound;
int *p;
int **pp;
ML blist,clist,bqlist,cqlist,rlist;
UM *b;
LUM fl,tl;
LUM *l;
if ( UDEG(f) == 1 ) {
*listp = blist = MLALLOC(1); blist->n = 1; blist->c[0] = 0;
return;
}
berle(index,count,f,&blist);
if ( blist->n == 1 ) {
*listp = blist = MLALLOC(1); blist->n = 1; blist->c[0] = 0;
return;
}
gcdgen(f,blist,&clist); henprep(f,blist,clist,&bqlist,&cqlist);
n = bqlist->n; q = bqlist->mod;
bqlist->bound = cqlist->bound = bound = mignotte(q,f);
if ( bound == 1 ) {
*listp = rlist = MLALLOC(n);
rlist->n = n; rlist->mod = q; rlist->bound = bound;
for ( i = 0, b = (UM *)bqlist->c, l = (LUM *)rlist->c; i < n; i++ ) {
tl = LUMALLOC(DEG(b[i]),1); l[i] = tl; p = COEF(b[i]);
for ( j = 0, pp = COEF(tl); j <= DEG(tl); j++ )
pp[j][0] = p[j];
}
} else {
W_LUMALLOC((int)UDEG(f),bound,fl);
ptolum(q,bound,f,fl); henmain(fl,bqlist,cqlist,listp);
}
}
void hensel2(int index,int count,P f,ML *listp)
{
register int i,j;
int mod,q,n,bound,dx;
ML blist,clist,bqlist,cqlist,rlist;
UM fm,qfm,gm,qgm,hm,qhm,qam,qbm,w;
UM *b;
LUM fl,tl;
int k;
dx = UDEG(f);
if ( dx == 1 ) {
*listp = blist = MLALLOC(1); blist->n = 1; blist->c[0] = 0;
return;
}
berle(index,count,f,&blist);
n = blist->n;
mod = blist->mod;
if ( n == 1 ) {
*listp = blist = MLALLOC(1); blist->n = 1; blist->c[0] = 0;
return;
}
/* find k s.t. mod^k <= 2^27 < mod^(k+1); set q = mod^k */
for ( k = 1, q = mod; q <= ((1<<27)/mod); q *= mod, k++ );
/* mignotte bound */
bound = mignotte(q,f);
*listp = rlist = MLALLOC(n);
rlist->n = n;
rlist->mod = q;
rlist->bound = bound;
if ( bound == 1 ) {
gcdgen(f,blist,&clist);
henprep(f,blist,clist,&bqlist,&cqlist);
for ( i = 0, b = (UM *)bqlist->c; i < n; i++ ) {
COEF(rlist)[i] = tl = LUMALLOC(DEG(b[i]),1);
for ( j = 0; j <= DEG(tl); j++ )
COEF(tl)[j][0] = COEF(b[i])[j];
COEF(rlist)[i] = tl;
}
} else {
/* fl = f mod q */
fl = LUMALLOC(dx,bound);
ptolum(q,bound,f,fl);
/* fm = f mod mod */
fm = W_UMALLOC(dx);
ptoum(mod,f,fm);
/* fm = f mod q */
qfm = W_UMALLOC(dx);
ptoum(q,f,qfm);
gm = W_UMALLOC(dx); qgm = W_UMALLOC(dx);
hm = W_UMALLOC(dx); qhm = W_UMALLOC(dx);
qam = W_UMALLOC(dx); qbm = W_UMALLOC(dx);
w = W_UMALLOC(dx);
for ( i = 0; i < n-1; i++ ) {
cpyum(COEF(blist)[i],gm);
cpyum(fm,w);
divum(mod,w,gm,hm);
/* find am,bm s.t. qam*qgm+qbm*qhm=1 mod q, qgm=gm mod mod, qhm=hm mod mod */
henprep2(mod,q,k,qfm,gm,hm,qgm,qhm,qam,qbm);
henmain2(fl,qgm,qhm,qam,qbm,q,bound,&tl);
rlist->c[i] = (pointer)tl;
cpyum(hm,fm);
cpyum(qhm,qfm);
}
rlist->c[i] = fl;
}
}
/*
f = g0*h0 mod m -> f = gk*hk mod m^(bound), f is replaced by hk
*/
void henmain2(LUM f,UM g0,UM h0,UM a0,UM b0,int m,int bound,LUM *gp)
{
int n,dg,dh,i,k,j,dg1,dh1;
UM wu,wr,ws,wt,q,wh1,wg1,wc,wd,we,wz;
LUM wb0,wb1,wb2,fk,gk,hk;
n = DEG(f); dg = DEG(g0); dh = DEG(h0);
W_LUMALLOC(n,bound,wb0);
W_LUMALLOC(n,bound,wb1);
W_LUMALLOC(n,bound,wb2);
wt = W_UMALLOC(2*n); ws = W_UMALLOC(2*n);
wr = W_UMALLOC(2*n); wu = W_UMALLOC(2*n);
q = W_UMALLOC(2*n);
wh1 = W_UMALLOC(2*n); wg1 = W_UMALLOC(2*n);
/* gk = g0 */
gk = LUMALLOC(n,bound);
DEG(gk) = dg;
for ( i = 0; i <= dg; i++ )
COEF(gk)[i][0] = COEF(g0)[i];
/* hk = h0 */
W_LUMALLOC(n,bound,hk);
DEG(hk) = dh;
for ( i = 0; i <= dh; i++ )
COEF(hk)[i][0] = COEF(h0)[i];
/* fk = gk*hk */
W_LUMALLOC(n,bound,fk);
mullum(m,bound,gk,hk,fk);
wc = W_UMALLOC(2*n); wd = W_UMALLOC(2*n);
we = W_UMALLOC(2*n); wz = W_UMALLOC(2*n);
#if 0
mulum(m,a0,g0,wc);
mulum(m,b0,h0,wd);
addum(m,wc,wd,wz);
if ( DEG(wz) != 0 || COEF(wz)[0] != 1 )
error("henmain2 : cannot happen(extgcd)");
#endif
#if 1
fprintf(stderr,"bound=%d\n",bound);
#endif
for ( k = 1; k < bound; k++ ) {
#if 1
fprintf(stderr,".");
#endif
#if defined(VISUAL) || defined(__MINGW32__) || defined(__MINGW64__)
check_intr();
#endif
/* at this point, f = gk*hk mod y^k */
#if 0
for ( j = 0; j < k; j++ )
for ( i = 0; i <= n; i++ )
if ( COEF(f)[i][j] != COEF(f)[i][j] )
error("henmain2 : cannot happen(f=fk)");
#endif
/* wt = (f-gk*hk)/y^k */
for ( i = 0; i <= n; i++ )
COEF(ws)[i] = COEF(f)[i][k];
degum(ws,n);
for ( i = 0; i <= n; i++ )
COEF(wu)[i] = COEF(fk)[i][k];
degum(wu,n);
subum(m,ws,wu,wt);
/* compute wf1,wg1 s.t. wh1*g0+wg1*h0 = wt */
mulum(m,a0,wt,wh1); DEG(wh1) = divum(m,wh1,h0,q);
mulum(m,wh1,g0,wc); subum(m,wt,wc,wd); DEG(wd) = divum(m,wd,h0,wg1);
/* check */
#if 0
if ( DEG(wd) >= 0 || DEG(wg1) > dg )
error("henmain2 : cannot happen(adj)");
mulum(m,wg1,h0,wc); mulum(m,wh1,g0,wd); addum(m,wc,wd,we);
subum(m,we,wt,wz);
if ( DEG(wz) >= 0 )
error("henmain2 : cannot happen(coef)");
#endif
/* fk += ((wg1*hk+wh1*gk)*y^k+wg1*wh1*y^(2*k) mod m^bound */
/* wb0 = wh1*y^k */
clearlum(n,bound,wb0);
DEG(wb0) = dh1 = DEG(wh1);
for ( i = 0; i <= dh1; i++ )
COEF(wb0)[i][k] = COEF(wh1)[i];
/* wb2 = gk*wb0 mod y^bound */
clearlum(n,bound,wb2);
mullum(m,bound,gk,wb0,wb2);
/* fk += wb2 */
addtolum(m,bound,wb2,fk);
/* wb1 = wg1*y^k */
clearlum(n,bound,wb1);
DEG(wb1) = dg1 = DEG(wg1);
for ( i = 0; i <= n; i++ )
COEF(wb1)[i][k] = COEF(wg1)[i];
/* wb2 = hk*wb1 mod y^bound */
clearlum(n,bound,wb2);
mullum(m,bound,hk,wb1,wb2);
/* fk += wb2 */
addtolum(m,bound,wb2,fk);
/* fk += wg1*wh1*y^(2*k) mod y^bound) */
if ( 2*k < bound ) {
clearlum(n,bound,wb2);
mullum(m,bound,wb0,wb1,wb2);
addtolum(m,bound,wb2,fk);
}
/* gk += wg1*y^k, hk += wh1*y^k */
for ( i = 0; i <= DEG(wg1); i++ )
COEF(gk)[i][k] = COEF(wg1)[i];
for ( i = 0; i <= DEG(wh1); i++ )
COEF(hk)[i][k] = COEF(wh1)[i];
}
#if 1
fprintf(stderr,"\n");
#if defined(__MINGW32__) || defined(__MINGW64__)
fflush(stderr);
#endif
#endif
*gp = gk;
clearlum(n,bound,f);
DEG(f) = dh;
for ( i = 0; i <= dh; i++ )
for ( j = 0; j < bound; j++ )
COEF(f)[i][j] = COEF(hk)[i][j];
}
void clearlum(int n,int bound,LUM f)
{
int i;
for ( i = 0; i <= n; i++ )
bzero(COEF(f)[i],bound*sizeof(int));
}
/* g += f */
void addtolum(int m,int bound,LUM f,LUM g)
{
int n,i;
n = DEG(f);
for ( i = 0; i <= n; i++ )
addpadic(m,bound,COEF(f)[i],COEF(g)[i]);
}
void hsq(int index,int count,P f,int *nindex,DCP *dcp)
{
register int i,j,k;
register int **pp,**fpp;
register int *px,*py;
int **wpp;
int n,dr,tmp,m,b,e,np,dt;
LUM fpa,wb0,wb1,lcpa,tpa,tlum;
struct oDUM *dct;
UM wt,wq0,wq,wr,wm,wm0,wa,ws,wb;
LUM *llist,*ll;
UM *dlist,*l,*c;
ML list,fp,cfp;
DCP dc;
sqfrum(index,count,f,nindex,&dct,&fp);
np = fp->n; m = fp->mod;
if ( ( np == 1 ) && ( dct[0].n == 1 ) ) {
NEWDC(dc); DEG(dc) = ONE; COEF(dc) = f; NEXT(dc) = 0; *dcp = dc;
return;
}
for ( i = 0, dt = 0; i < np; i++ )
dt = MAX(DEG(dct[i].f),dt);
b = mig(m,dt,f); fp->bound = b;
if ( np == 1 ) {
nthrootchk(f,dct,fp,dcp);
return;
}
list = W_MLALLOC(np); list->n = np; list->mod = m; list->bound = 1;
for ( i = 0, ll = (LUM *)list->c; i < np; i++ ) {
W_LUMALLOC(DEG(dct[i].f),b,ll[i]);
for ( j = 0, px = COEF(dct[i].f), pp = COEF(ll[i]);
j <= DEG(ll[i]); j++ )
pp[j][0] = px[j];
}
dtestsql(f,list,dct,&dc);
if ( dc ) {
*dcp = dc;
return;
}
n = UDEG(f);
W_LUMALLOC(n,b,fpa); W_LUMALLOC(0,b,lcpa);
W_LUMALLOC(n,b,wb0); W_LUMALLOC(n,b,wb1);
W_LUMALLOC(n,b,tpa);
wt = W_UMALLOC(n); ws = W_UMALLOC(n);
wr = W_UMALLOC(n);
wq = W_UMALLOC(2*n); wq0 = W_UMALLOC(n);
wm = W_UMALLOC(2*n); wm0 = W_UMALLOC(2*n);
wa = W_UMALLOC(2*n);
ptolum(m,b,f,fpa); DEG(lcpa) = 0;
for ( i = 0, pp = COEF(lcpa), fpp = COEF(fpa); i < b; i++ )
pp[0][i] = fpp[n][i];
gcdgen(f,fp,&cfp);
llist = (LUM *) ALLOCA(np*sizeof(LUM));
dlist = (UM *) ALLOCA(np*sizeof(UM));
l = (UM *)fp->c; c = (UM *)cfp->c;
for ( i = 0; i < np; i++ ) {
W_LUMALLOC(DEG(l[i]),b,llist[i]);
for ( j = DEG(l[i]), pp = COEF(llist[i]), px = COEF(l[i]); j >= 0; j-- )
pp[j][0] = px[j];
if ( ( e = dct[i].n ) != 1 ) {
wb = dct[i].f;
dlist[i] = W_UMALLOC(DEG(wb)*e); cpyum(l[i],dlist[i]);
divum(m,dlist[i],wb,wq); DEG(dlist[i])= DEG(wq);
for ( k = 0; k <= DEG(wq); k++ )
COEF(dlist[i])[k] = dmb(m,COEF(wq)[k],e,&tmp);
}
}
for ( i = 1; i < b; i++ ) {
mullum(m,i+1,lcpa,llist[0],wb0);
for ( j = 1; j < np; j++ ) {
mullum(m,i+1,llist[j],wb0,wb1);
tlum = wb0; wb0 = wb1; wb1 = tlum;
}
for ( j = n, px = COEF(wt), pp = COEF(fpa), wpp = COEF(wb0);
j >= 0; j-- )
px[j] = ( pp[j][i] - wpp[j][i] + m ) % m;
degum(wt,n);
for ( j = n, px = COEF(wq0); j >= 0; j-- )
px[j] = 0;
for ( j = 1; j < np; j++ ) {
mulum(m,wt,c[j],wm); dr = divum(m,wm,l[j],wq);
for ( k = DEG(wq), px = COEF(wq0), py = COEF(wq); k >= 0; k-- )
px[k] = ( px[k] + py[k] ) % m;
for ( k = dr, pp = COEF(llist[j]), px = COEF(wm); k >= 0; k-- )
pp[k][i] = px[k];
}
degum(wq0,n); mulum(m,wq0,l[0],wm);
mulum(m,wt,c[0],wm0); addum(m,wm,wm0,wa);
for ( j = DEG(wa), pp = COEF(llist[0]), px = COEF(wa); j >= 0; j-- )
pp[j][i] = px[j];
for ( j = n, px = COEF(wq0); j >= 0; j-- )
px[j] = 0;
for ( j = 0; j < np; j++ )
if ( dct[j].n == 1 )
for ( k = 0,
pp = COEF(llist[j]),
wpp = COEF(((LUM *)list->c)[j]);
k <= DEG(llist[j]); k++ )
wpp[k][i] = pp[k][i];
else {
pwrlum(m,i+1,((LUM *)list->c)[j],dct[j].n,tpa);
for ( k = 0,
pp = COEF(llist[j]),
wpp = COEF(tpa);
k <= DEG(l[j]); k++ )
COEF(wt)[k] = (pp[k][i]-wpp[k][i]+m)%m;
degum(wt,DEG(l[j])); dr = divum(m,wt,dlist[j],ws);
if ( dr >= 0 ) {
*dcp = 0;
return;
} else
for ( k = 0,
pp = COEF(((LUM *)list->c)[j]);
k <= DEG(ws); k++ )
pp[k][i] = COEF(ws)[k];
}
list->bound = i+1; dtestsql(f,list,dct,&dc);
if ( dc ) {
*dcp = dc;
return;
}
}
*dcp = 0;
}
void gcdgen(P f,ML blist,ML *clistp)
{
register int i;
int n,d,mod,np;
UM wf,wm,wx,wy,wu,wv,wa,wb,wg,q,tum;
UM *in,*out;
ML clist;
n = UDEG(f); mod = blist->mod; np = blist->n;
d = 2*n;
q = W_UMALLOC(d); wf = W_UMALLOC(d);
wm = W_UMALLOC(d); wx = W_UMALLOC(d);
wy = W_UMALLOC(d); wu = W_UMALLOC(d);
wv = W_UMALLOC(d); wg = W_UMALLOC(d);
wa = W_UMALLOC(d); wb = W_UMALLOC(d);
ptoum(mod,f,wf); DEG(wg) = 0; COEF(wg)[0] = 1;
*clistp = clist = MLALLOC(np); clist->mod = mod; clist->n = np;
for ( i = 0, in = (UM *)blist->c, out = (UM *)clist->c; i < np; i++ ) {
divum(mod,wf,in[i],q); tum = wf; wf = q; q = tum;
cpyum(wf,wx); cpyum(in[i],wy);
eucum(mod,wx,wy,wa,wb); mulum(mod,wa,wg,wm);
DEG(wm) = divum(mod,wm,in[i],q); out[i] = UMALLOC(DEG(wm));
cpyum(wm,out[i]); mulum(mod,q,wf,wu);
mulum(mod,wg,wb,wv); addum(mod,wu,wv,wg);
}
}
/* find a,b s.t. qa*qg+qb*qh=1 mod q, qg=g mod mod, qh=h mod mod */
/* q = mod^k */
void henprep2(int mod,int q,int k,UM f,UM g,UM h,UM qg,UM qh,UM qa,UM qb)
{
int n;
UM wg,wh,wa,wb;
ML bl,cl,bql,cql;
P ff;
n = DEG(f);
wg = W_UMALLOC(2*n); wh = W_UMALLOC(2*n);
wa = W_UMALLOC(2*n); wb = W_UMALLOC(2*n);
cpyum(g,wg); cpyum(h,wh);
/* wa*g+wb*h = 1 mod mod */
eucum(mod,wg,wh,wa,wb);
#if 0
/* check */
wt = W_UMALLOC(2*n); ws = W_UMALLOC(2*n); wu = W_UMALLOC(2*n);
mulum(mod,wa,g,wt);
mulum(mod,wb,h,ws);
addum(mod,wt,ws,wu);
if ( DEG(wu) != 0 || COEF(wu)[0] != 1 )
error("henprep 1");
#endif
bl = MLALLOC(2); bl->n = 2; bl->mod = mod; bl->c[0] = g; bl->c[1] = h;
cl = MLALLOC(2); cl->n = 2; cl->mod = mod; cl->c[0] = wb; cl->c[1] = wa;
umtop(CO->v,f,&ff); /* XXX */
henprep(ff,bl,cl,&bql,&cql); /* XXX */
cpyum(bql->c[0],qg); cpyum(bql->c[1],qh);
cpyum(cql->c[0],qb); cpyum(cql->c[1],qa);
#if 0
/* check */
mulum(q,qa,qg,wt);
mulum(q,qb,qh,ws);
addum(q,wt,ws,wu);
if ( DEG(wu) != 0 || COEF(wu)[0] != 1 )
error("henprep 2");
#endif
}
/*
henprep(f,blist,clist,&bqlist,&cqlist);
*/
void henprep(P f,ML blist,ML clist,ML *bqlistp,ML *cqlistp)
{
register int i,j,k,*px,*py,*pz;
int n,pmax,dr,tmp,p,p1,mod,np,b,q;
UM w,wm,wn,wa,wt,wq,wf,quot,tum,*in,*inc,*out,*outc;
ML bqlist,cqlist;
n = UDEG(f); p = mod = blist->mod; np = blist->n;
/* for ( b = 1, q = mod; q <= (unsigned int)(LBASE / (L)mod); q *= mod, b++ ); */
for ( b = 1, q = mod; q <= ((1<<27) / mod); q *= mod, b++ );
w = W_UMALLOC(n); ptoum(q,f,w);
wm = W_UMALLOC(2*n); wn = W_UMALLOC(2*n);
wa = W_UMALLOC(2*n); wt = W_UMALLOC(2*n);
wq = W_UMALLOC(2*n); wf = W_UMALLOC(2*n);
quot = W_UMALLOC(2*n);
*bqlistp = bqlist = MLALLOC(np); *cqlistp = cqlist = MLALLOC(np);
for ( i = 0; i < n+2; i++ )
COEF(wq)[i] = 0;
for ( i = 0,
in = (UM *)blist->c, inc = (UM *)clist->c,
out = (UM *)bqlist->c, outc = (UM *)cqlist->c;
i < np; i++ ) {
out[i] = C_UMALLOC(n+1); cpyum(in[i],out[i]);
outc[i] = C_UMALLOC(n+1); cpyum(inc[i],outc[i]);
}
for ( pmax = 1, i = b; i > 0; i-- )
pmax *= mod;
for ( i = 1; i < b; i++, p = p1 ) {
cpyum(out[0],wm);
for ( j = 1; j < np; j++ ) {
mulum(pmax,wm,out[j],wn);
tum = wm; wm = wn; wn = tum;
}
for ( j = n, px = COEF(w), py = COEF(wm), pz = COEF(wt); j >= 0; j-- ) {
tmp = ( ( px[j] - py[j] ) / p ) % mod;
pz[j] = ( tmp >= 0? tmp : tmp + mod );
}
degum(wt,n);
for ( j = 1; j < np; j++ ) {
mulum(mod,wt,inc[j],wm); dr = divum(mod,wm,in[j],quot);
for ( k = DEG(quot); k >= 0; k-- )
COEF(wq)[k] = ( COEF(wq)[k] + COEF(quot)[k] ) % mod;
for ( k = dr, px = COEF(out[j]), py = COEF(wm); k >= 0; k-- )
px[k] += p * py[k];
}
degum(wq,n); mulum(mod,wq,in[0],wm);
mulum(mod,wt,inc[0],wn); addum(mod,wm,wn,wa);
for ( j = DEG(wa), px = COEF(out[0]), py = COEF(wa); j >= 0; j-- )
px[j] += p * py[j];
for ( j = n, px = COEF(wq); j >= 0; j-- )
px[j] = 0;
p1 = p * mod;
for ( j = n, px = COEF(wt); j >= 1; j-- )
px[j] = 0;
px[0] = 1;
for ( j = 0; j < np; j++ ) {
cpyum(w,wf);
for ( k = DEG(wf), px = COEF(wf); k >= 0; k-- )
px[k] %= p1;
divum(p1,wf,out[j],quot); mulum(p1,outc[j],quot,wm);
for ( k = DEG(wm), px = COEF(wt), py = COEF(wm); k >= 0; k-- )
px[k] = ( px[k] - py[k] ) % p1;
}
degum(wt,n);
for ( j = DEG(wt), px = COEF(wt); j >= 0; j-- )
px[j] = ((tmp=(px[j]/p)%mod)>= 0?tmp:tmp + mod);
for ( j = 0; j < np; j++ ) {
mulum(mod,wt,outc[j],wm); dr = divum(mod,wm,in[j],quot);
for ( k = dr, px = COEF(outc[j]), py = COEF(wm); k >= 0; k-- )
px[k] += p * py[k];
degum(outc[j],MAX(DEG(outc[j]),dr));
}
}
bqlist->n = cqlist->n = np;
bqlist->mod = cqlist->mod = q;
}
/*
henmain(fl,bqlist,cqlist,listp)
*/
void henmain(LUM f,ML bqlist,ML cqlist,ML *listp)
{
register int i,j,k;
int *px,*py;
int **pp,**pp1;
int n,np,mod,bound,dr,tmp;
UM wt,wq0,wq,wr,wm,wm0,wa,q;
LUM wb0,wb1,tlum;
UM *b,*c;
LUM *l;
ML list;
n = DEG(f); np = bqlist->n; mod = bqlist->mod; bound = bqlist->bound;
*listp = list = MLALLOC(n);
list->n = np; list->mod = mod; list->bound = bound;
W_LUMALLOC(n,bound,wb0); W_LUMALLOC(n,bound,wb1);
wt = W_UMALLOC(n); wq0 = W_UMALLOC(n); wq = W_UMALLOC(n);
wr = W_UMALLOC(n); wm = W_UMALLOC(2*n); wm0 = W_UMALLOC(2*n);
wa = W_UMALLOC(2*n); q = W_UMALLOC(2*n);
b = (UM *)bqlist->c; c = (UM *)cqlist->c; l = (LUM *)list->c;
for ( i = 0; i < np; i++ ) {
l[i] = LUMALLOC(DEG(b[i]),bound);
for ( j = DEG(b[i]), pp = COEF(l[i]), px = COEF(b[i]); j >= 0; j-- )
pp[j][0] = px[j];
}
#if 0
fprintf(stderr,"bound=%d\n",bound);
#endif
for ( i = 1; i < bound; i++ ) {
#if 0
fprintf(stderr,".");
#endif
#if defined(VISUAL) || defined(__MINGW32__) || defined(__MINGW64__)
check_intr();
#endif
mullum(mod,i+1,l[0],l[1],wb0);
for ( j = 2; j < np; j++ ) {
mullum(mod,i+1,l[j],wb0,wb1);
tlum = wb0; wb0 = wb1; wb1 = tlum;
}
for ( j = n, px = COEF(wt); j >= 0; j-- )
px[j] = 0;
for ( j = n, pp = COEF(f), pp1 = COEF(wb0); j >= 0; j-- ) {
tmp = ( pp[j][i] - pp1[j][i] ) % mod;
COEF(wt)[j] = ( tmp < 0 ? tmp + mod : tmp );
}
degum(wt,n);
for ( j = n, px = COEF(wq0); j >= 0; j-- )
px[j] = 0;
for ( j = 1; j < np; j++ ) {
mulum(mod,wt,c[j],wm); dr = divum(mod,wm,b[j],q);
for ( k = DEG(q), px = COEF(wq0), py = COEF(q); k >= 0; k-- )
px[k] = ( px[k] + py[k] ) % mod;
for ( k = dr, pp = COEF(l[j]), px = COEF(wm); k >= 0; k-- )
pp[k][i] = px[k];
}
degum(wq0,n); mulum(mod,wq0,b[0],wm);
mulum(mod,wt,c[0],wm0); addum(mod,wm,wm0,wa);
for ( j = DEG(wa), pp = COEF(l[0]), px = COEF(wa); j >= 0; j-- )
pp[j][i] = px[j];
for ( j = n, px = COEF(wq0); j >= 0; j-- )
px[j] = 0;
}
#if 0
fprintf(stderr,"\n");
#endif
}
/*
henmain_incremental(fl,bqlist,cqlist,start)
fl = bqlist[0]*... mod q^start
*/
void henmain_incremental(LUM f,LUM *bqlist,ML cqlist,
int np, int mod, int start, int bound)
{
register int i,j,k;
int *px,*py;
int **pp,**pp1;
int n,dr,tmp;
UM wt,wq0,wq,wr,wm,wm0,wa,q;
LUM wb0,wb1,tlum;
UM *b,*c;
LUM *l;
ML list;
n = DEG(f);
W_LUMALLOC(n,bound,wb0); W_LUMALLOC(n,bound,wb1);
wt = W_UMALLOC(n); wq0 = W_UMALLOC(n); wq = W_UMALLOC(n);
wr = W_UMALLOC(n); wm = W_UMALLOC(2*n); wm0 = W_UMALLOC(2*n);
wa = W_UMALLOC(2*n); q = W_UMALLOC(2*n);
c = (UM *)cqlist->c; l = bqlist;
b = (UM *)ALLOCA(n*sizeof(UM));
for ( i = 0; i < np; i++ ) {
j = DEG(l[i]);
b[i] = W_UMALLOC(j);
DEG(b[i]) = j;
for ( pp = COEF(l[i]), px = COEF(b[i]); j >= 0; j-- )
px[j] = pp[j][0];
}
#if 0
fprintf(stderr,"bound=%d\n",bound);
#endif
for ( i = start; i < bound; i++ ) {
#if 0
fprintf(stderr,".");
#endif
mullum(mod,i+1,l[0],l[1],wb0);
for ( j = 2; j < np; j++ ) {
mullum(mod,i+1,l[j],wb0,wb1);
tlum = wb0; wb0 = wb1; wb1 = tlum;
}
for ( j = n, px = COEF(wt); j >= 0; j-- )
px[j] = 0;
for ( j = n, pp = COEF(f), pp1 = COEF(wb0); j >= 0; j-- ) {
tmp = ( pp[j][i] - pp1[j][i] ) % mod;
COEF(wt)[j] = ( tmp < 0 ? tmp + mod : tmp );
}
degum(wt,n);
for ( j = n, px = COEF(wq0); j >= 0; j-- )
px[j] = 0;
for ( j = 1; j < np; j++ ) {
mulum(mod,wt,c[j],wm); dr = divum(mod,wm,b[j],q);
for ( k = DEG(q), px = COEF(wq0), py = COEF(q); k >= 0; k-- )
px[k] = ( px[k] + py[k] ) % mod;
for ( k = dr, pp = COEF(l[j]), px = COEF(wm); k >= 0; k-- )
pp[k][i] = px[k];
}
degum(wq0,n); mulum(mod,wq0,b[0],wm);
mulum(mod,wt,c[0],wm0); addum(mod,wm,wm0,wa);
for ( j = DEG(wa), pp = COEF(l[0]), px = COEF(wa); j >= 0; j-- )
pp[j][i] = px[j];
for ( j = n, px = COEF(wq0); j >= 0; j-- )
px[j] = 0;
}
#if 0
fprintf(stderr,"\n");
#endif
}
static double M;
static int E;
int mignotte(int q,P f)
{
int p;
unsigned int *b;
N c;
DCP dc;
for ( dc = DC(f), M = 0, E = 0; dc; dc = NEXT(dc) ) {
c = NM((Q)COEF(dc)); p = PL(c); b = BD(c);
sqad(b[p-1],(p-1)*BSH);
}
if (E % 2) M *= 2; M = ceil(sqrt(M)); E /= 2;
c = NM((Q)COEF(DC(f))); p = PL(c); M *= ((double)BD(c)[p-1]+1.0); E += (p-1) * BSH;
return (int)ceil( (0.31*(E+UDEG(f)+1)+log10((double)M)) / log10((double)q) );
}
int mig(int q,int d,P f)
{
int p;
unsigned int *b;
N c;
DCP dc;
for ( dc = DC(f), M = 0, E = 0; dc; dc = NEXT(dc) ) {
c = NM((Q)COEF(dc)); p = PL(c); b = BD(c);
sqad(b[p-1],(p-1)*BSH);
}
if (E % 2) M *= 2; M = ceil(sqrt(M)); E /= 2;
c = NM((Q)COEF(DC(f))); p = PL(c);
M *= (BD(c)[p-1]+1); E += (p-1) * BSH;
return (int)ceil( (0.31*(E+d+1)+log10((double)M)) / log10((double)q) );
}
void sqad(unsigned int man,int exp)
{
int e,sqe;
unsigned int t;
double man1,d,sqm;
int diff;
if ( man == BMASK ) {
e = BSH; man1 = 1.0;
} else {
man += 1;
for ( e = 0, t = man; t; e++, t >>= 1 );
e--; d = (double)(1<<e);
man1 = ((double)man)/d;
}
exp += e; sqm = man1 * man1; sqe = 2 * exp;
if ( sqm >= 2.0 ) {
sqm /= 2.0; sqe++;
}
diff = E - sqe;
if ( diff > 18 )
return;
if ( diff < -18 ) {
M = sqm; E = sqe;
return;
}
if ( diff >= 0 )
M += (sqm / (double)(1<<diff));
else {
M = ( ( M / (double)(1<<-diff)) + sqm ); E = sqe;
}
if ( M >= 2.0 ) {
M /= 2.0; E++;
}
}
void ptolum(int q,int bound,P f,LUM fl)
{
DCP dc;
int i,j;
int **pp;
int d,br,s;
unsigned int r;
int *c;
unsigned int *m,*w;
for ( dc = DC(f), pp = COEF(fl); dc; dc = NEXT(dc) ) {
d = PL(NM((Q)COEF(dc))); m = BD(NM((Q)COEF(dc)));
c = pp[QTOS(DEG(dc))]; w = (unsigned int *)W_ALLOC(d);
for ( i = 0; i < d; i++ )
w[i] = m[i];
for ( i = 0; i < bound && d >= 1; ) {
for ( j = d - 1, r = 0; j >= 0; j-- ) {
DSAB(q,r,w[j],w[j],r)
}
c[i++] = (int)r;
if ( !w[d-1] )
d--;
}
if ( SGN((Q)COEF(dc)) < 0 )
for (i = 0, br = 0; i < bound; i++ )
if ( ( s = -(c[i] + br) ) < 0 ) {
c[i] = s + q; br = 1;
} else {
c[i] = 0; br = 0;
}
}
}
void modfctrp(P p,int mod,int flag,DCP *dcp)
{
int cm,n,i,j,k;
DCP dc,dc0;
P zp;
Q c,q;
UM mp;
UM *tl;
struct oDUM *udc,*udc1;
if ( !p ) {
*dcp = 0; return;
}
ptozp(p,1,&c,&zp);
if ( DN(c) || !(cm = rem(NM(c),mod)) ) {
*dcp = 0; return;
}
mp = W_UMALLOC(UDEG(p));
ptoum(mod,zp,mp);
if ( (n = DEG(mp)) < 0 ) {
*dcp = 0; return;
} else if ( n == 0 ) {
cm = dmar(cm,COEF(mp)[0],0,mod); STOQ(cm,q);
NEWDC(dc); COEF(dc) = (P)q; DEG(dc) = ONE;
NEXT(dc) = 0; *dcp = dc;
return;
}
if ( COEF(mp)[n] != 1 ) {
cm = dmar(cm,COEF(mp)[n],0,mod);
i = invm(COEF(mp)[n],mod);
for ( j = 0; j <= n; j++ )
COEF(mp)[j] = dmar(COEF(mp)[j],i,0,mod);
}
W_CALLOC(n+1,struct oDUM,udc);
gensqfrum(mod,mp,udc);
switch ( flag ) {
case FCTR:
tl = (UM *)ALLOCA((n+1)*sizeof(UM));
W_CALLOC(DEG(mp)+1,struct oDUM,udc1);
for ( i = 0,j = 0; udc[i].f; i++ )
if ( DEG(udc[i].f) == 1 ) {
udc1[j].f = udc[i].f; udc1[j].n = udc[i].n; j++;
} else {
bzero((char *)tl,(n+1)*sizeof(UM));
berlemain(mod,udc[i].f,tl);
for ( k = 0; tl[k]; k++, j++ ) {
udc1[j].f = tl[k]; udc1[j].n = udc[i].n;
}
}
udc = udc1; break;
case SQFR:
break;
case DDD:
tl = (UM *)ALLOCA((n+1)*sizeof(UM));
W_CALLOC(DEG(mp)+1,struct oDUM,udc1);
for ( i = 0,j = 0; udc[i].f; i++ )
if ( DEG(udc[i].f) == 1 ) {
udc1[j].f = udc[i].f; udc1[j].n = udc[i].n; j++;
} else {
bzero((char *)tl,(n+1)*sizeof(UM));
ddd(mod,udc[i].f,tl);
for ( k = 0; tl[k]; k++, j++ ) {
udc1[j].f = tl[k]; udc1[j].n = udc[i].n;
}
}
udc = udc1; break;
case NEWDDD:
tl = (UM *)ALLOCA((n+1)*sizeof(UM));
W_CALLOC(DEG(mp)+1,struct oDUM,udc1);
for ( i = 0,j = 0; udc[i].f; i++ )
if ( DEG(udc[i].f) == 1 ) {
udc1[j].f = udc[i].f; udc1[j].n = udc[i].n; j++;
} else {
bzero((char *)tl,(n+1)*sizeof(UM));
if ( mod == 2 )
berlemain(mod,udc[i].f,tl);
else
newddd(mod,udc[i].f,tl);
for ( k = 0; tl[k]; k++, j++ ) {
udc1[j].f = tl[k]; udc1[j].n = udc[i].n;
}
}
udc = udc1; break;
}
NEWDC(dc0); STOQ(cm,q); COEF(dc0) = (P)q; DEG(dc0) = ONE; dc = dc0;
for ( n = 0; udc[n].f; n++ ) {
NEWDC(NEXT(dc)); dc = NEXT(dc);
STOQ(udc[n].n,DEG(dc)); umtop(VR(p),udc[n].f,&COEF(dc));
}
NEXT(dc) = 0; *dcp = dc0;
}
void gensqfrum(int mod,UM p,struct oDUM *dc)
{
int n,i,j,d;
UM t,s,g,f,f1,b;
if ( (n = DEG(p)) == 1 ) {
dc[0].f = UMALLOC(DEG(p)); cpyum(p,dc[0].f); dc[0].n = 1;
return;
}
t = W_UMALLOC(n); s = W_UMALLOC(n); g = W_UMALLOC(n);
f = W_UMALLOC(n); f1 = W_UMALLOC(n); b = W_UMALLOC(n);
diffum(mod,p,t); cpyum(p,s); Gcdum(mod,t,s,g);
if ( !DEG(g) ) {
dc[0].f = UMALLOC(DEG(p)); cpyum(p,dc[0].f); dc[0].n = 1;
return;
}
cpyum(p,b); cpyum(p,t); Divum(mod,t,g,f);
for ( i = 0, d = 0; DEG(f); i++ ) {
while ( 1 ) {
cpyum(b,t);
if ( Divum(mod,t,f,s) >= 0 )
break;
else {
cpyum(s,b); d++;
}
}
cpyum(b,t); cpyum(f,s); Gcdum(mod,t,s,f1);
Divum(mod,f,f1,s); cpyum(f1,f);
dc[i].f = UMALLOC(DEG(s)); cpyum(s,dc[i].f); dc[i].n = d;
}
if ( DEG(b) > 0 ) {
d = 1;
while ( 1 ) {
cpyum(b,t);
for ( j = DEG(t); j >= 0; j-- )
if ( COEF(t)[j] && (j % mod) )
break;
if ( j >= 0 )
break;
else {
DEG(s) = DEG(t)/mod;
for ( j = 0; j <= DEG(t); j++ )
COEF(s)[j] = COEF(t)[j*mod];
cpyum(s,b); d *= mod;
}
}
gensqfrum(mod,b,dc+i);
for ( j = i; dc[j].f; j++ )
dc[j].n *= d;
}
}
#if 0
void srchum(int mod,UM p1,UM p2,UM gr)
{
UM m,m1,m2,q,r,t,g1,g2;
int lc,d,d1,d2,i,j,k,l,l1,l2,l3,tmp,adj;
V v;
d = MAX(DEG(p1),DEG(p2));
g1 = W_UMALLOC(d); g2 = W_UMALLOC(d);
bzero((char *)g1,(d+2)*sizeof(int)); bzero((char *)g2,(d+2)*sizeof(int));
if ( d == DEG(p1) ) {
cpyum(p1,g1); cpyum(p2,g2);
} else {
cpyum(p1,g2); cpyum(p2,g1);
}
if ( ( d1 = DEG(g1) ) > ( d2 = DEG(g2) ) ) {
j = d1 - 1; adj = 1;
} else
j = d2;
lc = 1;
r = W_UMALLOC(d1+d2); q = W_UMALLOC(d1+d2);
m1 = W_UMALLOC(d1+d2); t = W_UMALLOC(d1+d2);
bzero((char *)r,(d1+d2+2)*sizeof(int)); bzero((char *)q,(d1+d2+2)*sizeof(int));
bzero((char *)m1,(d1+d2+2)*sizeof(int)); bzero((char *)t,(d1+d2+2)*sizeof(int));
m = W_UMALLOC(0); bzero((char *)m,2*sizeof(int));
adj = pwrm(mod,COEF(g2)[DEG(g2)],DEG(g1));
DEG(m) = 0; COEF(m)[0] = invm(COEF(g2)[DEG(g2)],mod);
Mulum(mod,g2,m,r); cpyum(r,g2);
while ( 1 ) {
if ( ( k = DEG(g2) ) < 0 ) {
DEG(gr) = -1;
return;
}
if ( k == j ) {
if ( k == 0 ) {
DEG(m) = 0; COEF(m)[0] = adj;
Mulum(mod,g2,m,gr);
return;
} else {
DEG(m) = 0;
COEF(m)[0] = pwrm(mod,COEF(g2)[k],DEG(g1)-k+1);
Mulum(mod,g1,m,r); DEG(r) = Divum(mod,r,g2,t);
DEG(m) = 0; COEF(m)[0] = dmb(mod,lc,lc,&tmp);
Divum(mod,r,m,q); cpyum(g2,g1); cpyum(q,g2);
lc = COEF(g1)[DEG(g1)]; j = k - 1;
}
} else {
d = j - k;
DEG(m) = 0; COEF(m)[0] = pwrm(mod,COEF(g2)[DEG(g2)],d);
Mulum(mod,g2,m,m1); l = pwrm(mod,lc,d);
DEG(m) = 0; COEF(m)[0] = l; Divum(mod,m1,m,t);
if ( k == 0 ) {
DEG(m) = 0; COEF(m)[0] = adj;
Mulum(mod,t,m,gr);
return;
} else {
DEG(m) = 0;
COEF(m)[0] = pwrm(mod,COEF(g2)[k],DEG(g1)-k+1);
Mulum(mod,g1,m,r); DEG(r) = Divum(mod,r,g2,q);
l1 = dmb(mod,lc,lc,&tmp); l2 = dmb(mod,l,l1,&tmp);
DEG(m) = 0; COEF(m)[0] = l2;
Divum(mod,r,m,q); cpyum(t,g1); cpyum(q,g2);
if ( d % 2 )
for ( i = DEG(g2); i >= 0; i-- )
COEF(g2)[i] = ( mod - COEF(g2)[i] ) % mod;
lc = COEF(g1)[DEG(g1)]; j = k - 1;
}
}
}
}
UM *resberle(int mod,UM f,UM *fp)
{
UM w,wg,ws,wf,f0,gcd,q,res;
int n;
register int i;
n = DEG(f); wg = W_UMALLOC(n); mini(mod,f,wg);
if ( DEG(wg) <= 0 ) {
f0 = UMALLOC(n); cpyum(f,f0); *fp++ = f0;
return ( fp );
}
f0 = W_UMALLOC(n); cpyum(f,f0);
ws = W_UMALLOC(n); wf = W_UMALLOC(n);
q = W_UMALLOC(n); gcd = W_UMALLOC(n);
res = W_UMALLOC(2*n);
srchum(mod,f,wg,res);
for ( i = 0; i < mod; i++ ) {
if ( substum(mod,res,i) )
continue;
cpyum(f0,wf); cpyum(wg,ws);
COEF(ws)[0] = ( COEF(ws)[0] + mod - i ) % mod;
Gcdum(mod,wf,ws,gcd);
if ( DEG(gcd) > 0 ) {
if ( DEG(gcd) < n ) {
Divum(mod,f0,gcd,q); f0 = q; fp = resberle(mod,gcd,fp);
}
break;
}
}
fp = resberle(mod,f0,fp);
return ( fp );
}
int substum(int mod,UM p,int a)
{
int i,j,s;
int *c;
if ( DEG(p) < 0 )
return 0;
if ( DEG(p) == 0 )
return COEF(p)[0];
for ( i = DEG(p), c = COEF(p), s = c[i]; i >= 0; ) {
for ( j = i--; (i>=0) && !c[i]; i-- );
if ( i >= 0 )
s = (s*pwrm(mod,a,j-i)%mod+c[i])%mod;
else
s = s*pwrm(mod,a,j)%mod;
}
return s;
}
#endif
void ddd(int mod,UM f,UM *r)
{
register int i,j;
int d,n;
UM q,s,t,u,v,w,g,x,m;
UM *base;
n = DEG(f);
if ( n == 1 ) {
r[0] = UMALLOC(1); cpyum(f,r[0]); r[1] = 0; return;
}
base = (UM *)ALLOCA(n*sizeof(UM)); bzero((char *)base,n*sizeof(UM));
w = W_UMALLOC(2*n); q = W_UMALLOC(2*n); m = W_UMALLOC(2*n);
base[0] = W_UMALLOC(0); DEG(base[0]) = 0; COEF(base[0])[0] = 1;
t = W_UMALLOC(1); DEG(t) = 1; COEF(t)[0] = 0; COEF(t)[1] = 1;
pwrmodum(mod,t,mod,f,w); base[1] = W_UMALLOC(DEG(w)); cpyum(w,base[1]);
for ( i = 2; i < n; i++ ) {
mulum(mod,base[i-1],base[1],m); DEG(m) = divum(mod,m,f,q);
base[i] = W_UMALLOC(DEG(m)); cpyum(m,base[i]);
}
v = W_UMALLOC(n); cpyum(f,v);
DEG(w) = 1; COEF(w)[0] = 0; COEF(w)[1] = 1;
x = W_UMALLOC(1); DEG(x) = 1; COEF(x)[0] = 0; COEF(x)[1] = 1;
t = W_UMALLOC(n); s = W_UMALLOC(n); u = W_UMALLOC(n); g = W_UMALLOC(n);
for ( j = 0, d = 1; 2*d <= DEG(v); d++ ) {
for ( DEG(t) = -1, i = 0; i <= DEG(w); i++ )
if ( COEF(w)[i] ) {
Mulsum(mod,base[i],COEF(w)[i],s);
addum(mod,s,t,u); cpyum(u,t);
}
cpyum(t,w); cpyum(v,s); subum(mod,w,x,t); Gcdum(mod,s,t,g);
if ( DEG(g) >= 1 ) {
canzas(mod,g,d,base,r+j); j += DEG(g)/d;
Divum(mod,v,g,q); cpyum(q,v);
DEG(w) = Divum(mod,w,v,q);
for ( i = 0; i < DEG(v); i++ )
DEG(base[i]) = Divum(mod,base[i],v,q);
}
}
if ( DEG(v) ) {
r[j] = UMALLOC(DEG(v)); cpyum(v,r[j]); j++;
}
r[j] = 0;
}
#if 0
void canzas(int mod,UM f,int d,UM *base,UM *r)
{
UM t,s,u,w,g,o,q;
N n1,n2,n3,n4,n5;
UM *b;
int n,m,i;
if ( DEG(f) == d ) {
r[0] = UMALLOC(d); cpyum(f,r[0]);
return;
} else {
n = DEG(f); b = (UM *)ALLOCA(n*sizeof(UM)); bzero((char *)b,n*sizeof(UM));
for ( i = 0, m = 0; i < n; i++ )
m = MAX(m,DEG(base[i]));
q = W_UMALLOC(m);
for ( i = 0; i < n; i++ ) {
b[i] = W_UMALLOC(DEG(base[i])); cpyum(base[i],b[i]);
DEG(b[i]) = Divum(mod,b[i],f,q);
}
t = W_UMALLOC(2*d);
s = W_UMALLOC(DEG(f)); u = W_UMALLOC(DEG(f));
w = W_UMALLOC(DEG(f)); g = W_UMALLOC(DEG(f));
o = W_UMALLOC(0); DEG(o) = 0; COEF(o)[0] = 1;
STON(mod,n1); pwrn(n1,d,&n2); subn(n2,ONEN,&n3);
STON(2,n4); divsn(n3,n4,&n5);
while ( 1 ) {
randum(mod,2*d,t); spwrum(mod,f,b,t,n5,s);
subum(mod,s,o,u); cpyum(f,w); Gcdum(mod,w,u,g);
if ( (DEG(g) >= 1) && (DEG(g) < DEG(f)) ) {
canzas(mod,g,d,b,r);
cpyum(f,w); Divum(mod,w,g,s);
canzas(mod,s,d,b,r+DEG(g)/d);
return;
}
}
}
}
#else
void canzas(int mod,UM f,int d,UM *base,UM *r)
{
UM t,s,u,w,g,o,q;
N n1,n2,n3,n4,n5;
UM *b;
int n,m,i;
if ( DEG(f) == d ) {
r[0] = UMALLOC(d); cpyum(f,r[0]);
return;
} else {
n = DEG(f); b = (UM *)ALLOCA(n*sizeof(UM)); bzero((char *)b,n*sizeof(UM));
for ( i = 0, m = 0; i < n; i++ )
m = MAX(m,DEG(base[i]));
q = W_UMALLOC(m);
for ( i = 0; i < n; i++ ) {
b[i] = W_UMALLOC(DEG(base[i])); cpyum(base[i],b[i]);
DEG(b[i]) = Divum(mod,b[i],f,q);
}
t = W_UMALLOC(2*d);
s = W_UMALLOC(DEG(f)); u = W_UMALLOC(DEG(f));
w = W_UMALLOC(DEG(f)); g = W_UMALLOC(DEG(f));
o = W_UMALLOC(0); DEG(o) = 0; COEF(o)[0] = 1;
STON(mod,n1); pwrn(n1,d,&n2); subn(n2,ONEN,&n3);
STON(2,n4); divsn(n3,n4,&n5);
while ( 1 ) {
randum(mod,2*d,t); spwrum0(mod,f,t,n5,s);
subum(mod,s,o,u); cpyum(f,w); Gcdum(mod,w,u,g);
if ( (DEG(g) >= 1) && (DEG(g) < DEG(f)) ) {
canzas(mod,g,d,b,r);
cpyum(f,w); Divum(mod,w,g,s);
canzas(mod,s,d,b,r+DEG(g)/d);
return;
}
}
}
}
#endif
void randum(int mod,int d,UM p)
{
unsigned int n;
int i;
n = ((unsigned int)random()) % d; DEG(p) = n; COEF(p)[n] = 1;
for ( i = 0; i < (int)n; i++ )
COEF(p)[i] = ((unsigned int)random()) % mod;
}
void pwrmodum(int mod,UM p,int e,UM f,UM pr)
{
UM wt,ws,q;
if ( e == 0 ) {
DEG(pr) = 0; COEF(pr)[0] = 1;
} else if ( DEG(p) < 0 )
DEG(pr) = -1;
else if ( e == 1 ) {
q = W_UMALLOC(DEG(p)); cpyum(p,pr);
DEG(pr) = divum(mod,pr,f,q);
} else if ( DEG(p) == 0 ) {
DEG(pr) = 0; COEF(pr)[0] = pwrm(mod,COEF(p)[0],e);
} else {
wt = W_UMALLOC(2*DEG(f)); ws = W_UMALLOC(2*DEG(f));
q = W_UMALLOC(2*DEG(f));
pwrmodum(mod,p,e/2,f,wt);
if ( !(e%2) ) {
mulum(mod,wt,wt,pr); DEG(pr) = divum(mod,pr,f,q);
} else {
mulum(mod,wt,wt,ws); DEG(ws) = divum(mod,ws,f,q);
mulum(mod,ws,p,pr); DEG(pr) = divum(mod,pr,f,q);
}
}
}
void spwrum(int mod,UM m,UM *base,UM f,N e,UM r)
{
int a,n,i;
N e1,an;
UM t,s,u,q,r1,r2;
if ( !e ) {
DEG(r) = 0; COEF(r)[0] = 1;
} else if ( UNIN(e) )
cpyum(f,r);
else if ( (PL(e) == 1) && (BD(e)[0] < (unsigned int)mod) )
spwrum0(mod,m,f,e,r);
else {
a = divin(e,mod,&e1); STON(a,an);
n = DEG(m); t = W_UMALLOC(n); s = W_UMALLOC(n);
u = W_UMALLOC(2*n); q = W_UMALLOC(2*n);
for ( DEG(t) = -1, i = 0; i <= DEG(f); i++ )
if ( COEF(f)[i] ) {
Mulsum(mod,base[i],COEF(f)[i],s);
addum(mod,s,t,u); cpyum(u,t);
}
r1 = W_UMALLOC(n); spwrum0(mod,m,f,an,r1);
r2 = W_UMALLOC(n); spwrum(mod,m,base,t,e1,r2);
Mulum(mod,r1,r2,u); DEG(u) = Divum(mod,u,m,q);
cpyum(u,r);
}
}
void spwrum0(int mod,UM m,UM f,N e,UM r)
{
UM t,s,q;
N e1;
int a;
if ( !e ) {
DEG(r) = 0; COEF(r)[0] = 1;
} else if ( UNIN(e) )
cpyum(f,r);
else {
a = divin(e,2,&e1);
t = W_UMALLOC(2*DEG(m)); spwrum0(mod,m,f,e1,t);
s = W_UMALLOC(2*DEG(m)); q = W_UMALLOC(2*DEG(m));
Mulum(mod,t,t,s); DEG(s) = Divum(mod,s,m,q);
if ( a ) {
Mulum(mod,s,f,t); DEG(t) = Divum(mod,t,m,q); cpyum(t,r);
} else
cpyum(s,r);
}
}
#if 0
void Mulum(int mod,UM p1,UM p2,UM pr)
{
register int *pc1,*pcr;
register int mul,i,j,d1,d2;
int *c1,*c2,*cr;
if ( ( (d1 = DEG(p1)) < 0) || ( (d2 = DEG(p2)) < 0 ) ) {
DEG(pr) = -1;
return;
}
c1 = COEF(p1); c2 = COEF(p2); cr = COEF(pr);
bzero((char *)cr,(d1+d2+1)*sizeof(int));
for ( i = 0; i <= d2; i++, cr++ )
if ( mul = *c2++ )
for ( j = 0, pc1 = c1, pcr = cr; j <= d1; j++, pc1++, pcr++ )
*pcr = (*pc1 * mul + *pcr) % mod;
DEG(pr) = d1 + d2;
}
void Mulsum(int mod,UM p,int n,UM pr)
{
register int *sp,*dp;
register int i;
for ( i = DEG(pr) = DEG(p), sp = COEF(p)+i, dp = COEF(pr)+i;
i >= 0; i--, dp--, sp-- )
*dp = (*sp * n) % mod;
}
int Divum(int mod,UM p1,UM p2,UM pq)
{
register int *pc1,*pct;
register int tmp,i,j,inv;
int *c1,*c2,*ct;
int d1,d2,dd,hd;
if ( (d1 = DEG(p1)) < (d2 = DEG(p2)) ) {
DEG(pq) = -1;
return( d1 );
}
c1 = COEF(p1); c2 = COEF(p2); dd = d1-d2;
if ( ( hd = c2[d2] ) != 1 ) {
inv = invm(hd,mod);
for ( pc1 = c2 + d2; pc1 >= c2; pc1-- )
*pc1 = (*pc1 * inv) % mod;
} else
inv = 1;
for ( i = dd, ct = c1+d1; i >= 0; i-- )
if ( tmp = *ct-- ) {
tmp = mod - tmp;
for ( j = d2-1, pct = ct, pc1 = c2+j; j >= 0; j--, pct--, pc1-- )
*pct = (*pc1 * tmp + *pct) % mod;
}
if ( inv != 1 ) {
for ( pc1 = c1+d2, pct = c1+d1; pc1 <= pct; pc1++ )
*pc1 = (*pc1 * inv) % mod;
for ( pc1 = c2, pct = c2+d2, inv = hd; pc1 <= pct; pc1++ )
*pc1 = (*pc1 * inv) % mod;
}
for ( i = d2-1, pc1 = c1+i; i >= 0 && !(*pc1); pc1--, i-- );
for ( DEG(pq) = j = dd, pc1 = c1+d1, pct = COEF(pq)+j; j >= 0; j-- )
*pct-- = *pc1--;
return( i );
}
void Gcdum(int mod,UM p1,UM p2,UM pr)
{
register int *sp,*dp;
register int i,inv;
UM t1,t2,q,tum;
int drem;
if ( DEG(p1) < 0 )
cpyum(p2,pr);
else if ( DEG(p2) < 0 )
cpyum(p1,pr);
else {
if ( DEG(p1) >= DEG(p2) ) {
t1 = p1; t2 = p2;
} else {
t1 = p2; t2 = p1;
}
q = W_UMALLOC(DEG(t1));
while ( ( drem = Divum(mod,t1,t2,q) ) >= 0 ) {
tum = t1; t1 = t2; t2 = tum; DEG(t2) = drem;
}
inv = invm(COEF(t2)[DEG(t2)],mod);
Mulsum(mod,t2,inv,pr);
}
}
#endif
void mult_mod_tab(UM p,int mod,UM *tab,UM r,int d)
{
UM w,w1,c;
int n,i;
int *pc;
w = W_UMALLOC(d); w1 = W_UMALLOC(d);
c = W_UMALLOC(1); DEG(c) = 0;
n = DEG(p); DEG(r) = -1;
for ( i = 0, pc = COEF(p); i <= n; i++ )
if ( pc[i] ) {
COEF(c)[0] = pc[i];
mulum(mod,tab[i],c,w);
addum(mod,r,w,w1);
cpyum(w1,r);
}
}
void make_qmat(UM p,int mod,UM *tab,int ***mp)
{
int n,i,j;
int *c;
UM q,r;
int **mat;
n = DEG(p);
*mp = mat = almat(n,n);
for ( j = 0; j < n; j++ ) {
r = W_UMALLOC(DEG(tab[j])); q = W_UMALLOC(DEG(tab[j]));
cpyum(tab[j],r); DEG(r) = divum(mod,r,p,q);
for ( i = 0, c = COEF(r); i <= DEG(r); i++ )
mat[i][j] = c[i];
}
for ( i = 0; i < n; i++ )
mat[i][i] = (mat[i][i]+mod-1) % mod;
}
void null_mod(int **mat,int mod,int n,int *ind)
{
int i,j,l,s,h,inv;
int *t,*u;
bzero((char *)ind,n*sizeof(int));
ind[0] = 0;
for ( i = j = 0; j < n; i++, j++ ) {
for ( ; j < n; j++ ) {
for ( l = i; l < n; l++ )
if ( mat[l][j] )
break;
if ( l < n ) {
t = mat[i]; mat[i] = mat[l]; mat[l] = t; break;
} else
ind[j] = 1;
}
if ( j == n )
break;
inv = invm(mat[i][j],mod);
for ( s = j, t = mat[i]; s < n; s++ )
t[s] = dmar(t[s],inv,0,mod);
for ( l = 0; l < n; l++ ) {
if ( l == i )
continue;
for ( s = j, u = mat[l], h = (mod-u[j])%mod; s < n; s++ )
u[s] = dmar(h,t[s],u[s],mod);
}
}
}
void null_to_sol(int **mat,int *ind,int mod,int n,UM *r)
{
int i,j,k,l;
int *c;
UM w;
for ( i = 0, l = 0; i < n; i++ ) {
if ( !ind[i] )
continue;
w = UMALLOC(n);
for ( j = k = 0, c = COEF(w); j < n; j++ )
if ( ind[j] )
c[j] = 0;
else
c[j] = mat[k++][i];
c[i] = mod-1;
for ( j = n; j >= 0; j-- )
if ( c[j] )
break;
DEG(w) = j;
r[l++] = w;
}
}
/*
make_qmat(p,mod,tab,mp)
null_mod(mat,mod,n,ind)
null_to_sol(mat,ind,mod,n,r)
*/
void newddd(int mod,UM f,UM *r)
{
register int i,j;
int d,n;
UM q,s,t,u,v,w,g,x,m;
UM *base;
n = DEG(f); base = (UM *)ALLOCA(n*sizeof(UM)); bzero((char *)base,n*sizeof(UM));
w = W_UMALLOC(2*n); q = W_UMALLOC(2*n); m = W_UMALLOC(2*n);
base[0] = W_UMALLOC(0); DEG(base[0]) = 0; COEF(base[0])[0] = 1;
t = W_UMALLOC(1); DEG(t) = 1; COEF(t)[0] = 0; COEF(t)[1] = 1;
pwrmodum(mod,t,mod,f,w); base[1] = W_UMALLOC(DEG(w)); cpyum(w,base[1]);
for ( i = 2; i < n; i++ ) {
/* fprintf(stderr,"i=%d\n",i); */
mulum(mod,base[i-1],base[1],m); DEG(m) = divum(mod,m,f,q);
base[i] = W_UMALLOC(DEG(m)); cpyum(m,base[i]);
}
v = W_UMALLOC(n); cpyum(f,v);
DEG(w) = 1; COEF(w)[0] = 0; COEF(w)[1] = 1;
x = W_UMALLOC(1); DEG(x) = 1; COEF(x)[0] = 0; COEF(x)[1] = 1;
t = W_UMALLOC(n); s = W_UMALLOC(n); u = W_UMALLOC(n); g = W_UMALLOC(n);
for ( j = 0, d = 1; 2*d <= DEG(v); d++ ) {
/* fprintf(stderr,"d=%d\n",d); */
for ( DEG(t) = -1, i = 0; i <= DEG(w); i++ )
if ( COEF(w)[i] ) {
Mulsum(mod,base[i],COEF(w)[i],s);
addum(mod,s,t,u); cpyum(u,t);
}
cpyum(t,w); cpyum(v,s); subum(mod,w,x,t); Gcdum(mod,s,t,g);
if ( DEG(g) >= 1 ) {
berlekamp(g,mod,d,base,r+j); j += DEG(g)/d;
Divum(mod,v,g,q); cpyum(q,v);
DEG(w) = Divum(mod,w,v,q);
for ( i = 0; i < DEG(v); i++ )
DEG(base[i]) = Divum(mod,base[i],v,q);
}
}
if ( DEG(v) ) {
r[j] = UMALLOC(DEG(v)); cpyum(v,r[j]); j++;
}
r[j] = 0;
}
int nfctr_mod(UM f,int mod)
{
register int i,j;
int d,n;
UM q,s,t,u,v,w,g,x,m;
UM *base;
n = DEG(f); base = (UM *)ALLOCA(n*sizeof(UM)); bzero((char *)base,n*sizeof(UM));
w = W_UMALLOC(2*n); q = W_UMALLOC(2*n); m = W_UMALLOC(2*n);
base[0] = W_UMALLOC(0); DEG(base[0]) = 0; COEF(base[0])[0] = 1;
t = W_UMALLOC(1); DEG(t) = 1; COEF(t)[0] = 0; COEF(t)[1] = 1;
pwrmodum(mod,t,mod,f,w); base[1] = W_UMALLOC(DEG(w)); cpyum(w,base[1]);
for ( i = 2; i < n; i++ ) {
/* fprintf(stderr,"i=%d\n",i); */
mulum(mod,base[i-1],base[1],m); DEG(m) = divum(mod,m,f,q);
base[i] = W_UMALLOC(DEG(m)); cpyum(m,base[i]);
}
v = W_UMALLOC(n); cpyum(f,v);
DEG(w) = 1; COEF(w)[0] = 0; COEF(w)[1] = 1;
x = W_UMALLOC(1); DEG(x) = 1; COEF(x)[0] = 0; COEF(x)[1] = 1;
t = W_UMALLOC(n); s = W_UMALLOC(n); u = W_UMALLOC(n); g = W_UMALLOC(n);
for ( j = 0, d = 1; 2*d <= DEG(v); d++ ) {
/* fprintf(stderr,"d=%d\n",d); */
for ( DEG(t) = -1, i = 0; i <= DEG(w); i++ )
if ( COEF(w)[i] ) {
Mulsum(mod,base[i],COEF(w)[i],s);
addum(mod,s,t,u); cpyum(u,t);
}
cpyum(t,w); cpyum(v,s); subum(mod,w,x,t); Gcdum(mod,s,t,g);
if ( DEG(g) >= 1 ) {
j += DEG(g)/d;
Divum(mod,v,g,q); cpyum(q,v);
DEG(w) = Divum(mod,w,v,q);
for ( i = 0; i < DEG(v); i++ )
DEG(base[i]) = Divum(mod,base[i],v,q);
}
}
if ( DEG(v) ) j++;
return j;
}
int irred_check(UM f,int mod)
{
register int i,j;
int d,n;
UM q,s,t,u,v,w,g,x,m,f1,b;
UM *base;
if ( (n = DEG(f)) == 1 )
return 1;
t = W_UMALLOC(n); s = W_UMALLOC(n); g = W_UMALLOC(n);
f1 = W_UMALLOC(n); b = W_UMALLOC(n);
diffum(mod,f,t); cpyum(f,s); Gcdum(mod,t,s,g);
if ( DEG(g) )
return 0;
base = (UM *)ALLOCA(n*sizeof(UM)); bzero((char *)base,n*sizeof(UM));
w = W_UMALLOC(2*n); q = W_UMALLOC(2*n); m = W_UMALLOC(2*n);
base[0] = W_UMALLOC(0); DEG(base[0]) = 0; COEF(base[0])[0] = 1;
t = W_UMALLOC(1); DEG(t) = 1; COEF(t)[0] = 0; COEF(t)[1] = 1;
pwrmodum(mod,t,mod,f,w); base[1] = W_UMALLOC(DEG(w)); cpyum(w,base[1]);
for ( i = 2; i < n; i++ ) {
/* fprintf(stderr,"i=%d\n",i); */
mulum(mod,base[i-1],base[1],m); DEG(m) = divum(mod,m,f,q);
base[i] = W_UMALLOC(DEG(m)); cpyum(m,base[i]);
}
v = W_UMALLOC(n); cpyum(f,v);
DEG(w) = 1; COEF(w)[0] = 0; COEF(w)[1] = 1;
x = W_UMALLOC(1); DEG(x) = 1; COEF(x)[0] = 0; COEF(x)[1] = 1;
t = W_UMALLOC(n); s = W_UMALLOC(n); u = W_UMALLOC(n); g = W_UMALLOC(n);
for ( j = 0, d = 1; 2*d <= n; d++ ) {
/* fprintf(stderr,"d=%d\n",d); */
for ( DEG(t) = -1, i = 0; i <= DEG(w); i++ )
if ( COEF(w)[i] ) {
Mulsum(mod,base[i],COEF(w)[i],s);
addum(mod,s,t,u); cpyum(u,t);
}
cpyum(t,w); cpyum(v,s); subum(mod,w,x,t); Gcdum(mod,s,t,g);
if ( DEG(g) >= 1 )
return 0;
}
return 1;
}
int berlekamp(UM p,int mod,int df,UM *tab,UM *r)
{
int n,i,j,k,nf,d,nr;
int **mat;
int *ind;
UM mp,w,q,gcd,w1,w2;
UM *u;
int *root;
n = DEG(p);
ind = ALLOCA(n*sizeof(int));
make_qmat(p,mod,tab,&mat);
null_mod(mat,mod,n,ind);
for ( i = 0, d = 0; i < n; i++ )
if ( ind[i] )
d++;
if ( d == 1 ) {
r[0] = UMALLOC(n); cpyum(p,r[0]); return 1;
}
u = ALLOCA(d*sizeof(UM *));
r[0] = UMALLOC(n); cpyum(p,r[0]);
null_to_sol(mat,ind,mod,n,u);
root = ALLOCA(d*sizeof(int));
w = W_UMALLOC(n); mp = W_UMALLOC(d);
w1 = W_UMALLOC(n); w2 = W_UMALLOC(n);
for ( i = 1, nf = 1; i < d; i++ ) {
minipoly_mod(mod,u[i],p,mp);
nr = find_root(mod,mp,root);
for ( j = 0; j < nf; j++ ) {
if ( DEG(r[j]) == df )
continue;
for ( k = 0; k < nr; k++ ) {
cpyum(u[i],w1); cpyum(r[j],w2);
COEF(w1)[0] = (mod-root[k]) % mod;
gcdum(mod,w1,w2,w);
if ( DEG(w) > 0 && DEG(w) < DEG(r[j]) ) {
gcd = UMALLOC(DEG(w));
q = UMALLOC(DEG(r[j])-DEG(w));
cpyum(w,gcd); divum(mod,r[j],w,q);
r[j] = q; r[nf++] = gcd;
}
if ( nf == d )
return d;
}
}
}
/* NOTREACHED */
error("berlekamp : cannot happen");
return -1;
}
void minipoly_mod(int mod,UM f,UM p,UM mp)
{
struct p_pair *list,*l,*l1,*lprev;
int n,d;
UM u,p0,p1,np0,np1,q,w;
list = (struct p_pair *)MALLOC(sizeof(struct p_pair));
list->p0 = u = W_UMALLOC(0); DEG(u) = 0; COEF(u)[0] = 1;
list->p1 = W_UMALLOC(0); cpyum(list->p0,list->p1);
list->next = 0;
n = DEG(p); w = UMALLOC(2*n);
p0 = UMALLOC(2*n); cpyum(list->p0,p0);
p1 = UMALLOC(2*n); cpyum(list->p1,p1);
q = W_UMALLOC(2*n);
while ( 1 ) {
COEF(p0)[DEG(p0)] = 0; DEG(p0)++; COEF(p0)[DEG(p0)] = 1;
mulum(mod,f,p1,w); DEG(w) = divum(mod,w,p,q); cpyum(w,p1);
np0 = UMALLOC(n); np1 = UMALLOC(n);
lnf_mod(mod,n,p0,p1,list,np0,np1);
if ( DEG(np1) < 0 ) {
cpyum(np0,mp); return;
} else {
l1 = (struct p_pair *)MALLOC(sizeof(struct p_pair));
l1->p0 = np0; l1->p1 = np1;
for ( l = list, lprev = 0, d = DEG(np1);
l && (DEG(l->p1) > d); lprev = l, l = l->next );
if ( lprev ) {
lprev->next = l1; l1->next = l;
} else {
l1->next = list; list = l1;
}
}
}
}
void lnf_mod(int mod,int n,UM p0,UM p1,struct p_pair *list,UM np0,UM np1)
{
int inv,h,d1;
UM t0,t1,s0,s1;
struct p_pair *l;
cpyum(p0,np0); cpyum(p1,np1);
t0 = W_UMALLOC(n); t1 = W_UMALLOC(n);
s0 = W_UMALLOC(n); s1 = W_UMALLOC(n);
for ( l = list; l; l = l->next ) {
d1 = DEG(np1);
if ( d1 == DEG(l->p1) ) {
inv = invm((mod-COEF(l->p1)[d1])%mod,mod);
h = dmar(COEF(np1)[d1],inv,0,mod);
mulsum(mod,l->p0,h,t0); addum(mod,np0,t0,s0); cpyum(s0,np0);
mulsum(mod,l->p1,h,t1); addum(mod,np1,t1,s1); cpyum(s1,np1);
}
}
}
int find_root(int mod,UM p,int *root)
{
UM *r;
int i,j;
r = ALLOCA((DEG(p)+1)*sizeof(UM));
ddd(mod,p,r);
for ( i = 0, j = 0; r[i]; i++ )
if ( DEG(r[i]) == 1 )
root[j++] = (mod - COEF(r[i])[0]) % mod;
return j;
}
void showum(UM p)
{
int i;
int *c;
for ( i = DEG(p), c = COEF(p); i >= 0; i-- )
if ( c[i] )
printf("+%dx^%d",c[i],i);
printf("\n");
}
void showumat(int **mat,int n)
{
int i,j;
for ( i = 0; i < n; i++ ) {
for ( j = 0; j < n; j++ )
printf("%d ",mat[i][j]);
printf("\n");
}
}