improved for numbers that can be represented exactly

This commit is contained in:
ceriel 1990-12-04 11:33:01 +00:00
parent 0b35c31e5b
commit 8468609eda
3 changed files with 55 additions and 15 deletions

View file

@ -17,3 +17,4 @@ misc.h
ucmp.c
split.c
flt_arith.3
test.c

View file

@ -26,6 +26,10 @@ OBJ = flt_ar2flt.$(SUF) flt_div.$(SUF) flt_flt2ar.$(SUF) flt_modf.$(SUF) \
all: $(LIBFLT)
test: $(LIBFLT) test.c
$(CC) -o test test.c $(LIBFLT)
./test
$(LIBFLT): $(OBJ)
rm -f $(LIBFLT)
$(AR) r $(LIBFLT) $(OBJ)

View file

@ -300,6 +300,7 @@ flt_ecvt(e, ndigit, decpt, sign)
static char buf[NDIGITS+1];
register char *p = buf;
register char *pe;
register int findex = 0;
if (ndigit < 0) ndigit = 0;
if (ndigit > NDIGITS) ndigit = NDIGITS;
@ -316,21 +317,26 @@ flt_ecvt(e, ndigit, decpt, sign)
if (e->m1 != 0) {
register flt_arith *pp = &big_10pow[1];
findex = 1;
while (flt_cmp(e, &big_10pow[BIGSZ-1]) >= 0) {
flt_mul(e,&r_big_10pow[BIGSZ-1],e);
*decpt += (BIGSZ-1)*SMALLSZ;
}
while (flt_cmp(e,pp) >= 0) {
pp++;
findex++;
}
pp--;
flt_mul(e,&r_big_10pow[pp-big_10pow],e);
*decpt += (pp - big_10pow)*SMALLSZ;
findex--;
flt_mul(e,&r_big_10pow[findex],e);
*decpt += findex*SMALLSZ;
pp = &s10pow[1];
while (pp < &s10pow[SMALLSZ] && flt_cmp(e, pp) >= 0) pp++;
pp--;
flt_mul(e, &r_10pow[pp-s10pow], e);
*decpt += pp - s10pow;
findex = 1;
while (pp < &s10pow[SMALLSZ] && flt_cmp(e, pp) >= 0) {
pp++;
findex++;
}
findex--;
*decpt += findex;
if (flt_cmp(e, &s10pow[0]) < 0) {
while (flt_cmp(e, &r_big_10pow[BIGSZ-1]) < 0) {
@ -338,21 +344,50 @@ flt_ecvt(e, ndigit, decpt, sign)
*decpt -= (BIGSZ-1)*SMALLSZ;
}
pp = &r_big_10pow[1];
while(flt_cmp(e,pp) < 0) pp++;
pp--;
flt_mul(e,&big_10pow[pp-r_big_10pow],e);
*decpt -= (pp - r_big_10pow)*SMALLSZ;
findex = 1;
while(flt_cmp(e,pp) < 0) {
pp++;
findex++;
}
findex--;
flt_mul(e,&big_10pow[findex],e);
*decpt -= findex*SMALLSZ;
/* here, value >= 10 ** -28 */
flt_mul(e, &s10pow[1], e);
(*decpt)--;
pp = &r_10pow[0];
while(flt_cmp(e, pp) < 0) pp++;
flt_mul(e, &s10pow[pp-r_10pow], e);
*decpt -= pp - r_10pow;
findex = 0;
while(flt_cmp(e, pp) < 0) {
pp++;
findex++;
}
flt_mul(e, &s10pow[findex], e);
*decpt -= findex;
findex = 0;
}
(*decpt)++; /* because now value in [1.0, 10.0) */
}
while (p <= pe) {
if (findex) {
flt_arith tc, oldtc;
int count = 0;
oldtc.flt_exp = 0;
oldtc.flt_sign = 0;
oldtc.m1 = 0;
oldtc.m2 = 0;
tc = s10pow[findex];
while (flt_cmp(e, &tc) >= 0) {
oldtc = tc;
flt_add(&tc, &s10pow[findex], &tc);
count++;
}
*p++ = count + '0';
oldtc.flt_sign = 1;
flt_add(e, &oldtc, e);
findex--;
continue;
}
if (e->flt_exp >= 0 && e->m1 != 0) {
flt_arith x;
@ -367,7 +402,7 @@ flt_ecvt(e, ndigit, decpt, sign)
}
}
else *p++ = '0';
flt_mul(e, &s10pow[1], e);
if (e->m1) flt_mul(e, &s10pow[1], e);
}
if (pe >= buf) {
p = pe;