static VALUE
bug_str_cstr_term(VALUE str)
{
long len;
char *s;
int c;
rb_encoding *enc;
len = RSTRING_LEN(str);
s = StringValueCStr(str);
rb_gc();
enc = rb_enc_get(str);
c = rb_enc_codepoint(&s[len], &s[len+rb_enc_mbminlen(enc)], enc);
return INT2NUM(c);
}
VALUE
rb_str_format(int argc, const VALUE *argv, VALUE fmt)
{
rb_encoding *enc;
const char *p, *end;
char *buf;
int blen, bsiz;
VALUE result;
int width, prec, flags = FNONE;
int nextarg = 1;
int posarg = 0;
int tainted = 0;
VALUE nextvalue;
VALUE tmp;
VALUE str;
volatile VALUE hash = Qundef;
#define CHECK_FOR_WIDTH(f) \
if ((f) & FWIDTH) { \
rb_raise(rb_eArgError, "width given twice"); \
} \
if ((f) & FPREC0) { \
rb_raise(rb_eArgError, "width after precision"); \
}
#define CHECK_FOR_FLAGS(f) \
if ((f) & FWIDTH) { \
rb_raise(rb_eArgError, "flag after width"); \
} \
if ((f) & FPREC0) { \
rb_raise(rb_eArgError, "flag after precision"); \
}
++argc;
--argv;
if (OBJ_TAINTED(fmt)) tainted = 1;
StringValue(fmt);
enc = rb_enc_get(fmt);
fmt = rb_str_new4(fmt);
p = RSTRING_PTR(fmt);
end = p + RSTRING_LEN(fmt);
blen = 0;
bsiz = 120;
result = rb_str_buf_new(bsiz);
rb_enc_copy(result, fmt);
buf = RSTRING_PTR(result);
memset(buf, 0, bsiz);
for (; p < end; p++) {
const char *t;
int n;
for (t = p; t < end && *t != '%'; t++) ;
PUSH(p, t - p);
if (t >= end) {
/* end of fmt string */
goto sprint_exit;
}
p = t + 1; /* skip `%' */
width = prec = -1;
nextvalue = Qundef;
retry:
switch (*p) {
default:
if (rb_enc_isprint(*p, enc))
rb_raise(rb_eArgError, "malformed format string - %%%c", *p);
else
rb_raise(rb_eArgError, "malformed format string");
break;
case ' ':
CHECK_FOR_FLAGS(flags);
flags |= FSPACE;
p++;
goto retry;
case '#':
CHECK_FOR_FLAGS(flags);
flags |= FSHARP;
p++;
goto retry;
case '+':
CHECK_FOR_FLAGS(flags);
flags |= FPLUS;
p++;
goto retry;
case '-':
CHECK_FOR_FLAGS(flags);
flags |= FMINUS;
p++;
goto retry;
case '0':
CHECK_FOR_FLAGS(flags);
flags |= FZERO;
p++;
goto retry;
case '1': case '2': case '3': case '4':
case '5': case '6': case '7': case '8': case '9':
n = 0;
GETNUM(n, width);
if (*p == '$') {
if (nextvalue != Qundef) {
rb_raise(rb_eArgError, "value given twice - %d$", n);
}
nextvalue = GETPOSARG(n);
p++;
goto retry;
}
CHECK_FOR_WIDTH(flags);
width = n;
flags |= FWIDTH;
goto retry;
case '<':
case '{':
{
const char *start = p;
char term = (*p == '<') ? '>' : '}';
ID id;
for (; p < end && *p != term; ) {
p += rb_enc_mbclen(p, end, enc);
}
if (p >= end) {
rb_raise(rb_eArgError, "malformed name - unmatched parenthesis");
}
id = rb_intern3(start + 1, p - start - 1, enc);
nextvalue = GETNAMEARG(ID2SYM(id));
if (term == '}') goto format_s;
p++;
goto retry;
}
case '*':
CHECK_FOR_WIDTH(flags);
flags |= FWIDTH;
GETASTER(width);
if (width < 0) {
flags |= FMINUS;
width = -width;
}
p++;
goto retry;
case '.':
if (flags & FPREC0) {
rb_raise(rb_eArgError, "precision given twice");
}
flags |= FPREC|FPREC0;
prec = 0;
p++;
if (*p == '*') {
GETASTER(prec);
if (prec < 0) { /* ignore negative precision */
flags &= ~FPREC;
}
p++;
goto retry;
}
GETNUM(prec, precision);
goto retry;
case '\n':
case '\0':
p--;
case '%':
if (flags != FNONE) {
rb_raise(rb_eArgError, "invalid format character - %%");
}
PUSH("%", 1);
break;
case 'c':
{
VALUE val = GETARG();
VALUE tmp;
unsigned int c;
int n;
tmp = rb_check_string_type(val);
if (!NIL_P(tmp)) {
if (rb_enc_strlen(RSTRING_PTR(tmp),RSTRING_END(tmp),enc) != 1) {
rb_raise(rb_eArgError, "%%c requires a character");
}
c = rb_enc_codepoint(RSTRING_PTR(tmp), RSTRING_END(tmp), enc);
}
else {
c = NUM2INT(val);
}
n = rb_enc_codelen(c, enc);
if (n <= 0) {
rb_raise(rb_eArgError, "invalid character");
}
if (!(flags & FWIDTH)) {
CHECK(n);
rb_enc_mbcput(c, &buf[blen], enc);
blen += n;
}
else if ((flags & FMINUS)) {
CHECK(n);
rb_enc_mbcput(c, &buf[blen], enc);
blen += n;
FILL(' ', width-1);
}
else {
FILL(' ', width-1);
CHECK(n);
rb_enc_mbcput(c, &buf[blen], enc);
blen += n;
}
}
break;
case 's':
case 'p':
format_s:
{
VALUE arg = GETARG();
long len, slen;
if (*p == 'p') arg = rb_inspect(arg);
str = rb_obj_as_string(arg);
if (OBJ_TAINTED(str)) tainted = 1;
len = RSTRING_LEN(str);
rb_str_set_len(result, blen);
enc = rb_enc_check(result, str);
if (flags&(FPREC|FWIDTH)) {
slen = rb_enc_strlen(RSTRING_PTR(str),RSTRING_END(str),enc);
if (slen < 0) {
rb_raise(rb_eArgError, "invalid mbstring sequence");
}
if ((flags&FPREC) && (prec < slen)) {
char *p = rb_enc_nth(RSTRING_PTR(str), RSTRING_END(str),
prec, enc);
slen = prec;
len = p - RSTRING_PTR(str);
}
/* need to adjust multi-byte string pos */
if ((flags&FWIDTH) && (width > slen)) {
width -= slen;
if (!(flags&FMINUS)) {
CHECK(width);
while (width--) {
buf[blen++] = ' ';
}
}
CHECK(len);
memcpy(&buf[blen], RSTRING_PTR(str), len);
blen += len;
if (flags&FMINUS) {
CHECK(width);
while (width--) {
buf[blen++] = ' ';
}
}
rb_enc_associate(result, enc);
break;
}
}
PUSH(RSTRING_PTR(str), len);
rb_enc_associate(result, enc);
}
break;
case 'd':
case 'i':
case 'o':
case 'x':
case 'X':
case 'b':
case 'B':
case 'u':
{
volatile VALUE tmp1;
volatile VALUE val = GETARG();
char fbuf[32], nbuf[64], *s;
const char *prefix = 0;
int sign = 0, dots = 0;
char sc = 0;
long v = 0;
int base, bignum = 0;
int len, pos;
switch (*p) {
case 'd':
case 'i':
case 'u':
sign = 1; break;
case 'o':
case 'x':
case 'X':
case 'b':
case 'B':
if (flags&(FPLUS|FSPACE)) sign = 1;
break;
}
if (flags & FSHARP) {
switch (*p) {
case 'o':
prefix = "0"; break;
case 'x':
prefix = "0x"; break;
case 'X':
prefix = "0X"; break;
case 'b':
prefix = "0b"; break;
case 'B':
prefix = "0B"; break;
}
}
bin_retry:
switch (TYPE(val)) {
case T_FLOAT:
if (FIXABLE(RFLOAT_VALUE(val))) {
val = LONG2FIX((long)RFLOAT_VALUE(val));
goto bin_retry;
}
val = rb_dbl2big(RFLOAT_VALUE(val));
if (FIXNUM_P(val)) goto bin_retry;
bignum = 1;
break;
case T_STRING:
val = rb_str_to_inum(val, 0, Qtrue);
goto bin_retry;
case T_BIGNUM:
bignum = 1;
break;
case T_FIXNUM:
v = FIX2LONG(val);
break;
default:
val = rb_Integer(val);
goto bin_retry;
}
switch (*p) {
case 'o':
base = 8; break;
case 'x':
case 'X':
base = 16; break;
case 'b':
case 'B':
base = 2; break;
case 'u':
case 'd':
case 'i':
default:
base = 10; break;
}
if (!bignum) {
if (base == 2) {
val = rb_int2big(v);
goto bin_retry;
}
if (sign) {
char c = *p;
if (c == 'i') c = 'd'; /* %d and %i are identical */
if (v < 0) {
v = -v;
sc = '-';
width--;
}
else if (flags & FPLUS) {
sc = '+';
width--;
}
else if (flags & FSPACE) {
sc = ' ';
width--;
}
snprintf(fbuf, sizeof(fbuf), "%%l%c", c);
snprintf(nbuf, sizeof(nbuf), fbuf, v);
s = nbuf;
}
else {
s = nbuf;
if (v < 0) {
dots = 1;
}
snprintf(fbuf, sizeof(fbuf), "%%l%c", *p == 'X' ? 'x' : *p);
snprintf(++s, sizeof(nbuf) - 1, fbuf, v);
if (v < 0) {
char d = 0;
s = remove_sign_bits(s, base);
switch (base) {
case 16:
d = 'f'; break;
case 8:
d = '7'; break;
}
if (d && *s != d) {
*--s = d;
}
}
}
}
else {
if (sign) {
tmp = rb_big2str(val, base);
s = RSTRING_PTR(tmp);
if (s[0] == '-') {
s++;
sc = '-';
width--;
}
else if (flags & FPLUS) {
sc = '+';
width--;
}
else if (flags & FSPACE) {
sc = ' ';
width--;
}
}
else {
if (!RBIGNUM_SIGN(val)) {
val = rb_big_clone(val);
rb_big_2comp(val);
}
tmp1 = tmp = rb_big2str0(val, base, RBIGNUM_SIGN(val));
s = RSTRING_PTR(tmp);
if (*s == '-') {
dots = 1;
if (base == 10) {
rb_warning("negative number for %%u specifier");
}
s = remove_sign_bits(++s, base);
switch (base) {
case 16:
if (s[0] != 'f') *--s = 'f'; break;
case 8:
if (s[0] != '7') *--s = '7'; break;
case 2:
if (s[0] != '1') *--s = '1'; break;
}
}
}
}
pos = -1;
len = strlen(s);
if (dots) {
prec -= 2;
width -= 2;
}
if (*p == 'X') {
char *pp = s;
int c;
while ((c = (int)(unsigned char)*pp) != 0) {
*pp = rb_enc_toupper(c, enc);
pp++;
}
}
if (prefix && !prefix[1]) { /* octal */
if (dots) {
prefix = 0;
}
else if (len == 1 && *s == '0') {
len = 0;
if (flags & FPREC) prec--;
}
else if ((flags & FPREC) && (prec > len)) {
prefix = 0;
}
}
else if (len == 1 && *s == '0') {
prefix = 0;
}
if (prefix) {
width -= strlen(prefix);
}
if ((flags & (FZERO|FMINUS|FPREC)) == FZERO) {
prec = width;
width = 0;
}
else {
if (prec < len) {
if (!prefix && prec == 0 && len == 1 && *s == '0') len = 0;
prec = len;
}
width -= prec;
}
if (!(flags&FMINUS)) {
CHECK(width);
while (width-- > 0) {
buf[blen++] = ' ';
}
}
if (sc) PUSH(&sc, 1);
if (prefix) {
int plen = strlen(prefix);
PUSH(prefix, plen);
}
CHECK(prec - len);
if (dots) PUSH("..", 2);
if (!bignum && v < 0) {
char c = sign_bits(base, p);
while (len < prec--) {
buf[blen++] = c;
}
}
else if ((flags & (FMINUS|FPREC)) != FMINUS) {
char c;
if (!sign && bignum && !RBIGNUM_SIGN(val))
c = sign_bits(base, p);
else
c = '0';
while (len < prec--) {
buf[blen++] = c;
}
}
PUSH(s, len);
CHECK(width);
while (width-- > 0) {
buf[blen++] = ' ';
}
}
break;
case 'f':
case 'g':
case 'G':
case 'e':
case 'E':
{
VALUE val = GETARG();
double fval;
int i, need = 6;
char fbuf[32];
fval = RFLOAT_VALUE(rb_Float(val));
if (isnan(fval) || isinf(fval)) {
const char *expr;
if (isnan(fval)) {
expr = "NaN";
}
else {
expr = "Inf";
}
need = strlen(expr);
if ((!isnan(fval) && fval < 0.0) || (flags & FPLUS))
need++;
if ((flags & FWIDTH) && need < width)
need = width;
CHECK(need + 1);
snprintf(&buf[blen], need + 1, "%*s", need, "");
if (flags & FMINUS) {
if (!isnan(fval) && fval < 0.0)
buf[blen++] = '-';
else if (flags & FPLUS)
buf[blen++] = '+';
else if (flags & FSPACE)
blen++;
strncpy(&buf[blen], expr, strlen(expr));
}
else {
if (!isnan(fval) && fval < 0.0)
buf[blen + need - strlen(expr) - 1] = '-';
else if (flags & FPLUS)
buf[blen + need - strlen(expr) - 1] = '+';
else if ((flags & FSPACE) && need > width)
blen++;
strncpy(&buf[blen + need - strlen(expr)], expr,
strlen(expr));
}
blen += strlen(&buf[blen]);
break;
}
fmt_setup(fbuf, sizeof(fbuf), *p, flags, width, prec);
need = 0;
if (*p != 'e' && *p != 'E') {
i = INT_MIN;
frexp(fval, &i);
if (i > 0)
need = BIT_DIGITS(i);
}
need += (flags&FPREC) ? prec : 6;
if ((flags&FWIDTH) && need < width)
need = width;
need += 20;
CHECK(need);
snprintf(&buf[blen], need, fbuf, fval);
blen += strlen(&buf[blen]);
}
break;
}
flags = FNONE;
}
sprint_exit:
/* XXX - We cannot validate the number of arguments if (digit)$ style used.
*/
if (posarg >= 0 && nextarg < argc) {
const char *mesg = "too many arguments for format string";
if (RTEST(ruby_debug)) rb_raise(rb_eArgError, "%s", mesg);
if (RTEST(ruby_verbose)) rb_warn("%s", mesg);
}
rb_str_resize(result, blen);
if (tainted) OBJ_TAINT(result);
return result;
}