diff options
Diffstat (limited to 'sprintf.c')
| -rw-r--r-- | sprintf.c | 1367 |
1 files changed, 684 insertions, 683 deletions
@@ -44,13 +44,13 @@ sign_bits(int base, const char *p) switch (base) { case 16: - if (*p == 'X') c = 'F'; - else c = 'f'; - break; + if (*p == 'X') c = 'F'; + else c = 'f'; + break; case 8: - c = '7'; break; + c = '7'; break; case 2: - c = '1'; break; + c = '1'; break; } return c; } @@ -68,8 +68,8 @@ sign_bits(int base, const char *p) #define CHECK(l) do {\ int cr = ENC_CODERANGE(result);\ while ((l) >= bsiz - blen) {\ - bsiz*=2;\ - if (bsiz<0) rb_raise(rb_eArgError, "too big specifier");\ + bsiz*=2;\ + if (bsiz<0) rb_raise(rb_eArgError, "too big specifier");\ }\ rb_str_resize(result, bsiz);\ ENC_CODERANGE_SET(result, cr);\ @@ -97,8 +97,8 @@ sign_bits(int base, const char *p) blen += (l);\ } while (0) -#define GETARG() (nextvalue != Qundef ? nextvalue : \ - GETNEXTARG()) +#define GETARG() (!UNDEF_P(nextvalue) ? nextvalue : \ + GETNEXTARG()) #define GETNEXTARG() ( \ check_next_arg(posarg, nextarg), \ @@ -124,11 +124,11 @@ sign_bits(int base, const char *p) n = 0; \ GETNUM(n, val); \ if (*p == '$') { \ - tmp = GETPOSARG(n); \ + tmp = GETPOSARG(n); \ } \ else { \ - tmp = GETNEXTARG(); \ - p = t; \ + tmp = GETNEXTARG(); \ + p = t; \ } \ (val) = NUM2INT(tmp); \ } while (0) @@ -138,15 +138,15 @@ get_num(const char *p, const char *end, rb_encoding *enc, int *valp) { int next_n = *valp; for (; p < end && rb_enc_isdigit(*p, enc); p++) { - if (MUL_OVERFLOW_INT_P(10, next_n)) - return NULL; - next_n *= 10; - if (INT_MAX - (*p - '0') < next_n) - return NULL; - next_n += *p - '0'; + if (MUL_OVERFLOW_INT_P(10, next_n)) + return NULL; + next_n *= 10; + if (INT_MAX - (*p - '0') < next_n) + return NULL; + next_n += *p - '0'; } if (p >= end) { - rb_raise(rb_eArgError, "malformed format string - %%*[0-9]"); + rb_raise(rb_eArgError, "malformed format string - %%*[0-9]"); } *valp = next_n; return p; @@ -157,9 +157,9 @@ check_next_arg(int posarg, int nextarg) { switch (posarg) { case -1: - rb_raise(rb_eArgError, "unnumbered(%d) mixed with numbered", nextarg); + rb_raise(rb_eArgError, "unnumbered(%d) mixed with numbered", nextarg); case -2: - rb_raise(rb_eArgError, "unnumbered(%d) mixed with named", nextarg); + rb_raise(rb_eArgError, "unnumbered(%d) mixed with named", nextarg); } } @@ -167,13 +167,13 @@ static void check_pos_arg(int posarg, int n) { if (posarg > 0) { - rb_raise(rb_eArgError, "numbered(%d) after unnumbered(%d)", n, posarg); + rb_raise(rb_eArgError, "numbered(%d) after unnumbered(%d)", n, posarg); } if (posarg == -2) { - rb_raise(rb_eArgError, "numbered(%d) after named", n); + rb_raise(rb_eArgError, "numbered(%d) after named", n); } if (n < 1) { - rb_raise(rb_eArgError, "invalid index - %d$", n); + rb_raise(rb_eArgError, "invalid index - %d$", n); } } @@ -181,10 +181,10 @@ static void check_name_arg(int posarg, const char *name, int len, rb_encoding *enc) { if (posarg > 0) { - rb_enc_raise(enc, rb_eArgError, "named%.*s after unnumbered(%d)", len, name, posarg); + rb_enc_raise(enc, rb_eArgError, "named%.*s after unnumbered(%d)", len, name, posarg); } if (posarg == -1) { - rb_enc_raise(enc, rb_eArgError, "named%.*s after numbered", len, name); + rb_enc_raise(enc, rb_eArgError, "named%.*s after numbered", len, name); } } @@ -193,13 +193,13 @@ get_hash(volatile VALUE *hash, int argc, const VALUE *argv) { VALUE tmp; - if (*hash != Qundef) return *hash; + if (!UNDEF_P(*hash)) return *hash; if (argc != 2) { - rb_raise(rb_eArgError, "one hash required"); + rb_raise(rb_eArgError, "one hash required"); } tmp = rb_check_hash_type(argv[1]); if (NIL_P(tmp)) { - rb_raise(rb_eArgError, "one hash required"); + rb_raise(rb_eArgError, "one hash required"); } return (*hash = tmp); } @@ -221,7 +221,7 @@ rb_str_format(int argc, const VALUE *argv, VALUE fmt) VALUE result; long scanned = 0; - int coderange = ENC_CODERANGE_7BIT; + enum ruby_coderange_type coderange = ENC_CODERANGE_7BIT; int width, prec, flags = FNONE; int nextarg = 1; int posarg = 0; @@ -233,23 +233,34 @@ rb_str_format(int argc, const VALUE *argv, VALUE fmt) #define CHECK_FOR_WIDTH(f) \ if ((f) & FWIDTH) { \ - rb_raise(rb_eArgError, "width given twice"); \ + rb_raise(rb_eArgError, "width given twice"); \ } \ if ((f) & FPREC0) { \ - rb_raise(rb_eArgError, "width after precision"); \ + rb_raise(rb_eArgError, "width after precision"); \ } #define CHECK_FOR_FLAGS(f) \ if ((f) & FWIDTH) { \ - rb_raise(rb_eArgError, "flag after width"); \ + rb_raise(rb_eArgError, "flag after width"); \ } \ if ((f) & FPREC0) { \ - rb_raise(rb_eArgError, "flag after precision"); \ + rb_raise(rb_eArgError, "flag after precision"); \ } +#define update_coderange(partial) do { \ + if (coderange != ENC_CODERANGE_BROKEN && scanned < blen \ + && rb_enc_to_index(enc) /* != ENCINDEX_ASCII_8BIT */) { \ + int cr = coderange; \ + scanned += rb_str_coderange_scan_restartable(buf+scanned, buf+blen, enc, &cr); \ + ENC_CODERANGE_SET(result, \ + (partial && cr == ENC_CODERANGE_UNKNOWN ? \ + ENC_CODERANGE_BROKEN : (coderange = cr))); \ + } \ + } while (0) ++argc; --argv; StringValue(fmt); enc = rb_enc_get(fmt); + rb_must_asciicompat(fmt); orig = fmt; fmt = rb_str_tmp_frozen_acquire(fmt); p = RSTRING_PTR(fmt); @@ -263,362 +274,360 @@ rb_str_format(int argc, const VALUE *argv, VALUE fmt) ENC_CODERANGE_SET(result, coderange); for (; p < end; p++) { - const char *t; - int n; - VALUE sym = Qnil; - - for (t = p; t < end && *t != '%'; t++) ; - if (t + 1 == end) { - rb_raise(rb_eArgError, "incomplete format specifier; use %%%% (double %%) instead"); - } - PUSH(p, t - p); - if (coderange != ENC_CODERANGE_BROKEN && scanned < blen) { - scanned += rb_str_coderange_scan_restartable(buf+scanned, buf+blen, enc, &coderange); - ENC_CODERANGE_SET(result, coderange); - } - if (t >= end) { - /* end of fmt string */ - goto sprint_exit; - } - p = t + 1; /* skip `%' */ - - width = prec = -1; - nextvalue = Qundef; + const char *t; + int n; + VALUE sym = Qnil; + + for (t = p; t < end && *t != '%'; t++) ; + if (t + 1 == end) { + rb_raise(rb_eArgError, "incomplete format specifier; use %%%% (double %%) instead"); + } + PUSH(p, t - p); + update_coderange(FALSE); + 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$", n); - } - nextvalue = GETPOSARG(n); - p++; - goto retry; - } - CHECK_FOR_WIDTH(flags); - width = n; - flags |= FWIDTH; - goto retry; - - case '<': - case '{': - { - const char *start = p; - char term = (*p == '<') ? '>' : '}'; - int len; - - for (; p < end && *p != term; ) { - p += rb_enc_mbclen(p, end, enc); - } - if (p >= end) { - rb_raise(rb_eArgError, "malformed name - unmatched parenthesis"); - } + switch (*p) { + default: + if (rb_enc_isprint(*p, enc)) + rb_raise(rb_eArgError, "malformed format string - %%%c", *p); + else + rb_raise(rb_eArgError, "malformed format string"); + break; + + case ' ': + CHECK_FOR_FLAGS(flags); + flags |= FSPACE; + p++; + goto retry; + + case '#': + CHECK_FOR_FLAGS(flags); + flags |= FSHARP; + p++; + goto retry; + + case '+': + CHECK_FOR_FLAGS(flags); + flags |= FPLUS; + p++; + goto retry; + + case '-': + CHECK_FOR_FLAGS(flags); + flags |= FMINUS; + p++; + goto retry; + + case '0': + CHECK_FOR_FLAGS(flags); + flags |= FZERO; + p++; + goto retry; + + case '1': case '2': case '3': case '4': + case '5': case '6': case '7': case '8': case '9': + n = 0; + GETNUM(n, width); + if (*p == '$') { + if (!UNDEF_P(nextvalue)) { + 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 == '<') ? '>' : '}'; + int len; + + for (; p < end && *p != term; ) { + p += rb_enc_mbclen(p, end, enc); + } + if (p >= end) { + rb_raise(rb_eArgError, "malformed name - unmatched parenthesis"); + } #if SIZEOF_INT < SIZEOF_SIZE_T - if ((size_t)(p - start) >= INT_MAX) { - const int message_limit = 20; - len = (int)(rb_enc_right_char_head(start, start + message_limit, p, enc) - start); - rb_enc_raise(enc, rb_eArgError, - "too long name (%"PRIuSIZE" bytes) - %.*s...%c", - (size_t)(p - start - 2), len, start, term); - } + if ((size_t)(p - start) >= INT_MAX) { + const int message_limit = 20; + len = (int)(rb_enc_right_char_head(start, start + message_limit, p, enc) - start); + rb_enc_raise(enc, rb_eArgError, + "too long name (%"PRIuSIZE" bytes) - %.*s...%c", + (size_t)(p - start - 2), len, start, term); + } #endif - len = (int)(p - start + 1); /* including parenthesis */ - if (sym != Qnil) { - rb_enc_raise(enc, rb_eArgError, "named%.*s after <%"PRIsVALUE">", - len, start, rb_sym2str(sym)); - } - CHECKNAMEARG(start, len, enc); - get_hash(&hash, argc, argv); - sym = rb_check_symbol_cstr(start + 1, - len - 2 /* without parenthesis */, - enc); - if (!NIL_P(sym)) nextvalue = rb_hash_lookup2(hash, sym, Qundef); - if (nextvalue == Qundef) { - if (NIL_P(sym)) { - sym = rb_sym_intern(start + 1, - len - 2 /* without parenthesis */, - enc); - } - nextvalue = rb_hash_default_value(hash, sym); - if (NIL_P(nextvalue)) { - rb_key_err_raise(rb_enc_sprintf(enc, "key%.*s not found", len, start), hash, sym); - } - } - if (term == '}') goto format_s; - p++; - goto retry; - } - - case '*': - CHECK_FOR_WIDTH(flags); - flags |= FWIDTH; - GETASTER(width); - if (width < 0) { - flags |= FMINUS; - width = -width; - if (width < 0) rb_raise(rb_eArgError, "width too big"); - } - 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--; + len = (int)(p - start + 1); /* including parenthesis */ + if (sym != Qnil) { + rb_enc_raise(enc, rb_eArgError, "named%.*s after <%"PRIsVALUE">", + len, start, rb_sym2str(sym)); + } + CHECKNAMEARG(start, len, enc); + get_hash(&hash, argc, argv); + sym = rb_check_symbol_cstr(start + 1, + len - 2 /* without parenthesis */, + enc); + if (!NIL_P(sym)) nextvalue = rb_hash_lookup2(hash, sym, Qundef); + if (UNDEF_P(nextvalue)) { + if (NIL_P(sym)) { + sym = rb_sym_intern(start + 1, + len - 2 /* without parenthesis */, + enc); + } + nextvalue = rb_hash_default_value(hash, sym); + if (NIL_P(nextvalue)) { + rb_key_err_raise(rb_enc_sprintf(enc, "key%.*s not found", len, start), hash, sym); + } + } + if (term == '}') goto format_s; + p++; + goto retry; + } + + case '*': + CHECK_FOR_WIDTH(flags); + flags |= FWIDTH; + GETASTER(width); + if (width < 0) { + flags |= FMINUS; + width = -width; + if (width < 0) rb_raise(rb_eArgError, "width too big"); + } + 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--; /* fall through */ - 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_len(RSTRING_PTR(tmp), RSTRING_END(tmp), &n, enc); - RB_GC_GUARD(tmp); - } - 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; - if (width > 1) FILL(' ', width-1); - } - else { - if (width > 1) 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') { - str = rb_inspect(arg); - } - else { - str = rb_obj_as_string(arg); - } - len = RSTRING_LEN(str); - rb_str_set_len(result, blen); - if (coderange != ENC_CODERANGE_BROKEN && scanned < blen) { - int cr = coderange; - scanned += rb_str_coderange_scan_restartable(buf+scanned, buf+blen, enc, &cr); - ENC_CODERANGE_SET(result, - (cr == ENC_CODERANGE_UNKNOWN ? - ENC_CODERANGE_BROKEN : (coderange = cr))); - } - 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 -= (int)slen; - if (!(flags&FMINUS)) { - FILL(' ', width); - width = 0; - } - CHECK(len); - memcpy(&buf[blen], RSTRING_PTR(str), len); - RB_GC_GUARD(str); - blen += len; - if (flags&FMINUS) { - FILL(' ', width); - } - rb_enc_associate(result, enc); - break; - } - } - PUSH(RSTRING_PTR(str), len); - RB_GC_GUARD(str); - rb_enc_associate(result, enc); - } - break; - - case 'd': - case 'i': - case 'o': - case 'x': - case 'X': - case 'b': - case 'B': - case 'u': - { - volatile VALUE val = GETARG(); + 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)) { + flags |= FPREC; + prec = 1; + str = tmp; + goto format_s1; + } + n = NUM2INT(val); + if (n >= 0) n = rb_enc_codelen((c = n), enc); + if (n <= 0) { + rb_raise(rb_eArgError, "invalid character"); + } + int encidx = rb_ascii8bit_appendable_encoding_index(enc, c); + if (encidx >= 0 && encidx != rb_enc_to_index(enc)) { + /* special case */ + rb_enc_associate_index(result, encidx); + enc = rb_enc_from_index(encidx); + coderange = ENC_CODERANGE_VALID; + } + if (!(flags & FWIDTH)) { + CHECK(n); + rb_enc_mbcput(c, &buf[blen], enc); + blen += n; + } + else if ((flags & FMINUS)) { + --width; + CHECK(n + (width > 0 ? width : 0)); + rb_enc_mbcput(c, &buf[blen], enc); + blen += n; + if (width > 0) FILL_(' ', width); + } + else { + --width; + CHECK(n + (width > 0 ? width : 0)); + if (width > 0) FILL_(' ', width); + 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') { + str = rb_inspect(arg); + } + else { + str = rb_obj_as_string(arg); + } + format_s1: + len = RSTRING_LEN(str); + rb_str_set_len(result, blen); + update_coderange(TRUE); + 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 -= (int)slen; + CHECK(len + width); + if (!(flags&FMINUS)) { + FILL_(' ', width); + width = 0; + } + memcpy(&buf[blen], RSTRING_PTR(str), len); + RB_GC_GUARD(str); + blen += len; + if (flags&FMINUS) { + FILL_(' ', width); + } + rb_enc_associate(result, enc); + break; + } + } + PUSH(RSTRING_PTR(str), len); + RB_GC_GUARD(str); + rb_enc_associate(result, enc); + } + break; + + case 'd': + case 'i': + case 'o': + case 'x': + case 'X': + case 'b': + case 'B': + case 'u': + { + volatile VALUE val = GETARG(); int valsign; - char nbuf[BIT_DIGITS(SIZEOF_LONG*CHAR_BIT)+2], *s; - const char *prefix = 0; - int sign = 0, dots = 0; - char sc = 0; - long v = 0; - int base, bignum = 0; - int len; - - 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, TRUE); - 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; - } + char nbuf[BIT_DIGITS(SIZEOF_LONG*CHAR_BIT)+2], *s; + const char *prefix = 0; + int sign = 0, dots = 0; + char sc = 0; + long v = 0; + int base, bignum = 0; + int len; + + 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, TRUE); + 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 (base != 10) { int numbits = ffs(base)-1; @@ -690,10 +699,10 @@ rb_str_format(int argc, const VALUE *argv, VALUE fmt) sc = ' '; width--; } - s = ruby_ultoa((unsigned long)v, nbuf + sizeof(nbuf), 10, 0); - len = (int)(nbuf + sizeof(nbuf) - s); - } - else { + s = ruby_ultoa((unsigned long)v, nbuf + sizeof(nbuf), 10, 0); + len = (int)(nbuf + sizeof(nbuf) - s); + } + else { tmp = rb_big2str(val, 10); s = RSTRING_PTR(tmp); valsign = 1; @@ -711,230 +720,231 @@ rb_str_format(int argc, const VALUE *argv, VALUE fmt) sc = ' '; width--; } - len = rb_long2int(RSTRING_END(tmp) - 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 -= (int)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)) { - FILL(' ', width); - width = 0; - } - if (sc) PUSH(&sc, 1); - if (prefix) { - int plen = (int)strlen(prefix); - PUSH(prefix, plen); - } - if (dots) PUSH("..", 2); - if (prec > len) { - CHECK(prec - len); - if (!sign && valsign < 0) { - char c = sign_bits(base, p); - FILL_(c, prec - len); - } - else if ((flags & (FMINUS|FPREC)) != FMINUS) { - FILL_('0', prec - len); - } - } - PUSH(s, len); - RB_GC_GUARD(tmp); - FILL(' ', width); - } - break; - - case 'f': - { - VALUE val = GETARG(), num, den; - int sign = (flags&FPLUS) ? 1 : 0, zero = 0; - long len, fill; - if (RB_INTEGER_TYPE_P(val)) { - den = INT2FIX(1); - num = val; - } - else if (RB_TYPE_P(val, T_RATIONAL)) { - den = rb_rational_den(val); - num = rb_rational_num(val); - } - else { - nextvalue = val; - goto float_value; - } - if (!(flags&FPREC)) prec = default_float_precision; - if (FIXNUM_P(num)) { - if ((SIGNED_VALUE)num < 0) { - long n = -FIX2LONG(num); - num = LONG2FIX(n); - sign = -1; - } - } - else if (BIGNUM_NEGATIVE_P(num)) { - sign = -1; - num = rb_big_uminus(num); - } - if (den != INT2FIX(1)) { - num = rb_int_mul(num, rb_int_positive_pow(10, prec)); - num = rb_int_plus(num, rb_int_idiv(den, INT2FIX(2))); - num = rb_int_idiv(num, den); - } - else if (prec >= 0) { - zero = prec; - } - val = rb_int2str(num, 10); - len = RSTRING_LEN(val) + zero; - if (prec >= len) len = prec + 1; /* integer part 0 */ - if (sign || (flags&FSPACE)) ++len; - if (prec > 0) ++len; /* period */ - fill = width > len ? width - len : 0; - CHECK(fill + len); - if (fill && !(flags&(FMINUS|FZERO))) { - FILL_(' ', fill); - } - if (sign || (flags&FSPACE)) { - buf[blen++] = sign > 0 ? '+' : sign < 0 ? '-' : ' '; - } - if (fill && (flags&(FMINUS|FZERO)) == FZERO) { - FILL_('0', fill); - } - len = RSTRING_LEN(val) + zero; - t = RSTRING_PTR(val); - if (len > prec) { - PUSH_(t, len - prec); - } - else { - buf[blen++] = '0'; - } - if (prec > 0) { - buf[blen++] = '.'; - } - if (zero) { - FILL_('0', zero); - } - else if (prec > len) { - FILL_('0', prec - len); - PUSH_(t, len); - } - else if (prec > 0) { - PUSH_(t + len - prec, prec); - } - if (fill && (flags&FMINUS)) { - FILL_(' ', fill); - } - RB_GC_GUARD(val); - break; - } - case 'g': - case 'G': - case 'e': - case 'E': - /* TODO: rational support */ - case 'a': - case 'A': - float_value: - { - VALUE val = GETARG(); - double fval; - - fval = RFLOAT_VALUE(rb_Float(val)); - if (!isfinite(fval)) { - const char *expr; - int need; - int elen; - char sign = '\0'; - - if (isnan(fval)) { - expr = "NaN"; - } - else { - expr = "Inf"; - } - need = (int)strlen(expr); - elen = need; - if (!isnan(fval) && fval < 0.0) - sign = '-'; - else if (flags & (FPLUS|FSPACE)) - sign = (flags & FPLUS) ? '+' : ' '; - if (sign) - ++need; - if ((flags & FWIDTH) && need < width) - need = width; - - FILL(' ', need); - if (flags & FMINUS) { - if (sign) - buf[blen - need--] = sign; - memcpy(&buf[blen - need], expr, elen); - } - else { - if (sign) - buf[blen - elen - 1] = sign; - memcpy(&buf[blen - elen], expr, elen); - } - break; - } - else { - int cr = ENC_CODERANGE(result); - char fbuf[2*BIT_DIGITS(SIZEOF_INT*CHAR_BIT)+10]; - char *fmt = fmt_setup(fbuf, sizeof(fbuf), *p, flags, width, prec); - rb_str_set_len(result, blen); - rb_str_catf(result, fmt, fval); - ENC_CODERANGE_SET(result, cr); - bsiz = rb_str_capacity(result); - RSTRING_GETMEM(result, buf, blen); - } - } - break; - } - flags = FNONE; + len = rb_long2int(RSTRING_END(tmp) - 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 -= (int)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)) { + FILL(' ', width); + width = 0; + } + if (sc) PUSH(&sc, 1); + if (prefix) { + int plen = (int)strlen(prefix); + PUSH(prefix, plen); + } + if (dots) PUSH("..", 2); + if (prec > len) { + CHECK(prec - len); + if (!sign && valsign < 0) { + char c = sign_bits(base, p); + FILL_(c, prec - len); + } + else if ((flags & (FMINUS|FPREC)) != FMINUS) { + FILL_('0', prec - len); + } + } + PUSH(s, len); + RB_GC_GUARD(tmp); + FILL(' ', width); + } + break; + + case 'f': + { + VALUE val = GETARG(), num, den; + int sign = (flags&FPLUS) ? 1 : 0, zero = 0; + long len, fill; + if (RB_INTEGER_TYPE_P(val)) { + den = INT2FIX(1); + num = val; + } + else if (RB_TYPE_P(val, T_RATIONAL)) { + den = rb_rational_den(val); + num = rb_rational_num(val); + } + else { + nextvalue = val; + goto float_value; + } + if (!(flags&FPREC)) prec = default_float_precision; + if (FIXNUM_P(num)) { + if ((SIGNED_VALUE)num < 0) { + long n = -FIX2LONG(num); + num = LONG2FIX(n); + sign = -1; + } + } + else if (BIGNUM_NEGATIVE_P(num)) { + sign = -1; + num = rb_big_uminus(num); + } + if (den != INT2FIX(1)) { + num = rb_int_mul(num, rb_int_positive_pow(10, prec)); + num = rb_int_plus(num, rb_int_idiv(den, INT2FIX(2))); + num = rb_int_idiv(num, den); + } + else if (prec >= 0) { + zero = prec; + } + val = rb_int2str(num, 10); + len = RSTRING_LEN(val) + zero; + if (prec >= len) len = prec + 1; /* integer part 0 */ + if (sign || (flags&FSPACE)) ++len; + if (prec > 0) ++len; /* period */ + fill = width > len ? width - len : 0; + CHECK(fill + len); + if (fill && !(flags&(FMINUS|FZERO))) { + FILL_(' ', fill); + } + if (sign || (flags&FSPACE)) { + buf[blen++] = sign > 0 ? '+' : sign < 0 ? '-' : ' '; + } + if (fill && (flags&(FMINUS|FZERO)) == FZERO) { + FILL_('0', fill); + } + len = RSTRING_LEN(val) + zero; + t = RSTRING_PTR(val); + if (len > prec) { + PUSH_(t, len - prec); + } + else { + buf[blen++] = '0'; + } + if (prec > 0) { + buf[blen++] = '.'; + } + if (zero) { + FILL_('0', zero); + } + else if (prec > len) { + FILL_('0', prec - len); + PUSH_(t, len); + } + else if (prec > 0) { + PUSH_(t + len - prec, prec); + } + if (fill && (flags&FMINUS)) { + FILL_(' ', fill); + } + RB_GC_GUARD(val); + break; + } + case 'g': + case 'G': + case 'e': + case 'E': + /* TODO: rational support */ + case 'a': + case 'A': + float_value: + { + VALUE val = GETARG(); + double fval; + + fval = RFLOAT_VALUE(rb_Float(val)); + if (!isfinite(fval)) { + const char *expr; + int need; + int elen; + char sign = '\0'; + + if (isnan(fval)) { + expr = "NaN"; + } + else { + expr = "Inf"; + } + need = (int)strlen(expr); + elen = need; + if (!isnan(fval) && fval < 0.0) + sign = '-'; + else if (flags & (FPLUS|FSPACE)) + sign = (flags & FPLUS) ? '+' : ' '; + if (sign) + ++need; + if ((flags & FWIDTH) && need < width) + need = width; + + FILL(' ', need); + if (flags & FMINUS) { + if (sign) + buf[blen - need--] = sign; + memcpy(&buf[blen - need], expr, elen); + } + else { + if (sign) + buf[blen - elen - 1] = sign; + memcpy(&buf[blen - elen], expr, elen); + } + break; + } + else { + int cr = ENC_CODERANGE(result); + char fbuf[2*BIT_DIGITS(SIZEOF_INT*CHAR_BIT)+10]; + char *fmt = fmt_setup(fbuf, sizeof(fbuf), *p, flags, width, prec); + rb_str_set_len(result, blen); + rb_str_catf(result, fmt, fval); + ENC_CODERANGE_SET(result, cr); + bsiz = rb_str_capacity(result); + RSTRING_GETMEM(result, buf, blen); + } + } + break; + } + flags = FNONE; } + update_coderange(FALSE); sprint_exit: rb_str_tmp_frozen_release(orig, fmt); /* 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); + 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); @@ -949,12 +959,12 @@ fmt_setup(char *buf, size_t size, int c, int flags, int width, int prec) *--buf = c; if (flags & FPREC) { - buf = ruby_ultoa(prec, buf, 10, 0); - *--buf = '.'; + buf = ruby_ultoa(prec, buf, 10, 0); + *--buf = '.'; } if (flags & FWIDTH) { - buf = ruby_ultoa(width, buf, 10, 0); + buf = ruby_ultoa(width, buf, 10, 0); } if (flags & FSPACE) *--buf = ' '; @@ -1008,7 +1018,7 @@ int ruby_vsnprintf(char *str, size_t n, const char *fmt, va_list ap) { if (str && (ssize_t)n < 1) - return (EOF); + return (EOF); return ruby_do_vsnprintf(str, n, fmt, ap); } @@ -1038,7 +1048,7 @@ ruby_snprintf(char *str, size_t n, char const *fmt, ...) va_list ap; if (str && (ssize_t)n < 1) - return (EOF); + return (EOF); va_start(ap, fmt); ret = ruby_do_vsnprintf(str, n, fmt, ap); @@ -1061,22 +1071,22 @@ ruby__sfvwrite(register rb_printf_buffer *fp, register struct __suio *uio) long blen = buf - RSTRING_PTR(result), bsiz = fp->_w; if (RBASIC(result)->klass) { - rb_raise(rb_eRuntimeError, "rb_vsprintf reentered"); + rb_raise(rb_eRuntimeError, "rb_vsprintf reentered"); } if (uio->uio_resid == 0) - return 0; + return 0; #if SIZE_MAX > LONG_MAX if (uio->uio_resid >= LONG_MAX) - rb_raise(rb_eRuntimeError, "too big string"); + rb_raise(rb_eRuntimeError, "too big string"); #endif len = (long)uio->uio_resid; CHECK(len); buf += blen; fp->_w = bsiz; for (iov = uio->uio_iov; len > 0; ++iov) { - MEMCPY(buf, iov->iov_base, char, n = iov->iov_len); - buf += n; - len -= n; + MEMCPY(buf, iov->iov_base, char, n = iov->iov_len); + buf += n; + len -= n; } fp->_p = (unsigned char *)buf; rb_str_set_len(result, buf - RSTRING_PTR(result)); @@ -1093,51 +1103,41 @@ ruby__sfvextra(rb_printf_buffer *fp, size_t valsize, void *valp, long *sz, int s if (valsize != sizeof(VALUE)) return 0; value = *(VALUE *)valp; if (RBASIC(result)->klass) { - rb_raise(rb_eRuntimeError, "rb_vsprintf reentered"); + rb_raise(rb_eRuntimeError, "rb_vsprintf reentered"); } if (sign == '+') { - if (RB_TYPE_P(value, T_CLASS)) { # define LITERAL(str) (*sz = rb_strlen_lit(str), str) - - if (value == rb_cNilClass) { - return LITERAL("nil"); - } - else if (value == rb_cInteger) { - return LITERAL("Integer"); - } - else if (value == rb_cSymbol) { - return LITERAL("Symbol"); - } - else if (value == rb_cTrueClass) { - return LITERAL("true"); - } - else if (value == rb_cFalseClass) { - return LITERAL("false"); - } + /* optimize special const cases */ + switch (value) { +# define LITERAL_CASE(x) case Q##x: return LITERAL(#x) + LITERAL_CASE(nil); + LITERAL_CASE(true); + LITERAL_CASE(false); +# undef LITERAL_CASE + } # undef LITERAL - } - value = rb_inspect(value); + value = rb_inspect(value); } else if (SYMBOL_P(value)) { - value = rb_sym2str(value); - if (sign == ' ' && !rb_str_symname_p(value)) { - value = rb_str_escape(value); - } + value = rb_sym2str(value); + if (sign == ' ' && !rb_str_symname_p(value)) { + value = rb_str_escape(value); + } } else { - value = rb_obj_as_string(value); - if (sign == ' ') value = QUOTE(value); + value = rb_obj_as_string(value); + if (sign == ' ') value = QUOTE(value); } enc = rb_enc_compatible(result, value); if (enc) { - rb_enc_associate(result, enc); + rb_enc_associate(result, enc); } else { - enc = rb_enc_get(result); - value = rb_str_conv_enc_opts(value, rb_enc_get(value), enc, - ECONV_UNDEF_REPLACE|ECONV_INVALID_REPLACE, - Qnil); - *(volatile VALUE *)valp = value; + enc = rb_enc_get(result); + value = rb_str_conv_enc_opts(value, rb_enc_get(value), enc, + ECONV_UNDEF_REPLACE|ECONV_INVALID_REPLACE, + Qnil); + *(volatile VALUE *)valp = value; } StringValueCStr(value); RSTRING_GETMEM(value, cp, *sz); @@ -1145,36 +1145,54 @@ ruby__sfvextra(rb_printf_buffer *fp, size_t valsize, void *valp, long *sz, int s return cp; } -VALUE -rb_enc_vsprintf(rb_encoding *enc, const char *fmt, va_list ap) +static void +ruby_vsprintf0(VALUE result, char *p, const char *fmt, va_list ap) { rb_printf_buffer_extra buffer; #define f buffer.base - VALUE result; + VALUE klass = RBASIC(result)->klass; + int coderange = ENC_CODERANGE(result); + long scanned = 0; + + if (coderange != ENC_CODERANGE_UNKNOWN) scanned = p - RSTRING_PTR(result); f._flags = __SWR | __SSTR; f._bf._size = 0; - f._w = 120; - result = rb_str_buf_new(f._w); - if (enc) { - if (rb_enc_mbminlen(enc) > 1) { - /* the implementation deeply depends on plain char */ - rb_raise(rb_eArgError, "cannot construct wchar_t based encoding string: %s", - rb_enc_name(enc)); - } - rb_enc_associate(result, enc); - } + f._w = rb_str_capacity(result); f._bf._base = (unsigned char *)result; - f._p = (unsigned char *)RSTRING_PTR(result); + f._p = (unsigned char *)p; RBASIC_CLEAR_CLASS(result); f.vwrite = ruby__sfvwrite; f.vextra = ruby__sfvextra; buffer.value = 0; BSD_vfprintf(&f, fmt, ap); - RBASIC_SET_CLASS_RAW(result, rb_cString); - rb_str_resize(result, (char *)f._p - RSTRING_PTR(result)); + RBASIC_SET_CLASS_RAW(result, klass); + p = RSTRING_PTR(result); + long blen = (char *)f._p - p; + if (scanned < blen) { + rb_str_coderange_scan_restartable(p + scanned, p + blen, rb_enc_get(result), &coderange); + ENC_CODERANGE_SET(result, coderange); + } + rb_str_resize(result, blen); #undef f +} +VALUE +rb_enc_vsprintf(rb_encoding *enc, const char *fmt, va_list ap) +{ + const int initial_len = 120; + VALUE result; + + result = rb_str_buf_new(initial_len); + if (enc) { + if (rb_enc_mbminlen(enc) > 1) { + /* the implementation deeply depends on plain char */ + rb_raise(rb_eArgError, "cannot construct wchar_t based encoding string: %s", + rb_enc_name(enc)); + } + rb_enc_associate(result, enc); + } + ruby_vsprintf0(result, RSTRING_PTR(result), fmt, ap); return result; } @@ -1213,26 +1231,9 @@ rb_sprintf(const char *format, ...) VALUE rb_str_vcatf(VALUE str, const char *fmt, va_list ap) { - rb_printf_buffer_extra buffer; -#define f buffer.base - VALUE klass; - StringValue(str); rb_str_modify(str); - f._flags = __SWR | __SSTR; - f._bf._size = 0; - f._w = rb_str_capacity(str); - f._bf._base = (unsigned char *)str; - f._p = (unsigned char *)RSTRING_END(str); - klass = RBASIC(str)->klass; - RBASIC_CLEAR_CLASS(str); - f.vwrite = ruby__sfvwrite; - f.vextra = ruby__sfvextra; - buffer.value = 0; - BSD_vfprintf(&f, fmt, ap); - RBASIC_SET_CLASS_RAW(str, klass); - rb_str_resize(str, (char *)f._p - RSTRING_PTR(str)); -#undef f + ruby_vsprintf0(str, RSTRING_END(str), fmt, ap); return str; } |