diff options
Diffstat (limited to 'mrbgems/mruby-sprintf/src/sprintf.c')
| -rw-r--r-- | mrbgems/mruby-sprintf/src/sprintf.c | 694 |
1 files changed, 352 insertions, 342 deletions
diff --git a/mrbgems/mruby-sprintf/src/sprintf.c b/mrbgems/mruby-sprintf/src/sprintf.c index 3fce60237..8de3a4541 100644 --- a/mrbgems/mruby-sprintf/src/sprintf.c +++ b/mrbgems/mruby-sprintf/src/sprintf.c @@ -4,23 +4,19 @@ ** See Copyright Notice in mruby.h */ -#include "mruby.h" - -#include <limits.h> -#include <stdio.h> +#include <mruby.h> +#include <mruby/string.h> +#include <mruby/hash.h> +#include <mruby/numeric.h> +#include <mruby/presym.h> #include <string.h> -#include "mruby/string.h" -#include "mruby/hash.h" -#include "mruby/numeric.h" -#include <math.h> #include <ctype.h> #define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */ #define BITSPERDIG MRB_INT_BIT -#define EXTENDSIGN(n, l) (((~0 << (n)) >> (((n)*(l)) % BITSPERDIG)) & ~(~0 << (n))) +#define EXTENDSIGN(n, l) (((~0U << (n)) >> (((n)*(l)) % BITSPERDIG)) & ~(~0U << (n))) -mrb_value mrb_str_format(mrb_state *, int, const mrb_value *, mrb_value); -static void fmt_setup(char*,size_t,int,int,mrb_int,mrb_int); +mrb_value mrb_str_format(mrb_state *, mrb_int, const mrb_value *, mrb_value); static char* remove_sign_bits(char *str, int base) @@ -68,44 +64,38 @@ sign_bits(int base, const char *p) return c; } -static mrb_value -mrb_fix2binstr(mrb_state *mrb, mrb_value x, int base) +static char * +mrb_uint_to_cstr(char *buf, size_t len, mrb_int num, int base) { - char buf[64], *b = buf + sizeof buf; - mrb_int num = mrb_fixnum(x); - unsigned long val = (unsigned long)num; + char *b = buf + len - 1; + const int mask = base-1; + int shift; + mrb_uint val = (uint64_t)num; char d; - if (base != 2) { - mrb_raisef(mrb, E_ARGUMENT_ERROR, "invalid radix %S", mrb_fixnum_value(base)); + if (num == 0) { + buf[0] = '0'; buf[1] = '\0'; + return buf; } - - if (val >= (1 << 10)) - val &= 0x3ff; - - if (val == 0) { - return mrb_str_new_lit(mrb, "0"); + switch (base) { + case 16: d = 'f'; shift = 4; break; + case 8: d = '7'; shift = 3; break; + case 2: d = '1'; shift = 1; break; + default: return NULL; } *--b = '\0'; do { - *--b = mrb_digitmap[(int)(val % base)]; - } while (val /= base); + *--b = mrb_digitmap[(int)(val & mask)]; + } while (val >>= shift); if (num < 0) { b = remove_sign_bits(b, base); - switch (base) { - case 16: d = 'f'; break; - case 8: d = '7'; break; - case 2: d = '1'; break; - default: d = 0; break; - } - if (d && *b != d) { *--b = d; } } - return mrb_str_new_cstr(mrb, b); + return b; } #define FNONE 0 @@ -118,20 +108,64 @@ mrb_fix2binstr(mrb_state *mrb, mrb_value x, int base) #define FPREC 64 #define FPREC0 128 -#define CHECK(l) do {\ -/* int cr = ENC_CODERANGE(result);*/\ - while (blen + (l) >= bsiz) {\ +#ifndef MRB_NO_FLOAT +static int +fmt_float(char *buf, size_t buf_size, char fmt, int flags, mrb_int width, int prec, mrb_float f) +{ + char sign = '\0'; + int left_align = 0; + int zero_pad = 0; + + if (flags & FSHARP) fmt |= 0x80; + if (flags & FPLUS) sign = '+'; + if (flags & FMINUS) left_align = 1; + if (flags & FZERO) zero_pad = 1; + if (flags & FSPACE) sign = ' '; + + int len = mrb_format_float(f, buf, buf_size, fmt, prec, sign); + + // buf[0] < '0' returns true if the first character is space, + or - + // buf[1] < '9' matches a digit, and doesn't match when we get back +nan or +inf + if (buf[0] < '0' && buf[1] <= '9' && zero_pad) { + buf++; + width--; + len--; + } + if (*buf < '0' || *buf >= '9') { + // For inf or nan, we don't want to zero pad. + zero_pad = 0; + } + if (len >= width) { + return len; + } + buf[width] = '\0'; + if (left_align) { + memset(&buf[len], ' ', width - len); + return width; + } + memmove(&buf[width - len], buf, len); + if (zero_pad) { + memset(buf, '0', width - len); + } else { + memset(buf, ' ', width - len); + } + return width; +} +#endif + +#define CHECK(l) do { \ + while ((l) >= bsiz - blen) {\ + if (bsiz > MRB_INT_MAX/2) mrb_raise(mrb, E_ARGUMENT_ERROR, "too big specifier"); \ bsiz*=2;\ }\ mrb_str_resize(mrb, result, bsiz);\ -/* ENC_CODERANGE_SET(result, cr);*/\ buf = RSTRING_PTR(result);\ } while (0) #define PUSH(s, l) do { \ CHECK(l);\ memcpy(&buf[blen], s, l);\ - blen += (l);\ + blen += (mrb_int)(l);\ } while (0) #define FILL(c, l) do { \ @@ -140,71 +174,108 @@ mrb_fix2binstr(mrb_state *mrb, mrb_value x, int base) blen += (l);\ } while (0) -#define GETARG() (!mrb_undef_p(nextvalue) ? nextvalue : \ - posarg == -1 ? \ - (mrb_raisef(mrb, E_ARGUMENT_ERROR, "unnumbered(%S) mixed with numbered", mrb_fixnum_value(nextarg)), mrb_undef_value()) : \ - posarg == -2 ? \ - (mrb_raisef(mrb, E_ARGUMENT_ERROR, "unnumbered(%S) mixed with named", mrb_fixnum_value(nextarg)), mrb_undef_value()) : \ +static void +check_next_arg(mrb_state *mrb, int posarg, int nextarg) +{ + switch (posarg) { + case -1: + mrb_raisef(mrb, E_ARGUMENT_ERROR, "unnumbered(%d) mixed with numbered", nextarg); + break; + case -2: + mrb_raisef(mrb, E_ARGUMENT_ERROR, "unnumbered(%d) mixed with named", nextarg); + break; + default: + break; + } +} + +static void +check_pos_arg(mrb_state *mrb, int posarg, mrb_int n) +{ + if (posarg > 0) { + mrb_raisef(mrb, E_ARGUMENT_ERROR, "numbered(%i) after unnumbered(%d)", + n, posarg); + } + if (posarg == -2) { + mrb_raisef(mrb, E_ARGUMENT_ERROR, "numbered(%i) after named", n); + } + if (n < 1) { + mrb_raisef(mrb, E_ARGUMENT_ERROR, "invalid index - %i$", n); + } +} + +static void +check_name_arg(mrb_state *mrb, int posarg, const char *name, size_t len) +{ + if (posarg > 0) { + mrb_raisef(mrb, E_ARGUMENT_ERROR, "named%l after unnumbered(%d)", + name, len, posarg); + } + if (posarg == -1) { + mrb_raisef(mrb, E_ARGUMENT_ERROR, "named%l after numbered", name, len); + } +} + +#define GETNEXTARG() (\ + check_next_arg(mrb, posarg, nextarg),\ (posarg = nextarg++, GETNTHARG(posarg))) -#define GETPOSARG(n) (posarg > 0 ? \ - (mrb_raisef(mrb, E_ARGUMENT_ERROR, "numbered(%S) after unnumbered(%S)", mrb_fixnum_value(n), mrb_fixnum_value(posarg)), mrb_undef_value()) : \ - posarg == -2 ? \ - (mrb_raisef(mrb, E_ARGUMENT_ERROR, "numbered(%S) after named", mrb_fixnum_value(n)), mrb_undef_value()) : \ - ((n < 1) ? \ - (mrb_raisef(mrb, E_ARGUMENT_ERROR, "invalid index - %S$", mrb_fixnum_value(n)), mrb_undef_value()) : \ - (posarg = -1, GETNTHARG(n)))) +#define GETARG() (!mrb_undef_p(nextvalue) ? nextvalue : GETNEXTARG()) + +#define GETPOSARG(n) (\ + check_pos_arg(mrb, posarg, n),\ + (posarg = -1, GETNTHARG(n))) #define GETNTHARG(nth) \ ((nth >= argc) ? (mrb_raise(mrb, E_ARGUMENT_ERROR, "too few arguments"), mrb_undef_value()) : argv[nth]) -#define GETNAMEARG(id, name, len) ( \ - posarg > 0 ? \ - (mrb_raisef(mrb, E_ARGUMENT_ERROR, "named%S after unnumbered(%S)", mrb_str_new(mrb, (name), (len)), mrb_fixnum_value(posarg)), mrb_undef_value()) : \ - posarg == -1 ? \ - (mrb_raisef(mrb, E_ARGUMENT_ERROR, "named%S after numbered", mrb_str_new(mrb, (name), (len))), mrb_undef_value()) : \ - (posarg = -2, mrb_hash_fetch(mrb, get_hash(mrb, &hash, argc, argv), id, mrb_undef_value()))) - -#define GETNUM(n, val) \ - for (; p < end && ISDIGIT(*p); p++) {\ - int next_n = 10 * n + (*p - '0'); \ - if (next_n / 10 != n) {\ - mrb_raise(mrb, E_ARGUMENT_ERROR, #val " too big"); \ - } \ - n = next_n; \ - } \ - if (p >= end) { \ - mrb_raise(mrb, E_ARGUMENT_ERROR, "malformed format string - %*[0-9]"); \ - } +#define CHECKNAMEARG(name, len) (\ + check_name_arg(mrb, posarg, name, len),\ + posarg = -2) + +#define GETNUM(n, val) do { \ + if (!(p = get_num(mrb, p, end, &(n)))) \ + mrb_raise(mrb, E_ARGUMENT_ERROR, #val " too big 1"); \ +} while(0) #define GETASTER(num) do { \ + mrb_value tmp_v; \ t = p++; \ - n = 0; \ GETNUM(n, val); \ if (*p == '$') { \ - tmp = GETPOSARG(n); \ + tmp_v = GETPOSARG(n); \ } \ else { \ - tmp = GETARG(); \ + tmp_v = GETNEXTARG(); \ p = t; \ } \ - num = mrb_fixnum(tmp); \ + num = mrb_as_int(mrb, tmp_v); \ } while (0) -static mrb_value -get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) +static const char * +get_num(mrb_state *mrb, const char *p, const char *end, int *valp) +{ + char *e; + mrb_int n = mrb_int_read(p, end, &e); + if (e == NULL || n > INT_MAX) return NULL; + *valp = (int)n; + return e; +} + +static void +get_hash(mrb_state *mrb, mrb_value *hash, mrb_int argc, const mrb_value *argv) { mrb_value tmp; - if (!mrb_undef_p(*hash)) return *hash; + if (!mrb_undef_p(*hash)) return; if (argc != 2) { mrb_raise(mrb, E_ARGUMENT_ERROR, "one hash required"); } - tmp = mrb_check_convert_type(mrb, argv[1], MRB_TT_HASH, "Hash", "to_hash"); + tmp = mrb_check_hash_type(mrb, argv[1]); if (mrb_nil_p(tmp)) { mrb_raise(mrb, E_ARGUMENT_ERROR, "one hash required"); } - return (*hash = tmp); + *hash = tmp; } /* @@ -233,41 +304,37 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * ------+-------------------------------------------------------------- * b | Convert argument as a binary number. * | Negative numbers will be displayed as a two's complement - * | prefixed with `..1'. - * B | Equivalent to `b', but uses an uppercase 0B for prefix + * | prefixed with '..1'. + * B | Equivalent to 'b', but uses an uppercase 0B for prefix * | in the alternative format by #. * d | Convert argument as a decimal number. - * i | Identical to `d'. + * i | Identical to 'd'. * o | Convert argument as an octal number. * | Negative numbers will be displayed as a two's complement - * | prefixed with `..7'. - * u | Identical to `d'. + * | prefixed with '..7'. + * u | Identical to 'd'. * x | Convert argument as a hexadecimal number. * | Negative numbers will be displayed as a two's complement - * | prefixed with `..f' (representing an infinite string of + * | prefixed with '..f' (representing an infinite string of * | leading 'ff's). - * X | Equivalent to `x', but uses uppercase letters. + * X | Equivalent to 'x', but uses uppercase letters. * * Field | Float Format * ------+-------------------------------------------------------------- - * e | Convert floating point argument into exponential notation + * e | Convert floating-point argument into exponential notation * | with one digit before the decimal point as [-]d.dddddde[+-]dd. * | The precision specifies the number of digits after the decimal * | point (defaulting to six). - * E | Equivalent to `e', but uses an uppercase E to indicate + * E | Equivalent to 'e', but uses an uppercase E to indicate * | the exponent. - * f | Convert floating point argument as [-]ddd.dddddd, + * f | Convert floating-point argument as [-]ddd.dddddd, * | where the precision specifies the number of digits after * | the decimal point. - * g | Convert a floating point number using exponential form + * g | Convert a floating-point number using exponential form * | if the exponent is less than -4 or greater than or * | equal to the precision, or in dd.dddd form otherwise. * | The precision specifies the number of significant digits. - * G | Equivalent to `g', but use an uppercase `E' in exponent form. - * a | Convert floating point argument as [-]0xh.hhhhp[+-]dd, - * | which is consisted from optional sign, "0x", fraction part - * | as hexadecimal, "p", and exponential part as decimal. - * A | Equivalent to `a', but use uppercase `X' and `P'. + * G | Equivalent to 'g', but use an uppercase 'E' in exponent form. * * Field | Other Format * ------+-------------------------------------------------------------- @@ -286,7 +353,7 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * ---------+---------------+----------------------------------------- * space | bBdiouxX | Leave a space at the start of * | aAeEfgG | non-negative numbers. - * | (numeric fmt) | For `o', `x', `X', `b' and `B', use + * | (numeric fmt) | For 'o', 'x', 'X', 'b' and 'B', use * | | a minus sign with absolute value for * | | negative values. * ---------+---------------+----------------------------------------- @@ -296,27 +363,27 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * | | sprintf string. * ---------+---------------+----------------------------------------- * # | bBoxX | Use an alternative format. - * | aAeEfgG | For the conversions `o', increase the precision - * | | until the first digit will be `0' if + * | aAeEfgG | For the conversions 'o', increase the precision + * | | until the first digit will be '0' if * | | it is not formatted as complements. - * | | For the conversions `x', `X', `b' and `B' - * | | on non-zero, prefix the result with ``0x'', - * | | ``0X'', ``0b'' and ``0B'', respectively. - * | | For `a', `A', `e', `E', `f', `g', and 'G', + * | | For the conversions 'x', 'X', 'b' and 'B' + * | | on non-zero, prefix the result with "0x", + * | | "0X", "0b" and "0B", respectively. + * | | For 'e', 'E', 'f', 'g', and 'G', * | | force a decimal point to be added, * | | even if no digits follow. - * | | For `g' and 'G', do not remove trailing zeros. + * | | For 'g' and 'G', do not remove trailing zeros. * ---------+---------------+----------------------------------------- * + | bBdiouxX | Add a leading plus sign to non-negative * | aAeEfgG | numbers. - * | (numeric fmt) | For `o', `x', `X', `b' and `B', use + * | (numeric fmt) | For 'o', 'x', 'X', 'b' and 'B', use * | | a minus sign with absolute value for * | | negative values. * ---------+---------------+----------------------------------------- * - | all | Left-justify the result of this conversion. * ---------+---------------+----------------------------------------- * 0 (zero) | bBdiouxX | Pad with zeros, not spaces. - * | aAeEfgG | For `o', `x', `X', `b' and `B', radix-1 + * | aAeEfgG | For 'o', 'x', 'X', 'b' and 'B', radix-1 * | (numeric fmt) | is used for negative numbers formatted as * | | complements. * ---------+---------------+----------------------------------------- @@ -327,21 +394,21 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * * Examples of flags: * - * # `+' and space flag specifies the sign of non-negative numbers. + * # '+' and space flag specifies the sign of non-negative numbers. * sprintf("%d", 123) #=> "123" * sprintf("%+d", 123) #=> "+123" * sprintf("% d", 123) #=> " 123" * - * # `#' flag for `o' increases number of digits to show `0'. - * # `+' and space flag changes format of negative numbers. + * # '#' flag for 'o' increases number of digits to show '0'. + * # '+' and space flag changes format of negative numbers. * sprintf("%o", 123) #=> "173" * sprintf("%#o", 123) #=> "0173" * sprintf("%+o", -123) #=> "-173" * sprintf("%o", -123) #=> "..7605" * sprintf("%#o", -123) #=> "..7605" * - * # `#' flag for `x' add a prefix `0x' for non-zero numbers. - * # `+' and space flag disables complements for negative numbers. + * # '#' flag for 'x' add a prefix '0x' for non-zero numbers. + * # '+' and space flag disables complements for negative numbers. * sprintf("%x", 123) #=> "7b" * sprintf("%#x", 123) #=> "0x7b" * sprintf("%+x", -123) #=> "-7b" @@ -349,12 +416,12 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * sprintf("%#x", -123) #=> "0x..f85" * sprintf("%#x", 0) #=> "0" * - * # `#' for `X' uses the prefix `0X'. + * # '#' for 'X' uses the prefix '0X'. * sprintf("%X", 123) #=> "7B" * sprintf("%#X", 123) #=> "0X7B" * - * # `#' flag for `b' add a prefix `0b' for non-zero numbers. - * # `+' and space flag disables complements for negative numbers. + * # '#' flag for 'b' add a prefix '0b' for non-zero numbers. + * # '+' and space flag disables complements for negative numbers. * sprintf("%b", 123) #=> "1111011" * sprintf("%#b", 123) #=> "0b1111011" * sprintf("%+b", -123) #=> "-1111011" @@ -362,19 +429,19 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * sprintf("%#b", -123) #=> "0b..10000101" * sprintf("%#b", 0) #=> "0" * - * # `#' for `B' uses the prefix `0B'. + * # '#' for 'B' uses the prefix '0B'. * sprintf("%B", 123) #=> "1111011" * sprintf("%#B", 123) #=> "0B1111011" * - * # `#' for `e' forces to show the decimal point. + * # '#' for 'e' forces to show the decimal point. * sprintf("%.0e", 1) #=> "1e+00" * sprintf("%#.0e", 1) #=> "1.e+00" * - * # `#' for `f' forces to show the decimal point. + * # '#' for 'f' forces to show the decimal point. * sprintf("%.0f", 1234) #=> "1234" * sprintf("%#.0f", 1234) #=> "1234." * - * # `#' for `g' forces to show the decimal point. + * # '#' for 'g' forces to show the decimal point. * # It also disables stripping lowest zeros. * sprintf("%g", 123.4) #=> "123.4" * sprintf("%#g", 123.4) #=> "123.400" @@ -408,7 +475,7 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * * Examples of precisions: * - * # precision for `d', 'o', 'x' and 'b' is + * # precision for 'd', 'o', 'x' and 'b' is * # minimum number of digits <------> * sprintf("%20.8d", 123) #=> " 00000123" * sprintf("%20.8o", 123) #=> " 00000173" @@ -419,8 +486,8 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * sprintf("%20.8x", -123) #=> " ..ffff85" * sprintf("%20.8b", -11) #=> " ..110101" * - * # "0x" and "0b" for `#x' and `#b' is not counted for - * # precision but "0" for `#o' is counted. <------> + * # "0x" and "0b" for '#x' and '#b' is not counted for + * # precision but "0" for '#o' is counted. <------> * sprintf("%#20.8d", 123) #=> " 00000123" * sprintf("%#20.8o", 123) #=> " 00000173" * sprintf("%#20.8x", 123) #=> " 0x0000007b" @@ -430,22 +497,22 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * sprintf("%#20.8x", -123) #=> " 0x..ffff85" * sprintf("%#20.8b", -11) #=> " 0b..110101" * - * # precision for `e' is number of + * # precision for 'e' is number of * # digits after the decimal point <------> * sprintf("%20.8e", 1234.56789) #=> " 1.23456789e+03" * - * # precision for `f' is number of + * # precision for 'f' is number of * # digits after the decimal point <------> * sprintf("%20.8f", 1234.56789) #=> " 1234.56789000" * - * # precision for `g' is number of + * # precision for 'g' is number of * # significant digits <-------> * sprintf("%20.8g", 1234.56789) #=> " 1234.5679" * * # <-------> * sprintf("%20.8g", 123456789) #=> " 1.2345679e+08" * - * # precision for `s' is + * # precision for 's' is * # maximum number of characters <------> * sprintf("%20.8s", "string test") #=> " string t" * @@ -461,18 +528,18 @@ get_hash(mrb_state *mrb, mrb_value *hash, int argc, const mrb_value *argv) * For more complex formatting, Ruby supports a reference by name. * %<name>s style uses format style, but %{name} style doesn't. * - * Exapmles: + * Examples: * sprintf("%<foo>d : %<bar>f", { :foo => 1, :bar => 2 }) * #=> 1 : 2.000000 * sprintf("%{foo}f", { :foo => 1 }) * # => "1f" */ -mrb_value +static mrb_value mrb_f_sprintf(mrb_state *mrb, mrb_value obj) { mrb_int argc; - mrb_value *argv; + const mrb_value *argv; mrb_get_args(mrb, "*", &argv, &argc); @@ -486,60 +553,60 @@ mrb_f_sprintf(mrb_state *mrb, mrb_value obj) } mrb_value -mrb_str_format(mrb_state *mrb, int argc, const mrb_value *argv, mrb_value fmt) +mrb_str_format(mrb_state *mrb, mrb_int argc, const mrb_value *argv, mrb_value fmt) { const char *p, *end; char *buf; mrb_int blen; mrb_int bsiz; mrb_value result; - mrb_int n; - mrb_int width; - mrb_int prec; - int flags = FNONE; + int n; + int width; + int prec; int nextarg = 1; int posarg = 0; mrb_value nextvalue; - mrb_value tmp; mrb_value str; mrb_value hash = mrb_undef_value(); -#define CHECK_FOR_WIDTH(f) \ - if ((f) & FWIDTH) { \ - mrb_raise(mrb, E_ARGUMENT_ERROR, "width given twice"); \ - } \ - if ((f) & FPREC0) { \ - mrb_raise(mrb, E_ARGUMENT_ERROR, "width after precision"); \ +#define CHECK_FOR_WIDTH(f) \ + if ((f) & FWIDTH) { \ + mrb_raise(mrb, E_ARGUMENT_ERROR, "width given twice"); \ + } \ + if ((f) & FPREC0) { \ + mrb_raise(mrb, E_ARGUMENT_ERROR, "width after precision"); \ } -#define CHECK_FOR_FLAGS(f) \ - if ((f) & FWIDTH) { \ - mrb_raise(mrb, E_ARGUMENT_ERROR, "flag after width"); \ - } \ - if ((f) & FPREC0) { \ - mrb_raise(mrb, E_ARGUMENT_ERROR, "flag after precision"); \ +#define CHECK_FOR_FLAGS(f) \ + if ((f) & FWIDTH) { \ + mrb_raise(mrb, E_ARGUMENT_ERROR, "flag after width"); \ + } \ + if ((f) & FPREC0) { \ + mrb_raise(mrb, E_ARGUMENT_ERROR, "flag after precision"); \ } ++argc; --argv; - fmt = mrb_str_to_str(mrb, fmt); + mrb_ensure_string_type(mrb, fmt); p = RSTRING_PTR(fmt); end = p + RSTRING_LEN(fmt); blen = 0; bsiz = 120; - result = mrb_str_buf_new(mrb, bsiz); + result = mrb_str_new_capa(mrb, bsiz); buf = RSTRING_PTR(result); memset(buf, 0, bsiz); for (; p < end; p++) { const char *t; mrb_sym id = 0; + int flags = FNONE; for (t = p; t < end && *t != '%'; t++) ; + if (t + 1 == end) ++t; PUSH(p, t - p); if (t >= end) goto sprint_exit; /* end of fmt string */ - p = t + 1; /* skip `%' */ + p = t + 1; /* skip '%' */ width = prec = -1; nextvalue = mrb_undef_value(); @@ -547,7 +614,7 @@ mrb_str_format(mrb_state *mrb, int argc, const mrb_value *argv, mrb_value fmt) retry: switch (*p) { default: - mrb_raisef(mrb, E_ARGUMENT_ERROR, "malformed format string - \\%%S", mrb_str_new(mrb, p, 1)); + mrb_raisef(mrb, E_ARGUMENT_ERROR, "malformed format string - %%%c", *p); break; case ' ': @@ -582,11 +649,10 @@ 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 (!mrb_undef_p(nextvalue)) { - mrb_raisef(mrb, E_ARGUMENT_ERROR, "value given twice - %S$", mrb_fixnum_value(n)); + mrb_raisef(mrb, E_ARGUMENT_ERROR, "value given twice - %i$", n); } nextvalue = GETPOSARG(n); p++; @@ -601,19 +667,21 @@ retry: case '{': { const char *start = p; char term = (*p == '<') ? '>' : '}'; - mrb_value symname; for (; p < end && *p != term; ) p++; if (id) { - mrb_raisef(mrb, E_ARGUMENT_ERROR, "name%S after <%S>", - mrb_str_new(mrb, start, p - start + 1), mrb_sym2str(mrb, id)); + mrb_raisef(mrb, E_ARGUMENT_ERROR, "name%l after <%n>", + start, p - start + 1, id); + } + CHECKNAMEARG(start, p - start + 1); + get_hash(mrb, &hash, argc, argv); + id = mrb_intern_check(mrb, start + 1, p - start - 1); + if (id) { + nextvalue = mrb_hash_fetch(mrb, hash, mrb_symbol_value(id), mrb_undef_value()); } - symname = mrb_str_new(mrb, start + 1, p - start - 1); - id = mrb_intern_str(mrb, symname); - nextvalue = GETNAMEARG(mrb_symbol_value(id), start, (int)(p - start + 1)); - if (mrb_undef_p(nextvalue)) { - mrb_raisef(mrb, E_KEY_ERROR, "key%S not found", mrb_str_new(mrb, start, p - start + 1)); + if (!id || mrb_undef_p(nextvalue)) { + mrb_raisef(mrb, E_KEY_ERROR, "key%l not found", start, p - start + 1); } if (term == '}') goto format_s; p++; @@ -624,6 +692,9 @@ retry: CHECK_FOR_WIDTH(flags); flags |= FWIDTH; GETASTER(width); + if (width > INT16_MAX || INT16_MIN > width) { + mrb_raise(mrb, E_ARGUMENT_ERROR, "width too big"); + } if (width < 0) { flags |= FMINUS; width = -width; @@ -637,7 +708,6 @@ retry: } flags |= FPREC|FPREC0; - prec = 0; p++; if (*p == '*') { GETASTER(prec); @@ -647,7 +717,6 @@ retry: p++; goto retry; } - GETNUM(prec, precision); goto retry; @@ -669,12 +738,29 @@ retry: tmp = mrb_check_string_type(mrb, val); if (!mrb_nil_p(tmp)) { - if (mrb_fixnum(mrb_funcall(mrb, tmp, "size", 0)) != 1 ) { + if (RSTRING_LEN(tmp) != 1) { mrb_raise(mrb, E_ARGUMENT_ERROR, "%c requires a character"); } } - else if (mrb_fixnum_p(val)) { - tmp = mrb_funcall(mrb, val, "chr", 0); + else if (mrb_integer_p(val)) { + mrb_int n = mrb_integer(val); +#ifndef MRB_UTF8_STRING + char buf[1]; + + buf[0] = (char)n&0xff; + tmp = mrb_str_new(mrb, buf, 1); +#else + if (n < 0x80) { + char buf[1]; + + buf[0] = (char)n; + tmp = mrb_str_new(mrb, buf, 1); + } + else { + tmp = mrb_funcall_id(mrb, val, MRB_SYM(chr), 0); + mrb_check_type(mrb, tmp, MRB_TT_STRING); + } +#endif } else { mrb_raise(mrb, E_ARGUMENT_ERROR, "invalid character"); @@ -682,21 +768,15 @@ retry: c = RSTRING_PTR(tmp); n = RSTRING_LEN(tmp); if (!(flags & FWIDTH)) { - CHECK(n); - memcpy(buf+blen, c, n); - blen += n; + PUSH(c, n); } else if ((flags & FMINUS)) { - CHECK(n); - memcpy(buf+blen, c, n); - blen += n; - FILL(' ', width-1); + PUSH(c, n); + if (width>0) FILL(' ', width-1); } else { - FILL(' ', width-1); - CHECK(n); - memcpy(buf+blen, c, n); - blen += n; + if (width>0) FILL(' ', width-1); + PUSH(c, n); } } break; @@ -716,7 +796,8 @@ retry: mrb_int tmp_n = len; RSTRING(result)->flags &= ~MRB_STR_EMBED_LEN_MASK; RSTRING(result)->flags |= tmp_n << MRB_STR_EMBED_LEN_SHIFT; - } else { + } + else { RSTRING(result)->as.heap.len = blen; } if (flags&(FPREC|FWIDTH)) { @@ -727,25 +808,17 @@ retry: if ((flags&FPREC) && (prec < slen)) { char *p = RSTRING_PTR(str) + prec; slen = prec; - len = p - RSTRING_PTR(str); + len = (mrb_int)(p - RSTRING_PTR(str)); } /* need to adjust multi-byte string pos */ if ((flags&FWIDTH) && (width > slen)) { width -= (int)slen; if (!(flags&FMINUS)) { - CHECK(width); - while (width--) { - buf[blen++] = ' '; - } + FILL(' ', width); } - CHECK(len); - memcpy(&buf[blen], RSTRING_PTR(str), len); - blen += len; + PUSH(RSTRING_PTR(str), len); if (flags&FMINUS) { - CHECK(width); - while (width--) { - buf[blen++] = ' '; - } + FILL(' ', width); } break; } @@ -763,29 +836,14 @@ retry: case 'B': case 'u': { mrb_value val = GETARG(); - char fbuf[32], nbuf[64], *s; + char nbuf[69], *s; const char *prefix = NULL; int sign = 0, dots = 0; char sc = 0; - mrb_int v = 0, org_v = 0; + mrb_int v = 0; int base; mrb_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; - default: - break; - } if (flags & FSHARP) { switch (*p) { case 'o': prefix = "0"; break; @@ -799,23 +857,20 @@ retry: bin_retry: switch (mrb_type(val)) { +#ifndef MRB_NO_FLOAT case MRB_TT_FLOAT: - if (FIXABLE(mrb_float(val))) { - val = mrb_fixnum_value((mrb_int)mrb_float(val)); - goto bin_retry; - } - val = mrb_flo_to_fixnum(mrb, val); - if (mrb_fixnum_p(val)) goto bin_retry; - break; + val = mrb_float_to_integer(mrb, val); + goto bin_retry; +#endif case MRB_TT_STRING: - val = mrb_str_to_inum(mrb, val, 0, TRUE); + val = mrb_str_to_integer(mrb, val, 0, TRUE); goto bin_retry; - case MRB_TT_FIXNUM: - v = mrb_fixnum(val); + case MRB_TT_INTEGER: + v = mrb_integer(val); break; default: - val = mrb_Integer(mrb, val); - goto bin_retry; + v = mrb_as_int(mrb, val); + break; } switch (*p) { @@ -830,67 +885,36 @@ retry: case 'u': case 'd': case 'i': + sign = 1; + /* fall through */ default: base = 10; break; } - if (base == 2) { - org_v = v; - if (v < 0 && !sign) { - val = mrb_fix2binstr(mrb, mrb_fixnum_value(v), base); - dots = 1; + if (sign) { + if (v >= 0) { + if (flags & FPLUS) { + sc = '+'; + width--; + } + else if (flags & FSPACE) { + sc = ' '; + width--; + } } else { - val = mrb_fixnum_to_str(mrb, mrb_fixnum_value(v), base); - } - v = mrb_fixnum(mrb_str_to_inum(mrb, val, 10, FALSE)); - } - if (sign) { - char c = *p; - if (c == 'i') c = 'd'; /* %d and %i are identical */ - if (base == 2) c = 'd'; - 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; + s = mrb_int_to_cstr(nbuf, sizeof(nbuf), v, base); + if (v < 0) s++; /* skip minus sign */ } else { - char c = *p; - if (c == 'X') c = 'x'; - if (base == 2) c = 'd'; - s = nbuf; + /* print as unsigned */ + s = mrb_uint_to_cstr(nbuf, sizeof(nbuf), v, base); if (v < 0) { dots = 1; } - snprintf(fbuf, sizeof(fbuf), "%%l%c", c); - snprintf(++s, sizeof(nbuf) - 1, fbuf, v); - if (v < 0) { - char d; - - s = remove_sign_bits(s, base); - switch (base) { - case 16: d = 'f'; break; - case 8: d = '7'; break; - case 2: d = '1'; break; - default: d = 0; break; - } - - if (d && *s != d) { - *--s = d; - } - } } { size_t size; @@ -899,11 +923,6 @@ retry: len = (mrb_int)size; } - if (dots) { - prec -= 2; - width -= 2; - } - if (*p == 'X') { char *pp = s; int c; @@ -949,11 +968,9 @@ retry: width -= prec; } - if (!(flags&FMINUS)) { - CHECK(width); - while (width-- > 0) { - buf[blen++] = ' '; - } + if (!(flags&FMINUS) && width > 0) { + FILL(' ', width); + width = 0; } if (sc) PUSH(&sc, 1); @@ -962,46 +979,44 @@ retry: int plen = (int)strlen(prefix); PUSH(prefix, plen); } - CHECK(prec - len); - if (dots) PUSH("..", 2); - - if (v < 0 || (base == 2 && org_v < 0)) { - char c = sign_bits(base, p); - while (len < prec--) { - buf[blen++] = c; - } + if (dots) { + prec -= 2; + width -= 2; + PUSH("..", 2); } - else if ((flags & (FMINUS|FPREC)) != FMINUS) { - char c = '0'; - while (len < prec--) { - buf[blen++] = c; + + if (prec > len) { + CHECK(prec - len); + if ((flags & (FMINUS|FPREC)) != FMINUS) { + char c = '0'; + FILL(c, prec - len); + } else if (v < 0) { + char c = sign_bits(base, p); + FILL(c, prec - len); } } - PUSH(s, len); - CHECK(width); - while (width-- > 0) { - buf[blen++] = ' '; + if (width > 0) { + FILL(' ', width); } } break; +#ifndef MRB_NO_FLOAT case 'f': case 'g': case 'G': case 'e': - case 'E': - case 'a': - case 'A': { + case 'E': { mrb_value val = GETARG(); double fval; - int i, need = 6; - char fbuf[32]; + mrb_int need = 6; - fval = mrb_float(mrb_Float(mrb, val)); + fval = mrb_as_float(mrb, val); if (!isfinite(fval)) { const char *expr; - const int elen = 3; + const mrb_int elen = 3; + char sign = '\0'; if (isnan(fval)) { expr = "NaN"; @@ -1010,53 +1025,61 @@ retry: expr = "Inf"; } need = elen; - if ((!isnan(fval) && fval < 0.0) || (flags & FPLUS)) - 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; - CHECK(need + 1); - snprintf(&buf[blen], need + 1, "%*s", need, ""); + if (need < 0) { + mrb_raise(mrb, E_ARGUMENT_ERROR, "width too big"); + } + FILL(' ', need); if (flags & FMINUS) { - if (!isnan(fval) && fval < 0.0) - buf[blen++] = '-'; - else if (flags & FPLUS) - buf[blen++] = '+'; - else if (flags & FSPACE) - blen++; - memcpy(&buf[blen], expr, elen); + if (sign) + buf[blen - need--] = sign; + memcpy(&buf[blen - need], expr, elen); } else { - if (!isnan(fval) && fval < 0.0) - buf[blen + need - elen - 1] = '-'; - else if (flags & FPLUS) - buf[blen + need - elen - 1] = '+'; - else if ((flags & FSPACE) && need > width) - blen++; - memcpy(&buf[blen + need - elen], expr, elen); + if (sign) + buf[blen - elen - 1] = sign; + memcpy(&buf[blen - elen], expr, elen); } - blen += strlen(&buf[blen]); break; } - fmt_setup(fbuf, sizeof(fbuf), *p, flags, width, prec); need = 0; if (*p != 'e' && *p != 'E') { - i = INT_MIN; + int i; frexp(fval, &i); if (i > 0) need = BIT_DIGITS(i); } + if (need > MRB_INT_MAX - ((flags&FPREC) ? prec : 6)) { + too_big_width_prec: + mrb_raise(mrb, E_ARGUMENT_ERROR, + (width > prec ? "width too big" : "prec too big")); + } need += (flags&FPREC) ? prec : 6; if ((flags&FWIDTH) && need < width) need = width; + if (need > MRB_INT_MAX - 20) { + goto too_big_width_prec; + } need += 20; CHECK(need); - n = snprintf(&buf[blen], need, fbuf, fval); + n = fmt_float(&buf[blen], need, *p, flags, width, prec, fval); + if (n < 0 || n >= need) { + mrb_raise(mrb, E_RUNTIME_ERROR, "formatting error"); + } blen += n; } break; +#endif } flags = FNONE; } @@ -1068,7 +1091,7 @@ retry: if (posarg >= 0 && nextarg < argc) { const char *mesg = "too many arguments for format string"; if (mrb_test(ruby_debug)) mrb_raise(mrb, E_ARGUMENT_ERROR, mesg); - if (mrb_test(ruby_verbose)) mrb_warn(mrb, "%S", mrb_str_new_cstr(mrb, mesg)); + if (mrb_test(ruby_verbose)) mrb_warn(mrb, "%s", mesg); } #endif mrb_str_resize(mrb, result, blen); @@ -1076,29 +1099,16 @@ retry: return result; } -static void -fmt_setup(char *buf, size_t size, int c, int flags, mrb_int width, mrb_int prec) -{ - char *end = buf + size; - int n; - - *buf++ = '%'; - if (flags & FSHARP) *buf++ = '#'; - if (flags & FPLUS) *buf++ = '+'; - if (flags & FMINUS) *buf++ = '-'; - if (flags & FZERO) *buf++ = '0'; - if (flags & FSPACE) *buf++ = ' '; - - if (flags & FWIDTH) { - n = snprintf(buf, end - buf, "%d", (int)width); - buf += n; - } - if (flags & FPREC) { - n = snprintf(buf, end - buf, ".%d", (int)prec); - buf += n; - } +void +mrb_mruby_sprintf_gem_init(mrb_state *mrb) +{ + struct RClass *krn = mrb->kernel_module; + mrb_define_module_function(mrb, krn, "sprintf", mrb_f_sprintf, MRB_ARGS_ANY()); + mrb_define_module_function(mrb, krn, "format", mrb_f_sprintf, MRB_ARGS_ANY()); +} - *buf++ = c; - *buf = '\0'; +void +mrb_mruby_sprintf_gem_final(mrb_state *mrb) +{ } |