/**********************************************************************
string.c -
$Author$
created at: Mon Aug 9 17:12:58 JST 1993
Copyright (C) 1993-2007 Yukihiro Matsumoto
Copyright (C) 2000 Network Applied Communication Laboratory, Inc.
Copyright (C) 2000 Information-technology Promotion Agency, Japan
**********************************************************************/
#include "ruby/ruby.h"
#include "ruby/re.h"
#include "ruby/encoding.h"
#include <assert.h>
#define BEG(no) regs->beg[no]
#define END(no) regs->end[no]
#include <math.h>
#include <ctype.h>
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif
#define numberof(array) (int)(sizeof(array) / sizeof((array)[0]))
#undef rb_str_new_cstr
#undef rb_tainted_str_new_cstr
#undef rb_usascii_str_new_cstr
#undef rb_external_str_new_cstr
#undef rb_locale_str_new_cstr
#undef rb_str_new2
#undef rb_str_new3
#undef rb_str_new4
#undef rb_str_new5
#undef rb_tainted_str_new2
#undef rb_usascii_str_new2
#undef rb_str_dup_frozen
#undef rb_str_buf_new_cstr
#undef rb_str_buf_new2
#undef rb_str_buf_cat2
#undef rb_str_cat2
VALUE rb_cString;
VALUE rb_cSymbol;
#define RUBY_MAX_CHAR_LEN 16
#define STR_TMPLOCK FL_USER7
#define STR_NOEMBED FL_USER1
#define STR_SHARED FL_USER2 /* = ELTS_SHARED */
#define STR_ASSOC FL_USER3
#define STR_SHARED_P(s) FL_ALL(s, STR_NOEMBED|ELTS_SHARED)
#define STR_ASSOC_P(s) FL_ALL(s, STR_NOEMBED|STR_ASSOC)
#define STR_NOCAPA (STR_NOEMBED|ELTS_SHARED|STR_ASSOC)
#define STR_NOCAPA_P(s) (FL_TEST(s,STR_NOEMBED) && FL_ANY(s,ELTS_SHARED|STR_ASSOC))
#define STR_UNSET_NOCAPA(s) do {\
if (FL_TEST(s,STR_NOEMBED)) FL_UNSET(s,(ELTS_SHARED|STR_ASSOC));\
} while (0)
#define STR_SET_NOEMBED(str) do {\
FL_SET(str, STR_NOEMBED);\
STR_SET_EMBED_LEN(str, 0);\
} while (0)
#define STR_SET_EMBED(str) FL_UNSET(str, STR_NOEMBED)
#define STR_EMBED_P(str) (!FL_TEST(str, STR_NOEMBED))
#define STR_SET_EMBED_LEN(str, n) do { \
long tmp_n = (n);\
RBASIC(str)->flags &= ~RSTRING_EMBED_LEN_MASK;\
RBASIC(str)->flags |= (tmp_n) << RSTRING_EMBED_LEN_SHIFT;\
} while (0)
#define STR_SET_LEN(str, n) do { \
if (STR_EMBED_P(str)) {\
STR_SET_EMBED_LEN(str, n);\
}\
else {\
RSTRING(str)->as.heap.len = (n);\
}\
} while (0)
#define STR_DEC_LEN(str) do {\
if (STR_EMBED_P(str)) {\
long n = RSTRING_LEN(str);\
n--;\
STR_SET_EMBED_LEN(str, n);\
}\
else {\
RSTRING(str)->as.heap.len--;\
}\
} while (0)
#define RESIZE_CAPA(str,capacity) do {\
if (STR_EMBED_P(str)) {\
if ((capacity) > RSTRING_EMBED_LEN_MAX) {\
char *tmp = ALLOC_N(char, capacity+1);\
memcpy(tmp, RSTRING_PTR(str), RSTRING_LEN(str));\
RSTRING(str)->as.heap.ptr = tmp;\
RSTRING(str)->as.heap.len = RSTRING_LEN(str);\
STR_SET_NOEMBED(str);\
RSTRING(str)->as.heap.aux.capa = (capacity);\
}\
}\
else {\
REALLOC_N(RSTRING(str)->as.heap.ptr, char, (capacity)+1);\
if (!STR_NOCAPA_P(str))\
RSTRING(str)->as.heap.aux.capa = (capacity);\
}\
} while (0)
#define is_ascii_string(str) (rb_enc_str_coderange(str) == ENC_CODERANGE_7BIT)
#define is_broken_string(str) (rb_enc_str_coderange(str) == ENC_CODERANGE_BROKEN)
#define STR_ENC_GET(str) rb_enc_from_index(ENCODING_GET(str))
static inline int
single_byte_optimizable(VALUE str)
{
rb_encoding *enc;
/* Conservative. It may be ENC_CODERANGE_UNKNOWN. */
if (ENC_CODERANGE(str) == ENC_CODERANGE_7BIT)
return 1;
enc = STR_ENC_GET(str);
if (rb_enc_mbmaxlen(enc) == 1)
return 1;
/* Conservative. Possibly single byte.
* "\xa1" in Shift_JIS for example. */
return 0;
}
VALUE rb_fs;
static inline const char *
search_nonascii(const char *p, const char *e)
{
#if SIZEOF_VALUE == 8
# define NONASCII_MASK 0x8080808080808080LL
#elif SIZEOF_VALUE == 4
# define NONASCII_MASK 0x80808080UL
#endif
#ifdef NONASCII_MASK
if ((int)sizeof(VALUE) * 2 < e - p) {
const VALUE *s, *t;
const VALUE lowbits = sizeof(VALUE) - 1;
s = (const VALUE*)(~lowbits & ((VALUE)p + lowbits));
while (p < (const char *)s) {
if (!ISASCII(*p))
return p;
p++;
}
t = (const VALUE*)(~lowbits & (VALUE)e);
while (s < t) {
if (*s & NONASCII_MASK) {
t = s;
break;
}
s++;
}
p = (const char *)t;
}
#endif
while (p < e) {
if (!ISASCII(*p))
return p;
p++;
}
return NULL;
}
static int
coderange_scan(const char *p, long len, rb_encoding *enc)
{
const char *e = p + len;
if (rb_enc_to_index(enc) == 0) {
/* enc is ASCII-8BIT. ASCII-8BIT string never be broken. */
p = search_nonascii(p, e);
return p ? ENC_CODERANGE_VALID : ENC_CODERANGE_7BIT;
}
if (rb_enc_asciicompat(enc)) {
p = search_nonascii(p, e);
if (!p) {
return ENC_CODERANGE_7BIT;
}
while (p < e) {
int ret = rb_enc_precise_mbclen(p, e, enc);
if (!MBCLEN_CHARFOUND_P(ret)) {
return ENC_CODERANGE_BROKEN;
}
p += MBCLEN_CHARFOUND_LEN(ret);
if (p < e) {
p = search_nonascii(p, e);
if (!p) {
return ENC_CODERANGE_VALID;
}
}
}
if (e < p) {
return ENC_CODERANGE_BROKEN;
}
return ENC_CODERANGE_VALID;
}
while (p < e) {
int ret = rb_enc_precise_mbclen(p, e, enc);
if (!MBCLEN_CHARFOUND_P(ret)) {
return ENC_CODERANGE_BROKEN;
}
p += MBCLEN_CHARFOUND_LEN(ret);
}
if (e < p) {
return ENC_CODERANGE_BROKEN;
}
return ENC_CODERANGE_VALID;
}
long
rb_str_coderange_scan_restartable(const char *s, const char *e, rb_encoding *enc, int *cr)
{
const char *p = s;
if (*cr == ENC_CODERANGE_BROKEN)
return e - s;
if (rb_enc_to_index(enc) == 0) {
/* enc is ASCII-8BIT. ASCII-8BIT string never be broken. */
p = search_nonascii(p, e);
*cr = (!p && *cr != ENC_CODERANGE_VALID) ? ENC_CODERANGE_7BIT : ENC_CODERANGE_VALID;
return e - s;
}
else if (rb_enc_asciicompat(enc)) {
p = search_nonascii(p, e);
if (!p) {
if (*cr != ENC_CODERANGE_VALID) *cr = ENC_CODERANGE_7BIT;
return e - s;
}
while (p < e) {
int ret = rb_enc_precise_mbclen(p, e, enc);
if (!MBCLEN_CHARFOUND_P(ret)) {
*cr = MBCLEN_INVALID_P(ret) ? ENC_CODERANGE_BROKEN: ENC_CODERANGE_UNKNOWN;
return p - s;
}
p += MBCLEN_CHARFOUND_LEN(ret);
if (p < e) {
p = search_nonascii(p, e);
if (!p) {
*cr = ENC_CODERANGE_VALID;
return e - s;
}
}
}
*cr = e < p ? ENC_CODERANGE_BROKEN: ENC_CODERANGE_VALID;
return p - s;
}
else {
while (p < e) {
int ret = rb_enc_precise_mbclen(p, e, enc);
if (!MBCLEN_CHARFOUND_P(ret)) {
*cr = MBCLEN_INVALID_P(ret) ? ENC_CODERANGE_BROKEN: ENC_CODERANGE_UNKNOWN;
return p - s;
}
p += MBCLEN_CHARFOUND_LEN(ret);
}
*cr = e < p ? ENC_CODERANGE_BROKEN: ENC_CODERANGE_VALID;
return p - s;
}
}
static inline void
str_enc_copy(VALUE str1, VALUE str2)
{
rb_enc_set_index(str1, ENCODING_GET(str2));
}
static void
rb_enc_cr_str_copy_for_substr(VALUE dest, VALUE src)
{
/* this function is designed for copying encoding and coderange
* from src to new string "dest" which is made from the part of src.
*/
str_enc_copy(dest, src);
switch (ENC_CODERANGE(src)) {
case ENC_CODERANGE_7BIT:
ENC_CODERANGE_SET(dest, ENC_CODERANGE_7BIT);
break;
case ENC_CODERANGE_VALID:
if (!rb_enc_asciicompat(STR_ENC_GET(src)) ||
search_nonascii(RSTRING_PTR(dest), RSTRING_END(dest)))
ENC_CODERANGE_SET(dest, ENC_CODERANGE_VALID);
else
ENC_CODERANGE_SET(dest, ENC_CODERANGE_7BIT);
break;
default:
if (RSTRING_LEN(dest) == 0) {
if (!rb_enc_asciicompat(STR_ENC_GET(src)))
ENC_CODERANGE_SET(dest, ENC_CODERANGE_VALID);
else
ENC_CODERANGE_SET(dest, ENC_CODERANGE_7BIT);
}
break;
}
}
static void
rb_enc_cr_str_exact_copy(VALUE dest, VALUE src)
{
str_enc_copy(dest, src);
ENC_CODERANGE_SET(dest, ENC_CODERANGE(src));
}
int
rb_enc_str_coderange(VALUE str)
{
int cr = ENC_CODERANGE(str);
if (cr == ENC_CODERANGE_UNKNOWN) {
rb_encoding *enc = STR_ENC_GET(str);
cr = coderange_scan(RSTRING_PTR(str), RSTRING_LEN(str), enc);
ENC_CODERANGE_SET(str, cr);
}
return cr;
}
int
rb_enc_str_asciionly_p(VALUE str)
{
rb_encoding *enc = STR_ENC_GET(str);
if (!rb_enc_asciicompat(enc))
return FALSE;
else if (rb_enc_str_coderange(str) == ENC_CODERANGE_7BIT)
return TRUE;
return FALSE;
}
static inline void
str_mod_check(VALUE s, const char *p, long len)
{
if (RSTRING_PTR(s) != p || RSTRING_LEN(s) != len){
rb_raise(rb_eRuntimeError, "string modified");
}
}
static inline void
str_frozen_check(VALUE s)
{
if (OBJ_FROZEN(s)) {
rb_raise(rb_eRuntimeError, "string frozen");
}
}
size_t
rb_str_capacity(VALUE str)
{
if (STR_EMBED_P(str)) {
return RSTRING_EMBED_LEN_MAX;
}
else if (STR_NOCAPA_P(str)) {
return RSTRING(str)->as.heap.len;
}
else {
return RSTRING(str)->as.heap.aux.capa;
}
}
static inline VALUE
str_alloc(VALUE klass)
{
NEWOBJ(str, struct RString);
OBJSETUP(str, klass, T_STRING);
str->as.heap.ptr = 0;
str->as.heap.len = 0;
str->as.heap.aux.capa = 0;
return (VALUE)str;
}
static VALUE
str_new(VALUE klass, const char *ptr, long len)
{
VALUE str;
if (len < 0) {
rb_raise(rb_eArgError, "negative string size (or size too big)");
}
str = str_alloc(klass);
if (len > RSTRING_EMBED_LEN_MAX) {
RSTRING(str)->as.heap.aux.capa = len;
RSTRING(str)->as.heap.ptr = ALLOC_N(char,len+1);
STR_SET_NOEMBED(str);
}
else if (len == 0) {
ENC_CODERANGE_SET(str, ENC_CODERANGE_7BIT);
}
if (ptr) {
memcpy(RSTRING_PTR(str), ptr, len);
}
STR_SET_LEN(str, len);
RSTRING_PTR(str)[len] = '\0';
return str;
}
VALUE
rb_str_new(const char *ptr, long len)
{
return str_new(rb_cString, ptr, len);
}
VALUE
rb_usascii_str_new(const char *ptr, long len)
{
VALUE str = rb_str_new(ptr, len);
ENCODING_CODERANGE_SET(str, rb_usascii_encindex(), ENC_CODERANGE_7BIT);
return str;
}
VALUE
rb_enc_str_new(const char *ptr, long len, rb_encoding *enc)
{
VALUE str = rb_str_new(ptr, len);
rb_enc_associate(str, enc);
return str;
}
VALUE
rb_str_new_cstr(const char *ptr)
{
if (!ptr) {
rb_raise(rb_eArgError, "NULL pointer given");
}
return rb_str_new(ptr, strlen(ptr));
}
RUBY_ALIAS_FUNCTION(rb_str_new2(const char *ptr), rb_str_new_cstr, (ptr))
#define rb_str_new2 rb_str_new_cstr
VALUE
rb_usascii_str_new_cstr(const char *ptr)
{
VALUE str = rb_str_new2(ptr);
ENCODING_CODERANGE_SET(str, rb_usascii_encindex(), ENC_CODERANGE_7BIT);
return str;
}
RUBY_ALIAS_FUNCTION(rb_usascii_str_new2(const char *ptr), rb_usascii_str_new_cstr, (ptr))
#define rb_usascii_str_new2 rb_usascii_str_new_cstr
VALUE
rb_tainted_str_new(const char *ptr, long len)
{
VALUE str = rb_str_new(ptr, len);
OBJ_TAINT(str);
return str;
}
VALUE
rb_tainted_str_new_cstr(const char *ptr)
{
VALUE str = rb_str_new2(ptr);
OBJ_TAINT(str);
return str;
}
RUBY_ALIAS_FUNCTION(rb_tainted_str_new2(const char *ptr), rb_tainted_str_new_cstr, (ptr))
#define rb_tainted_str_new2 rb_tainted_str_new_cstr
VALUE
rb_str_conv_enc_opts(VALUE str, rb_encoding *from, rb_encoding *to, int ecflags, VALUE ecopts)
{
rb_econv_t *ec;
rb_econv_result_t ret;
long len;
VALUE newstr;
const unsigned char *sp;
unsigned char *dp;
if (!to) return str;
if (from == to) return str;
if ((rb_enc_asciicompat(to) && ENC_CODERANGE(str) == ENC_CODERANGE_7BIT) ||
to == rb_ascii8bit_encoding()) {
if (STR_ENC_GET(str) != to) {
str = rb_str_dup(str);
rb_enc_associate(str, to);
}
return str;
}
len = RSTRING_LEN(str);
newstr = rb_str_new(0, len);
retry:
ec = rb_econv_open_opts(from->name, to->name, ecflags, ecopts);
if (!ec) return str;
sp = (unsigned char*)RSTRING_PTR(str);
dp = (unsigned char*)RSTRING_PTR(newstr);
ret = rb_econv_convert(ec, &sp, (unsigned char*)RSTRING_END(str),
&dp, (unsigned char*)RSTRING_END(newstr), 0);
rb_econv_close(ec);
switch (ret) {
case econv_destination_buffer_full:
/* destination buffer short */
len *= 2;
rb_str_resize(newstr, len);
goto retry;
case econv_finished:
len = dp - (unsigned char*)RSTRING_PTR(newstr);
rb_str_set_len(newstr, len);
rb_enc_associate(newstr, to);
return newstr;
default:
/* some error, return original */
return str;
}
}
VALUE
rb_str_conv_enc(VALUE str, rb_encoding *from, rb_encoding *to)
{
return rb_str_conv_enc_opts(str, from, to, 0, Qnil);
}
VALUE
rb_external_str_new_with_enc(const char *ptr, long len, rb_encoding *eenc)
{
VALUE str;
str = rb_tainted_str_new(ptr, len);
if (eenc == rb_usascii_encoding() &&
rb_enc_str_coderange(str) != ENC_CODERANGE_7BIT) {
rb_enc_associate(str, rb_ascii8bit_encoding());
return str;
}
rb_enc_associate(str, eenc);
return rb_str_conv_enc(str, eenc, rb_default_internal_encoding());
}
VALUE
rb_external_str_new(const char *ptr, long len)
{
return rb_external_str_new_with_enc(ptr, len, rb_default_external_encoding());
}
VALUE
rb_external_str_new_cstr(const char *ptr)
{
return rb_external_str_new_with_enc(ptr, strlen(ptr), rb_default_external_encoding());
}
VALUE
rb_locale_str_new(const char *ptr, long len)
{
return rb_external_str_new_with_enc(ptr, len, rb_locale_encoding());
}
VALUE
rb_locale_str_new_cstr(const char *ptr)
{
return rb_external_str_new_with_enc(ptr, strlen(ptr), rb_locale_encoding());
}
VALUE
rb_filesystem_str_new(const char *ptr, long len)
{
return rb_external_str_new_with_enc(ptr, len, rb_filesystem_encoding());
}
VALUE
rb_filesystem_str_new_cstr(const char *ptr)
{
return rb_external_str_new_with_enc(ptr, strlen(ptr), rb_filesystem_encoding());
}
VALUE
rb_str_export(VALUE str)
{
return rb_str_conv_enc(str, STR_ENC_GET(str), rb_default_external_encoding());
}
VALUE
rb_str_export_locale(VALUE str)
{
return rb_str_conv_enc(str, STR_ENC_GET(str), rb_locale_encoding());
}
VALUE
rb_str_export_to_enc(VALUE str, rb_encoding *enc)
{
return rb_str_conv_enc(str, STR_ENC_GET(str), enc);
}
static VALUE
str_replace_shared(VALUE str2, VALUE str)
{
if (RSTRING_LEN(str) <= RSTRING_EMBED_LEN_MAX) {
STR_SET_EMBED(str2);
memcpy(RSTRING_PTR(str2), RSTRING_PTR(str), RSTRING_LEN(str)+1);
STR_SET_EMBED_LEN(str2, RSTRING_LEN(str));
}
else {
FL_SET(str2, STR_NOEMBED);
str = rb_str_new_frozen(str);
RSTRING(str2)->as.heap.len = RSTRING_LEN(str);
RSTRING(str2)->as.heap.ptr = RSTRING_PTR(str);
RSTRING(str2)->as.heap.aux.shared = str;
FL_SET(str2, ELTS_SHARED);
}
rb_enc_cr_str_exact_copy(str2, str);
return str2;
}
static VALUE
str_new_shared(VALUE klass, VALUE str)
{
return str_replace_shared(str_alloc(klass), str);
}
static VALUE
str_new3(VALUE klass, VALUE str)
{
return str_new_shared(klass, str);
}
VALUE
rb_str_new_shared(VALUE str)
{
VALUE str2 = str_new3(rb_obj_class(str), str);
OBJ_INFECT(str2, str);
return str2;
}
RUBY_ALIAS_FUNCTION(rb_str_new3(VALUE str), rb_str_new_shared, (str))
#define rb_str_new3 rb_str_new_shared
static VALUE
str_new4(VALUE klass, VALUE str)
{
VALUE str2;
str2 = str_alloc(klass);
STR_SET_NOEMBED(str2);
RSTRING(str2)->as.heap.len = RSTRING_LEN(str);
RSTRING(str2)->as.heap.ptr = RSTRING_PTR(str);
if (STR_SHARED_P(str)) {
VALUE shared = RSTRING(str)->as.heap.aux.shared;
assert(OBJ_FROZEN(shared));
FL_SET(str2, ELTS_SHARED);
RSTRING(str2)->as.heap.aux.shared = shared;
}
else {
FL_SET(str, ELTS_SHARED);
RSTRING(str)->as.heap.aux.shared = str2;
}
rb_enc_cr_str_exact_copy(str2, str);
OBJ_INFECT(str2, str);
return str2;
}
VALUE
rb_str_new_frozen(VALUE orig)
{
VALUE klass, str;
if (OBJ_FROZEN(orig)) return orig;
klass = rb_obj_class(orig);
if (STR_SHARED_P(orig) && (str = RSTRING(orig)->as.heap.aux.shared)) {
long ofs;
assert(OBJ_FROZEN(str));
ofs = RSTRING_LEN(str) - RSTRING_LEN(orig);
if ((ofs > 0) || (klass != RBASIC(str)->klass) ||
(!OBJ_TAINTED(str) && OBJ_TAINTED(orig)) ||
ENCODING_GET(str) != ENCODING_GET(orig)) {
str = str_new3(klass, str);
RSTRING(str)->as.heap.ptr += ofs;
RSTRING(str)->as.heap.len -= ofs;
rb_enc_cr_str_exact_copy(str, orig);
OBJ_INFECT(str, orig);
}
}
else if (STR_EMBED_P(orig)) {
str = str_new(klass, RSTRING_PTR(orig), RSTRING_LEN(orig));
rb_enc_cr_str_exact_copy(str, orig);
OBJ_INFECT(str, orig);
}
else if (STR_ASSOC_P(orig)) {
VALUE assoc = RSTRING(orig)->as.heap.aux.shared;
FL_UNSET(orig, STR_ASSOC);
str = str_new4(klass, orig);
FL_SET(str, STR_ASSOC);
RSTRING(str)->as.heap.aux.shared = assoc;
}
else {
str = str_new4(klass, orig);
}
OBJ_FREEZE(str);
return str;
}
RUBY_ALIAS_FUNCTION(rb_str_new4(VALUE orig), rb_str_new_frozen, (orig))
#define rb_str_new4 rb_str_new_frozen
VALUE
rb_str_new_with_class(VALUE obj, const char *ptr, long len)
{
return str_new(rb_obj_class(obj), ptr, len);
}
RUBY_ALIAS_FUNCTION(rb_str_new5(VALUE obj, const char *ptr, long len),
rb_str_new_with_class, (obj, ptr, len))
#define rb_str_new5 rb_str_new_with_class
#define STR_BUF_MIN_SIZE 128
VALUE
rb_str_buf_new(long capa)
{
VALUE str = str_alloc(rb_cString);
if (capa < STR_BUF_MIN_SIZE) {
capa = STR_BUF_MIN_SIZE;
}
FL_SET(str, STR_NOEMBED);
RSTRING(str)->as.heap.aux.capa = capa;
RSTRING(str)->as.heap.ptr = ALLOC_N(char, capa+1);
RSTRING(str)->as.heap.ptr[0] = '\0';
return str;
}
VALUE
rb_str_buf_new_cstr(const char *ptr)
{
VALUE str;
long len = strlen(ptr);
str = rb_str_buf_new(len);
rb_str_buf_cat(str, ptr, len);
return str;
}
RUBY_ALIAS_FUNCTION(rb_str_buf_new2(const char *ptr), rb_str_buf_new_cstr, (ptr))
#define rb_str_buf_new2 rb_str_buf_new_cstr
VALUE
rb_str_tmp_new(long len)
{
return str_new(0, 0, len);
}
void
rb_str_free(VALUE str)
{
if (!STR_EMBED_P(str) && !STR_SHARED_P(str)) {
xfree(RSTRING(str)->as.heap.ptr);
}
}
size_t
rb_str_memsize(VALUE str)
{
if (!STR_EMBED_P(str) && !STR_SHARED_P(str)) {
return RSTRING(str)->as.heap.aux.capa;
}
else {
return 0;
}
}
VALUE
rb_str_to_str(VALUE str)
{
return rb_convert_type(str, T_STRING, "String", "to_str");
}
static inline void str_discard(VALUE str);
void
rb_str_shared_replace(VALUE str, VALUE str2)
{
rb_encoding *enc;
int cr;
| 