ffd8cf8348
git-svn-id: svn://svn.openwrt.org/openwrt/packages@24870 3c298f89-4303-0410-b956-a3cf2f4a3e73
305 lines
6.2 KiB
C
305 lines
6.2 KiB
C
#include <iconv.h>
|
|
#include <errno.h>
|
|
#include <wchar.h>
|
|
#include <string.h>
|
|
#include <strings.h>
|
|
#include <stdlib.h>
|
|
#include <limits.h>
|
|
|
|
#include <dirent.h>
|
|
#include <fcntl.h>
|
|
#include <sys/mman.h>
|
|
#include <sys/stat.h>
|
|
#include <unistd.h>
|
|
#include <stdint.h>
|
|
|
|
/* builtin charmaps */
|
|
#include "charmaps.h"
|
|
|
|
/* only 0-7 are valid as dest charset */
|
|
#define UTF_16BE 000
|
|
#define UTF_16LE 001
|
|
#define UTF_32BE 002
|
|
#define UTF_32LE 003
|
|
#define WCHAR_T 004
|
|
#define UTF_8 005
|
|
#define US_ASCII 006
|
|
#define LATIN_1 007
|
|
|
|
/* additional charsets with algorithmic conversion */
|
|
#define LATIN_9 010
|
|
#define TIS_620 011
|
|
#define JIS_0201 012
|
|
|
|
/* some programs like php need this */
|
|
int _libiconv_version = _LIBICONV_VERSION;
|
|
|
|
/* these must match the constants above */
|
|
static const unsigned char charsets[] =
|
|
"\005" "UTF-8" "\0"
|
|
"\004" "WCHAR_T" "\0"
|
|
"\000" "UTF-16BE" "\0"
|
|
"\001" "UTF-16LE" "\0"
|
|
"\002" "UTF-32BE" "\0"
|
|
"\003" "UTF-32LE" "\0"
|
|
"\006" "US-ASCII" "\0"
|
|
"\007" "ISO-8859-1" "\0"
|
|
"\007" "LATIN1" "\0"
|
|
"\010" "ISO-8859-15""\0"
|
|
"\010" "LATIN9" "\0"
|
|
"\011" "ISO-8859-11""\0"
|
|
"\011" "TIS-620" "\0"
|
|
"\012" "JIS-0201" "\0"
|
|
"\377";
|
|
|
|
/* separate identifiers for sbcs/dbcs/etc map type */
|
|
#define UCS2_8BIT 000
|
|
#define UCS3_8BIT 001
|
|
#define EUC 002
|
|
#define EUC_TW 003
|
|
#define SHIFT_JIS 004
|
|
#define BIG5 005
|
|
#define GBK 006
|
|
|
|
/* FIXME: these are not implemented yet
|
|
// EUC: A1-FE A1-FE
|
|
// GBK: 81-FE 40-7E,80-FE
|
|
// Big5: A1-FE 40-7E,A1-FE
|
|
*/
|
|
|
|
static const unsigned short maplen[] = {
|
|
[UCS2_8BIT] = 4+ 2* 128,
|
|
[UCS3_8BIT] = 4+ 3* 128,
|
|
[EUC] = 4+ 2* 94*94,
|
|
[SHIFT_JIS] = 4+ 2* 94*94,
|
|
[BIG5] = 4+ 2* 94*157,
|
|
[GBK] = 4+ 2* 126*190,
|
|
[EUC_TW] = 4+ 2* 2*94*94,
|
|
};
|
|
|
|
static int find_charmap(const char *name)
|
|
{
|
|
int i;
|
|
for (i = 0; i < (sizeof(charmaps) / sizeof(charmaps[0])); i++)
|
|
if (!strcasecmp(charmaps[i].name, name))
|
|
return i;
|
|
return -1;
|
|
}
|
|
|
|
static int find_charset(const char *name)
|
|
{
|
|
const unsigned char *s;
|
|
for (s=charsets; *s<0xff && strcasecmp(s+1, name); s+=strlen(s)+1);
|
|
return *s;
|
|
}
|
|
|
|
iconv_t iconv_open(const char *to, const char *from)
|
|
{
|
|
unsigned f, t;
|
|
int m;
|
|
|
|
if ((t = find_charset(to)) >= 8)
|
|
return -1;
|
|
|
|
if ((f = find_charset(from)) < 255)
|
|
return 0 | (t<<1) | (f<<4);
|
|
|
|
if ((m = find_charmap(from)) > -1)
|
|
return 1 | (t<<1) | (m<<4);
|
|
|
|
return -1;
|
|
}
|
|
|
|
int iconv_close(iconv_t cd)
|
|
{
|
|
return 0;
|
|
}
|
|
|
|
static inline wchar_t get_16(const unsigned char *s, int endian)
|
|
{
|
|
endian &= 1;
|
|
return s[endian]<<8 | s[endian^1];
|
|
}
|
|
|
|
static inline void put_16(unsigned char *s, wchar_t c, int endian)
|
|
{
|
|
endian &= 1;
|
|
s[endian] = c>>8;
|
|
s[endian^1] = c;
|
|
}
|
|
|
|
size_t iconv(iconv_t cd, char **in, size_t *inb, char **out, size_t *outb)
|
|
{
|
|
size_t x=0;
|
|
unsigned char to = (cd>>1)&7;
|
|
unsigned char from = 255;
|
|
const unsigned char *map = 0;
|
|
mbstate_t st = {0};
|
|
char tmp[MB_LEN_MAX];
|
|
wchar_t c, d;
|
|
size_t k, l;
|
|
int err;
|
|
|
|
if (!in || !*in || !*inb) return 0;
|
|
|
|
if (cd & 1)
|
|
map = charmaps[cd>>4].map;
|
|
else
|
|
from = cd>>4;
|
|
|
|
for (; *inb; *in+=l, *inb-=l) {
|
|
c = *(unsigned char *)*in;
|
|
l = 1;
|
|
if (from >= UTF_8 && c < 0x80) goto charok;
|
|
switch (from) {
|
|
case WCHAR_T:
|
|
l = sizeof(wchar_t);
|
|
if (*inb < l) goto starved;
|
|
c = *(wchar_t *)*in;
|
|
break;
|
|
case UTF_8:
|
|
l = mbrtowc(&c, *in, *inb, &st);
|
|
if (!l) l++;
|
|
else if (l == (size_t)-1) goto ilseq;
|
|
else if (l == (size_t)-2) goto starved;
|
|
break;
|
|
case US_ASCII:
|
|
goto ilseq;
|
|
case LATIN_9:
|
|
if ((unsigned)c - 0xa4 <= 0xbe - 0xa4) {
|
|
static const unsigned char map[] = {
|
|
0, 0x60, 0, 0x61, 0, 0, 0, 0, 0, 0, 0,
|
|
0, 0, 0, 0, 0x7d, 0, 0, 0, 0x7e, 0, 0, 0,
|
|
0x52, 0x53, 0x78
|
|
};
|
|
if (c == 0xa4) c = 0x20ac;
|
|
else if (map[c-0xa5]) c = 0x100 | map[c-0xa5];
|
|
}
|
|
case LATIN_1:
|
|
goto charok;
|
|
case TIS_620:
|
|
if (c >= 0xa1) c += 0x0e01-0xa1;
|
|
goto charok;
|
|
case JIS_0201:
|
|
if (c >= 0xa1) {
|
|
if (c <= 0xdf) c += 0xff61-0xa1;
|
|
else goto ilseq;
|
|
}
|
|
goto charok;
|
|
case UTF_16BE:
|
|
case UTF_16LE:
|
|
l = 2;
|
|
if (*inb < 2) goto starved;
|
|
c = get_16(*in, from);
|
|
if ((unsigned)(c-0xdc00) < 0x400) goto ilseq;
|
|
if ((unsigned)(c-0xd800) < 0x400) {
|
|
l = 4;
|
|
if (*inb < 4) goto starved;
|
|
d = get_16(*in + 2, from);
|
|
if ((unsigned)(c-0xdc00) >= 0x400) goto ilseq;
|
|
c = ((c-0xd800)<<10) | (d-0xdc00);
|
|
}
|
|
break;
|
|
case UTF_32BE:
|
|
case UTF_32LE:
|
|
l = 4;
|
|
if (*inb < 4) goto starved;
|
|
// FIXME
|
|
// c = get_32(*in, from);
|
|
break;
|
|
default:
|
|
/* only support ascii supersets */
|
|
if (c < 0x80) break;
|
|
switch (map[0]) {
|
|
case UCS2_8BIT:
|
|
c -= 0x80;
|
|
break;
|
|
case EUC:
|
|
if ((unsigned)c - 0xa1 >= 94) goto ilseq;
|
|
if ((unsigned)in[0][1] - 0xa1 >= 94) goto ilseq;
|
|
c = (c-0xa1)*94 + (in[0][1]-0xa1);
|
|
l = 2;
|
|
break;
|
|
case SHIFT_JIS:
|
|
if ((unsigned)c - 0xa1 <= 0xdf-0xa1) {
|
|
c += 0xff61-0xa1;
|
|
goto charok;
|
|
}
|
|
// FIXME...
|
|
l = 2;
|
|
break;
|
|
default:
|
|
goto badf;
|
|
}
|
|
c = get_16(map + 4 + 2*c, 0);
|
|
if (c == 0xffff) goto ilseq;
|
|
goto charok;
|
|
}
|
|
|
|
if ((unsigned)c - 0xd800 < 0x800 || (unsigned)c >= 0x110000)
|
|
goto ilseq;
|
|
charok:
|
|
switch (to) {
|
|
case WCHAR_T:
|
|
if (*outb < sizeof(wchar_t)) goto toobig;
|
|
*(wchar_t *)*out = c;
|
|
*out += sizeof(wchar_t);
|
|
*outb -= sizeof(wchar_t);
|
|
break;
|
|
case UTF_8:
|
|
if (*outb < 4) {
|
|
k = wctomb(tmp, c);
|
|
if (*outb < k) goto toobig;
|
|
memcpy(*out, tmp, k);
|
|
} else k = wctomb(*out, c);
|
|
*out += k;
|
|
*outb -= k;
|
|
break;
|
|
case US_ASCII:
|
|
if (c > 0x7f) c = 0xfffd;
|
|
/* fall thru and count replacement in latin1 case */
|
|
case LATIN_1:
|
|
if (!*outb) goto toobig;
|
|
if (c < 0x100) **out = c;
|
|
else x++, **out = '*'; //FIXME: translit?
|
|
++*out;
|
|
--*outb;
|
|
break;
|
|
case UTF_16BE:
|
|
case UTF_16LE:
|
|
if (c < 0x10000) {
|
|
if (*outb < 2) goto toobig;
|
|
put_16(*out, c, to);
|
|
*out += 2;
|
|
*outb -= 2;
|
|
break;
|
|
}
|
|
if (*outb < 4) goto toobig;
|
|
put_16(*out, (c>>10)|0xd800, to);
|
|
put_16(*out + 2, (c&0x3ff)|0xdc00, to);
|
|
*out += 4;
|
|
*outb -= 4;
|
|
break;
|
|
default:
|
|
goto badf;
|
|
}
|
|
}
|
|
return x;
|
|
ilseq:
|
|
err = EILSEQ;
|
|
x = -1;
|
|
goto end;
|
|
badf:
|
|
err = EBADF;
|
|
x = -1;
|
|
goto end;
|
|
toobig:
|
|
err = E2BIG;
|
|
goto end;
|
|
starved:
|
|
err = EINVAL;
|
|
end:
|
|
errno = err;
|
|
return x;
|
|
}
|