#include <stdlib.h>
#include <limits.h>
#include <stdint.h>
+#include "locale_impl.h"
#define UTF_32BE 0300
#define UTF_16LE 0301
#define UTF_32LE 0303
#define UCS2BE 0304
#define UCS2LE 0305
-#define US_ASCII 0306
-#define WCHAR_T 0307
+#define WCHAR_T 0306
+#define US_ASCII 0307
#define UTF_8 0310
#define EUC_JP 0320
#define SHIFT_JIS 0321
#define GB18030 0330
#define GBK 0331
#define GB2312 0332
-
-/* 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
- */
+#define BIG5 0340
+#define EUC_KR 0350
/* Definitions of charmaps. Each charmap consists of:
* 1. Empty-string-terminated list of null-terminated aliases.
* 3. Character table (size determined by field 2). */
static const unsigned char charmaps[] =
-"utf8\0\0\310"
-"wchart\0\0\307"
+"utf8\0char\0\0\310"
+"wchart\0\0\306"
"ucs2\0ucs2be\0\0\304"
"ucs2le\0\0\305"
"utf16\0utf16be\0\0\302"
"utf16le\0\0\301"
"ucs4\0ucs4be\0utf32\0utf32be\0\0\300"
"ucs4le\0utf32le\0\0\303"
-"ascii\0usascii\0iso646\0iso646us\0\0\306"
+"ascii\0usascii\0iso646\0iso646us\0\0\307"
"eucjp\0\0\320"
"shiftjis\0sjis\0\0\321"
"gb18030\0\0\330"
"gbk\0\0\331"
"gb2312\0\0\332"
+"big5\0bigfive\0cp950\0big5hkscs\0\0\340"
+"euckr\0ksc5601\0ksx1001\0cp949\0\0\350"
#include "codepages.h"
;
#include "gb18030.h"
};
+static const unsigned short big5[89][157] = {
+#include "big5.h"
+};
+
+static const unsigned short hkscs[] = {
+#include "hkscs.h"
+};
+
+static const unsigned short ksc[93][94] = {
+#include "ksc.h"
+};
+
static int fuzzycmp(const unsigned char *a, const unsigned char *b)
{
for (; *a && *b; a++, b++) {
static size_t find_charmap(const void *name)
{
const unsigned char *s;
+ if (!*(char *)name) name=charmaps; /* "utf8" */
for (s=charmaps; *s; ) {
if (!fuzzycmp(name, s)) {
for (; *s; s+=strlen((void *)s)+1);
#define mbrtowc_utf8 mbrtowc
#define wctomb_utf8 wctomb
-size_t iconv(iconv_t cd0, char **in, size_t *inb, char **out, size_t *outb)
+size_t iconv(iconv_t cd0, char **restrict in, size_t *restrict inb, char **restrict out, size_t *restrict outb)
{
size_t x=0;
unsigned long cd = (unsigned long)cd0;
int err;
unsigned char type = map[-1];
unsigned char totype = tomap[-1];
+ locale_t *ploc = &CURRENT_LOCALE, loc = *ploc;
if (!in || !*in || !*inb) return 0;
+ *ploc = UTF8_LOCALE;
+
for (; *inb; *in+=l, *inb-=l) {
c = *(unsigned char *)*in;
l = 1;
- if (c >= 128) switch (type) {
+ if (c >= 128 || type-UTF_32BE < 7U) switch (type) {
case UTF_8:
l = mbrtowc_utf8(&wc, *in, *inb, &st);
if (!l) l++;
if (type-UCS2BE < 2U) goto ilseq;
l = 4;
if (*inb < 4) goto starved;
- d = get_16((void *)(*in + 2), from);
- if ((unsigned)(c-0xdc00) >= 0x400) goto ilseq;
- c = ((c-0xd800)<<10) | (d-0xdc00);
+ d = get_16((void *)(*in + 2), type);
+ if ((unsigned)(d-0xdc00) >= 0x400) goto ilseq;
+ c = ((c-0xd7c0)<<10) + (d-0xdc00);
}
break;
case SHIFT_JIS:
if (d>63) d--;
c = gb18030[c][d];
break;
+ case BIG5:
+ l = 2;
+ if (*inb < 2) goto starved;
+ d = *((unsigned char *)*in + 1);
+ if (d-0x40>=0xff-0x40 || d-0x7f<0xa1-0x7f) goto ilseq;
+ d -= 0x40;
+ if (d > 0x3e) d -= 0x22;
+ if (c-0xa1>=0xfa-0xa1) {
+ if (c-0x87>=0xff-0x87) goto ilseq;
+ if (c < 0xa1) c -= 0x87;
+ else c -= 0x87 + (0xfa-0xa1);
+ c = (hkscs[4867+(c*157+d)/16]>>(c*157+d)%16)%2<<17
+ | hkscs[c*157+d];
+ /* A few HKSCS characters map to pairs of UCS
+ * characters. These are mapped to surrogate
+ * range in the hkscs table then hard-coded
+ * here. Ugly, yes. */
+ if (c/256 == 0xdc) {
+ if (totype-0300U > 8) k = 2;
+ else k = "\10\4\4\10\4\4\10\2\4"[totype-0300];
+ if (k > *outb) goto toobig;
+ x += iconv((iconv_t)(uintptr_t)to,
+ &(char *){"\303\212\314\204"
+ "\303\212\314\214"
+ "\303\252\314\204"
+ "\303\252\314\214"
+ +c%256}, &(size_t){4},
+ out, outb);
+ continue;
+ }
+ if (!c) goto ilseq;
+ break;
+ }
+ c -= 0xa1;
+ c = big5[c][d]|(c==0x27&&(d==0x3a||d==0x3c||d==0x42))<<17;
+ if (!c) goto ilseq;
+ break;
+ case EUC_KR:
+ l = 2;
+ if (*inb < 2) goto starved;
+ d = *((unsigned char *)*in + 1);
+ c -= 0xa1;
+ d -= 0xa1;
+ if (c >= 93 || d >= 94) {
+ c += (0xa1-0x81);
+ d += 0xa1;
+ if (c >= 93 || c>=0xc6-0x81 && d>0x52)
+ goto ilseq;
+ if (d-'A'<26) d = d-'A';
+ else if (d-'a'<26) d = d-'a'+26;
+ else if (d-0x81<0xff-0x81) d = d-0x81+52;
+ else goto ilseq;
+ if (c < 0x20) c = 178*c + d;
+ else c = 178*0x20 + 84*(c-0x20) + d;
+ c += 0xac00;
+ for (d=0xac00; d<=c; ) {
+ k = 0;
+ for (int i=0; i<93; i++)
+ for (int j=0; j<94; j++)
+ if (ksc[i][j]-d <= c-d)
+ k++;
+ d = c+1;
+ c += k;
+ }
+ break;
+ }
+ c = ksc[c][d];
+ if (!c) goto ilseq;
+ break;
default:
if (c < 128+type) break;
c -= 128+type;
}
d = c;
for (c=0; c<128-totype; c++) {
- if (d == legacy_chars[ map[c*5/4]>>2*c%8 |
- map[c*5/4+1]<<8-2*c%8 & 1023 ]) {
+ if (d == legacy_chars[ tomap[c*5/4]>>2*c%8 |
+ tomap[c*5/4+1]<<8-2*c%8 & 1023 ]) {
c += 128;
goto revout;
}
case UCS2LE:
case UTF_16BE:
case UTF_16LE:
- if (c < 0x10000) {
+ if (c < 0x10000 || type-UCS2BE < 2U) {
+ if (c >= 0x10000) c = 0xFFFD;
if (*outb < 2) goto toobig;
put_16((void *)*out, c, totype);
*out += 2;
*outb -= 2;
break;
}
- if (type-UCS2BE < 2U) goto ilseq;
if (*outb < 4) goto toobig;
c -= 0x10000;
put_16((void *)*out, (c>>10)|0xd800, totype);
break;
}
}
+ *ploc = loc;
return x;
ilseq:
err = EILSEQ;
x = -1;
end:
errno = err;
+ *ploc = loc;
return x;
}
projects / musl / blobdiff