/* This routine is designed to replace the following routine:
mz_euc2euc
mz_b52b5
mz_cns2cns
mz_ksc2ksc
mz_sjis2sjis
mz_utf82utf8
It should also replace
mz_gb2gb
but currently mz_gb2gb also handle hz to gb. We need to move that functionality out of mz_gb2gb
*/
PRIVATE unsigned char *
mz_mbNullConv(CCCDataObject obj, const unsigned char *buf, int32 bufsz)
{
int32 left_over;
int32 len;
unsigned char *p;
unsigned char *ret;
int32 total;
intl_CharLenFunc CharLenFunc = intl_char_len_func[INTL_GetCCCCvtflag(obj)];
int charlen = 0;
/* Get the unconverted buffer */
unsigned char *uncvtbuf = INTL_GetCCCUncvtbuf(obj);
int32 uncvtsz = strlen((char *)uncvtbuf);
/* return in the input is nonsense */
if ((!obj) || (! buf) || (bufsz < 0))
return NULL;
/* Allocate Output Buffer */
total = uncvtsz + bufsz;
ret = (unsigned char *) XP_ALLOC(total + 1);
if (!ret)
{
INTL_SetCCCRetval(obj, MK_OUT_OF_MEMORY);
return NULL;
}
/* Copy unconverted buffer into the output bufer */
memcpy(ret, uncvtbuf, uncvtsz);
/* Copy the current input buffer into the output buffer */
memcpy(ret+uncvtsz, buf, bufsz);
/* Walk through the buffer and figure out the left_over length */
for (p=ret, len=total, left_over=0; len > 0; p += charlen, len -= charlen)
{
if((charlen = CharLenFunc(*p)) > 1)
{ /* count left_over only if it is multibyte char */
if(charlen > len) /* count left_over only if the len is less than charlen */
left_over = len;
};
}
/* Copy the left over into the uncvtbuf */
if(left_over)
memcpy(uncvtbuf, p - charlen, left_over);
/* Null terminated the uncvtbuf */
uncvtbuf[left_over] = '\0';
/* Null terminate the return buffer and set the length */
INTL_SetCCCLen(obj, total - left_over);
ret[total - left_over] = 0;
return ret;
}
MODULE_PRIVATE unsigned char *
mz_sjis2euc( CCCDataObject obj,
const unsigned char *sjisbuf, /* SJIS buf for conversion */
int32 sjisbufsz) /* SJIS buf size in bytes */
{
unsigned char *tobuf = NULL;
int32 tobufsz;
register unsigned char *sjisp, *tobufp; /* current byte in bufs */
register unsigned char *sjisep, *tobufep; /* end of buffers */
int32 uncvtlen;
unsigned char *uncvtbuf = INTL_GetCCCUncvtbuf(obj);
/* Allocate a EUC buffer: */
/* In the worst case ( all Half-width Kanas ), the converted */
/* EUC will be 2X the size of the SJIS + 1 for nul byte */
uncvtlen = strlen((char *)uncvtbuf);
tobufsz = ((sjisbufsz + uncvtlen) << 1) + 1;
if ((tobuf = (unsigned char *)XP_ALLOC(tobufsz)) == (unsigned char *)NULL) {
INTL_SetCCCRetval(obj, MK_OUT_OF_MEMORY);
return(NULL);
}
/* Initialize pointers, etc. */
sjisp = (unsigned char *)sjisbuf;
sjisep = sjisp + sjisbufsz - 1;
#define uncvtp tobufp /* use tobufp as temp */
/* If prev. unconverted chars, append unconverted
* chars w/new chars and try to process.
*/
if (uncvtbuf[0] != '\0') {
uncvtp = uncvtbuf + uncvtlen;
while (uncvtp < (uncvtbuf + UNCVTBUF_SIZE) &&
sjisp <= sjisep)
*uncvtp++ = *sjisp++;
*uncvtp = '\0'; /* nul terminate */
sjisp = uncvtbuf; /* process unconverted first */
sjisep = uncvtp - 1;
}
#undef uncvtp
tobufp = tobuf;
tobufep = tobufp + tobufsz - 1; /* save space for terminating null */
WHILELOOP:
/* While SJIS data && space in EUC buf. */
while ((sjisp <= sjisep) && (tobufp <= tobufep)) {
if (*sjisp < 0x80) {
/* ASCII/JIS-Roman */
*tobufp++ = *sjisp++;
} else if (*sjisp < 0xA0) {
/* 1st byte of 2-byte low SJIS. */
if (sjisp+1 > sjisep) /* No 2nd byte in SJIS buffer? */
break;
TwoByteSJIS2EUC(sjisp, tobufp, 0x70);
} else if (*sjisp==0xA0) {
/* SJIS half-width space. */
/* Just treat like Roman?? */
*tobufp++ = *sjisp++;
} else if (*sjisp < 0xE0) {
/* SJIS half-width katakana */
*tobufp++ = SS2;
*tobufp++ = *sjisp | 0x80; /* Set 8th bit for EUC & SJIS */
sjisp++;
} else if (*sjisp < 0xF0) {
/* 1st byte of 2-byte high SJIS */
if (sjisp+1 > sjisep) /* No 2nd byte in SJIS buffer? */
break;
TwoByteSJIS2EUC(sjisp, tobufp, 0xB0);
} else {
/* User Defined SJIS: copy bytes */
if (sjisp+1 > sjisep) /* No 2nd byte in SJIS buf? */
break;
*tobufp++ = *sjisp++; /* Just copy 2 bytes. */
*tobufp++ = *sjisp++;
}
}
if (uncvtbuf[0] != '\0') {
/* jisp pts to 1st unprocessed char in
* jisbuf. Some may have been processed
* while processing unconverted chars,
* so set up ptrs not to process them
* twice.
*/
sjisp = (unsigned char *)sjisbuf + (sjisp - uncvtbuf - uncvtlen);
sjisep = (unsigned char *)sjisbuf + sjisbufsz - 1;
uncvtbuf[0] = '\0'; /* No more uncoverted chars. */
goto WHILELOOP; /* Process new data */
}
*tobufp = '\0'; /* null terminate EUC data */
INTL_SetCCCLen(obj, tobufp - tobuf); /* length not counting null */
if (sjisp <= sjisep) { /* uncoverted SJIS? */
tobufp = uncvtbuf; /* reuse the tobufp as a TEMP */
while (sjisp <= sjisep)
*tobufp++ = *sjisp++;
*tobufp = '\0'; /* null terminate */
}
return(tobuf);
}