diff options
Diffstat (limited to 'iconvdata/eucjp.c')
| -rw-r--r-- | iconvdata/eucjp.c | 480 |
1 files changed, 182 insertions, 298 deletions
diff --git a/iconvdata/eucjp.c b/iconvdata/eucjp.c index e6a71cc..24ebed9 100644 --- a/iconvdata/eucjp.c +++ b/iconvdata/eucjp.c @@ -18,306 +18,190 @@ write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ -#include <gconv.h> #include <stdint.h> -#include <string.h> -#include <wchar.h> +#include <gconv.h> #include <jis0201.h> #include <jis0208.h> #include <jis0212.h> -/* Direction of the transformation. */ -static int to_eucjp_object; -static int from_eucjp_object; - - -int -gconv_init (struct gconv_step *step) -{ - /* Determine which direction. */ - if (strcasestr (step->from_name, "EUC-JP") != NULL) - step->data = &from_eucjp_object; - else if (strcasestr (step->to_name, "EUC-JP") != NULL) - step->data = &to_eucjp_object; - else - return GCONV_NOCONV; - - return GCONV_OK; -} - - -void -gconv_end (struct gconv_step *data) -{ - /* Nothing to do. */ -} - - -int -gconv (struct gconv_step *step, struct gconv_step_data *data, - const char *inbuf, size_t *inbufsize, size_t *written, int do_flush) -{ - struct gconv_step *next_step = step + 1; - struct gconv_step_data *next_data = data + 1; - gconv_fct fct = next_step->fct; - size_t do_write; - int result; - - /* If the function is called with no input this means we have to reset - to the initial state. The possibly partly converted input is - dropped. */ - if (do_flush) - { - do_write = 0; - - /* Call the steps down the chain if there are any. */ - if (data->is_last) - result = GCONV_OK; - else - { - struct gconv_step *next_step = step + 1; - struct gconv_step_data *next_data = data + 1; - - result = (*fct) (next_step, next_data, NULL, 0, written, 1); - - /* Clear output buffer. */ - data->outbufavail = 0; - } - } - else - { - do_write = 0; - - do - { - result = GCONV_OK; - - if (step->data == &from_eucjp_object) - { - size_t inchars = *inbufsize; - size_t outwchars = data->outbufavail; - char *outbuf = data->outbuf; - size_t cnt = 0; - - while (cnt < inchars - && (outwchars + sizeof (wchar_t) <= data->outbufsize)) - { - int inchar = (unsigned char) inbuf[cnt]; - wchar_t ch; - - if (inchar <= 0x7f) - ch = (wchar_t) inchar; - else if ((inchar <= 0xa0 || inchar > 0xfe) - && inchar != 0x8e && inchar != 0x8f) - /* This is illegal. */ - ch = L'\0'; - else - { - /* Two or more byte character. First test whether the - next character is also available. */ - int inchar2; - - if (cnt + 1 >= inchars) - { - /* The second character is not available. Store - the intermediate result. */ - result = GCONV_INCOMPLETE_INPUT; - break; - } - - inchar2 = (unsigned char) inbuf[++cnt]; - - /* All second bytes of a multibyte character must be - >= 0xa1. */ - if (inchar2 < 0xa1) - { - /* This is an illegal character. */ - --cnt; - result = GCONV_ILLEGAL_INPUT; - break; - } - - if (inchar == '\x8e') - /* This is code set 2: half-width katakana. */ - ch = jisx0201_to_ucs4 (inchar2); - else if (inchar == '\x8f') - { - /* This is code set 3: JIS X 0212-1990. */ - const char *endp = &inbuf[cnt]; - - ch = jisx0212_to_ucs4 (&endp, 1 + inchars - cnt, - 0x80); - cnt = endp - inbuf; - } - else - { - /* This is code set 1: JIS X 0208. */ - const char *endp = &inbuf[cnt - 1]; - - ch = jisx0208_to_ucs4 (&endp, 2 + inchars - cnt, - 0x80); - if (ch != L'\0') - ++cnt; - } - - if (ch == UNKNOWN_10646_CHAR) - ch = L'\0'; - - if (ch == L'\0') - --cnt; - } - - if (ch == L'\0' && inbuf[cnt] != '\0') - { - /* This is an illegal character. */ - result = GCONV_ILLEGAL_INPUT; - break; - } - - *((wchar_t *) (outbuf + outwchars)) = ch; - ++do_write; - outwchars += sizeof (wchar_t); - ++cnt; - } - *inbufsize -= cnt; - inbuf += cnt; - data->outbufavail = outwchars; - } - else - { - size_t inwchars = *inbufsize; - size_t outchars = data->outbufavail; - char *outbuf = data->outbuf; - size_t cnt = 0; - int extra = 0; - - while (inwchars >= cnt + sizeof (wchar_t) - && outchars < data->outbufsize) - { - wchar_t ch = *((wchar_t *) (inbuf + cnt)); - - if (ch <= L'\x7f') - /* It's plain ASCII. */ - outbuf[outchars] = ch; - else - { - /* Try the JIS character sets. */ - size_t found; - - found = ucs4_to_jisx0201 (ch, &outbuf[outchars]); - - if (found == UNKNOWN_10646_CHAR) - { - /* No JIS 0201 character. */ - found = ucs4_to_jisx0208 (ch, &outbuf[outchars], - (data->outbufsize - - outchars)); - if (found == 0) - { - /* We ran out of space. */ - extra = 2; - break; - } - else if (found != UNKNOWN_10646_CHAR) - { - /* It's a JIS 0208 character, adjust it for - EUC-JP. */ - outbuf[outchars++] += 0x80; - outbuf[outchars] += 0x80; - } - else - { - /* No JIS 0208 character. */ - found = ucs4_to_jisx0212 (ch, &outbuf[outchars], - (data->outbufsize - - outchars)); - - if (found == 0) - { - /* We ran out of space. */ - extra = 2; - break; - } - else if (found != UNKNOWN_10646_CHAR) - { - /* It's a JIS 0212 character, adjust it for - EUC-JP. */ - outbuf[outchars++] += 0x80; - outbuf[outchars] += 0x80; - } - else - /* Illegal character. */ - break; - } - } - } - - ++do_write; - ++outchars; - cnt += sizeof (wchar_t); - } - *inbufsize -= cnt; - inbuf += cnt; - data->outbufavail = outchars; - - if (outchars + extra < data->outbufsize) - { - /* If there is still room in the output buffer something - is wrong with the input. */ - if (inwchars >= cnt + sizeof (wchar_t)) - { - /* An error occurred. */ - result = GCONV_ILLEGAL_INPUT; - break; - } - if (inwchars != cnt) - { - /* There are some unprocessed bytes at the end of the - input buffer. */ - result = GCONV_INCOMPLETE_INPUT; - break; - } - } - } - - if (result != GCONV_OK) - break; - - if (data->is_last) - { - /* This is the last step. */ - result = (*inbufsize > (step->data == &from_eucjp_object - ? 0 : sizeof (wchar_t) - 1) - ? GCONV_FULL_OUTPUT : GCONV_EMPTY_INPUT); - break; - } - - /* Status so far. */ - result = GCONV_EMPTY_INPUT; - - if (data->outbufavail > 0) - { - /* Call the functions below in the chain. */ - size_t newavail = data->outbufavail; - - result = (*fct) (next_step, next_data, data->outbuf, &newavail, - written, 0); - - /* Correct the output buffer. */ - if (newavail != data->outbufavail && newavail > 0) - { - memmove (data->outbuf, - &data->outbuf[data->outbufavail - newavail], - newavail); - data->outbufavail = newavail; - } - } - } - while (*inbufsize > 0 && result == GCONV_EMPTY_INPUT); - } - - if (written != NULL && data->is_last) - *written = do_write; - - return result; -} +/* Definitions used in the body of the `gconv' function. */ +#define CHARSET_NAME "EUC-JP" +#define FROM_LOOP from_euc_jp +#define TO_LOOP to_euc_jp +#define DEFINE_INIT 1 +#define DEFINE_FINI 1 +#define MIN_NEEDED_FROM 1 +#define MAX_NEEDED_FROM 3 +#define MIN_NEEDED_TO 4 + + +/* First define the conversion function from EUC-JP to UCS4. */ +#define MIN_NEEDED_INPUT MIN_NEEDED_FROM +#define MAX_NEEDED_INPUT MAX_NEEDED_FROM +#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO +#define LOOPFCT FROM_LOOP +#define BODY \ + { \ + uint32_t ch = *inptr; \ + \ + if (ch <= 0x7f) \ + ++inptr; \ + else if ((ch <= 0xa0 || ch > 0xfe) && ch != 0x8e && ch != 0x8f) \ + { \ + /* This is illegal. */ \ + result = GCONV_ILLEGAL_INPUT; \ + break; \ + } \ + else \ + { \ + /* Two or more byte character. First test whether the next \ + character is also available. */ \ + int ch2; \ + \ + if (NEED_LENGTH_TEST && inptr + 1 >= inend) \ + { \ + /* The second character is not available. Store the \ + intermediate result. */ \ + result = GCONV_INCOMPLETE_INPUT; \ + break; \ + } \ + \ + ch2 = inptr[1]; \ + \ + /* All second bytes of a multibyte character must be >= 0xa1. */ \ + if (ch2 < 0xa1) \ + { \ + /* This is an illegal character. */ \ + result = GCONV_ILLEGAL_INPUT; \ + break; \ + } \ + \ + if (ch == 0x8e) \ + { \ + /* This is code set 2: half-width katakana. */ \ + ch = jisx0201_to_ucs4 (ch2); \ + inptr += 2; \ + } \ + else \ + { \ + const unsigned char *endp; \ + \ + if (ch == 0x8f) \ + { \ + /* This is code set 3: JIS X 0212-1990. */ \ + endp = inptr + 1; \ + \ + ch = jisx0212_to_ucs4 (&endp, \ + NEED_LENGTH_TEST ? inend - endp : 2, \ + 0x80); \ + } \ + else \ + { \ + /* This is code set 1: JIS X 0208. */ \ + endp = inptr; \ + \ + ch = jisx0208_to_ucs4 (&endp, \ + NEED_LENGTH_TEST ? inend - inptr : 2, \ + 0x80); \ + } \ + \ + if (NEED_LENGTH_TEST && ch == 0) \ + { \ + /* Not enough input available. */ \ + result = GCONV_INCOMPLETE_INPUT; \ + break; \ + } \ + if (ch == UNKNOWN_10646_CHAR) \ + { \ + /* Illegal character. */ \ + result = GCONV_ILLEGAL_INPUT; \ + break; \ + } \ + inptr = endp; \ + } \ + } \ + \ + *((uint32_t *) outptr)++ = ch; \ + } +#include <iconv/loop.c> + + +/* Next, define the other direction. */ +#define MIN_NEEDED_INPUT MIN_NEEDED_TO +#define MIN_NEEDED_OUTPUT MIN_NEEDED_FROM +#define MAX_NEEDED_OUTPUT MAX_NEEDED_FROM +#define LOOPFCT TO_LOOP +#define BODY \ + { \ + uint32_t ch = *((uint32_t *) inptr); \ + \ + if (ch <= 0x7f) \ + /* It's plain ASCII. */ \ + *outptr++ = ch; \ + else \ + { \ + /* Try the JIS character sets. */ \ + size_t found; \ + \ + /* See whether we have room for at least two characters. */ \ + if (NEED_LENGTH_TEST && outptr + 1 >= outend) \ + { \ + result = GCONV_FULL_OUTPUT; \ + break; \ + } \ + \ + found = ucs4_to_jisx0201 (ch, outptr + 1); \ + if (found != UNKNOWN_10646_CHAR) \ + { \ + /* Yes, it's a JIS 0201 character. Store the shift byte. */ \ + *outptr = 0x8e; \ + outptr += 2; \ + } \ + else \ + { \ + /* No JIS 0201 character. */ \ + found = ucs4_to_jisx0208 (ch, outptr, 2); \ + /* Please note that we always have enough room for the output. */ \ + if (found != UNKNOWN_10646_CHAR) \ + { \ + /* It's a JIS 0208 character, adjust it for EUC-JP. */ \ + *outptr++ += 0x80; \ + *outptr++ += 0x80; \ + } \ + else \ + { \ + /* No JIS 0208 character. */ \ + found = ucs4_to_jisx0212 (ch, outptr + 1, \ + (NEED_LENGTH_TEST \ + ? outend - outptr - 1 : 2)); \ + \ + if (found == 0) \ + { \ + /* We ran out of space. */ \ + result = GCONV_FULL_OUTPUT; \ + break; \ + } \ + else if (found != UNKNOWN_10646_CHAR) \ + { \ + /* It's a JIS 0212 character, adjust it for EUC-JP. */ \ + *outptr++ = 0x8f; \ + *outptr++ += 0x80; \ + *outptr++ += 0x80; \ + } \ + else \ + { \ + /* Illegal character. */ \ + result = GCONV_ILLEGAL_INPUT; \ + break; \ + } \ + } \ + } \ + } \ + \ + inptr += 4; \ + } +#include <iconv/loop.c> + + +/* Now define the toplevel functions. */ +#include <iconv/skeleton.c> |