diff options
author | Ulrich Drepper <drepper@redhat.com> | 2002-04-20 08:14:24 +0000 |
---|---|---|
committer | Ulrich Drepper <drepper@redhat.com> | 2002-04-20 08:14:24 +0000 |
commit | 02779eaafe9065ecafa307f33ddbb8b5eac999b7 (patch) | |
tree | 255451e8aea47f26bc18215497173fe97d6e56d9 /iconvdata/euc-jisx0213.c | |
parent | 93a568aaba1d3da3497102dfe1eeb81070a24c1c (diff) | |
download | glibc-02779eaafe9065ecafa307f33ddbb8b5eac999b7.tar.gz |
Update.
2002-04-18 Bruno Haible <bruno@clisp.org>
* iconvdata/euc-jisx0213.c (EMIT_SHIFT_TO_INIT, BODY for
FROM_DIRECTION): Make the FROM direction stateless.
* iconvdata/shift_jisx0213.c (EMIT_SHIFT_TO_INIT, BODY for
FROM_DIRECTION): Likewise.
* iconvdata/cvs11643l1.c: Update comments.
* sysdeps/unix/sysv/linux/s390/s390-64/syscalls.list: Likewise.
Diffstat (limited to 'iconvdata/euc-jisx0213.c')
-rw-r--r-- | iconvdata/euc-jisx0213.c | 204 |
1 files changed, 93 insertions, 111 deletions
diff --git a/iconvdata/euc-jisx0213.c b/iconvdata/euc-jisx0213.c index b0836b173d..c696f94ee8 100644 --- a/iconvdata/euc-jisx0213.c +++ b/iconvdata/euc-jisx0213.c @@ -62,9 +62,7 @@ *statep = saved_state -/* During EUC-JISX0213 to UCS-4 conversion, the COUNT element of the state - contains the last UCS-4 character, shifted by 3 bits. - During UCS-4 to EUC-JISX0213 conversion, the COUNT element of the state +/* During UCS-4 to EUC-JISX0213 conversion, the COUNT element of the state contains the last two bytes to be output, shifted by 3 bits. */ /* Since this is a stateful encoding we have to provide code which resets @@ -74,17 +72,8 @@ if (data->__statep->__count != 0) \ { \ if (FROM_DIRECTION) \ - { \ - if (__builtin_expect (outbuf + 4 <= outend, 1)) \ - { \ - /* Write out the last character. */ \ - *((uint32_t *) outbuf)++ = data->__statep->__count >> 3; \ - data->__statep->__count = 0; \ - } \ - else \ - /* We don't have enough room in the output buffer. */ \ - status = __GCONV_FULL_OUTPUT; \ - } \ + /* We don't use shift states in the FROM_DIRECTION. */ \ + data->__statep->__count = 0; \ else \ { \ if (__builtin_expect (outbuf + 2 <= outend, 1)) \ @@ -109,33 +98,44 @@ #define LOOPFCT FROM_LOOP #define BODY \ { \ - uint32_t ch; \ + uint32_t ch = *inptr; \ \ - /* Determine whether there is a buffered character pending. */ \ - ch = *statep >> 3; \ - if (__builtin_expect (ch == 0, 1)) \ + if (ch < 0x80) \ + /* Plain ASCII character. */ \ + ++inptr; \ + else if ((ch >= 0xa1 && ch <= 0xfe) || ch == 0x8e || ch == 0x8f) \ { \ - /* No - so look at the next input byte. */ \ - ch = *inptr; \ - if (ch < 0x80) \ - /* Plain ASCII character. */ \ - ++inptr; \ - else if ((ch >= 0xa1 && ch <= 0xfe) || ch == 0x8e || ch == 0x8f) \ + /* Two or three byte character. */ \ + uint32_t ch2; \ + \ + if (__builtin_expect (inptr + 1 >= inend, 0)) \ { \ - /* Two or three byte character. */ \ - uint32_t ch2; \ + /* The second byte is not available. */ \ + result = __GCONV_INCOMPLETE_INPUT; \ + break; \ + } \ + \ + ch2 = inptr[1]; \ \ - if (__builtin_expect (inptr + 1 >= inend, 0)) \ + /* The second byte must be >= 0xa1 and <= 0xfe. */ \ + if (__builtin_expect (ch2 < 0xa1 || ch2 > 0xfe, 0)) \ + { \ + /* This is an illegal character. */ \ + if (! ignore_errors_p ()) \ { \ - /* The second byte is not available. */ \ - result = __GCONV_INCOMPLETE_INPUT; \ + result = __GCONV_ILLEGAL_INPUT; \ break; \ } \ \ - ch2 = inptr[1]; \ + ++inptr; \ + ++*irreversible; \ + break; \ + } \ \ - /* The second byte must be >= 0xa1 and <= 0xfe. */ \ - if (__builtin_expect (ch2 < 0xa1 || ch2 > 0xfe, 0)) \ + if (ch == 0x8e) \ + { \ + /* Half-width katakana. */ \ + if (__builtin_expect (ch2 > 0xdf, 0)) \ { \ /* This is an illegal character. */ \ if (! ignore_errors_p ()) \ @@ -149,107 +149,89 @@ break; \ } \ \ - if (ch == 0x8e) \ + ch = ch2 + 0xfec0; \ + inptr += 2; \ + } \ + else \ + { \ + const unsigned char *endp; \ + \ + if (ch == 0x8f) \ { \ - /* Half-width katakana. */ \ - if (__builtin_expect (ch2 > 0xdf, 0)) \ + /* JISX 0213 plane 2. */ \ + uint32_t ch3; \ + \ + if (__builtin_expect (inptr + 2 >= inend, 0)) \ { \ - /* This is an illegal character. */ \ - if (! ignore_errors_p ()) \ - { \ - result = __GCONV_ILLEGAL_INPUT; \ - break; \ - } \ - \ - ++inptr; \ - ++*irreversible; \ + /* The third byte is not available. */ \ + result = __GCONV_INCOMPLETE_INPUT; \ break; \ } \ \ - ch = ch2 + 0xfec0; \ - inptr += 2; \ + ch3 = inptr[2]; \ + endp = inptr + 3; \ + \ + ch = jisx0213_to_ucs4 (0x200 - 0x80 + ch2, ch3 ^ 0x80); \ } \ else \ { \ - const unsigned char *endp; \ + /* JISX 0213 plane 1. */ \ + endp = inptr + 2; \ \ - if (ch == 0x8f) \ - { \ - /* JISX 0213 plane 2. */ \ - uint32_t ch3; \ + ch = jisx0213_to_ucs4 (0x100 - 0x80 + ch, ch2 ^ 0x80); \ + } \ \ - if (__builtin_expect (inptr + 2 >= inend, 0)) \ - { \ - /* The third byte is not available. */ \ - result = __GCONV_INCOMPLETE_INPUT; \ - break; \ - } \ + if (ch == 0) \ + { \ + /* This is an illegal character. */ \ + if (! ignore_errors_p ()) \ + { \ + result = __GCONV_ILLEGAL_INPUT; \ + break; \ + } \ \ - ch3 = inptr[2]; \ - endp = inptr + 3; \ + ++inptr; \ + ++*irreversible; \ + break; \ + } \ \ - ch = jisx0213_to_ucs4 (0x200 - 0x80 + ch2, ch3 ^ 0x80); \ - } \ - else \ - { \ - /* JISX 0213 plane 1. */ \ - endp = inptr + 2; \ + inptr = endp; \ \ - ch = jisx0213_to_ucs4 (0x100 - 0x80 + ch, ch2 ^ 0x80); \ - } \ + if (ch < 0x80) \ + { \ + /* It's a combining character. */ \ + uint32_t u1 = __jisx0213_to_ucs_combining[ch - 1][0]; \ + uint32_t u2 = __jisx0213_to_ucs_combining[ch - 1][1]; \ \ - if (ch == 0) \ + /* See whether we have room for two characters. */ \ + if (outptr + 8 <= outend) \ { \ - /* This is an illegal character. */ \ - if (! ignore_errors_p ()) \ - { \ - result = __GCONV_ILLEGAL_INPUT; \ - break; \ - } \ - \ - ++inptr; \ - ++*irreversible; \ - break; \ + put32 (outptr, u1); \ + outptr += 4; \ + put32 (outptr, u2); \ + outptr += 4; \ + continue; \ } \ - \ - inptr = endp; \ - \ - if (ch < 0x80) \ + else \ { \ - /* It's a combining character. */ \ - uint32_t u1 = __jisx0213_to_ucs_combining[ch - 1][0]; \ - uint32_t u2 = __jisx0213_to_ucs_combining[ch - 1][1]; \ - \ - /* See whether we have room for two characters. */ \ - if (outptr + 8 <= outend) \ - { \ - put32 (outptr, u1); \ - outptr += 4; \ - put32 (outptr, u2); \ - outptr += 4; \ - continue; \ - } \ - \ - /* Otherwise store only the first character now, and \ - put the second one into the queue. */ \ - ch = u1; \ - *statep = u2 << 3; \ + result = __GCONV_FULL_OUTPUT; \ + break; \ } \ } \ } \ - else \ + } \ + else \ + { \ + /* This is illegal. */ \ + if (! ignore_errors_p ()) \ { \ - /* This is illegal. */ \ - if (! ignore_errors_p ()) \ - { \ - result = __GCONV_ILLEGAL_INPUT; \ - break; \ - } \ - \ - ++inptr; \ - ++*irreversible; \ - continue; \ + result = __GCONV_ILLEGAL_INPUT; \ + break; \ } \ + \ + ++inptr; \ + ++*irreversible; \ + continue; \ } \ \ put32 (outptr, ch); \ |