/* Simple transformations functions. Copyright (C) 1997, 1998, 1999, 2000 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper , 1997. The GNU C Library is free software; you can redistribute it and/or modify it under the terms of the GNU Library General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. The GNU C Library is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Library General Public License for more details. You should have received a copy of the GNU Library General Public License along with the GNU C Library; see the file COPYING.LIB. If not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ #include #include #include #include #include #include #include #include #include #ifndef EILSEQ # define EILSEQ EINVAL #endif /* These are definitions used by some of the functions for handling UTF-8 encoding below. */ static const uint32_t encoding_mask[] = { ~0x7ff, ~0xffff, ~0x1fffff, ~0x3ffffff }; static const unsigned char encoding_byte[] = { 0xc0, 0xe0, 0xf0, 0xf8, 0xfc }; /* Transform from the internal, UCS4-like format, to UCS4. The difference between the internal ucs4 format and the real UCS4 format is, if any, the endianess. The Unicode/ISO 10646 says that unless some higher protocol specifies it differently, the byte order is big endian.*/ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 4 #define MIN_NEEDED_TO 4 #define FROM_DIRECTION 1 #define FROM_LOOP internal_ucs4_loop #define TO_LOOP internal_ucs4_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_internal_ucs4 static inline int internal_ucs4_loop (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { const unsigned char *inptr = *inptrp; unsigned char *outptr = *outptrp; size_t n_convert = MIN (inend - inptr, outend - outptr) / 4; int result; #if __BYTE_ORDER == __LITTLE_ENDIAN /* Sigh, we have to do some real work. */ size_t cnt; for (cnt = 0; cnt < n_convert; ++cnt, inptr += 4) *((uint32_t *) outptr)++ = bswap_32 (*(uint32_t *) inptr); *inptrp = inptr; *outptrp = outptr; #elif __BYTE_ORDER == __BIG_ENDIAN /* Simply copy the data. */ *inptrp = inptr + n_convert * 4; *outptrp = __mempcpy (outptr, inptr, n_convert * 4); #else # error "This endianess is not supported." #endif /* Determine the status. */ if (*outptrp == outend) result = __GCONV_FULL_OUTPUT; else if (*inptrp == inend) result = __GCONV_EMPTY_INPUT; else result = __GCONV_INCOMPLETE_INPUT; return result; } #ifndef _STRING_ARCH_unaligned static inline int internal_ucs4_loop_unaligned (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { const unsigned char *inptr = *inptrp; unsigned char *outptr = *outptrp; size_t n_convert = MIN (inend - inptr, outend - outptr) / 4; int result; # if __BYTE_ORDER == __LITTLE_ENDIAN /* Sigh, we have to do some real work. */ size_t cnt; for (cnt = 0; cnt < n_convert; ++cnt, inptr += 4, outptr += 4) { outptr[0] = inptr[3]; outptr[1] = inptr[2]; outptr[2] = inptr[1]; outptr[3] = inptr[0]; } *inptrp = inptr; *outptrp = outptr; # elif __BYTE_ORDER == __BIG_ENDIAN /* Simply copy the data. */ *inptrp = inptr + n_convert * 4; *outptrp = __mempcpy (outptr, inptr, n_convert * 4); # else # error "This endianess is not supported." # endif /* Determine the status. */ if (*outptrp == outend) result = __GCONV_FULL_OUTPUT; else if (*inptrp == inend) result = __GCONV_EMPTY_INPUT; else result = __GCONV_INCOMPLETE_INPUT; return result; } #endif static inline int internal_ucs4_loop_single (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { size_t cnt = state->__count & 7; while (*inptrp < inend && cnt < 4) state->__value.__wchb[cnt++] = *(*inptrp)++; if (cnt < 4) { /* Still not enough bytes. Store the ones in the input buffer. */ state->__count &= ~7; state->__count |= cnt; return __GCONV_INCOMPLETE_INPUT; } #if __BYTE_ORDER == __LITTLE_ENDIAN (*outptrp)[0] = state->__value.__wchb[3]; (*outptrp)[1] = state->__value.__wchb[2]; (*outptrp)[2] = state->__value.__wchb[1]; (*outptrp)[3] = state->__value.__wchb[0]; #elif __BYTE_ORDER == __BIG_ENDIAN /* XXX unaligned */ *(*((uint32_t **) outptrp)++) = state->__value.__wch; #else # error "This endianess is not supported." #endif /* Clear the state buffer. */ state->__count &= ~7; return __GCONV_OK; } #include /* Transform from UCS4 to the internal, UCS4-like format. Unlike for the other direction we have to check for correct values here. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 4 #define MIN_NEEDED_TO 4 #define FROM_DIRECTION 1 #define FROM_LOOP ucs4_internal_loop #define TO_LOOP ucs4_internal_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_ucs4_internal static inline int ucs4_internal_loop (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { const unsigned char *inptr = *inptrp; unsigned char *outptr = *outptrp; size_t n_convert = MIN (inend - inptr, outend - outptr) / 4; int result; size_t cnt; for (cnt = 0; cnt < n_convert; ++cnt, inptr += 4) { uint32_t inval; #if __BYTE_ORDER == __LITTLE_ENDIAN inval = bswap_32 (*(uint32_t *) inptr); #else inval = *(uint32_t *) inptr; #endif if (inval > 0x7fffffff) { *inptrp = inptr; *outptrp = outptr; return __GCONV_ILLEGAL_INPUT; } *((uint32_t *) outptr)++ = bswap_32 (*(uint32_t *) inptr); } *inptrp = inptr; *outptrp = outptr; /* Determine the status. */ if (*outptrp == outend) result = __GCONV_FULL_OUTPUT; else if (*inptrp == inend) result = __GCONV_EMPTY_INPUT; else result = __GCONV_INCOMPLETE_INPUT; return result; } #ifndef _STRING_ARCH_unaligned static inline int ucs4_internal_loop_unaligned (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { const unsigned char *inptr = *inptrp; unsigned char *outptr = *outptrp; size_t n_convert = MIN (inend - inptr, outend - outptr) / 4; int result; size_t cnt; for (cnt = 0; cnt < n_convert; ++cnt, inptr += 4, outptr += 4) { if (inptr[0] > 0x80) { /* The value is too large. */ *inptrp = inptr; *outptrp = outptr; return __GCONV_ILLEGAL_INPUT; } # if __BYTE_ORDER == __LITTLE_ENDIAN outptr[3] = inptr[0]; outptr[2] = inptr[1]; outptr[1] = inptr[2]; outptr[0] = inptr[3]; # else outptr[0] = inptr[0]; outptr[1] = inptr[1]; outptr[2] = inptr[2]; outptr[3] = inptr[3]; # endif # if __BYTE_ORDER == __LITTLE_ENDIAN outptr[3] = inptr[0]; outptr[2] = inptr[1]; outptr[1] = inptr[2]; outptr[0] = inptr[3]; # else outptr[0] = inptr[0]; outptr[1] = inptr[1]; outptr[2] = inptr[2]; outptr[3] = inptr[3]; # endif } *inptrp = inptr; *outptrp = outptr; /* Determine the status. */ if (*outptrp == outend) result = __GCONV_FULL_OUTPUT; else if (*inptrp == inend) result = __GCONV_EMPTY_INPUT; else result = __GCONV_INCOMPLETE_INPUT; return result; } #endif static inline int ucs4_internal_loop_single (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { size_t cnt = state->__count & 7; while (*inptrp < inend && cnt < 4) state->__value.__wchb[cnt++] = *(*inptrp)++; if (cnt < 4) { /* Still not enough bytes. Store the ones in the input buffer. */ state->__count &= ~7; state->__count |= cnt; return __GCONV_INCOMPLETE_INPUT; } if (((unsigned char *) state->__value.__wchb)[0] > 0x80) /* The value is too large. */ return __GCONV_ILLEGAL_INPUT; #if __BYTE_ORDER == __LITTLE_ENDIAN (*outptrp)[0] = state->__value.__wchb[3]; (*outptrp)[1] = state->__value.__wchb[2]; (*outptrp)[2] = state->__value.__wchb[1]; (*outptrp)[3] = state->__value.__wchb[0]; #elif __BYTE_ORDER == __BIG_ENDIAN (*outptrp)[0] = state->__value.__wchb[0]; (*outptrp)[1] = state->__value.__wchb[1]; (*outptrp)[2] = state->__value.__wchb[2]; (*outptrp)[3] = state->__value.__wchb[3]; #endif /* Clear the state buffer. */ state->__count &= ~7; return __GCONV_OK; } #include /* Similarly for the little endian form. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 4 #define MIN_NEEDED_TO 4 #define FROM_DIRECTION 1 #define FROM_LOOP internal_ucs4le_loop #define TO_LOOP internal_ucs4le_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_internal_ucs4le static inline int internal_ucs4le_loop (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { const unsigned char *inptr = *inptrp; unsigned char *outptr = *outptrp; size_t n_convert = MIN (inend - inptr, outend - outptr) / 4; int result; #if __BYTE_ORDER == __BIG_ENDIAN /* Sigh, we have to do some real work. */ size_t cnt; for (cnt = 0; cnt < n_convert; ++cnt, inptr += 4) *((uint32_t *) outptr)++ = bswap_32 (*(uint32_t *) inptr); *inptrp = inptr; *outptrp = outptr; #elif __BYTE_ORDER == __LITTLE_ENDIAN /* Simply copy the data. */ *inptrp = inptr + n_convert * 4; *outptrp = __mempcpy (outptr, inptr, n_convert * 4); #else # error "This endianess is not supported." #endif /* Determine the status. */ if (*outptrp == outend) result = __GCONV_FULL_OUTPUT; else if (*inptrp == inend) result = __GCONV_EMPTY_INPUT; else result = __GCONV_INCOMPLETE_INPUT; return result; } #ifndef _STRING_ARCH_unaligned static inline int internal_ucs4le_loop_unaligned (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { const unsigned char *inptr = *inptrp; unsigned char *outptr = *outptrp; size_t n_convert = MIN (inend - inptr, outend - outptr) / 4; int result; # if __BYTE_ORDER == __BIG_ENDIAN /* Sigh, we have to do some real work. */ size_t cnt; for (cnt = 0; cnt < n_convert; ++cnt, inptr += 4) { outptr[0] = inptr[3]; outptr[1] = inptr[2]; outptr[2] = inptr[1]; outptr[3] = inptr[0]; } *inptrp = inptr; *outptrp = outptr; # elif __BYTE_ORDER == __LITTLE_ENDIAN /* Simply copy the data. */ *inptrp = inptr + n_convert * 4; *outptrp = __mempcpy (outptr, inptr, n_convert * 4); # else # error "This endianess is not supported." # endif /* Determine the status. */ if (*outptrp == outend) result = __GCONV_FULL_OUTPUT; else if (*inptrp == inend) result = __GCONV_EMPTY_INPUT; else result = __GCONV_INCOMPLETE_INPUT; return result; } #endif static inline int internal_ucs4le_loop_single (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { size_t cnt = state->__count & 7; while (*inptrp < inend && cnt < 4) state->__value.__wchb[cnt++] = *(*inptrp)++; if (cnt < 4) { /* Still not enough bytes. Store the ones in the input buffer. */ state->__count &= ~7; state->__count |= cnt; return __GCONV_INCOMPLETE_INPUT; } #if __BYTE_ORDER == __BIG_ENDIAN (*outptrp)[0] = state->__value.__wchb[3]; (*outptrp)[1] = state->__value.__wchb[2]; (*outptrp)[2] = state->__value.__wchb[1]; (*outptrp)[3] = state->__value.__wchb[0]; #else /* XXX unaligned */ *(*((uint32_t **) outptrp)++) = state->__value.__wch; #endif /* Clear the state buffer. */ state->__count &= ~7; return __GCONV_OK; } #include /* And finally from UCS4-LE to the internal encoding. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 4 #define MIN_NEEDED_TO 4 #define FROM_DIRECTION 1 #define FROM_LOOP ucs4le_internal_loop #define TO_LOOP ucs4le_internal_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_ucs4le_internal static inline int ucs4le_internal_loop (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { const unsigned char *inptr = *inptrp; unsigned char *outptr = *outptrp; size_t n_convert = MIN (inend - inptr, outend - outptr) / 4; int result; size_t cnt; for (cnt = 0; cnt < n_convert; ++cnt, inptr += 4) { uint32_t inval; #if __BYTE_ORDER == __BIG_ENDIAN inval = bswap_32 (*(uint32_t *) inptr); #else inval = *(uint32_t *) inptr; #endif if (inval > 0x7fffffff) return __GCONV_ILLEGAL_INPUT; *((uint32_t *) outptr)++ = bswap_32 (*(uint32_t *) inptr); } *inptrp = inptr; *outptrp = outptr; /* Determine the status. */ if (*outptrp == outend) result = __GCONV_FULL_OUTPUT; else if (*inptrp == inend) result = __GCONV_EMPTY_INPUT; else result = __GCONV_INCOMPLETE_INPUT; return result; } #ifndef _STRING_ARCH_unaligned static inline int ucs4le_internal_loop_unaligned (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { const unsigned char *inptr = *inptrp; unsigned char *outptr = *outptrp; size_t n_convert = MIN (inend - inptr, outend - outptr) / 4; int result; size_t cnt; for (cnt = 0; cnt < n_convert; ++cnt, inptr += 4) { if (inptr[3] > 0x80) { /* The value is too large. */ *inptrp = inptr; *outptrp = outptr; return __GCONV_ILLEGAL_INPUT; } # if __BYTE_ORDER == __BIG_ENDIAN outptr[3] = inptr[0]; outptr[2] = inptr[1]; outptr[1] = inptr[2]; outptr[0] = inptr[3]; # else outptr[0] = inptr[0]; outptr[1] = inptr[1]; outptr[2] = inptr[2]; outptr[3] = inptr[3]; # endif } *inptrp = inptr; *outptrp = outptr; /* Determine the status. */ if (*outptrp == outend) result = __GCONV_FULL_OUTPUT; else if (*inptrp == inend) result = __GCONV_EMPTY_INPUT; else result = __GCONV_INCOMPLETE_INPUT; return result; } #endif static inline int ucs4le_internal_loop_single (const unsigned char **inptrp, const unsigned char *inend, unsigned char **outptrp, unsigned char *outend, mbstate_t *state, void *data, size_t *converted) { size_t cnt = state->__count & 7; while (*inptrp < inend && cnt < 4) state->__value.__wchb[cnt++] = *(*inptrp)++; if (cnt < 4) { /* Still not enough bytes. Store the ones in the input buffer. */ state->__count &= ~7; state->__count |= cnt; return __GCONV_INCOMPLETE_INPUT; } if (((unsigned char *) state->__value.__wchb)[3] > 0x80) /* The value is too large. */ return __GCONV_ILLEGAL_INPUT; #if __BYTE_ORDER == __BIG_ENDIAN (*outptrp)[0] = state->__value.__wchb[3]; (*outptrp)[1] = state->__value.__wchb[2]; (*outptrp)[2] = state->__value.__wchb[1]; (*outptrp)[3] = state->__value.__wchb[0]; #elif __BYTE_ORDER == __BIG_ENDIAN (*outptrp)[0] = state->__value.__wchb[0]; (*outptrp)[1] = state->__value.__wchb[1]; (*outptrp)[2] = state->__value.__wchb[2]; (*outptrp)[3] = state->__value.__wchb[3]; #endif /* Clear the state buffer. */ state->__count &= ~7; return __GCONV_OK; } #include /* Convert from ISO 646-IRV to the internal (UCS4-like) format. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 1 #define MIN_NEEDED_TO 4 #define FROM_DIRECTION 1 #define FROM_LOOP ascii_internal_loop #define TO_LOOP ascii_internal_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_ascii_internal #define ONE_DIRECTION 1 #define MIN_NEEDED_INPUT MIN_NEEDED_FROM #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO #define LOOPFCT FROM_LOOP #define BODY \ { \ if (*inptr > '\x7f') \ { \ /* This is no correct ANSI_X3.4-1968 character. */ \ result = __GCONV_ILLEGAL_INPUT; \ break; \ } \ \ /* It's an one byte sequence. */ \ /* XXX unaligned. */ \ *((uint32_t *) outptr)++ = *inptr++; \ } #include #include /* Convert from the internal (UCS4-like) format to ISO 646-IRV. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 4 #define MIN_NEEDED_TO 1 #define FROM_DIRECTION 1 #define FROM_LOOP internal_ascii_loop #define TO_LOOP internal_ascii_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_internal_ascii #define ONE_DIRECTION 1 #define MIN_NEEDED_INPUT MIN_NEEDED_FROM #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO #define LOOPFCT FROM_LOOP #define BODY \ { \ if (*((uint32_t *) inptr) > 0x7f) \ { \ /* This is no correct ANSI_X3.4-1968 character. */ \ result = __GCONV_ILLEGAL_INPUT; \ break; \ } \ \ /* It's an one byte sequence. */ \ *outptr++ = *((uint32_t *) inptr)++; \ } #include #include /* Convert from the internal (UCS4-like) format to UTF-8. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 4 #define MIN_NEEDED_TO 1 #define MAX_NEEDED_TO 6 #define FROM_DIRECTION 1 #define FROM_LOOP internal_utf8_loop #define TO_LOOP internal_utf8_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_internal_utf8 #define ONE_DIRECTION 1 #define MIN_NEEDED_INPUT MIN_NEEDED_FROM #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO #define MAX_NEEDED_OUTPUT MAX_NEEDED_TO #define LOOPFCT FROM_LOOP #define BODY \ { \ uint32_t wc = *((uint32_t *) inptr); \ \ /* Since we control every character we read this cannot happen. */ \ assert (wc <= 0x7fffffff); \ \ if (wc < 0x80) \ /* It's an one byte sequence. */ \ *outptr++ = (unsigned char) wc; \ else \ { \ size_t step; \ char *start; \ \ for (step = 2; step < 6; ++step) \ if ((wc & encoding_mask[step - 2]) == 0) \ break; \ \ if (outptr + step > outend) \ { \ /* Too long. */ \ result = __GCONV_FULL_OUTPUT; \ break; \ } \ \ start = outptr; \ *outptr = encoding_byte[step - 2]; \ outptr += step; \ --step; \ do \ { \ start[step] = 0x80 | (wc & 0x3f); \ wc >>= 6; \ } \ while (--step > 0); \ start[0] |= wc; \ } \ \ inptr += 4; \ } #include #include /* Convert from UTF-8 to the internal (UCS4-like) format. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 1 #define MAX_NEEDED_FROM 6 #define MIN_NEEDED_TO 4 #define FROM_DIRECTION 1 #define FROM_LOOP utf8_internal_loop #define TO_LOOP utf8_internal_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_utf8_internal #define ONE_DIRECTION 1 #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; \ uint_fast32_t cnt; \ uint_fast32_t i; \ \ /* Next input byte. */ \ ch = *inptr; \ \ if (ch < 0x80) \ { \ /* One byte sequence. */ \ cnt = 1; \ ++inptr; \ } \ else \ { \ if (ch >= 0xc2 && ch < 0xe0) \ { \ /* We expect two bytes. The first byte cannot be 0xc0 or 0xc1, \ otherwise the wide character could have been represented \ using a single byte. */ \ cnt = 2; \ ch &= 0x1f; \ } \ else if ((ch & 0xf0) == 0xe0) \ { \ /* We expect three bytes. */ \ cnt = 3; \ ch &= 0x0f; \ } \ else if ((ch & 0xf8) == 0xf0) \ { \ /* We expect four bytes. */ \ cnt = 4; \ ch &= 0x07; \ } \ else if ((ch & 0xfc) == 0xf8) \ { \ /* We expect five bytes. */ \ cnt = 5; \ ch &= 0x03; \ } \ else if ((ch & 0xfe) == 0xfc) \ { \ /* We expect six bytes. */ \ cnt = 6; \ ch &= 0x01; \ } \ else \ { \ /* This is an illegal encoding. */ \ result = __GCONV_ILLEGAL_INPUT; \ break; \ } \ \ if (NEED_LENGTH_TEST && inptr + cnt > inend) \ { \ /* We don't have enough input. But before we report that check \ that all the bytes are correct. */ \ for (i = 1; inptr + i < inend; ++i) \ if ((inptr[i] & 0xc0) != 0x80) \ break; \ result = (inptr + i == inend \ ? __GCONV_INCOMPLETE_INPUT : __GCONV_ILLEGAL_INPUT); \ break; \ } \ \ /* Read the possible remaining bytes. */ \ for (i = 1; i < cnt; ++i) \ { \ uint32_t byte = inptr[i]; \ \ if ((byte & 0xc0) != 0x80) \ /* This is an illegal encoding. */ \ break; \ \ ch <<= 6; \ ch |= byte & 0x3f; \ } \ \ /* If i < cnt, some trail byte was not >= 0x80, < 0xc0. \ If cnt > 2 and ch < 2^(5*cnt-4), the wide character ch could \ have been represented with fewer than cnt bytes. */ \ if (i < cnt || (cnt > 2 && (ch >> (5 * cnt - 4)) == 0)) \ { \ /* This is an illegal encoding. */ \ result = __GCONV_ILLEGAL_INPUT; \ break; \ } \ \ inptr += cnt; \ } \ \ /* Now adjust the pointers and store the result. */ \ *((uint32_t *) outptr)++ = ch; \ } #define STORE_REST \ { \ /* We store the remaining bytes while converting them into the UCS4 \ format. We can assume that the first byte in the buffer is \ correct and that it requires a larger number of bytes than there \ are in the input buffer. */ \ wint_t ch = **inptrp; \ size_t cnt; \ \ state->__count = inend - *inptrp; \ \ if (ch >= 0xc2 && ch < 0xe0) \ { \ /* We expect two bytes. The first byte cannot be 0xc0 or \ 0xc1, otherwise the wide character could have been \ represented using a single byte. */ \ cnt = 2; \ ch &= 0x1f; \ } \ else if ((ch & 0xf0) == 0xe0) \ { \ /* We expect three bytes. */ \ cnt = 3; \ ch &= 0x0f; \ } \ else if ((ch & 0xf8) == 0xf0) \ { \ /* We expect four bytes. */ \ cnt = 4; \ ch &= 0x07; \ } \ else if ((ch & 0xfc) == 0xf8) \ { \ /* We expect five bytes. */ \ cnt = 5; \ ch &= 0x03; \ } \ else \ { \ /* We expect six bytes. */ \ cnt = 6; \ ch &= 0x01; \ } \ \ /* The first byte is already consumed. */ \ --cnt; \ while (++(*inptrp) < inend) \ { \ ch <<= 6; \ ch |= **inptrp & 0x3f; \ --cnt; \ } \ \ /* Shift for the so far missing bytes. */ \ ch <<= cnt * 6; \ \ /* Store the value. */ \ state->__value.__wch = ch; \ } #define UNPACK_BYTES \ { \ wint_t wch = state->__value.__wch; \ inlen = state->__count; \ \ if (state->__value.__wch <= 0x7ff) \ bytebuf[0] = 0xc0; \ else if (state->__value.__wch <= 0xffff) \ bytebuf[0] = 0xe0; \ else if (state->__value.__wch <= 0x1fffff) \ bytebuf[0] = 0xf0; \ else if (state->__value.__wch <= 0x3ffffff) \ bytebuf[0] = 0xf8; \ else \ bytebuf[0] = 0xfc; \ \ while (inlen-- > 1) \ bytebuf[inlen] = 0x80 | (wch & 0x3f); \ \ bytebuf[0] |= wch; \ } #include #include /* Convert from UCS2 to the internal (UCS4-like) format. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 2 #define MIN_NEEDED_TO 4 #define FROM_DIRECTION 1 #define FROM_LOOP ucs2_internal_loop #define TO_LOOP ucs2_internal_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_ucs2_internal #define ONE_DIRECTION 1 #define MIN_NEEDED_INPUT MIN_NEEDED_FROM #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO #define LOOPFCT FROM_LOOP #define BODY \ *((uint32_t *) outptr)++ = *((uint16_t *) inptr)++; #include #include /* Convert from the internal (UCS4-like) format to UCS2. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 4 #define MIN_NEEDED_TO 2 #define FROM_DIRECTION 1 #define FROM_LOOP internal_ucs2_loop #define TO_LOOP internal_ucs2_loop /* This is not used. */ #define FUNCTION_NAME __gconv_transform_internal_ucs2 #define ONE_DIRECTION 1 #define MIN_NEEDED_INPUT MIN_NEEDED_FROM #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO #define LOOPFCT FROM_LOOP #define BODY \ { \ if (*((uint32_t *) inptr) >= 0x10000) \ { \ result = __GCONV_ILLEGAL_INPUT; \ break; \ } \ *((uint16_t *) outptr)++ = *((uint32_t *) inptr)++; \ } #include #include /* Convert from UCS2 in other endianness to the internal (UCS4-like) format. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 2 #define MIN_NEEDED_TO 4 #define FROM_DIRECTION 1 #define FROM_LOOP ucs2reverse_internal_loop #define TO_LOOP ucs2reverse_internal_loop/* This is not used.*/ #define FUNCTION_NAME __gconv_transform_ucs2reverse_internal #define ONE_DIRECTION 1 #define MIN_NEEDED_INPUT MIN_NEEDED_FROM #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO #define LOOPFCT FROM_LOOP #define BODY \ *((uint32_t *) outptr)++ = bswap_16 (*(uint16_t *) inptr); \ inptr += 2; #include #include /* Convert from the internal (UCS4-like) format to UCS2 in other endianness. */ #define DEFINE_INIT 0 #define DEFINE_FINI 0 #define MIN_NEEDED_FROM 4 #define MIN_NEEDED_TO 2 #define FROM_DIRECTION 1 #define FROM_LOOP internal_ucs2reverse_loop #define TO_LOOP internal_ucs2reverse_loop/* This is not used.*/ #define FUNCTION_NAME __gconv_transform_internal_ucs2reverse #define ONE_DIRECTION 1 #define MIN_NEEDED_INPUT MIN_NEEDED_FROM #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO #define LOOPFCT FROM_LOOP #define BODY \ { \ uint32_t val = *((uint32_t *) inptr); \ if (val >= 0x10000) \ { \ result = __GCONV_ILLEGAL_INPUT; \ break; \ } \ *((uint16_t *) outptr)++ = bswap_16 (val); \ inptr += 4; \ } #include #include