diff options
| -rw-r--r-- | ChangeLog | 8 | ||||
| -rw-r--r-- | iconv/gconv_builtin.h | 14 | ||||
| -rw-r--r-- | iconv/gconv_int.h | 2 | ||||
| -rw-r--r-- | iconv/gconv_simple.c | 172 |
4 files changed, 193 insertions, 3 deletions
@@ -1,3 +1,9 @@ +1999-04-24 Ulrich Drepper <drepper@cygnus.com> + + * iconv/gconv_builtin.h: Add definitions for UTF16 builtins. + * iconv/gconv_int.h: Declare UTF16 functions. + * iconv/gconv_simple.c: Add UTF16 conversion functions. + 1999-04-20 Andreas Jaeger <aj@arthur.rhein-neckar.de> * posix/wordexp.c (parse_param): Fix type of offset to allow it to @@ -37,7 +43,7 @@ 1999-04-17 Thorsten Kukuk <kukuk@suse.de> - * timezone/zic.c (dolink): Append complete path ot only filename + * timezone/zic.c (dolink): Append complete path not only filename to ../ list. 1999-04-17 Andreas Jaeger <aj@arthur.rhein-neckar.de> diff --git a/iconv/gconv_builtin.h b/iconv/gconv_builtin.h index e12f1e46ee..0c31890a37 100644 --- a/iconv/gconv_builtin.h +++ b/iconv/gconv_builtin.h @@ -1,5 +1,5 @@ /* Builtin transformations. - Copyright (C) 1997, 1998 Free Software Foundation, Inc. + Copyright (C) 1997, 1998, 1999 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997. @@ -79,3 +79,15 @@ BUILTIN_TRANSFORMATION (NULL, "INTERNAL", 8, "UNICODELITTLE//", 1, "=INTERNAL->ucs2little", __gconv_transform_internal_ucs2little, NULL, NULL, 4, 4, 2, 2) + +BUILTIN_ALIAS ("UTF-16//", "UTF16//") + +BUILTIN_TRANSFORMATION (NULL, "INTERNAL", 8, "UTF16//", + 1, "=INTERNAL->utf16", + __gconv_transform_internal_utf16, NULL, NULL, + 4, 4, 2, 4) + +BUILTIN_TRANSFORMATION (NULL, "UTF16//", 7, "INTERNAL", + 1, "=utf16->INTERNAL", + __gconv_transform_utf16_internal, NULL, NULL, + 2, 4, 4, 4) diff --git a/iconv/gconv_int.h b/iconv/gconv_int.h index d4ffa551d6..9b00e6522c 100644 --- a/iconv/gconv_int.h +++ b/iconv/gconv_int.h @@ -168,6 +168,8 @@ __BUILTIN_TRANS (__gconv_transform_internal_ucs2); __BUILTIN_TRANS (__gconv_transform_ucs2little_internal); __BUILTIN_TRANS (__gconv_transform_internal_ucs2little); __BUILTIN_TRANS (__gconv_transform_internal_ucs4); +__BUILTIN_TRANS (__gconv_transform_internal_utf16); +__BUILTIN_TRANS (__gconv_transform_utf16_internal); # undef __BUITLIN_TRANS #endif diff --git a/iconv/gconv_simple.c b/iconv/gconv_simple.c index 4084d04b44..74dbfc0356 100644 --- a/iconv/gconv_simple.c +++ b/iconv/gconv_simple.c @@ -1,5 +1,5 @@ /* Simple transformations functions. - Copyright (C) 1997, 1998 Free Software Foundation, Inc. + Copyright (C) 1997, 1998, 1999 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997. @@ -451,3 +451,173 @@ internal_ucs4_loop (const unsigned char **inptrp, const unsigned char *inend, #endif #include <iconv/loop.c> #include <iconv/skeleton.c> + + +/* Convert from the internal (UCS4-like) format to UTF-16. */ +#define DEFINE_INIT 0 +#define DEFINE_FINI 0 +#define MIN_NEEDED_FROM 4 +#define MIN_NEEDED_TO 2 +#define MAX_NEEDED_TO 4 +#define FROM_DIRECTION 1 +#define FROM_LOOP internal_utf16_loop +#define TO_LOOP internal_utf16_loop /* This is not used. */ +#define FUNCTION_NAME __gconv_transform_internal_utf16 + +#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 +#if __BYTE_ORDER == __LITTLE_ENDIAN +# define BODY \ + { \ + if (*((uint32_t *) inptr) >= 0x10000) \ + { \ + if (*((uint32_t *) inptr) >= 0x110000) \ + { \ + result = GCONV_ILLEGAL_INPUT; \ + break; \ + } \ + \ + /* Generate a surrogate character. */ \ + if (NEED_LENGTH_TEST && outptr + 4 > outend) \ + { \ + /* Overflow in the output buffer. */ \ + result = GCONV_FULL_OUTPUT; \ + break; \ + } \ + \ + *((uint16_t *) outptr)++ = bswap_16 (0xd7c0 \ + + (*((uint32_t *) inptr) >> 10));\ + *((uint16_t *) outptr)++ = bswap_16 (0xdc00 \ + + (*((uint32_t *) inptr) \ + & 0x3ff)); \ + } \ + else \ + /* Please note that we use the `uint32_t' from-pointer as an `uint16_t' \ + pointer which works since we are on a little endian machine. */ \ + *((uint16_t *) outptr)++ = bswap_16 (*((uint16_t *) inptr)); \ + inptr += 4; \ + } +#else +# define BODY \ + { \ + if (*((uint32_t *) inptr) >= 0x10000) \ + { \ + if (*((uint32_t *) inptr) >= 0x110000) \ + { \ + result = GCONV_ILLEGAL_INPUT; \ + break; \ + } \ + \ + /* Generate a surrogate character. */ \ + if (NEED_LENGTH_TEST && outptr + 4 > outend) \ + { \ + /* Overflow in the output buffer. */ \ + result = GCONV_FULL_OUTPUT; \ + break; \ + } \ + \ + *((uint16_t *) outptr)++ = 0xd7c0 + (*((uint32_t *) inptr) >> 10); \ + *((uint16_t *) outptr)++ = 0xdc00 + (*((uint32_t *) inptr) & 0x3ff); \ + } \ + else \ + *((uint16_t *) outptr)++ = *((uint32_t *) inptr)++; \ + } +#endif +#include <iconv/loop.c> +#include <iconv/skeleton.c> + + +/* Convert from UTF-16 to the internal (UCS4-like) format. */ +#define DEFINE_INIT 0 +#define DEFINE_FINI 0 +#define MIN_NEEDED_FROM 2 +#define MAX_NEEDED_FROM 4 +#define MIN_NEEDED_TO 4 +#define FROM_DIRECTION 1 +#define FROM_LOOP utf16_internal_loop +#define TO_LOOP utf16_internal_loop /* This is not used.*/ +#define FUNCTION_NAME __gconv_transform_utf16_internal + +#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 +#if __BYTE_ORDER == __LITTLE_ENDIAN +# define BODY \ + { \ + uint16_t u1 = bswap_16 (*(uint16_t *) inptr); \ + \ + if (u1 < 0xd800 || u1 > 0xdfff) \ + { \ + /* No surrogate. */ \ + *((uint32_t *) outptr)++ = u1; \ + inptr += 2; \ + } \ + else \ + { \ + uint16_t u2; \ + \ + /* It's a surrogate character. At least the first word says \ + it is. */ \ + if (NEED_LENGTH_TEST && inptr + 4 > inend) \ + { \ + /* We don't have enough input for another complete input \ + character. */ \ + result = GCONV_INCOMPLETE_INPUT; \ + break; \ + } \ + \ + u2 = bswap_16 (((uint16_t *) inptr)[1]); \ + if (u2 < 0xdc00 || u2 >= 0xdfff) \ + { \ + /* This is no valid second word for a surrogate. */ \ + result = GCONV_ILLEGAL_INPUT; \ + break; \ + } \ + \ + *((uint32_t *) outptr)++ = ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00); \ + inptr += 4; \ + } \ + } +#else +# define BODY \ + { \ + uint16_t u1 = *(uint16_t *) inptr; \ + \ + if (u1 < 0xd800 || u1 > 0xdfff) \ + { \ + /* No surrogate. */ \ + *((uint32_t *) outptr)++ = u1; \ + inptr += 2; \ + } \ + else \ + { \ + uint16_t u2; \ + \ + /* It's a surrogate character. At least the first word says \ + it is. */ \ + if (NEED_LENGTH_TEST && inptr + 4 > inend) \ + { \ + /* We don't have enough input for another complete input \ + character. */ \ + result = GCONV_INCOMPLETE_INPUT; \ + break; \ + } \ + \ + u2 = ((uint16_t *) inptr)[1]; \ + if (u2 < 0xdc00 || u2 >= 0xdfff) \ + { \ + /* This is no valid second word for a surrogate. */ \ + result = GCONV_ILLEGAL_INPUT; \ + break; \ + } \ + \ + *((uint32_t *) outptr)++ = ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00); \ + inptr += 4; \ + } \ + } +#endif +#include <iconv/loop.c> +#include <iconv/skeleton.c> |