gconv_simple.c

Newlib 嵌入式 C库 标准实现代码

# error "This endianess is not supported."
#endif

  /* Determine the status.  */
  if (*inptrp == inend)
    result = __GCONV_EMPTY_INPUT;
  else if (*outptrp + 4 > outend)
    result = __GCONV_FULL_OUTPUT;
  else
    result = __GCONV_INCOMPLETE_INPUT;

  return result;
}

#ifndef _STRING_ARCH_unaligned
static inline int
internal_ucs4le_loop_unaligned (struct __gconv_step *step,
				struct __gconv_step_data *step_data,
				const unsigned char **inptrp,
				const unsigned char *inend,
				unsigned char **outptrp, unsigned char *outend,
				size_t *irreversible)
{
  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 += 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 = memcpy (outptr, inptr, n_convert * 4);
  *outptrp += n_convert * 4;
# else
#  error "This endianess is not supported."
# endif

  /* Determine the status.  */
  if (*inptrp + 4 > inend)
    result = __GCONV_EMPTY_INPUT;
  else if (*outptrp + 4 > outend)
    result = __GCONV_FULL_OUTPUT;
  else
    result = __GCONV_INCOMPLETE_INPUT;

  return result;
}
#endif


static inline int
internal_ucs4le_loop_single (struct __gconv_step *step,
			     struct __gconv_step_data *step_data,
			     const unsigned char **inptrp,
			     const unsigned char *inend,
			     unsigned char **outptrp, unsigned char *outend,
			     size_t *irreversible)
{
  mbstate_t *state = step_data->__statep;
  size_t cnt = state->__count & 7;

  while (*inptrp < inend && cnt < 4)
    state->__value.__wchb[cnt++] = *(*inptrp)++;

  if (__builtin_expect (cnt < 4, 0))
    {
      /* 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];

  *outptrp += 4;
#else
  {
  /* XXX unaligned */
  uint32_t **p = (uint32_t **)outptrp;
  uint32_t *q = *p;
  *q = state->__value.__wch;
  outptrp = (unsigned char **)(p + 1);
  }
#endif

  /* Clear the state buffer.  */
  state->__count &= ~7;

  return __GCONV_OK;
}

#include <iconv/skeleton.c>


/* 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 (struct __gconv_step *step,
		      struct __gconv_step_data *step_data,
		      const unsigned char **inptrp, const unsigned char *inend,
		      unsigned char **outptrp, unsigned char *outend,
		      size_t *irreversible)
{
  int flags = step_data->__flags;
  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 (*(const uint32_t *) inptr);
#else
      inval = *(const uint32_t *) inptr;
#endif

      if (__builtin_expect (inval > 0x7fffffff, 0))
	{
	  /* The value is too large.  We don't try transliteration here since
	     this is not an error because of the lack of possibilities to
	     represent the result.  This is a genuine bug in the input since
	     UCS4 does not allow such values.  */
	  if (irreversible == NULL)
	    /* We are transliterating, don't try to correct anything.  */
	    return __GCONV_ILLEGAL_INPUT;

	  if (flags & __GCONV_IGNORE_ERRORS)
	    {
	      /* Just ignore this character.  */
	      ++*irreversible;
	      continue;
	    }

	  return __GCONV_ILLEGAL_INPUT;
	}

      *((uint32_t *) outptr) = inval;
      outptr = (unsigned char *)((uint32_t *) outptr + 1);
    }

  *inptrp = inptr;
  *outptrp = outptr;

  /* Determine the status.  */
  if (*inptrp == inend)
    result = __GCONV_EMPTY_INPUT;
  else if (*outptrp + 4 > outend)
    result = __GCONV_FULL_OUTPUT;
  else
    result = __GCONV_INCOMPLETE_INPUT;

  return result;
}

#ifndef _STRING_ARCH_unaligned
static inline int
ucs4le_internal_loop_unaligned (struct __gconv_step *step,
				struct __gconv_step_data *step_data,
				const unsigned char **inptrp,
				const unsigned char *inend,
				unsigned char **outptrp, unsigned char *outend,
				size_t *irreversible)
{
  int flags = step_data->__flags;
  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 (__builtin_expect (inptr[3] > 0x80, 0))
	{
	  /* The value is too large.  We don't try transliteration here since
	     this is not an error because of the lack of possibilities to
	     represent the result.  This is a genuine bug in the input since
	     UCS4 does not allow such values.  */
	  if (irreversible == NULL)
	    /* We are transliterating, don't try to correct anything.  */
	    return __GCONV_ILLEGAL_INPUT;

	  if (flags & __GCONV_IGNORE_ERRORS)
	    {
	      /* Just ignore this character.  */
	      ++*irreversible;
	      continue;
	    }

	  *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

      outptr += 4;
    }

  *inptrp = inptr;
  *outptrp = outptr;

  /* Determine the status.  */
  if (*inptrp == inend)
    result = __GCONV_EMPTY_INPUT;
  else if (*outptrp + 4 > outend)
    result = __GCONV_FULL_OUTPUT;
  else
    result = __GCONV_INCOMPLETE_INPUT;

  return result;
}
#endif


static inline int
ucs4le_internal_loop_single (struct __gconv_step *step,
			     struct __gconv_step_data *step_data,
			     const unsigned char **inptrp,
			     const unsigned char *inend,
			     unsigned char **outptrp, unsigned char *outend,
			     size_t *irreversible)
{
  mbstate_t *state = step_data->__statep;
  int flags = step_data->__flags;
  size_t cnt = state->__count & 7;

  while (*inptrp < inend && cnt < 4)
    state->__value.__wchb[cnt++] = *(*inptrp)++;

  if (__builtin_expect (cnt < 4, 0))
    {
      /* Still not enough bytes.  Store the ones in the input buffer.  */
      state->__count &= ~7;
      state->__count |= cnt;

      return __GCONV_INCOMPLETE_INPUT;
    }

  if (__builtin_expect (((unsigned char *) state->__value.__wchb)[3] > 0x80,
			0))
    {
      /* The value is too large.  We don't try transliteration here since
	 this is not an error because of the lack of possibilities to
	 represent the result.  This is a genuine bug in the input since
	 UCS4 does not allow such values.  */
      if (!(flags & __GCONV_IGNORE_ERRORS))
	return __GCONV_ILLEGAL_INPUT;
    }
  else
    {
#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

      *outptrp += 4;
    }

  /* Clear the state buffer.  */
  state->__count &= ~7;

  return __GCONV_OK;
}

#include <iconv/skeleton.c>


/* 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 (__builtin_expect (*inptr > '\x7f', 0))				      \
      {									      \
	/* The value is too large.  We don't try transliteration here since   \
	   this is not an error because of the lack of possibilities to	      \
	   represent the result.  This is a genuine bug in the input since    \
	   ASCII does not allow such values.  */			      \
	if (! ignore_errors_p ())					      \
	  {								      \
	    /* This is no correct ANSI_X3.4-1968 character.  */		      \
	    result = __GCONV_ILLEGAL_INPUT;				      \
	    break;							      \
	  }								      \
									      \
	*irreversible = *irreversible + 1;				      \
	++inptr;							      \
      }									      \
    else								      \
      { 								      \
        /* It's an one byte sequence.  */				      \
        *((uint32_t *) outptr) = *inptr;				      \
        ++inptr;							      \
        outptr = (unsigned char *)((uint32_t *) outptr + 1);		      \
      }									      \
  }
#define LOOP_NEED_FLAGS
#include <iconv/loop.c>
#include <iconv/skeleton.c>


/* 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 (__builtin_expect (*((const uint32_t *) inptr) > 0x7f, 0))	      \
      {									      \
	UNICODE_TAG_HANDLER (*((const uint32_t *) inptr), 4);		      \
	STANDARD_ERR_HANDLER (4);					      \
      }									      \
    else								      \
      { 								      \
        /* It's an one byte sequence.  */				      \
        *outptr++ = *((const uint32_t *) inptr);			      \
        inptr = ((const uint32_t *) inptr + 1);				      \
      }									      \
  }
#define LOOP_NEED_FLAGS
#include <iconv/loop.c>
#include <iconv/skeleton.c>


/* 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 = *((const uint32_t *) inptr);				      \
									      \
    if (wc < 0x80)							      \
      /* It's an one byte sequence.  */					      \
      *outptr++ = (unsigned char) wc;					      \
    else if (__builtin_expect (wc <= 0x7fffffff, 1))			      \
      {									      \
	size_t step;							      \
	char *start;							      \
									      \
	for (step = 2; step < 6; ++step)				      \
	  if ((wc & (~(uint32_t)0 << (5 * step + 1))) == 0)		      \
	    break;							      \
									      \
	if (__builtin_expect (outptr + step > outend, 0))		      \
	  {								      \
	    /* Too long.  */						      \
	    result = __GCONV_FULL_OUTPUT;				      \
	    break;							      \
	  }								      \
									      \
	start = outptr;							      \
	*outptr = (unsigned char) (~0xff >> step);			      \
	outptr += step;							      \
	--step;								      \
	do								      \
	  {								      \
	    start[step] = 0x80 | (wc & 0x3f);				      \
	    wc >>= 6;							      \
	  }								      \
	while (--step > 0);						      \
	start[0] |= wc;							      \
      }									      \
    else								      \
      {									      \
	STANDARD_ERR_HANDLER (4);					      \
      }									      \
									      \
    inptr += 4;								      \