regex_internal.c

gnu tar 源码包。 tar 软件是 Unix 系统下的一个打包软件

/* -*- buffer-read-only: t -*- vi: set ro: */
/* DO NOT EDIT! GENERATED AUTOMATICALLY! */
/* Extended regular expression matching and search library.
   Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007 Free Software
   Foundation, Inc.
   This file is part of the GNU C Library.
   Contributed by Isamu Hasegawa <[e-mail]>.

   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 3, or (at your option)
   any later version.

   This program 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 General Public License for more details.

   You should have received a copy of the GNU General Public License along
   with this program; if not, write to the Free Software Foundation,
   Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */

static void re_string_construct_common (const char *str, Idx len,
					re_string_t *pstr,
					RE_TRANSLATE_TYPE trans, bool icase,
					const re_dfa_t *dfa) internal_function;
static re_dfastate_t *create_ci_newstate (const re_dfa_t *dfa,
					  const re_node_set *nodes,
					  re_hashval_t hash) internal_function;
static re_dfastate_t *create_cd_newstate (const re_dfa_t *dfa,
					  const re_node_set *nodes,
					  unsigned int context,
					  re_hashval_t hash) internal_function;

/* Functions for string operation.  */

/* This function allocate the buffers.  It is necessary to call
   re_string_reconstruct before using the object.  */

static reg_errcode_t
internal_function
re_string_allocate (re_string_t *pstr, const char *str, Idx len, Idx init_len,
		    RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa)
{
  reg_errcode_t ret;
  Idx init_buf_len;

  /* Ensure at least one character fits into the buffers.  */
  if (init_len < dfa->mb_cur_max)
    init_len = dfa->mb_cur_max;
  init_buf_len = (len + 1 < init_len) ? len + 1: init_len;
  re_string_construct_common (str, len, pstr, trans, icase, dfa);

  ret = re_string_realloc_buffers (pstr, init_buf_len);
  if (BE (ret != REG_NOERROR, 0))
    return ret;

  pstr->word_char = dfa->word_char;
  pstr->word_ops_used = dfa->word_ops_used;
  pstr->mbs = pstr->mbs_allocated ? pstr->mbs : (unsigned char *) str;
  pstr->valid_len = (pstr->mbs_allocated || dfa->mb_cur_max > 1) ? 0 : len;
  pstr->valid_raw_len = pstr->valid_len;
  return REG_NOERROR;
}

/* This function allocate the buffers, and initialize them.  */

static reg_errcode_t
internal_function
re_string_construct (re_string_t *pstr, const char *str, Idx len,
		     RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa)
{
  reg_errcode_t ret;
  memset (pstr, '\0', sizeof (re_string_t));
  re_string_construct_common (str, len, pstr, trans, icase, dfa);

  if (len > 0)
    {
      ret = re_string_realloc_buffers (pstr, len + 1);
      if (BE (ret != REG_NOERROR, 0))
	return ret;
    }
  pstr->mbs = pstr->mbs_allocated ? pstr->mbs : (unsigned char *) str;

  if (icase)
    {
#ifdef RE_ENABLE_I18N
      if (dfa->mb_cur_max > 1)
	{
	  while (1)
	    {
	      ret = build_wcs_upper_buffer (pstr);
	      if (BE (ret != REG_NOERROR, 0))
		return ret;
	      if (pstr->valid_raw_len >= len)
		break;
	      if (pstr->bufs_len > pstr->valid_len + dfa->mb_cur_max)
		break;
	      ret = re_string_realloc_buffers (pstr, pstr->bufs_len * 2);
	      if (BE (ret != REG_NOERROR, 0))
		return ret;
	    }
	}
      else
#endif /* RE_ENABLE_I18N  */
	build_upper_buffer (pstr);
    }
  else
    {
#ifdef RE_ENABLE_I18N
      if (dfa->mb_cur_max > 1)
	build_wcs_buffer (pstr);
      else
#endif /* RE_ENABLE_I18N  */
	{
	  if (trans != NULL)
	    re_string_translate_buffer (pstr);
	  else
	    {
	      pstr->valid_len = pstr->bufs_len;
	      pstr->valid_raw_len = pstr->bufs_len;
	    }
	}
    }

  return REG_NOERROR;
}

/* Helper functions for re_string_allocate, and re_string_construct.  */

static reg_errcode_t
internal_function
re_string_realloc_buffers (re_string_t *pstr, Idx new_buf_len)
{
#ifdef RE_ENABLE_I18N
  if (pstr->mb_cur_max > 1)
    {
      wint_t *new_wcs;

      /* Avoid overflow.  */
      size_t max_object_size = MAX (sizeof (wint_t), sizeof (Idx));
      if (BE (SIZE_MAX / max_object_size < new_buf_len, 0))
	return REG_ESPACE;

      new_wcs = re_realloc (pstr->wcs, wint_t, new_buf_len);
      if (BE (new_wcs == NULL, 0))
	return REG_ESPACE;
      pstr->wcs = new_wcs;
      if (pstr->offsets != NULL)
	{
	  Idx *new_offsets = re_realloc (pstr->offsets, Idx, new_buf_len);
	  if (BE (new_offsets == NULL, 0))
	    return REG_ESPACE;
	  pstr->offsets = new_offsets;
	}
    }
#endif /* RE_ENABLE_I18N  */
  if (pstr->mbs_allocated)
    {
      unsigned char *new_mbs = re_realloc (pstr->mbs, unsigned char,
					   new_buf_len);
      if (BE (new_mbs == NULL, 0))
	return REG_ESPACE;
      pstr->mbs = new_mbs;
    }
  pstr->bufs_len = new_buf_len;
  return REG_NOERROR;
}


static void
internal_function
re_string_construct_common (const char *str, Idx len, re_string_t *pstr,
			    RE_TRANSLATE_TYPE trans, bool icase,
			    const re_dfa_t *dfa)
{
  pstr->raw_mbs = (const unsigned char *) str;
  pstr->len = len;
  pstr->raw_len = len;
  pstr->trans = trans;
  pstr->icase = icase;
  pstr->mbs_allocated = (trans != NULL || icase);
  pstr->mb_cur_max = dfa->mb_cur_max;
  pstr->is_utf8 = dfa->is_utf8;
  pstr->map_notascii = dfa->map_notascii;
  pstr->stop = pstr->len;
  pstr->raw_stop = pstr->stop;
}

#ifdef RE_ENABLE_I18N

/* Build wide character buffer PSTR->WCS.
   If the byte sequence of the string are:
     <mb1>(0), <mb1>(1), <mb2>(0), <mb2>(1), <sb3>
   Then wide character buffer will be:
     <wc1>   , WEOF    , <wc2>   , WEOF    , <wc3>
   We use WEOF for padding, they indicate that the position isn't
   a first byte of a multibyte character.

   Note that this function assumes PSTR->VALID_LEN elements are already
   built and starts from PSTR->VALID_LEN.  */

static void
internal_function
build_wcs_buffer (re_string_t *pstr)
{
#ifdef _LIBC
  unsigned char buf[MB_LEN_MAX];
  assert (MB_LEN_MAX >= pstr->mb_cur_max);
#else
  unsigned char buf[64];
#endif
  mbstate_t prev_st;
  Idx byte_idx, end_idx, remain_len;
  size_t mbclen;

  /* Build the buffers from pstr->valid_len to either pstr->len or
     pstr->bufs_len.  */
  end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len;
  for (byte_idx = pstr->valid_len; byte_idx < end_idx;)
    {
      wchar_t wc;
      const char *p;

      remain_len = end_idx - byte_idx;
      prev_st = pstr->cur_state;
      /* Apply the translation if we need.  */
      if (BE (pstr->trans != NULL, 0))
	{
	  int i, ch;

	  for (i = 0; i < pstr->mb_cur_max && i < remain_len; ++i)
	    {
	      ch = pstr->raw_mbs [pstr->raw_mbs_idx + byte_idx + i];
	      buf[i] = pstr->mbs[byte_idx + i] = pstr->trans[ch];
	    }
	  p = (const char *) buf;
	}
      else
	p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + byte_idx;
      mbclen = mbrtowc (&wc, p, remain_len, &pstr->cur_state);
      if (BE (mbclen == (size_t) -2, 0))
	{
	  /* The buffer doesn't have enough space, finish to build.  */
	  pstr->cur_state = prev_st;
	  break;
	}
      else if (BE (mbclen == (size_t) -1 || mbclen == 0, 0))
	{
	  /* We treat these cases as a singlebyte character.  */
	  mbclen = 1;
	  wc = (wchar_t) pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx];
	  if (BE (pstr->trans != NULL, 0))
	    wc = pstr->trans[wc];
	  pstr->cur_state = prev_st;
	}

      /* Write wide character and padding.  */
      pstr->wcs[byte_idx++] = wc;
      /* Write paddings.  */
      for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;)
	pstr->wcs[byte_idx++] = WEOF;
    }
  pstr->valid_len = byte_idx;
  pstr->valid_raw_len = byte_idx;
}

/* Build wide character buffer PSTR->WCS like build_wcs_buffer,
   but for REG_ICASE.  */

static reg_errcode_t
internal_function
build_wcs_upper_buffer (re_string_t *pstr)
{
  mbstate_t prev_st;
  Idx src_idx, byte_idx, end_idx, remain_len;
  size_t mbclen;
#ifdef _LIBC
  char buf[MB_LEN_MAX];
  assert (MB_LEN_MAX >= pstr->mb_cur_max);
#else
  char buf[64];
#endif

  byte_idx = pstr->valid_len;
  end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len;

  /* The following optimization assumes that ASCII characters can be
     mapped to wide characters with a simple cast.  */
  if (! pstr->map_notascii && pstr->trans == NULL && !pstr->offsets_needed)
    {
      while (byte_idx < end_idx)
	{
	  wchar_t wc;

	  if (isascii (pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx])
	      && mbsinit (&pstr->cur_state))
	    {
	      /* In case of a singlebyte character.  */
	      pstr->mbs[byte_idx]
		= toupper (pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]);
	      /* The next step uses the assumption that wchar_t is encoded
		 ASCII-safe: all ASCII values can be converted like this.  */
	      pstr->wcs[byte_idx] = (wchar_t) pstr->mbs[byte_idx];
	      ++byte_idx;
	      continue;
	    }

	  remain_len = end_idx - byte_idx;
	  prev_st = pstr->cur_state;
	  mbclen = mbrtowc (&wc,
			    ((const char *) pstr->raw_mbs + pstr->raw_mbs_idx
			     + byte_idx), remain_len, &pstr->cur_state);
	  if (BE (mbclen < (size_t) -2, 1))
	    {
	      wchar_t wcu = wc;
	      if (iswlower (wc))
		{
		  size_t mbcdlen;

		  wcu = towupper (wc);
		  mbcdlen = wcrtomb (buf, wcu, &prev_st);
		  if (BE (mbclen == mbcdlen, 1))
		    memcpy (pstr->mbs + byte_idx, buf, mbclen);
		  else
		    {
		      src_idx = byte_idx;
		      goto offsets_needed;
		    }
		}
	      else
		memcpy (pstr->mbs + byte_idx,
			pstr->raw_mbs + pstr->raw_mbs_idx + byte_idx, mbclen);
	      pstr->wcs[byte_idx++] = wcu;
	      /* Write paddings.  */
	      for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;)
		pstr->wcs[byte_idx++] = WEOF;
	    }
	  else if (mbclen == (size_t) -1 || mbclen == 0)
	    {
	      /* It is an invalid character or '\0'.  Just use the byte.  */
	      int ch = pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx];
	      pstr->mbs[byte_idx] = ch;
	      /* And also cast it to wide char.  */
	      pstr->wcs[byte_idx++] = (wchar_t) ch;
	      if (BE (mbclen == (size_t) -1, 0))
		pstr->cur_state = prev_st;
	    }
	  else
	    {
	      /* The buffer doesn't have enough space, finish to build.  */
	      pstr->cur_state = prev_st;
	      break;
	    }
	}
      pstr->valid_len = byte_idx;
      pstr->valid_raw_len = byte_idx;
      return REG_NOERROR;
    }
  else
    for (src_idx = pstr->valid_raw_len; byte_idx < end_idx;)
      {
	wchar_t wc;
	const char *p;
      offsets_needed:
	remain_len = end_idx - byte_idx;
	prev_st = pstr->cur_state;
	if (BE (pstr->trans != NULL, 0))
	  {
	    int i, ch;

	    for (i = 0; i < pstr->mb_cur_max && i < remain_len; ++i)
	      {
		ch = pstr->raw_mbs [pstr->raw_mbs_idx + src_idx + i];
		buf[i] = pstr->trans[ch];
	      }
	    p = (const char *) buf;
	  }
	else
	  p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + src_idx;
	mbclen = mbrtowc (&wc, p, remain_len, &pstr->cur_state);
	if (BE (mbclen < (size_t) -2, 1))
	  {
	    wchar_t wcu = wc;
	    if (iswlower (wc))
	      {
		size_t mbcdlen;

		wcu = towupper (wc);
		mbcdlen = wcrtomb ((char *) buf, wcu, &prev_st);
		if (BE (mbclen == mbcdlen, 1))
		  memcpy (pstr->mbs + byte_idx, buf, mbclen);
		else if (mbcdlen != (size_t) -1)
		  {
		    size_t i;

		    if (byte_idx + mbcdlen > pstr->bufs_len)
		      {
			pstr->cur_state = prev_st;
			break;
		      }

		    if (pstr->offsets == NULL)
		      {
			pstr->offsets = re_malloc (Idx, pstr->bufs_len);

			if (pstr->offsets == NULL)
			  return REG_ESPACE;
		      }
		    if (!pstr->offsets_needed)
		      {
			for (i = 0; i < (size_t) byte_idx; ++i)
			  pstr->offsets[i] = i;
			pstr->offsets_needed = 1;
		      }

		    memcpy (pstr->mbs + byte_idx, buf, mbcdlen);
		    pstr->wcs[byte_idx] = wcu;
		    pstr->offsets[byte_idx] = src_idx;
		    for (i = 1; i < mbcdlen; ++i)
		      {
			pstr->offsets[byte_idx + i]
			  = src_idx + (i < mbclen ? i : mbclen - 1);
			pstr->wcs[byte_idx + i] = WEOF;
		      }
		    pstr->len += mbcdlen - mbclen;
		    if (pstr->raw_stop > src_idx)
		      pstr->stop += mbcdlen - mbclen;
		    end_idx = (pstr->bufs_len > pstr->len)
			      ? pstr->len : pstr->bufs_len;
		    byte_idx += mbcdlen;
		    src_idx += mbclen;
		    continue;
		  }
                else
                  memcpy (pstr->mbs + byte_idx, p, mbclen);