/* Extended regular expression matching and search library.
- Copyright (C) 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
+ Copyright (C) 2002-2014 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
- License along with the GNU C Library; if not, write to the Free
- Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
- 02111-1307 USA. */
+ License along with the GNU C Library; if not, see
+ <http://www.gnu.org/licenses/>. */
static void re_string_construct_common (const char *str, int len,
re_string_t *pstr,
RE_TRANSLATE_TYPE trans, int icase,
const re_dfa_t *dfa) internal_function;
-#ifdef RE_ENABLE_I18N
-static int re_string_skip_chars (re_string_t *pstr, int new_raw_idx,
- wint_t *last_wc) internal_function;
-#endif /* RE_ENABLE_I18N */
-static reg_errcode_t register_state (const re_dfa_t *dfa,
- re_dfastate_t *newstate,
- unsigned int hash) internal_function;
static re_dfastate_t *create_ci_newstate (const re_dfa_t *dfa,
const re_node_set *nodes,
unsigned int hash) internal_function;
const re_node_set *nodes,
unsigned int context,
unsigned int hash) internal_function;
-static inline unsigned int calc_state_hash (const re_node_set *nodes,
- unsigned int context) internal_function;
\f
/* Functions for string operation. */
re_string_reconstruct before using the object. */
static reg_errcode_t
-re_string_allocate (pstr, str, len, init_len, trans, icase, dfa)
- re_string_t *pstr;
- const char *str;
- int len, init_len, icase;
- RE_TRANSLATE_TYPE trans;
- const re_dfa_t *dfa;
+internal_function __attribute_warn_unused_result__
+re_string_allocate (re_string_t *pstr, const char *str, int len, int init_len,
+ RE_TRANSLATE_TYPE trans, int icase, const re_dfa_t *dfa)
{
reg_errcode_t ret;
int init_buf_len;
/* This function allocate the buffers, and initialize them. */
static reg_errcode_t
-re_string_construct (pstr, str, len, trans, icase, dfa)
- re_string_t *pstr;
- const char *str;
- int len, icase;
- RE_TRANSLATE_TYPE trans;
- const re_dfa_t *dfa;
+internal_function __attribute_warn_unused_result__
+re_string_construct (re_string_t *pstr, const char *str, int len,
+ RE_TRANSLATE_TYPE trans, int icase, const re_dfa_t *dfa)
{
reg_errcode_t ret;
memset (pstr, '\0', sizeof (re_string_t));
/* Helper functions for re_string_allocate, and re_string_construct. */
static reg_errcode_t
-re_string_realloc_buffers (pstr, new_buf_len)
- re_string_t *pstr;
- int new_buf_len;
+internal_function __attribute_warn_unused_result__
+re_string_realloc_buffers (re_string_t *pstr, int new_buf_len)
{
#ifdef RE_ENABLE_I18N
if (pstr->mb_cur_max > 1)
{
- wint_t *new_wcs = re_realloc (pstr->wcs, wint_t, new_buf_len);
+ wint_t *new_wcs;
+
+ /* Avoid overflow in realloc. */
+ const size_t max_object_size = MAX (sizeof (wint_t), sizeof (int));
+ 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;
static void
-re_string_construct_common (str, len, pstr, trans, icase, dfa)
- const char *str;
- int len;
- re_string_t *pstr;
- RE_TRANSLATE_TYPE trans;
- int icase;
- const re_dfa_t *dfa;
+internal_function
+re_string_construct_common (const char *str, int len, re_string_t *pstr,
+ RE_TRANSLATE_TYPE trans, int icase,
+ const re_dfa_t *dfa)
{
pstr->raw_mbs = (const unsigned char *) str;
pstr->len = len;
pstr->raw_len = len;
- pstr->trans = (unsigned RE_TRANSLATE_TYPE) trans;
+ pstr->trans = trans;
pstr->icase = icase ? 1 : 0;
pstr->mbs_allocated = (trans != NULL || icase);
pstr->mb_cur_max = dfa->mb_cur_max;
built and starts from PSTR->VALID_LEN. */
static void
-build_wcs_buffer (pstr)
- re_string_t *pstr;
+internal_function
+build_wcs_buffer (re_string_t *pstr)
{
#ifdef _LIBC
unsigned char buf[MB_LEN_MAX];
}
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))
+ mbclen = __mbrtowc (&wc, p, remain_len, &pstr->cur_state);
+ if (BE (mbclen == (size_t) -1 || mbclen == 0
+ || (mbclen == (size_t) -2 && pstr->bufs_len >= pstr->len), 0))
{
/* We treat these cases as a singlebyte character. */
mbclen = 1;
wc = pstr->trans[wc];
pstr->cur_state = prev_st;
}
+ else if (BE (mbclen == (size_t) -2, 0))
+ {
+ /* The buffer doesn't have enough space, finish to build. */
+ pstr->cur_state = prev_st;
+ break;
+ }
/* Write wide character and padding. */
pstr->wcs[byte_idx++] = wc;
/* Build wide character buffer PSTR->WCS like build_wcs_buffer,
but for REG_ICASE. */
-static int
-build_wcs_upper_buffer (pstr)
- re_string_t *pstr;
+static reg_errcode_t
+internal_function __attribute_warn_unused_result__
+build_wcs_upper_buffer (re_string_t *pstr)
{
mbstate_t prev_st;
int src_idx, byte_idx, end_idx, remain_len;
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);
+ mbclen = __mbrtowc (&wc,
+ ((const char *) pstr->raw_mbs + pstr->raw_mbs_idx
+ + byte_idx), remain_len, &pstr->cur_state);
if (BE (mbclen + 2 > 2, 1))
{
wchar_t wcu = wc;
for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;)
pstr->wcs[byte_idx++] = WEOF;
}
- else if (mbclen == (size_t) -1 || mbclen == 0)
+ else if (mbclen == (size_t) -1 || mbclen == 0
+ || (mbclen == (size_t) -2 && pstr->bufs_len >= pstr->len))
{
- /* It is an invalid character or '\0'. Just use the byte. */
+ /* It is an invalid character, an incomplete character
+ at the end of the string, 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. */
}
else
p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + src_idx;
- mbclen = mbrtowc (&wc, p, remain_len, &pstr->cur_state);
+ mbclen = __mbrtowc (&wc, p, remain_len, &pstr->cur_state);
if (BE (mbclen + 2 > 2, 1))
{
wchar_t wcu = wc;
src_idx += mbclen;
continue;
}
- else
- memcpy (pstr->mbs + byte_idx, p, mbclen);
+ else
+ memcpy (pstr->mbs + byte_idx, p, mbclen);
}
else
memcpy (pstr->mbs + byte_idx, p, mbclen);
for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;)
pstr->wcs[byte_idx++] = WEOF;
}
- else if (mbclen == (size_t) -1 || mbclen == 0)
+ else if (mbclen == (size_t) -1 || mbclen == 0
+ || (mbclen == (size_t) -2 && pstr->bufs_len >= pstr->len))
{
/* It is an invalid character or '\0'. Just use the byte. */
int ch = pstr->raw_mbs[pstr->raw_mbs_idx + src_idx];
Return the index. */
static int
-re_string_skip_chars (pstr, new_raw_idx, last_wc)
- re_string_t *pstr;
- int new_raw_idx;
- wint_t *last_wc;
+internal_function
+re_string_skip_chars (re_string_t *pstr, int new_raw_idx, wint_t *last_wc)
{
mbstate_t prev_st;
int rawbuf_idx;
size_t mbclen;
- wchar_t wc = 0;
+ wint_t wc = WEOF;
/* Skip the characters which are not necessary to check. */
for (rawbuf_idx = pstr->raw_mbs_idx + pstr->valid_raw_len;
rawbuf_idx < new_raw_idx;)
{
- int remain_len;
- remain_len = pstr->len - rawbuf_idx;
+ wchar_t wc2;
+ int remain_len = pstr->raw_len - rawbuf_idx;
prev_st = pstr->cur_state;
- mbclen = mbrtowc (&wc, (const char *) pstr->raw_mbs + rawbuf_idx,
- remain_len, &pstr->cur_state);
- if (BE (mbclen == (size_t) -2 || mbclen == (size_t) -1 || mbclen == 0, 0))
+ mbclen = __mbrtowc (&wc2, (const char *) pstr->raw_mbs + rawbuf_idx,
+ remain_len, &pstr->cur_state);
+ if (BE ((ssize_t) mbclen <= 0, 0))
{
- /* We treat these cases as a singlebyte character. */
+ /* We treat these cases as a single byte character. */
+ if (mbclen == 0 || remain_len == 0)
+ wc = L'\0';
+ else
+ wc = *(unsigned char *) (pstr->raw_mbs + rawbuf_idx);
mbclen = 1;
pstr->cur_state = prev_st;
}
+ else
+ wc = (wint_t) wc2;
/* Then proceed the next character. */
rawbuf_idx += mbclen;
}
- *last_wc = (wint_t) wc;
+ *last_wc = wc;
return rawbuf_idx;
}
#endif /* RE_ENABLE_I18N */
This function is used in case of REG_ICASE. */
static void
-build_upper_buffer (pstr)
- re_string_t *pstr;
+internal_function
+build_upper_buffer (re_string_t *pstr)
{
int char_idx, end_idx;
end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len;
/* Apply TRANS to the buffer in PSTR. */
static void
-re_string_translate_buffer (pstr)
- re_string_t *pstr;
+internal_function
+re_string_translate_buffer (re_string_t *pstr)
{
int buf_idx, end_idx;
end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len;
convert to upper case in case of REG_ICASE, apply translation. */
static reg_errcode_t
-re_string_reconstruct (pstr, idx, eflags)
- re_string_t *pstr;
- int idx, eflags;
+internal_function __attribute_warn_unused_result__
+re_string_reconstruct (re_string_t *pstr, int idx, int eflags)
{
int offset = idx - pstr->raw_mbs_idx;
if (BE (offset < 0, 0))
if (BE (offset != 0, 1))
{
- /* Are the characters which are already checked remain? */
- if (BE (offset < pstr->valid_raw_len, 1)
-#ifdef RE_ENABLE_I18N
- /* Handling this would enlarge the code too much.
- Accept a slowdown in that case. */
- && pstr->offsets_needed == 0
-#endif
- )
+ /* Should the already checked characters be kept? */
+ if (BE (offset < pstr->valid_raw_len, 1))
{
/* Yes, move them to the front of the buffer. */
- pstr->tip_context = re_string_context_at (pstr, offset - 1, eflags);
#ifdef RE_ENABLE_I18N
- if (pstr->mb_cur_max > 1)
- memmove (pstr->wcs, pstr->wcs + offset,
- (pstr->valid_len - offset) * sizeof (wint_t));
+ if (BE (pstr->offsets_needed, 0))
+ {
+ int low = 0, high = pstr->valid_len, mid;
+ do
+ {
+ mid = (high + low) / 2;
+ if (pstr->offsets[mid] > offset)
+ high = mid;
+ else if (pstr->offsets[mid] < offset)
+ low = mid + 1;
+ else
+ break;
+ }
+ while (low < high);
+ if (pstr->offsets[mid] < offset)
+ ++mid;
+ pstr->tip_context = re_string_context_at (pstr, mid - 1,
+ eflags);
+ /* This can be quite complicated, so handle specially
+ only the common and easy case where the character with
+ different length representation of lower and upper
+ case is present at or after offset. */
+ if (pstr->valid_len > offset
+ && mid == offset && pstr->offsets[mid] == offset)
+ {
+ memmove (pstr->wcs, pstr->wcs + offset,
+ (pstr->valid_len - offset) * sizeof (wint_t));
+ memmove (pstr->mbs, pstr->mbs + offset, pstr->valid_len - offset);
+ pstr->valid_len -= offset;
+ pstr->valid_raw_len -= offset;
+ for (low = 0; low < pstr->valid_len; low++)
+ pstr->offsets[low] = pstr->offsets[low + offset] - offset;
+ }
+ else
+ {
+ /* Otherwise, just find out how long the partial multibyte
+ character at offset is and fill it with WEOF/255. */
+ pstr->len = pstr->raw_len - idx + offset;
+ pstr->stop = pstr->raw_stop - idx + offset;
+ pstr->offsets_needed = 0;
+ while (mid > 0 && pstr->offsets[mid - 1] == offset)
+ --mid;
+ while (mid < pstr->valid_len)
+ if (pstr->wcs[mid] != WEOF)
+ break;
+ else
+ ++mid;
+ if (mid == pstr->valid_len)
+ pstr->valid_len = 0;
+ else
+ {
+ pstr->valid_len = pstr->offsets[mid] - offset;
+ if (pstr->valid_len)
+ {
+ for (low = 0; low < pstr->valid_len; ++low)
+ pstr->wcs[low] = WEOF;
+ memset (pstr->mbs, 255, pstr->valid_len);
+ }
+ }
+ pstr->valid_raw_len = pstr->valid_len;
+ }
+ }
+ else
+#endif
+ {
+ pstr->tip_context = re_string_context_at (pstr, offset - 1,
+ eflags);
+#ifdef RE_ENABLE_I18N
+ if (pstr->mb_cur_max > 1)
+ memmove (pstr->wcs, pstr->wcs + offset,
+ (pstr->valid_len - offset) * sizeof (wint_t));
#endif /* RE_ENABLE_I18N */
- if (BE (pstr->mbs_allocated, 0))
- memmove (pstr->mbs, pstr->mbs + offset,
- pstr->valid_len - offset);
- pstr->valid_len -= offset;
- pstr->valid_raw_len -= offset;
+ if (BE (pstr->mbs_allocated, 0))
+ memmove (pstr->mbs, pstr->mbs + offset,
+ pstr->valid_len - offset);
+ pstr->valid_len -= offset;
+ pstr->valid_raw_len -= offset;
#if DEBUG
- assert (pstr->valid_len > 0);
+ assert (pstr->valid_len > 0);
#endif
+ }
}
else
{
/* No, skip all characters until IDX. */
+ int prev_valid_len = pstr->valid_len;
+
#ifdef RE_ENABLE_I18N
if (BE (pstr->offsets_needed, 0))
{
}
#endif
pstr->valid_len = 0;
- pstr->valid_raw_len = 0;
#ifdef RE_ENABLE_I18N
if (pstr->mb_cur_max > 1)
{
if (pstr->is_utf8)
{
- const unsigned char *raw, *p, *q, *end;
+ const unsigned char *raw, *p, *end;
/* Special case UTF-8. Multi-byte chars start with any
byte other than 0x80 - 0xbf. */
raw = pstr->raw_mbs + pstr->raw_mbs_idx;
end = raw + (offset - pstr->mb_cur_max);
+ if (end < pstr->raw_mbs)
+ end = pstr->raw_mbs;
p = raw + offset - 1;
#ifdef _LIBC
/* We know the wchar_t encoding is UCS4, so for the simple
if (isascii (*p) && BE (pstr->trans == NULL, 1))
{
memset (&pstr->cur_state, '\0', sizeof (mbstate_t));
- pstr->valid_len = 0;
+ /* pstr->valid_len = 0; */
wc = (wchar_t) *p;
}
else
unsigned char buf[6];
size_t mbclen;
- q = p;
+ const unsigned char *pp = p;
if (BE (pstr->trans != NULL, 0))
{
int i = mlen < 6 ? mlen : 6;
while (--i >= 0)
buf[i] = pstr->trans[p[i]];
- q = buf;
+ pp = buf;
}
/* XXX Don't use mbrtowc, we know which conversion
to use (UTF-8 -> UCS4). */
memset (&cur_state, 0, sizeof (cur_state));
- mbclen = mbrtowc (&wc2, (const char *) p, mlen,
- &cur_state);
+ mbclen = __mbrtowc (&wc2, (const char *) pp, mlen,
+ &cur_state);
if (raw + offset - p <= mbclen
&& mbclen < (size_t) -2)
{
if (wc == WEOF)
pstr->valid_len = re_string_skip_chars (pstr, idx, &wc) - idx;
+ if (wc == WEOF)
+ pstr->tip_context
+ = re_string_context_at (pstr, prev_valid_len - 1, eflags);
+ else
+ pstr->tip_context = ((BE (pstr->word_ops_used != 0, 0)
+ && IS_WIDE_WORD_CHAR (wc))
+ ? CONTEXT_WORD
+ : ((IS_WIDE_NEWLINE (wc)
+ && pstr->newline_anchor)
+ ? CONTEXT_NEWLINE : 0));
if (BE (pstr->valid_len, 0))
{
for (wcs_idx = 0; wcs_idx < pstr->valid_len; ++wcs_idx)
memset (pstr->mbs, 255, pstr->valid_len);
}
pstr->valid_raw_len = pstr->valid_len;
- pstr->tip_context = ((BE (pstr->word_ops_used != 0, 0)
- && IS_WIDE_WORD_CHAR (wc))
- ? CONTEXT_WORD
- : ((IS_WIDE_NEWLINE (wc)
- && pstr->newline_anchor)
- ? CONTEXT_NEWLINE : 0));
}
else
#endif /* RE_ENABLE_I18N */
{
int c = pstr->raw_mbs[pstr->raw_mbs_idx + offset - 1];
+ pstr->valid_raw_len = 0;
if (pstr->trans)
c = pstr->trans[c];
pstr->tip_context = (bitset_contain (pstr->word_char, c)
{
if (pstr->icase)
{
- int ret = build_wcs_upper_buffer (pstr);
+ reg_errcode_t ret = build_wcs_upper_buffer (pstr);
if (BE (ret != REG_NOERROR, 0))
return ret;
}
}
static unsigned char
-re_string_peek_byte_case (pstr, idx)
- const re_string_t *pstr;
- int idx;
+internal_function __attribute ((pure))
+re_string_peek_byte_case (const re_string_t *pstr, int idx)
{
int ch, off;
}
static unsigned char
-re_string_fetch_byte_case (pstr)
- re_string_t *pstr;
+internal_function
+re_string_fetch_byte_case (re_string_t *pstr)
{
if (BE (!pstr->mbs_allocated, 1))
return re_string_fetch_byte (pstr);
}
static void
-re_string_destruct (pstr)
- re_string_t *pstr;
+internal_function
+re_string_destruct (re_string_t *pstr)
{
#ifdef RE_ENABLE_I18N
re_free (pstr->wcs);
/* Return the context at IDX in INPUT. */
static unsigned int
-re_string_context_at (input, idx, eflags)
- const re_string_t *input;
- int idx, eflags;
+internal_function
+re_string_context_at (const re_string_t *input, int idx, int eflags)
{
int c;
if (BE (idx < 0, 0))
/* Functions for set operation. */
static reg_errcode_t
-re_node_set_alloc (set, size)
- re_node_set *set;
- int size;
+internal_function __attribute_warn_unused_result__
+re_node_set_alloc (re_node_set *set, int size)
{
set->alloc = size;
set->nelem = 0;
}
static reg_errcode_t
-re_node_set_init_1 (set, elem)
- re_node_set *set;
- int elem;
+internal_function __attribute_warn_unused_result__
+re_node_set_init_1 (re_node_set *set, int elem)
{
set->alloc = 1;
set->nelem = 1;
}
static reg_errcode_t
-re_node_set_init_2 (set, elem1, elem2)
- re_node_set *set;
- int elem1, elem2;
+internal_function __attribute_warn_unused_result__
+re_node_set_init_2 (re_node_set *set, int elem1, int elem2)
{
set->alloc = 2;
set->elems = re_malloc (int, 2);
}
static reg_errcode_t
-re_node_set_init_copy (dest, src)
- re_node_set *dest;
- const re_node_set *src;
+internal_function __attribute_warn_unused_result__
+re_node_set_init_copy (re_node_set *dest, const re_node_set *src)
{
dest->nelem = src->nelem;
if (src->nelem > 0)
Note: We assume dest->elems is NULL, when dest->alloc is 0. */
static reg_errcode_t
-re_node_set_add_intersect (dest, src1, src2)
- re_node_set *dest;
- const re_node_set *src1, *src2;
+internal_function __attribute_warn_unused_result__
+re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1,
+ const re_node_set *src2)
{
int i1, i2, is, id, delta, sbase;
if (src1->nelem == 0 || src2->nelem == 0)
int new_alloc = src1->nelem + src2->nelem + dest->alloc;
int *new_elems = re_realloc (dest->elems, int, new_alloc);
if (BE (new_elems == NULL, 0))
- return REG_ESPACE;
+ return REG_ESPACE;
dest->elems = new_elems;
dest->alloc = new_alloc;
}
while (id >= 0 && dest->elems[id] > src1->elems[i1])
--id;
- if (id < 0 || dest->elems[id] != src1->elems[i1])
- dest->elems[--sbase] = src1->elems[i1];
+ if (id < 0 || dest->elems[id] != src1->elems[i1])
+ dest->elems[--sbase] = src1->elems[i1];
if (--i1 < 0 || --i2 < 0)
break;
if (delta > 0 && id >= 0)
for (;;)
{
- if (dest->elems[is] > dest->elems[id])
- {
- /* Copy from the top. */
- dest->elems[id + delta--] = dest->elems[is--];
- if (delta == 0)
- break;
- }
- else
- {
- /* Slide from the bottom. */
- dest->elems[id + delta] = dest->elems[id];
- if (--id < 0)
- break;
- }
+ if (dest->elems[is] > dest->elems[id])
+ {
+ /* Copy from the top. */
+ dest->elems[id + delta--] = dest->elems[is--];
+ if (delta == 0)
+ break;
+ }
+ else
+ {
+ /* Slide from the bottom. */
+ dest->elems[id + delta] = dest->elems[id];
+ if (--id < 0)
+ break;
+ }
}
/* Copy remaining SRC elements. */
DEST. Return value indicate the error code or REG_NOERROR if succeeded. */
static reg_errcode_t
-re_node_set_init_union (dest, src1, src2)
- re_node_set *dest;
- const re_node_set *src1, *src2;
+internal_function __attribute_warn_unused_result__
+re_node_set_init_union (re_node_set *dest, const re_node_set *src1,
+ const re_node_set *src2)
{
int i1, i2, id;
if (src1 != NULL && src1->nelem > 0 && src2 != NULL && src2->nelem > 0)
DEST. Return value indicate the error code or REG_NOERROR if succeeded. */
static reg_errcode_t
-re_node_set_merge (dest, src)
- re_node_set *dest;
- const re_node_set *src;
+internal_function __attribute_warn_unused_result__
+re_node_set_merge (re_node_set *dest, const re_node_set *src)
{
int is, id, sbase, delta;
if (src == NULL || src->nelem == 0)
is = src->nelem - 1, id = dest->nelem - 1; is >= 0 && id >= 0; )
{
if (dest->elems[id] == src->elems[is])
- is--, id--;
+ is--, id--;
else if (dest->elems[id] < src->elems[is])
- dest->elems[--sbase] = src->elems[is--];
+ dest->elems[--sbase] = src->elems[is--];
else /* if (dest->elems[id] > src->elems[is]) */
- --id;
+ --id;
}
if (is >= 0)
for (;;)
{
if (dest->elems[is] > dest->elems[id])
- {
+ {
/* Copy from the top. */
- dest->elems[id + delta--] = dest->elems[is--];
+ dest->elems[id + delta--] = dest->elems[is--];
if (delta == 0)
break;
}
else
- {
- /* Slide from the bottom. */
- dest->elems[id + delta] = dest->elems[id];
+ {
+ /* Slide from the bottom. */
+ dest->elems[id + delta] = dest->elems[id];
if (--id < 0)
{
/* Copy remaining SRC elements. */
memcpy (dest->elems, dest->elems + sbase,
- delta * sizeof (int));
+ delta * sizeof (int));
break;
}
}
return -1 if an error is occured, return 1 otherwise. */
static int
-re_node_set_insert (set, elem)
- re_node_set *set;
- int elem;
+internal_function __attribute_warn_unused_result__
+re_node_set_insert (re_node_set *set, int elem)
{
int idx;
/* In case the set is empty. */
{
idx = 0;
for (idx = set->nelem; idx > 0; idx--)
- set->elems[idx] = set->elems[idx - 1];
+ set->elems[idx] = set->elems[idx - 1];
}
else
{
for (idx = set->nelem; set->elems[idx - 1] > elem; idx--)
- set->elems[idx] = set->elems[idx - 1];
+ set->elems[idx] = set->elems[idx - 1];
}
/* Insert the new element. */
Return -1 if an error is occured, return 1 otherwise. */
static int
-re_node_set_insert_last (set, elem)
- re_node_set *set;
- int elem;
+internal_function __attribute_warn_unused_result__
+re_node_set_insert_last (re_node_set *set, int elem)
{
/* Realloc if we need. */
if (set->alloc == set->nelem)
return 1 if SET1 and SET2 are equivalent, return 0 otherwise. */
static int
-re_node_set_compare (set1, set2)
- const re_node_set *set1, *set2;
+internal_function __attribute ((pure))
+re_node_set_compare (const re_node_set *set1, const re_node_set *set2)
{
int i;
if (set1 == NULL || set2 == NULL || set1->nelem != set2->nelem)
/* Return (idx + 1) if SET contains the element ELEM, return 0 otherwise. */
static int
-re_node_set_contains (set, elem)
- const re_node_set *set;
- int elem;
+internal_function __attribute ((pure))
+re_node_set_contains (const re_node_set *set, int elem)
{
unsigned int idx, right, mid;
if (set->nelem <= 0)
}
static void
-re_node_set_remove_at (set, idx)
- re_node_set *set;
- int idx;
+internal_function
+re_node_set_remove_at (re_node_set *set, int idx)
{
if (idx < 0 || idx >= set->nelem)
return;
Or return -1, if an error will be occured. */
static int
-re_dfa_add_node (dfa, token)
- re_dfa_t *dfa;
- re_token_t token;
+internal_function
+re_dfa_add_node (re_dfa_t *dfa, re_token_t token)
{
int type = token.type;
if (BE (dfa->nodes_len >= dfa->nodes_alloc, 0))
re_node_set *new_edests, *new_eclosures;
re_token_t *new_nodes;
- /* Avoid overflows. */
- if (BE (new_nodes_alloc < dfa->nodes_alloc, 0))
+ /* Avoid overflows in realloc. */
+ const size_t max_object_size = MAX (sizeof (re_token_t),
+ MAX (sizeof (re_node_set),
+ sizeof (int)));
+ if (BE (SIZE_MAX / max_object_size < new_nodes_alloc, 0))
return -1;
new_nodes = re_realloc (dfa->nodes, re_token_t, new_nodes_alloc);
}
static inline unsigned int
-calc_state_hash (nodes, context)
- const re_node_set *nodes;
- unsigned int context;
+internal_function
+calc_state_hash (const re_node_set *nodes, unsigned int context)
{
unsigned int hash = nodes->nelem + context;
int i;
- We never return non-NULL value in case of any errors, it is for
optimization. */
-static re_dfastate_t*
-re_acquire_state (err, dfa, nodes)
- reg_errcode_t *err;
- const re_dfa_t *dfa;
- const re_node_set *nodes;
+static re_dfastate_t *
+internal_function __attribute_warn_unused_result__
+re_acquire_state (reg_errcode_t *err, const re_dfa_t *dfa,
+ const re_node_set *nodes)
{
unsigned int hash;
re_dfastate_t *new_state;
- We never return non-NULL value in case of any errors, it is for
optimization. */
-static re_dfastate_t*
-re_acquire_state_context (err, dfa, nodes, context)
- reg_errcode_t *err;
- const re_dfa_t *dfa;
- const re_node_set *nodes;
- unsigned int context;
+static re_dfastate_t *
+internal_function __attribute_warn_unused_result__
+re_acquire_state_context (reg_errcode_t *err, const re_dfa_t *dfa,
+ const re_node_set *nodes, unsigned int context)
{
unsigned int hash;
re_dfastate_t *new_state;
indicates the error code if failed. */
static reg_errcode_t
-register_state (dfa, newstate, hash)
- const re_dfa_t *dfa;
- re_dfastate_t *newstate;
- unsigned int hash;
+__attribute_warn_unused_result__
+register_state (const re_dfa_t *dfa, re_dfastate_t *newstate,
+ unsigned int hash)
{
struct re_state_table_entry *spot;
reg_errcode_t err;
{
int elem = newstate->nodes.elems[i];
if (!IS_EPSILON_NODE (dfa->nodes[elem].type))
- re_node_set_insert_last (&newstate->non_eps_nodes, elem);
+ if (re_node_set_insert_last (&newstate->non_eps_nodes, elem) < 0)
+ return REG_ESPACE;
}
spot = dfa->state_table + (hash & dfa->state_hash_mask);
return REG_NOERROR;
}
+static void
+free_state (re_dfastate_t *state)
+{
+ re_node_set_free (&state->non_eps_nodes);
+ re_node_set_free (&state->inveclosure);
+ if (state->entrance_nodes != &state->nodes)
+ {
+ re_node_set_free (state->entrance_nodes);
+ re_free (state->entrance_nodes);
+ }
+ re_node_set_free (&state->nodes);
+ re_free (state->word_trtable);
+ re_free (state->trtable);
+ re_free (state);
+}
+
/* Create the new state which is independ of contexts.
Return the new state if succeeded, otherwise return NULL. */
static re_dfastate_t *
-create_ci_newstate (dfa, nodes, hash)
- const re_dfa_t *dfa;
- const re_node_set *nodes;
- unsigned int hash;
+internal_function __attribute_warn_unused_result__
+create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes,
+ unsigned int hash)
{
int i;
reg_errcode_t err;
Return the new state if succeeded, otherwise return NULL. */
static re_dfastate_t *
-create_cd_newstate (dfa, nodes, context, hash)
- const re_dfa_t *dfa;
- const re_node_set *nodes;
- unsigned int context, hash;
+internal_function __attribute_warn_unused_result__
+create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes,
+ unsigned int context, unsigned int hash)
{
int i, nctx_nodes = 0;
reg_errcode_t err;
for (i = 0 ; i < nodes->nelem ; i++)
{
- unsigned int constraint = 0;
re_token_t *node = dfa->nodes + nodes->elems[i];
re_token_type_t type = node->type;
- if (node->constraint)
- constraint = node->constraint;
+ unsigned int constraint = node->constraint;
if (type == CHARACTER && !constraint)
continue;
newstate->halt = 1;
else if (type == OP_BACK_REF)
newstate->has_backref = 1;
- else if (type == ANCHOR)
- constraint = node->opr.ctx_type;
if (constraint)
{
free_state (newstate);
return NULL;
}
- re_node_set_init_copy (newstate->entrance_nodes, nodes);
+ if (re_node_set_init_copy (newstate->entrance_nodes, nodes)
+ != REG_NOERROR)
+ return NULL;
nctx_nodes = 0;
newstate->has_constraint = 1;
}
}
return newstate;
}
-
-static void
-free_state (state)
- re_dfastate_t *state;
-{
- re_node_set_free (&state->non_eps_nodes);
- re_node_set_free (&state->inveclosure);
- if (state->entrance_nodes != &state->nodes)
- {
- re_node_set_free (state->entrance_nodes);
- re_free (state->entrance_nodes);
- }
- re_node_set_free (&state->nodes);
- re_free (state->word_trtable);
- re_free (state->trtable);
- re_free (state);
-}