X-Git-Url: http://erislabs.net/gitweb/?a=blobdiff_plain;f=lib%2Fregex_internal.c;h=0343ee6e36a0bbb597440467b33c88a77a36a542;hb=1276a2c5f24c0c932426aca9c899fa524d2443f2;hp=57ba44d10656b4b1292fa6ef91ac46a38241109a;hpb=9581d12ccba50fb818aa2ad5857a54a51a305993;p=gnulib.git diff --git a/lib/regex_internal.c b/lib/regex_internal.c index 57ba44d10..0343ee6e3 100644 --- a/lib/regex_internal.c +++ b/lib/regex_internal.c @@ -1,25 +1,25 @@ /* 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 . - 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 2, or (at your option) - any later version. + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. - This program is distributed in the hope that it will be useful, + 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 General Public License for more details. + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser 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. */ + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ static void re_string_construct_common (const char *str, Idx len, re_string_t *pstr, - REG_TRANSLATE_TYPE trans, bool icase, + 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, @@ -35,9 +35,9 @@ static re_dfastate_t *create_cd_newstate (const re_dfa_t *dfa, re_string_reconstruct before using the object. */ static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ re_string_allocate (re_string_t *pstr, const char *str, Idx len, Idx init_len, - REG_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa) + RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa) { reg_errcode_t ret; Idx init_buf_len; @@ -63,9 +63,9 @@ re_string_allocate (re_string_t *pstr, const char *str, Idx len, Idx init_len, /* This function allocate the buffers, and initialize them. */ static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ re_string_construct (re_string_t *pstr, const char *str, Idx len, - REG_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa) + RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa) { reg_errcode_t ret; memset (pstr, '\0', sizeof (re_string_t)); @@ -126,13 +126,20 @@ re_string_construct (re_string_t *pstr, const char *str, Idx len, /* Helper functions for re_string_allocate, and re_string_construct. */ static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ 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 = 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 (Idx)); + if (BE (MIN (IDX_MAX, 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; @@ -161,13 +168,13 @@ re_string_realloc_buffers (re_string_t *pstr, Idx new_buf_len) static void internal_function re_string_construct_common (const char *str, Idx len, re_string_t *pstr, - REG_TRANSLATE_TYPE trans, bool icase, + 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 = (unsigned REG_TRANSLATE_TYPE) trans; + pstr->trans = trans; pstr->icase = icase; pstr->mbs_allocated = (trans != NULL || icase); pstr->mb_cur_max = dfa->mb_cur_max; @@ -228,14 +235,9 @@ build_wcs_buffer (re_string_t *pstr) } 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; @@ -244,6 +246,12 @@ build_wcs_buffer (re_string_t *pstr) 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; @@ -259,7 +267,7 @@ build_wcs_buffer (re_string_t *pstr) but for REG_ICASE. */ static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ build_wcs_upper_buffer (re_string_t *pstr) { mbstate_t prev_st; @@ -298,10 +306,10 @@ build_wcs_upper_buffer (re_string_t *pstr) 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 ((size_t) (mbclen + 2) > 2, 1)) + 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)) @@ -326,9 +334,11 @@ build_wcs_upper_buffer (re_string_t *pstr) 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. */ @@ -368,8 +378,8 @@ build_wcs_upper_buffer (re_string_t *pstr) } else p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + src_idx; - mbclen = mbrtowc (&wc, p, remain_len, &pstr->cur_state); - if (BE ((size_t) (mbclen + 2) > 2, 1)) + mbclen = __mbrtowc (&wc, p, remain_len, &pstr->cur_state); + if (BE (mbclen < (size_t) -2, 1)) { wchar_t wcu = wc; if (iswlower (wc)) @@ -422,8 +432,8 @@ build_wcs_upper_buffer (re_string_t *pstr) 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); @@ -441,7 +451,8 @@ build_wcs_upper_buffer (re_string_t *pstr) 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]; @@ -481,27 +492,33 @@ re_string_skip_chars (re_string_t *pstr, Idx new_raw_idx, wint_t *last_wc) mbstate_t prev_st; Idx 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;) { - Idx remain_len; - remain_len = pstr->len - rawbuf_idx; + wchar_t wc2; + Idx 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); + mbclen = __mbrtowc (&wc2, (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)) { - /* 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 = wc2; /* Then proceed the next character. */ rawbuf_idx += mbclen; } - *last_wc = (wint_t) wc; + *last_wc = wc; return rawbuf_idx; } #endif /* RE_ENABLE_I18N */ @@ -554,7 +571,7 @@ re_string_translate_buffer (re_string_t *pstr) convert to upper case in case of REG_ICASE, apply translation. */ static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) { Idx offset; @@ -583,35 +600,99 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) 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)) + { + Idx 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. */ #ifdef RE_ENABLE_I18N + /* No, skip all characters until IDX. */ + Idx prev_valid_len = pstr->valid_len; + if (BE (pstr->offsets_needed, 0)) { pstr->len = pstr->raw_len - idx + offset; @@ -620,7 +701,6 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) } #endif pstr->valid_len = 0; - pstr->valid_raw_len = 0; #ifdef RE_ENABLE_I18N if (pstr->mb_cur_max > 1) { @@ -629,47 +709,72 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) 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); - for (p = raw + offset - 1; p >= end; --p) - if ((*p & 0xc0) != 0x80) - { - mbstate_t cur_state; - wchar_t wc2; - Idx mlen = raw + pstr->len - p; - unsigned char buf[6]; - size_t mbclen; - - q = p; - if (BE (pstr->trans != NULL, 0)) - { - int i = mlen < 6 ? mlen : 6; - while (--i >= 0) - buf[i] = pstr->trans[p[i]]; - q = 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); - if (raw + offset - p <= mbclen && mbclen < (size_t) -2) - { - memset (&pstr->cur_state, '\0', - sizeof (mbstate_t)); - pstr->valid_len = mbclen - (raw + offset - p); - wc = wc2; - } - break; - } + 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 + case, ASCII characters, skip the conversion step. */ + if (isascii (*p) && BE (pstr->trans == NULL, 1)) + { + memset (&pstr->cur_state, '\0', sizeof (mbstate_t)); + /* pstr->valid_len = 0; */ + wc = (wchar_t) *p; + } + else +#endif + for (; p >= end; --p) + if ((*p & 0xc0) != 0x80) + { + mbstate_t cur_state; + wchar_t wc2; + Idx mlen = raw + pstr->len - p; + unsigned char buf[6]; + size_t mbclen; + + 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]]; + 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 *) pp, mlen, + &cur_state); + if (raw + offset - p <= mbclen + && mbclen < (size_t) -2) + { + memset (&pstr->cur_state, '\0', + sizeof (mbstate_t)); + pstr->valid_len = mbclen - (raw + offset - p); + wc = wc2; + } + break; + } } 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) @@ -678,17 +783,12 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) 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) @@ -719,22 +819,22 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) } else #endif /* RE_ENABLE_I18N */ - if (BE (pstr->mbs_allocated, 0)) - { - if (pstr->icase) - build_upper_buffer (pstr); - else if (pstr->trans != NULL) - re_string_translate_buffer (pstr); - } - else - pstr->valid_len = pstr->len; + if (BE (pstr->mbs_allocated, 0)) + { + if (pstr->icase) + build_upper_buffer (pstr); + else if (pstr->trans != NULL) + re_string_translate_buffer (pstr); + } + else + pstr->valid_len = pstr->len; pstr->cur_idx = 0; return REG_NOERROR; } static unsigned char -internal_function __attribute ((pure)) +internal_function __attribute__ ((pure)) re_string_peek_byte_case (const re_string_t *pstr, Idx idx) { int ch; @@ -771,7 +871,7 @@ re_string_peek_byte_case (const re_string_t *pstr, Idx idx) } static unsigned char -internal_function __attribute ((pure)) +internal_function re_string_fetch_byte_case (re_string_t *pstr) { if (BE (!pstr->mbs_allocated, 1)) @@ -868,19 +968,19 @@ re_string_context_at (const re_string_t *input, Idx idx, int eflags) /* Functions for set operation. */ static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ re_node_set_alloc (re_node_set *set, Idx size) { set->alloc = size; set->nelem = 0; set->elems = re_malloc (Idx, size); - if (BE (set->elems == NULL, 0)) + if (BE (set->elems == NULL, 0) && (MALLOC_0_IS_NONNULL || size != 0)) return REG_ESPACE; return REG_NOERROR; } static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ re_node_set_init_1 (re_node_set *set, Idx elem) { set->alloc = 1; @@ -896,7 +996,7 @@ re_node_set_init_1 (re_node_set *set, Idx elem) } static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ re_node_set_init_2 (re_node_set *set, Idx elem1, Idx elem2) { set->alloc = 2; @@ -926,7 +1026,7 @@ re_node_set_init_2 (re_node_set *set, Idx elem1, Idx elem2) } static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ re_node_set_init_copy (re_node_set *dest, const re_node_set *src) { dest->nelem = src->nelem; @@ -939,7 +1039,7 @@ re_node_set_init_copy (re_node_set *dest, const re_node_set *src) dest->alloc = dest->nelem = 0; return REG_ESPACE; } - memcpy (dest->elems, src->elems, src->nelem * sizeof dest->elems[0]); + memcpy (dest->elems, src->elems, src->nelem * sizeof (Idx)); } else re_node_set_init_empty (dest); @@ -951,7 +1051,7 @@ re_node_set_init_copy (re_node_set *dest, const re_node_set *src) Note: We assume dest->elems is NULL, when dest->alloc is 0. */ static reg_errcode_t -internal_function +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) { @@ -966,7 +1066,7 @@ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1, Idx new_alloc = src1->nelem + src2->nelem + dest->alloc; Idx *new_elems = re_realloc (dest->elems, Idx, new_alloc); if (BE (new_elems == NULL, 0)) - return REG_ESPACE; + return REG_ESPACE; dest->elems = new_elems; dest->alloc = new_alloc; } @@ -1016,24 +1116,24 @@ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1, if (delta > 0 && REG_VALID_INDEX (id)) 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 (! REG_VALID_INDEX (--id)) - 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 (! REG_VALID_INDEX (--id)) + break; + } } /* Copy remaining SRC elements. */ - memcpy (dest->elems, dest->elems + sbase, delta * sizeof dest->elems[0]); + memcpy (dest->elems, dest->elems + sbase, delta * sizeof (Idx)); return REG_NOERROR; } @@ -1042,7 +1142,7 @@ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1, DEST. Return value indicate the error code or REG_NOERROR if succeeded. */ static reg_errcode_t -internal_function +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) { @@ -1078,13 +1178,13 @@ re_node_set_init_union (re_node_set *dest, const re_node_set *src1, if (i1 < src1->nelem) { memcpy (dest->elems + id, src1->elems + i1, - (src1->nelem - i1) * sizeof dest->elems[0]); + (src1->nelem - i1) * sizeof (Idx)); id += src1->nelem - i1; } else if (i2 < src2->nelem) { memcpy (dest->elems + id, src2->elems + i2, - (src2->nelem - i2) * sizeof dest->elems[0]); + (src2->nelem - i2) * sizeof (Idx)); id += src2->nelem - i2; } dest->nelem = id; @@ -1095,7 +1195,7 @@ re_node_set_init_union (re_node_set *dest, const re_node_set *src1, DEST. Return value indicate the error code or REG_NOERROR if succeeded. */ static reg_errcode_t -internal_function +internal_function __attribute_warn_unused_result__ re_node_set_merge (re_node_set *dest, const re_node_set *src) { Idx is, id, sbase, delta; @@ -1114,7 +1214,7 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src) if (BE (dest->nelem == 0, 0)) { dest->nelem = src->nelem; - memcpy (dest->elems, src->elems, src->nelem * sizeof dest->elems[0]); + memcpy (dest->elems, src->elems, src->nelem * sizeof (Idx)); return REG_NOERROR; } @@ -1125,19 +1225,18 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src) REG_VALID_INDEX (is) && REG_VALID_INDEX (id); ) { 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 (REG_VALID_INDEX (is)) { /* If DEST is exhausted, the remaining items of SRC must be unique. */ sbase -= is + 1; - memcpy (dest->elems + sbase, src->elems, - (is + 1) * sizeof dest->elems[0]); + memcpy (dest->elems + sbase, src->elems, (is + 1) * sizeof (Idx)); } id = dest->nelem - 1; @@ -1152,21 +1251,21 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src) 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 (! REG_VALID_INDEX (--id)) { /* Copy remaining SRC elements. */ memcpy (dest->elems, dest->elems + sbase, - delta * sizeof dest->elems[0]); + delta * sizeof (Idx)); break; } } @@ -1180,13 +1279,13 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src) Return true if successful. */ static bool -internal_function +internal_function __attribute_warn_unused_result__ re_node_set_insert (re_node_set *set, Idx elem) { Idx idx; /* In case the set is empty. */ if (set->alloc == 0) - return re_node_set_init_1 (set, elem) == REG_NOERROR; + return BE (re_node_set_init_1 (set, elem) == REG_NOERROR, 1); if (BE (set->nelem, 0) == 0) { @@ -1213,12 +1312,12 @@ re_node_set_insert (re_node_set *set, Idx elem) { 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. */ @@ -1232,7 +1331,7 @@ re_node_set_insert (re_node_set *set, Idx elem) Return true if successful. */ static bool -internal_function +internal_function __attribute_warn_unused_result__ re_node_set_insert_last (re_node_set *set, Idx elem) { /* Realloc if we need. */ @@ -1255,7 +1354,7 @@ re_node_set_insert_last (re_node_set *set, Idx elem) Return true if SET1 and SET2 are equivalent. */ static bool -internal_function __attribute ((pure)) +internal_function __attribute__ ((pure)) re_node_set_compare (const re_node_set *set1, const re_node_set *set2) { Idx i; @@ -1270,7 +1369,7 @@ re_node_set_compare (const re_node_set *set1, const re_node_set *set2) /* Return (idx + 1) if SET contains the element ELEM, return 0 otherwise. */ static Idx -internal_function __attribute ((pure)) +internal_function __attribute__ ((pure)) re_node_set_contains (const re_node_set *set, Idx elem) { __re_size_t idx, right, mid; @@ -1310,15 +1409,21 @@ static Idx 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)) { - Idx new_nodes_alloc = dfa->nodes_alloc * 2; + size_t new_nodes_alloc = dfa->nodes_alloc * 2; Idx *new_nexts, *new_indices; re_node_set *new_edests, *new_eclosures; + re_token_t *new_nodes; + + /* Avoid overflows in realloc. */ + const size_t max_object_size = MAX (sizeof (re_token_t), + MAX (sizeof (re_node_set), + sizeof (Idx))); + if (BE (MIN (IDX_MAX, SIZE_MAX / max_object_size) < new_nodes_alloc, 0)) + return REG_MISSING; - re_token_t *new_nodes = re_realloc (dfa->nodes, re_token_t, - new_nodes_alloc); + new_nodes = re_realloc (dfa->nodes, re_token_t, new_nodes_alloc); if (BE (new_nodes == NULL, 0)) return REG_MISSING; dfa->nodes = new_nodes; @@ -1339,7 +1444,8 @@ re_dfa_add_node (re_dfa_t *dfa, re_token_t token) dfa->nodes[dfa->nodes_len].constraint = 0; #ifdef RE_ENABLE_I18N dfa->nodes[dfa->nodes_len].accept_mb = - (type == OP_PERIOD && dfa->mb_cur_max > 1) || type == COMPLEX_BRACKET; + ((token.type == OP_PERIOD && dfa->mb_cur_max > 1) + || token.type == COMPLEX_BRACKET); #endif dfa->nexts[dfa->nodes_len] = REG_MISSING; re_node_set_init_empty (dfa->edests + dfa->nodes_len); @@ -1347,7 +1453,7 @@ re_dfa_add_node (re_dfa_t *dfa, re_token_t token) return dfa->nodes_len++; } -static inline re_hashval_t +static re_hashval_t internal_function calc_state_hash (const re_node_set *nodes, unsigned int context) { @@ -1367,9 +1473,10 @@ calc_state_hash (const re_node_set *nodes, unsigned int context) - We never return non-NULL value in case of any errors, it is for optimization. */ -static re_dfastate_t* -internal_function -re_acquire_state (reg_errcode_t *err, 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) { re_hashval_t hash; re_dfastate_t *new_state; @@ -1398,13 +1505,10 @@ re_acquire_state (reg_errcode_t *err, re_dfa_t *dfa, const re_node_set *nodes) /* There are no appropriate state in the dfa, create the new one. */ new_state = create_ci_newstate (dfa, nodes, hash); - if (BE (new_state != NULL, 1)) - return new_state; - else - { - *err = REG_ESPACE; - return NULL; - } + if (BE (new_state == NULL, 0)) + *err = REG_ESPACE; + + return new_state; } /* Search for the state whose node_set is equivalent to NODES and @@ -1417,9 +1521,9 @@ re_acquire_state (reg_errcode_t *err, re_dfa_t *dfa, const re_node_set *nodes) - We never return non-NULL value in case of any errors, it is for optimization. */ -static re_dfastate_t* -internal_function -re_acquire_state_context (reg_errcode_t *err, re_dfa_t *dfa, +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) { re_hashval_t hash; @@ -1446,15 +1550,12 @@ re_acquire_state_context (reg_errcode_t *err, re_dfa_t *dfa, && re_node_set_compare (state->entrance_nodes, nodes)) return state; } - /* There are no appropriate state in `dfa', create the new one. */ + /* There are no appropriate state in 'dfa', create the new one. */ new_state = create_cd_newstate (dfa, nodes, context, hash); - if (BE (new_state != NULL, 1)) - return new_state; - else - { - *err = REG_ESPACE; - return NULL; - } + if (BE (new_state == NULL, 0)) + *err = REG_ESPACE; + + return new_state; } /* Finish initialization of the new state NEWSTATE, and using its hash value @@ -1462,8 +1563,9 @@ re_acquire_state_context (reg_errcode_t *err, re_dfa_t *dfa, indicates the error code if failed. */ static reg_errcode_t -internal_function -register_state (const re_dfa_t *dfa, re_dfastate_t *newstate, re_hashval_t hash) +__attribute_warn_unused_result__ +register_state (const re_dfa_t *dfa, re_dfastate_t *newstate, + re_hashval_t hash) { struct re_state_table_entry *spot; reg_errcode_t err; @@ -1477,11 +1579,8 @@ register_state (const re_dfa_t *dfa, re_dfastate_t *newstate, re_hashval_t hash) { Idx elem = newstate->nodes.elems[i]; if (!IS_EPSILON_NODE (dfa->nodes[elem].type)) - { - bool ok = re_node_set_insert_last (&newstate->non_eps_nodes, elem); - if (BE (! ok, 0)) - return REG_ESPACE; - } + if (! re_node_set_insert_last (&newstate->non_eps_nodes, elem)) + return REG_ESPACE; } spot = dfa->state_table + (hash & dfa->state_hash_mask); @@ -1499,11 +1598,27 @@ register_state (const re_dfa_t *dfa, re_dfastate_t *newstate, re_hashval_t hash) return REG_NOERROR; } -/* Create the new state which is independ of contexts. +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 independent of contexts. Return the new state if succeeded, otherwise return NULL. */ static re_dfastate_t * -internal_function +internal_function __attribute_warn_unused_result__ create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes, re_hashval_t hash) { @@ -1511,7 +1626,7 @@ create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes, reg_errcode_t err; re_dfastate_t *newstate; - newstate = re_calloc (re_dfastate_t, 1); + newstate = (re_dfastate_t *) calloc (sizeof (re_dfastate_t), 1); if (BE (newstate == NULL, 0)) return NULL; err = re_node_set_init_copy (&newstate->nodes, nodes); @@ -1553,7 +1668,7 @@ create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes, Return the new state if succeeded, otherwise return NULL. */ static re_dfastate_t * -internal_function +internal_function __attribute_warn_unused_result__ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, unsigned int context, re_hashval_t hash) { @@ -1561,7 +1676,7 @@ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, reg_errcode_t err; re_dfastate_t *newstate; - newstate = re_calloc (re_dfastate_t, 1); + newstate = (re_dfastate_t *) calloc (sizeof (re_dfastate_t), 1); if (BE (newstate == NULL, 0)) return NULL; err = re_node_set_init_copy (&newstate->nodes, nodes); @@ -1576,11 +1691,9 @@ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, 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; @@ -1593,8 +1706,6 @@ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, newstate->halt = 1; else if (type == OP_BACK_REF) newstate->has_backref = 1; - else if (type == ANCHOR) - constraint = node->opr.ctx_type; if (constraint) { @@ -1606,7 +1717,9 @@ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, 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; } @@ -1626,20 +1739,3 @@ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, } return newstate; } - -static void -internal_function -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); -}