From a405fc138596f552581d2011fd6de02d5c8186c4 Mon Sep 17 00:00:00 2001 From: RincewindsHat <12514511+RincewindsHat@users.noreply.github.com> Date: Wed, 17 Nov 2021 01:05:54 +0100 Subject: Sync with the latest Gnulib code (1a268176f) --- gl/regex_internal.c | 350 +++++++++++++++++++++++++++------------------------- 1 file changed, 181 insertions(+), 169 deletions(-) (limited to 'gl/regex_internal.c') diff --git a/gl/regex_internal.c b/gl/regex_internal.c index 899b0ae6..aefcfa2f 100644 --- a/gl/regex_internal.c +++ b/gl/regex_internal.c @@ -1,33 +1,43 @@ /* Extended regular expression matching and search library. - Copyright (C) 2002-2013 Free Software Foundation, Inc. + Copyright (C) 2002-2021 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Isamu Hasegawa . The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU General Public + modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either - version 3 of the License, or (at your option) any later version. + version 2.1 of the License, or (at your option) any later version. 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. + Lesser General Public License for more details. - You should have received a copy of the GNU General Public + 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, RE_TRANSLATE_TYPE trans, bool icase, - const re_dfa_t *dfa) internal_function; + const re_dfa_t *dfa); static re_dfastate_t *create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes, - re_hashval_t hash) internal_function; + re_hashval_t hash); 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; + re_hashval_t hash); +static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr, + Idx new_buf_len); +#ifdef RE_ENABLE_I18N +static void build_wcs_buffer (re_string_t *pstr); +static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr); +#endif /* RE_ENABLE_I18N */ +static void build_upper_buffer (re_string_t *pstr); +static void re_string_translate_buffer (re_string_t *pstr); +static unsigned int re_string_context_at (const re_string_t *input, Idx idx, + int eflags) __attribute__ ((pure)); /* Functions for string operation. */ @@ -35,7 +45,7 @@ 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 __attribute_warn_unused_result__ +__attribute_warn_unused_result__ 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) { @@ -49,7 +59,7 @@ re_string_allocate (re_string_t *pstr, const char *str, Idx len, Idx 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)) + if (__glibc_unlikely (ret != REG_NOERROR)) return ret; pstr->word_char = dfa->word_char; @@ -63,7 +73,7 @@ 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 __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_string_construct (re_string_t *pstr, const char *str, Idx len, RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa) { @@ -74,7 +84,7 @@ re_string_construct (re_string_t *pstr, const char *str, Idx len, if (len > 0) { ret = re_string_realloc_buffers (pstr, len + 1); - if (BE (ret != REG_NOERROR, 0)) + if (__glibc_unlikely (ret != REG_NOERROR)) return ret; } pstr->mbs = pstr->mbs_allocated ? pstr->mbs : (unsigned char *) str; @@ -87,14 +97,14 @@ re_string_construct (re_string_t *pstr, const char *str, Idx len, while (1) { ret = build_wcs_upper_buffer (pstr); - if (BE (ret != REG_NOERROR, 0)) + if (__glibc_unlikely (ret != REG_NOERROR)) 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)) + if (__glibc_unlikely (ret != REG_NOERROR)) return ret; } } @@ -126,7 +136,7 @@ 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 __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_string_realloc_buffers (re_string_t *pstr, Idx new_buf_len) { #ifdef RE_ENABLE_I18N @@ -136,17 +146,18 @@ re_string_realloc_buffers (re_string_t *pstr, Idx new_buf_len) /* 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)) + if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / max_object_size) + < new_buf_len)) return REG_ESPACE; new_wcs = re_realloc (pstr->wcs, wint_t, new_buf_len); - if (BE (new_wcs == NULL, 0)) + if (__glibc_unlikely (new_wcs == NULL)) 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)) + if (__glibc_unlikely (new_offsets == NULL)) return REG_ESPACE; pstr->offsets = new_offsets; } @@ -156,7 +167,7 @@ re_string_realloc_buffers (re_string_t *pstr, Idx new_buf_len) { unsigned char *new_mbs = re_realloc (pstr->mbs, unsigned char, new_buf_len); - if (BE (new_mbs == NULL, 0)) + if (__glibc_unlikely (new_mbs == NULL)) return REG_ESPACE; pstr->mbs = new_mbs; } @@ -166,7 +177,6 @@ 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, RE_TRANSLATE_TYPE trans, bool icase, const re_dfa_t *dfa) @@ -198,12 +208,11 @@ re_string_construct_common (const char *str, Idx len, re_string_t *pstr, 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); + DEBUG_ASSERT (MB_LEN_MAX >= pstr->mb_cur_max); #else unsigned char buf[64]; #endif @@ -222,7 +231,7 @@ build_wcs_buffer (re_string_t *pstr) remain_len = end_idx - byte_idx; prev_st = pstr->cur_state; /* Apply the translation if we need. */ - if (BE (pstr->trans != NULL, 0)) + if (__glibc_unlikely (pstr->trans != NULL)) { int i, ch; @@ -236,17 +245,18 @@ 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) -1 || mbclen == 0 - || (mbclen == (size_t) -2 && pstr->bufs_len >= pstr->len), 0)) + if (__glibc_unlikely (mbclen == (size_t) -1 || mbclen == 0 + || (mbclen == (size_t) -2 + && pstr->bufs_len >= pstr->len))) { /* 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)) + if (__glibc_unlikely (pstr->trans != NULL)) wc = pstr->trans[wc]; pstr->cur_state = prev_st; } - else if (BE (mbclen == (size_t) -2, 0)) + else if (__glibc_unlikely (mbclen == (size_t) -2)) { /* The buffer doesn't have enough space, finish to build. */ pstr->cur_state = prev_st; @@ -267,7 +277,7 @@ build_wcs_buffer (re_string_t *pstr) but for REG_ICASE. */ static reg_errcode_t -internal_function __attribute_warn_unused_result__ +__attribute_warn_unused_result__ build_wcs_upper_buffer (re_string_t *pstr) { mbstate_t prev_st; @@ -275,7 +285,7 @@ build_wcs_upper_buffer (re_string_t *pstr) size_t mbclen; #ifdef _LIBC char buf[MB_LEN_MAX]; - assert (MB_LEN_MAX >= pstr->mb_cur_max); + DEBUG_ASSERT (pstr->mb_cur_max <= MB_LEN_MAX); #else char buf[64]; #endif @@ -290,18 +300,20 @@ build_wcs_upper_buffer (re_string_t *pstr) while (byte_idx < end_idx) { wchar_t wc; + unsigned char ch = pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]; - if (isascii (pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]) - && mbsinit (&pstr->cur_state)) + if (isascii (ch) && 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; + wchar_t wcu = __towupper (ch); + if (isascii (wcu)) + { + pstr->mbs[byte_idx] = wcu; + pstr->wcs[byte_idx] = wcu; + byte_idx++; + continue; + } } remain_len = end_idx - byte_idx; @@ -309,16 +321,15 @@ build_wcs_upper_buffer (re_string_t *pstr) 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)) + if (__glibc_likely (0 < mbclen && mbclen < (size_t) -2)) { - wchar_t wcu = wc; - if (iswlower (wc)) + wchar_t wcu = __towupper (wc); + if (wcu != wc) { size_t mbcdlen; - wcu = towupper (wc); - mbcdlen = wcrtomb (buf, wcu, &prev_st); - if (BE (mbclen == mbcdlen, 1)) + mbcdlen = __wcrtomb (buf, wcu, &prev_st); + if (__glibc_likely (mbclen == mbcdlen)) memcpy (pstr->mbs + byte_idx, buf, mbclen); else { @@ -339,11 +350,10 @@ build_wcs_upper_buffer (re_string_t *pstr) { /* 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. */ pstr->wcs[byte_idx++] = (wchar_t) ch; - if (BE (mbclen == (size_t) -1, 0)) + if (__glibc_unlikely (mbclen == (size_t) -1)) pstr->cur_state = prev_st; } else @@ -365,7 +375,7 @@ build_wcs_upper_buffer (re_string_t *pstr) offsets_needed: remain_len = end_idx - byte_idx; prev_st = pstr->cur_state; - if (BE (pstr->trans != NULL, 0)) + if (__glibc_unlikely (pstr->trans != NULL)) { int i, ch; @@ -379,16 +389,15 @@ 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 (mbclen < (size_t) -2, 1)) + if (__glibc_likely (0 < mbclen && mbclen < (size_t) -2)) { - wchar_t wcu = wc; - if (iswlower (wc)) + wchar_t wcu = __towupper (wc); + if (wcu != wc) { size_t mbcdlen; - wcu = towupper (wc); - mbcdlen = wcrtomb ((char *) buf, wcu, &prev_st); - if (BE (mbclen == mbcdlen, 1)) + mbcdlen = __wcrtomb ((char *) buf, wcu, &prev_st); + if (__glibc_likely (mbclen == mbcdlen)) memcpy (pstr->mbs + byte_idx, buf, mbclen); else if (mbcdlen != (size_t) -1) { @@ -438,7 +447,7 @@ build_wcs_upper_buffer (re_string_t *pstr) else memcpy (pstr->mbs + byte_idx, p, mbclen); - if (BE (pstr->offsets_needed != 0, 0)) + if (__glibc_unlikely (pstr->offsets_needed != 0)) { size_t i; for (i = 0; i < mbclen; ++i) @@ -457,17 +466,17 @@ build_wcs_upper_buffer (re_string_t *pstr) /* It is an invalid character or '\0'. Just use the byte. */ int ch = pstr->raw_mbs[pstr->raw_mbs_idx + src_idx]; - if (BE (pstr->trans != NULL, 0)) + if (__glibc_unlikely (pstr->trans != NULL)) ch = pstr->trans [ch]; pstr->mbs[byte_idx] = ch; - if (BE (pstr->offsets_needed != 0, 0)) + if (__glibc_unlikely (pstr->offsets_needed != 0)) pstr->offsets[byte_idx] = src_idx; ++src_idx; /* And also cast it to wide char. */ pstr->wcs[byte_idx++] = (wchar_t) ch; - if (BE (mbclen == (size_t) -1, 0)) + if (__glibc_unlikely (mbclen == (size_t) -1)) pstr->cur_state = prev_st; } else @@ -486,7 +495,6 @@ build_wcs_upper_buffer (re_string_t *pstr) Return the index. */ static Idx -internal_function re_string_skip_chars (re_string_t *pstr, Idx new_raw_idx, wint_t *last_wc) { mbstate_t prev_st; @@ -503,7 +511,8 @@ re_string_skip_chars (re_string_t *pstr, Idx new_raw_idx, wint_t *last_wc) prev_st = 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)) + if (__glibc_unlikely (mbclen == (size_t) -2 || mbclen == (size_t) -1 + || mbclen == 0)) { /* We treat these cases as a single byte character. */ if (mbclen == 0 || remain_len == 0) @@ -527,7 +536,6 @@ re_string_skip_chars (re_string_t *pstr, Idx new_raw_idx, wint_t *last_wc) This function is used in case of REG_ICASE. */ static void -internal_function build_upper_buffer (re_string_t *pstr) { Idx char_idx, end_idx; @@ -536,12 +544,9 @@ build_upper_buffer (re_string_t *pstr) for (char_idx = pstr->valid_len; char_idx < end_idx; ++char_idx) { int ch = pstr->raw_mbs[pstr->raw_mbs_idx + char_idx]; - if (BE (pstr->trans != NULL, 0)) + if (__glibc_unlikely (pstr->trans != NULL)) ch = pstr->trans[ch]; - if (islower (ch)) - pstr->mbs[char_idx] = toupper (ch); - else - pstr->mbs[char_idx] = ch; + pstr->mbs[char_idx] = toupper (ch); } pstr->valid_len = char_idx; pstr->valid_raw_len = char_idx; @@ -550,7 +555,6 @@ build_upper_buffer (re_string_t *pstr) /* Apply TRANS to the buffer in PSTR. */ static void -internal_function re_string_translate_buffer (re_string_t *pstr) { Idx buf_idx, end_idx; @@ -571,12 +575,12 @@ 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 __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) { Idx offset; - if (BE (pstr->raw_mbs_idx <= idx, 0)) + if (__glibc_unlikely (pstr->raw_mbs_idx <= idx)) offset = idx - pstr->raw_mbs_idx; else { @@ -598,14 +602,14 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) offset = idx; } - if (BE (offset != 0, 1)) + if (__glibc_likely (offset != 0)) { /* Should the already checked characters be kept? */ - if (BE (offset < pstr->valid_raw_len, 1)) + if (__glibc_likely (offset < pstr->valid_raw_len)) { /* Yes, move them to the front of the buffer. */ #ifdef RE_ENABLE_I18N - if (BE (pstr->offsets_needed, 0)) + if (__glibc_unlikely (pstr->offsets_needed)) { Idx low = 0, high = pstr->valid_len, mid; do @@ -677,14 +681,12 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) memmove (pstr->wcs, pstr->wcs + offset, (pstr->valid_len - offset) * sizeof (wint_t)); #endif /* RE_ENABLE_I18N */ - if (BE (pstr->mbs_allocated, 0)) + if (__glibc_unlikely (pstr->mbs_allocated)) 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); -#endif + DEBUG_ASSERT (pstr->valid_len > 0); } } else @@ -693,7 +695,7 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) /* No, skip all characters until IDX. */ Idx prev_valid_len = pstr->valid_len; - if (BE (pstr->offsets_needed, 0)) + if (__glibc_unlikely (pstr->offsets_needed)) { pstr->len = pstr->raw_len - idx + offset; pstr->stop = pstr->raw_stop - idx + offset; @@ -721,7 +723,7 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) #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)) + if (isascii (*p) && __glibc_likely (pstr->trans == NULL)) { memset (&pstr->cur_state, '\0', sizeof (mbstate_t)); /* pstr->valid_len = 0; */ @@ -739,7 +741,7 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) size_t mbclen; const unsigned char *pp = p; - if (BE (pstr->trans != NULL, 0)) + if (__glibc_unlikely (pstr->trans != NULL)) { int i = mlen < 6 ? mlen : 6; while (--i >= 0) @@ -769,13 +771,13 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) pstr->tip_context = re_string_context_at (pstr, prev_valid_len - 1, eflags); else - pstr->tip_context = ((BE (pstr->word_ops_used != 0, 0) + pstr->tip_context = ((__glibc_unlikely (pstr->word_ops_used != 0) && IS_WIDE_WORD_CHAR (wc)) ? CONTEXT_WORD : ((IS_WIDE_NEWLINE (wc) && pstr->newline_anchor) ? CONTEXT_NEWLINE : 0)); - if (BE (pstr->valid_len, 0)) + if (__glibc_unlikely (pstr->valid_len)) { for (wcs_idx = 0; wcs_idx < pstr->valid_len; ++wcs_idx) pstr->wcs[wcs_idx] = WEOF; @@ -797,7 +799,7 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) ? CONTEXT_NEWLINE : 0)); } } - if (!BE (pstr->mbs_allocated, 0)) + if (!__glibc_unlikely (pstr->mbs_allocated)) pstr->mbs += offset; } pstr->raw_mbs_idx = idx; @@ -811,7 +813,7 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) if (pstr->icase) { reg_errcode_t ret = build_wcs_upper_buffer (pstr); - if (BE (ret != REG_NOERROR, 0)) + if (__glibc_unlikely (ret != REG_NOERROR)) return ret; } else @@ -819,7 +821,7 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) } else #endif /* RE_ENABLE_I18N */ - if (BE (pstr->mbs_allocated, 0)) + if (__glibc_unlikely (pstr->mbs_allocated)) { if (pstr->icase) build_upper_buffer (pstr); @@ -834,14 +836,14 @@ re_string_reconstruct (re_string_t *pstr, Idx idx, int eflags) } static unsigned char -internal_function __attribute__ ((pure)) +__attribute__ ((pure)) re_string_peek_byte_case (const re_string_t *pstr, Idx idx) { int ch; Idx off; /* Handle the common (easiest) cases first. */ - if (BE (!pstr->mbs_allocated, 1)) + if (__glibc_likely (!pstr->mbs_allocated)) return re_string_peek_byte (pstr, idx); #ifdef RE_ENABLE_I18N @@ -871,10 +873,9 @@ re_string_peek_byte_case (const re_string_t *pstr, Idx idx) } static unsigned char -internal_function re_string_fetch_byte_case (re_string_t *pstr) { - if (BE (!pstr->mbs_allocated, 1)) + if (__glibc_likely (!pstr->mbs_allocated)) return re_string_fetch_byte (pstr); #ifdef RE_ENABLE_I18N @@ -909,7 +910,6 @@ re_string_fetch_byte_case (re_string_t *pstr) } static void -internal_function re_string_destruct (re_string_t *pstr) { #ifdef RE_ENABLE_I18N @@ -923,15 +923,14 @@ re_string_destruct (re_string_t *pstr) /* Return the context at IDX in INPUT. */ static unsigned int -internal_function re_string_context_at (const re_string_t *input, Idx idx, int eflags) { int c; - if (BE (! REG_VALID_INDEX (idx), 0)) + if (__glibc_unlikely (idx < 0)) /* In this case, we use the value stored in input->tip_context, since we can't know the character in input->mbs[-1] here. */ return input->tip_context; - if (BE (idx == input->len, 0)) + if (__glibc_unlikely (idx == input->len)) return ((eflags & REG_NOTEOL) ? CONTEXT_ENDBUF : CONTEXT_NEWLINE | CONTEXT_ENDBUF); #ifdef RE_ENABLE_I18N @@ -941,16 +940,14 @@ re_string_context_at (const re_string_t *input, Idx idx, int eflags) Idx wc_idx = idx; while(input->wcs[wc_idx] == WEOF) { -#ifdef DEBUG - /* It must not happen. */ - assert (REG_VALID_INDEX (wc_idx)); -#endif + DEBUG_ASSERT (wc_idx >= 0); --wc_idx; - if (! REG_VALID_INDEX (wc_idx)) + if (wc_idx < 0) return input->tip_context; } wc = input->wcs[wc_idx]; - if (BE (input->word_ops_used != 0, 0) && IS_WIDE_WORD_CHAR (wc)) + if (__glibc_unlikely (input->word_ops_used != 0) + && IS_WIDE_WORD_CHAR (wc)) return CONTEXT_WORD; return (IS_WIDE_NEWLINE (wc) && input->newline_anchor ? CONTEXT_NEWLINE : 0); @@ -968,25 +965,26 @@ re_string_context_at (const re_string_t *input, Idx idx, int eflags) /* Functions for set operation. */ static reg_errcode_t -internal_function __attribute_warn_unused_result__ +__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) && (MALLOC_0_IS_NONNULL || size != 0)) + if (__glibc_unlikely (set->elems == NULL) + && (MALLOC_0_IS_NONNULL || size != 0)) return REG_ESPACE; return REG_NOERROR; } static reg_errcode_t -internal_function __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_node_set_init_1 (re_node_set *set, Idx elem) { set->alloc = 1; set->nelem = 1; set->elems = re_malloc (Idx, 1); - if (BE (set->elems == NULL, 0)) + if (__glibc_unlikely (set->elems == NULL)) { set->alloc = set->nelem = 0; return REG_ESPACE; @@ -996,12 +994,12 @@ re_node_set_init_1 (re_node_set *set, Idx elem) } static reg_errcode_t -internal_function __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_node_set_init_2 (re_node_set *set, Idx elem1, Idx elem2) { set->alloc = 2; set->elems = re_malloc (Idx, 2); - if (BE (set->elems == NULL, 0)) + if (__glibc_unlikely (set->elems == NULL)) return REG_ESPACE; if (elem1 == elem2) { @@ -1026,7 +1024,7 @@ re_node_set_init_2 (re_node_set *set, Idx elem1, Idx elem2) } static reg_errcode_t -internal_function __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_node_set_init_copy (re_node_set *dest, const re_node_set *src) { dest->nelem = src->nelem; @@ -1034,7 +1032,7 @@ re_node_set_init_copy (re_node_set *dest, const re_node_set *src) { dest->alloc = dest->nelem; dest->elems = re_malloc (Idx, dest->alloc); - if (BE (dest->elems == NULL, 0)) + if (__glibc_unlikely (dest->elems == NULL)) { dest->alloc = dest->nelem = 0; return REG_ESPACE; @@ -1051,7 +1049,7 @@ 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 __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1, const re_node_set *src2) { @@ -1065,7 +1063,7 @@ 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)) + if (__glibc_unlikely (new_elems == NULL)) return REG_ESPACE; dest->elems = new_elems; dest->alloc = new_alloc; @@ -1082,25 +1080,25 @@ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1, if (src1->elems[i1] == src2->elems[i2]) { /* Try to find the item in DEST. Maybe we could binary search? */ - while (REG_VALID_INDEX (id) && dest->elems[id] > src1->elems[i1]) + while (id >= 0 && dest->elems[id] > src1->elems[i1]) --id; - if (! REG_VALID_INDEX (id) || dest->elems[id] != src1->elems[i1]) + if (id < 0 || dest->elems[id] != src1->elems[i1]) dest->elems[--sbase] = src1->elems[i1]; - if (! REG_VALID_INDEX (--i1) || ! REG_VALID_INDEX (--i2)) + if (--i1 < 0 || --i2 < 0) break; } /* Lower the highest of the two items. */ else if (src1->elems[i1] < src2->elems[i2]) { - if (! REG_VALID_INDEX (--i2)) + if (--i2 < 0) break; } else { - if (! REG_VALID_INDEX (--i1)) + if (--i1 < 0) break; } } @@ -1113,7 +1111,7 @@ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1, DEST elements are already in place; this is more or less the same loop that is in re_node_set_merge. */ dest->nelem += delta; - if (delta > 0 && REG_VALID_INDEX (id)) + if (delta > 0 && id >= 0) for (;;) { if (dest->elems[is] > dest->elems[id]) @@ -1127,7 +1125,7 @@ re_node_set_add_intersect (re_node_set *dest, const re_node_set *src1, { /* Slide from the bottom. */ dest->elems[id + delta] = dest->elems[id]; - if (! REG_VALID_INDEX (--id)) + if (--id < 0) break; } } @@ -1142,7 +1140,7 @@ 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 __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_node_set_init_union (re_node_set *dest, const re_node_set *src1, const re_node_set *src2) { @@ -1151,7 +1149,7 @@ re_node_set_init_union (re_node_set *dest, const re_node_set *src1, { dest->alloc = src1->nelem + src2->nelem; dest->elems = re_malloc (Idx, dest->alloc); - if (BE (dest->elems == NULL, 0)) + if (__glibc_unlikely (dest->elems == NULL)) return REG_ESPACE; } else @@ -1195,7 +1193,7 @@ 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 __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_node_set_merge (re_node_set *dest, const re_node_set *src) { Idx is, id, sbase, delta; @@ -1205,14 +1203,18 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src) { Idx new_alloc = 2 * (src->nelem + dest->alloc); Idx *new_buffer = re_realloc (dest->elems, Idx, new_alloc); - if (BE (new_buffer == NULL, 0)) + if (__glibc_unlikely (new_buffer == NULL)) return REG_ESPACE; dest->elems = new_buffer; dest->alloc = new_alloc; } - if (BE (dest->nelem == 0, 0)) + if (__glibc_unlikely (dest->nelem == 0)) { + /* Although we already guaranteed above that dest->alloc != 0 and + therefore dest->elems != NULL, add a debug assertion to pacify + GCC 11.2.1's -fanalyzer. */ + DEBUG_ASSERT (dest->elems); dest->nelem = src->nelem; memcpy (dest->elems, src->elems, src->nelem * sizeof (Idx)); return REG_NOERROR; @@ -1221,8 +1223,7 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src) /* Copy into the top of DEST the items of SRC that are not found in DEST. Maybe we could binary search in DEST? */ for (sbase = dest->nelem + 2 * src->nelem, - is = src->nelem - 1, id = dest->nelem - 1; - REG_VALID_INDEX (is) && REG_VALID_INDEX (id); ) + is = src->nelem - 1, id = dest->nelem - 1; is >= 0 && id >= 0; ) { if (dest->elems[id] == src->elems[is]) is--, id--; @@ -1232,7 +1233,7 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src) --id; } - if (REG_VALID_INDEX (is)) + if (is >= 0) { /* If DEST is exhausted, the remaining items of SRC must be unique. */ sbase -= is + 1; @@ -1261,7 +1262,7 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src) { /* Slide from the bottom. */ dest->elems[id + delta] = dest->elems[id]; - if (! REG_VALID_INDEX (--id)) + if (--id < 0) { /* Copy remaining SRC elements. */ memcpy (dest->elems, dest->elems + sbase, @@ -1279,17 +1280,20 @@ re_node_set_merge (re_node_set *dest, const re_node_set *src) Return true if successful. */ static bool -internal_function __attribute_warn_unused_result__ +__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 BE (re_node_set_init_1 (set, elem) == REG_NOERROR, 1); + return __glibc_likely (re_node_set_init_1 (set, elem) == REG_NOERROR); - if (BE (set->nelem, 0) == 0) + if (__glibc_unlikely (set->nelem) == 0) { - /* We already guaranteed above that set->alloc != 0. */ + /* Although we already guaranteed above that set->alloc != 0 and + therefore set->elems != NULL, add a debug assertion to pacify + GCC 11.2 -fanalyzer. */ + DEBUG_ASSERT (set->elems); set->elems[0] = elem; ++set->nelem; return true; @@ -1301,7 +1305,7 @@ re_node_set_insert (re_node_set *set, Idx elem) Idx *new_elems; set->alloc = set->alloc * 2; new_elems = re_realloc (set->elems, Idx, set->alloc); - if (BE (new_elems == NULL, 0)) + if (__glibc_unlikely (new_elems == NULL)) return false; set->elems = new_elems; } @@ -1310,7 +1314,6 @@ re_node_set_insert (re_node_set *set, Idx elem) first element separately to skip a check in the inner loop. */ if (elem < set->elems[0]) { - idx = 0; for (idx = set->nelem; idx > 0; idx--) set->elems[idx] = set->elems[idx - 1]; } @@ -1318,6 +1321,7 @@ re_node_set_insert (re_node_set *set, Idx elem) { for (idx = set->nelem; set->elems[idx - 1] > elem; idx--) set->elems[idx] = set->elems[idx - 1]; + DEBUG_ASSERT (set->elems[idx - 1] < elem); } /* Insert the new element. */ @@ -1331,7 +1335,7 @@ re_node_set_insert (re_node_set *set, Idx elem) Return true if successful. */ static bool -internal_function __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_node_set_insert_last (re_node_set *set, Idx elem) { /* Realloc if we need. */ @@ -1340,7 +1344,7 @@ re_node_set_insert_last (re_node_set *set, Idx elem) Idx *new_elems; set->alloc = (set->alloc + 1) * 2; new_elems = re_realloc (set->elems, Idx, set->alloc); - if (BE (new_elems == NULL, 0)) + if (__glibc_unlikely (new_elems == NULL)) return false; set->elems = new_elems; } @@ -1354,13 +1358,13 @@ 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)) +__attribute__ ((pure)) re_node_set_compare (const re_node_set *set1, const re_node_set *set2) { Idx i; if (set1 == NULL || set2 == NULL || set1->nelem != set2->nelem) return false; - for (i = set1->nelem ; REG_VALID_INDEX (--i) ; ) + for (i = set1->nelem ; --i >= 0 ; ) if (set1->elems[i] != set2->elems[i]) return false; return true; @@ -1369,11 +1373,11 @@ 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)) +__attribute__ ((pure)) re_node_set_contains (const re_node_set *set, Idx elem) { __re_size_t idx, right, mid; - if (! REG_VALID_NONZERO_INDEX (set->nelem)) + if (set->nelem <= 0) return 0; /* Binary search the element. */ @@ -1391,7 +1395,6 @@ re_node_set_contains (const re_node_set *set, Idx elem) } static void -internal_function re_node_set_remove_at (re_node_set *set, Idx idx) { if (idx < 0 || idx >= set->nelem) @@ -1403,13 +1406,12 @@ re_node_set_remove_at (re_node_set *set, Idx idx) /* Add the token TOKEN to dfa->nodes, and return the index of the token. - Or return REG_MISSING if an error occurred. */ + Or return -1 if an error occurred. */ static Idx -internal_function re_dfa_add_node (re_dfa_t *dfa, re_token_t token) { - if (BE (dfa->nodes_len >= dfa->nodes_alloc, 0)) + if (__glibc_unlikely (dfa->nodes_len >= dfa->nodes_alloc)) { size_t new_nodes_alloc = dfa->nodes_alloc * 2; Idx *new_nexts, *new_indices; @@ -1420,20 +1422,27 @@ re_dfa_add_node (re_dfa_t *dfa, re_token_t token) 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; + if (__glibc_unlikely (MIN (IDX_MAX, SIZE_MAX / max_object_size) + < new_nodes_alloc)) + return -1; new_nodes = re_realloc (dfa->nodes, re_token_t, new_nodes_alloc); - if (BE (new_nodes == NULL, 0)) - return REG_MISSING; + if (__glibc_unlikely (new_nodes == NULL)) + return -1; dfa->nodes = new_nodes; new_nexts = re_realloc (dfa->nexts, Idx, new_nodes_alloc); new_indices = re_realloc (dfa->org_indices, Idx, new_nodes_alloc); new_edests = re_realloc (dfa->edests, re_node_set, new_nodes_alloc); new_eclosures = re_realloc (dfa->eclosures, re_node_set, new_nodes_alloc); - if (BE (new_nexts == NULL || new_indices == NULL - || new_edests == NULL || new_eclosures == NULL, 0)) - return REG_MISSING; + if (__glibc_unlikely (new_nexts == NULL || new_indices == NULL + || new_edests == NULL || new_eclosures == NULL)) + { + re_free (new_nexts); + re_free (new_indices); + re_free (new_edests); + re_free (new_eclosures); + return -1; + } dfa->nexts = new_nexts; dfa->org_indices = new_indices; dfa->edests = new_edests; @@ -1447,14 +1456,13 @@ re_dfa_add_node (re_dfa_t *dfa, re_token_t token) ((token.type == OP_PERIOD && dfa->mb_cur_max > 1) || token.type == COMPLEX_BRACKET); #endif - dfa->nexts[dfa->nodes_len] = REG_MISSING; + dfa->nexts[dfa->nodes_len] = -1; re_node_set_init_empty (dfa->edests + dfa->nodes_len); re_node_set_init_empty (dfa->eclosures + dfa->nodes_len); return dfa->nodes_len++; } static re_hashval_t -internal_function calc_state_hash (const re_node_set *nodes, unsigned int context) { re_hashval_t hash = nodes->nelem + context; @@ -1474,7 +1482,7 @@ calc_state_hash (const re_node_set *nodes, unsigned int context) optimization. */ static re_dfastate_t * -internal_function __attribute_warn_unused_result__ +__attribute_warn_unused_result__ re_acquire_state (reg_errcode_t *err, const re_dfa_t *dfa, const re_node_set *nodes) { @@ -1482,11 +1490,11 @@ re_acquire_state (reg_errcode_t *err, const re_dfa_t *dfa, re_dfastate_t *new_state; struct re_state_table_entry *spot; Idx i; -#ifdef lint +#if defined GCC_LINT || defined lint /* Suppress bogus uninitialized-variable warnings. */ *err = REG_NOERROR; #endif - if (BE (nodes->nelem == 0, 0)) + if (__glibc_unlikely (nodes->nelem == 0)) { *err = REG_NOERROR; return NULL; @@ -1505,7 +1513,7 @@ re_acquire_state (reg_errcode_t *err, const re_dfa_t *dfa, /* 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, 0)) + if (__glibc_unlikely (new_state == NULL)) *err = REG_ESPACE; return new_state; @@ -1522,7 +1530,7 @@ re_acquire_state (reg_errcode_t *err, const re_dfa_t *dfa, optimization. */ static re_dfastate_t * -internal_function __attribute_warn_unused_result__ +__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) { @@ -1530,7 +1538,7 @@ re_acquire_state_context (reg_errcode_t *err, const re_dfa_t *dfa, re_dfastate_t *new_state; struct re_state_table_entry *spot; Idx i; -#ifdef lint +#if defined GCC_LINT || defined lint /* Suppress bogus uninitialized-variable warnings. */ *err = REG_NOERROR; #endif @@ -1552,7 +1560,7 @@ re_acquire_state_context (reg_errcode_t *err, const re_dfa_t *dfa, } /* 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, 0)) + if (__glibc_unlikely (new_state == NULL)) *err = REG_ESPACE; return new_state; @@ -1573,7 +1581,7 @@ register_state (const re_dfa_t *dfa, re_dfastate_t *newstate, newstate->hash = hash; err = re_node_set_alloc (&newstate->non_eps_nodes, newstate->nodes.nelem); - if (BE (err != REG_NOERROR, 0)) + if (__glibc_unlikely (err != REG_NOERROR)) return REG_ESPACE; for (i = 0; i < newstate->nodes.nelem; i++) { @@ -1584,12 +1592,12 @@ register_state (const re_dfa_t *dfa, re_dfastate_t *newstate, } spot = dfa->state_table + (hash & dfa->state_hash_mask); - if (BE (spot->alloc <= spot->num, 0)) + if (__glibc_unlikely (spot->alloc <= spot->num)) { Idx new_alloc = 2 * spot->num + 2; re_dfastate_t **new_array = re_realloc (spot->array, re_dfastate_t *, new_alloc); - if (BE (new_array == NULL, 0)) + if (__glibc_unlikely (new_array == NULL)) return REG_ESPACE; spot->array = new_array; spot->alloc = new_alloc; @@ -1618,7 +1626,7 @@ free_state (re_dfastate_t *state) Return the new state if succeeded, otherwise return NULL. */ static re_dfastate_t * -internal_function __attribute_warn_unused_result__ +__attribute_warn_unused_result__ create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes, re_hashval_t hash) { @@ -1627,10 +1635,10 @@ create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes, re_dfastate_t *newstate; newstate = (re_dfastate_t *) calloc (sizeof (re_dfastate_t), 1); - if (BE (newstate == NULL, 0)) + if (__glibc_unlikely (newstate == NULL)) return NULL; err = re_node_set_init_copy (&newstate->nodes, nodes); - if (BE (err != REG_NOERROR, 0)) + if (__glibc_unlikely (err != REG_NOERROR)) { re_free (newstate); return NULL; @@ -1656,7 +1664,7 @@ create_ci_newstate (const re_dfa_t *dfa, const re_node_set *nodes, newstate->has_constraint = 1; } err = register_state (dfa, newstate, hash); - if (BE (err != REG_NOERROR, 0)) + if (__glibc_unlikely (err != REG_NOERROR)) { free_state (newstate); newstate = NULL; @@ -1668,7 +1676,7 @@ 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 __attribute_warn_unused_result__ +__attribute_warn_unused_result__ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, unsigned int context, re_hashval_t hash) { @@ -1677,10 +1685,10 @@ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, re_dfastate_t *newstate; newstate = (re_dfastate_t *) calloc (sizeof (re_dfastate_t), 1); - if (BE (newstate == NULL, 0)) + if (__glibc_unlikely (newstate == NULL)) return NULL; err = re_node_set_init_copy (&newstate->nodes, nodes); - if (BE (err != REG_NOERROR, 0)) + if (__glibc_unlikely (err != REG_NOERROR)) { re_free (newstate); return NULL; @@ -1711,15 +1719,19 @@ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, { if (newstate->entrance_nodes == &newstate->nodes) { - newstate->entrance_nodes = re_malloc (re_node_set, 1); - if (BE (newstate->entrance_nodes == NULL, 0)) + re_node_set *entrance_nodes = re_malloc (re_node_set, 1); + if (__glibc_unlikely (entrance_nodes == NULL)) { free_state (newstate); return NULL; } + newstate->entrance_nodes = entrance_nodes; if (re_node_set_init_copy (newstate->entrance_nodes, nodes) != REG_NOERROR) - return NULL; + { + free_state (newstate); + return NULL; + } nctx_nodes = 0; newstate->has_constraint = 1; } @@ -1732,7 +1744,7 @@ create_cd_newstate (const re_dfa_t *dfa, const re_node_set *nodes, } } err = register_state (dfa, newstate, hash); - if (BE (err != REG_NOERROR, 0)) + if (__glibc_unlikely (err != REG_NOERROR)) { free_state (newstate); newstate = NULL; -- cgit v1.2.3-74-g34f1