mirrors/cosmopolitan
1
0
Fork 0
mirror of https://github.com/jart/cosmopolitan.git synced 2025-05-28 08:12:28 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Initial import

Browse source
This commit is contained in:
Justine Tunney 2020-06-15 07:18:57 -07:00
commit c91b3c5006
14915 changed files with 590219 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

294
third_party/regex/glob.c.todo vendored Normal file
View file

@ -0,0 +1,294 @@
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
│vi: set net ft=c ts=8 sts=2 sw=2 fenc=utf-8 :vi│
╚─────────────────────────────────────────────────────────────────────────────*/
/* clang-format off */
/*
Musl Libc
Copyright © 2005-2014 Rich Felker, et al.
Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:
The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
#define _BSD_SOURCE
#include <dirent.h>
#include <errno.h>
#include <fnmatch.h>
#include <glob.h>
#include <limits.h>
#include <stddef.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
asm(".ident\t\"\\n\\n\
Musl Libc » glob (MIT License)\\n\
Copyright 2005-2014 Rich Felker\"");
struct match
{
struct match *next;
char name[];
};
static int append(struct match **tail, const char *name, size_t len, int mark)
{
struct match *new = malloc(sizeof(struct match) + len + 2);
if (!new) return -1;
(*tail)->next = new;
new->next = NULL;
memcpy(new->name, name, len+1);
if (mark && len && name[len-1]!='/') {
new->name[len] = '/';
new->name[len+1] = 0;
}
*tail = new;
return 0;
}
static int do_glob(char *buf, size_t pos, int type, char *pat, int flags, int (*errfunc)(const char *path, int err), struct match **tail)
{
/* If GLOB_MARK is unused, we don't care about type. */
if (!type && !(flags & GLOB_MARK)) type = DT_REG;
/* Special-case the remaining pattern being all slashes, in
* which case we can use caller-passed type if it's a dir. */
if (*pat && type!=DT_DIR) type = 0;
while (pos+1 < PATH_MAX && *pat=='/') buf[pos++] = *pat++;
/* Consume maximal [escaped-]literal prefix of pattern, copying
* and un-escaping it to the running buffer as we go. */
ptrdiff_t i=0, j=0;
int in_bracket = 0, overflow = 0;
for (; pat[i]!='*' && pat[i]!='?' && (!in_bracket || pat[i]!=']'); i++) {
if (!pat[i]) {
if (overflow) return 0;
pat += i;
pos += j;
i = j = 0;
break;
} else if (pat[i] == '[') {
in_bracket = 1;
} else if (pat[i] == '\\' && !(flags & GLOB_NOESCAPE)) {
/* Backslashes inside a bracket are (at least by
* our interpretation) non-special, so if next
* char is ']' we have a complete expression. */
if (in_bracket && pat[i+1]==']') break;
/* Unpaired final backslash never matches. */
if (!pat[i+1]) return 0;
i++;
}
if (pat[i] == '/') {
if (overflow) return 0;
in_bracket = 0;
pat += i+1;
i = -1;
pos += j+1;
j = -1;
}
/* Only store a character if it fits in the buffer, but if
* a potential bracket expression is open, the overflow
* must be remembered and handled later only if the bracket
* is unterminated (and thereby a literal), so as not to
* disallow long bracket expressions with short matches. */
if (pos+(j+1) < PATH_MAX) {
buf[pos+j++] = pat[i];
} else if (in_bracket) {
overflow = 1;
} else {
return 0;
}
/* If we consume any new components, the caller-passed type
* or dummy type from above is no longer valid. */
type = 0;
}
buf[pos] = 0;
if (!*pat) {
/* If we consumed any components above, or if GLOB_MARK is
* requested and we don't yet know if the match is a dir,
* we must call stat to confirm the file exists and/or
* determine its type. */
struct stat st;
if ((flags & GLOB_MARK) && type==DT_LNK) type = 0;
if (!type && stat(buf, &st)) {
if (errno!=ENOENT && (errfunc(buf, errno) || (flags & GLOB_ERR)))
return GLOB_ABORTED;
return 0;
}
if (!type && S_ISDIR(st.st_mode)) type = DT_DIR;
if (append(tail, buf, pos, (flags & GLOB_MARK) && type==DT_DIR))
return GLOB_NOSPACE;
return 0;
}
char *p2 = strchr(pat, '/'), saved_sep = '/';
/* Check if the '/' was escaped and, if so, remove the escape char
* so that it will not be unpaired when passed to fnmatch. */
if (p2 && !(flags & GLOB_NOESCAPE)) {
char *p;
for (p=p2; p>pat && p[-1]=='\\'; p--);
if ((p2-p)%2) {
p2--;
saved_sep = '\\';
}
}
DIR *dir = opendir(pos ? buf : ".");
if (!dir) {
if (errfunc(buf, errno) || (flags & GLOB_ERR))
return GLOB_ABORTED;
return 0;
}
int old_errno = errno;
struct dirent *de;
while (errno=0, de=readdir(dir)) {
/* Quickly skip non-directories when there's pattern left. */
if (p2 && de->d_type && de->d_type!=DT_DIR && de->d_type!=DT_LNK)
continue;
size_t l = strlen(de->d_name);
if (l >= PATH_MAX-pos) continue;
if (p2) *p2 = 0;
int fnm_flags= ((flags & GLOB_NOESCAPE) ? FNM_NOESCAPE : 0)
| ((!(flags & GLOB_PERIOD)) ? FNM_PERIOD : 0);
if (fnmatch(pat, de->d_name, fnm_flags))
continue;
/* With GLOB_PERIOD, don't allow matching . or .. unless
* fnmatch would match them with FNM_PERIOD rules in effect. */
if (p2 && (flags & GLOB_PERIOD) && de->d_name[0]=='.'
&& (!de->d_name[1] || de->d_name[1]=='.' && !de->d_name[2])
&& fnmatch(pat, de->d_name, fnm_flags | FNM_PERIOD))
continue;
memcpy(buf+pos, de->d_name, l+1);
if (p2) *p2 = saved_sep;
int r = do_glob(buf, pos+l, de->d_type, p2 ? p2 : "", flags, errfunc, tail);
if (r) {
closedir(dir);
return r;
}
}
int readerr = errno;
if (p2) *p2 = saved_sep;
closedir(dir);
if (readerr && (errfunc(buf, errno) || (flags & GLOB_ERR)))
return GLOB_ABORTED;
errno = old_errno;
return 0;
}
static int ignore_err(const char *path, int err)
{
return 0;
}
static void freelist(struct match *head)
{
struct match *match, *next;
for (match=head->next; match; match=next) {
next = match->next;
free(match);
}
}
static int sort(const void *a, const void *b)
{
return strcmp(*(const char **)a, *(const char **)b);
}
int glob(const char *restrict pat, int flags, int (*errfunc)(const char *path, int err), glob_t *restrict g)
{
struct match head = { .next = NULL }, *tail = &head;
size_t cnt, i;
size_t offs = (flags & GLOB_DOOFFS) ? g->gl_offs : 0;
int error = 0;
char buf[PATH_MAX];
if (!errfunc) errfunc = ignore_err;
if (!(flags & GLOB_APPEND)) {
g->gl_offs = offs;
g->gl_pathc = 0;
g->gl_pathv = NULL;
}
if (*pat) {
char *p = strdup(pat);
if (!p) return GLOB_NOSPACE;
buf[0] = 0;
error = do_glob(buf, 0, 0, p, flags, errfunc, &tail);
free(p);
}
if (error == GLOB_NOSPACE) {
freelist(&head);
return error;
}
for (cnt=0, tail=head.next; tail; tail=tail->next, cnt++);
if (!cnt) {
if (flags & GLOB_NOCHECK) {
tail = &head;
if (append(&tail, pat, strlen(pat), 0))
return GLOB_NOSPACE;
cnt++;
} else
return GLOB_NOMATCH;
}
if (flags & GLOB_APPEND) {
char **pathv = realloc(g->gl_pathv, (offs + g->gl_pathc + cnt + 1) * sizeof(char *));
if (!pathv) {
freelist(&head);
return GLOB_NOSPACE;
}
g->gl_pathv = pathv;
offs += g->gl_pathc;
} else {
g->gl_pathv = malloc((offs + cnt + 1) * sizeof(char *));
if (!g->gl_pathv) {
freelist(&head);
return GLOB_NOSPACE;
}
for (i=0; i<offs; i++)
g->gl_pathv[i] = NULL;
}
for (i=0, tail=head.next; i<cnt; tail=tail->next, i++)
g->gl_pathv[offs + i] = tail->name;
g->gl_pathv[offs + i] = NULL;
g->gl_pathc += cnt;
if (!(flags & GLOB_NOSORT))
qsort(g->gl_pathv+offs, cnt, sizeof(char *), sort);
return error;
}
void globfree(glob_t *g)
{
size_t i;
for (i=0; i<g->gl_pathc; i++)
free(g->gl_pathv[g->gl_offs + i] - offsetof(struct match, name));
free(g->gl_pathv);
g->gl_pathc = 0;
g->gl_pathv = NULL;
}

10
third_party/regex/notice.inc vendored Normal file
View file

@ -0,0 +1,10 @@
asm(".ident\t\"\\n\\n\
Musl Libc (MIT License)\\n\
Copyright 2005-2014 Rich Felker\"");
asm(".include \"libc/disclaimer.inc\"");
asm(".ident\t\"\\n\\n\
TRE regex (BSD-2 License)\\n\
Copyright 2001-2009 Ville Laurikari <vl@iki.fi>\\n\
Copyright 2016 Szabolcs Nagy\"");
asm(".include \"libc/disclaimer.inc\"");

2595
third_party/regex/regcomp.c vendored Normal file
View file

File diff suppressed because it is too large Load diff

61
third_party/regex/regerror.c vendored Normal file
View file

@ -0,0 +1,61 @@
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
│vi: set net ft=c ts=8 sts=2 sw=2 fenc=utf-8 :vi│
╚──────────────────────────────────────────────────────────────────────────────╝
│ │
│ Musl Libc │
│ Copyright © 2005-2014 Rich Felker, et al. │
│ │
│ Permission is hereby granted, free of charge, to any person obtaining │
│ a copy of this software and associated documentation files (the │
│ "Software"), to deal in the Software without restriction, including │
│ without limitation the rights to use, copy, modify, merge, publish, │
│ distribute, sublicense, and/or sell copies of the Software, and to │
│ permit persons to whom the Software is furnished to do so, subject to │
│ the following conditions: │
│ │
│ The above copyright notice and this permission notice shall be │
│ included in all copies or substantial portions of the Software. │
│ │
│ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, │
│ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF │
│ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. │
│ IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY │
│ CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, │
│ TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE │
│ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. │
│ │
╚─────────────────────────────────────────────────────────────────────────────*/
#include "libc/fmt/fmt.h"
#include "third_party/regex/tre.inc"
/* Error message strings for error codes listed in `regex.h'. This list
needs to be in sync with the codes listed there, naturally. */
/* Converted to single string by Rich Felker to remove the need for
* data relocations at runtime, 27 Feb 2006. */
static const char messages[] = {
"No error\0"
"No match\0"
"Invalid regexp\0"
"Unknown collating element\0"
"Unknown character class name\0"
"Trailing backslash\0"
"Invalid back reference\0"
"Missing ']'\0"
"Missing ')'\0"
"Missing '}'\0"
"Invalid contents of {}\0"
"Invalid character range\0"
"Out of memory\0"
"Repetition not preceded by valid expression\0"
"\0Unknown error"};
size_t regerror(int e, const regex_t *restrict preg, char *restrict buf,
size_t size) {
const char *s;
for (s = messages; e && *s; e--, s += strlen(s) + 1)
;
if (!*s) s++;
return 1 + snprintf(buf, size, "%s", s);
}

60
third_party/regex/regex.h vendored Normal file
View file

@ -0,0 +1,60 @@
#ifndef COSMOPOLITAN_LIBC_REGEX_REGEX_H_
#define COSMOPOLITAN_LIBC_REGEX_REGEX_H_
#if !(__ASSEMBLER__ + __LINKER__ + 0)
COSMOPOLITAN_C_START_
#if 0
/*───────────────────────────────────────────────────────────────────────────│─╗
│ cosmopolitan § regular expressions ─╬─│┼
╚────────────────────────────────────────────────────────────────────────────│*/
#endif
#define REG_EXTENDED 1
#define REG_ICASE 2
#define REG_NEWLINE 4
#define REG_NOSUB 8
#define REG_NOTBOL 1
#define REG_NOTEOL 2
#define REG_OK 0
#define REG_NOMATCH 1
#define REG_BADPAT 2
#define REG_ECOLLATE 3
#define REG_ECTYPE 4
#define REG_EESCAPE 5
#define REG_ESUBREG 6
#define REG_EBRACK 7
#define REG_EPAREN 8
#define REG_EBRACE 9
#define REG_BADBR 10
#define REG_ERANGE 11
#define REG_ESPACE 12
#define REG_BADRPT 13
#define REG_ENOSYS -1
typedef long regoff_t;
typedef struct re_pattern_buffer {
size_t re_nsub;
void *__opaque, *__padding[4];
size_t __nsub2;
char __padding2;
} regex_t;
typedef struct {
regoff_t rm_so;
regoff_t rm_eo;
} regmatch_t;
int regcomp(regex_t *preg, const char *regex, int flags);
int regexec(const regex_t *preg, const char *string, size_t nmatch,
regmatch_t pmatch[], int eflags);
void regfree(regex_t *preg);
size_t regerror(int errcode, const regex_t *preg, char *errbuf,
size_t errbufsize);
COSMOPOLITAN_C_END_
#endif /* !(__ASSEMBLER__ + __LINKER__ + 0) */
#endif /* COSMOPOLITAN_LIBC_REGEX_REGEX_H_ */

56
third_party/regex/regex.mk vendored Normal file
View file

@ -0,0 +1,56 @@
#-*-mode:makefile-gmake;indent-tabs-mode:t;tab-width:8;coding:utf-8-*-┐
#───vi: set et ft=make ts=8 tw=8 fenc=utf-8 :vi───────────────────────┘
PKGS += THIRD_PARTY_REGEX
THIRD_PARTY_REGEX_ARTIFACTS += THIRD_PARTY_REGEX_A
THIRD_PARTY_REGEX = $(THIRD_PARTY_REGEX_A_DEPS) $(THIRD_PARTY_REGEX_A)
THIRD_PARTY_REGEX_A = o/$(MODE)/third_party/regex/regex.a
THIRD_PARTY_REGEX_A_FILES := $(wildcard third_party/regex/*)
THIRD_PARTY_REGEX_A_HDRS = $(filter %.h,$(THIRD_PARTY_REGEX_A_FILES))
THIRD_PARTY_REGEX_A_SRCS = $(filter %.c,$(THIRD_PARTY_REGEX_A_FILES))
THIRD_PARTY_REGEX_A_OBJS = \
$(THIRD_PARTY_REGEX_A_SRCS:%=o/$(MODE)/%.zip.o) \
$(THIRD_PARTY_REGEX_A_SRCS:%.c=o/$(MODE)/%.o)
THIRD_PARTY_REGEX_A_DIRECTDEPS = \
LIBC_ALG \
LIBC_FMT \
LIBC_MEM \
LIBC_NEXGEN32E \
LIBC_STR \
LIBC_STUBS
THIRD_PARTY_REGEX_A_DEPS := \
$(call uniq,$(foreach x,$(THIRD_PARTY_REGEX_A_DIRECTDEPS),$($(x))))
THIRD_PARTY_REGEX_A_CHECKS = \
$(THIRD_PARTY_REGEX_A).pkg \
$(THIRD_PARTY_REGEX_A_HDRS:%=o/$(MODE)/%.ok)
$(THIRD_PARTY_REGEX_A): \
third_party/regex/ \
$(THIRD_PARTY_REGEX_A).pkg \
$(THIRD_PARTY_REGEX_A_OBJS)
$(THIRD_PARTY_REGEX_A).pkg: \
$(THIRD_PARTY_REGEX_A_OBJS) \
$(foreach x,$(THIRD_PARTY_REGEX_A_DIRECTDEPS),$($(x)_A).pkg)
THIRD_PARTY_REGEX_LIBS = $(foreach x,$(THIRD_PARTY_REGEX_ARTIFACTS),$($(x)))
THIRD_PARTY_REGEX_SRCS = $(foreach x,$(THIRD_PARTY_REGEX_ARTIFACTS),$($(x)_SRCS))
THIRD_PARTY_REGEX_CHECKS = $(foreach x,$(THIRD_PARTY_REGEX_ARTIFACTS),$($(x)_CHECKS))
THIRD_PARTY_REGEX_OBJS = $(foreach x,$(THIRD_PARTY_REGEX_ARTIFACTS),$($(x)_OBJS))
$(THIRD_PARTY_REGEX_OBJS): $(BUILD_FILES) third_party/regex/regex.mk
o/$(MODE)/third_party/regex/regcomp.o \
o/$(MODE)/third_party/regex/regerror.o \
o/$(MODE)/third_party/regex/regexec.o \
o/$(MODE)/third_party/regex/tre-mem.o: \
OVERRIDE_COPTS += \
$(OLD_CODE)
.PHONY: o/$(MODE)/third_party/regex
o/$(MODE)/third_party/regex: $(THIRD_PARTY_REGEX_CHECKS)

898
third_party/regex/regexec.c vendored Normal file
View file

@ -0,0 +1,898 @@
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
│vi: set net ft=c ts=2 sts=2 sw=2 fenc=utf-8 :vi│
╚──────────────────────────────────────────────────────────────────────────────╝
│ │
│ regexec.c - TRE POSIX compatible matching functions (and more). │
│ │
│ Copyright (c) 2001-2009 Ville Laurikari <vl@iki.fi> │
│ All rights reserved. │
│ │
│ Redistribution and use in source and binary forms, with or without │
│ modification, are permitted provided that the following conditions │
│ are met: │
│ │
│ 1. Redistributions of source code must retain the above copyright │
│ notice, this list of conditions and the following disclaimer. │
│ │
│ 2. Redistributions in binary form must reproduce the above copyright │
│ notice, this list of conditions and the following disclaimer in │
│ the documentation and/or other materials provided with the │
│ distribution. │
│ │
│ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER AND CONTRIBUTORS │
│ ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT │
│ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR │
│ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT │
│ HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, │
│ SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT │
│ LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, │
│ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY │
│ THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT │
│ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE │
│ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. │
│ │
│──────────────────────────────────────────────────────────────────────────────│
│ │
│ Musl Libc │
│ Copyright © 2005-2014 Rich Felker, et al. │
│ │
│ Permission is hereby granted, free of charge, to any person obtaining │
│ a copy of this software and associated documentation files (the │
│ "Software"), to deal in the Software without restriction, including │
│ without limitation the rights to use, copy, modify, merge, publish, │
│ distribute, sublicense, and/or sell copies of the Software, and to │
│ permit persons to whom the Software is furnished to do so, subject to │
│ the following conditions: │
│ │
│ The above copyright notice and this permission notice shall be │
│ included in all copies or substantial portions of the Software. │
│ │
│ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, │
│ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF │
│ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. │
│ IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY │
│ CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, │
│ TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE │
│ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. │
│ │
╚─────────────────────────────────────────────────────────────────────────────*/
#include "libc/limits.h"
#include "third_party/regex/tre.inc"
static void tre_fill_pmatch(size_t nmatch, regmatch_t pmatch[], int cflags,
const tre_tnfa_t *tnfa, regoff_t *tags,
regoff_t match_eo);
/***********************************************************************
from tre-match-utils.h
***********************************************************************/
#define GET_NEXT_WCHAR() \
do { \
prev_c = next_c; \
pos += pos_add_next; \
if ((pos_add_next = mbtowc(&next_c, str_byte, MB_LEN_MAX)) <= 0) { \
if (pos_add_next < 0) { \
ret = REG_NOMATCH; \
goto error_exit; \
} else \
pos_add_next++; \
} \
str_byte += pos_add_next; \
} while (0)
#define IS_WORD_CHAR(c) ((c) == L'_' || tre_isalnum(c))
#define CHECK_ASSERTIONS(assertions) \
(((assertions & ASSERT_AT_BOL) && (pos > 0 || reg_notbol) && \
(prev_c != L'\n' || !reg_newline)) || \
((assertions & ASSERT_AT_EOL) && (next_c != L'\0' || reg_noteol) && \
(next_c != L'\n' || !reg_newline)) || \
((assertions & ASSERT_AT_BOW) && \
(IS_WORD_CHAR(prev_c) || !IS_WORD_CHAR(next_c))) || \
((assertions & ASSERT_AT_EOW) && \
(!IS_WORD_CHAR(prev_c) || IS_WORD_CHAR(next_c))) || \
((assertions & ASSERT_AT_WB) && \
(pos != 0 && next_c != L'\0' && \
IS_WORD_CHAR(prev_c) == IS_WORD_CHAR(next_c))) || \
((assertions & ASSERT_AT_WB_NEG) && \
(pos == 0 || next_c == L'\0' || \
IS_WORD_CHAR(prev_c) != IS_WORD_CHAR(next_c))))
#define CHECK_CHAR_CLASSES(trans_i, tnfa, eflags) \
(((trans_i->assertions & ASSERT_CHAR_CLASS) && \
!(tnfa->cflags & REG_ICASE) && \
!tre_isctype((tre_cint_t)prev_c, trans_i->u.class)) || \
((trans_i->assertions & ASSERT_CHAR_CLASS) && (tnfa->cflags & REG_ICASE) && \
!tre_isctype(tre_tolower((tre_cint_t)prev_c), trans_i->u.class) && \
!tre_isctype(tre_toupper((tre_cint_t)prev_c), trans_i->u.class)) || \
((trans_i->assertions & ASSERT_CHAR_CLASS_NEG) && \
tre_neg_char_classes_match(trans_i->neg_classes, (tre_cint_t)prev_c, \
tnfa->cflags & REG_ICASE)))
/* Returns 1 if `t1' wins `t2', 0 otherwise. */
static int tre_tag_order(int num_tags, tre_tag_direction_t *tag_directions,
regoff_t *t1, regoff_t *t2) {
int i;
for (i = 0; i < num_tags; i++) {
if (tag_directions[i] == TRE_TAG_MINIMIZE) {
if (t1[i] < t2[i]) return 1;
if (t1[i] > t2[i]) return 0;
} else {
if (t1[i] > t2[i]) return 1;
if (t1[i] < t2[i]) return 0;
}
}
/* assert(0);*/
return 0;
}
static int tre_neg_char_classes_match(tre_ctype_t *classes, tre_cint_t wc,
int icase) {
while (*classes != (tre_ctype_t)0)
if ((!icase && tre_isctype(wc, *classes)) ||
(icase && (tre_isctype(tre_toupper(wc), *classes) ||
tre_isctype(tre_tolower(wc), *classes))))
return 1; /* Match. */
else
classes++;
return 0; /* No match. */
}
/***********************************************************************
from tre-match-parallel.c
***********************************************************************/
/*
This algorithm searches for matches basically by reading characters
in the searched string one by one, starting at the beginning. All
matching paths in the TNFA are traversed in parallel. When two or
more paths reach the same state, exactly one is chosen according to
tag ordering rules; if returning submatches is not required it does
not matter which path is chosen.
The worst case time required for finding the leftmost and longest
match, or determining that there is no match, is always linearly
dependent on the length of the text being searched.
This algorithm cannot handle TNFAs with back referencing nodes.
See `tre-match-backtrack.c'.
*/
typedef struct {
tre_tnfa_transition_t *state;
regoff_t *tags;
} tre_tnfa_reach_t;
typedef struct {
regoff_t pos;
regoff_t **tags;
} tre_reach_pos_t;
static reg_errcode_t tre_tnfa_run_parallel(const tre_tnfa_t *tnfa,
const void *string,
regoff_t *match_tags, int eflags,
regoff_t *match_end_ofs) {
/* State variables required by GET_NEXT_WCHAR. */
tre_char_t prev_c = 0, next_c = 0;
const char *str_byte = string;
regoff_t pos = -1;
regoff_t pos_add_next = 1;
#ifdef TRE_MBSTATE
mbstate_t mbstate;
#endif /* TRE_MBSTATE */
int reg_notbol = eflags & REG_NOTBOL;
int reg_noteol = eflags & REG_NOTEOL;
int reg_newline = tnfa->cflags & REG_NEWLINE;
reg_errcode_t ret;
char *buf;
tre_tnfa_transition_t *trans_i;
tre_tnfa_reach_t *reach, *reach_next, *reach_i, *reach_next_i;
tre_reach_pos_t *reach_pos;
int *tag_i;
int num_tags, i;
regoff_t match_eo = -1; /* end offset of match (-1 if no match found yet) */
int new_match = 0;
regoff_t *tmp_tags = NULL;
regoff_t *tmp_iptr;
#ifdef TRE_MBSTATE
memset(&mbstate, '\0', sizeof(mbstate));
#endif /* TRE_MBSTATE */
if (!match_tags)
num_tags = 0;
else
num_tags = tnfa->num_tags;
/* Allocate memory for temporary data required for matching. This needs to
be done for every matching operation to be thread safe. This allocates
everything in a single large block with calloc(). */
{
size_t tbytes, rbytes, pbytes, xbytes, total_bytes;
char *tmp_buf;
/* Ensure that tbytes and xbytes*num_states cannot overflow, and that
* they don't contribute more than 1/8 of SIZE_MAX to total_bytes. */
if (num_tags > SIZE_MAX / (8 * sizeof(regoff_t) * tnfa->num_states))
return REG_ESPACE;
/* Likewise check rbytes. */
if (tnfa->num_states + 1 > SIZE_MAX / (8 * sizeof(*reach_next)))
return REG_ESPACE;
/* Likewise check pbytes. */
if (tnfa->num_states > SIZE_MAX / (8 * sizeof(*reach_pos)))
return REG_ESPACE;
/* Compute the length of the block we need. */
tbytes = sizeof(*tmp_tags) * num_tags;
rbytes = sizeof(*reach_next) * (tnfa->num_states + 1);
pbytes = sizeof(*reach_pos) * tnfa->num_states;
xbytes = sizeof(regoff_t) * num_tags;
total_bytes = (sizeof(long) - 1) * 4 /* for alignment paddings */
+ (rbytes + xbytes * tnfa->num_states) * 2 + tbytes + pbytes;
/* Allocate the memory. */
buf = calloc(total_bytes, 1);
if (buf == NULL) return REG_ESPACE;
/* Get the various pointers within tmp_buf (properly aligned). */
tmp_tags = (void *)buf;
tmp_buf = buf + tbytes;
tmp_buf += ALIGN(tmp_buf, long);
reach_next = (void *)tmp_buf;
tmp_buf += rbytes;
tmp_buf += ALIGN(tmp_buf, long);
reach = (void *)tmp_buf;
tmp_buf += rbytes;
tmp_buf += ALIGN(tmp_buf, long);
reach_pos = (void *)tmp_buf;
tmp_buf += pbytes;
tmp_buf += ALIGN(tmp_buf, long);
for (i = 0; i < tnfa->num_states; i++) {
reach[i].tags = (void *)tmp_buf;
tmp_buf += xbytes;
reach_next[i].tags = (void *)tmp_buf;
tmp_buf += xbytes;
}
}
for (i = 0; i < tnfa->num_states; i++) reach_pos[i].pos = -1;
GET_NEXT_WCHAR();
pos = 0;
reach_next_i = reach_next;
while (1) {
/* If no match found yet, add the initial states to `reach_next'. */
if (match_eo < 0) {
trans_i = tnfa->initial;
while (trans_i->state != NULL) {
if (reach_pos[trans_i->state_id].pos < pos) {
if (trans_i->assertions && CHECK_ASSERTIONS(trans_i->assertions)) {
trans_i++;
continue;
}
reach_next_i->state = trans_i->state;
for (i = 0; i < num_tags; i++) reach_next_i->tags[i] = -1;
tag_i = trans_i->tags;
if (tag_i)
while (*tag_i >= 0) {
if (*tag_i < num_tags) reach_next_i->tags[*tag_i] = pos;
tag_i++;
}
if (reach_next_i->state == tnfa->final) {
match_eo = pos;
new_match = 1;
for (i = 0; i < num_tags; i++)
match_tags[i] = reach_next_i->tags[i];
}
reach_pos[trans_i->state_id].pos = pos;
reach_pos[trans_i->state_id].tags = &reach_next_i->tags;
reach_next_i++;
}
trans_i++;
}
reach_next_i->state = NULL;
} else {
if (num_tags == 0 || reach_next_i == reach_next)
/* We have found a match. */
break;
}
/* Check for end of string. */
if (!next_c) break;
GET_NEXT_WCHAR();
/* Swap `reach' and `reach_next'. */
reach_i = reach;
reach = reach_next;
reach_next = reach_i;
/* For each state in `reach', weed out states that don't fulfill the
minimal matching conditions. */
if (tnfa->num_minimals && new_match) {
new_match = 0;
reach_next_i = reach_next;
for (reach_i = reach; reach_i->state; reach_i++) {
int skip = 0;
for (i = 0; tnfa->minimal_tags[i] >= 0; i += 2) {
int end = tnfa->minimal_tags[i];
int start = tnfa->minimal_tags[i + 1];
if (end >= num_tags) {
skip = 1;
break;
} else if (reach_i->tags[start] == match_tags[start] &&
reach_i->tags[end] < match_tags[end]) {
skip = 1;
break;
}
}
if (!skip) {
reach_next_i->state = reach_i->state;
tmp_iptr = reach_next_i->tags;
reach_next_i->tags = reach_i->tags;
reach_i->tags = tmp_iptr;
reach_next_i++;
}
}
reach_next_i->state = NULL;
/* Swap `reach' and `reach_next'. */
reach_i = reach;
reach = reach_next;
reach_next = reach_i;
}
/* For each state in `reach' see if there is a transition leaving with
the current input symbol to a state not yet in `reach_next', and
add the destination states to `reach_next'. */
reach_next_i = reach_next;
for (reach_i = reach; reach_i->state; reach_i++) {
for (trans_i = reach_i->state; trans_i->state; trans_i++) {
/* Does this transition match the input symbol? */
if (trans_i->code_min <= (tre_cint_t)prev_c &&
trans_i->code_max >= (tre_cint_t)prev_c) {
if (trans_i->assertions &&
(CHECK_ASSERTIONS(trans_i->assertions) ||
CHECK_CHAR_CLASSES(trans_i, tnfa, eflags))) {
continue;
}
/* Compute the tags after this transition. */
for (i = 0; i < num_tags; i++) tmp_tags[i] = reach_i->tags[i];
tag_i = trans_i->tags;
if (tag_i != NULL)
while (*tag_i >= 0) {
if (*tag_i < num_tags) tmp_tags[*tag_i] = pos;
tag_i++;
}
if (reach_pos[trans_i->state_id].pos < pos) {
/* Found an unvisited node. */
reach_next_i->state = trans_i->state;
tmp_iptr = reach_next_i->tags;
reach_next_i->tags = tmp_tags;
tmp_tags = tmp_iptr;
reach_pos[trans_i->state_id].pos = pos;
reach_pos[trans_i->state_id].tags = &reach_next_i->tags;
if (reach_next_i->state == tnfa->final &&
(match_eo == -1 ||
(num_tags > 0 && reach_next_i->tags[0] <= match_tags[0]))) {
match_eo = pos;
new_match = 1;
for (i = 0; i < num_tags; i++)
match_tags[i] = reach_next_i->tags[i];
}
reach_next_i++;
} else {
assert(reach_pos[trans_i->state_id].pos == pos);
/* Another path has also reached this state. We choose
the winner by examining the tag values for both
paths. */
if (tre_tag_order(num_tags, tnfa->tag_directions, tmp_tags,
*reach_pos[trans_i->state_id].tags)) {
/* The new path wins. */
tmp_iptr = *reach_pos[trans_i->state_id].tags;
*reach_pos[trans_i->state_id].tags = tmp_tags;
if (trans_i->state == tnfa->final) {
match_eo = pos;
new_match = 1;
for (i = 0; i < num_tags; i++) match_tags[i] = tmp_tags[i];
}
tmp_tags = tmp_iptr;
}
}
}
}
}
reach_next_i->state = NULL;
}
*match_end_ofs = match_eo;
ret = match_eo >= 0 ? REG_OK : REG_NOMATCH;
error_exit:
free(buf), buf = NULL;
return ret;
}
/***********************************************************************
from tre-match-backtrack.c
***********************************************************************/
/*
This matcher is for regexps that use back referencing. Regexp matching
with back referencing is an NP-complete problem on the number of back
references. The easiest way to match them is to use a backtracking
routine which basically goes through all possible paths in the TNFA
and chooses the one which results in the best (leftmost and longest)
match. This can be spectacularly expensive and may run out of stack
space, but there really is no better known generic algorithm. Quoting
Henry Spencer from comp.compilers:
<URL: http://compilers.iecc.com/comparch/article/93-03-102>
POSIX.2 REs require longest match, which is really exciting to
implement since the obsolete ("basic") variant also includes
\<digit>. I haven't found a better way of tackling this than doing
a preliminary match using a DFA (or simulation) on a modified RE
that just replicates subREs for \<digit>, and then doing a
backtracking match to determine whether the subRE matches were
right. This can be rather slow, but I console myself with the
thought that people who use \<digit> deserve very slow execution.
(Pun unintentional but very appropriate.)
*/
typedef struct {
regoff_t pos;
const char *str_byte;
tre_tnfa_transition_t *state;
int state_id;
int next_c;
regoff_t *tags;
#ifdef TRE_MBSTATE
mbstate_t mbstate;
#endif /* TRE_MBSTATE */
} tre_backtrack_item_t;
typedef struct tre_backtrack_struct {
tre_backtrack_item_t item;
struct tre_backtrack_struct *prev;
struct tre_backtrack_struct *next;
} * tre_backtrack_t;
#ifdef TRE_MBSTATE
#define BT_STACK_MBSTATE_IN stack->item.mbstate = (mbstate)
#define BT_STACK_MBSTATE_OUT (mbstate) = stack->item.mbstate
#else /* !TRE_MBSTATE */
#define BT_STACK_MBSTATE_IN
#define BT_STACK_MBSTATE_OUT
#endif /* !TRE_MBSTATE */
#define tre_bt_mem_new tre_mem_new
#define tre_bt_mem_alloc tre_mem_alloc
#define tre_bt_mem_destroy tre_mem_destroy
#define BT_STACK_PUSH(_pos, _str_byte, _str_wide, _state, _state_id, _next_c, \
_tags, _mbstate) \
do { \
int i; \
if (!stack->next) { \
tre_backtrack_t s; \
s = tre_bt_mem_alloc(mem, sizeof(*s)); \
if (!s) { \
tre_bt_mem_destroy(mem); \
if (tags) free(tags), tags = NULL; \
if (pmatch) free(pmatch), pmatch = NULL; \
if (states_seen) free(states_seen), states_seen = NULL; \
return REG_ESPACE; \
} \
s->prev = stack; \
s->next = NULL; \
s->item.tags = tre_bt_mem_alloc(mem, sizeof(*tags) * tnfa->num_tags); \
if (!s->item.tags) { \
tre_bt_mem_destroy(mem); \
if (tags) free(tags), tags = NULL; \
if (pmatch) free(pmatch), pmatch = NULL; \
if (states_seen) free(states_seen), states_seen = NULL; \
return REG_ESPACE; \
} \
stack->next = s; \
stack = s; \
} else \
stack = stack->next; \
stack->item.pos = (_pos); \
stack->item.str_byte = (_str_byte); \
stack->item.state = (_state); \
stack->item.state_id = (_state_id); \
stack->item.next_c = (_next_c); \
for (i = 0; i < tnfa->num_tags; i++) stack->item.tags[i] = (_tags)[i]; \
BT_STACK_MBSTATE_IN; \
} while (0)
#define BT_STACK_POP() \
do { \
int i; \
assert(stack->prev); \
pos = stack->item.pos; \
str_byte = stack->item.str_byte; \
state = stack->item.state; \
next_c = stack->item.next_c; \
for (i = 0; i < tnfa->num_tags; i++) tags[i] = stack->item.tags[i]; \
BT_STACK_MBSTATE_OUT; \
stack = stack->prev; \
} while (0)
#undef MIN
#define MIN(a, b) ((a) <= (b) ? (a) : (b))
static reg_errcode_t tre_tnfa_run_backtrack(const tre_tnfa_t *tnfa,
const void *string,
regoff_t *match_tags, int eflags,
regoff_t *match_end_ofs) {
/* State variables required by GET_NEXT_WCHAR. */
tre_char_t prev_c = 0, next_c = 0;
const char *str_byte = string;
regoff_t pos = 0;
regoff_t pos_add_next = 1;
#ifdef TRE_MBSTATE
mbstate_t mbstate;
#endif /* TRE_MBSTATE */
int reg_notbol = eflags & REG_NOTBOL;
int reg_noteol = eflags & REG_NOTEOL;
int reg_newline = tnfa->cflags & REG_NEWLINE;
/* These are used to remember the necessary values of the above
variables to return to the position where the current search
started from. */
int next_c_start;
const char *str_byte_start;
regoff_t pos_start = -1;
#ifdef TRE_MBSTATE
mbstate_t mbstate_start;
#endif /* TRE_MBSTATE */
/* End offset of best match so far, or -1 if no match found yet. */
regoff_t match_eo = -1;
/* Tag arrays. */
int *next_tags;
regoff_t *tags = NULL;
/* Current TNFA state. */
tre_tnfa_transition_t *state;
int *states_seen = NULL;
/* Memory allocator to for allocating the backtracking stack. */
tre_mem_t mem = tre_bt_mem_new();
/* The backtracking stack. */
tre_backtrack_t stack;
tre_tnfa_transition_t *trans_i;
regmatch_t *pmatch = NULL;
int ret;
#ifdef TRE_MBSTATE
memset(&mbstate, '\0', sizeof(mbstate));
#endif /* TRE_MBSTATE */
if (!mem) return REG_ESPACE;
stack = tre_bt_mem_alloc(mem, sizeof(*stack));
if (!stack) {
ret = REG_ESPACE;
goto error_exit;
}
stack->prev = NULL;
stack->next = NULL;
if (tnfa->num_tags) {
tags = malloc(sizeof(*tags) * tnfa->num_tags);
if (!tags) {
ret = REG_ESPACE;
goto error_exit;
}
}
if (tnfa->num_submatches) {
pmatch = malloc(sizeof(*pmatch) * tnfa->num_submatches);
if (!pmatch) {
ret = REG_ESPACE;
goto error_exit;
}
}
if (tnfa->num_states) {
states_seen = malloc(sizeof(*states_seen) * tnfa->num_states);
if (!states_seen) {
ret = REG_ESPACE;
goto error_exit;
}
}
retry : {
int i;
for (i = 0; i < tnfa->num_tags; i++) {
tags[i] = -1;
if (match_tags) match_tags[i] = -1;
}
for (i = 0; i < tnfa->num_states; i++) states_seen[i] = 0;
}
state = NULL;
pos = pos_start;
GET_NEXT_WCHAR();
pos_start = pos;
next_c_start = next_c;
str_byte_start = str_byte;
#ifdef TRE_MBSTATE
mbstate_start = mbstate;
#endif /* TRE_MBSTATE */
/* Handle initial states. */
next_tags = NULL;
for (trans_i = tnfa->initial; trans_i->state; trans_i++) {
if (trans_i->assertions && CHECK_ASSERTIONS(trans_i->assertions)) {
continue;
}
if (state == NULL) {
/* Start from this state. */
state = trans_i->state;
next_tags = trans_i->tags;
} else {
/* Backtrack to this state. */
BT_STACK_PUSH(pos, str_byte, 0, trans_i->state, trans_i->state_id, next_c,
tags, mbstate);
{
int *tmp = trans_i->tags;
if (tmp)
while (*tmp >= 0) stack->item.tags[*tmp++] = pos;
}
}
}
if (next_tags)
for (; *next_tags >= 0; next_tags++) tags[*next_tags] = pos;
if (state == NULL) goto backtrack;
while (1) {
tre_tnfa_transition_t *next_state;
int empty_br_match;
if (state == tnfa->final) {
if (match_eo < pos || (match_eo == pos && match_tags &&
tre_tag_order(tnfa->num_tags, tnfa->tag_directions,
tags, match_tags))) {
int i;
/* This match wins the previous match. */
match_eo = pos;
if (match_tags)
for (i = 0; i < tnfa->num_tags; i++) match_tags[i] = tags[i];
}
/* Our TNFAs never have transitions leaving from the final state,
so we jump right to backtracking. */
goto backtrack;
}
/* Go to the next character in the input string. */
empty_br_match = 0;
trans_i = state;
if (trans_i->state && trans_i->assertions & ASSERT_BACKREF) {
/* This is a back reference state. All transitions leaving from
this state have the same back reference "assertion". Instead
of reading the next character, we match the back reference. */
regoff_t so, eo;
int bt = trans_i->u.backref;
regoff_t bt_len;
int result;
/* Get the substring we need to match against. Remember to
turn off REG_NOSUB temporarily. */
tre_fill_pmatch(bt + 1, pmatch, tnfa->cflags & ~REG_NOSUB, tnfa, tags,
pos);
so = pmatch[bt].rm_so;
eo = pmatch[bt].rm_eo;
bt_len = eo - so;
result = strncmp((const char *)string + so, str_byte - 1, (size_t)bt_len);
if (result == 0) {
/* Back reference matched. Check for infinite loop. */
if (bt_len == 0) empty_br_match = 1;
if (empty_br_match && states_seen[trans_i->state_id]) {
goto backtrack;
}
states_seen[trans_i->state_id] = empty_br_match;
/* Advance in input string and resync `prev_c', `next_c'
and pos. */
str_byte += bt_len - 1;
pos += bt_len - 1;
GET_NEXT_WCHAR();
} else {
goto backtrack;
}
} else {
/* Check for end of string. */
if (next_c == L'\0') goto backtrack;
/* Read the next character. */
GET_NEXT_WCHAR();
}
next_state = NULL;
for (trans_i = state; trans_i->state; trans_i++) {
if (trans_i->code_min <= (tre_cint_t)prev_c &&
trans_i->code_max >= (tre_cint_t)prev_c) {
if (trans_i->assertions &&
(CHECK_ASSERTIONS(trans_i->assertions) ||
CHECK_CHAR_CLASSES(trans_i, tnfa, eflags))) {
continue;
}
if (next_state == NULL) {
/* First matching transition. */
next_state = trans_i->state;
next_tags = trans_i->tags;
} else {
/* Second matching transition. We may need to backtrack here
to take this transition instead of the first one, so we
push this transition in the backtracking stack so we can
jump back here if needed. */
BT_STACK_PUSH(pos, str_byte, 0, trans_i->state, trans_i->state_id,
next_c, tags, mbstate);
{
int *tmp;
for (tmp = trans_i->tags; tmp && *tmp >= 0; tmp++)
stack->item.tags[*tmp] = pos;
}
#if 0 /* XXX - it's important not to look at all transitions here to keep \
the stack small! */
break;
#endif
}
}
}
if (next_state != NULL) {
/* Matching transitions were found. Take the first one. */
state = next_state;
/* Update the tag values. */
if (next_tags)
while (*next_tags >= 0) tags[*next_tags++] = pos;
} else {
backtrack:
/* A matching transition was not found. Try to backtrack. */
if (stack->prev) {
if (stack->item.state->assertions & ASSERT_BACKREF) {
states_seen[stack->item.state_id] = 0;
}
BT_STACK_POP();
} else if (match_eo < 0) {
/* Try starting from a later position in the input string. */
/* Check for end of string. */
if (next_c == L'\0') {
break;
}
next_c = next_c_start;
#ifdef TRE_MBSTATE
mbstate = mbstate_start;
#endif /* TRE_MBSTATE */
str_byte = str_byte_start;
goto retry;
} else {
break;
}
}
}
ret = match_eo >= 0 ? REG_OK : REG_NOMATCH;
*match_end_ofs = match_eo;
error_exit:
tre_bt_mem_destroy(mem);
#ifndef TRE_USE_ALLOCA
if (tags) free(tags), tags = NULL;
if (pmatch) free(pmatch), pmatch = NULL;
if (states_seen) free(states_seen), states_seen = NULL;
#endif /* !TRE_USE_ALLOCA */
return ret;
}
/***********************************************************************
from regexec.c
***********************************************************************/
/* Fills the POSIX.2 regmatch_t array according to the TNFA tag and match
endpoint values. */
static void tre_fill_pmatch(size_t nmatch, regmatch_t pmatch[], int cflags,
const tre_tnfa_t *tnfa, regoff_t *tags,
regoff_t match_eo) {
tre_submatch_data_t *submatch_data;
unsigned int i, j;
int *parents;
i = 0;
if (match_eo >= 0 && !(cflags & REG_NOSUB)) {
/* Construct submatch offsets from the tags. */
submatch_data = tnfa->submatch_data;
while (i < tnfa->num_submatches && i < nmatch) {
if (submatch_data[i].so_tag == tnfa->end_tag)
pmatch[i].rm_so = match_eo;
else
pmatch[i].rm_so = tags[submatch_data[i].so_tag];
if (submatch_data[i].eo_tag == tnfa->end_tag)
pmatch[i].rm_eo = match_eo;
else
pmatch[i].rm_eo = tags[submatch_data[i].eo_tag];
/* If either of the endpoints were not used, this submatch
was not part of the match. */
if (pmatch[i].rm_so == -1 || pmatch[i].rm_eo == -1)
pmatch[i].rm_so = pmatch[i].rm_eo = -1;
i++;
}
/* Reset all submatches that are not within all of their parent
submatches. */
i = 0;
while (i < tnfa->num_submatches && i < nmatch) {
if (pmatch[i].rm_eo == -1) assert(pmatch[i].rm_so == -1);
assert(pmatch[i].rm_so <= pmatch[i].rm_eo);
parents = submatch_data[i].parents;
if (parents != NULL)
for (j = 0; parents[j] >= 0; j++) {
if (pmatch[i].rm_so < pmatch[parents[j]].rm_so ||
pmatch[i].rm_eo > pmatch[parents[j]].rm_eo)
pmatch[i].rm_so = pmatch[i].rm_eo = -1;
}
i++;
}
}
while (i < nmatch) {
pmatch[i].rm_so = -1;
pmatch[i].rm_eo = -1;
i++;
}
}
/*
Wrapper functions for POSIX compatible regexp matching.
*/
int regexec(const regex_t *restrict preg, const char *restrict string,
size_t nmatch, regmatch_t pmatch[restrict], int eflags) {
tre_tnfa_t *tnfa = (void *)preg->TRE_REGEX_T_FIELD;
reg_errcode_t status;
regoff_t *tags = NULL, eo;
if (tnfa->cflags & REG_NOSUB) nmatch = 0;
if (tnfa->num_tags > 0 && nmatch > 0) {
tags = malloc(sizeof(*tags) * tnfa->num_tags);
if (tags == NULL) return REG_ESPACE;
}
/* Dispatch to the appropriate matcher. */
if (tnfa->have_backrefs) {
/* The regex has back references, use the backtracking matcher. */
status = tre_tnfa_run_backtrack(tnfa, string, tags, eflags, &eo);
} else {
/* Exact matching, no back references, use the parallel matcher. */
status = tre_tnfa_run_parallel(tnfa, string, tags, eflags, &eo);
}
if (status == REG_OK) /* A match was found, so fill the submatch registers. */
tre_fill_pmatch(nmatch, pmatch, tnfa->cflags, tnfa, tags, eo);
if (tags) free(tags), tags = NULL;
return status;
}

158
third_party/regex/tre-mem.c vendored Normal file
View file

@ -0,0 +1,158 @@
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
│vi: set net ft=c ts=2 sts=2 sw=2 fenc=utf-8 :vi│
╚──────────────────────────────────────────────────────────────────────────────╝
│ │
│ tre-mem.c - TRE memory allocator │
│ │
│ Copyright (c) 2001-2009 Ville Laurikari <vl@iki.fi> │
│ All rights reserved. │
│ │
│ Redistribution and use in source and binary forms, with or without │
│ modification, are permitted provided that the following conditions │
│ are met: │
│ │
│ 1. Redistributions of source code must retain the above copyright │
│ notice, this list of conditions and the following disclaimer. │
│ │
│ 2. Redistributions in binary form must reproduce the above copyright │
│ notice, this list of conditions and the following disclaimer in │
│ the documentation and/or other materials provided with the │
│ distribution. │
│ │
│ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER AND CONTRIBUTORS │
│ ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT │
│ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR │
│ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT │
│ HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, │
│ SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT │
│ LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, │
│ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY │
│ THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT │
│ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE │
│ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. │
│ │
│──────────────────────────────────────────────────────────────────────────────│
│ │
│ Musl Libc │
│ Copyright © 2005-2014 Rich Felker, et al. │
│ │
│ Permission is hereby granted, free of charge, to any person obtaining │
│ a copy of this software and associated documentation files (the │
│ "Software"), to deal in the Software without restriction, including │
│ without limitation the rights to use, copy, modify, merge, publish, │
│ distribute, sublicense, and/or sell copies of the Software, and to │
│ permit persons to whom the Software is furnished to do so, subject to │
│ the following conditions: │
│ │
│ The above copyright notice and this permission notice shall be │
│ included in all copies or substantial portions of the Software. │
│ │
│ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, │
│ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF │
│ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. │
│ IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY │
│ CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, │
│ TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE │
│ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. │
│ │
╚─────────────────────────────────────────────────────────────────────────────*/
#include "third_party/regex/tre.inc"
/*
This memory allocator is for allocating small memory blocks efficiently
in terms of memory overhead and execution speed. The allocated blocks
cannot be freed individually, only all at once. There can be multiple
allocators, though.
*/
/*
This memory allocator is for allocating small memory blocks efficiently
in terms of memory overhead and execution speed. The allocated blocks
cannot be freed individually, only all at once. There can be multiple
allocators, though.
*/
/* Returns a new memory allocator or NULL if out of memory. */
tre_mem_t tre_mem_new_impl(int provided, void *provided_block) {
tre_mem_t mem;
if (provided) {
mem = provided_block;
memset(mem, 0, sizeof(*mem));
} else
mem = calloc(1, sizeof(*mem));
if (mem == NULL) return NULL;
return mem;
}
/* Frees the memory allocator and all memory allocated with it. */
void tre_mem_destroy(tre_mem_t mem) {
tre_list_t *tmp, *l = mem->blocks;
while (l != NULL) {
free(l->data), l->data = NULL;
tmp = l->next;
free(l), l = tmp;
}
free(mem), mem = NULL;
}
/* Allocates a block of `size' bytes from `mem'. Returns a pointer to the
allocated block or NULL if an underlying malloc() failed. */
void *tre_mem_alloc_impl(tre_mem_t mem, int provided, void *provided_block,
int zero, size_t size) {
void *ptr;
if (mem->failed) {
return NULL;
}
if (mem->n < size) {
/* We need more memory than is available in the current block.
Allocate a new block. */
tre_list_t *l;
if (provided) {
if (provided_block == NULL) {
mem->failed = 1;
return NULL;
}
mem->ptr = provided_block;
mem->n = TRE_MEM_BLOCK_SIZE;
} else {
int block_size;
if (size * 8 > TRE_MEM_BLOCK_SIZE)
block_size = size * 8;
else
block_size = TRE_MEM_BLOCK_SIZE;
l = malloc(sizeof(*l));
if (l == NULL) {
mem->failed = 1;
return NULL;
}
l->data = malloc(block_size);
if (l->data == NULL) {
free(l), l = NULL;
mem->failed = 1;
return NULL;
}
l->next = NULL;
if (mem->current != NULL) mem->current->next = l;
if (mem->blocks == NULL) mem->blocks = l;
mem->current = l;
mem->ptr = l->data;
mem->n = block_size;
}
}
/* Make sure the next pointer will be aligned. */
size += ALIGN(mem->ptr + size, long);
/* Allocate from current block. */
ptr = mem->ptr;
mem->ptr += size;
mem->n -= size;
/* Set to zero if needed. */
if (zero) memset(ptr, 0, size);
return ptr;
}

248
third_party/regex/tre.inc vendored Normal file
View file

@ -0,0 +1,248 @@
/*-*- mode:c;indent-tabs-mode:nil;c-basic-offset:2;tab-width:8;coding:utf-8 -*-│
│vi: set net ft=c ts=2 sts=2 sw=2 fenc=utf-8 :vi│
╚──────────────────────────────────────────────────────────────────────────────╝
│ │
│ Musl Libc │
│ Copyright © 2005-2014 Rich Felker, et al. │
│ │
│ Permission is hereby granted, free of charge, to any person obtaining │
│ a copy of this software and associated documentation files (the │
│ "Software"), to deal in the Software without restriction, including │
│ without limitation the rights to use, copy, modify, merge, publish, │
│ distribute, sublicense, and/or sell copies of the Software, and to │
│ permit persons to whom the Software is furnished to do so, subject to │
│ the following conditions: │
│ │
│ The above copyright notice and this permission notice shall be │
│ included in all copies or substantial portions of the Software. │
│ │
│ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, │
│ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF │
│ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. │
│ IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY │
│ CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, │
│ TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE │
│ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. │
│ │
│──────────────────────────────────────────────────────────────────────────────│
│ │
│ tre-internal.h - TRE internal definitions │
│ │
│ Copyright (c) 2001-2009 Ville Laurikari <vl@iki.fi> │
│ All rights reserved. │
│ │
│ Redistribution and use in source and binary forms, with or without │
│ modification, are permitted provided that the following conditions │
│ are met: │
│ │
│ 1. Redistributions of source code must retain the above copyright │
│ notice, this list of conditions and the following disclaimer. │
│ │
│ 2. Redistributions in binary form must reproduce the above copyright │
│ notice, this list of conditions and the following disclaimer in │
│ the documentation and/or other materials provided with the │
│ distribution. │
│ │
│ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER AND CONTRIBUTORS │
│ ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT │
│ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR │
│ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT │
│ HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, │
│ SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT │
│ LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, │
│ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY │
│ THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT │
│ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE │
│ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. │
│ │
╚─────────────────────────────────────────────────────────────────────────────*/
#include "libc/alg/alg.h"
#include "libc/assert.h"
#include "libc/mem/mem.h"
#include "libc/str/str.h"
#include "third_party/regex/notice.inc"
#include "third_party/regex/regex.h"
#undef TRE_MBSTATE
#define TRE_REGEX_T_FIELD __opaque
typedef int reg_errcode_t;
typedef wchar_t tre_char_t;
#define DPRINT(msg) \
do { \
} while (0)
#define elementsof(x) (sizeof(x) / sizeof(x[0]))
#define tre_mbrtowc(pwc, s, n, ps) (mbtowc((pwc), (s), (n)))
/* Wide characters. */
typedef wint_t tre_cint_t;
#define TRE_CHAR_MAX 0x10ffff
#define tre_isalnum iswalnum
#define tre_isalpha iswalpha
#define tre_isblank iswblank
#define tre_iscntrl iswcntrl
#define tre_isdigit iswdigit
#define tre_isgraph iswgraph
#define tre_islower iswlower
#define tre_isprint iswprint
#define tre_ispunct iswpunct
#define tre_isspace iswspace
#define tre_isupper iswupper
#define tre_isxdigit iswxdigit
#define tre_tolower towlower
#define tre_toupper towupper
#define tre_strlen wcslen
/* Use system provided iswctype() and wctype(). */
typedef wctype_t tre_ctype_t;
#define tre_isctype iswctype
#define tre_ctype wctype
/* Returns number of bytes to add to (char *)ptr to make it
properly aligned for the type. */
#define ALIGN(ptr, type) \
((((long)ptr) % sizeof(type)) \
? (sizeof(type) - (((long)ptr) % sizeof(type))) \
: 0)
#undef MAX
#undef MIN
#define MAX(a, b) (((a) >= (b)) ? (a) : (b))
#define MIN(a, b) (((a) <= (b)) ? (a) : (b))
/* TNFA transition type. A TNFA state is an array of transitions,
the terminator is a transition with NULL `state'. */
typedef struct tnfa_transition tre_tnfa_transition_t;
struct tnfa_transition {
/* Range of accepted characters. */
tre_cint_t code_min;
tre_cint_t code_max;
/* Pointer to the destination state. */
tre_tnfa_transition_t *state;
/* ID number of the destination state. */
int state_id;
/* -1 terminated array of tags (or NULL). */
int *tags;
/* Assertion bitmap. */
int assertions;
/* Assertion parameters. */
union {
/* Character class assertion. */
tre_ctype_t class;
/* Back reference assertion. */
int backref;
} u;
/* Negative character class assertions. */
tre_ctype_t *neg_classes;
};
/* Assertions. */
#define ASSERT_AT_BOL 1 /* Beginning of line. */
#define ASSERT_AT_EOL 2 /* End of line. */
#define ASSERT_CHAR_CLASS 4 /* Character class in `class'. */
#define ASSERT_CHAR_CLASS_NEG 8 /* Character classes in `neg_classes'. */
#define ASSERT_AT_BOW 16 /* Beginning of word. */
#define ASSERT_AT_EOW 32 /* End of word. */
#define ASSERT_AT_WB 64 /* Word boundary. */
#define ASSERT_AT_WB_NEG 128 /* Not a word boundary. */
#define ASSERT_BACKREF 256 /* A back reference in `backref'. */
#define ASSERT_LAST 256
/* Tag directions. */
typedef enum { TRE_TAG_MINIMIZE = 0, TRE_TAG_MAXIMIZE = 1 } tre_tag_direction_t;
/* Instructions to compute submatch register values from tag values
after a successful match. */
struct tre_submatch_data {
/* Tag that gives the value for rm_so (submatch start offset). */
int so_tag;
/* Tag that gives the value for rm_eo (submatch end offset). */
int eo_tag;
/* List of submatches this submatch is contained in. */
int *parents;
};
typedef struct tre_submatch_data tre_submatch_data_t;
/* TNFA definition. */
typedef struct tnfa tre_tnfa_t;
struct tnfa {
tre_tnfa_transition_t *transitions;
unsigned int num_transitions;
tre_tnfa_transition_t *initial;
tre_tnfa_transition_t *final;
tre_submatch_data_t *submatch_data;
char *firstpos_chars;
int first_char;
unsigned int num_submatches;
tre_tag_direction_t *tag_directions;
int *minimal_tags;
int num_tags;
int num_minimals;
int end_tag;
int num_states;
int cflags;
int have_backrefs;
int have_approx;
};
/* from tre-mem.h: */
#define TRE_MEM_BLOCK_SIZE 1024
typedef struct tre_list {
void *data;
struct tre_list *next;
} tre_list_t;
typedef struct tre_mem_struct {
tre_list_t *blocks;
tre_list_t *current;
char *ptr;
size_t n;
int failed;
void **provided;
} * tre_mem_t;
#define tre_mem_new_impl __tre_mem_new_impl
#define tre_mem_alloc_impl __tre_mem_alloc_impl
#define tre_mem_destroy __tre_mem_destroy
tre_mem_t tre_mem_new_impl(int provided, void *provided_block) hidden;
void *tre_mem_alloc_impl(tre_mem_t mem, int provided, void *provided_block,
int zero, size_t size) hidden;
/* Returns a new memory allocator or NULL if out of memory. */
#define tre_mem_new() tre_mem_new_impl(0, NULL)
/* Allocates a block of `size' bytes from `mem'. Returns a pointer to the
allocated block or NULL if an underlying malloc() failed. */
#define tre_mem_alloc(mem, size) tre_mem_alloc_impl(mem, 0, NULL, 0, size)
/* Allocates a block of `size' bytes from `mem'. Returns a pointer to the
allocated block or NULL if an underlying malloc() failed. The memory
is set to zero. */
#define tre_mem_calloc(mem, size) tre_mem_alloc_impl(mem, 0, NULL, 1, size)
#ifdef TRE_USE_ALLOCA
/* alloca() versions. Like above, but memory is allocated with alloca()
instead of malloc(). */
#define tre_mem_newa() \
tre_mem_new_impl(1, alloca(sizeof(struct tre_mem_struct)))
#define tre_mem_alloca(mem, size) \
((mem)->n >= (size) \
? tre_mem_alloc_impl((mem), 1, NULL, 0, (size)) \
: tre_mem_alloc_impl((mem), 1, alloca(TRE_MEM_BLOCK_SIZE), 0, (size)))
#endif /* TRE_USE_ALLOCA */
/* Frees the memory allocator and all memory allocated with it. */
hidden void tre_mem_destroy(tre_mem_t mem);