2011-04-07 13:03:35 +00:00
|
|
|
/* $NetBSD: util.c,v 1.9 2011/02/27 17:33:37 joerg Exp $ */
|
|
|
|
/* $FreeBSD$ */
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
/* $OpenBSD: util.c,v 1.39 2010/07/02 22:18:03 tedu Exp $ */
|
|
|
|
|
|
|
|
/*-
|
2017-11-27 15:37:16 +00:00
|
|
|
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
|
|
|
|
*
|
2010-08-19 09:28:59 +00:00
|
|
|
* Copyright (c) 1999 James Howard and Dag-Erling Coïdan Smørgrav
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
* Copyright (C) 2008-2010 Gabor Kovesdan <gabor@FreeBSD.org>
|
2017-07-06 19:53:30 +00:00
|
|
|
* Copyright (C) 2017 Kyle Evans <kevans@FreeBSD.org>
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
* 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 AUTHOR 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 AUTHOR 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 <sys/cdefs.h>
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
|
|
|
#include <sys/stat.h>
|
|
|
|
#include <sys/types.h>
|
|
|
|
|
|
|
|
#include <ctype.h>
|
|
|
|
#include <err.h>
|
|
|
|
#include <errno.h>
|
|
|
|
#include <fnmatch.h>
|
|
|
|
#include <fts.h>
|
|
|
|
#include <libgen.h>
|
2010-07-29 00:11:14 +00:00
|
|
|
#include <stdbool.h>
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
#include <stdio.h>
|
|
|
|
#include <stdlib.h>
|
|
|
|
#include <string.h>
|
|
|
|
#include <unistd.h>
|
|
|
|
#include <wchar.h>
|
|
|
|
#include <wctype.h>
|
|
|
|
|
|
|
|
#include "grep.h"
|
|
|
|
|
2017-05-02 20:39:33 +00:00
|
|
|
static bool first_match = true;
|
|
|
|
|
2018-04-20 03:08:46 +00:00
|
|
|
/*
|
|
|
|
* Match printing context
|
|
|
|
*/
|
|
|
|
struct mprintc {
|
|
|
|
long long tail; /* Number of trailing lines to record */
|
|
|
|
int last_outed; /* Number of lines since last output */
|
|
|
|
bool doctx; /* Printing context? */
|
|
|
|
bool printmatch; /* Printing matches? */
|
|
|
|
bool same_file; /* Same file as previously printed? */
|
|
|
|
};
|
|
|
|
|
2018-04-20 18:06:03 +00:00
|
|
|
static void procmatch_match(struct mprintc *mc, struct parsec *pc);
|
|
|
|
static void procmatch_nomatch(struct mprintc *mc, struct parsec *pc);
|
2018-04-20 03:08:46 +00:00
|
|
|
static bool procmatches(struct mprintc *mc, struct parsec *pc, bool matched);
|
2017-08-24 01:23:33 +00:00
|
|
|
#ifdef WITH_INTERNAL_NOSPEC
|
|
|
|
static int litexec(const struct pat *pat, const char *string,
|
|
|
|
size_t nmatch, regmatch_t pmatch[]);
|
|
|
|
#endif
|
2018-06-09 18:11:46 +00:00
|
|
|
static bool procline(struct parsec *pc);
|
2017-05-02 20:39:33 +00:00
|
|
|
static void printline(struct parsec *pc, int sep);
|
|
|
|
static void printline_metadata(struct str *line, int sep);
|
2017-04-17 13:36:30 +00:00
|
|
|
|
2010-07-29 00:11:14 +00:00
|
|
|
bool
|
|
|
|
file_matching(const char *fname)
|
|
|
|
{
|
2016-07-28 15:19:47 +00:00
|
|
|
char *fname_base, *fname_buf;
|
2010-07-29 00:11:14 +00:00
|
|
|
bool ret;
|
|
|
|
|
|
|
|
ret = finclude ? false : true;
|
2016-07-28 15:19:47 +00:00
|
|
|
fname_buf = strdup(fname);
|
|
|
|
if (fname_buf == NULL)
|
|
|
|
err(2, "strdup");
|
|
|
|
fname_base = basename(fname_buf);
|
2010-07-29 00:11:14 +00:00
|
|
|
|
|
|
|
for (unsigned int i = 0; i < fpatterns; ++i) {
|
2011-04-07 13:01:03 +00:00
|
|
|
if (fnmatch(fpattern[i].pat, fname, 0) == 0 ||
|
2018-04-21 13:46:07 +00:00
|
|
|
fnmatch(fpattern[i].pat, fname_base, 0) == 0)
|
|
|
|
/*
|
|
|
|
* The last pattern matched wins exclusion/inclusion
|
|
|
|
* rights, so we can't reasonably bail out early here.
|
|
|
|
*/
|
2018-04-21 01:33:13 +00:00
|
|
|
ret = (fpattern[i].mode != EXCL_PAT);
|
2010-07-29 00:11:14 +00:00
|
|
|
}
|
2016-07-28 15:19:47 +00:00
|
|
|
free(fname_buf);
|
2010-07-29 00:11:14 +00:00
|
|
|
return (ret);
|
|
|
|
}
|
|
|
|
|
2010-08-15 22:15:04 +00:00
|
|
|
static inline bool
|
2010-07-29 00:11:14 +00:00
|
|
|
dir_matching(const char *dname)
|
|
|
|
{
|
|
|
|
bool ret;
|
|
|
|
|
|
|
|
ret = dinclude ? false : true;
|
|
|
|
|
|
|
|
for (unsigned int i = 0; i < dpatterns; ++i) {
|
2018-04-21 01:33:13 +00:00
|
|
|
if (dname != NULL && fnmatch(dpattern[i].pat, dname, 0) == 0)
|
2018-04-21 13:46:07 +00:00
|
|
|
/*
|
|
|
|
* The last pattern matched wins exclusion/inclusion
|
|
|
|
* rights, so we can't reasonably bail out early here.
|
|
|
|
*/
|
|
|
|
ret = (dpattern[i].mode != EXCL_PAT);
|
2010-07-29 00:11:14 +00:00
|
|
|
}
|
|
|
|
return (ret);
|
|
|
|
}
|
|
|
|
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
/*
|
|
|
|
* Processes a directory when a recursive search is performed with
|
|
|
|
* the -R option. Each appropriate file is passed to procfile().
|
|
|
|
*/
|
2018-06-07 18:27:58 +00:00
|
|
|
bool
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
grep_tree(char **argv)
|
|
|
|
{
|
|
|
|
FTS *fts;
|
|
|
|
FTSENT *p;
|
2018-06-07 18:36:12 +00:00
|
|
|
int fts_flags;
|
2018-06-07 18:27:58 +00:00
|
|
|
bool matched, ok;
|
2017-04-17 13:22:39 +00:00
|
|
|
const char *wd[] = { ".", NULL };
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
2018-06-07 18:27:58 +00:00
|
|
|
matched = false;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
2018-06-07 18:38:48 +00:00
|
|
|
/* This switch effectively initializes 'fts_flags' */
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
switch(linkbehave) {
|
|
|
|
case LINK_EXPLICIT:
|
|
|
|
fts_flags = FTS_COMFOLLOW;
|
|
|
|
break;
|
|
|
|
case LINK_SKIP:
|
|
|
|
fts_flags = FTS_PHYSICAL;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
fts_flags = FTS_LOGICAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
fts_flags |= FTS_NOSTAT | FTS_NOCHDIR;
|
|
|
|
|
2017-04-17 13:22:39 +00:00
|
|
|
fts = fts_open((argv[0] == NULL) ?
|
|
|
|
__DECONST(char * const *, wd) : argv, fts_flags, NULL);
|
|
|
|
if (fts == NULL)
|
2010-07-23 19:36:11 +00:00
|
|
|
err(2, "fts_open");
|
2020-12-08 23:38:26 +00:00
|
|
|
while (errno = 0, (p = fts_read(fts)) != NULL) {
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
switch (p->fts_info) {
|
|
|
|
case FTS_DNR:
|
|
|
|
/* FALLTHROUGH */
|
|
|
|
case FTS_ERR:
|
2011-12-07 12:25:28 +00:00
|
|
|
file_err = true;
|
2011-11-28 20:04:26 +00:00
|
|
|
if(!sflag)
|
|
|
|
warnx("%s: %s", p->fts_path, strerror(p->fts_errno));
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
break;
|
|
|
|
case FTS_D:
|
|
|
|
/* FALLTHROUGH */
|
|
|
|
case FTS_DP:
|
2011-08-17 13:58:39 +00:00
|
|
|
if (dexclude || dinclude)
|
|
|
|
if (!dir_matching(p->fts_name) ||
|
|
|
|
!dir_matching(p->fts_path))
|
|
|
|
fts_set(fts, p, FTS_SKIP);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
break;
|
|
|
|
case FTS_DC:
|
|
|
|
/* Print a warning for recursive directory loop */
|
|
|
|
warnx("warning: %s: recursive directory loop",
|
2018-04-21 01:33:13 +00:00
|
|
|
p->fts_path);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
/* Check for file exclusion/inclusion */
|
|
|
|
ok = true;
|
2010-07-29 00:11:14 +00:00
|
|
|
if (fexclude || finclude)
|
|
|
|
ok &= file_matching(p->fts_path);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
2018-06-07 18:27:58 +00:00
|
|
|
if (ok && procfile(p->fts_path))
|
|
|
|
matched = true;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2020-12-08 23:38:26 +00:00
|
|
|
if (errno != 0)
|
|
|
|
err(2, "fts_read");
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
2010-07-23 19:36:11 +00:00
|
|
|
fts_close(fts);
|
2018-06-07 18:27:58 +00:00
|
|
|
return (matched);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
|
|
|
|
2018-04-20 18:06:03 +00:00
|
|
|
static void
|
|
|
|
procmatch_match(struct mprintc *mc, struct parsec *pc)
|
|
|
|
{
|
|
|
|
|
|
|
|
if (mc->doctx) {
|
|
|
|
if (!first_match && (!mc->same_file || mc->last_outed > 0))
|
|
|
|
printf("--\n");
|
|
|
|
if (Bflag > 0)
|
|
|
|
printqueue();
|
|
|
|
mc->tail = Aflag;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Print the matching line, but only if not quiet/binary */
|
|
|
|
if (mc->printmatch) {
|
|
|
|
printline(pc, ':');
|
|
|
|
while (pc->matchidx >= MAX_MATCHES) {
|
|
|
|
/* Reset matchidx and try again */
|
|
|
|
pc->matchidx = 0;
|
2019-09-25 17:14:43 +00:00
|
|
|
if (procline(pc) == !vflag)
|
2018-04-20 18:06:03 +00:00
|
|
|
printline(pc, ':');
|
|
|
|
else
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
first_match = false;
|
|
|
|
mc->same_file = true;
|
|
|
|
mc->last_outed = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
procmatch_nomatch(struct mprintc *mc, struct parsec *pc)
|
|
|
|
{
|
|
|
|
|
|
|
|
/* Deal with any -A context as needed */
|
|
|
|
if (mc->tail > 0) {
|
|
|
|
grep_printline(&pc->ln, '-');
|
|
|
|
mc->tail--;
|
|
|
|
if (Bflag > 0)
|
|
|
|
clearqueue();
|
|
|
|
} else if (Bflag == 0 || (Bflag > 0 && enqueue(&pc->ln)))
|
|
|
|
/*
|
|
|
|
* Enqueue non-matching lines for -B context. If we're not
|
|
|
|
* actually doing -B context or if the enqueue resulted in a
|
|
|
|
* line being rotated out, then go ahead and increment
|
|
|
|
* last_outed to signify a gap between context/match.
|
|
|
|
*/
|
|
|
|
++mc->last_outed;
|
|
|
|
}
|
|
|
|
|
2018-04-20 03:08:46 +00:00
|
|
|
/*
|
|
|
|
* Process any matches in the current parsing context, return a boolean
|
|
|
|
* indicating whether we should halt any further processing or not. 'true' to
|
|
|
|
* continue processing, 'false' to halt.
|
|
|
|
*/
|
|
|
|
static bool
|
|
|
|
procmatches(struct mprintc *mc, struct parsec *pc, bool matched)
|
|
|
|
{
|
|
|
|
|
grep: fix -A handling in conjunction with -m match limitation
The basic issue here is that grep, when given -m 1, would stop all
line processing once it hit the match count and exit immediately. The
problem with exiting immediately is that -A processing only happens when
subsequent lines are processed and do not match.
The fix here is relatively easy; when bsdgrep matches a line, it resets
the 'tail' of the matching context to the value supplied to -A and
dumps anything that's been queued up for -B. After the current line has
been printed and tail is reset, we check our mcount and do what's
needed. Therefore, at the time that we decide we're doing nothing, we
know that 'tail' of the context is correct and we can simply continue
on if there's still more to pick up.
With this change, we still bail out immediately if there's been no -A
flag. If -A was supplied, we signal that we should continue on. However,
subsequent lines will not even bothere to try and process the line. We
have reached the match count, so even if the next line would match then
we must process it if it hadn't. Thus, the loop in procfile() can
short-circuit and just process the line as a non-match until
procmatches() indicates that it's safe to stop.
A test has been added to reflect both that we should be picking up the
next line and that the next line should be considered a non-match even
if it should have been.
PR: 253350
MFC-after: 3 days
2021-02-08 18:31:17 +00:00
|
|
|
if (mflag && mcount <= 0) {
|
|
|
|
/*
|
|
|
|
* We already hit our match count, but we need to keep dumping
|
|
|
|
* lines until we've lost our tail.
|
|
|
|
*/
|
|
|
|
grep_printline(&pc->ln, '-');
|
|
|
|
mc->tail--;
|
|
|
|
return (mc->tail != 0);
|
|
|
|
}
|
|
|
|
|
2018-04-20 03:29:06 +00:00
|
|
|
/*
|
|
|
|
* XXX TODO: This should loop over pc->matches and handle things on a
|
|
|
|
* line-by-line basis, setting up a `struct str` as needed.
|
|
|
|
*/
|
2018-04-20 03:08:46 +00:00
|
|
|
/* Deal with any -B context or context separators */
|
2018-04-20 03:11:51 +00:00
|
|
|
if (matched) {
|
2018-04-20 18:06:03 +00:00
|
|
|
procmatch_match(mc, pc);
|
2018-04-20 03:08:46 +00:00
|
|
|
|
2018-04-20 03:11:51 +00:00
|
|
|
/* Count the matches if we have a match limit */
|
|
|
|
if (mflag) {
|
|
|
|
/* XXX TODO: Decrement by number of matched lines */
|
|
|
|
mcount -= 1;
|
2018-06-08 01:25:07 +00:00
|
|
|
if (mcount <= 0)
|
grep: fix -A handling in conjunction with -m match limitation
The basic issue here is that grep, when given -m 1, would stop all
line processing once it hit the match count and exit immediately. The
problem with exiting immediately is that -A processing only happens when
subsequent lines are processed and do not match.
The fix here is relatively easy; when bsdgrep matches a line, it resets
the 'tail' of the matching context to the value supplied to -A and
dumps anything that's been queued up for -B. After the current line has
been printed and tail is reset, we check our mcount and do what's
needed. Therefore, at the time that we decide we're doing nothing, we
know that 'tail' of the context is correct and we can simply continue
on if there's still more to pick up.
With this change, we still bail out immediately if there's been no -A
flag. If -A was supplied, we signal that we should continue on. However,
subsequent lines will not even bothere to try and process the line. We
have reached the match count, so even if the next line would match then
we must process it if it hadn't. Thus, the loop in procfile() can
short-circuit and just process the line as a non-match until
procmatches() indicates that it's safe to stop.
A test has been added to reflect both that we should be picking up the
next line and that the next line should be considered a non-match even
if it should have been.
PR: 253350
MFC-after: 3 days
2021-02-08 18:31:17 +00:00
|
|
|
return (mc->tail != 0);
|
2018-04-20 03:11:51 +00:00
|
|
|
}
|
2018-04-20 18:06:03 +00:00
|
|
|
} else if (mc->doctx)
|
|
|
|
procmatch_nomatch(mc, pc);
|
2018-04-20 03:08:46 +00:00
|
|
|
|
|
|
|
return (true);
|
|
|
|
}
|
|
|
|
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
/*
|
|
|
|
* Opens a file and processes it. Each file is processed line-by-line
|
|
|
|
* passing the lines to procline().
|
|
|
|
*/
|
2018-06-07 18:27:58 +00:00
|
|
|
bool
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
procfile(const char *fn)
|
|
|
|
{
|
2017-05-02 20:39:33 +00:00
|
|
|
struct parsec pc;
|
2018-04-20 03:08:46 +00:00
|
|
|
struct mprintc mc;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
struct file *f;
|
|
|
|
struct stat sb;
|
|
|
|
mode_t s;
|
2018-06-09 18:11:46 +00:00
|
|
|
int lines;
|
|
|
|
bool line_matched;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
|
|
|
if (strcmp(fn, "-") == 0) {
|
2018-06-06 23:12:35 +00:00
|
|
|
fn = label != NULL ? label : errstr[1];
|
2010-08-18 17:40:10 +00:00
|
|
|
f = grep_open(NULL);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
} else {
|
2018-04-21 01:02:35 +00:00
|
|
|
if (stat(fn, &sb) == 0) {
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
/* Check if we need to process the file */
|
|
|
|
s = sb.st_mode & S_IFMT;
|
2018-04-21 01:02:35 +00:00
|
|
|
if (dirbehave == DIR_SKIP && s == S_IFDIR)
|
2018-06-07 18:27:58 +00:00
|
|
|
return (false);
|
2018-04-21 01:02:35 +00:00
|
|
|
if (devbehave == DEV_SKIP && (s == S_IFIFO ||
|
|
|
|
s == S_IFCHR || s == S_IFBLK || s == S_IFSOCK))
|
2018-06-07 18:27:58 +00:00
|
|
|
return (false);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
|
|
|
f = grep_open(fn);
|
|
|
|
}
|
|
|
|
if (f == NULL) {
|
2011-12-07 12:25:28 +00:00
|
|
|
file_err = true;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
if (!sflag)
|
|
|
|
warn("%s", fn);
|
2018-06-07 18:27:58 +00:00
|
|
|
return (false);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
|
|
|
|
2018-04-21 14:58:45 +00:00
|
|
|
pc.ln.file = grep_strdup(fn);
|
2017-05-02 20:39:33 +00:00
|
|
|
pc.ln.line_no = 0;
|
|
|
|
pc.ln.len = 0;
|
2017-05-20 11:20:03 +00:00
|
|
|
pc.ln.boff = 0;
|
2017-05-02 20:39:33 +00:00
|
|
|
pc.ln.off = -1;
|
|
|
|
pc.binary = f->binary;
|
2018-06-08 01:25:07 +00:00
|
|
|
pc.cntlines = false;
|
2018-04-20 03:08:46 +00:00
|
|
|
memset(&mc, 0, sizeof(mc));
|
|
|
|
mc.printmatch = true;
|
2017-05-05 17:35:05 +00:00
|
|
|
if ((pc.binary && binbehave == BINFILE_BIN) || cflag || qflag ||
|
|
|
|
lflag || Lflag)
|
2018-04-20 03:08:46 +00:00
|
|
|
mc.printmatch = false;
|
|
|
|
if (mc.printmatch && (Aflag != 0 || Bflag != 0))
|
|
|
|
mc.doctx = true;
|
2018-06-08 01:25:07 +00:00
|
|
|
if (mc.printmatch && (Aflag != 0 || Bflag != 0 || mflag || nflag))
|
|
|
|
pc.cntlines = true;
|
2017-05-03 13:47:02 +00:00
|
|
|
mcount = mlimit;
|
|
|
|
|
2018-06-07 18:27:58 +00:00
|
|
|
for (lines = 0; lines == 0 || !(lflag || qflag); ) {
|
2018-04-20 03:29:06 +00:00
|
|
|
/*
|
|
|
|
* XXX TODO: We need to revisit this in a chunking world. We're
|
|
|
|
* not going to be doing per-line statistics because of the
|
|
|
|
* overhead involved. procmatches can figure that stuff out as
|
|
|
|
* needed. */
|
2017-05-20 11:20:03 +00:00
|
|
|
/* Reset per-line statistics */
|
|
|
|
pc.printed = 0;
|
2017-05-02 20:39:33 +00:00
|
|
|
pc.matchidx = 0;
|
2017-05-20 03:51:31 +00:00
|
|
|
pc.lnstart = 0;
|
2017-05-20 11:20:03 +00:00
|
|
|
pc.ln.boff = 0;
|
2017-05-02 20:39:33 +00:00
|
|
|
pc.ln.off += pc.ln.len + 1;
|
2018-04-20 03:29:06 +00:00
|
|
|
/* XXX TODO: Grab a chunk */
|
2018-06-08 01:25:07 +00:00
|
|
|
if ((pc.ln.dat = grep_fgetln(f, &pc)) == NULL ||
|
2017-07-25 01:50:37 +00:00
|
|
|
pc.ln.len == 0)
|
|
|
|
break;
|
2017-05-02 20:39:33 +00:00
|
|
|
|
|
|
|
if (pc.ln.len > 0 && pc.ln.dat[pc.ln.len - 1] == fileeol)
|
|
|
|
--pc.ln.len;
|
|
|
|
pc.ln.line_no++;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
|
|
|
/* Return if we need to skip a binary file */
|
2017-05-02 20:39:33 +00:00
|
|
|
if (pc.binary && binbehave == BINFILE_SKIP) {
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
grep_close(f);
|
2017-05-02 20:39:33 +00:00
|
|
|
free(pc.ln.file);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
free(f);
|
|
|
|
return (0);
|
|
|
|
}
|
2017-04-17 13:36:30 +00:00
|
|
|
|
grep: fix -A handling in conjunction with -m match limitation
The basic issue here is that grep, when given -m 1, would stop all
line processing once it hit the match count and exit immediately. The
problem with exiting immediately is that -A processing only happens when
subsequent lines are processed and do not match.
The fix here is relatively easy; when bsdgrep matches a line, it resets
the 'tail' of the matching context to the value supplied to -A and
dumps anything that's been queued up for -B. After the current line has
been printed and tail is reset, we check our mcount and do what's
needed. Therefore, at the time that we decide we're doing nothing, we
know that 'tail' of the context is correct and we can simply continue
on if there's still more to pick up.
With this change, we still bail out immediately if there's been no -A
flag. If -A was supplied, we signal that we should continue on. However,
subsequent lines will not even bothere to try and process the line. We
have reached the match count, so even if the next line would match then
we must process it if it hadn't. Thus, the loop in procfile() can
short-circuit and just process the line as a non-match until
procmatches() indicates that it's safe to stop.
A test has been added to reflect both that we should be picking up the
next line and that the next line should be considered a non-match even
if it should have been.
PR: 253350
MFC-after: 3 days
2021-02-08 18:31:17 +00:00
|
|
|
if (mflag && mcount <= 0) {
|
|
|
|
/*
|
|
|
|
* Short-circuit, already hit match count and now we're
|
|
|
|
* just picking up any remaining pieces.
|
|
|
|
*/
|
|
|
|
if (!procmatches(&mc, &pc, false))
|
|
|
|
break;
|
|
|
|
continue;
|
|
|
|
}
|
2019-09-25 17:14:43 +00:00
|
|
|
line_matched = procline(&pc) == !vflag;
|
2018-06-09 18:11:46 +00:00
|
|
|
if (line_matched)
|
2018-06-07 18:27:58 +00:00
|
|
|
++lines;
|
2017-05-02 20:39:33 +00:00
|
|
|
|
2018-04-20 03:08:46 +00:00
|
|
|
/* Halt processing if we hit our match limit */
|
2018-06-09 18:11:46 +00:00
|
|
|
if (!procmatches(&mc, &pc, line_matched))
|
2018-04-20 03:08:46 +00:00
|
|
|
break;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
|
|
|
if (Bflag > 0)
|
|
|
|
clearqueue();
|
|
|
|
grep_close(f);
|
|
|
|
|
|
|
|
if (cflag) {
|
|
|
|
if (!hflag)
|
2017-05-02 20:39:33 +00:00
|
|
|
printf("%s:", pc.ln.file);
|
2018-06-07 18:27:58 +00:00
|
|
|
printf("%u\n", lines);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
2018-06-07 18:27:58 +00:00
|
|
|
if (lflag && !qflag && lines != 0)
|
2011-11-28 20:00:31 +00:00
|
|
|
printf("%s%c", fn, nullflag ? 0 : '\n');
|
2018-06-07 18:27:58 +00:00
|
|
|
if (Lflag && !qflag && lines == 0)
|
2011-11-28 20:00:31 +00:00
|
|
|
printf("%s%c", fn, nullflag ? 0 : '\n');
|
2018-06-07 18:27:58 +00:00
|
|
|
if (lines != 0 && !cflag && !lflag && !Lflag &&
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
binbehave == BINFILE_BIN && f->binary && !qflag)
|
2018-06-06 23:12:35 +00:00
|
|
|
printf(errstr[7], fn);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
2017-05-02 20:39:33 +00:00
|
|
|
free(pc.ln.file);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
free(f);
|
2018-06-07 18:27:58 +00:00
|
|
|
return (lines != 0);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
|
|
|
|
2017-08-24 01:23:33 +00:00
|
|
|
#ifdef WITH_INTERNAL_NOSPEC
|
|
|
|
/*
|
|
|
|
* Internal implementation of literal string search within a string, modeled
|
|
|
|
* after regexec(3), for use when the regex(3) implementation doesn't offer
|
|
|
|
* either REG_NOSPEC or REG_LITERAL. This does not apply in the default FreeBSD
|
|
|
|
* config, but in other scenarios such as building against libgnuregex or on
|
|
|
|
* some non-FreeBSD OSes.
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
litexec(const struct pat *pat, const char *string, size_t nmatch,
|
|
|
|
regmatch_t pmatch[])
|
|
|
|
{
|
|
|
|
char *(*strstr_fn)(const char *, const char *);
|
|
|
|
char *sub, *subject;
|
|
|
|
const char *search;
|
|
|
|
size_t idx, n, ofs, stringlen;
|
|
|
|
|
|
|
|
if (cflags & REG_ICASE)
|
|
|
|
strstr_fn = strcasestr;
|
|
|
|
else
|
|
|
|
strstr_fn = strstr;
|
|
|
|
idx = 0;
|
|
|
|
ofs = pmatch[0].rm_so;
|
|
|
|
stringlen = pmatch[0].rm_eo;
|
|
|
|
if (ofs >= stringlen)
|
|
|
|
return (REG_NOMATCH);
|
|
|
|
subject = strndup(string, stringlen);
|
|
|
|
if (subject == NULL)
|
|
|
|
return (REG_ESPACE);
|
|
|
|
for (n = 0; ofs < stringlen;) {
|
|
|
|
search = (subject + ofs);
|
|
|
|
if ((unsigned long)pat->len > strlen(search))
|
|
|
|
break;
|
|
|
|
sub = strstr_fn(search, pat->pat);
|
|
|
|
/*
|
|
|
|
* Ignoring the empty string possibility due to context: grep optimizes
|
|
|
|
* for empty patterns and will never reach this point.
|
|
|
|
*/
|
|
|
|
if (sub == NULL)
|
|
|
|
break;
|
|
|
|
++n;
|
|
|
|
/* Fill in pmatch if necessary */
|
|
|
|
if (nmatch > 0) {
|
|
|
|
pmatch[idx].rm_so = ofs + (sub - search);
|
|
|
|
pmatch[idx].rm_eo = pmatch[idx].rm_so + pat->len;
|
|
|
|
if (++idx == nmatch)
|
|
|
|
break;
|
|
|
|
ofs = pmatch[idx].rm_so + 1;
|
|
|
|
} else
|
|
|
|
/* We only needed to know if we match or not */
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
free(subject);
|
|
|
|
if (n > 0 && nmatch > 0)
|
|
|
|
for (n = idx; n < nmatch; ++n)
|
|
|
|
pmatch[n].rm_so = pmatch[n].rm_eo = -1;
|
|
|
|
|
|
|
|
return (n > 0 ? 0 : REG_NOMATCH);
|
|
|
|
}
|
|
|
|
#endif /* WITH_INTERNAL_NOSPEC */
|
|
|
|
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
#define iswword(x) (iswalnum((x)) || (x) == L'_')
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Processes a line comparing it with the specified patterns. Each pattern
|
|
|
|
* is looped to be compared along with the full string, saving each and every
|
|
|
|
* match, which is necessary to colorize the output and to count the
|
|
|
|
* matches. The matching lines are passed to printline() to display the
|
|
|
|
* appropriate output.
|
|
|
|
*/
|
2018-06-09 18:11:46 +00:00
|
|
|
static bool
|
2017-05-02 20:39:33 +00:00
|
|
|
procline(struct parsec *pc)
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
{
|
2017-05-02 20:39:33 +00:00
|
|
|
regmatch_t pmatch, lastmatch, chkmatch;
|
|
|
|
wchar_t wbegin, wend;
|
2017-05-20 03:51:31 +00:00
|
|
|
size_t st, nst;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
unsigned int i;
|
2018-06-09 18:11:46 +00:00
|
|
|
int r = 0, leflags = eflags;
|
2017-05-02 20:39:33 +00:00
|
|
|
size_t startm = 0, matchidx;
|
2017-05-20 03:51:31 +00:00
|
|
|
unsigned int retry;
|
2018-06-09 18:11:46 +00:00
|
|
|
bool lastmatched, matched;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
2017-05-02 20:39:33 +00:00
|
|
|
matchidx = pc->matchidx;
|
|
|
|
|
2021-02-04 21:26:45 +00:00
|
|
|
/* Null pattern shortcuts. */
|
2019-09-25 17:14:43 +00:00
|
|
|
if (matchall) {
|
2021-02-04 21:26:45 +00:00
|
|
|
if (xflag && pc->ln.len == 0) {
|
|
|
|
/* Matches empty lines (-x). */
|
2018-06-09 18:11:46 +00:00
|
|
|
return (true);
|
2021-02-04 21:26:45 +00:00
|
|
|
} else if (!wflag && !xflag) {
|
|
|
|
/* Matches every line (no -w or -x). */
|
2018-06-09 18:11:46 +00:00
|
|
|
return (true);
|
2021-02-04 21:26:45 +00:00
|
|
|
}
|
2019-09-25 17:14:43 +00:00
|
|
|
|
|
|
|
/*
|
2021-02-04 21:26:45 +00:00
|
|
|
* If we only have the NULL pattern, whether we match or not
|
|
|
|
* depends on if we got here with -w or -x. If either is set,
|
|
|
|
* the answer is no. If we have other patterns, we'll defer
|
|
|
|
* to them.
|
2019-09-25 17:14:43 +00:00
|
|
|
*/
|
2021-02-04 21:26:45 +00:00
|
|
|
if (patterns == 0) {
|
|
|
|
return (!(wflag || xflag));
|
|
|
|
}
|
|
|
|
} else if (patterns == 0) {
|
|
|
|
/* Pattern file with no patterns. */
|
|
|
|
return (false);
|
2019-09-25 17:14:43 +00:00
|
|
|
}
|
2017-05-02 20:39:33 +00:00
|
|
|
|
2018-06-09 18:11:46 +00:00
|
|
|
matched = false;
|
2017-05-20 03:51:31 +00:00
|
|
|
st = pc->lnstart;
|
|
|
|
nst = 0;
|
2017-04-04 13:34:19 +00:00
|
|
|
/* Initialize to avoid a false positive warning from GCC. */
|
|
|
|
lastmatch.rm_so = lastmatch.rm_eo = 0;
|
|
|
|
|
2011-10-05 09:56:43 +00:00
|
|
|
/* Loop to process the whole line */
|
2017-05-02 20:39:33 +00:00
|
|
|
while (st <= pc->ln.len) {
|
2018-06-09 18:11:46 +00:00
|
|
|
lastmatched = false;
|
2017-05-02 20:39:33 +00:00
|
|
|
startm = matchidx;
|
2017-05-02 02:32:10 +00:00
|
|
|
retry = 0;
|
2017-05-26 02:30:26 +00:00
|
|
|
if (st > 0 && pc->ln.dat[st - 1] != fileeol)
|
2017-04-03 23:16:51 +00:00
|
|
|
leflags |= REG_NOTBOL;
|
2011-10-05 09:56:43 +00:00
|
|
|
/* Loop to compare with all the patterns */
|
|
|
|
for (i = 0; i < patterns; i++) {
|
2017-04-03 23:16:51 +00:00
|
|
|
pmatch.rm_so = st;
|
2017-05-02 20:39:33 +00:00
|
|
|
pmatch.rm_eo = pc->ln.len;
|
2017-08-24 01:23:33 +00:00
|
|
|
#ifdef WITH_INTERNAL_NOSPEC
|
|
|
|
if (grepbehave == GREP_FIXED)
|
|
|
|
r = litexec(&pattern[i], pc->ln.dat, 1, &pmatch);
|
|
|
|
else
|
|
|
|
#endif
|
2018-05-04 03:13:25 +00:00
|
|
|
r = regexec(&r_pattern[i], pc->ln.dat, 1, &pmatch,
|
|
|
|
leflags);
|
2017-05-02 20:39:33 +00:00
|
|
|
if (r != 0)
|
2011-10-05 09:56:43 +00:00
|
|
|
continue;
|
|
|
|
/* Check for full match */
|
2017-05-02 20:39:33 +00:00
|
|
|
if (xflag && (pmatch.rm_so != 0 ||
|
|
|
|
(size_t)pmatch.rm_eo != pc->ln.len))
|
|
|
|
continue;
|
2011-10-05 09:56:43 +00:00
|
|
|
/* Check for whole word match */
|
2017-05-02 20:39:33 +00:00
|
|
|
if (wflag) {
|
2011-10-05 09:56:43 +00:00
|
|
|
wbegin = wend = L' ';
|
|
|
|
if (pmatch.rm_so != 0 &&
|
2017-05-02 20:39:33 +00:00
|
|
|
sscanf(&pc->ln.dat[pmatch.rm_so - 1],
|
2011-10-05 09:56:43 +00:00
|
|
|
"%lc", &wbegin) != 1)
|
|
|
|
r = REG_NOMATCH;
|
|
|
|
else if ((size_t)pmatch.rm_eo !=
|
2017-05-02 20:39:33 +00:00
|
|
|
pc->ln.len &&
|
|
|
|
sscanf(&pc->ln.dat[pmatch.rm_eo],
|
2011-10-05 09:56:43 +00:00
|
|
|
"%lc", &wend) != 1)
|
|
|
|
r = REG_NOMATCH;
|
|
|
|
else if (iswword(wbegin) ||
|
|
|
|
iswword(wend))
|
|
|
|
r = REG_NOMATCH;
|
2017-05-02 02:32:10 +00:00
|
|
|
/*
|
|
|
|
* If we're doing whole word matching and we
|
|
|
|
* matched once, then we should try the pattern
|
2017-05-02 20:39:33 +00:00
|
|
|
* again after advancing just past the start of
|
2017-05-02 02:32:10 +00:00
|
|
|
* the earliest match. This allows the pattern
|
|
|
|
* to match later on in the line and possibly
|
|
|
|
* still match a whole word.
|
|
|
|
*/
|
|
|
|
if (r == REG_NOMATCH &&
|
2017-05-20 03:51:31 +00:00
|
|
|
(retry == pc->lnstart ||
|
2017-08-17 13:40:45 +00:00
|
|
|
(unsigned int)pmatch.rm_so + 1 < retry))
|
2017-05-02 02:32:10 +00:00
|
|
|
retry = pmatch.rm_so + 1;
|
2017-05-02 20:39:33 +00:00
|
|
|
if (r == REG_NOMATCH)
|
|
|
|
continue;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
2018-06-09 18:11:46 +00:00
|
|
|
lastmatched = true;
|
2017-05-02 20:39:33 +00:00
|
|
|
lastmatch = pmatch;
|
|
|
|
|
|
|
|
if (matchidx == 0)
|
2018-06-09 18:11:46 +00:00
|
|
|
matched = true;
|
2017-05-02 20:39:33 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Replace previous match if the new one is earlier
|
|
|
|
* and/or longer. This will lead to some amount of
|
|
|
|
* extra work if -o/--color are specified, but it's
|
|
|
|
* worth it from a correctness point of view.
|
|
|
|
*/
|
|
|
|
if (matchidx > startm) {
|
|
|
|
chkmatch = pc->matches[matchidx - 1];
|
|
|
|
if (pmatch.rm_so < chkmatch.rm_so ||
|
|
|
|
(pmatch.rm_so == chkmatch.rm_so &&
|
|
|
|
(pmatch.rm_eo - pmatch.rm_so) >
|
|
|
|
(chkmatch.rm_eo - chkmatch.rm_so))) {
|
|
|
|
pc->matches[matchidx - 1] = pmatch;
|
|
|
|
nst = pmatch.rm_eo;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
/* Advance as normal if not */
|
|
|
|
pc->matches[matchidx++] = pmatch;
|
|
|
|
nst = pmatch.rm_eo;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
2017-05-02 20:39:33 +00:00
|
|
|
/* avoid excessive matching - skip further patterns */
|
|
|
|
if ((color == NULL && !oflag) || qflag || lflag ||
|
2017-05-26 02:30:26 +00:00
|
|
|
matchidx >= MAX_MATCHES) {
|
2017-05-20 03:51:31 +00:00
|
|
|
pc->lnstart = nst;
|
2018-06-09 18:11:46 +00:00
|
|
|
lastmatched = false;
|
2017-05-02 20:39:33 +00:00
|
|
|
break;
|
2017-05-20 03:51:31 +00:00
|
|
|
}
|
2011-10-05 09:56:43 +00:00
|
|
|
}
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
2017-05-02 02:32:10 +00:00
|
|
|
/*
|
|
|
|
* Advance to just past the start of the earliest match, try
|
|
|
|
* again just in case we still have a chance to match later in
|
|
|
|
* the string.
|
|
|
|
*/
|
2018-06-09 18:11:46 +00:00
|
|
|
if (!lastmatched && retry > pc->lnstart) {
|
2017-05-02 02:32:10 +00:00
|
|
|
st = retry;
|
|
|
|
continue;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
2011-10-05 09:56:43 +00:00
|
|
|
|
2018-04-20 03:29:06 +00:00
|
|
|
/* XXX TODO: We will need to keep going, since we're chunky */
|
2011-10-05 09:56:43 +00:00
|
|
|
/* One pass if we are not recording matches */
|
2014-08-18 12:29:28 +00:00
|
|
|
if (!wflag && ((color == NULL && !oflag) || qflag || lflag || Lflag))
|
2011-10-05 09:56:43 +00:00
|
|
|
break;
|
|
|
|
|
2017-04-03 23:16:51 +00:00
|
|
|
/* If we didn't have any matches or REG_NOSUB set */
|
2018-06-09 18:11:46 +00:00
|
|
|
if (!lastmatched || (cflags & REG_NOSUB))
|
2017-05-02 20:39:33 +00:00
|
|
|
nst = pc->ln.len;
|
2017-04-03 23:16:51 +00:00
|
|
|
|
2018-06-09 18:11:46 +00:00
|
|
|
if (!lastmatched)
|
2017-04-03 23:16:51 +00:00
|
|
|
/* No matches */
|
|
|
|
break;
|
|
|
|
else if (st == nst && lastmatch.rm_so == lastmatch.rm_eo)
|
|
|
|
/* Zero-length match -- advance one more so we don't get stuck */
|
|
|
|
nst++;
|
|
|
|
|
|
|
|
/* Advance st based on previous matches */
|
|
|
|
st = nst;
|
2017-05-20 03:51:31 +00:00
|
|
|
pc->lnstart = st;
|
2011-10-05 09:56:43 +00:00
|
|
|
}
|
|
|
|
|
2017-05-02 20:39:33 +00:00
|
|
|
/* Reflect the new matchidx in the context */
|
|
|
|
pc->matchidx = matchidx;
|
2018-06-09 18:11:46 +00:00
|
|
|
return matched;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Safe malloc() for internal use.
|
|
|
|
*/
|
|
|
|
void *
|
|
|
|
grep_malloc(size_t size)
|
|
|
|
{
|
|
|
|
void *ptr;
|
|
|
|
|
|
|
|
if ((ptr = malloc(size)) == NULL)
|
|
|
|
err(2, "malloc");
|
|
|
|
return (ptr);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Safe calloc() for internal use.
|
|
|
|
*/
|
|
|
|
void *
|
|
|
|
grep_calloc(size_t nmemb, size_t size)
|
|
|
|
{
|
|
|
|
void *ptr;
|
|
|
|
|
|
|
|
if ((ptr = calloc(nmemb, size)) == NULL)
|
|
|
|
err(2, "calloc");
|
|
|
|
return (ptr);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Safe realloc() for internal use.
|
|
|
|
*/
|
|
|
|
void *
|
|
|
|
grep_realloc(void *ptr, size_t size)
|
|
|
|
{
|
|
|
|
|
|
|
|
if ((ptr = realloc(ptr, size)) == NULL)
|
|
|
|
err(2, "realloc");
|
|
|
|
return (ptr);
|
|
|
|
}
|
|
|
|
|
2010-07-29 00:11:14 +00:00
|
|
|
/*
|
|
|
|
* Safe strdup() for internal use.
|
|
|
|
*/
|
|
|
|
char *
|
|
|
|
grep_strdup(const char *str)
|
|
|
|
{
|
|
|
|
char *ret;
|
|
|
|
|
|
|
|
if ((ret = strdup(str)) == NULL)
|
|
|
|
err(2, "strdup");
|
|
|
|
return (ret);
|
|
|
|
}
|
|
|
|
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
/*
|
2017-05-02 20:39:33 +00:00
|
|
|
* Print an entire line as-is, there are no inline matches to consider. This is
|
|
|
|
* used for printing context.
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
*/
|
2017-05-02 20:39:33 +00:00
|
|
|
void grep_printline(struct str *line, int sep) {
|
|
|
|
printline_metadata(line, sep);
|
|
|
|
fwrite(line->dat, line->len, 1, stdout);
|
|
|
|
putchar(fileeol);
|
|
|
|
}
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
|
2017-05-02 20:39:33 +00:00
|
|
|
static void
|
|
|
|
printline_metadata(struct str *line, int sep)
|
|
|
|
{
|
|
|
|
bool printsep;
|
2017-04-03 23:16:51 +00:00
|
|
|
|
2017-05-02 20:39:33 +00:00
|
|
|
printsep = false;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
if (!hflag) {
|
2011-11-28 20:00:31 +00:00
|
|
|
if (!nullflag) {
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
fputs(line->file, stdout);
|
2017-05-02 20:39:33 +00:00
|
|
|
printsep = true;
|
2011-11-28 20:00:31 +00:00
|
|
|
} else {
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
printf("%s", line->file);
|
|
|
|
putchar(0);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (nflag) {
|
2017-05-02 20:39:33 +00:00
|
|
|
if (printsep)
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
putchar(sep);
|
|
|
|
printf("%d", line->line_no);
|
2017-05-02 20:39:33 +00:00
|
|
|
printsep = true;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
|
|
|
if (bflag) {
|
2017-05-02 20:39:33 +00:00
|
|
|
if (printsep)
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
putchar(sep);
|
2017-05-20 11:20:03 +00:00
|
|
|
printf("%lld", (long long)(line->off + line->boff));
|
2017-05-02 20:39:33 +00:00
|
|
|
printsep = true;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|
2017-05-02 20:39:33 +00:00
|
|
|
if (printsep)
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
putchar(sep);
|
2017-05-02 20:39:33 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Prints a matching line according to the command line options.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
printline(struct parsec *pc, int sep)
|
|
|
|
{
|
|
|
|
size_t a = 0;
|
|
|
|
size_t i, matchidx;
|
|
|
|
regmatch_t match;
|
|
|
|
|
|
|
|
/* If matchall, everything matches but don't actually print for -o */
|
|
|
|
if (oflag && matchall)
|
|
|
|
return;
|
|
|
|
|
|
|
|
matchidx = pc->matchidx;
|
|
|
|
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
/* --color and -o */
|
2017-05-02 20:39:33 +00:00
|
|
|
if ((oflag || color) && matchidx > 0) {
|
2017-05-20 11:20:03 +00:00
|
|
|
/* Only print metadata once per line if --color */
|
|
|
|
if (!oflag && pc->printed == 0)
|
|
|
|
printline_metadata(&pc->ln, sep);
|
2017-05-02 20:39:33 +00:00
|
|
|
for (i = 0; i < matchidx; i++) {
|
|
|
|
match = pc->matches[i];
|
2017-04-17 14:59:55 +00:00
|
|
|
/* Don't output zero length matches */
|
2017-05-02 20:39:33 +00:00
|
|
|
if (match.rm_so == match.rm_eo)
|
2017-04-17 14:59:55 +00:00
|
|
|
continue;
|
2017-05-20 11:20:03 +00:00
|
|
|
/*
|
|
|
|
* Metadata is printed on a per-line basis, so every
|
|
|
|
* match gets file metadata with the -o flag.
|
|
|
|
*/
|
|
|
|
if (oflag) {
|
|
|
|
pc->ln.boff = match.rm_so;
|
|
|
|
printline_metadata(&pc->ln, sep);
|
|
|
|
} else
|
2017-05-02 20:39:33 +00:00
|
|
|
fwrite(pc->ln.dat + a, match.rm_so - a, 1,
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
stdout);
|
2017-05-02 20:39:33 +00:00
|
|
|
if (color)
|
2017-04-04 14:17:50 +00:00
|
|
|
fprintf(stdout, "\33[%sm\33[K", color);
|
2017-05-02 20:39:33 +00:00
|
|
|
fwrite(pc->ln.dat + match.rm_so,
|
|
|
|
match.rm_eo - match.rm_so, 1, stdout);
|
|
|
|
if (color)
|
2017-04-04 14:17:50 +00:00
|
|
|
fprintf(stdout, "\33[m\33[K");
|
2017-05-02 20:39:33 +00:00
|
|
|
a = match.rm_eo;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
if (oflag)
|
|
|
|
putchar('\n');
|
|
|
|
}
|
|
|
|
if (!oflag) {
|
2017-05-02 20:39:33 +00:00
|
|
|
if (pc->ln.len - a > 0)
|
|
|
|
fwrite(pc->ln.dat + a, pc->ln.len - a, 1,
|
|
|
|
stdout);
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
putchar('\n');
|
|
|
|
}
|
2017-05-02 20:39:33 +00:00
|
|
|
} else
|
|
|
|
grep_printline(&pc->ln, sep);
|
2017-05-20 11:20:03 +00:00
|
|
|
pc->printed++;
|
Add BSD grep to the base system and make it our default grep.
Deliverables: Small and clean code (1,4 KSLOC vs GNU's 8,5 KSLOC),
lower memory usage than GNU grep, GNU compatibility,
BSD license.
TODO: Performance is somewhat behind GNU grep but it is only
significant for bigger searches. The reason is complex, the
most important factor is that GNU grep uses lots of
optimizations to improve the speed of the regex library.
First, we need a modern regex library (practically by adopting
TRE), add support for GNU-style non-standard regexes and then
reevalute the performance issues and look for bottlenecks. In
the meantime, for those, who need better performance, it is
possible to build GNU grep by setting WITH_GNU_GREP.
Approved by: delphij (mentor)
Obtained from: OpenBSD (http://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/grep/),
freegrep (http://github.com/howardjp/freegrep)
Sponsored by: Google SoC 2008
Portbuild tests run by: kris, pav, erwin
Acknowledgements to: fjoe (as SoC 2008 mentor),
everyone who helped in reviewing and testing
2010-07-22 19:11:57 +00:00
|
|
|
}
|