Add the csplit(1) utility, which splits files based on context, as specified

by SUSv3.

PR:		36191
Reviewed by:	mike
This commit is contained in:
Tim J. Robbins 2002-05-02 07:46:36 +00:00
parent c3bdc05fb9
commit ea9e961cb2
Notes: svn2git 2020-12-20 02:59:44 +00:00
svn path=/head/; revision=95926
3 changed files with 589 additions and 0 deletions

View File

@ -32,6 +32,7 @@ SUBDIR= apply \
comm \
compile_et \
compress \
csplit \
ctags \
cut \
dirname \

134
usr.bin/csplit/csplit.1 Normal file
View File

@ -0,0 +1,134 @@
.\" Copyright (c) 2002 Tim J. Robbins.
.\" 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.
.\"
.\" $FreeBSD$
.\"
.Dd April 28, 2002
.Dt CSPLIT 1
.Os
.Sh NAME
.Nm csplit
.Nd split files based on context
.Sh SYNOPSIS
.Nm
.Op Fl ks
.Op Fl f Ar prefix
.Op Fl n Ar number
.Ar file
.Bk
.Op Ar args ...
.Ek
.Sh DESCRIPTION
The
.Nm
utility splits
.Ar file
into pieces using the patterns
.Ar args... .
If
.Ar file
is
a dash
.Pq Ql \&- ,
.Nm
reads from standard input.
.Pp
The options are as follows:
.Bl -tag -width indent
.It Fl f Ar prefix
Give created files names beginning with
.Ar prefix .
The default is
.Sq Pa xx .
.It Fl k
Do not remove output files if an error occurs or a
.Dv HUP ,
.Dv INT
or
.Dv TERM
signal is received.
.It Fl n Ar number
Use
.Ar number
of decimal digits after the
.Ar prefix
to form the file name. The default is 2.
.It Fl s
Do not write the size of each output file to standard output as it is
created.
.El
.Pp
The
.Ar args...
operands may be a combination of the following patterns:
.Bl -tag -width "line_no"
.It Cm / Ns Ar regexp Ns Cm / Ns Oo Oo +|- Oc Ns Ar offset Oc
Create a file containing the input from the current line to (but not including)
the next line matching the given basic regular expression.
An optional
.Ar offset
from the line that matched may be specified.
.It Cm % Ns Ar regexp Ns Cm % Ns Oo Oo +|- Oc Ns Ar offset Oc
Same as above but a file is not created for the output.
.It Ar line_no
Create containing the input from the current line to (but not including)
the specified line number.
.It Cm { Ns Ar num Ns Cm }
Repeat the previous pattern the specified number of times.
If it follows a line number pattern, a new file will be created for each
.Em line_no
lines,
.Em num
times.
The first line of the file is line number 1 for historic reasons.
.El
.Pp
After all the patterns have been processed, the remaining input data
(if there is any) will be written to a new file.
.Pp
Requesting to split at a line before the current line number or past the
end of the file will result in an error.
.Sh EXAMPLES
Split the
.Xr mdoc 7
file
.Pa foo.1
into one file for each section (up to 20):
.Pp
.D1 Ic csplit -k foo.1 '%^\e.Sh%' '/^\e.Sh/' '{20}'
.Pp
Split standard input after the first 99 lines and every 100 lines thereafter:
.Pp
.D1 Ic csplit -k - 100 '{19}'
.Sh DIAGNOSTICS
.Ex -std
.Sh SEE ALSO
.Xr sed 1 ,
.Xr split 1 ,
.Xr re_format 7
.Sh STANDARDS
The
.Nm
utility conforms to
.St -p1003.1-2001 .

454
usr.bin/csplit/csplit.c Normal file
View File

@ -0,0 +1,454 @@
/*-
* Copyright (c) 2002 Tim J. Robbins.
* 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.
*/
/*
* csplit -- split files based on context
*
* This utility splits its input into numbered output files by line number
* or by a regular expression. Regular expression matches have an optional
* offset with them, allowing the split to occur a specified number of
* lines before or after the match.
*
* To handle negative offsets, we stop reading when the match occurs and
* store the offset that the file should have been split at, then use
* this output file as input until all the "overflowed" lines have been read.
* The file is then closed and truncated to the correct length.
*
* We assume that the output files can be seeked upon (ie. they cannot be
* symlinks to named pipes or character devices), but make no such
* assumption about the input.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/types.h>
#include <ctype.h>
#include <err.h>
#include <errno.h>
#include <limits.h>
#include <regex.h>
#include <signal.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
void cleanup(void);
void do_lineno(const char *);
void do_rexp(const char *);
char *getline(void);
void handlesig(int);
FILE *newfile(void);
void toomuch(FILE *, long);
void usage(void);
/*
* Command line options
*/
const char *prefix; /* File name prefix */
long sufflen; /* Number of decimal digits for suffix */
int sflag; /* Suppress output of file names */
int kflag; /* Keep output if error occurs */
/*
* Other miscellaneous globals (XXX too many)
*/
long lineno; /* Current line number in input file */
long reps; /* Number of repetitions for this pattern */
long nfiles; /* Number of files output so far */
long maxfiles; /* Maximum number of files we can create */
char currfile[PATH_MAX]; /* Current output file */
const char *infn; /* Name of the input file */
FILE *infile; /* Input file handle */
FILE *overfile; /* Overflow file for toomuch() */
off_t truncofs; /* Offset this file should be truncated at */
int doclean; /* Should cleanup() remove output? */
int
main(int argc, char *argv[])
{
long i;
int ch;
const char *expr;
char *ep, *p;
FILE *ofp;
kflag = sflag = 0;
prefix = "xx";
sufflen = 2;
while ((ch = getopt(argc, argv, "ksf:n:")) > 0) {
switch (ch) {
case 'f':
prefix = optarg;
break;
case 'k':
kflag = 1;
break;
case 'n':
errno = 0;
sufflen = strtol(optarg, &ep, 10);
if (sufflen <= 0 || *ep != '\0' || errno != 0)
errx(1, "%s: bad suffix length", optarg);
break;
case 's':
sflag = 1;
break;
default:
usage();
/*NOTREACHED*/
}
}
if (sufflen + strlen(prefix) >= PATH_MAX)
errx(1, "name too long");
argc -= optind;
argv += optind;
if ((infn = *argv++) == NULL)
usage();
if (strcmp(infn, "-") == 0) {
infile = stdin;
infn = "stdin";
} else if ((infile = fopen(infn, "r")) == NULL)
err(1, "%s", infn);
if (!kflag) {
doclean = 1;
atexit(cleanup);
signal(SIGHUP, handlesig);
signal(SIGINT, handlesig);
signal(SIGTERM, handlesig);
}
lineno = 0;
nfiles = 0;
truncofs = 0;
overfile = NULL;
/* Ensure 10^sufflen < LONG_MAX. */
for (maxfiles = 1, i = 0; i < sufflen; i++) {
if (maxfiles > LONG_MAX / 10)
errx(1, "%ld: suffix too long (limit %ld)",
sufflen, i);
maxfiles *= 10;
}
/* Create files based on supplied patterns. */
while (nfiles < maxfiles - 1 && (expr = *argv++) != NULL) {
/* Look ahead & see if this pattern has any repetitions. */
if (*argv != NULL && **argv == '{') {
errno = 0;
reps = strtol(*argv + 1, &ep, 10);
if (reps < 0 || *ep != '}' || errno != 0)
errx(1, "%s: bad repetition count", *argv + 1);
argv++;
} else
reps = 0;
if (*expr == '/' || *expr == '%') {
do
do_rexp(expr);
while (reps-- != 0 && nfiles < maxfiles - 1);
} else if (isdigit((unsigned char)*expr))
do_lineno(expr);
else
errx(1, "%s: unrecognised pattern", expr);
}
/* Copy the rest into a new file. */
if (!feof(infile)) {
ofp = newfile();
while ((p = getline()) != NULL && fputs(p, ofp) == 0)
;
if (!sflag)
printf("%jd\n", (intmax_t)ftello(ofp));
if (fclose(ofp) != 0)
err(1, "%s", currfile);
}
toomuch(NULL, 0);
doclean = 0;
return (0);
}
void
usage(void)
{
fprintf(stderr,
"usage: csplit [-ks] [-f prefix] [-n number] file [args ...]\n");
exit(1);
}
void
handlesig(int sig __unused)
{
const char msg[] = "csplit: caught signal, cleaning up\n";
write(STDERR_FILENO, msg, sizeof(msg) - 1);
cleanup();
_exit(2);
}
/* Create a new output file. */
FILE *
newfile(void)
{
FILE *fp;
snprintf(currfile, sizeof(currfile), "%s%0*ld", prefix, (int)sufflen,
nfiles);
if ((fp = fopen(currfile, "w+")) == NULL)
err(1, "%s", currfile);
nfiles++;
return (fp);
}
/* Remove partial output, called before exiting. */
void
cleanup(void)
{
char fnbuf[PATH_MAX];
long i;
if (!doclean)
return;
/*
* NOTE: One cannot portably assume to be able to call snprintf()
* from inside a signal handler. It does, however, appear to be safe
* to do on FreeBSD. The solution to this problem is worse than the
* problem itself.
*/
for (i = 0; i < nfiles; i++) {
snprintf(fnbuf, sizeof(fnbuf), "%s%0*ld", prefix,
(int)sufflen, i);
unlink(fnbuf);
}
}
/* Read a line from the input into a static buffer. */
char *
getline(void)
{
static char lbuf[LINE_MAX];
FILE *src;
src = overfile != NULL ? overfile : infile;
again: if (fgets(lbuf, sizeof(lbuf), src) == NULL) {
if (src == overfile) {
src = infile;
goto again;
}
return (NULL);
}
if (ferror(src))
err(1, "%s", infn);
lineno++;
return (lbuf);
}
/* Conceptually rewind the input (as obtained by getline()) back `n' lines. */
void
toomuch(FILE *ofp, long n)
{
char buf[BUFSIZ];
size_t i, nread;
if (overfile != NULL) {
/*
* Truncate the previous file we overflowed into back to
* the correct length, close it.
*/
if (fflush(overfile) != 0)
err(1, "overflow");
if (ftruncate(fileno(overfile), truncofs) != 0)
err(1, "overflow");
if (fclose(overfile) != 0)
err(1, "overflow");
overfile = NULL;
}
if (n == 0)
/* Just tidying up */
return;
lineno -= n;
/*
* Wind the overflow file backwards to `n' lines before the
* current one.
*/
do {
if (ftello(ofp) < (off_t)sizeof(buf))
rewind(ofp);
else
fseek(ofp, -(long)sizeof(buf), SEEK_CUR);
if (ferror(ofp))
errx(1, "%s: can't seek", currfile);
if ((nread = fread(buf, 1, sizeof(buf), ofp)) == 0)
errx(1, "can't read overflowed output");
if (fseek(ofp, -(long)nread, SEEK_CUR) != 0)
err(1, "%s", currfile);
for (i = 1; i <= nread; i++)
if (buf[nread - i] == '\n' && n-- == 0)
break;
} while (n > 0);
if (fseek(ofp, nread - i + 1, SEEK_CUR) != 0)
err(1, "%s", currfile);
/*
* getline() will read from here. Next call will truncate to
* truncofs in this file.
*/
overfile = ofp;
truncofs = ftello(overfile);
}
/* Handle splits for /regexp/ and %regexp% patterns. */
void
do_rexp(const char *expr)
{
regex_t cre;
intmax_t nwritten;
long ofs;
int first;
char *ecopy, *ep, *p, *pofs, *re;
FILE *ofp;
if ((ecopy = strdup(expr)) == NULL)
err(1, "strdup");
re = ecopy + 1;
if ((pofs = strrchr(ecopy, *expr)) == NULL || pofs[-1] == '\\')
errx(1, "%s: missing trailing %c", expr, *expr);
*pofs++ = '\0';
if (*pofs != '\0') {
errno = 0;
ofs = strtol(pofs, &ep, 10);
if (*ep != '\0' || errno != 0)
errx(1, "%s: bad offset", pofs);
} else
ofs = 0;
if (regcomp(&cre, re, REG_BASIC|REG_NOSUB) != 0)
errx(1, "%s: bad regular expression", re);
if (*expr == '/')
/* /regexp/: Save results to a file. */
ofp = newfile();
else {
/* %regexp%: Make a temporary file for overflow. */
if ((ofp = tmpfile()) == NULL)
err(1, "tmpfile");
}
/* Read and output lines until we get a match. */
first = 1;
while ((p = getline()) != NULL) {
if (fputs(p, ofp) != 0)
break;
if (!first && regexec(&cre, p, 0, NULL, 0) == 0)
break;
first = 0;
}
if (p == NULL)
errx(1, "%s: no match", re);
if (ofs <= 0) {
/*
* Negative (or zero) offset: throw back any lines we should
* not have read yet.
*/
if (p != NULL) {
toomuch(ofp, -ofs + 1);
nwritten = (intmax_t)truncofs;
} else
nwritten = (intmax_t)ftello(ofp);
} else {
/*
* Positive offset: copy the requested number of lines
* after the match.
*/
while (--ofs > 0 && (p = getline()) != NULL)
fputs(p, ofp);
toomuch(NULL, 0);
nwritten = (intmax_t)ftello(ofp);
if (fclose(ofp) != 0)
err(1, "%s", currfile);
}
if (!sflag && *expr == '/')
printf("%jd\n", nwritten);
regfree(&cre);
free(ecopy);
}
/* Handle splits based on line number. */
void
do_lineno(const char *expr)
{
long lastline, tgtline;
char *ep, *p;
FILE *ofp;
errno = 0;
tgtline = strtol(expr, &ep, 10);
if (tgtline <= 0 || errno != 0 || *ep != '\0')
errx(1, "%s: bad line number", expr);
lastline = tgtline;
if (lastline <= lineno)
errx(1, "%s: can't go backwards", expr);
while (nfiles < maxfiles - 1) {
ofp = newfile();
while (lineno + 1 != lastline) {
if ((p = getline()) == NULL)
errx(1, "%ld: out of range", lastline);
if (fputs(p, ofp) != 0)
break;
}
if (!sflag)
printf("%jd\n", (intmax_t)ftello(ofp));
if (fclose(ofp) != 0)
err(1, "%s", currfile);
if (reps-- == 0)
break;
lastline += tgtline;
}
}