rsalz@uunet.uu.net (Rich Salz) (05/04/90)
Submitted-by: David B Rosen <rosen@bucasb.bu.edu>
Posting-number: Volume 22, Issue 11
Archive-name: tpipe
tpipe is a simple utility program that can be used to split a unix
pipeline into two pipelines. That is, the output of one pipeline can
be replicated and supplied as the input to two other pipelines
executing simultaneously.
For example
cmd1 <infile | tpipe "cmd2 | cmd3 >outfile" | cmd4
will look like this diagramatically:
--> cmd2 --> cmd3 --> outfile
/
infile --> cmd1 -<
\\
--> cmd4 --> (standard output)
#! /bin/sh
# This is a shell archive. Remove anything before this line, then unpack
# it by saving it into a file and typing "sh file". To overwrite existing
# files, type "sh file -c". You can also feed this as standard input via
# unshar, or by typing "sh <file", e.g.. If this archive is complete, you
# will see the following message at the end:
# "End of shell archive."
# Contents: Makefile README tpipe.1 tpipe.c
# Wrapped by rsalz@papaya.bbn.com on Thu May 3 16:40:10 1990
PATH=/bin:/usr/bin:/usr/ucb ; export PATH
if test -f 'Makefile' -a "${1}" != "-c" ; then
echo shar: Will not clobber existing file \"'Makefile'\"
else
echo shar: Extracting \"'Makefile'\" \(107 characters\)
sed "s/^X//" >'Makefile' <<'END_OF_FILE'
XCC = cc
XCFLAGS = -O
XTPIPE = tpipe
XCFILES = tpipe.c
X
X$(TPIPE): $(CFILES)
X $(CC) $(CFLAGS) $(CFILES) -o $@
END_OF_FILE
if test 107 -ne `wc -c <'Makefile'`; then
echo shar: \"'Makefile'\" unpacked with wrong size!
fi
# end of 'Makefile'
fi
if test -f 'README' -a "${1}" != "-c" ; then
echo shar: Will not clobber existing file \"'README'\"
else
echo shar: Extracting \"'README'\" \(2118 characters\)
sed "s/^X//" >'README' <<'END_OF_FILE'
X-------------------------------------------------------------
Xtpipe -- replicate standard output to an additional pipeline
X version 1.02 6 Mar. 1990
X-------------------------------------------------------------
X
Xtpipe is a simple utility program that can be used to split a unix
Xpipeline into two pipelines. That is, the output of one pipeline can
Xbe replicated and supplied as the input to two other pipelines
Xexecuting simultaneously.
X
XLike tee(1), tpipe transcribes its standard input to its standard
Xoutput. But where tee(1) writes an additional copy of its input to a
Xfile, tpipe writes the additional copy to the input of another
Xpipeline, which is specified as the argument to tpipe. In a typical
Xuse, this pipeline will eventually write to a file. The standard
Xoutput of tpipe is typically piped into another pipeline, whose output
X(if any) may go to the user's terminal or anywhere at all.
X
XI wrote tpipe because I was processing image files (using pbmplus),
Xand I wanted to apply more than one pipeline to the same input file,
Xbut the early parts of the pipeline were the same. I did not want to
Xhave to execute the early parts multiple times. I did not have enough
Xdisk space to write what would have been a huge intermediate file.
XFrankly, I did not know about teeing to named pipes, but anyway it's
Xnice not to have to bother with them or worry about name conflicts,
Xespecially if you have more than one job wanting to do this at the
Xsame time...
X
XSee the manual page for tpipe(1) for information about its use,
Xincluding an artificial example.
X
XThis distribution contains four files:
X Readme (this is it)
X Makefile (it is trivial)
X tpipe.1 (the manual page)
X tpipe.c (the code!)
X
XTo read the manual page, just `nroff -man tpipe.1'.
X
XTo compile tpipe, just `make'. Put tpipe.1 in the man1/ directory in
Xyour MANPATH or the system man path.
X
X--
XDavid B Rosen, Cognitive & Neural Systems rosen@bucasb.bu.edu
XCenter for Adaptive Systems rosen%bucasb@{buacca,bu-it}.bu.edu
XBoston University {mit-eddie,harvard,uunet}!bu.edu!thalamus!rosen
END_OF_FILE
if test 2118 -ne `wc -c <'README'`; then
echo shar: \"'README'\" unpacked with wrong size!
fi
# end of 'README'
fi
if test -f 'tpipe.1' -a "${1}" != "-c" ; then
echo shar: Will not clobber existing file \"'tpipe.1'\"
else
echo shar: Extracting \"'tpipe.1'\" \(2465 characters\)
sed "s/^X//" >'tpipe.1' <<'END_OF_FILE'
X.TH TPIPE 1 "29 January 1990"
X.\" @(#)tpipe.1 1.0 90/01/29; David B Rosen (rosen@bucasb.bu.edu)
X.SH NAME
Xtpipe \- replicate the standard output into an additional pipeline
Xthat is run in a subshell
X.SH SYNOPSIS
X.B tpipe
X[
X.I pipeline
X]
X.IX "tpipe command" "" "\fLtpipe\fP \(em copy standard output to many pipelines"
X.IX "copy" "standard output to many pipelines \(em \fLtpipe\fP"
X.IX "standard output" "copy to many pipelines" "" "copy to many pipelines \(em \fLtpipe\fP"
X.IX pipelines "copy standard output to many" "" "copy standard output to many \(em \fLtpipe\fP"
X.SH DESCRIPTION
X.B tpipe
Xtranscribes the standard input to the
Xstandard output and simultaneously writes an additional copy to
Xa specified
X.IR pipeline .
X.PP
Xtpipe is similar to tee(1) except that tpipe writes the duplicate copy
Xof its standard input to a command or pipeline instead of a file. This
Xcan help you avoid re-executing earlier commands in the pipeline,
Xwriting temporary files, or resorting to the use of named pipes.
X.PP
XThe specified
X.IR pipeline
Xis always executed in a subshell by sh(1), regardless of your current
Xshell. If a non-empty string is supplied as the argument, it must be
Xa valid pipeline or command for sh(1). Normally, you will want to enclose
Xthe
X.IR pipeline
Xargument in quotes ('' or ""). The type of quotes you choose
Xwill affect variable substitution by your shell (see the man page
Xfor your shell, such as csh(1), for details).
X.PP
XIf the subshell pipeline writes to its standard output, this output
Xwill go to the standard output of tpipe, where it will be interspersed
Xin an unpredictable way with the other copy of standard input.
XNormally, this is not what you want. Instead, you would typically
Xspecify a subshell
X.IR pipeline
Xwhose output is redirected to a file (as in the example
Xbelow) or has some other effect.
X.SH EXAMPLE
X.IP
X% cmd1 <infile | tpipe "cmd2 | cmd3 >outfile" | cmd4
X.PP
Xwhich has the effect of running the output of command cmd1
Xsimultaneously through two pipelines, "cmd2 | cmd3 >outfile" and cmd4.
XDiagramatically, in this example (this will look wrong with
Xa variable-spaced font):
X
X.nf
X --> cmd2 --> cmd3 --> outfile
X /
Xinfile --> cmd1 -<
X \\
X --> cmd4 --> (standard output)
X.fi
X.SH SEE ALSO
X.BR tee (1),
X.BR sh (1),
X.BR cat (1)
X
X.SH AUTHOR
XDavid B Rosen
X.PP
XTHIS SOFTWARE IS PROVIDED "AS IS" AND WITHOUT EXPRESS OR IMPLIED WARRANTY.
END_OF_FILE
if test 2465 -ne `wc -c <'tpipe.1'`; then
echo shar: \"'tpipe.1'\" unpacked with wrong size!
fi
# end of 'tpipe.1'
fi
if test -f 'tpipe.c' -a "${1}" != "-c" ; then
echo shar: Will not clobber existing file \"'tpipe.c'\"
else
echo shar: Extracting \"'tpipe.c'\" \(1603 characters\)
sed "s/^X//" >'tpipe.c' <<'END_OF_FILE'
X/* tpipe.c -- tee a pipeline into two pipelines. Like tee(1) but
X argument is a command or pipeline rather than a file.
X
X See the man page tpipe(1) supplied with this software.
X
X This version uses the unix system calls popen(3), read(2), and
X write(2). It uses write(2) to write directly to the fileno() of
X of the file pointer stream returned by popen.
X
X I've tried it out under BSD, System V, and an older version of unix,
X but:
X
X THIS SOFTWARE IS PROVIDED "AS IS" AND WITHOUT EXPRESS OR IMPLIED
X WARRANTY.
X
X Version 1.02 (4 Mar 1989) (Use fileno())
X
X--
XDavid B Rosen, Cognitive & Neural Systems rosen@bucasb.bu.edu
XCenter for Adaptive Systems rosen%bucasb@{buacca,bu-it}.bu.edu
XBoston University {mit-eddie,harvard,uunet}!bu.edu!thalamus!rosen
X
X*/
X
X#include <stdio.h>
X
X/*#define NOHACK*/
X
X#ifndef BUFSIZ
X#define BUFSIZ 2048
X#endif /*BUFSIZ*/
X
Xint main(argc, argv)
X int argc;
X char *argv[];
X{
X char buf[BUFSIZ];
X register FILE *subpipeline = NULL;
X register unsigned n;
X
X if (argc == 2){
X if (*argv[1]) {
X if ((subpipeline = popen(argv[1],"w")) == NULL) {
X fprintf(stderr, "%s: can't create subpipeline %s\n", argv[0], argv[1]);
X exit(1);
X }
X }
X } else if (argc > 2) {
X fprintf(stderr, "usage: %s [pipeline]\n", argv[0]);
X exit(2);
X }
X
X while ((n = read(0, buf, BUFSIZ)) > 0) {
X write(1, buf, n); /* write to standard output */
X if (subpipeline) { /* write to subpipeline: */
X write((int)fileno(subpipeline), buf, n);
X }
X }
X
X if (subpipeline) pclose(subpipeline);
X return 0;
X}
END_OF_FILE
if test 1603 -ne `wc -c <'tpipe.c'`; then
echo shar: \"'tpipe.c'\" unpacked with wrong size!
fi
# end of 'tpipe.c'
fi
echo shar: End of shell archive.
exit 0
exit 0 # Just in case...
--
Please send comp.sources.unix-related mail to rsalz@uunet.uu.net.
Use a domain-based address or give alternate paths, or you may lose out.