Linux and UNIX Man Pages

Linux & Unix Commands - Search Man Pages

multitee(1) [debian man page]

multitee(1)						      General Commands Manual						       multitee(1)

NAME
multitee - send multiple inputs to multiple outputs SYNTAX
multitee [ -bsize ] [ -vQq ] [ fd-fd,fd,fd... ] ... DESCRIPTION
multitee sends multiple inputs to multiple outputs. Given an argument of the form fdin-fdout,fdout,fdout... it will send all input on file descriptor fdin to each descriptor fdout. It will exit when all fdin are closed. Several arguments may specify outputs from the same fdin. -fdout and ,fdout are equivalent. If there is an error of any sort (including SIGPIPE) in writing to fdout, multitee prints a warning on stderr and forgets fdout entirely. (This doesn't affect reads on fdin.) If -fdout is replaced by :fdout then multitee will exit upon any SIGPIPEs from that descriptor. Furthermore, efd means that as soon as fdin reaches end of file, fd is considered to reach EOF as well. multitee will warn about any input errors and then treat them like EOF. Unlike tee, multitee tries its best to continue processing all descriptors even while some of them are blocked. However, it will get stuck reading if someone else is reading the descriptor and grabs the input first; it will get stuck writing if an input packet does not fit in an output pipe. (If the output descriptor has NDELAY set, and multitee receives EWOULDBLOCK, it writes one byte at a time to avoid pipe synchronization problems.) While it is tempting to set the descriptors to non-blocking mode, this is dangerous: other processes using the same open file may not be able to deal with NDELAY. It is incredible that none of the major UNIX vendors or standards committees has come up with true per-process non-blocking I/O. (Under BSD 4.3 and its variants, multitee could send timer signals to itself rapidly to inter- rupt any blocking I/O. However, this cannot work under BSD 4.2, and is generally more trouble than it's worth.) A program can set NDELAY before invoking multitee if it knows that no other processes will use the same open file. multitee will also temporarily stop reading an input descriptor if more than 8192 bytes are pending on one of its output descriptors. This does not affect independent fdin-fdout pairs. multitee has several flags: -bsize Change input buffer size from 8192 to size. Unlike the previous version of multitee, this version does not require output buf- fers, and does not copy bytes anywhere between read() and write(). -v Verbose. -q Quiet. multitee will not use stderr in any way (except, of course, if descriptor 2 is specified in an argument). -Q Normal level of verbosity. EXIT VALUE
0 normally. 1 for usage messages. 3 if multitee runs out of memory. 4 in various impossible situations. DIAGNOSTICS
fatal: out of memory multitee has run out of memory. warning: cannot read descriptor Self-explanatory. warning: cannot write descriptor Self-explanatory. EXAMPLES
multitee 0-1,4,5 4>foo 5>bar Same as tee foo bar except for better blocking behavior. multitee 0:1 3:1 4:1,2 6:7 Merge several sources into the output, meanwhile copying 6 to 7 and recording 4's input in 2. tcpclient servermachine smtp multitee 0:7 6:1e0 Same as mconnect on Suns. The e0 tells multitee to quit as soon as the network connection closes. RESTRICTIONS
multitee expects all descriptors involved to be open. Currently a closed descriptor acts like an open descriptor which can never be writ- ten to. BUGS
None known. VERSION
multitee version 3.0, 7/22/91. AUTHOR
Placed into the public domain by Daniel J. Bernstein. SEE ALSO
tee(1) multitee(1)

Check Out this Related Man Page

PIPE(2) 						      BSD System Calls Manual							   PIPE(2)

NAME
pipe -- create descriptor pair for interprocess communication LIBRARY
Standard C Library (libc, -lc) SYNOPSIS
#include <unistd.h> int pipe(int fildes[2]); int pipe2(int fildes[2], int flags); DESCRIPTION
The pipe() function creates a pipe, which is an object allowing unidirectional data flow, and allocates a pair of file descriptors. The first descriptor connects to the read end of the pipe, and the second connects to the write end, so that data written to fildes[1] appears on (i.e., can be read from) fildes[0]. This allows the output of one program to be sent to another program: the source's standard output is set up to be the write end of the pipe, and the sink's standard input is set up to be the read end of the pipe. The pipe itself persists until all its associated descriptors are closed. A pipe whose read or write end has been closed is considered widowed. Writing on such a pipe causes the writing process to receive a SIGPIPE signal. Widowing a pipe is the only way to deliver end-of-file to a reader: after the reader consumes any buffered data, reading a widowed pipe returns a zero count. The pipe2() function behaves exactly like pipe() only it allows extra flags to be set on the returned file descriptor. The following flags are valid: O_CLOEXEC Set the ``close-on-exec'' property. O_NONBLOCK Sets non-blocking I/O. O_NOSIGPIPE Return EPIPE instead of raising SIGPIPE. RETURN VALUES
On successful creation of the pipe, zero is returned. Otherwise, a value of -1 is returned and the variable errno set to indicate the error. ERRORS
The pipe() and pipe2() calls will fail if: [EFAULT] The fildes buffer is in an invalid area of the process's address space. The reliable detection of this error cannot be guaranteed; when not detected, a signal may be delivered to the process, indicating an address violation. [EMFILE] Too many descriptors are active. [ENFILE] The system file table is full. pipe2() will also fail if: [EINVAL] flags is other than O_NONBLOCK or O_CLOEXEC. SEE ALSO
sh(1), fork(2), read(2), socketpair(2), write(2) STANDARDS
The pipe() function conforms to ISO/IEC 9945-1:1990 (``POSIX.1''). HISTORY
A pipe() function call appeared in Version 6 AT&T UNIX. The pipe2() function is inspired from Linux and appeared in NetBSD 6.0. BSD
January 23, 2012 BSD
Man Page