IPC::Open3
NAME
IPC::Open3 - open a process for reading, writing, and error handling using open3()
SYNOPSIS
DESCRIPTION
Extremely similar to open2(), open3() spawns the given $cmd and connects CHLD_OUT for reading from the child, CHLD_IN for writing to the child, and CHLD_ERR for errors. If CHLD_ERR is false, or the same file descriptor as CHLD_OUT, then STDOUT and STDERR of the child are on the same filehandle (this means that an autovivified lexical cannot be used for the STDERR filehandle, see SYNOPSIS). The CHLD_IN will have autoflush turned on.
If CHLD_IN begins with <&
, then CHLD_IN will be closed in the
parent, and the child will read from it directly. If CHLD_OUT or
CHLD_ERR begins with >&
, then the child will send output
directly to that filehandle. In both cases, there will be a dup(2)
instead of a pipe(2) made.
If either reader or writer is the null string, this will be replaced by an autogenerated filehandle. If so, you must pass a valid lvalue in the parameter slot so it can be overwritten in the caller, or an exception will be raised.
The filehandles may also be integers, in which case they are understood as file descriptors.
open3() returns the process ID of the child process. It doesn't return on
failure: it just raises an exception matching /^open3:/
. However,
exec
failures in the child (such as no such file or permission denied),
are just reported to CHLD_ERR under Windows and OS/2, as it is not possible
to trap them.
If the child process dies for any reason, the next write to CHLD_IN is likely to generate a SIGPIPE in the parent, which is fatal by default. So you may wish to handle this signal.
Note if you specify -
as the command, in an analogous fashion to
open(FOO, "-|")
the child process will just be the forked Perl
process rather than an external command. This feature isn't yet
supported on Win32 platforms.
open3() does not wait for and reap the child process after it exits.
Except for short programs where it's acceptable to let the operating system
take care of this, you need to do this yourself. This is normally as
simple as calling waitpid $pid, 0
when you're done with the process.
Failing to do this can result in an accumulation of defunct or "zombie"
processes. See waitpid for more information.
If you try to read from the child's stdout writer and their stderr writer, you'll have problems with blocking, which means you'll want to use select() or the IO::Select, which means you'd best use sysread() instead of readline() for normal stuff.
This is very dangerous, as you may block forever. It assumes it's going to talk to something like bc, both writing to it and reading from it. This is presumably safe because you "know" that commands like bc will read a line at a time and output a line at a time. Programs like sort that read their entire input stream first, however, are quite apt to cause deadlock.
The big problem with this approach is that if you don't have control
over source code being run in the child process, you can't control
what it does with pipe buffering. Thus you can't just open a pipe to
cat -v
and continually read and write a line from it.
See Also
- IPC::Open2
Like Open3 but without STDERR capture.
- IPC::Run
This is a CPAN module that has better error handling and more facilities than Open3.
WARNING
The order of arguments differs from that of open2().