Controlling Subprocesses

Octave includes some high-level commands like system and popen for starting subprocesses. If you want to run another program to perform some task and then look at its output, you will probably want to use these functions.

Octave also provides several very low-level Unix-like functions which can also be used for starting subprocesses, but you should probably only use them if you can't find any way to do what you need with the higher-level functions.

— Built-in Function: system (string, return_output, type)

Execute a shell command specified by string. The second argument is optional. If type is "async", the process is started in the background and the process id of the child process is returned immediately. Otherwise, the process is started, and Octave waits until it exits. If type argument is omitted, a value of "sync" is assumed.
If two input arguments are given (the actual value of return_output is irrelevant) and the subprocess is started synchronously, or if system is called with one input argument and one or more output arguments, the output from the command is returned. Otherwise, if the subprocess is executed synchronously, its output is sent to the standard output. To send the output of a command executed with system through the pager, use a command like
          disp (system (cmd, 1));
     
or
          printf ("%s
          ", system (cmd, 1));
     
The system function can return two values. The first is any output from the command that was written to the standard output stream, and the second is the output status of the command. For example,
          [output, status] = system ("echo foo; exit 2");
     
will set the variable output to the string `foo', and the variable status to the integer `2'.

— Built-in Function: fid = popen (command, mode)

Start a process and create a pipe. The name of the command to run is given by command. The file identifier corresponding to the input or output stream of the process is returned in fid. The argument mode may be

"r"
The pipe will be connected to the standard output of the process, and open for reading.
"w"
The pipe will be connected to the standard input of the process, and open for writing.

For example,
          fid = popen ("ls -ltr / | tail -3", "r");
          while (isstr (s = fgets (fid)))
            fputs (stdout, s);
          endwhile
               -| drwxr-xr-x  33 root  root  3072 Feb 15 13:28 etc
               -| drwxr-xr-x   3 root  root  1024 Feb 15 13:28 lib
               -| drwxrwxrwt  15 root  root  2048 Feb 17 14:53 tmp
     

— Built-in Function: pclose (fid)

Close a file identifier that was opened by popen. You may also use fclose for the same purpose.

— Function File: [in, out, pid] = popen2 (command, args)

Start a subprocess with two-way communication. The name of the process is given by command, and args is an array of strings containing options for the command. The file identifiers for the input and output streams of the subprocess are returned in in and out. If execution of the command is successful, pid contains the process ID of the subprocess. Otherwise, pid is −1.
For example,
          [in, out, pid] = popen2 ("sort", "-nr");
          fputs (in, "these\nare\nsome\nstrings\n");
          fclose (in);
          while (isstr (s = fgets (out)))
            fputs (stdout, s);
          endwhile
          fclose (out);
          -| are
          -| some
          -| strings
          -| these
     

— Built-in Variable: EXEC_PATH

The variable EXEC_PATH is a colon separated list of directories to search when executing external programs. Its initial value is taken from the environment variable OCTAVE_EXEC_PATH (if it exists) or PATH, but that value can be overridden by the command line argument --exec-path PATH, or by setting the value of EXEC_PATH in a startup script. If the value of EXEC_PATH begins (ends) with a colon, the directories
          octave-home/libexec/octave/site/exec/arch
          octave-home/libexec/octave/version/exec/arch
     
are prepended (appended) to EXEC_PATH, where octave-home is the top-level directory where all of Octave is installed (the default value is /usr). If you don't specify a value for EXEC_PATH explicitly, these special directories are prepended to your shell path.

In most cases, the following functions simply decode their arguments and make the corresponding Unix system calls. For a complete example of how they can be used, look at the definition of the function popen2.

— Built-in Function: [pid, msg] = fork ()

Create a copy of the current process.
Fork can return one of the following values:

> 0
You are in the parent process. The value returned from fork is the process id of the child process. You should probably arrange to wait for any child processes to exit.
0
You are in the child process. You can call exec to start another process. If that fails, you should probably call exit.
< 0
The call to fork failed for some reason. You must take evasive action. A system dependent error message will be waiting in msg.

— Built-in Function: [err, msg] = exec (file, args)

Replace current process with a new process. Calling exec without first calling fork will terminate your current Octave process and replace it with the program named by file. For example,
          exec ("ls" "-l")
     
will run ls and return you to your shell prompt.
If successful, exec does not return. If exec does return, err will be nonzero, and msg will contain a system-dependent error message.

— Built-in Function: [file_ids, err, msg] = pipe ()

Create a pipe and return the vector file_ids, which corresponding to the reading and writing ends of the pipe.
If successful, err is 0 and msg is an empty string. Otherwise, err is nonzero and msg contains a system-dependent error message.

— Built-in Function: [fid, msg] = dup2 (old, new)

Duplicate a file descriptor.
If successful, fid is greater than zero and contains the new file ID. Otherwise, fid is negative and msg contains a system-dependent error message.

— Built-in Function: [pid, msg] = waitpid (pid, options)

Wait for process pid to terminate. The pid argument can be:

−1
Wait for any child process.
0
Wait for any child process whose process group ID is equal to that of the Octave interpreter process.
> 0
Wait for termination of the child process with ID pid.

The options argument can be:

0
Wait until signal is received or a child process exits (this is the default if the options argument is missing).
1
Do not hang if status is not immediately available.
2
Report the status of any child processes that are stopped, and whose status has not yet been reported since they stopped.
3
Implies both 1 and 2.

If the returned value of pid is greater than 0, it is the process ID of the child process that exited. If an error occurs, pid will be less than zero and msg will contain a system-dependent error message.

— Built-in Function: [err, msg] = fcntl (fid, request, arg)

Change the properties of the open file fid. The following values may be passed as request:

F_DUPFD
Return a duplicate file descriptor.
F_GETFD
Return the file descriptor flags for fid.
F_SETFD
Set the file descriptor flags for fid.
F_GETFL
Return the file status flags for fid. The following codes may be returned (some of the flags may be undefined on some systems).

O_RDONLY
Open for reading only.
O_WRONLY
Open for writing only.
O_RDWR
Open for reading and writing.
O_APPEND
Append on each write.
O_CREAT
Create the file if it does not exist.
O_NONBLOCK
Nonblocking mode.
O_SYNC
Wait for writes to complete.
O_ASYNC
Asynchronous I/O.

F_SETFL
Set the file status flags for fid to the value specified by arg. The only flags that can be changed are O_APPEND and O_NONBLOCK.

If successful, err is 0 and msg is an empty string. Otherwise, err is nonzero and msg contains a system-dependent error message.

34.3 Controlling Subprocesses