Method: Kernel#spawn

Defined in:

process.c

#spawn([env,][,options]) ⇒ Object #spawn([env,][,options]) ⇒ Object

spawn executes specified command and return its pid.

pid = spawn("tar xf ruby-2.0.0-p195.tar.bz2")
Process.wait pid

pid = spawn(RbConfig.ruby, "-eputs'Hello, world!'")
Process.wait pid

This method is similar to Kernel#system but it doesn’t wait for the command to finish.

The parent process should use Process.wait to collect the termination status of its child or use Process.detach to register disinterest in their status; otherwise, the operating system may accumulate zombie processes.

spawn has bunch of options to specify process attributes:

env: hash
  name => val : set the environment variable
  name => nil : unset the environment variable

  the keys and the values except for +nil+ must be strings.
command...:
  commandline                 : command line string which is passed to the standard shell
  cmdname, arg1, ...          : command name and one or more arguments (This form does not use the shell. See below for caveats.)
  [cmdname, argv0], arg1, ... : command name, argv[0] and zero or more arguments (no shell)
options: hash
  clearing environment variables:
    :unsetenv_others => true   : clear environment variables except specified by env
    :unsetenv_others => false  : don't clear (default)
  process group:
    :pgroup => true or 0 : make a new process group
    :pgroup => pgid      : join the specified process group
    :pgroup => nil       : don't change the process group (default)
  create new process group: Windows only
    :new_pgroup => true  : the new process is the root process of a new process group
    :new_pgroup => false : don't create a new process group (default)
  resource limit: resourcename is core, cpu, data, etc.  See Process.setrlimit.
    :rlimit_resourcename => limit
    :rlimit_resourcename => [cur_limit, max_limit]
  umask:
    :umask => int
  redirection:
    key:
      FD              : single file descriptor in child process
      [FD, FD, ...]   : multiple file descriptor in child process
    value:
      FD                        : redirect to the file descriptor in parent process
      string                    : redirect to file with open(string, "r" or "w")
      [string]                  : redirect to file with open(string, File::RDONLY)
      [string, open_mode]       : redirect to file with open(string, open_mode, 0644)
      [string, open_mode, perm] : redirect to file with open(string, open_mode, perm)
      [:child, FD]              : redirect to the redirected file descriptor
      :close                    : close the file descriptor in child process
    FD is one of follows
      :in     : the file descriptor 0 which is the standard input
      :out    : the file descriptor 1 which is the standard output
      :err    : the file descriptor 2 which is the standard error
      integer : the file descriptor of specified the integer
      io      : the file descriptor specified as io.fileno
  file descriptor inheritance: close non-redirected non-standard fds (3, 4, 5, ...) or not
    :close_others => false  : inherit
  current directory:
    :chdir => str

The cmdname, arg1, ... form does not use the shell. However, on different OSes, different things are provided as built-in commands. An example of this is ‘echo’, which is a built-in on Windows, but is a normal program on Linux and Mac OS X. This means that Process.spawn 'echo', '%Path%' will display the contents of the %Path% environment variable on Windows, but Process.spawn 'echo', '$PATH' prints the literal $PATH.

If a hash is given as env, the environment is updated by env before exec(2) in the child process. If a pair in env has nil as the value, the variable is deleted.

# set FOO as BAR and unset BAZ.
pid = spawn({"FOO"=>"BAR", "BAZ"=>nil}, command)

If a hash is given as options, it specifies process group, create new process group, resource limit, current directory, umask and redirects for the child process. Also, it can be specified to clear environment variables.

The :unsetenv_others key in options specifies to clear environment variables, other than specified by env.

pid = spawn(command, :unsetenv_others=>true) # no environment variable
pid = spawn({"FOO"=>"BAR"}, command, :unsetenv_others=>true) # FOO only

The :pgroup key in options specifies a process group. The corresponding value should be true, zero, a positive integer, or nil. true and zero cause the process to be a process leader of a new process group. A non-zero positive integer causes the process to join the provided process group. The default value, nil, causes the process to remain in the same process group.

pid = spawn(command, :pgroup=>true) # process leader
pid = spawn(command, :pgroup=>10) # belongs to the process group 10

The :new_pgroup key in options specifies to pass CREATE_NEW_PROCESS_GROUP flag to CreateProcessW() that is Windows API. This option is only for Windows. true means the new process is the root process of the new process group. The new process has CTRL+C disabled. This flag is necessary for Process.kill(:SIGINT, pid) on the subprocess. :new_pgroup is false by default.

pid = spawn(command, :new_pgroup=>true)  # new process group
pid = spawn(command, :new_pgroup=>false) # same process group

The :rlimit_foo key specifies a resource limit. foo should be one of resource types such as core. The corresponding value should be an integer or an array which have one or two integers: same as cur_limit and max_limit arguments for Process.setrlimit.

cur, max = Process.getrlimit(:CORE)
pid = spawn(command, :rlimit_core=>[0,max]) # disable core temporary.
pid = spawn(command, :rlimit_core=>max) # enable core dump
pid = spawn(command, :rlimit_core=>0) # never dump core.

The :umask key in options specifies the umask.

pid = spawn(command, :umask=>077)

The :in, :out, :err, an integer, an IO and an array key specifies a redirection. The redirection maps a file descriptor in the child process.

For example, stderr can be merged into stdout as follows:

pid = spawn(command, :err=>:out)
pid = spawn(command, 2=>1)
pid = spawn(command, STDERR=>:out)
pid = spawn(command, STDERR=>STDOUT)

The hash keys specifies a file descriptor in the child process started by #spawn. :err, 2 and STDERR specifies the standard error stream (stderr).

The hash values specifies a file descriptor in the parent process which invokes #spawn. :out, 1 and STDOUT specifies the standard output stream (stdout).

In the above example, the standard output in the child process is not specified. So it is inherited from the parent process.

The standard input stream (stdin) can be specified by :in, 0 and STDIN.

A filename can be specified as a hash value.

pid = spawn(command, :in=>"/dev/null") # read mode
pid = spawn(command, :out=>"/dev/null") # write mode
pid = spawn(command, :err=>"log") # write mode
pid = spawn(command, [:out, :err]=>"/dev/null") # write mode
pid = spawn(command, 3=>"/dev/null") # read mode

For stdout and stderr (and combination of them), it is opened in write mode. Otherwise read mode is used.

For specifying flags and permission of file creation explicitly, an array is used instead.

pid = spawn(command, :in=>["file"]) # read mode is assumed
pid = spawn(command, :in=>["file", "r"])
pid = spawn(command, :out=>["log", "w"]) # 0644 assumed
pid = spawn(command, :out=>["log", "w", 0600])
pid = spawn(command, :out=>["log", File::WRONLY|File::EXCL|File::CREAT, 0600])

The array specifies a filename, flags and permission. The flags can be a string or an integer. If the flags is omitted or nil, File::RDONLY is assumed. The permission should be an integer. If the permission is omitted or nil, 0644 is assumed.

If an array of IOs and integers are specified as a hash key, all the elements are redirected.

# stdout and stderr is redirected to log file.
# The file "log" is opened just once.
pid = spawn(command, [:out, :err]=>["log", "w"])

Another way to merge multiple file descriptors is [:child, fd]. [:child, fd] means the file descriptor in the child process. This is different from fd. For example, :err=>:out means redirecting child stderr to parent stdout. But :err=>[:child, :out] means redirecting child stderr to child stdout. They differ if stdout is redirected in the child process as follows.

# stdout and stderr is redirected to log file.
# The file "log" is opened just once.
pid = spawn(command, :out=>["log", "w"], :err=>[:child, :out])

[:child, :out] can be used to merge stderr into stdout in IO.popen. In this case, IO.popen redirects stdout to a pipe in the child process and [:child, :out] refers the redirected stdout.

io = IO.popen(["sh", "-c", "echo out; echo err >&2", :err=>[:child, :out]])
p io.read #=> "out\nerr\n"

The :chdir key in options specifies the current directory.

pid = spawn(command, :chdir=>"/var/tmp")

spawn closes all non-standard unspecified descriptors by default. The “standard” descriptors are 0, 1 and 2. This behavior is specified by :close_others option. :close_others doesn’t affect the standard descriptors which are closed only if :close is specified explicitly.

pid = spawn(command, :close_others=>true)  # close 3,4,5,... (default)
pid = spawn(command, :close_others=>false) # don't close 3,4,5,...

:close_others is false by default for spawn and IO.popen.

Note that fds which close-on-exec flag is already set are closed regardless of :close_others option.

So IO.pipe and spawn can be used as IO.popen.

# similar to r = IO.popen(command)
r, w = IO.pipe
pid = spawn(command, :out=>w)   # r, w is closed in the child process.
w.close

:close is specified as a hash value to close a fd individually.

f = open(foo)
system(command, f=>:close)        # don't inherit f.

If a file descriptor need to be inherited, io=>io can be used.

# valgrind has --log-fd option for log destination.
# log_w=>log_w indicates log_w.fileno inherits to child process.
log_r, log_w = IO.pipe
pid = spawn("valgrind", "--log-fd=#{log_w.fileno}", "echo", "a", log_w=>log_w)
log_w.close
p log_r.read

It is also possible to exchange file descriptors.

pid = spawn(command, :out=>:err, :err=>:out)

The hash keys specify file descriptors in the child process. The hash values specifies file descriptors in the parent process. So the above specifies exchanging stdout and stderr. Internally, spawn uses an extra file descriptor to resolve such cyclic file descriptor mapping.

See Kernel.exec for the standard shell.

# File 'process.c', line 5059

static VALUE
rb_f_spawn(int argc, VALUE *argv, VALUE _)
{
    rb_pid_t pid;
    char errmsg[CHILD_ERRMSG_BUFLEN] = { '\0' };
    VALUE execarg_obj, fail_str;
    struct rb_execarg *eargp;

    execarg_obj = rb_execarg_new(argc, argv, TRUE, FALSE);
    eargp = rb_execarg_get(execarg_obj);
    fail_str = eargp->use_shell ? eargp->invoke.sh.shell_script : eargp->invoke.cmd.command_name;

    pid = rb_execarg_spawn(execarg_obj, errmsg, sizeof(errmsg));

    if (pid == -1) {
	int err = errno;
	rb_exec_fail(eargp, err, errmsg);
	RB_GC_GUARD(execarg_obj);
	rb_syserr_fail_str(err, fail_str);
    }
#if defined(HAVE_WORKING_FORK) || defined(HAVE_SPAWNV)
    return PIDT2NUM(pid);
#else
    return Qnil;
#endif
}