class PTY
Creates and manages pseudo terminals (PTYs). See also en.wikipedia.org/wiki/Pseudo_terminal
PTY
allows you to allocate new terminals using ::open
or ::spawn
a new terminal with a specific command.
Example¶ ↑
In this example we will change the buffering type in the factor
command, assuming that factor uses stdio for stdout buffering.
If IO.pipe
is used instead of PTY.open
, this code deadlocks because factor’s stdout is fully buffered.
# start by requiring the standard library PTY require 'pty' master, slave = PTY.open read, write = IO.pipe pid = spawn("factor", :in=>read, :out=>slave) read.close # we dont need the read slave.close # or the slave # pipe "42" to the factor command write.puts "42" # output the response from factor p master.gets #=> "42: 2 3 7\n" # pipe "144" to factor and print out the response write.puts "144" p master.gets #=> "144: 2 2 2 2 3 3\n" write.close # close the pipe # The result of read operation when pty slave is closed is platform # dependent. ret = begin master.gets # FreeBSD returns nil. rescue Errno::EIO # GNU/Linux raises EIO. nil end p ret #=> nil
License¶ ↑
© Copyright 1998 by Akinori Ito.
This software may be redistributed freely for this purpose, in full or in part, provided that this entire copyright notice is included on any copies of this software and applications and derivations thereof.
This software is provided on an “as is” basis, without warranty of any kind, either expressed or implied, as to any matter including, but not limited to warranty of fitness of purpose, or merchantability, or results obtained from use of this software.
Public Class Methods
Source
static VALUE pty_check(int argc, VALUE *argv, VALUE self) { VALUE pid, exc; rb_pid_t cpid; int status; const int flag = #ifdef WNOHANG WNOHANG| #endif #ifdef WUNTRACED WUNTRACED| #endif 0; rb_scan_args(argc, argv, "11", &pid, &exc); cpid = rb_waitpid(NUM2PIDT(pid), &status, flag); if (cpid == -1 || cpid == 0) return Qnil; if (!RTEST(exc)) return rb_last_status_get(); raise_from_check(cpid, status); UNREACHABLE_RETURN(Qnil); }
Checks the status of the child process specified by pid
. Returns nil
if the process is still alive.
If the process is not alive, and raise
was true, a PTY::ChildExited
exception will be raised. Otherwise it will return a Process::Status
instance.
pid
-
The process id of the process to check
raise
-
If
true
and the process identified bypid
is no longer alive aPTY::ChildExited
is raised.
Source
static VALUE pty_open(VALUE klass) { int master_fd, slave_fd; char slavename[DEVICELEN]; getDevice(&master_fd, &slave_fd, slavename, 1); VALUE master_path = rb_obj_freeze(rb_sprintf("masterpty:%s", slavename)); VALUE master_io = rb_io_open_descriptor(rb_cIO, master_fd, FMODE_READWRITE | FMODE_SYNC | FMODE_DUPLEX, master_path, RUBY_IO_TIMEOUT_DEFAULT, NULL); VALUE slave_path = rb_obj_freeze(rb_str_new_cstr(slavename)); VALUE slave_file = rb_io_open_descriptor(rb_cFile, slave_fd, FMODE_READWRITE | FMODE_SYNC | FMODE_DUPLEX | FMODE_TTY, slave_path, RUBY_IO_TIMEOUT_DEFAULT, NULL); VALUE assoc = rb_assoc_new(master_io, slave_file); if (rb_block_given_p()) { return rb_ensure(rb_yield, assoc, pty_close_pty, assoc); } return assoc; }
Allocates a pty (pseudo-terminal).
In the block form, yields an array of two elements (master_io, slave_file
) and the value of the block is returned from open
.
The IO
and File
are both closed after the block completes if they haven’t been already closed.
PTY.open {|master, slave| p master #=> #<IO:masterpty:/dev/pts/1> p slave #=> #<File:/dev/pts/1> p slave.path #=> "/dev/pts/1" }
In the non-block form, returns a two element array, [master_io, slave_file]
.
master, slave = PTY.open # do something with master for IO, or the slave file
The arguments in both forms are:
master_io
-
the master of the pty, as an
IO
. slave_file
-
the slave of the pty, as a
File
. The path to the terminal device is available viaslave_file.path
IO#raw!
is usable to disable newline conversions:
require 'io/console' PTY.open {|m, s| s.raw! # ... }
Source
static VALUE pty_getpty(int argc, VALUE *argv, VALUE self) { VALUE res; struct pty_info info; char SlaveName[DEVICELEN]; establishShell(argc, argv, &info, SlaveName); VALUE pty_path = rb_obj_freeze(rb_str_new_cstr(SlaveName)); VALUE rport = rb_io_open_descriptor( rb_cFile, info.fd, FMODE_READABLE, pty_path, RUBY_IO_TIMEOUT_DEFAULT, NULL ); int wpty_fd = rb_cloexec_dup(info.fd); if (wpty_fd == -1) { rb_sys_fail("dup()"); } VALUE wport = rb_io_open_descriptor( rb_cFile, wpty_fd, FMODE_WRITABLE | FMODE_TRUNC | FMODE_CREATE | FMODE_SYNC, pty_path, RUBY_IO_TIMEOUT_DEFAULT, NULL ); res = rb_ary_new2(3); rb_ary_store(res, 0, rport); rb_ary_store(res, 1, wport); rb_ary_store(res,2,PIDT2NUM(info.child_pid)); if (rb_block_given_p()) { rb_ensure(rb_yield, res, pty_detach_process, (VALUE)&info); return Qnil; } return res; }
Spawns the specified command on a newly allocated pty. You can also use the alias ::getpty
.
The command’s controlling tty is set to the slave device of the pty and its standard input/output/error is redirected to the slave device.
env
is an optional hash that provides additional environment variables to the spawned pty.
# sets FOO to "bar" PTY.spawn({"FOO"=>"bar"}, "printenv", "FOO") do |r, w, pid| p r.read #=> "bar\r\n" ensure r.close; w.close; Process.wait(pid) end # unsets FOO PTY.spawn({"FOO"=>nil}, "printenv", "FOO") do |r, w, pid| p r.read #=> "" ensure r.close; w.close; Process.wait(pid) end
command
and command_line
are the full commands to run, given a String
. Any additional arguments
will be passed to the command.
Return values¶ ↑
In the non-block form this returns an array of size three, [r, w, pid]
.
In the block form these same values will be yielded to the block:
r
-
A readable
IO
that contains the command’s standard output and standard error w
-
A writable
IO
that is the command’s standard input pid
-
The process identifier for the command.
Clean up¶ ↑
This method does not clean up like closing IOs or waiting for child process, except that the process is detached in the block form to prevent it from becoming a zombie (see Process.detach
). Any other cleanup is the responsibility of the caller. If waiting for pid
, be sure to close both r
and w
before doing so; doing it in the reverse order may cause deadlock on some OSes.