Expect::Simple - wrapper around the Expect module


Expect-Simple documentation Contained in the Expect-Simple distribution.

Index


Code Index:

NAME

Top

Expect::Simple - wrapper around the Expect module

SYNOPSIS

Top

  use Expect::Simple;

  my $obj = new Expect::Simple 
        { Cmd => [ dmcoords => 'verbose=1', "infile=$infile"],
	  Prompt => [ -re => 'dmcoords>:\s+' ],
	  DisconnectCmd => 'q',
	  Verbose => 0,
	  Debug => 0,
	  Timeout => 100
	};

  $obj->send( $cmd );
  print $obj->before;
  print $obj->after;
  print $obj->match_str, "\n";
  print $obj->match_idx, "\n";
  print $obj->error_expect;
  print $obj->error;

  $expect_object = $obj->expect_handle;

DESCRIPTION

Top

Expect::Simple is a wrapper around the Expect module which should suffice for simple applications. It hides most of the Expect machinery; the Expect object is available for tweaking if need be.

Generally, one starts by creating an Expect::Simple object using new. This will start up the target program, and will wait until one of the specified prompts is output by the target. At that point the caller should send() commands to the program; the results are available via the before, after, match_str, and match_idx methods. Since Expect simulates a terminal, there will be extra \r characters at the end of each line in the result (on UNIX at least). This is easily fixed:

    ($res = $obj->before) =~ tr/\r//d;
    @lines = split( "\n", $res );

This is not done automatically.

Exceptions will be thrown on error (match with /Expect::Simple/). Errors from Expect are available via the error_expect method. More human readable errors are available via the error method.

The connection is automatically broken (by sending the specified disconnect command to the target) when the Expect::Simple object is destroyed.

Methods

new
    $obj = Expect::Simple->new( \%attr );

This creates a new object, starting up the program with which to communicate (using the Expect spawn method) and waiting for a prompt. The passed hash reference must contain at least the Prompt, DisconnectCmd, and Cmd elements. The available attributes are:

Cmd
  Cmd => $command,
  Cmd => [ $command, $arg1, $arg2, ... ],

The command to which to connect. The passed command may either be a scalar or an array.

Prompt

This specifies one or more prompts to scan for. For a single prompt, the value may be a scalar; for more, or for matching of regular expressions, it should be an array reference. For example,

  Prompt => 'prompt1> ',
  Prompt => [ 'prompt1> ', 'prompt2> ', -re => 'prompt\d+>\s+' ]

All prompts are taken literally, unless immediately preceded by a -re flag, in which case they are regular expressions.

DisconnectCmd

This is the command to be sent to the target program which will cause it to exit.

RawPty

If set, then underlying Expect object's pty mode is set to raw mode (see Expect::raw_pty()).

Timeout

The time in seconds to wait until giving up on the target program responding. This is used during program startup and when any commands are sent to the program. It defaults to 1000 seconds.

Debug

The value is passed to Expect via its debug method.

Verbose

This results in various messages printed to the STDERR stream. If greater than 3, it turns on Expect's logging to STDOUT (via the log_stdout Expect method.

send
   $obj->send( $cmd );
   $obj->send( @cmds );

Send one or more commands to the target. After each command is sent, it waits for a prompt from the target. Only the output resulting from the last command is available via the after, before, etc. methods.

match_idx

This returns a unary based index indicating which prompt (in the list of prompts specified via the Prompt attribute to the new method) was received after the last command was sent. It will be undef if none was returned.

match_str

This returns the prompt which was matched after the last command was sent.

before

This returns the string received before the prompt. If no prompt was seen, it returns all output accumulated. This is usually what the caller wants to parse. Note that the first line will (usually) be the command that was sent to the target, because of echoing. Check this out to be sure!

after

This returns the 'after' string. Please read the Expect docs for more enlightenment.

error

This returns a cleaned up, more humanly readable version of the errors from Expect. It'll be undef if there was no error.

error_expect

This returns the original Expect error.

expect_handle

This returns the Expect object, in case further tweaking is necessary.

BUGS

Top

If the command to be run does not exist (or not in the current execution path), it's quite possible that the new method will not throw an exception. It's up to the caller to make sure that the command will run! There's no known workaround for this.

LICENSE

Top

This software is released under the GNU General Public License. You may find a copy at

   http://www.fsf.org/copyleft/gpl.html

AUTHOR

Top

Diab Jerius (djerius@cpan.org)


Expect-Simple documentation Contained in the Expect-Simple distribution.

package Expect::Simple;

use strict;
use warnings;

use Carp;
use Expect;

our $VERSION = '0.04';


sub new {
  my $proto = shift;
  my $class = ref($proto) || $proto;
  my $obj  = {
	      Timeout => 1000,
	      Debug => 0,
	      Verbose => 0,
	      Prompt => undef,
	      DisconnectCmd => undef,
	      Cmd => undef,
	      RawPty => 0,
	     };

  bless ($obj, $class);

  my $attr = shift or
    croak( __PACKAGE__, ': must specify some attributes!' );

  while( my ( $attr, $val ) = each %{$attr} )
  {
    croak( __PACKAGE__, ": attribute error : `$attr' is not recognized" )
      unless exists $obj->{$attr};

    $obj->{$attr} = $val;
  }


  # ensure all the attribures are set
  foreach ( keys %$obj )
  {
    croak( __PACKAGE__, ": must specify attribute `$_'" )
      unless defined $obj->{$_};
  }

  # rework prompt
  $obj->{Prompt} = [ 'ARRAY' eq ref $obj->{Prompt} ?
		   @{$obj->{Prompt}} : $obj->{Prompt} ];


  eval { $obj->_connect; };
  if ( $@ )
  {
      chomp $@;
      croak (__PACKAGE__, ': ', $@);
  }

  return $obj;
}

# _connect - start up the cmd
#
# creates an Expect object which talks to the specified command.  It dies with
# an appropriate message upon error.

sub _connect
{
  my $obj = shift;

  print STDERR "Running command..."
    if $obj->{Verbose};

  # don't store the connection in the object until we're sure it's
  # up. otherwise DESTROY will try to disconnect, which won't work...
  delete $obj->{_conn};
  my $conn = Expect->new();

  $conn->raw_pty(1) if $obj->{RawPty};

  # Expect docs say that upon failure of spawn, one can get the error
  # from the next call to expect().  Unfortunately, that message is
  # generated by a die in the forked child process, which means it has
  # propagated up through the eval in Expect::Simple::new and
  # had lots of cruft slapped onto it, front and back.
  #
  # instead, I use a horrible hack to grab the (undocumented) warning
  # emitted by Expect if $^W is turned on.

  {
      my $error;
      local $^W = 1;
      local $SIG{__WARN__} = sub { chomp( $error = $_[0]); return; };

      my $success =
        $conn->spawn( 'ARRAY' eq ref($obj->{Cmd})
                      ? @{$obj->{Cmd}}
                      :   $obj->{Cmd} );

      if ( ! $success )
      {
          chomp $error;
          die( "error spawning command: $error\n" );
      }
  }

  print STDERR "done.\n"
    if $obj->{Verbose};

  $conn->debug( $obj->{Debug} );

  $conn->log_stdout( $obj->{Verbose} > 3 ? 1 : 0 );

  $obj->{_conn} = $conn;
  $obj->_expect( @{$obj->{Prompt}} ) 
    or die( __PACKAGE__, ": couldn't find prompt\n");
}

sub _disconnect
{
  my $obj = shift;

  return unless
    $obj->{_conn} && !defined $obj->{_conn}->exitstatus;

  print STDERR "Disconnecting.\n"
    if $obj->{Verbose};

  $obj->{_conn}->print( $obj->{DisconnectCmd}, "\n" );
  $obj->_expect( 'the unexpected' );
  croak( __PACKAGE__, ": disconnection error" )
    unless $obj->{_conn}->exp_error =~ /^(2|3)/;

  $obj->{_conn} = undef;
}


# send( @commands )
#
# send commands to the server. each command is sent independently.
# it waits for the  prompt to indicate success.
#
# it croaks if there was an error.  $obj->error returns
# the results of the communication to
# the server which caused the error.

sub send
{
  my $obj = shift;

  foreach ( @_ )
  {
    print STDERR "Sending `$_'\n"
      if $obj->{Verbose} && ! $obj->{_conn}->log_stdout;
    $obj->{_conn}->print( $_, "\n");

    $obj->_expect( @{$obj->{Prompt}} ) || 
      croak( __PACKAGE__, ": couldn't find prompt after send");
  }
}


# _expect( @match_patterns )
#
# match output of the server.The error message is massaged to
# make it more obvious.
#
# it returns 1 upon success, undef if there was an error.

sub _expect
{
  my $obj = shift;

  my $match = $obj->{_conn}->expect( $obj->{Timeout}, @_ ); 

  $obj->{_error} = undef;

  unless ( defined $match )
  {
    local $_ = $obj->{_conn}->exp_error;

    if ( /^1/ )
    {
      $obj->{_error} = 'connection timed out';
    }
    elsif ( /^(2|3)/ )
    {
      $obj->{_error} = 'connection unexpectedly terminated';
    }
    else
    {
      my ( $errno, $errmsg) = /(\d):(.*)/;

      $obj->{_error} = "error in communications: $errmsg";
    }

    return undef;
  }

  1;
}

sub error        { shift()->{_error} }
sub error_expect { shift()->{_conn}->exp_error }
sub match_idx    { shift()->{_conn}->exp_match_number }
sub match_str    { shift()->{_conn}->exp_match }
sub before       { shift()->{_conn}->exp_before }
sub after        { shift()->{_conn}->exp_after }
sub expect_handle{ shift()->{_conn} }

sub DESTROY { shift()->_disconnect }


# Autoload methods go after =cut, and are processed by the autosplit program.

1;
__END__