expectperl-discuss — Discussions about Perl modules Expect.pm, IO::Tty and IO::Stty

Re: [Expectperl-discuss] Need Help with Capturing Output

[<RGi...@a1...>]

> I am using perl 5.6 with Expect module 1.12 and trying to get the 
> size of a file on a remote server.

You don't need Expect for this, set up ssh with RSA key authentication 
without passphrase (so ssh doesn't prompt for a password) and use ssh 
directly:

  $filesize = (split " ", qx{ssh user@host ls -l /path/to/file})[4];

or

  $filesize = qx{ssh user@host perl -e 'print -s "/path/to/file"'};
  chomp $filesize;

From the FAQ:

  I want to automate password entry for su/ssh/scp/rsh/...
  
  You shouldn't use Expect for this.  Putting passwords,
  especially root passwords, into scripts in clear text can
  mean severe security problems.  I strongly recommend using
  other means.  For 'su', consider switching to 'sudo', which
  gives you root access on a per-command and per-user basis
  without the need to enter passwords.  'ssh'/'scp' can be set
  up with RSA authentication without passwords.  'rsh' can use
  the .rhost mechanism, but I'd strongly suggest to switch to
  'ssh'; to mention 'rsh' and 'security' in the same sentence
  makes an oxymoron.
  
  It will work for 'telnet', though, and there are valid uses
  for it, but you still might want to consider using 'ssh', as
  keeping cleartext passwords around is very insecure.
  
  I want to use Expect to automate [anything with a buzzword]...
  
  Are you sure there is no other, easier way?  As a rule of
  thumb, Expect is useful for automating things that expect to
  talk to a human, where no formal standard applies.  For
  other tasks that do follow a well-defined protocol, there
  are often better-suited modules that already can handle
  those protocols.  Don't try to do HTTP requests by spawning
  telnet to port 80, use LWP instead.  You don't use a
  screwdriver to hammer in your nails either, or do you?

Hope this helps,

Roland
--
RGi...@cp...

Re: [Expectperl-discuss] Need Help with Capturing Output

[Austin S. <te...@of...>]

On Wed, Mar 13, 2002 at 02:01:12PM -0600, al....@ac... wrote:
> 
> Austin,
> 
> Thank you.  Your comments worked.  Now I can see the remote file size on my
> screen.  I need to assign
> that value to a variable so that I can compare it against the size of my
> other files.  How can I assign the
> returned value to a local variable?
> 

	Take a look at $object->before() in the docs, or $before_match in
the $object->expect() section. Basically the idea is to use the output
before the matched prompt. In other words you would do something like

$ssh->expect($timeout, $prompt);

print $ssh "$command\r";

$ssh->expect($timeout, $prompt);

$output = $ssh->before();

	Austin

Re: [Expectperl-discuss] Need Help with Capturing Output

[<al....@ac...>]

Austin,

Thank you.  Your comments worked.  Now I can see the remote file size on my
screen.  I need to assign
that value to a variable so that I can compare it against the size of my
other files.  How can I assign the
returned value to a local variable?


Al

*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
Al Pashazadeh
Accenture
CIO Technology Services - Network Services
Dallas, Infomart
Phone:  (214) 672-4255    VPN: 573-4255
*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*


                                                                                                                 
            Austin Schutz                                                                                        
            <te...@of...>                To:     al....@ac...                                     
                                         cc:     exp...@li...                        
            03/13/2002 12:20 PM          Subject:     Re: [Expectperl-discuss] Need Help with Capturing Output   
                                                                                                                 
                                                                                                                 



On Wed, Mar 13, 2002 at 12:05:50PM -0600, al....@ac...
wrote:
> Hello,
>
> I am using perl 5.6 with Expect module 1.12 and trying to get the size of
a
> file on a remote server.
> I tried piping the output of ls -l on the remote  server to awk but that
> only returned the output of the ls -l
> commend.  I searched the archived messages and noticed that this issue
was
> reported as a bug.
> Does anyone knows if this bug is fixed?  For now, I created a short
script
> on the remote server to
> display the size of the file.  The scritpt works fine when it is excuted
> from within the spawned ssh
> but I am not sure how to capture its output.  I need to compare the
> returned value (from my script)
> against the size of a similar file on the local server.  Any help is
> appreciated.
>
>

#!/usr/bin/perl
use Expect;

$ssh = Expect->spawn('ssh host.com');

$ssh->expect(60,"d: ");

print $ssh 'password' . "\n";

$ssh->expect(60,'> ');

print $ssh 'ls -l | awk \'{print $5}\''."\n";

$ssh->expect(60,'> ');


           I dunno, works for me.

           Austin





This message is for the designated recipient only and may contain
privileged, proprietary, or otherwise private information.  If you have
received it in error, please notify the sender immediately and delete the
original.  Any other use of the email by you is prohibited.

Re: [Expectperl-discuss] Need Help with Capturing Output

[Austin S. <te...@of...>]

On Wed, Mar 13, 2002 at 12:05:50PM -0600, al....@ac... wrote:
> Hello,
> 
> I am using perl 5.6 with Expect module 1.12 and trying to get the size of a
> file on a remote server.
> I tried piping the output of ls -l on the remote  server to awk but that
> only returned the output of the ls -l
> commend.  I searched the archived messages and noticed that this issue was
> reported as a bug.
> Does anyone knows if this bug is fixed?  For now, I created a short script
> on the remote server to
> display the size of the file.  The scritpt works fine when it is excuted
> from within the spawned ssh
> but I am not sure how to capture its output.  I need to compare the
> returned value (from my script)
> against the size of a similar file on the local server.  Any help is
> appreciated.
> 
> 

#!/usr/bin/perl
use Expect;

$ssh = Expect->spawn('ssh host.com');

$ssh->expect(60,"d: ");

print $ssh 'password' . "\n";

$ssh->expect(60,'> ');

print $ssh 'ls -l | awk \'{print $5}\''."\n";

$ssh->expect(60,'> ');


	I dunno, works for me.

	Austin

[Expectperl-discuss] Need Help with Capturing Output

[<al....@ac...>]

Hello,

I am using perl 5.6 with Expect module 1.12 and trying to get the size of a
file on a remote server.
I tried piping the output of ls -l on the remote  server to awk but that
only returned the output of the ls -l
commend.  I searched the archived messages and noticed that this issue was
reported as a bug.
Does anyone knows if this bug is fixed?  For now, I created a short script
on the remote server to
display the size of the file.  The scritpt works fine when it is excuted
from within the spawned ssh
but I am not sure how to capture its output.  I need to compare the
returned value (from my script)
against the size of a similar file on the local server.  Any help is
appreciated.


Al



This message is for the designated recipient only and may contain
privileged, proprietary, or otherwise private information.  If you have
received it in error, please notify the sender immediately and delete the
original.  Any other use of the email by you is prohibited.

[Expectperl-discuss] Sending Control Character

[<al....@ac...>]

Roland,

Thank you for your help.   kill TERM => $exp->pid;   worked as you
described.


Al

*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*
Al Pashazadeh
Accenture
CIO Technology Services - Network Services
Dallas, Infomart
Phone:  (214) 672-4255    VPN: 573-4255
*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*


This message is for the designated recipient only and may contain
privileged, proprietary, or otherwise private information.  If you have
received it in error, please notify the sender immediately and delete the
original.  Any other use of the email by you is prohibited.

Re: [Expectperl-discuss] Sending Control Character

[<RGi...@a1...>]

> I need
> to do is to send a kill signal to a child process,
> that is waiting for input, on the remote server.  I don't know how 
> to find
> the child process id on the remote server so I thought
> trying  by sending something similar to  CTRL+C.  Any comment?

Actually, signal propagation is the job of the communicating entity.  
Just send it to the spawned telnet/ssh client and he should do the rest.

  kill TERM => $exp->pid;

should do the trick.

HTH,

Roland
--
RGi...@cp...

Re: [Expectperl-discuss] Sending Control Character

[Austin S. <te...@of...>]

On Tue, Mar 12, 2002 at 04:36:46PM -0600, al....@ac... wrote:
> David,
> 
> Thank you for your comment.  I tried that approach before but what I need
> to do is to send a kill signal to a child process,
> that is waiting for input, on the remote server.  I don't know how to find
> the child process id on the remote server so I thought
> trying  by sending something similar to  CTRL+C.  Any comment?
> 

	Seems like it should work. You may also be able to send it
an EOF ("\cd"). Most programs will respond to that also. Something
we do here for routers is explicitly send them "exit\r" to make them
close a connection, since they've been known to get confused otherwise.

	Austin

[Expectperl-discuss] Sending Control Character

[<al....@ac...>]

David,

Thank you for your comment.  I tried that approach before but what I need
to do is to send a kill signal to a child process,
that is waiting for input, on the remote server.  I don't know how to find
the child process id on the remote server so I thought
trying  by sending something similar to  CTRL+C.  Any comment?

Al




This message is for the designated recipient only and may contain
privileged, proprietary, or otherwise private information.  If you have
received it in error, please notify the sender immediately and delete the
original.  Any other use of the email by you is prohibited.

RE: [Expectperl-discuss] Sending Control Character

[Blackstone, J. D. <jda...@ci...>]

  You can get control-anything with the special sequence "\cX" in a
double-quoted string, where X is the character you want the control of.

  But that's a Perl question, not an Expect.pm question. :)

jdb

> -----Original Message-----
> From: al....@ac... [mailto:al....@ac...]
> Sent: Tuesday, March 12, 2002 3:22 PM
> To: exp...@li...
> Subject: [Expectperl-discuss] Sending Control Character
> 
> 
> Hello,
> 
> I need to send  a CTRL+C  (or CTRL+D) from within an Expect 
> module, running
> on my local server, to a child process that is
> waiting for an input on my remote server.  Does anybody know 
> what is the
> equivalent character to CTRL+C or CTRL+D that can be
> sent to a child process?  Thank you.
> 
> 
> Al
> 
> 
> 
> 
> This message is for the designated recipient only and may contain
> privileged, proprietary, or otherwise private information.  
> If you have
> received it in error, please notify the sender immediately 
> and delete the
> original.  Any other use of the email by you is prohibited.
> 
> 
> _______________________________________________
> Expectperl-discuss mailing list
> Exp...@li...
> https://lists.sourceforge.net/lists/listinfo/expectperl-discuss
> 

[Expectperl-discuss] Sending Control Character

[<al....@ac...>]

Hello,

I need to send  a CTRL+C  (or CTRL+D) from within an Expect module, running
on my local server, to a child process that is
waiting for an input on my remote server.  Does anybody know what is the
equivalent character to CTRL+C or CTRL+D that can be
sent to a child process?  Thank you.


Al




This message is for the designated recipient only and may contain
privileged, proprietary, or otherwise private information.  If you have
received it in error, please notify the sender immediately and delete the
original.  Any other use of the email by you is prohibited.

Re: [Expectperl-discuss] Sending Control Characters

[Austin S. <te...@of...>]

On Mon, Mar 11, 2002 at 04:15:35PM -0800, Wim Verhaert wrote:
> Does anyone know how to send any control sequences (eg. CTRL+D) to the 
> expect module?
> 
> Thanks
> 

	Expect objects are filehandles.

$telnet = Expect->spawn('telnet');

print $telnet "\cd";

	Austin

[Expectperl-discuss] Sending Control Characters

[Wim V. <wve...@re...>]

Does anyone know how to send any control sequences (eg. CTRL+D) to the 
expect module?

Thanks

Wim

Re: [Expectperl-discuss] Using Expect with other programs

[Hal V. <ha...@th...>]

Yes, it's Term::VT102.  Oops.

The system I'm dealing with has a habit of freezing at times, so VT102 by 
itself would not work, since I would have no way of detecting a timeout.  I 
agree having a callback for a specific screen X,Y co-ord would be helpful, 
but it would not help me for this specific system.  It always returns the 
cursor to 80,24 when posting a screen (on the other hand -- that could prove 
helpful...).

If I were used to data structures and OOP, I'd look at finding a way to 
combine the two -- that would be extremely powerful.  One could get callbacks 
if certain prompts were sent through the VT102, but could also check the 
exact screen locations, to get more information (with this system that would 
be a HUGE help).

Off the top of my head, I was thinking I could do an expect statement and, 
after it's matched, feed the exp_before string into VT102.  Assuming there's 
only been one screen of info (or less) since the last expect, that should put 
it all on the virtual screen, right?  I'm going to look into that.

Thanks for the comments.

Hal


On Friday 08 March 2002 08:04 am, RGi...@a1... wrote:
> > I've been using Expect for Perl to handle a connection with
> > another system
> > over a modem and it's been working well, but I've recently found
> > IO::VT102.
>
> That's Term::VT102, right?
>
> > I've been using IO:VT102 in post processing the capture files.
> > Now I'm wondering if it's possible to use this mod with Expect.
> > That would let me extract the need information while online.
> >
> > To use VT102, I need to be able to send it the input strings, like
> > $vt->process($mydata).
>
> An Expect object is just a file handle that you can read from and write
> to, so it should be easy to link that to Term::VT102, just read a chunk
> and call process() with it.  The OUTPUT callback should simply do a
> syswrite to the Expect object.
>
> > Is there some way to feed data coming from the modem into both
> > Expect and VT102 at the same time?  It would mean I can use
> > Expect to tell when data I need is on the screen, then I could
> > pull it straight from the virtual VT102 terminal.
>
> No, you can't use the expect() method concurrently, as this would
> consume the characters.  Instead, you have to register a ROWCHANGE
> callback and do your processing there via normal regexp matches.
>
> [BTW, what I think is missing from Term::VT102 is the ability to get a
> callback when the cursor position enters certain coordinates (to be
> able to react to a prompt).]
>
> Hope this helps,
>
> Roland

Re: [Expectperl-discuss] Using Expect with other programs

[<RGi...@a1...>]

> I've been using Expect for Perl to handle a connection with 
> another system 
> over a modem and it's been working well, but I've recently found 
> IO::VT102.  

That's Term::VT102, right?

> I've been using IO:VT102 in post processing the capture files.  
> Now I'm wondering if it's possible to use this mod with Expect.
> That would let me extract the need information while online.
> 
> To use VT102, I need to be able to send it the input strings, like 
> $vt->process($mydata).

An Expect object is just a file handle that you can read from and write 
to, so it should be easy to link that to Term::VT102, just read a chunk 
and call process() with it.  The OUTPUT callback should simply do a 
syswrite to the Expect object.

> Is there some way to feed data coming from the modem into both 
> Expect and VT102 at the same time?  It would mean I can use
> Expect to tell when data I need is on the screen, then I could
> pull it straight from the virtual VT102 terminal.

No, you can't use the expect() method concurrently, as this would 
consume the characters.  Instead, you have to register a ROWCHANGE 
callback and do your processing there via normal regexp matches.

[BTW, what I think is missing from Term::VT102 is the ability to get a 
callback when the cursor position enters certain coordinates (to be 
able to react to a prompt).]

Hope this helps,

Roland
--
RGi...@cp...

[Expectperl-discuss] Using Expect with other programs

[Hal V. <ha...@th...>]

I've been using Expect for Perl to handle a connection with another system 
over a modem and it's been working well, but I've recently found IO::VT102.  
I've been using IO:VT102 in post processing the capture files.  Now I'm 
wondering if it's possible to use this mod with Expect.  That would let me 
extract the need information while online.

To use VT102, I need to be able to send it the input strings, like 
$vt->process($mydata).  I'm sure if I were more experienced, the answer would 
be obvious, but I'm still getting used to perl and all the tools it and *nix 
have.

Is there some way to feed data coming from the modem into both Expect and 
VT102 at the same time?  It would mean I can use Expect to tell when data I 
need is on the screen, then I could pull it straight from the virtual VT102 
terminal.

Thanks for any thoughts or suggestions.

Hal

[Expectperl-discuss] perl test.pl eof test failed on solaris 8

[Yang Tj-A. <AT...@mo...>]

Hi, there 

I got Expect-1.11 with  IO-Stty-0.2,IO-Tty-0.92_02 to compiled and
packaged on my solaris 2.6 box and it pass the "perl test.pl".
But when I install these packages on solaris 8 and run perl test.pl
again, it failed as shown below  
1..7
ok 1
ok 2
ok 3
ok 4
not ok 5
# Failed test 5 in test.pl at line 50
ok 6
ok 7

test 5 is a test on getting eof from a process, should I concern
about this failing test ? my program seems working fine so far.


I am writing web page password changer using expect.pm. please
help and I will make up a easy one for inclusion in  faq or 
examples section.(if others think my example is good enough to
be included).

Thanks

tj

Install machine solaris 8 information
-----------------------------------------------------------------
test> perl -v

This is perl, version 5.005_02 built for sun4-solaris

Copyright 1987-1998, Larry Wall

Perl may be copied only under the terms of either the Artistic License or
the
GNU General Public License, which may be found in the Perl 5.0 source kit.

Complete documentation for Perl, including FAQ lists, should be found on
this system using `man perl' or `perldoc perl'.  If you have access to the
Internet, point your browser at http://www.perl.com/, the Perl Home Page.

test> uname -a
SunOS tango 5.8 Generic_108528-06 sun4u sparc SUNW,UltraAX-i2
test>

Compile machine information
----------------------------------------------------------------
root> uname -a
SunOS bingo 5.6 Generic_105181-12 sun4u sparc SUNW,Ultra-5_10
root> perl -v

This is perl, version 5.005_02 built for sun4-solaris

Copyright 1987-1998, Larry Wall

Perl may be copied only under the terms of either the Artistic License or
the
GNU General Public License, which may be found in the Perl 5.0 source kit.

Complete documentation for Perl, including FAQ lists, should be found on
this system using `man perl' or `perldoc perl'.  If you have access to the
Internet, point your browser at http://www.perl.com/, the Perl Home Page.

root> CC -V
CC: WorkShop Compilers 5.0 98/12/15 C++ 5.0
root> gcc -v
Reading specs from /opt/local/lib/gcc-lib/sparc-sun-solaris2.6/2.8.1/specs
gcc version 2.8.1
root> 
-- 
T.J. Yang     
 

RE: RE: [Expectperl-discuss] does timeout => sub{} pass filehandl e?

[Sorrell, A. <Al_...@tr...>]

Thanks! That's what I needed. Working great now.
Al

-----Original Message-----
From: RGi...@a1... [mailto:RGi...@a1...]
Sent: Tuesday, March 05, 2002 6:10 AM
To: Sorrell, Al
Cc: 'exp...@li...'
Subject: Re: RE: [Expectperl-discuss] does timeout => sub{} pass
filehandle?


> I'm still having a problem trying to "send" from inside a timeout 
> condition:

Oh, sorry for that mistake:  as the timeout is shared among all matches 
(which could be on different spawned programs), its callback doesn't 
get an expect object, you have to use one from the outer context, i.e.

  my $exp = new Expect;
  $exp->raw_pty(1);
  $exp->spawn(...);
  $exp->expect($timeout,
               ...
               [ timeout => sub { $exp->send() }]);

Hope this helps,

Roland
--
RGi...@cp...

Re: RE: [Expectperl-discuss] does timeout => sub{} pass filehandle?

[<RGi...@a1...>]

> I'm still having a problem trying to "send" from inside a timeout 
> condition:

Oh, sorry for that mistake:  as the timeout is shared among all matches 
(which could be on different spawned programs), its callback doesn't 
get an expect object, you have to use one from the outer context, i.e.

  my $exp = new Expect;
  $exp->raw_pty(1);
  $exp->spawn(...);
  $exp->expect($timeout,
               ...
               [ timeout => sub { $exp->send() }]);

Hope this helps,

Roland
--
RGi...@cp...

RE: [Expectperl-discuss] does timeout => sub{} pass filehandle?

[Sorrell, A. <Al_...@tr...>]

I loaded gcc 2.95.3 and resolved the cpp issue and so installed
Expect-1.13_08 & IO-Tty-0.97_01 successfully.  The $exp->raw_pty(1) works
fine now. Thanks!

I'm still having a problem trying to "send" from inside a timeout condition:

code frag:
  my $timeouts=0;
  $exp->send("\n");
  $exp->expect($timeout,
        ['OVL111', sub { my $fh = shift; $fh->send("login\n"); }],
        ['-ex','.', sub { my $fh = shift;
$fh->send("****\nlogo\n");exp_continue; }],
#        ['SCH', sub { my $fh = shift;
$fh->send("****\nlogo\n");exp_continue; }],
#        [timeout => sub {die "Timed out trying to login";}]
        [timeout => sub { $timeouts++; if($timeouts < 5)
                {my $fh = shift; $fh->send("****\nlogo\n");exp_continue; }
                else { die "Timed out trying to clear loging";}
                        }]
        );

results in:
TIMEOUT
Calling timeout function ARRAY(0x3831fc)...
Can't call method "send" on unblessed reference at get_pbx_routes.pl line
166.

line 166 is the timeout in the code above

=================
This is the original Expect script code that I'm trying to mimic:

send -- "\r"
expect {
        timeout         {send_error "System still logged in-sending
logout\n"; send "****\rlogo\r"}
        "."             {send_error "System still logged in-sending
logout\n"; send "****\rlogo\r"}
        "OVL111"
}
send -- "login\r"


-----Original Message-----
From: Sorrell, Al 
Sent: Thursday, February 28, 2002 11:54 AM
To: 'RGi...@cp...'
Subject: RE: [Expectperl-discuss] does timeout => sub{} pass filehandle?


Thanks for the new material.
1) In trying to perl Makefile.PL, I'm running into a problem. Looks like you
changed from 'cc' to 'cpp'. Is cpp really needed, and if so what version
(I'm assuming that's gcc's C++ compiler). I currently have 2.8.1 installed,
which creates a c++

2) I really was trying to match "." - this app is going to the console of a
Nortel Meridian PBX via a router's AUX port and the logged-in prompt for
this thing is "."

Thanks,
Al
-----Original Message-----
From: RGi...@a1... [mailto:RGi...@a1...]
Sent: Thursday, February 28, 2002 5:23 AM
To: Sorrell, Al
Cc: 'exp...@li...'
Subject: Re: [Expectperl-discuss] does timeout => sub{} pass filehandle?


> Solaris 2.6 Expect 1.12
> In all the examples, when a timeout is encountered you simply die();
> In my case, a device can get into a state where what is returned is
> unpredictable. There is a specific sequence that I can then send 
> to log out the previous action and get back to a known state. 
> However, the code below doesn't seem to work correctly
>  $exp->expect($timeout,
>        ['OVL111', sub { my $fh = shift; $fh->send("login\n"); }],
>        ['\.', sub { my $fh = shift; $fh-
> >send("****\nlogo\n");exp_continue;}],

Are you matching for a dot here?

>        [timeout => sub { my $fh = shift;
> $fh->send("****\nlogo\n");exp_continue; }]
>        );

exp_continue for timeout only started working a few beta versions ago. 

> Using a regex of '.*' would be another way - but risky - 
> especially until I can get "raw" working. Any suggestions?

I'm attaching the latest version, much improved and almost ready for 
release (I still need a few test results for IO-Tty, especially on 
HPUX, OSF and SCO).  Please try it (be sure to browse the docs for all 
the new features) and tell me if it solves your problems.

Thanks!

Roland
--
RGi...@cp...

Re: [Expectperl-discuss] Can I use expectperl from a sub-process?

[<RGi...@a1...>]

> Here's what I'd like to do with Expect.pm:
> 
> * Spawn a telnet to somewhere.
> * Invoke a subshell
> *   From within that subshell, call another Perl program one or more
> times to interact with the telnet process spawned by the parent 
> process.

Let me try to summarize what I think you want to accomplish:

* connect to another machine via telnet/whatever
* have the user either
1) interact with the other machine
2) or start procedures/macros/scripts/whatever to do more complicated 
things on the other machine.

So what you effectively want to do is to impose yourself between the 
user and the remote connection.  Sounds like a job for interact
()/interconnect()...

> My first try was to write a short Perl program that spawned a telnet,
> did an expect to get to the prompt, then did 'system("bash");' to get
> into the subshell.  I then invoked program#2 which tried to interact
> with the telnet.  I used the pty that the spawned telnet was using
> as fd's 0,1 and 2 (found via lsof)  as the filename to re-open with
> exp_init(), but that didn't work.  It did sent the data to the pty,
> but it was received by the parent program rather than the spawned 
> telnet.I suspect I'probably running afoul of the pty master/slave 
> business,which I'm not very familiar with.

Ugh, that sounds twisted.  Why do you want to start another shell?  If 
you just want to ask the user what procedure/macro to start, why not do 
it yourself?  Use Readline and just give the user a prompt to enter 
something, then send that to the other machine or evaluate it 
yourself.  Or use interact() to have the user directly talk to the 
other machine.

If you want to write different perl subroutines for some tasks on the 
other machine, just get a string from the user and eval() it (there is 
no need to invent another language/write another parser, you already 
have perl at hand).  That way, the user can call predefined procedures 
from a library.

If you want to write scripts or programs in other languages to handle 
such tasks, have them set up to use stdin/out to communicate with the 
other machine, then spawn them with Expect and use interconnect() to 
get them talking to the spawned telnet connection.

I hope I presented some new ideas to get you started into the right 
direction.  Remember: KISS: Keep It Small and Simple!

Roland
--
RGi...@cp...

Re: RE: [Expectperl-discuss] SOLVED: Need to get the very last thing output b y the program

[<RGi...@a1...>]

> This error occured because of the way I was creating the
> $exp object.  I wanted to create it before spawning the
> command so that I could do a $exp->log_user(0) before
> the process started spitting out stuff.

Ah, yes, you are a premature coder, this splitting of new() and spawn() 
was introduced in one of the latest beta versions and is not available 
in v1.12.  And BTW you are right to want to do it that way...

Just a few more days 'til the next release...

Roland
--
RGi...@cp...


  What I discovered
> is that you can go it after just fine, as below:
> 
> $exp = Expect->spawn ( $command, @params );
> $exp->log_user(0);
> $exp->expect (.....
> 
> So now things are working.  Thought this might help
> others trying to do something similar.
> 
> Patrick
> 
> _______________________________________________
> Expectperl-discuss mailing list
> Exp...@li...
> https://lists.sourceforge.net/lists/listinfo/expectperl-discuss
> 

Re: [Expectperl-discuss] Can I use expectperl from a sub-process?

[Austin S. <te...@of...>]

> 
> But for an ssh or Kerberized telnet session, that won't work - telnet or
> ssh has to stay running to encrypt/decrypt the data - the only way to
> talk to the remote router is via the pty of ssh/telnet.  I could have
> the perl/expect program start ssh/telnet, but my program is meant to be
> a tool that you run multiple times against the router, so I want the ssh/
> telnet to persist.
> 
> > 	Another way is to have an expect script that does interact() at
> > some point and uses named pipes as i/o. I think that's how things like
> > kibitz work. Performance-wise it's less efficient, but probably easier to
> > code.
> 
> Easy to code, I like that :-).  I had earlier tried using ksh93,
> which has a built-in for doing this - when you
> start a process and background it with "|&" instead of "&", you
> end up with two pipes (meant for use with ksh's "read -p" and "print -p"
> commands).  I suppose I can try hooking one of the pipes to the Expect
> object, and teaching my script to send the output data for the telnet/ssh
> session to the other?

	Something like that would probably be my approach.

	Austin

Re: [Expectperl-discuss] Can I use expectperl from a sub-process?

[Ed R. <er...@pa...>]

Austin Schutz writes:
> On Sat, Mar 02, 2002 at 12:40:25AM -0500, Ed Ravin wrote:
> > Here's what I'd like to do with Expect.pm:
> > * Spawn a telnet to somewhere.
> > * Invoke a subshell
> > *   From within that subshell, call another Perl program one or more
> > times to interact with the telnet process spawned by the parent process.

> 	Thirdly, you could have the called perl program call the telnet
> expect process and work with it as if it were any other process.  It's hard to
> tell (for me anyway) what you're trying to accomplish.

What I want to do (and why this option won't work) is to set up Kerberized
telnet or ssh session to a Cisco router, drop into a shell, and be able
to run shell commands that manipulate the router via the session.  I've
already got this working with regular (cleartext) telnet sessions - the
filehandle with the telnet session is inherited by telnet's children  - I
subshell from telnet and then have the Perl/Expect program use exp_init()
to hook up to fd3 and I'm all set.  I can call the Perl/Expect program
multiple times to manipulate the router via the Telnet session, I can call
the editor or other shell programs as needed, and when I'm done I exit
the shell and return back to the telnet prompt.

But for an ssh or Kerberized telnet session, that won't work - telnet or
ssh has to stay running to encrypt/decrypt the data - the only way to
talk to the remote router is via the pty of ssh/telnet.  I could have
the perl/expect program start ssh/telnet, but my program is meant to be
a tool that you run multiple times against the router, so I want the ssh/
telnet to persist.

> 	Another way is to have an expect script that does interact() at
> some point and uses named pipes as i/o. I think that's how things like
> kibitz work. Performance-wise it's less efficient, but probably easier to
> code.

Easy to code, I like that :-).  I had earlier tried using ksh93,
which has a built-in for doing this - when you
start a process and background it with "|&" instead of "&", you
end up with two pipes (meant for use with ksh's "read -p" and "print -p"
commands).  I suppose I can try hooking one of the pipes to the Expect
object, and teaching my script to send the output data for the telnet/ssh
session to the other?

Re: [Expectperl-discuss] Can I use expectperl from a sub-process?

[Hal V. <ha...@th...>]

I spent last week developing a program that does almost the same thing and 
this week I've been successfully testing and using it online with another 
system.

If you're going through the modem, or another device, you can open it like a 
file (open DEVICE, "+>/dev/modem/").  There are a few other commands to use 
to handle IO.  I can't remember them, but if you look at the MODEMINIT 
example in the tutorial, included with the Expect module, you will see all 
you need for initializing the connection with the modem and for closing it.  
This method, which I am using quite successfully (with only 1 problem, which 
I'll mention in a bit), lets you easily dial in and communicate with another 
computer using only perl, no telnet, no other programs at all.  I love it, 
since I was previously using TCL and the Expect that was developed using TCL. 
 That system controlled Minicom, which worked well, but I found TCL a pain to 
program in, and, although I like Minicom, I feel much more comfortable using 
only perl, instead of TCL, Expect, and Minicom.  Fewer programs means less to 
go wrong.

You send any commands to the modem (or whatever device you want to open) this 
way, which gives you full control.  If you're using any capture files, you 
may get many more escape and control codes in the capture than you're used 
to, but they can be purged with a simple filter subroutine (or awk, if you 
want).

There are examples in the tutorial files that also use telnet (I think).  If, 
for some reason, you don't have the tutorial files, you can download the 
lastest archive from CPAN, and they're in there.  If that doesn't work, I'll 
be glad to send them to you.

The only problem I'm having with this setup is when I try to capture 
screenfuls of ino.  The program I'm communicating with prints a screen of 
info, then I type in a new ID number, and it prints a new screen of info.  I 
use Expect to detect when each screen is finished displaying.  After the last 
screen, I terminate a loop, then close the log file.  I've found if I don't 
display the last screen again, and use Expect to wait for the display to 
complete again, the last screen is never caught in the capture file.  In this 
case the last screen is displayed twice to get it once.  It seems like the 
input for the last "expect" command is tossed and never makes it to the log 
file.  This can be worked around, so it isn't a huge problem, but I've 
noticed it, so I thought a heads-up was appropriate.

Hal


On Saturday 02 March 2002 12:40 am, you wrote:
> Here's what I'd like to do with Expect.pm:
>
> * Spawn a telnet to somewhere.
> * Invoke a subshell
> *   From within that subshell, call another Perl program one or more
> times to interact with the telnet process spawned by the parent process.
>
> My first try was to write a short Perl program that spawned a telnet,
> did an expect to get to the prompt, then did 'system("bash");' to get
> into the subshell.  I then invoked program#2 which tried to interact
> with the telnet.  I used the pty that the spawned telnet was using
> as fd's 0,1 and 2 (found via lsof)  as the filename to re-open with
> exp_init(), but that didn't work.  It did sent the data to the pty,
> but it was received by the parent program rather than the spawned telnet.
> I suspect I'probably running afoul of the pty master/slave business,
> which I'm not very familiar with.
>
> Any idea how to get what I want to happen?  Is it even possible?
>
> Thanks,
> 	-- Ed
>
> _______________________________________________
> Expectperl-discuss mailing list
> Exp...@li...
> https://lists.sourceforge.net/lists/listinfo/expectperl-discuss