Subprocess module: running an interactive shell
Joe Tyson
joe at routespot.com
Sun Mar 15 03:09:06 EDT 2009
On 2009-03-14 20:10:29 -0400, Karthik Gurusamy <kar1107 at gmail.com> said:
> On Mar 14, 3:03 am, Roman Medina-Heigl Hernandez <ro... at rs-labs.com>
> wrote:
>> Karthik Gurusamy escribió:
>>
>>
>>
>>> On Mar 13, 6:39 pm, Roman Medina-Heigl Hernandez <ro... at rs-labs.com>
>>> wrote:
>>>> Hi,
>>
>>>> I'm experimenting with Python and I need a little help with this. What
> I'd
>>>> like is to launch an interactive shell, having the chance to send firs
> t
>>>> several commands from python. I've written the following code:
>>
>>>> ============
>>
>>>> #!/usr/bin/env python
>>
>>>> import sys, subprocess
>>
>>>> exe = "/bin/sh"
>>>> params = "-i"
>>
>>> -i says shell to be interactive. So looks like it is directly trying
>>> to read from the terminal.
>>
>> Well, then the question will be: is there any way to tell python to
>> directly "map" the terminal to the subprocess?
>
> pexpect seems to be the solution for such problems :). [other
> applications include ssh which asks for password from terminal (not
> ssh's stdin)]
>
> http://pexpect.sourceforge.net/pexpect.html
>
>>
>>>> proc = subprocess.Popen([exe, params], stdin=subprocess.PIPE)
>>
>>> proc = subprocess.Popen([exe,], stdin=subprocess.PIPE)
>>
>>> works for me; but if there is an error 'sh' terminates.
>>
>>> If you want to simulate interactive, explore the pexpect module.
>>
>> I'll get it a try :)))
>>
>>>> proc.stdin.write("id\n")
>>
>>>> while True:
>>>> line = sys.stdin.readline()
>>>> if not line:
>>
>>> note that a simple enter terminates the shell which you may not want.
>>
>> Test my code and you'll see that this is not true :) When you hit enter
>> line will contain '\n' so it's not empty.
>
> You are right. I thought readline() strips the trailing \n (It doesn't
> and shouldn't as it's necessary for the case a file ends without a
> newline).
>
>>
>>>> break
>>>> proc.stdin.write(line)
>>
>> Btw, another curiosity I have: is it possible to make a print not
>> automatically add \n (which is the normal case) neither " " (which happen
> s
>> when you add a "," to the print sentence)? I found an alternative not
>> using print at all, eg: sys.stdout.write("KKKKK"). But it resulted strang
> e
>> to me having to do that trick :)
>
> I am also aware of only the sys.stdout.write solution.
>
> python3.0 has a way to do it.
>
>>>> help(print)
> Help on built-in function print in module builtins:
>
> print(...)
> print(value, ..., sep=' ', end='\n', file=sys.stdout)
>
> Prints the values to a stream, or to sys.stdout by default.
> Optional keyword arguments:
> file: a file-like object (stream); defaults to the current
> sys.stdout.
> sep: string inserted between values, default a space.
> end: string appended after the last value, default a newline.
>
>>>> print('hello', end='')
> hello>>>
>
> Karthik
>
>>
>> Thank you for all your comments and comprenhension.
>>
>> -r
>>
>>
>>
>>>> sys.exit()
>>
>>>> ============
>>
>>>> The problem is that when I launch it, python proggy is automatically
>>>> suspended. The output I got is:
>>
>>>> roman at rslabs:~/pruebas$ ./shell.py
>>>> roman at rslabs:~/pruebas$ uid=1000(roman) gid=1000(roman) groups=1
> 000(roman)
>>>> roman at rslabs:~/pruebas$
>>
>>>> [2]+ Stopped ./shell.py
>>>> roman at rslabs:~/pruebas$
>>
>>>> Why and how to fix it? Would you suggest a better and more elegant way
> to
>>>> do what I want?
>>
>>> As I see it, 'sh' is attempting to read from the keyboard and not from
>>> stdin.
>>
>>> Karthik
>>
>>>> Thank you.
>>
>>>> --
>>
>>>> Saludos,
>>>> -Roman
>>
>>>> PGP Fingerprint:
>>>> 09BB EFCD 21ED 4E79 25FB 29E1 E47F 8A7D EAD5 6742
>>>> [Key ID: 0xEAD56742. Available at KeyServ]
>>
>>> --
>>> http://mail.python.org/mailman/listinfo/python-list
>>
>> --
>>
>> Saludos,
>> -Roman
>>
>> PGP Fingerprint:
>> 09BB EFCD 21ED 4E79 25FB 29E1 E47F 8A7D EAD5 6742
>> [Key ID: 0xEAD56742. Available at KeyServ]
There is quite a bit involved in handling this. Check out this recipe
on activestate:
http://code.activestate.com/recipes/278731/
- Joe
More information about the Python-list
mailing list