From: Tim Golden on
On 05/08/2010 15:56, Chris Withers wrote:
> Tim Golden wrote:
>> On 05/08/2010 15:49, Chris Withers wrote:
>>> Tim Golden wrote:
>>>>>> See http://pexpect.sf.net for a python version.
>>>>>
>>>>> ...which doesn't work on Windows.
>>>>
>>>> There is a winpexpect:
>>>>
>>>> http://pypi.python.org/pypi/winpexpect/1.3
>>>
>>> Are the two api-identical?
>>
>> From the bitbucket page:
>>
>> http://bitbucket.org/geertj/winpexpect/wiki/Home
>>
>> """
>> The API of winpexpect is identical to that of pexpect.
>> The only difference is that you must use the class "winspawn"
>> instead of "spawn"
>
> ...then why aren't they in the same fs&@ing package? ;-)

He does explain there that he hopes to have the changes merged
upstream. Presumably it hasn't happened yet...

TJG
From: Nobody on
On Thu, 05 Aug 2010 12:23:35 +0100, Chris Withers wrote:

>>> ...then the output is indeed captured. So, what is svn doing
>>> differently? How is it escaping its jail?
>>
>> maybe it does not read from stdin but directly from /dev/tty
>
> But why only the request for auth credentials?

So that you can do e.g.:

ssh remotehost command arg arg ... <infile >outfile

From: Lawrence D'Oliveiro on
In message <mailman.1621.1281012420.1673.python-list(a)python.org>, Chris
Withers wrote:

> Wolfgang Rohdewald wrote:
>
>> On Donnerstag 05 August 2010, Chris Withers wrote:
>>
>>> But why only the request for auth credentials?
>>
>> for security reasons I suppose - make sure a human enters
>> the password
>
> Well yes, but what if you actually want to script it?

Set up a public/private key pair.