[Edu-sig] Help Testing IDLE patch

[John Zelle]

I think my first reply to this went only to David. For others who might 
be interested, here's the scoop.

The problem folks have been having is that IDLE sometimes loses 
communication with its subprocess. This can cause IDLE to freeze or be 
unable to bring the shell window up because the socket is in use. It's 
hard to recreate exactly what does this, as it seems to be a sporadic 
(race condition?) problem.

There is a situation where I can reliably cause IDLE to hang. Simply 
create a program that has an input statement such as: x = input("Testing ")
Run the program with <F5> and when the prompt "Testing " comes up, close 
the shell window to go back to the program window. Usually, this will 
hang IDLE. If it doesn't happen the first time, do <F5> - close again. 
Versions of IDLE since 1.0 will freeze after 1 or two rounds of this.

My "patch" just does the equivalent of a keyboard interrupt before 
closing the shell, since I've found that doing a Ctrl-c before closing 
seems to avoid the problem.

BTW, my fix is for IDLE 1.04 shipped with Python 2.3.3. I don't think it 
works for IDLE 1.1 in Python 2.4 :-(

--John


David Handy wrote:

>On Wed, Dec 15, 2004 at 12:33:49PM -0600, John Zelle wrote:
>
>  
>
>>I was more concerned about IDLE 1.04's instablility under Windows. I 
>>have a very simple test that causes IDLE to hang, and it does not appear 
>>to be fixed in IDLE 1.1.
>>    
>>
>
>I'd like to try out your fix on my Windows 98 machine, but I'm not sure if
>I'm seeing the same problem as you.
>
>Please post your simple test that causes IDLE to hang.
>
>Thanks,
>
>David H.
>_______________________________________________
>Edu-sig mailing list
>Edu-sig at python.org
>http://mail.python.org/mailman/listinfo/edu-sig
>
>
>  
>