IDLE processes don't close

[mensanator]

BPO	8093
Nosy	@terryjreedy, @serwy, @briancurtin, @skrah
Superseder	bpo-12540: "Restart Shell" command leaves pythonw.exe processes running

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2011-12-22.06:18:22.661>
created_at = <Date 2010-03-08.23:21:30.749>
labels = ['expert-IDLE', 'type-bug', 'OS-windows']
title = "IDLE processes don't close"
updated_at = <Date 2011-12-22.06:18:22.661>
user = 'https://bugs.python.org/mensanator'

bugs.python.org fields:

activity = <Date 2011-12-22.06:18:22.661>
actor = 'ned.deily'
assignee = 'none'
closed = True
closed_date = <Date 2011-12-22.06:18:22.661>
closer = 'ned.deily'
components = ['IDLE', 'Windows']
creation = <Date 2010-03-08.23:21:30.749>
creator = 'mensanator'
dependencies = []
files = []
hgrepos = []
issue_num = 8093
keywords = []
message_count = 5.0
messages = ['100679', '100735', '102818', '138814', '150044']
nosy_count = 6.0
nosy_names = ['terry.reedy', 'roger.serwy', 'brian.curtin', 'skrah', 'mensanator', 'Simon']
pr_nums = []
priority = 'normal'
resolution = 'duplicate'
stage = 'resolved'
status = 'closed'
superseder = '12540'
type = 'behavior'
url = 'https://bugs.python.org/issue8093'
versions = ['Python 3.1']

[mensanator]

So I started 3.1.2...

Python 3.1.2rc1 (r312rc1:78742, Mar 7 2010, 07:49:40) [MSC v.1500 32 bit (Intel)] on win32
Type "copyright", "credits" or "license()" for more information.

>>

And monitored Windows Task Manager (XP):

Image Name User Name CPU Mem Usage
---------------------------------------------
pythonw.exe paul_stadfeld 00 11,120 K
pythonw.exe paul_stadfeld 00 16,816 K

I assume this is normal.

I deliberately do something stupid:

>> a = 222222222222222222222222222
>> b = 5555555555555555555555555555
>> a**b

Ok, that will run for a while, long enough for me to watch the Task Manager:

Image Name User Name CPU Mem Usage
---------------------------------------------
pythonw.exe paul_stadfeld ~55 ~30,548 K (constantly changing)
pythonw.exe paul_stadfeld 00 16,828 K

So I do [Restart Shell] from IDLE's [Shell] menu:

>> ================================ RESTART ================================

And try it again.

>> a = 222222222222222222222222222
>> b = 5555555555555555555555555555
>> a**b

But now it's different:

Image Name User Name CPU Mem Usage
---------------------------------------------
pythonw.exe paul_stadfeld 00 13,716
pythonw.exe paul_stadfeld ~50 ~30,548 K (constantly changing)
pythonw.exe paul_stadfeld ~48 ~45,548 K (constantly changing)
pythonw.exe paul_stadfeld 00 16,892 K

Looks like the previous process was never stopped.

Trying to Restart Shell again and running same creates a 5th instance
of pythonw.exe.

>> ================================ RESTART ================================
>> a = 222222222222222222222222222
>> b = 5555555555555555555555555555
>> a**b

Image Name User Name CPU Mem Usage
---------------------------------------------
pythonw.exe paul_stadfeld 00 13,716
pythonw.exe paul_stadfeld ~50 ~54,548 K (constantly changing)
pythonw.exe paul_stadfeld ~25 ~42,548 K (constantly changing)
pythonw.exe paul_stadfeld ~25 ~54,548 K (constantly changing)
pythonw.exe paul_stadfeld 00 16,892 K

Only this time I got an error trying to re-start.

IDLE's subprocess didn't make connection. Either IDLE can't start a 
subprocess or personal firewall software is blocking the connection

Now I'm stuck, so I close the "Python Shell" window.

But closing the window only stopped one of the pythonw.exe instances.
The three active processes are still running full tilt (the one quiescent
process remains quiescent.)

This appears to be a process leak in addition to it being a memory leak.

Also, had one of the processes that won't stop opened a script from a USB
port, then Windows would refuse to eject the USB drive as it correctly
thinks some process is still using it. Of course, since the application
is closed, there's not suppoed to be any such process, so it's
not obvious where the problem is unless you look at the process table
in Windows Task Manager and manually halt all the renegade pythonw.exe
processes that are still running. At which point Windows will allow the
USB to be ejected.

[terryjreedy]

I reproduced this with 3.1.1 on xp. It took a while to get the shell menu to restart as the calculation process is hogging the cpu 99%.
KeyboardInterrupt (^C) would not stop the runaway process. There may have been other issues about this.

Two processes is normal - one for the shell and one for calculations. The 'normal' behavior of Restart Shell (^F6) is to start a third process. Watching in task manager, the abandoned calculation process dies in 3-4 seconds. A stuck process does not end on its own though. Bad bug.

[skrah]

For the record:

In 2.7-alpha you do not even get to press [Restart Shell], since IDLE is
not responding during the calculation.

[simon]

Not sure if this is related, but for me none of the F5 processes (command line "pythonw.exe -c __import__('idlelib.run').run.main(True) <num>") ever exit on either 2.7.2 or 3.2.1rc1.

[serwy]

This issue should be closed since bpo-12540 fixes it.

I verified that the bug exists with 3.1.4 on Vista. It does not exist with 3.2.2 on Vista.