Share

PyS60

Tracker: Bugs

6 ao_sleep crash - ID: 1550365
Last Update: Comment added ( jplauril )

Hi,
this crashes pys60 1.3.1 on N70 when joystick is
pressed sufficiently fast:

import appuifw, e32
appuifw.app.body=appuifw.Listbox([u'press joystick'],
lambda: e32.ao_sleep(100))
lock=e32.Ao_lock()
appuifw.app.exit_key_handler=lock.signal
lock.wait()
appuifw.app.exit_key_handler=None

I don't know if this is the same bug that I stumbled on
when using bluetooth or not: #1540691 (Crash using BT
socket and keypad)

Jarno


Jarno Seppänen ( jams ) - 2006-09-01 08:16

6

Closed

Rejected

Jukka Laurila

e32

v1.3.1

Public


Comments ( 5 )

Date: 2007-06-12 19:00
Sender: jplaurilProject Admin


Same problem with active scheduler nesting as in
http://sourceforge.net/tracker/index.php?func=detail&aid=1550413&group_id=154155&atid=790646

Fix: Don't have long-running callbacks.



Date: 2006-10-11 12:42
Sender: vbartoni

Logged In: YES
user_id=1178755

Correction. It's not needed to press fast, crash causes 19th
press on joystick's ok at N70 with 1.3.1 pyS60. Same
behaviour at 1.3.11 sdk. N80 needs about 40-ish presses.


Date: 2006-10-05 14:15
Sender: otsov

Logged In: YES
user_id=1371303

Correction, could be a duplicate of 1481966?


Date: 2006-10-05 13:14
Sender: otsov

Logged In: YES
user_id=1371303

Correction, could be a duplicate of 1481966?


Date: 2006-10-05 06:37
Sender: otsov

Logged In: YES
user_id=1371303

This is a known "feature" with "e32.ao_sleep()". There is an
active scheduler loop ongoing and you are most probably
exiting before this loop is finished. This results in stray
signal and hence panic.

This "feature" should have been solved by using
"e32.Ao_timer" which can be cancelled if there is a pending
call. Unfortunaly "e32.Ao_timer" contains also a bug at the
time of writing (search the Forum Nokia Discussion Board for
a temporary solution). See also API doc on "e32.Ao_timer" type.

"e32.Ao_timer" will be fixed - no fix will be available for
"e32.ao_sleep" - a programmer needs to be sure that there is
not an ongoing sleep prior exiting. This should be
documented also to "e32.ao_sleep" API description.


Attached File

No Files Currently Attached

Changes ( 9 )

Field Old Value Date By
close_date - 2007-06-12 19:00 jplauril
status_id Open 2007-06-12 19:00 jplauril
resolution_id None 2007-06-12 19:00 jplauril
category_id None 2007-06-12 19:00 jplauril
assigned_to vbartoni 2007-06-12 19:00 jplauril
priority 5 2007-02-19 17:42 jplauril
assigned_to nobody 2006-10-19 07:39 vbartoni
priority 6 2006-10-05 13:14 otsov
priority 5 2006-10-05 13:12 jplauril