[Dev] Chandler pegs CPU

[Katie Capps Parlante]

Hey folks,

Its probably time to take this conversation to the bug (Bug 4356), as 
there is a larger history there, and we don't need to distract everyone 
with the details. If you are interested in the bug conversation cc 
yourself to the bug.

It was useful to have a bit of this conversation on the list -- 
definitely a judement call about when to bring a bug conversation to the 
list or take a conversation about a bug to bugzilla. One metric is 
whether or not the conversation is converging and mainly about the bug 
-- then bugzilla is probably better.

Cheers,
Katie

Donn Denman wrote:
> I don't think it's related to the stamping bar because of this test -  
> create a new collection and select it, you get no detail view or  
> stamping bar.  The CPU usage is fine until you edit the name of the  
> collection in the sidebar, then it jumps.
> 
> On Oct 26, 2005, at 4:39 PM, John Anderson wrote:
> 
>> Actually, I now think the problem is related to the stamping bar.  
>> Whenever the last button is completely off the screen the problem  
>> goes away. And, if I remember correctly, David's problem before  
>> happened to be caused by the search box in the toolbar, which was  the 
>> last item in the toolbar.
>>
>> John
>>
>> John Anderson wrote:
>>
>>
>>> If had this recollection that the problem was caused by  overlapping 
>>> windows and david's offscreen draw trick. Sure enough,  if the detail 
>>> view's controls are no longer visible, by moving the  main splitter 
>>> right, the problem goes away.
>>>
>>> John
>>>
>>> John Anderson wrote:
>>>
>>>
>>>>
>>>>
>>>> David Surovell wrote:
>>>>
>>>>
>>>>> Folks,
>>>>>
>>>>> I've noticed the CPU problem, but it went away for me after a  
>>>>> check-in late Tuesday afternoon. The problem has been caused  
>>>>> before by the UpdateUI stuff, also by my wxMSW composited window  
>>>>> shim.
>>>>>
>>>>
>>>>
>>>>
>>>> UpdateUI shouldn't cause infinite loops updating windows -- but I  
>>>> double checked that it isn't causing the problem. Is there an  easy 
>>>> way to turn off wxMSW composited window shim to verify that  it 
>>>> isn't part of the infinite update problem?
>>>>
>>>>
>>>>>
>>>>> My (strong) belief is that this problem is due to Chandler code  
>>>>> changes, and not wx. The wx code hasn't changed at all for a  week, 
>>>>> and no significant low-level changes in wxMSW for some  days before.
>>>>>
>>>>> Regards,
>>>>> David S.
>>>>>
>>>>> =====================
>>>>>
>>>>> Katie Capps Parlante wrote:
>>>>>
>>>>>
>>>>>> Is there any way to automate a test for this?
>>>>>>
>>>>>> It would be great if we could be flagged when a new checkin  
>>>>>> reintroduces this kind of problem.
>>>>>>
>>>>>> Cheers,
>>>>>> Katie
>>>>>>
>>>>>> John Anderson wrote:
>>>>>>
>>>>>>
>>>>>>> Hi David:
>>>>>>>
>>>>>>> Morgen noticed that Chandler pegs the CPU on windows. Other  
>>>>>>> people have
>>>>>>> noticed it's started getting sluggish a few days ago. It looks  
>>>>>>> like were
>>>>>>> stuck in an infinite loop updating the screen. Seems to not  
>>>>>>> happen on
>>>>>>> win2k, which sounds like a bug you had awhile back with your  
>>>>>>> offscreen
>>>>>>> buffer fix. Any ideas?
>>>>>>>
>>>>>>> John
>>>>>>>
>>>>>>>
>>>>>>> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>>>>>>>
>>>>>>> Open Source Applications Foundation "Dev" mailing list
>>>>>>> http://lists.osafoundation.org/mailman/listinfo/dev
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>>>>
>>>> Open Source Applications Foundation "Dev" mailing list
>>>> http://lists.osafoundation.org/mailman/listinfo/dev
>>>>
>>>
>>>
>>> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>>>
>>> Open Source Applications Foundation "Dev" mailing list
>>> http://lists.osafoundation.org/mailman/listinfo/dev
>>>
>>
>> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>>
>> Open Source Applications Foundation "Dev" mailing list
>> http://lists.osafoundation.org/mailman/listinfo/dev
>>
> 
> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
> 
> Open Source Applications Foundation "Dev" mailing list
> http://lists.osafoundation.org/mailman/listinfo/dev