Re: WO bug fixes

Am 03.08.2005 um 13:01 schrieb John Pollard:

> Sounds a little complex, perhaps Apple could build this in to a  
> future WO release for me.

Actually it's rather simple: use PW which has this feature built-in  
(yeah, I know, I know, no more plugs:)

> What I did try was creating a new EC with new OSC for the long  
> query and it made simultaneous requests respond immediately, so at  
> least I can apply this technique for selected long queries - thank  
> you!
> I am applying lock/unlock to the new EC. Is there anything I need  
> to do to free up memory when the new temporary EC/OSC are created  
> and then finished with?

Sorry, I don't know. But you could - instead of creating a new OSC  
every time, create it only once and re-use that (presuming that your  
long running stuff doesn't occur that often). This is what I do.  
Other than that disposing of the EC should be enough.

Cheers, Anjo





> At 10:22 03/08/2005, Anjo Krank wrote:
>
>
>> You might want to consider to use some sort of pooling of
>> EOObjectStoreCoordinators... so you don't block *all* sessions, but
>> only the ones using the same OSC.
>>
>> Cheers, Anjo
>>
>> Am 03.08.2005 um 10:35 schrieb John Pollard:
>>
>>
>>> That's it! I have retracted my bug report. I was indeed being
>>> confused by a long EOF fetch with prefetching that blocks other
>>> activity. Multithreading is certainly working very effectively in
>>> my JC app I have established. In this case I will look at
>>> optimising the query. Thank you.
>>>
>
>