mac VM 3.2.5b7

John M McIntosh johnmci at smalltalkconsulting.com
Thu Mar 7 05:50:40 UTC 2002


>John,
>
>>  Note the EventSensor will fire every 500 ms. That should be changed
>>  to wait many seconds/forever?. Say we propose a VM attribute that
>>  indicates true event signalling then we could alter EventSensor...
>
>I firmly believe this code to be obsolete. It used to work around a
>problem which is fixed by now. So you should be able to just remove it
>and wait forever.
>
>Cheers,
>   - Andreas

Unfortunately we can't just remove it because older mac VM, mmm even 
the current Mac VM does not trigger the event sensor semaphore. This 
was based on a note from you way back... (see below) Hence  my 
comment about have some way to query the VM to see if it has true 
async event support, then adjust things.


At 11:14 PM -0700 8/17/00, Raab, Andreas wrote:
>As I told you before, do whatever you like. Signaling the input semaphore
>will only make the system slightly slower since you will be signaling when
>being called by the get next event primitive and signaling the semaphore at
>this point is useless but harmless.
>
>>  But in the early code I thought we waited on the semaphore
>>  using a timer, so it didn't matter if we didn't get the signal,
>>  and if we did then we could complete faster.
>
>You will only complete faster if you signal asynchronously. Otherwise you
>won't. And since you can only signal when ioProcessEvents() is called and
>since currently ioProcessEvents() is called either from
>#primitiveGetNextEvent or from #checkForInterrupts signaling won't help
>since you will either be reading the events already (in
>#primitiveGetNextEvent) or you'll check only every half a second. So there
>is simply no point in signaling the input semaphore unless you are in fact
>asynchronous (which you are not).
>
>>  However this means more polling activity for systems that
>>  are truly Async.
>
>By nowadays standards this argument is meaningless. All of todays machines,
>including PDAs, will not be affected by 50 additional cheap primitive calls
>per second.

-- 
--
===========================================================================
John M. McIntosh <johnmci at smalltalkconsulting.com> 1-800-477-2659
Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
===========================================================================



More information about the Squeak-dev mailing list