[BUG][FIX] interrupt driven EventSensor ( could somebody provide detailed review, please? )

Andreas Raab andreas.raab at gmx.de
Thu Jul 31 18:31:32 UTC 2003


I am not questioning the implementation I am questioning the conceptual
soundness of the fix. Should we hack EventSensor to accomodate for a bug in
SystemWindow's reframing code? No.

The fix adds an arbitrary delay to a very low-level place in the system of
which the implications are not at all clear to me. It _may_ work but it may
equally well have various nasty side effects. In addition, having two ways
of stimulating the input semaphore from Squeak (one with delay and one
without it) adds to the conceptual complexity of event sensor. Someone
looking at this will ask himself why the hell are there two variants, one
used from this set of methods and the other one from the other set of
methods. The answer: Well, of course, we need to accomodate for that
particular abuse in SystemWindow. Hello?! ;-)

Cheers,
  - Andreas

> -----Original Message-----
> From: squeak-dev-bounces at lists.squeakfoundation.org 
> [mailto:squeak-dev-bounces at lists.squeakfoundation.org] On 
> Behalf Of Brent Vukmer
> Sent: Thursday, July 31, 2003 6:48 PM
> To: The general-purpose Squeak developers list
> Subject: [BUG][FIX] interrupt driven EventSensor ( could 
> somebody provide detailed review, please? )
> 
> 
> Andreas( and all EventSensor gurus ) -
> 
> Could someone do a detailed review of John's submission with 
> comments on
> the soundness of his design and the 
> clarity/simplicity/robustness of his
> implementation?  Will John's changes make EventSensor less reliable?
> 
> Does this submission improve EventSensor code quality, or not?  If it
> does, why NOT put it in?  
> 
> 
> 
> 



More information about the Squeak-dev mailing list