Semaphore>>wait:

>>Dear ObjectArtists,
>>According to the method comment of #wait:ret: you are not allowed to
>>specify a timeout other then INFINITE or 0.
>>Does this (still) apply (to Win2K or XP)? And if yes, why?
>>Guessing from the constants the actual work is performed by a
>>WaitForSingleObject API call and that one does support millisecond
>>timeouts.
>>Or am I missing something?
>>
>
>
> The timeout parameter is reserved for future use as stated in the method
> comment, sorry.
>
> Regards
>
> Blair
>

> >>Dear ObjectArtists,
> >>According to the method comment of #wait:ret: you are not allowed to
> >>specify a timeout other then INFINITE or 0.
> >>Does this (still) apply (to Win2K or XP)? And if yes, why?
> >>Guessing from the constants the actual work is performed by a
> >>WaitForSingleObject API call and that one does support millisecond
> >>timeouts.
> >>Or am I missing something?
> >>
> >
> >
> > The timeout parameter is reserved for future use as stated in the method
> > comment, sorry.
> >
> > Regards
> >
> > Blair
> >
>
> Is it ok to subclass Semaphore (with instance variable)?

> Pieter,
>
>> No, it is not safe to subclass Semaphore. The VM sometimes checks
>> specifically for the class. However you can always use has-a instead of
>> is-a, and wrap up Semaphore inside another class as is done by Mutex.
>
>
> I'm starting to suspect that you are working on a (hopefully) solve
> problem, if only in part.  Take a look at the timed evaluators on my web
> site.  Feel free to use the code as-is or change it as necessary, but
> please let me know if you find any bugs.
>
> Have a good one,
>
> Bill
>