🚀 go-pugleaf

"Bob Gonder" <notbg@notmindspring.invalid> wrote in message 
news:r5pr915usa5ffe44er8fif4un793sol0gc@4ax.com...
> Matt Ross wrote:
>
>>
>>I modified my class TBaseObject to class TThread on both these classes, 
>>and
>>added all the requisite functions. I also modified the debug output 
>>function
>>which references back to the main class and the VCL to cope with shared
>>access. The way I intend to use them is have each thread start suspended 
>>and
>>use a timer on my main form to trigger the resuming of each thread. They 
>>go
>>off and do their thing, and suspend themselves when the work is done. The
>>next time the timer is activated it first checks the threads are all
>>suspended [i.e. they all finished doing their work last time round]. This 
>>is
>>the snag.
>
> How about this.
> Instead of using Resume, have the main thread set up 4 manual-reset
> event 2 for animals, 2 for flood.
> Each thread enters an outter loop the locks and incriments a static
> counter...psudo code might be faster to explain....
[snipped]

Yes it looks like I should be using events to determine when each thread has 
finished its current iteration.

> When main receives MSG_BarnFull, then it could either enable the
> timer, or just go ahead with the next round.
>
>>The problem I'm having is that most of the time some of these threads 
>>never
>>make it back to the suspended state. Each time I run the application it
>
> If this is like the game of life, where each node is dependant upon
> the state of it's neighbors, and you are locking access to neighbors
> who are locking access to you...could be a deadlock.

The agents do not interact with each other directly in this way. I have 
another class called Grid that handles where everyone/everything is located 
and each agent adds to a queue of things to add/move/delete for the current 
iteration. Once all the threads are suspended the Grid updates itself ready 
for the next iteration.

I will look at using Events to signal when the threads are ready and see how 
it goes.

Thanks,

-- 
Matt Ross

-=I haven't lost my mind; it's backed up on tape.=-