Re: [AD] Mini-synchronization API proposal for 4.1.x

[Thomas Fjellstrom <tfjellstrom@xxxxxxxxxx>]

On March 24, 2002 07:36 pm, you wrote:
> On 2002-03-24, Thomas Fjellstrom <tfjellstrom@xxxxxxxxxx> wrote:
> > void handler(void)
> > {
> >   al_mutex_lock(my_mutex);
> >   my_value = 0;
> >   al_mutex_unlock(my_mutex);
> > }
> >
> > int main()
> > {
> >   al_mutex_lock(my_mutex);
> >   /* handler called here */
> >   if(my_value) do something;
> >   al_mutex_unlock(my_mutex);
> >
> >   ...
> > }
> >
> >
> > if mutexes do nothing then wouldn't that be a bit odd? how would that be
> > handled? (can't wait in the handler...)
>
> Ouch, this is Very Bad.  On one hand if AL_MUTEXes do something in the
> pause-main-thread-for-callbacks case, then your program will deadlock.
> OTOH if they do nothing, then they don't do what you want (some values
> remaining constant between lock/unlock).

My thoughts exactly. I orriginally was going to say something about WE HAVE 
TO WAIT... But then I remembered you cant in a call back... Will we have to 
pause any and all timers when a mutex is locked? Of course thats not really 
acceptable. what can be done?

-- 
Thomas Fjellstrom
tfjellstrom@xxxxxxxxxx
http://strangesoft.net