Re: [sc-users] Document keyUp

[James McCartney <asynth@xxxxxx>]

On Dec 25, 2004, at 2:30 PM, Jan Trützschler wrote:

> as far as i can see this is called before the lock:
>
> 			PyrSymbol *method = getsym("keyUp");
>
> 			if (pobj) {
> 				pthread_mutex_lock (&gLangMutex);

yes, this is wrong. If another thread like a timer were to take the 
lock and define a new symbol, then the getsym could crash.
getsym must be done after the lock is taken.

> > compiledOK must be tested again inside the lock.
>
> like this?
>
> 			PyrSymbol *method = getsym("keyUp");
>
> 			if (pobj) {
> 				pthread_mutex_lock (&gLangMutex);
> 				if(compiledOK){
> 					....
>
> so is it neccessary to have check two times for  compiledOK ?
yes, because in the meantime the value may have changed unless you own 
the lock.