Subj : Re: Futexes are wrong! (Are they?)
To   : comp.programming.threads
From : David Schwartz
Date : Tue Jun 07 2005 10:49 am


"Maciej Sobczak" <no.spam@no.spam.com> wrote in message 
news:d84eig$8qt$1@sunnews.cern.ch...

> Interestingly, the scoped lock pattern (realized with the RAII idiom in 
> C++ and native in languages that support the "synchronized" keyword or 
> alike) promotes exactly this way of working:

    Personally, I think the scoped lock pattern is an abomination. A lock 
should be held for exactly as long as needed. If this perfectly corresponds 
to a language scope, then great. If not, to hold the lock longer than needed 
because it fits a pattern, well, down that path is recursive locks and all 
the other such abominations.

> {
>     scoped_lock lock(mtx);
>
>     // do something ...
>     condvar.signal();
> }
>
> Above, condvar is signalled before the mutex gets unlocked.
> Doing otherwise would require something like this:
>
> {
>     scoped_lock lock(mtx);
>
>     // do something ...
> }
> condvar.signal();
>
> which can become ugly in the outermost block (like when the block is the 
> whole funcion body).
>
> What about this?

    You should hold a lock for exactly as long as you need it. If this means 
a scope just for the lock, then so be it. What about if you have ten things 
to do that don't require the lock? Should you do them with the lock held 
just because this makes the block cleaner?

    DS

.