Re: [Threads-devel] reader/writer locks

From: Allen Pulsifer (pulsifer3_at_[hidden])
Date: 2007-08-03 10:16:39

> I believe that all the code there works. It passes all the 
> tests on Windows for certain, and I think it passes all the 
> test on Linux too. When I move it over to trunk I'll verify 
> that everything works.

Hello Anthony,

Thanks for the reply.

Just to double check, the warning at
http://www.boost.org/doc/html/boost/read_write_mutex.html said:

"...discussions on the mailing list led to the conclusion that the current
concepts need to be rethought. In particular the schedulings Inter-Class
Scheduling Policies are deemed unnecessary. There seems to be common belief
that a fair scheme suffices."

However, in the file
http://svn.boost.org/svn/boost/tags/thread_rewrite_1/boost/boost/thread/deta
il/read_write_scheduling_policy.hpp, I see:

        enum read_write_scheduling_policy_enum
        {
            writer_priority,
            reader_priority,
            alternating_many_reads,
            alternating_single_read,
            unspecified=writer_priority
        };

But in
http://svn.boost.org/svn/boost/tags/thread_rewrite_1/boost/boost/thread/read
_write_mutex.hpp, I see the following comment:

// Shared implementation construct for explicit Scheduling Policies
// This implementation is susceptible to self-deadlock, though....
template<typename Mutex>
struct read_write_mutex_impl
{

My question is, do the Scheduling Policies work, or is there any special
usage restrictions to be aware of?

Thanks,

Allen