Re: [boost] [thread] RFC standard proposed mutex, read-write mutex, condition

From: David Abrahams (dave_at_[hidden])
Date: 2007-08-27 15:29:19

on Mon Aug 27 2007, Howard Hinnant <howard.hinnant-AT-gmail.com> wrote:

> unique_lock<Mutex>
> ------------------
> exclusive_lock<Mutex>
> The boost name is Mutex::scoped_lock

I have no problem with unique_lock now, and I don't think
exclusive_lock adds anything of value.

> unique_lock<Mutex>::owns()
> --------------------------
> owned()

-1.  It's the mutex that's owned

> owns_lock()

-1.  Does a shared_lock "own" a lock on the mutex?

> held()

-1. It's the Mutex that's held.

> holds_lock()

-1. Cumbersome.  And a lock holds a lock?  Ugh.

> The boost name is locked()

-1 You gave the reasons why not.  Too many "locks" running around
also; it will make code hard to talk about.

I think acquired() might be best.  IMO nothing will be truly
satisfying because of the grammatical ambiguities (there's the lock
object which owns a lock on the mutex) and because the analogy to the
real-life thing called a lock is incomplete.

> try_to_lock
> -----------
> immediate
> The boost name is true

+1 for immediate in principle, though it might not be an identifier we
want to reserve in std.

>
> defer_lock
> ----------
> deferred
> Boost doesn't have this functionality

ditto.

> accept_ownership
> ----------------
> prelocked
> Boost doesn't have this functionality

prelocked is good.

-- 
Dave Abrahams
Boost Consulting
http://www.boost-consulting.com
The Astoria Seminar ==> http://www.astoriaseminar.com