[boost] [signals2] post-review rfc before moving to trunk

Subject: [boost] [signals2] post-review rfc before moving to trunk
From: Frank Mori Hess (frank.hess_at_[hidden])
Date: 2009-02-11 15:34:00

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi all,

I've made some changes to Boost.Signals2 in response to some of the 
suggestions made during the review.  Unless there is are show-stopper 
objections, I'd like to move Signals2 from the sandbox to the trunk in a week 
or two.  For context, Stjepan Rajko's review summary is here:

http://lists.boost.org/boost-announce/2008/11/0211.php

The current Signals2 documentation is online here:

http://www.comedi.org/projects/signals2/libs/signals2/doc/html/

Here is a summary of most of the changes made since the review:  
*Added a signals_vs_signals2_benchmark.cpp to the signals2/test subdirectory, 
and made some optimizations to signal invocation speed.  
*Made sure all macros were prefixed with BOOST_SIGNALS2_.
*Added tests for boost::signals2::mutex and dummy_mutex.
*Brought back boost::signals2::trackable class to ease porting of 
single-threaded code from Boost.Signals.
*Deprecated deconstruct_ptr/postconstructible/predestructible in favor
of deconstruct, a factory function similar to make_shared, which supports 
postconstructors and predestructors via ADL, similar to how the 
sp_accept_owner stuff works in boost smart_ptr trunk.
*Expanded FAQ entry on differences between Signals and Signals2 by adding some 
notes on how to deal with the changes while porting.  Changed the faq entry 
into a section.
*Added documentation section on thread-safety which describes how the library 
uses its mutexes internally.
*Added boost::signals2::signal_type metafunction which uses Boost.Parameter to 
allow named template type parameters for signals.
*Added sections to the tutorial on postconstructors, extended slots, and 
signal_type.
*Added examples section to docs, with compilable example programs extracted 
from tutorial, plus a few new ones.
*Added a couple paragraphs to the design rationale section about 
optional_last_value and the signals2 mutex classes.
*Added a small description to the optional_last_value reference page.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFJkzY45vihyNWuA4URAivcAKCSGti3YkLGnzjmTr3EfAeN3FfiUwCg13S4
Jp54J65svUx6MVKNOsW31nU=
=U39m
-----END PGP SIGNATURE-----