Re: [sigc] ANNOUNCE: libsigc++ 1.9.12 (aka libsigc++ 2.0 beta 1)

[Murray Cumming <murrayc usa net>]

On Thu, 2003-11-06 at 22:20, Martin Schulze wrote:
> Hi!
> 
> Could you give me some feedback about the latest announcement?
> Any typos, bad language, waffle, wrong/missing information, etc.?

see below.

> Unless somebody else does the job, I will finally copy&paste vast
> parts of this notice into the libsigc++-2.0 announcement in a few
> weeks

I hope you will wait for gtkmm 2.4 to be declared API stable.
  
> > libsigc++ implements a typesafe callback system for standard C++. It
> > allows you to define signals and to connect those signals to any
> > callback function, either global or a member function, regardless of
> > whether it is static or virtual.
> > 
> > libsigc++ is used by gtkmm to wrap the GTK+ signal system. It does not
> > depend on GTK or gtkmm.
> > 
> > libsigc++ 2.0 uses modern C++ mechanisms to achieve a highly flexible,

I would remove the word "mechanisms",

> > yet typesafe callback system. It supports all features of libsigc++ 

the features

> > 1.2
> > and improves upon it with a simpler and more powerful API. The old API
> > is preserved by means of a deprecated compatibility module.
> > 
> > In particular the API improvements from libsigc++ 1.2 to 2.0 include:
> > * Provides unnumbered signal and slot templates.

I would give a quick example here.

> > * Supports arbitrary functor types.

Again an example would show what this means. If you want to keep this
list short then you could put examples on a libsigc++ web page and link
to them. Tell me if you have any problems updating the libsigc++ web
page with libsigc++-1.2/docs (or something similar) 

> > * Supports implicit argument type conversions on signal emission.
> > * Adds support for overloaded functions.

Again, I'm not really sure what this means myself.

> > * Borrows powerful accumulator interface from boost::signal.

Again, though I personally have more idea this time.

> > * Simplifies use of adaptors: argument types need not be specified any
> >   more (except for sigc::retype_return()).

I would say "For instance, SigC::bind(), because people don't know what
an adaptor is, but that is the one that they use.

> > * More powerful adaptors: position of arguments to hide or bind can
> >   be chosen freely.
> > * Adds two useful adaptors: sigc::compose(), sigc::exception_catch().
> > * Adds all-in-one operator sigc::group() with support for lambdas
> >   featuring a complete set of lambda selectors and lambda operators.
> > * Highly optimizable:

Do you mean "highly optimized" or do you mean that there is something
that the application writer can do to make it faster?
]
>  performs only one indirect function call during
> >   slot invokation independant

s/independent/regardless/

>  of the number of adaptors used.
> > * Conforms to the STL naming scheme: only uses small letters.

I would mention the compatibility layer again here.
 
> > *** Prerequisits

Prerequisites, but "Supported Compilers" is probably better anyway, and
I would name exactly which compilers are known to work, and which we
think should be tested.

> > Compatible compilers must supports the following recent c++ 
> > techniques:
> > * Partial template (function) specialization.
> > * Arithmetics in template arguments.
> > * Explicit template (member) function instantiation.
> > * Treat void return as normal return.
> > 
> > 
> > *** Changes:
> > 
> > 1.9.12:
> > 
> > This is the first beta release of libsigc++ 2.0. No more API additions
> > should be made. Future work concentrates on supporting more compilers,
> > bug hunting and finishing the documentation. Changes in this release:
> > * Added adaptor retype(). With this final API addition all adaptors
> >   are in place that are available in libsigc++-1.2.
> > * Added negation lambda operator. Use STL names for lambda actions.
> > * Remove formerly disabled support for gcc extension typeof().
> > * Added project files for MS Visual Studio .Net 2003. (Roel Vanhout)
> > * Make libsigc++2 compile with .Net 2003. (Roel Vanhout,Martin 
> > Schulze)
> > * Build shared version of libsigc++2 by default. (Cedric Gustin)
> > * Add support for win32 platform. (Cedric Gustin)
> > * Install .m4 files. (requested by Ron Steinke)
> > * Cleaned up functors .m4 files for ptr_fun() and mem_fun().
> > * Restructured and completed documentation of the core library parts.
> > 
> > 
> > *** Download:
> > 
> > libsigc++ is available for download here:
> > http://libsigc.sourceforge.net/
> > 
> > A reference documentation for libsigc++ 2.0 can be found here:
> > http://libsigc.sourceforge.net/libsigc2/docs/
> > 
> > 
> > *** Known issues:
> > 
> > * This release has only been tested with gcc-3.2. It should also
> >   compile with:
> >   - gcc-3.3, gcc-3.3.x
> >   - cygwin (gcc >= 3.2)
> >   - mingw32
> >   - Microsoft Visual Studio .Net 2003
> >   - Sun Forte C++ compiler
> >   Please report any troubles you encounter with these compilers!
> >   You are also invited to try a different compiler.
> > 
> > * When the compatibility module is used with "using namespace SigC",
> >   gcc-3.2 reports conflicts of SigC::slot() with struct sigc::slot.
> >   A hackish work-around has been disabled to support the .Net 
> > compiler.
> > 
> > * Adaptors are not documented, yet.

-- 
Murray Cumming
murrayc usa net
www.murrayc.com