Re: dropping sigc++

On Mon, Jan 17, 2005 at 11:09:27PM +0000, Mark Thomas wrote:
> On Mon, 17 Jan 2005, Phil wrote:
> 
> > James Harr wrote:
> > > How do the need for c++, sigc++, and autoconf relate at all?
> >
> > You need a C++ compiler to build libsigc++ and those parts of Y which
> > use it.  Autoconf shouldn't need C++, but Y's configure script will
> > check for a C++ compiler.
> 
> That's mostly due to the current state of Y.  Y's only client library is
> the C++ one, so to get a functional version you need C++.
> 
> In the Distant Future, Y will have different client libraries, so
> configure will probably let you say --without-c++-library if you don't
> have/want C++.  Maybe it'll even automagically detect its presence.

I arranged things with a view to simply splitting it off into a
different tree, once the protocol between the two settles down and we don't
need atomic commits between them any more.

We'll need to get the basic server architecture complete before this
can happen (I have half a dozen protocol-critical points that I know
of). Until then C++ is as good a language as any for development
purposes.

> Before people start jumping to implement a c/perl/python/... client
> library, can I just suggest we keep it simple for now and just stick to
> the c++ one.  I've rewritten the C++ client lib once, and Andrew did it
> another time.

I didn't finish the job, btw, I just fixed the API so that it wouldn't
have to change again in the forseeable future. The actual
implementation will probably have to be done over; my only goal was to
avoid having to redesign the API *after* all the features had been
bolted onto it.

-- 
  .''`.  ** Debian GNU/Linux ** | Andrew Suffield
 : :' :  http://www.debian.org/ |
 `. `'                          |
   `-             -><-          |

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQFB7NWelpK98RSteX8RAqkQAJ9cKqg0nYj0Tj3I7qKqhAWpxdx33ACeMnqA
yvttABeLVAqObOQ9K5gdu4E=
=SOyF
-----END PGP SIGNATURE-----