Re: libsigc++2 in gtkmm-2.4? (was: Re: [gtkmm] ANNOUNCE: glibmm 2 .3.0)

[Murray Cumming <murrayc usa net>]

On Sat, 2003-10-18 at 14:27, Chris Vine wrote:
> On Wednesday 15 October 2003 8:18 am, Murray Cumming Comneon com wrote:
> > From: martin-ml hippogriff de [mailto:martin-ml hippogriff de]
> > > Well first of all sorry for answering this late.
> > > As to changing back to "SigC::Slot" in libsigc++2 I must say
> > > that I would really prefer a compatibility header:
> >
> > Yes, that would be OK if it can be done. We would have
> > #ifndef LIBSIGC_DISABLE_DEPRECATED
> > Around the API.
> 
> Changing the API by going from initial upper case letters for namespace names 
> to lower case letters seems to me to be a self-indulgence which should be 
> resisted

I agree, but it will not be a problem with the proposed compatibility
headers, if they work.

>  (but I have no expectation that my view will have any effect, given 
> previous exchanges of view on other API breakage issues).

The 2.0/2.2 -> 2.4 transition will be a lot easier. I'll be moving the
deprecated widgets into gtkmm_deprecated 2.4 and I'll invite you to
help. Notice that I'm doing the work of creating it for you, but you
still have to do some work if it's important enough to you.

> However, if initial capitals for namespaces causes aesthetic problems for the 
> maintainer why not just use a namespace alias, so allowing either to be used 
> without having to mess around with configure time definitions:
> 
> namespace sigc = SigC;
> 
> in the relevant header files.
>
> Of course this will not work if it is also proposed to change all type names 
> forming part of the public interface to initial lower capitals, but that 
> could be dealt with by using the equivalent typedefs and thus avoid breaking 
> code.

The proposed deprecation system seems to cover this OK. Let's see if it
works.

> But then, why not go the full distance and do the same for gtkmm, so really 
> impressing users about the aesthetic purity of the type names employed?

Let's deal with problems that actually exist, shall we?

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