Re: [gtkmm] retype() and retype_return()

[Murray Cumming <murrayc usa net>]

On Thu, 2002-07-25 at 17:49, Jarek Dukat wrote:
> > At the least:
> > 1. Someone wants to understand the library. They see lots of code. They
> > run away.
> > 2. Someone greps the code. They get lots of results. They don't see at a
> > glance what's relevant. The central role of that symbol is not clear.
> 
> I don't agree at all. It is not that hard to find interesting stuff, and
> it is absolutelly not necesary to understand how sigc++'s internals work
> to use the library. It is only necessary when you want to contribute
> patches or maintain it.

And that's the problem that I am trying to solve. You haven't offered me
any other way to partially solve it, and you have not suggested any
downside to my solution. The only downside that I can imagine is that
the small number of people who want extra features need to install an
extra library.

> > Less code is less code to understand. If a library has a smaller mandate
> > then it's easier to maintain.
> 
> Well, lets strip some not-so-often used Gtkmm's stuff then. Why not?
> Little people use it, most of them won't even notice that. Move that stuff
> to Gtkmm_extra. I think it's a great idea ;)

Gtkmm has a clear mandate - wrap GTK+. If it doesn't wrap GTK+ and it
isn't useful then it isn't in gtkmm. 

> > I have given you a way out - libsigc_extras. Don't complain about my
> > maintainership of libsigc++. Either maintain libsigc++ yourself or
> > maintain libsigc_extras.
> 
> Murray, this is becoming boring. Perhaps you just don't want to maintain
> it? Then quit it. I understand you don't have time to do all this stuff
> about maintaing sigc++, gtkmm, and many other libs. But stripping lib's
> functionality is definitelly not a good way to solve these problems.

You are
1. Not offering to maintain libsigc++ yourself.
2. Not saying what is wrong with the libsigc_extras plan.

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