Re: [gtkmm] Talk about a major rewrite

["P. Christeas" <p_christ hol gr>]

> > At this point of discussion I don't expect to look into the code. That's
> > why I don't refer to particular bugs. My questions are:
> > - do the design pitfals constitute the need for a re-design?
>
> What design pitfalls? Give us specifics.

[flame-starter..]
I personally avoid stl. It introduces too many code lines for the simplest 
things. The compiler then is pushed and more likely to mis-interpret things 
(the gcc 3.1 problems etc.). Moreover, the core code is less human-readable.

One more thing is that of iterators. By referring to an object with its 
pointer, bad coding results in havoc. One example is when I deleted an object 
and the iterator was still around. It is very diffficult to check the 
consistency of data holded by pointers. On the other hand, integer indexes of 
C lists are easier to debug. Giving a bad index can be recognized at once. 
Indexes are slower, but much safer. It will be the app's fault when such a 
thing happens, but the API must indicate the trouble.

I'm also working on a model to replace reference counting. As I mentioned 
before, I'm not planning to release anything before September. Then, you may 
look at my code and see if it makes things better. I 've been using C++ for 
many years and some other programs I've written (proprietary ones) seem 
amazingly stable. I wanted to contribute to make the toolkit (gtkmm or other) 
that stable.