Re: Gtkmm-PLplot

[Maciej Piotr Sauermann <m p sauermann zoho com>]

Hi,

I've just had a glance at some parts of the code and indeed
 
> For example, due to my C background, I rely quite a lot on pointers which I am
> sure is frowned upon, but I am not really sure how to solve this (Glib::RefPtr
> perhaps?).

the above is true.
I'd suggest using std::unique_ptr (and std::vector<std::unique_ptr>), writing
your own destructor which only goes through a vector and deletes pointers held
inside it seems pointless to me. But the first thing I would suggest is to
reconsider whether the instances really have to be pointers (the code I looked
at indeed needs them because of inheritance). I also saw calls to `g_realloc`
and `g_strfreev` - I'd also suggest wrapping them in a unique_ptr.

Canvas::get_plot's parameter's type is incorrect - it should be
std::vector::size_type. Moreover, I saw some code commented-out. Plus there are
some long methods. Finally, the code does not seem to be line-wrapped - I
mention this since I keep mine at 80 columns limit and I find it useful. 

Sorry if my remarks are too general.

Kind regards,

Maciej