Future programming language.
jralls at ceridwen.us
Tue Mar 4 12:16:25 EST 2014
On Mar 4, 2014, at 2:42 AM, Robert Fewell <14ubobit at gmail.com> wrote:
> Could some one clarify what future programming components, not sure thats
> the correct word, will be used. I was thinking about looking at the
> register tree view agian / gtk3 but I read that you are moving away from
> gtk so that seems a bit pointless.
> May be I am getting confused with backend and gui ?
At the moment, you're confusing the backend, core, and GUI. We're not planning to do much with the GUI this dev cycle, including not moving to Gtk3. But the Gtk team has been having trouble supporting M$Windows, never mind mobile platforms, so we've been discussing changing GUI backends in the *next* dev cycle and unless the situation changes with Gtk on MSWindows they're probably not a viable candidate.
That said, the existing register code uses an obsolete library, libgnomecanvas, for drawing its custom controls. Your conversion of the register to GtkTreeView gets rid of that dependency, so it may still have a point. On the other hand it might be simpler to just replace the few libgnomecanvas calls with either GdkDrawable or direct Cairo calls.
Either way, the GUI also needs better MVC separation; I know that's hard with GtkTreeView, which encourages one to mix the model implementation code in the same modules that implement GtkTreeModel, but having all of that "business logic" mixed with the GUI makes it hard to maintain never mind to implement a different GUI. That's a large part of what's made your work with GtkTreeView so difficult.
More information about the gnucash-devel