Re: [pygtk] Turning the PyGTK+ brand into something more than it currently is

[Simon Feltman <s feltman gmail com>]

Hi Giuseppe,

On Sat, Nov 17, 2012 at 1:26 PM, Giuseppe Penone <giuspen gmail com> wrote:

I read that there's no need of reference because gtk3 reference is enough, well I'm a gtk3 and gtkmm3 developer over a pygtk developer (my most used app is www.giuspen.com/cherrytree) but still I miss pygtk reference when doing something in PyGobject.

I'm curious where you read this as I think the idea should be squashed. Or perhaps the statement is being misinterpreted? The current gtk3 C docs/annotations should be "enough", but using them to generate python specific docs still needs to happen (more below).

I read somewhere that gtkmm generates the reference automatically from gtk, why can't pygobject do the same?

It can and should. I would even take it further and say we should be pulling translated (as in programming language translated) versions of the docs into the python bindings as __doc__ attributes (perhaps lazily to keep good import performance). Python override docs would be overlayed on top of these. With this idea, documentation would be available through all the standard mechanisms python programmers are used to (interactively, calltips, doc generators, etc..). There has been some advancement on this front recently but much more work is needed.

Gtk/GObject function signatures are now available in python:

https://bugzilla.gnome.org/show_bug.cgi?id=681967

Ticket regarding dynamic doc generation:

https://bugzilla.gnome.org/show_bug.cgi?id=625494

-Simon