[Glade-devel] Colaboration about changes in the glade format

[Tristan Van Berkom <tristan van berkom gmail com> (Tristan Van Berkom)]

On Fri, 04 Feb 2005 19:15:48 +0100, Lorenzo Gil Sanchez <lgs sicem biz> wrote:
> Hi,
>
> as some of you have already talk, the glade file format need to be
> changed to support GObjects aswell as GtkWidgets. This way we can store
> UIManager based Menubars and Toolbars, TreeView Columns definitions,
> SizeGroups and many other things.

Hmmm,
    I think that the portions of bug 161903 that deal with non-widgets
is sufficient for dealing with non-widget GObjects without changing
the glade file format, I guess it might be a drawback that an object
is still "called a widget" but I dont see that as an issue.

Note that I might be mistaken in regards with toplevel objects that are
not of type GtkWindow, there might be some assertions that need
to be removed.

> I'm the main developer of Gazpacho, an application very similar to
> Glade, but written in python and I have already done some work towards
> the UIManager support. I have a proposal for the changes I think are
> needed for glade and libglade. You can read it here:
>
> http://gruppy.sicem.biz/wiki/ProposalUIManagerLibGlade
>
> As Glade-3 has the same problem here as Gazpacho, I would like to work
> together to get a specification which is useful for both programs.

I have to admit that I feel a bit queezy about GtkUIManager support in 
glade/libglade, afaics its a dual implementation of libglade in gtk+

/me consults the api docs for UIManager...

    Still I am failing to see what GtkUIManager has to offer that can not 
be treated with the traditional glade file format + non-widget support
(clarifications on this are invited); and I dont see what would be so hard
about managing GtkActions/GtkActionGroups either (but I haven't studied
this in detail).

My fear of GtkUIManager is pretty much this:
    If we invent a standard that incorperates a dual implementation of UI's
generated by xml files we are making things complicated for no good reason;
and if we go through with it; it will just "be" and when the new generations of
glade/gazpacho reach maturity; people will just have to deal with it. There
has also been talk of incorperating glade file parsing into gtk+, if
this happens,
I would hope that it would either be an extention of the GtkUIManager
or it would
result in the deprication of the GtkUIManager.

My current take on the UI manager is pretty much, lets have a libglade
framework for generating UIs that will last a few years or lets have a 
GtkUIManager framework for generating UIs for the next few years, but 
lets not impose a dual standard of UI generation that people will just have
to live with over the next few years.

Thank you for brining this up for the record Lorenzo, I hope that this
subject gets treated with it's due attention from all parties involved ;-)

Cheers,
                                                     -Tristan