Re: GtkGrid design document
- From: Jonathan Blandford <jrb redhat com>
- To: Lorenzo Gil Sanchez <lgs sicem biz>
- Cc: otaylor redhat com, gtk-devel-list gnome org
- Subject: Re: GtkGrid design document
- Date: 19 Apr 2004 16:49:13 -0400
Lorenzo Gil Sanchez <lgs sicem biz> writes:
> I'd love to see that in GtkTreeView since that is what I was asking for
> some time ago. Don't get me wrong, I have enjoyed and learned a lot
> while writting the grid widget but at the beginning the only reason I
> start writing was because it was absolutely needed in my project and I
> got frustrated with the TreeView.
I would imagine that fixing TreeView will be easier than writing a new
widget from scratch. (-;
> I suppose I was (and I am) too unexperienced to add such behaviour to
> the TreeView but it people think it is worthy I will be happy to help
> enhacing the TreeView this way.
> > On the other hand, there are certainly cases where you want something
> > different from GtkTreeView:
> > - Efficient handling of many columns and sparse data
> Do you mean here that the header shouldn't be made from GtkButtons and
> only made of a regular gdkwindow? I guess that would allow to have 10k
> columns without memory worries.
Not necessarily the headers, but more importantly handling a sheet with
a large number of rows/columns, but only partially filled in. You don't
want to needlessly allocate memory for the entire area. But perhaps
this is more of a data-model issue. I haven't really thought through
the issues with this widget.
It does raise another issue. It seems like the GtkTreeModel interface
might be a bad fit for a grid widget, as it is inherently row based.
You might want to think about having a GtkGridModel interface (that the
GtkListStore could implement).
> > - Rectangular selection
> This is planned but still need to be studied since all the cells of the
> same column share the same type it will be difficult to allow random
> copy and paste of rectangular selections.
I imagine cut-n-paste working only a homogenous setup won't surprise the
> > But I don't have a good sense of what apps would use such a widget
> > at the moment. Having that idea is core to designing the widget.
> Just because we don't have any famous GNOME application that need it I
> don't think there is no neccesity for such a widget. Maybe home users
> don't really need this kind of programs but they are a must for company
> users. Here in Spain most of the small and medium size companies use
> some Access customized program to solve their needs so I guess the user
> base is potentially pretty.
Sounds like a reasonable start.
] [Thread Prev