Re: Inplace Tips for GtkTreeView (#346992)

["Michael Urman" <murman gmail com>]

On 7/19/06, Kristian Rietveld <kris gtk org> wrote:
> Supporting this in the generic tooltips API probably adds quite a bit of
> complexity.  That said, I am also still unsure whether it makes sense to
> have this in GTK+.

That fits my assessment. It looks a lot like a tooltip, but it doesn't
act much like one, and thus should be separate from tooltips. So what
more do you need to know to help you become sure whether or not this
belongs in GTK+?

Arguments I see for having it in GTK+:
 * It requires API addition to CellRenderer interface (the patch uses
up a reserved pointer) to determine the real size of a column's
contents.
 * It relies on a semi-private TreeViewColumn method to render the
contents on the tip. It could alternately leverage a way to copy
CellRenderers with attributes and functions from one CellLayout to
another, but I think this way is preferable.
 * It makes it really easy for all applications to use this - just
set a property on a TreeViewColumn and it gets inplace-tips for that
column.
 * From the user perspective, it catches us up to Windows usability
in TreeViews.

Arguments against:
 * It uses up one of two reserved pointers in the CellRenderer
Interface; even if it's doing so with a general function, there might
be something more important later.
 * It's unproven; only one app is using something close (but less
complete), and there have been reports it conflicts with button
presses in some environment. It hasn't gone through libegg.

I'd really like to see this capability in GTK+, but the sooner we get
to a solid yes or no, the happier I'll be.
--
Michael Urman  http://www.tortall.net/mu/blog