[Kst] DCOP, Regression, "Commands"

[C. Barth Netterfield]

We'll be adding a third, in the javascript stuff...

Let's talk.

cbn

On Wednesday 05 November 2003 05:20 pm, George Staikos wrote:
>   I've been fleshing out the regression test a bit to catch some of these
> problems that we've encountered recently.  In order to do this, I need to
> enhance the DCOP interface somewhat.  This is actually allowing me to
> implement some of the requested DCOP interfaces as well.  Anyhow, I'm
> starting to notice a pattern here and I think it calls for a new concept in
> Kst.  Effectively we have objects (vectors, data objects), plots, and
> interfaces to those.  The problem is that we have two interfaces now,
> dialogs and DCOP, and effectively those interfaces boil down to methods to
> execute "commands" on those objects.  I think we need to consider the idea
> of creating a class of "Kst commands" such as "load data vector", "create
> plot", etc.  Then we could avoid some of the code duplication that is
> creeping into the DCOP interface.  It's basically another refactoring.  Any
> thoughts?