KDevelop editor interfaces - ups

[Matthias Hoelzer-Kluepfel]

On Thu, 26 Apr 2001, Richard Dale wrote:

> But I don't think KDevelop HEAD should be turned into a test bed for getting
> various unusual window management UI's/editors working. Why do editors need to
> manage windows? I don't know, the discussion is all beyond me :-)...
>
> The current framework is already proving its worth. For instance, I tried the
> code reformat option last night - it works great on java, and just adds a
> single menu item to do something useful. And you can set up the plugins you
> want enabled quite easily with Mathias's option setting plugin. The java
> jdb debugger will fit in quite nicely when it's done, and be able to run
> concurrently with the gdb frontend. Both debuggers are optional and just add
> the menu items they need, and so on...

Richard,

these things are not as independent as they seem. The part
architecture is very good, but as an author of a part, you
need access to some other parts of the system. For example,
the reformat part I wrote needs access to the text currently
in the editor. I solved this by looking up the editor part and
casting it around to some class I just happened to know to be
the one use by kdevelop. But the right thing would be to have
a standardized interface to access this text, and that is what
part of all this fuss is about. It is not esoteric, it is
about providing a clean interface so the parts can be as
powerfull as possible. Believe me, I want to write
(non-editor!) parts, not interfaces :)


> The number of people who would move to KDevelop 3.0 because of emacs or vi
> support is much smaller, compared with the numbers you can get by doing, say,
> customised java or python programming environments. I would like the idea to
> be to try and make KDE programming easier for the average programmer, so we pull
> more programmers away from lesser apis :-). But too much emphasis on bells
> and whistles for the real programming wizards (ie emacs users) won't do that.

Again, you are correct, but getting an interface right does
not mean that it will be any harder to use for an average
user.  Most of the time, it will get easier.

And about editor fans: I happen to know one experienced
KDE developer who started to use KDevelop 3.0 after just
_showing_ him how he could integrate his loved nedit into the
framework. (I won't talk any more about that. He is following
this list closely ;)


Bye,
Matthias.



-
to unsubscribe from this list send an email to kdevelop-devel-request at kdevelop.org with the following body:
unsubscribe »your-email-address«