[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Mon, 05 Feb 2001 03:03:18 -0600
Yesterday, I experimentally obliterated all traces of the gmodel
code from gnustep-gui. This broke only NSColorPanel. Today, I did a
CVS update and found that Fred re-implemented NSColorPanel without my
gmodel. As we can see, gmodels aren't winning any popularity contests.
Instead of fixing up a few problems with the NSSavePanel and
NSColorPanel gmodels, we now have hard-coded versions. And none of the
apps released for GNUstep uses gmodels. Unless I'm missing something, a
gmodel doesn't have any advantage over a NIB-style object archive. Only
native GNUstep and OPENSTEP could possibly share gmodels, and even then
with a lot of hackish code to take into account different widget sizes.
This leads me to suggest removing libgmodel from gnustep-gui.
Nothing in gnustep-gui uses it any longer, and few (if any) other apps
use it. I think the best use for the code is as a seperate library for
any apps that still use it, and for nib2gmodel. Thus, a gmodel would be
a transfer format for Gorm to load for tweaking, before saving as a
Does this sound reasonable?
All persons, living or dead, are purely coincedental.
Re: Nuke libgmodel?, Gregory Casamento, 2001/02/05
- Nuke libgmodel?,
Jonathan Gapen <=