On Wed, 12 Jan 2000, Brian S. Julin wrote:
> On Wed, 12 Jan 2000, Mardy wrote:
> > Then, here is the make log (unfortunately, when the patch has already
> > been applied). Looks like KGI is missing two GGI file, which I think I
> > have somewhere, but I preferred not to move anything before having some
> > info from you. :-) Should I just copy those files? Shouldn't they be
> > included in KGI?
>
> To me it looks like the libggi directory is meant to compile
> on a system on which libggi has already been installed, (perhaps
> even to be copied over into the libggi tree?)
Right now that code is an exact copy of the KGI/Dali genkgi target
code, just as a placeholder for code I never got around to writing yet.
I should have removed it before I tarred up the tree, sorry about that.
Just ignore it for now.
As far as how the directory will 'work' when there is useful code
in it, I'm not sure yet. The LibGGI build system is sufficiently modular
that we _should_ be able to have a complete, functional standalone build
and install system for the KGI-0.9 display target. This would produce a
.so file which should be installable into an existing /usr/local/lib/ggi,
along with the necessary modifications to /usr/local/etc/ggi/libggi.conf
so that LibGGI 'sees' the new target.
The whole process is supposed to work without any modifications to
the base LibGGI sources. This is supposed to be possible with LibGGI, so
that hardware vendors can provide an easily-installable binary kernel
driver and the associated binary userspace LibGGI .so in one simple
package. AFAIK, though, no one has put this capability to the test yet.
Perhaps the KGI-0.9 target should be used to test/debug this ability of
LibGGI?
Jon
---
'Cloning and the reprogramming of DNA is the first serious step in
becoming one with God.'
- Scientist G. Richard Seed
