From: Jon M. Taylor <taylorj@gaia.ecs.csus.edu>
  To  : ggi-develop@eskimo.com
  Date: Thu, 10 Jun 1999 20:35:42 -0700 (PDT)

Re: generic-kgi renaming request

On Fri, 11 Jun 1999, Andreas Beck wrote:

> > 	There's only one problem:  The LibGGI fbdev target uses
> > "generic-kgicon" as its API name to use when loading the genkgi helper. 
> > However, unlike the LibGGI genkgi helper, the GGIMesa genkgi helper
> > implements card-specific acceleration APIs, and as such we get things like
> > tgt-fbdev-kgicon-generic-savage4-mesa
> 
> I'll take care of that when I have some time (after the 22nd). I'll then as
> well get the suikgi target up again and make that stuff something like
> generic-kgicommand.

	Don't worry about it, I worked around it by explicitly detecting the
presence of a Savage4 in the mesa-genkgi dlinit() function.  The mesa-genkgi
getapi() function now only has one case to handle now - if a supported accel
is detected, it loads the appropriate library.  Otherwise it does nothing,
since all the generic-* helpers have already been loaded by
display-fbdev-mesa.
 
> Your card-specific stuff should usually be loaded by the "vendor" section.
> I'll take care of reimplementing that functionality.

	That would be cool.  A separate vendor section would make the
distribution of the mesa-fbdev-kgicon-<vendor> libraries completely separable
from GGIMesa as a whole.  But, that will mean that _all_ the sublibs that
load a .conf file will need to use ggLoadConfig() and friends on their own -
currently, only the top-level extension code itself handles its own .conf
file.  All targets, helpers, etc which belong to that extension .include
their .conf files into the top-level .conf (ggimesa.conf, etc).

Jon

---
'Cloning and the reprogramming of DNA is the first serious step in 
becoming one with God.'
	- Scientist G. Richard Seed