Index: [thread] [date] [subject] [author]
  From: Cesar Crusius <crusius@leland.stanford.edu>
  To  : ggi-develop@eskimo.com
  Date: Mon, 16 Aug 1999 19:37:14 -0700

Re: kgicon C&T

Andreas Beck wrote:
> 
> Hi !
> 
> > Maybe, if I didn't work from 7am to 9pm everyday...
> 
> *grin* Change jobs :-).
> 
> > BUT if someone can convince me that I need kgicon (I have fbcon compiled
> > in my kernel)
> 
> there is a CT fbcon kernel ? Or are you using VESAFB ?

Sorry. VESAFb.

> > BTW, I have a probably stupid question: what do I need kgi for? I mean,
> > the kernel thing. I have fbcon compiled, but not the kgi patch.
> 
> KGIcon is no patch. kgicon ist just a wrapper to KGI drivers that makes
> loadable modules that look like fbcon drivers to the kernel.

I meant the KGI package (now in version 0.9). It patches the kernel and
you have to recompile it without Virtual Consoles support. I tried
but: neither the kernel (2.2.10) nor KGI compiled successfully.

> Advantage: They are accelerated generically and usually support the card
> a little better than the fbcon counterpart, if it exists.

Not in my case... (C&T) :(

> If you use VESAfb, you are much better off with a native driver, as the
> native driver can switch modes.

That means that I should stick to VESAfb ? Weird things were happening, and
I thought that it could be some driver problem. XGGI works fine (a bit slow
in some areas, but ok), but ggv segfaults.

> > some programs work (XGGI) using a fbdev display and some don't (cube3d).
> > If somebody can explain it to me, please do it...
> 
> Hmm - exactly what driver do you use ? cube3d for example asks for
> doublebuffering, which could be pretty impossible with some fbcon drivers
> (like VESAfb).

That explains it. Thanks!

> > Also, the configure.in files of libGGI and libGII hardcode the "-g"
> > compiler option, so I cannot compile them without it (even specifying
> > CFLAGS when ./configure-ing).
> 
> Uh - wasn't there an option to turn of all debugging ? Doesn't
> configure --disable-debug #don't build with run-time debugging (speed freaks)
> do that ?
> 
> Other than that strip --strip-unneded does wonders on that.

No. There's a line somewhere on configure.in that explicitly adds the "-g"
flag
together with some '-D' flags. I don't have the script here, but I'm sure it's
there.
--disable-debug and CFLAGS="-O2000" won't do the trick. I had to edit the
configure script myself to be able to compile the libraries without '-g' (and
we need that because we can't strip libraries).

> > For kgi, I have the kernel sources in another place (not /usr/src/linux),
> > but make doesn't recognize this. And 'make' doesn't complete successfuly
> > (but after some modifications it patches the kernel). Is that normal?
> 
> What KGI are you talking about ? Steffen's 0.0.9 ?

Yes...

-- 

Cesar Augusto Rorato Crusius      o__     o__     o__     o__     o__
Stanford University               ,>/'_   ,>/'_   ,>/'_   ,>/'_   ,>/'_
e-mail:c.crusius@ieee.org        (_)\(_) (_)\(_) (_)\(_) (_)\(_) (_)\(_)
www.stanford.edu/~crusius

HE WHO SACRIFICES FUNCTIONALITY FOR EASE OF USE
LOSES BOTH AND DESERVES NEITHER

Index: [thread] [date] [subject] [author]