Re: Can't build gcc-3.0.1 for m68k-coff

[Larry Gray <larry at greenmotor dot com>]

On Wednesday 17 October 2001 11:18, David Korn wrote:
> >-----Original Message-----
>
> From: Larry Gray [mailto:larry@greenmotor.com]
>
> >Sent: 17 October 2001 16:01
> >
> >There should be a listing of known working combinations of the
> >different
> >versions, along with the script (or instructions) that
> >produced them. That
> >way, the development system would "work straight from the
> >box". All that
> >should be required is the newlib port of your specific hardware.
>
>   There's no way in the world any kind of autoconfig process could examine
> the target board you are planning to use for your cross compiled code and
> find out how much ram and rom it has, what locations they can be addressed
> at, and what purposes they are intended for, so the toolchain is always
> going to need at least as much manual intervention as it takes to write a
> linker script as well.
>
>       DaveK

Hello,
	This is not what I meant, you're right that would be impossible. I thought 
that binutils, gcc, and newlib, could be built with the (for m68k-coff case) 
m68k-unknown-coff. What I'm suggesting, is a listing of known working 
combinations, with the hardware unknown. Then (of course) its up to the 
engineer/developer to port newlib to their hardware.
	Is it normal to have a complete toolset (binutil, gcc, newlib) for each 
board, even if they use the same cpu? We have one toolset, and change the 
specs file for the appropriate linker script and library for our different 
boards.

Thanks,
Larry Gray
larry@greenmotor.com

------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com