Re: [avr-libc-dev] build per arch

[E. Weddington]

On 23 Apr 2004 at 9:59, Theodore A. Roth wrote:

> I've started a new thread for this since I think it's worth
> discussing...
> 
> On Fri, 23 Apr 2004, Joerg Wunsch wrote:
> >
> > I wonder whether the current way of building per `architecture' is
> > still that optimal, or whether we should start a finer-grained build
> > scenario (which unfortunately would require changes to the linker).
> 
> At first I didn't think that we'd need linker changes, but doing a bit
> of digging shows that we would. :-(

What specifically? 
 
> We'd probably need to change things so that there would be a directory
> for each device which looked like this (relative to ${prefix}/avr/lib/)
> (showing only mega128 for brievity):
> 
>   ./atmega128/libc.a
>   ./atmega128/libm.a
>   ./atmega128/libprintf_min.a
>   ./atmega128/libprintf_flt.a
>   ./atmega128/libscanf_min.a
>   ./atmega128/libscanf_flt.a
>   ./atmega128/crtm128.o
> 
> or maybe like this:
> 
>   ./avr5/atmega128/libc.a
>   ./avr5/atmega128/libm.a
>   ./avr5/atmega128/libprintf_min.a
>   ./avr5/atmega128/libprintf_flt.a
>   ./avr5/atmega128/libscanf_min.a
>   ./avr5/atmega128/libscanf_flt.a
>   ./avr5/atmega128/crtm128.o


If its going to per-device, then I would vote to drop the architecture 
subdirectory as it seems to 
be arbitrary to begin with.
 
 
> Would we need a special linker script on a per device basis or could we
> still use the arch based ones? I don't know...

Good question....
 
> Now would be a good time to start on this since binutils 2.15 is nearing
> release (as far as I can tell)

I thought I keep hearing "within a week or two".

> Of course, this will cause the build to take a considerly long time, but
> who cares, we've all got 3 GHz systems by now, right? 8-)

I'm ok with it taking a long time. Right now, nothing compares with how long it 
takes to build 
gcc and gdb / insight. :-P

Eric