[Buildroot] svn commit: trunk/buildroot/target/device/Atmel/atngw100

[Ulf Samuelsson <ulf@atmel.com>]

ons 2007-09-26 klockan 23:02 +0200 skrev Bernhard Fischer:
> On Thu, Sep 20, 2007 at 09:44:16AM -0700, ulf at uclibc.org wrote:
> >Author: ulf
> >Date: 2007-09-20 09:44:15 -0700 (Thu, 20 Sep 2007)
> >New Revision: 19912
> >
> >Log:
> >Enable CONFIG_SEQ in busybox for ngw100
> >
> >Modified:
> >   trunk/buildroot/target/device/Atmel/atngw100/busybox-1.6.1.config
> 
> That's interresting. atngw100_defconfig doesn't build for me since at
> least some hundred revisions.

The toolset builds in the Atmel trunk at www.atmel.no/buildroot.
Something has broken after I checked in that toolset, and tested it.
I have been spending a lot of time trying to find out why.

When C++ is select, there is a problem with the current trunk.
The Atmel trunk, using the same source builds fine.
Something has happened with the toolchain makefile between July
and now, causing lots of problems building stdc++ver3.

Recently the uClibc for AVR32 build broke.
uclibc-0.9.28.config does not have the statement
# TARGET_avr32 is not set
I believe it was present in uclibc-0.9.28.3.config

I think that using the default uCLibc configuration file will
therefore use 386 architecture by default (you will get a question)
If the default is accepted, then things will break.

I just checked in a minimal build which does not exhibit the uClibc
problem.

make atngw100small_defconfig

should use a working uClibc config file

Will test building it during the night to see if there are other
problems.

BR
Ulf Samuelsson


> 
> Can you please show me the exact commands that are ment to build this
> config?
> 
> PS: In fact, what ATMEL defconfig can i use as a reference if *any*
> of these ATMEL targets is building successfully?