Summary: experiences with NanoBSD, successes and nits on a Soekris 4801

[Poul-Henning Kamp]

In message <m2fyuv75cj.wl%gnn at neville-neil.com>, gnn at freebsd.org writes:
>At Mon, 20 Jun 2005 08:36:07 +0200,
>Poul-Henning Kamp wrote:
>> The trouble with options like this is that they escape our normal
>> build tests.
>> 
>> A good example of this is the kernel option INET which in theory
>> is optional, but which on average only is it 10% of the time.
>
>The only way to get this to work, IMHO, is to take the full system,
>and generate a dependency graph, if that's even possible.  Then you
>know where to cut and where new APIs need to be defined to know where
>to cut.  I'll try to generate this as part of the scripts on
>code-speluking.org.

I played with developing that graph by removing lines from LINT
and see what compiled and what didn't.  Based on the progress I
made I would estimate the full graph will take about 1 CPU-year to
calculate by trial&error.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.