Build Options Survey run

[Poul-Henning Kamp]

--------
In message <B0B7A53D-2B33-40CE-97EB-F53A43B307E6 at bsdimp.com>, Warner Losh writes:

>> phk had asked me to run a build options survey again.  It took about 
>two weeks to go through all of them.  The results are here:

Many thanks!

>> There are plans to run it more often again and improve it a bit.  No 
>promises on the actual timeline but it'll happen.

It's actually a good beginner task if anybody wants to do something
for the project which requires just a machine and some shell
programming.

My old script can be improved in a lot of ways, the most productive
would probably be to make it incremental rather than one big batch
job.  (Something like:  Test any options for which there are no
results, otherwise retest the option with the oldest result.)

>While not all combinations tested were valid combinations, many that 
>failed are. Some of them look to be relatively easy to fix, while others 
>are much harder.

If one _really_ wanted to burn CPU cycles, my testing years back revealed
many combinations where options are incompatible.  (NB: Today we have almost
four times as many options as then and we're into N! territory here...)

-- 
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.