jhalfs X: reworked paco patch try2

Dan Nicholson dbn.lists at gmail.com
Fri Mar 31 10:23:00 PST 2006


On 3/31/06, M.Canales.es <manuel at linuxfromscratch.org> wrote:
>
> Then, just after finished the first full chapter06 build, add:
>
> iteration1:  chapter6
>   tar -cjf /tools-$BOOK_VERSION.tar.bz2 /tools
>   rm -rf /tools
>   prepare_ica
>   rm -f {list of touched files that need be redone for rebuild_system target}
>   touch iteration1
>
> iteration2:  iteration1  rebuild_system
>   prepare_ica
>   rm -f {list of touched files that need be redone for rebuild_system target}
>   touch iteration2
>
> iteration3: iteration2  rebuild_system
>   prepare_ica
>   do_ica_test
>   touch iteration3
>
> 122-bootscripts:  iteration3

That sounds pretty good.  Initially it may be easiest to force the
number of iterations to be 3, but it really should be configureable. 
I.e., if you want to do 2 or 5 iterations, that should be possible. 
Most of the time 2 iterations is enough unless you want to be certain
that something will always differ (libstdc++.so, etc.).

This may be more difficult to generate.  If you can't immediately
think of a way to handle a generic ITERATIONS, then focus on the case
of 3.  It will be more important to get the looping and testing
correct.

Actually, it looks like make is well suited to this task.  The simple
touching and removing of files for the targets should work well.

Also, I don't know how jhalfs handles logging, but the rotating of
logs would be helpful.  I.e., if do_ica_test shows that there are
differences between iteration1 bison and iteration2 bison, having
build logs of both will help considerably in debugging the issue. 
Again, this is a secondary feature, so only focus on this if you
immediately see a way to handle it.

Keep me updated, Manuel.  I need to get more of my BLFS tasks taken
care of, but I'll be happy to contribute soon.

--
Dan



More information about the alfs-discuss mailing list