jhalfs-X build status

George Boudreau georgeb at linuxfromscratch.org
Sun Mar 26 11:56:16 PST 2006



M.Canales.es wrote:
> El Domingo, 26 de Marzo de 2006 20:38, George Boudreau escribió:
> 
>>    I don't see the real value of this ability.. That is why I created
>> the config files, to eliminate the 20 switch command line.
> 
> When rendering the multiple variations for each book (CLFS requires 44 runs to 
> track properly the arch, method and testsites-level variations) to can 
> diffing the changes with previous ones, I find annoying to have to edit the 
> config files each time and to remember to return it to their default values 
> before doing a commit.
   I see your point..
> Of course, for final user could be most simple and failsafe to edit the config 
> files.
> 
>>    I also forgot the problem with CLFS 'boot method' makefile division.
>> Back in the cave I go, the yellow ball is supposed to return in my
>> lifetime.. *sigh*
> 
> The current implementation work for now?
   CLFS build works fine for method=chroot. There is a specific work 
method when using it with method=boot.
   At the moment once the building of the boot partition succeeds 
clfs-Makefile is forced to exit( by a very ugly method). The builder 
then transfers $BUILDDIR to the destination machine, and does a second 
make -f clfs-Makefile and the build process continues.

   What I want/wanted was
    make -f clfs-Makefile makeboot
    make -f clfs-Makefile makesys
   then there could be no change for errors.. (Yes, I did it once or 
twice and so will someone else)


> My point is that the current experimental code is better than the old one, 
> thus we should to promote it to trunk ASAP. 
> 



More information about the alfs-discuss mailing list