[RFC] Time to organize the LFS-6 development effort
Kevin P. Fleming
kpfleming at linuxfromscratch.org
Sun Apr 4 20:38:16 PDT 2004
James Robertson wrote:
> I would also like to add one small suggestion to the unstable branch (if
> that is the right term). Add a section on testing and maybe comparitive
> analysis. If you are in unstable territory and want to do testing, it
> would be a good place to bring all that info together. That way all
> hackers can do comparitive analysis and post results to help out.
> Kevin - can nALFS help with comparitive analysis?
In a sense yes, that's why I started this thread with the focus on
getting an ALFS profile built ASAP. We have testers (both "testing" and
"unstable" testers), but if they are not following the identical build
instructions their results are not directly comparable. Any tester who
wants to run more than one test is going to have use _some_ type of
automation, or they'll go out of their minds too quickly for our needs :-)
The analysis part is a little trickier; today it still has to be
human-based (or using some sort of scripting like Greg's ICA stuff). I
will be producing a new version of nALFS in the near future with
logfiles that _really_ document the entire build process, and those will
be directly comparable. However, that is at least a month or two off,
and I don't want to hold up this BE-LFS effort waiting for that.
More information about the alfs-discuss