2008/2/26, Guy Dalziel <[EMAIL PROTECTED]>:
> The complication I am referring to is the fact that you can build on
>  some hosts but not on others, there was a table at one point that listed
>  the distros that had been tested to just such an effect. If someone was
>  using a distro that was unable to be used as a host system then the user
>  would have 3 choices, give up, install another distro, use the livecd. I
>  expect this situation to have arisen for more than a few people.
>  Unfortunately it's not always as simple as having package X, X and X.

OK, let's review the past cases of this breakage. This all boils down
to the following:

1) the user attempted to use a 64-bit host distro, which can't work
(well, it does work with the linux32 wrapper and the next-generation
build method from diy-linux).

2) there was a flaw in the LFS build method, that manifested itself
during one of the first three packages (binutils, gcc, glibc). A good
example is the 4th bullet from
http://www.linuxfromscratch.org/lfs/errata/6.2/ and is, again,
expected to be fixed for good in the next-generation build method from
diy-linux.

3) there was an upstream bug in one of the first packages, e.g., see
the second bullet in the "Miscellaneous Errata" section in
http://www.linuxfromscratch.org/lfs/errata/6.1.1/

The first case should go away soon, and the rest can be removed with
the errata page as soon as they are discovered.

>  I understand about being able to do lfs from anywhere and that the
>  livecd has somewhat eclipsed that, but I feel that the livecd is being
>  solely blamed for it when some of us simply prefer to use the livecd. If
>  people want to test distros as a suitable host and then make a table of
>  the results for everyone else, then by all means do so.

IMHO, the discussion went away from its intended course. It talks
mostly about the LiveCD removing the prerequisites, but forgot about
its unmaintained and unmaintainable (due to the huge amount of
packages and deviations, "rebuild everything" style of package
management, and lack of community feedback) status.

>  I doubt I would be of much help with the livecd since I prefer to use
>  6.1-3, however if the interest is lacking then my suggestion would be to
>  make a sourceless livecd that won't need to be updated so frequently.

There are already two variants of such sourceless CD (one with X and
one without).

-- 
Alexander E. Patrakov
-- 
http://linuxfromscratch.org/mailman/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Reply via email to