Gerard Beekmans gerard at linuxfromscratch.org
Mon Jun 11 12:45:16 PDT 2001

On Sun, Jun 10, 2001 at 09:21:40AM -0400, Scot Mc Pherson wrote:
> To use an example that is already specified in the book....VIM does not have
> all features compiled when you finish the book and you must recompile it
> after you have installed other /non/LFS packages before all the features are
> utilized.
> It is no problem to recompile a piece of software, as well if one were to
> compile everything according to the book, they would find a lot of things
> that they could change and suggest according to scenario.
> BUT...
> Not everyone has DVD drives, and if we start beginning to say, "If you
> have...., then use this config, and if you have..., then use this other
> config" the book will become a set of if/then statements which will lead to
> a lot of people making configuration errors and building a non working LFS.
> We want to keep LFS simple, not make it more complicated, that is what keeps
> people from Linux distros to begin with and what will keep people from LFS
> if we go the same route.
> I state again, I think it is not within the spirit of the book to place
> configuration specific issues within the book's binder...As a hint I think
> it is indispensable, but I do not think its appropriate for the book...

That's one of the major reasons why I'm reluctant of adding too many
things. Right now we have a set of packages that everybody more or less
agrees we should install (i said more or less, i didn't say it's
perfect). Now, there's something to be said for adding util-linux's raw
program. then what about the missing gcc NLS, or missing ncurses
profiling, and dozens of other items that are missing. I prefer to just
install a package using it's default settings, not enabling or disabling
specific feature. If ncurses doesn't install profiling libs by default,
but Glibc does, well so be it. There's als somethin to be said for "LFS
provides a means to and end, but not _the_ definite guide. You are still
supposed to read all the docs with a package and make changes as you
require them". It's just that what's in the book right now most people
can live with, some can't (I know I can't - I don't follow the book
myself exactly, but that doesnt' mean I'll just going to change it so my
next LFS can be done braindead by copy&pasting everything without

The discussion can go on forever. On the 'raw' issue. If future
util-linux versions include that enable raw feature, then the book won't
disable it. 'raw' will then be installed by default.

Another point that's already been raised: if we start customizing one
package to build something extra that's not build normally speaking,
then we can't discriminate and not do the same for other packages. You
can't say that 'raw' is more important that ncurses' profiling libs just
because _you_ (im not pointing to anybody in specific) don't use profiling 

Gerard Beekmans

-*- If Linux doesn't have the solution, you have the wrong problem -*-
Unsubscribe: send email to lfs-apps-request at linuxfromscratch.org
and put unsubscribe in the subject header of the message

More information about the blfs-support mailing list