RE: stupid question
Quoting Greg Folkert <firstname.lastname@example.org>:
> On Tue, 2005-03-29 at 08:27 -0500, Roberto C. Sanchez wrote:
> > Quoting Armands Upmalis <email@example.com>:
> > > If U want the kernel form debian, then apt-cache search kernel-source |
> > > x.x (x.x is the number of kernel), when U are found the latest kernel U
> > > then reconfigure, rebuild the kernel...U can get aslo the newest kernel
> > > http://www.kernel.org!!!
> > >
> > > uncompress the new kernel in /usr/src, then follow tje steps:
> > > ln -s /usr/src/kernel-source-U-downloaded /usr/src/linux
> > <SNIP>
> > I forgot one other thing. According to Linus, you should never make the
> > /usr/src/linux symlink (unless you are a glibc developer), and you should
> > compile the kernel in your home directory:
> > http://linuxmafia.com/faq/Kernel/usr-src-linux-symlink.html
> Ahhhh, so now we are taking Rick Moen's writing as gospel now, are we?
> Please make sure you take off your rose-colored glasses before you make
> such statements.
Sorry. I don't know who Rick Moen is. The message I reference was posted by
> Personally, I like Rick and Karsten (he has a few things there too)(and
> those two and I have had a few (numerous?) rounds of arguments,
> elsewhere), I just have to make sure you understand. These two people
> are very good at pointing out problems. Very good at solving problems.
> Very good at explaining things at length. Very good a pointing out
> non-compliance at certain, traditionally done things.
> But, there is a point where you have to start thinking on your own about
> this type of thing. Making good judgment calls with things like Rick's
> and Karsten's articles as references... Making sure the way things are
> done are repeatable and understandable to you.
> Blanket statements are really too broad, even Linus Torvalds has his own
> way of doing things, you should too. As long as the outcome is what YOU
> want... therein lies the lesson. If you do things in a way you
> understand and is (subjectively)"good practice", you should have little
> to worry about.
Um, OK. But if you read this (admittedly old) thread from -devel, you will
see that the Debian developers agree with Linus' reasoning (at least as far as
the symlinks go):
It is enough for me that if the Debian glibc and/or kernel maintainers agree
with Linus' opinion on this point, that it would stand to reason that it is
probably a best practice for Debian.
Roberto C. Sanchez