Re: What's Debian's /usr/src policy
Hi,
>>"Martin" == Martin Mitchell <martin@debian.org> writes:
Martin> I don't find this rationale convincing, particularly when you
Martin> only quoted part of the relevant section of the FSSTND. Let me
Martin> quote those lines, immediately following that part you quoted:
Martin> ------------------------------------------------------------------------
Martin> /usr/include should contain links to these directories, named
Martin> asm and linux. Since they are needed by the C compiler, at
Martin> least those include files should always be distributed with
Martin> installations which include a C compiler. They should be
Martin> distributed in the /usr/src/linux directory so there are no
Martin> problems when system administrators upgrade their kernel
Martin> version for the first time.
We have decided a while ago to go against this part of
fsstnd. Yes, we do so at times, but only for a compelling technical
reason.
Martin> The FSSTND is showing its age in the above section. It
Martin> justifies the reason the /usr/src/linux directory exists by
Martin> saying /usr/include contains links to it. This is not the case
Martin> in Debian.
/usr/src/linux/include/ Debian still point to the latest
installed kernel headers (either from the source or header
packages). I have yet to see a technical reason to break this
invariant.
So far, the only reason that has been advanced is to condone,
nay, pander to, perpetuation of the behaviour of users to encroach
into vendor space.
Kernels do not an OS make. Debian distributes an OS. A well
oiled, working OS, which can work with hundreds of contributors
contributing parts because we follow rules.
If you put things into /usr in digital UNIX, DEC shall tell
you they are not responsible for any breakage either.
Debian is not a haphazrd do-it-yourself SLS system. It is an
honest to goodness vendor supported OS distribution. You can't
scribble over vendor directories at random.
Martin> Please remove the link for the kernel-headers package
Martin> only. Rather than blind adherence to an obsolete standard, I
Martin> feel we should take a more intelligent approach to our
Martin> interpretation of the standard for Debian.
We defied the standard when we had a technical reason to do
so, and ever since I have had to defend that action to people on the
mailing list. We are, then, not hide bound by convention; but we
don't make egriigious changes at our whim either.
Martin> The link is completely irrelevant in the kernel-headers
Martin> package now.
I happen to disagree. Have you any reason for this, or is this
a statement of belief?
manoj
who like his random sig generator
--
"Trust me. I know what I'm doing." Sledge Hammer
Manoj Srivastava <srivasta@acm.org> <http://www.datasync.com/%7Esrivasta/>
Key C7261095 fingerprint = CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: