Re: Debian and FHS
Steve Robbins <email@example.com> writes:
> May I suggest that the policy document clearly state that the aim is
> `compatibility'? For instance, by replacing the quoted sentence with
> something like:
> Debian packages must be compatible with version X.Y of the FHS.
> (See the FHS document for a definition of `compatible'.)
You can do better than suggest; you can make a formal proposal.
(Anyone in the world can offer policy proposals, though only Debian
members can second them.) And if you propose this, I'm strongly
inclined to second. If I propose it, then you can't second unless
you're a project member.
It's simple: just file a bug report against debian-policy with your
suggested wording. (Best -- though not required -- to make a unified
diff against /usr/share/doc/debian-policy/policy.sgml.gz.)
My first comment, though: I doubt that we want to specify the version
of the FHS. I realize that there's a little ambiguity if we don't,
but A) I don't think we want to revise policy every time the FHS is
updated, and B) we do include a copy of the FHS with the debian-policy
package, so the ambiguity really isn't *that* bad.
Second comment: if you do make a formal proposal, make sure that you
point out where policy currently prevents compliance. (The /usr/doc
symlinks are a good example.) Otherwise you'll get lots of irrelevent
objections from people who think that compliance is a better goal.
(Of course it *is*, but that hardly matters when policy forbids it at
present.) It would be nice if people paid attention, but often they
And do keep in mind that individual packages comply with specific
versions of policy, and we don't require packages to comply with the
very latest version at all times. We only require that they comply
with a reasonably recent version. Therefore, even if policy *did*
require compliance, we would still have non-FHS-compliant-but-legal
packages for a while. Probably for at least one more release cycle.
> A question now occurs to me: is Debian `policy' now fully compatible with
> the FHS?
To the best of my knowledge, yes. There have been several rounds of
checking for FHS issues. If any are left, it should probably be
considered a bug in policy. (Most of the ones we have found to date
have been treated as bugs and quickly fixed.)
Chris Waters firstname.lastname@example.org | I have a truly elegant proof of the
or email@example.com | above, but it is too long to fit into
http://www.dsp.net/xtifr | this .signature file.