Re: LSB Spec 1.2 criticism
On Sun, Jun 16, 2002 at 06:16:26PM +0200, Hendrik Visage wrote:
> On Mon, Jun 17, 2002 at 01:29:18AM +1000, Anthony Towns wrote:
> > Still in chapter 22, Facility Names, it's not clear what value $local_fs
>
> > In the next section, Script names, I'd like to again suggest that LSB scripts
> > be named either:
> >
> > * lsb-<foo> for LANANA allocated names
> > or * foo.com-<bar> for names based on the DNS
> >
>
> I'm still not seeing the need for this "registration" etc.
The reason why scripts names need to be registered is because
otherwise packages that chose the same /etc/init.d script name can't
be installed at the same time. Conflicts between script names used by
LSB programs and ones used by system distributions are also a problem,
although that can be solved by using a namespace, such as lsb-<foo> or
foo.com<bar>.
The reason why we didn't require lsb script names to be named
something like lsb-<foo> or foo.com-<bar> was the acknowledgement that
system administrators *do* type /etc/init.d/<foo> to start and stop
services, and it seemed unfair to require non-distribution packages to
need to type something longer and more heinous. (Although granted,
/etc/init.d/lsb-oracle isn't *that* bad...)
This was a tradeoff which I acknowledged when I first wrote the init.d
naming proposal, and while there wasn't a lot of discussion on this
subjection, I seem to remember that there was at least some
consideration before we adopted it.
> > The real reason I think that's a better way of doing things is that
> > otherwise Debian needs to register some 350 init.d script names [1],
> > and needs to worry about an additional 114 that conflict with the LSB's
> > guidelines. Personally, I think we should avoid this conflict while
> > we can. (For reference, there are about 40 pre-allocated names listed in
> > the spec at the moment)
The assumption is that LANANA would registered all of the script names
used by distributions, and reserve them as "reserved for
distributions", so that LSB applications wouldn't conflict with
distribution-used init.d names.
I assume the 114 script names that conflict with the LSB guidelines
are the ones which use the '-' or the '.' characters? If it would
help, we should be able to change the definition of the assigned
namespace so that '.' is permitted; the only characters we need to
really reserve are leading underscore characters, and the hyphen
character. And even for the hyphen character, if necessary we can
simply have LANANA reserve the certain hyphen prefixed --- for
example, "nfs-*" can be reserved for distribution use only, which
would take care of things like Debian's /etc/init.d/nfs-kernel-server.
So I think these problems can be solved, without causing too much
undue pain to the distributions. The other thing we *could* do is
consider re-evaluating the tradeoff which we made long ago to allow
applications provided by third-party LSB vendors to use "short"
/etc/init.d names. This is a change, though, and it's not at *all*
clear to me that we should make such a fundamental change the last few
days before LSB 1.2 freezes.
- Ted
--
To UNSUBSCRIBE, email to lsb-spec-request@lists.linuxbase.org
with subject of "unsubscribe". Trouble? Email listmaster@lists.linuxbase.org
Reply to: