Re: real LSB compliance
On Sun, 1 Jul 2001, Joey Hess wrote:
> Steve Langasek wrote:
> > Does the LSB allow vendors to depend on being able to invoke those startup
> > scripts using LANA names?
> The intent seems to be that they want to allow syadmins to invoke them
> using the L[AN]* names. This implies that the docs accompnying the
> package (which may be in dead tree form..), should probably be able to
> document that.
Ok, but the LSB is a standard intended to make it easier for third-party
software vendors to distribute their software to a larger Linux market, and
make it easier for Linux distro creators to make sure this third-party
software will work on this system. Neither goal implicates the system
administrator's experience, so the *intentions* of the LSB authors are not
really relevant here, are they? :)
IMHO, we're LSB-compliant if we provide a mechanism for installing and using
LSB packages. How those packages look once they're done installing is
somewhat out of scope. ;)
> LSB packages also need to be able to run a program similar to
> update-rc.d on the scripts, which does rely on the program being told
> the full path to it, according to their examples. But as you note, alien
> could hack around that.
Yep -- there's quite a bit of other manipulation that would need to be done
with an LSB package to make it interact respectably with Debian packages (more
than just 'rpm -i <file>' -- what happens if someone decides to provide an lsb
package of emacs that tries to overwrite the .deb? :P), so it's not
programmatically difficult to change the script name while we're at it.
Steve Langasek
postmodern programmer
Reply to: