On Wed, Jun 21, 2000 at 11:34:27PM -0700, H. Peter Anvin wrote:
> Daniel Bradley wrote:
> > > WHY should we change the name? We didn't change the name of
> > > /usr/sbin/sendmail to lsb-mta because we standardized the interface.
> > Yeah but after it was done I think a lot of people realized it was a
> > stupid thing to do.
> I believe it was exactly the right thing to do.
Except that it came about more because implementations just turned out to
be compatible in that manner. ie, smail, exim, postfix all have sendmail
binaries because that's what's useful for admins, not because the LSB
dictated that this was the way things should be.
> > > All you do by changing the name is breaking all existing scripts.
> > Oh come on!, all you need is a symlink from lsb-package to rpm.
> No, you have 2^n scripts out there trying to access "rpm". Now you have
> to rewrite them. Congrats.
And the 2^n scripts out there that try to access "dpkg"? Why should
either set of scripts be considered LSB compliant in their present form?
What of the 2^n scripts out there that use rpm features that dpkg doesn't
have, such as, say --percent or verifying and install direct from an ftp
site? Why should other distributions have to rewrite rpm to get these
features that aren't needed or wanted by their users?
What of the non-rpm systems that wish to be LSB compliant, but also make
.rpm packages, either directly or via alien, and thus want an "rpm" binary
that actually *is* rpm, even though they don't want to use rpm to update
their system?
We should be doing the minimum amount necessary so people can create
packages that can be installed on Red Hat, Debian and Slackware
systems. Not forcing distributors to redesign their products, or rewriting
package managers that already work.
Cheers,
aj, who wonders whatever happened to "release early, release often"
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
``We reject: kings, presidents, and voting.
We believe in: rough consensus and working code.''
-- Dave Clark
Attachment:
pgpHwTcM3v17F.pgp
Description: PGP signature