Remove BIND, Sendmail, Perl and etc from base?
Matthew Dillon
dillon at apollo.backplane.com
Wed Jul 23 10:56:11 PDT 2003
:On Wed, 23 Jul 2003 13:53:43 +0200, Simon 'corecode' Schubert wrote:
:
:> Lately Jeremy Messenger told:
:>
:>> Good, have the ability of build the source (like ports) is great.
:>> Because, many apps have the options to enable or disable in the compile
:>> such as Gnome apps. For example, if I want AbiWord to compile with
:>> Gnome support. If someone don't even want to have AbiWord with the
:>> Gnome support. I don't know how it works with Debian, I only use it for
:>> like a week long time ago.
:>
:> oh please, people, stop talking about debian. it has always been a PITA
:> to me.
:
:I didn't say anything about to copy Debian and I dislike Debian. I was
:just wondering how Debian does that get people happy without complain
:about all binaries are limited. But, I guess it's because Debian is
:creating too huge binaries database by abiword-gnome, abiword and
:abiword-plugins.
:
:--
:bsdforums.org 's moderator, mezz.
:
Well, Debian makes people happy because of this:
apt-get upgrade
And "it just works". It installs all available uprades for things
you already have installed, it restarts servers that need restarting,
it will even turn off nfs exports when installing a new nfs-kernel-server
then turn them back on, etc. At a minimum we would need to support
similar functionality.
FreeBSD's ports and pkg system doesn't even come close to the level
of functionality that Debian's apt-get represents, and tools like
portupgrade etc do not produce reliable results. Every time I try
to do a major upgrade of my installed ports on a FreeBSD box half
of them wind up broken.
That said, I believe that the FreeBSD ports infrastructure can be
used as a basis for the model that we've been discussing. I believe
it is possible to create a much stricter environment for ports to
be built, installed, and run in which properly maps and guarentees
the myrid dependancies.
-Matt
Matthew Dillon
<dillon at xxxxxxxxxxxxx>
More information about the Kernel
mailing list