Subject: Re: the netbsd way
To: None <tech-pkg@netbsd.org>
From: Joerg Sonnenberger <joerg@britannica.bec.de>
List: tech-pkg
Date: 10/31/2006 14:38:47
On Tue, Oct 31, 2006 at 11:32:08AM +0100, Marc Espie wrote:
> One thing that strikes me though is that, even though some of you
> talked about FreeBSD's package system, there is absolutely no mention
> of the way OpenBSD does things.

I certainly try to follow the OpenBSD development :-) Heck, I recently
even went as far as asking on IRC about the sudo support in OpenBSD as I
couldn't find it directly (no OpenBSD installed). When I pointed out
that the documentation in ports(7) is misleading, the responsive was not
really helpful and I went away again.

> We rewrote our pkgtools infrastructure.

I am aware of it and while I agree with some of the choices, i disagree
with others. One problem mentioned already is the use of Perl. Another
even bigger item is the separation of management infrastructure from
development in the tools. E.g. package lists are supposed to be static
for the tools, how they are generated is outside the scope of them. I am
a big fan of mostly static PLISTs for packages as they help finding
nasty bugs like uncompletly builds. This is not so much an issue for
OpenBSD as the supported software base is much more homogenous, but
hunting down issues e.g. with random limitations of native tools on four
Operating Systems can be daring.

> There's no reason you need to run into the same issues we did.

I hope to avoid that which is one reason why the process of pkg_install
is slower than it could be.

> In fact, our current clean-ups involve formal checking
> of file permissions and ownership, then turning on a more paranoid mode
> for pkg_create/pkg_add.

That's on my very list of things to do.

> I don't have high hopes there. I expect that this email will be ignored
> like most of the stuff I post to this list....

I read pretty much any mail to this list and even if I don't comment on
them, I don't ignore them either. I think we both arrived at similar
conclusions for many parts, so please -- don't be discouraged to speak
with us. I am interested in what other systems learned. I might not
agree with the conclusions, but that's a my good right.

BTW, will you come to Milan?

Joerg