On 16/09/2007, Peter Tribble <peter.tribble at gmail.com> wrote:
> On 9/16/07, Shawn Walker <swalker at opensolaris.org> wrote:
> >
> > That's why I think Danek's suggestion about integrating functionality
> > scripts need into the packaging system and making them use it instead
> > of being able to do their own thing is preferable.
>
> There was a time I liked this idea. Then the realization that it's
> not quite so simple dawns.
>
> Is it really feasible to expect the packaging system to understand all
> possible administrative actions that might be necessary for all software
> in all possible install contexts - many of which do not yet exist and
> will not be tested?

The idea here is to limit what the possible administrative actions
are. When it comes down to it, you really don't want packages doing a
bunch of things they're not supposed to be doing. It's easier to add
specific functionality to allow packages to do things (in my opinion)
than it is to try to keep them from doing a bunch of things.

> I worry about the burden imposed on software suppliers, who have
> a new mechanism to learn and support. They have to provide
> something that is different from prior releases, and that behaves
> differently from other platforms.

Every platform behaves differently, so that's not much of a
justification. In addition, I would expect that, no matter what is
adopted here, the mechanism for OpenSolaris is going to be wildly
different regardless of whether or not scripting is supported from
Solaris. As such, there is going to be a support burden regardless of
what happens. The package structure is certainly going to change from
everything I'm seeing.

> I think that we're going to have to support scripting during install,
> and that we should then provide the mechanisms to allow those
> writing install scripts to do so as easily and safely as possible.

I think the goal of not supporting arbitrary scripting is admirable,
and that it is worth trying a new approach out before succumbing to
what is supposed "inevitable defeat" and accepting arbitrary scripting
as a necessary part of packaging.

-- 
Shawn Walker, Software and Systems Analyst
binarycrusader at gmail.com - http://binarycrusader.blogspot.com/

"Beware of bugs in the above code; I have only proved it correct, not
tried it. " --Donald Knuth

Reply via email to