Subject: Re: Solaris packages from pkgsrc
To: Alistair Crooks <agc@pkgsrc.org>
From: Malcolm Herbert <mjch@mjch.net>
List: tech-pkg
Date: 12/10/2007 15:20:01
On Mon, Dec 10, 2007 at 12:17:12AM +0000, Alistair Crooks wrote:
|On Fri, Dec 07, 2007 at 05:00:24PM +1100, Malcolm Herbert wrote:
|> I've been playing with pkgtools/gensolpkg and it looks very nice ...
|>
[snip]
|>
|
|I haven't worked on Solaris packaging since gensolpkg was written,
|which must make it 7 or 8 years old. At the time I left it, it wasn't
|integrated into the build mechanism, and I was building (NetBSD-style)
|packages via pkgsrc, and then doing gensolpkg manually after that.
|
|Having said that, integrating it into pkgsrc would be superb to have,
|and I may be able to get my hands on some Sparc kit to do just that...
maybe it would be better to think of this in general terms - forget that
we are targetting Solaris in this case, as far as I can see there are
two main functions of pkgsrc - as a build environment and as a package
management environment.
If we were able to make the build environment independant of how
packages were managed, then we could easily generate native packages for
any OS, not just pkgsrc tarballs and Solaris packages.
Apart from creating some scripts to emulate pkg_add, pkg_info and pkg_rm
like functionality, the only issue may be package namespace collisions -
in the case of Solaris this is worked around by prepending each package
name with TNF. Similar tricks would be needed for rpm, deb or whatever.
I don't think it would be a good idea to continue to use the pkgsrc
package database and the Solaris package database at the same
time - I accidentally created a Solaris TNFperl package out of an
already-installed pkgsrc perl package and got myself into a knot when
Solaris' pkgrm removed it from underneath pkgsrc ... :/
|> ... which makes me wonder whether the bulk package builder would work
|> too ... 8)
|
|I think this is a SMOP, FSVO "small".
if the above were done, then this should just naturally happen without
further work, as far as I can see ...
|If you're interested in pursuing this, please contact me off list, and
|we can try to further refine that 'S'.
I think that would be a Good Thing ... :)
--
Malcolm Herbert This brain intentionally
mjch@mjch.net left blank