Danek Duvall wrote:
On Fri, Mar 06, 2009 at 07:13:13PM -0800, Ed McKnight wrote:
I'm not sure whether that's deliberate in the coding, but it's
decidedly confusing, and therefore busted.
Regardless, you're right that the right answer for this particular
situation is fat packages.
The situation is not variant content--it's what repo are packages going
to be drawn from. The fact that one set of packages is debug and has an
indicator in its version string was the give away that packages were
being drawn--incorrectly on two counts, IMHO: non-preferred repo; older
packages--from multiple repositories.
Or, are you saying that the .33284 in one of the version strings (older
package; non-preferred authority) makes that one appear newer to the
Planner?
I'm saying that despite the fact that the results you're getting are due to
a somewhat confusing and busted plan creation, you should actually be
avoiding the problem entirely by having debug and non-debug variants to
indicate debug and non-debug content, rather than using package versioning
to do so. You're using a hammer when you should be using a pair of pliers,
and although there is in fact something wrong with the hammer, you should
still be using the pair of pliers.
Danek, et. al.,
I'm a bit confused about how we would implement this with variants. I
understand how the arch variant works. But I can't find any information
on a "debug" variant or creating custom variants. Would we add tags like
"variant.release_type=debug"? Then how would IPS know whether we're
installing on a debug or non-debug image?
Any documentation pointers appreciated.
Thanks,
Nick
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
