Danek Duvall wrote:
On Fri, Mar 06, 2009 at 07:57:29PM -0800, Ed McKnight wrote:
2009.4,5.11-0.108.33284
is considered newer than
2009.4,5.11-0.108
I can accept that, though I didn't expect it.
Why? Isn't 2.0.1 newer than 2.0? How is this any different?
That's fine--looking at the definition of the fields in pkg(5) I can
see it this way. But I had come to think of the the specific value we
added (.33284) as arbitrary in its own field rather than being part of
the branch and indicating a version. I stand corrected. Is there any
thought of allowing non-integer values anywhere in here?
thx, --emk
Somewhere I got the idea that older packages from the preferred
authority would be 'preferred' over newer packages from a non-
preferred authority, but I can't locate that reference now.
There's been a *lot* of confusion in this area; I suspect that the only
person who doesn't think he's confused is Stephen. :) Shawn's wad is the
first step in clearing this up, but there's more that needs to be thought
through and (more importantly) written down. It will go in pkg(5) when it
does.
Danek
|
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
