Shawn Walker wrote:

> And yes, I'm aware they're separate now so that --require-new-be
> doesn't force you to provide the BE name.  However, I wonder how
> much value there is in that.

If you don't care about your BE names, and are happy with solaris-N, then
you don't use --be-name.  I think people who do that are mental.  :)

On the other hand, if I used --be-name and got a new BE, but was surprised
that I didn't actually get bits that required it, I would likely be that it
created the new BE for me anyway, and would want to re-run the command to
not produce the new BE.

We probably need to think about this a bit more, anyway.  Bart and I had a
discussion the other day about having successful non-BE-creating operations
leaving behind your previous BE (i.e., take a snapshot prior to the
operation and operate in the live image like a normal install, but instead
of deleting the snapshot on success, create a new BE based on that
snapshot).  Pretty slick for leaving a trail of BEs for posterity, which it
appears some customers might like.  But it introduces a whole new angle to
thinking about naming BEs -- am I naming the next one, the previous one, or
am I renaming the current one (and passing the current name off to the now
previous BE)?

Danek
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to