On Fri, Apr 04, 2008 at 09:07:19PM -0600, Tim Knitter wrote: > Well I wasn't able to convince myself :-) and changed it still.
Oh, no! :) I'm afraid I don't understand the new construct, because you're no longer ever comparing the BE mountpoint to root. I think you get it right for the "root == '/'" case, but not when root is somewhere else. In that case, don't you just end up picking the first inactive BE you run across? Which I guess leads me to ask a question I hadn't thought to ask before -- how does a person know what dataset to mount (and where) in order to point pkg at it, so that it knows to get all the BE stuff right? > One of my test system was always producing '/' for a -R arg of anything > but '/' when instantiating be: > > be = bootenv.BootEnv(img.get_root()) > > Not sure why img.get_root() was returning '/' for all cases. I hope it is > just an anomaly. I would hope so, too, but anomalies are scary, too. If you can point me at a directory where that's happening for you, please do so. > http://cr.opensolaris.org/~tsk/863.3/ client.py: - line 277, et al: Since you're not using "e", there's no need to specify it. bootenv.py: - line 82: You're comparing root against "/ "; I doubt that's what you meant. Given that, I'll request that you do a full testing of your code before you consider it done. In fact, I'd strongly recommend adding tests to the testsuite to help you with that. - line 148: four-space indent on a continuation line. - line 152: list + string == traceback - line 158: " ".join(cmd) is the idiom you want. And shouldn't that be self.clone_dir instead of self.root? In fact, I'd probably do "cmd += [ self.clone_dir ]" inside the if clause, and then just use cmd by itself in the two places. Use the other value in the else clause. - line 171: self.self? Danek _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
