On Tue, Dec 15, 2009 at 07:31:50PM -0600, Shawn Walker wrote: > [email protected] wrote: > >On Tue, Dec 15, 2009 at 04:56:21PM -0800, Danek Duvall wrote: > >>>>imageplan.py: > >>>> > >>>> - line 670: what does this have to do with the license stuff? > >>>My (perhaps naive) understanding was that I needed to do a transport > >>>stats reset to ensure that download speeds during preexecute() were > >>>correct. See lines 792-795. > >>> > >>>At the very least, it looks like I could remove 792-795 since I'm > >>>doing this earlier. > >>Okay; if johansen is comfortable with this, then I have no cause for > >>concern. It just seemed out of place. > > > >Was this actually changed? The stats.reset() caught my eye in the > >sdiff, but when I went to look at the wdiff, that line didn't appear to > >be changed. The stats.reset() is there to ensure that the client > >discards the information from the metadata downloads when considering > >content downloads. It's probably safe to reset any time before files > >are actually downloaded, but you should test that this doesn't have a > >big impact on repository/mirror selection when multiple hosts are > >configured for a publisher. > > This was an additional call to stats.reset() that I'm doing *before* > pulling license actions, but after the manifest retrieval is done.
If this is happening after manifest retrieval and you're restting the statistics again before downloading files, then this is probably fine. > I can't seem to find a happy place to download the license data > before plan execution. I'm generally happy with how this wad looks. I'm sorry if I was giving you an impression to the contrary. -j _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
