On 03/23/10 12:08 PM, Dave Miner wrote: > On 03/23/10 01:12 PM, sanjay nadkarni wrote: >> Please review the fix for: >> >> 15132 <http://defect.opensolaris.org/bz/show_bug.cgi?id=15132> >> client side AI installs can benefit from using batch pkg installs. >> The webrev is available at: >> http://cr.opensolaris.org/~nadkarni/bugfix/ >> > > im_pop.py, 569: why is putting this on stderr the right thing to do? > I could see stdout, maybe, so that it's in the detailed log, but this > seems to just add noise to the simple log. The block at 546 really > ought to be using with, I'd think, but I guess it's OK to leave it for > now. > okay > transfer_mod.py, 1209: comment here ("Changes to batching...") doesn't > quite make sense to me. > It does not make sense to me either ;-) I will fix that. > 1221-3 could use a comment on the theory. The thing that concerns me > is that the normalization for the key could be incomplete; would it be > possible to end up with a key like "pkg://opensolaris.org/entire" such > that the ordering algorithm at 1238 won't work correctly (this would > require full FMRI's to be expressed in the manifest, I guess, and I > don't recall whether that works or is prevented in some other way)? Indeed, code will not normalize the key correctly for pkgs expressed as a full FMRI. I will check and see if a manifest can have such an entry. > > 1249-52: I'd think you could turn this into simply > > pkgs = " ".join(order.values()) > That will not work, since order is a list of keys not a dictionary.
> Dave > _______________________________________________ > caiman-discuss mailing list > caiman-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
