On 03/25/10 08:11 AM, Shawn Walker wrote:
On 03/25/10 09:16 AM, Liane Praza wrote:
I'm really embarrassed. Please review the fix for:
15356 fix for 15346 breaks just_these_pkgs
at:
http://cr.opensolaris.org/~lianep/pkg-re/
The fix is as minimal as possible, since I'm worried about introducing
other breakage. Testing looks good so far, though.
So I don't understand the comment about the list being pruned previously.
I almost thought the block was redundant last night when I looked at
this, but then I noticed that while the pkgdict gets pruned initially,
afterwards, dependencies can be added back to pkgdict again which might
cause the need for it to be pruned again.
That's the part I went back and forth on being confused about. If
you're right, if I pick two packages with a dependency, and add them
both not_these_pkgs, the dependent one will get published regardless.
So I tried this with:
system/library/c++/sunpro archiver/unrar
and it didn't, despite the fact that unrar has a dependency on c++/sunpro.
I fully admit that I might be very confused about this, but I don't see
where dependencies cause new additions between the two removal calls.
Hm, unless if you're talking about the incorporation dependencies?
I'm also confused about the else clause causing just_these_pkgs to be
ignored since the eliding happened before that.
The else clause at 1652 set newpkgs from pkgdict unconditionally. Which
ignored the newpkgs pruning that happened at 1640. That's how
just_these_pkgs got ignored, because if not_these_pkgs was set, we
ignored the work that just_these_pkgs did.
liane
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
