On 02/27/12 11:45, Danek Duvall wrote:
Shawn Walker wrote:
On 02/24/12 14:52, Danek Duvall wrote:
Shawn Walker wrote:
https://cr.opensolaris.org/action/browse/pkg/swalker/pkg-7145863/webrev/
...
manifest.py:
...
- line 609: why just for hardlinks now?
I've added a comment here already, but it's because all actions except
hardlinks have had the leading '/' stripped from paths for quite some time
now during action __init__. Why we never stripped the leading '/' from
paths for hardlinks is a historical mystery at this point. How much it
actually matters since we have stripped them from the manifests that were
written out for a very long time also remains uncertain.
I think your changeset ab6a2324f73a misunderstood what was going on w.r.t.
hardlinks -- all paths were getting stripped up until that point, but since
hardlinks are subclassed from links, there wasn't any code in hardlink.py
to do that work. Sorry I didn't catch that back then.
Ah, yes, it now makes sense. Thanks for sourcing that. Yay; I've
deleted several bits of code thanks to this.
We could probably also torch 601-604, per the comment on line 597. If
nothing else, setting attrs and aname should come before the comment, which
only applies to 601-604.
So the problem is that I can't remove this logic unless we want to
de-support upgrading older format images. We can definitely remove it
for S12, but I was uncertain whether I should do so for U1.
If you think I should keep it, I'll update the comment to reflect why
its really being kept now.
I'll run the test suite again and then send out a diff of all the review
changes so far.
Thanks,
-Shawn
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
