Bart Smaalders wrote:
> http://cr.opensolaris.org/~barts/16066/
This all looks perfectly correct, though I agree with Shawn that the
refcount lists should be released when we're done with them. Some nits,
though ...
imageplan.py:
- line 113: trailing whitespace
- line 449-451: four-space indent for interior of comprehension, and
return to indent of 448 on 451.
- Can you explain why get_directories doesn't look just like
__get_symlinks() and __get_hardlinks()? It certainly looks to me like
they should be identical, except that directories have a small set of
hardcoded entries. I'd like to see these combined into a single
function parameterized by type, but that's not this wad.
- It might be worth a quick comment why you're unique_links isn't just a
set; you're using the side-effect of creating a dict to ensure that
only *some* of the link actions are even up for consideration, but even
then, I'm surprised that it's not arbitrary. This is definitely
sneaky.
- Do we ever, at install time, check for duplicate links and hardlinks
which don't have the same target? Or is this currently an
importer-only trick?
Danek
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
