On 01/31/13 10:16 AM, Shawn Walker wrote:
On 01/31/13 09:47, Erik Trauschke wrote:
On 01/31/13 09:37 AM, Bart Smaalders wrote:
This looks correct, Shawn. I take it the original tests failed
as expected when frozenset was changed to tuple?
Why is that changed to tuple from set anyway? If I understand this
correctly, akey pretty much represents a path, so how can there be
duplicates?
Since frozenset() doesn't guarantee order, it's possible that a path may
match a mediator-implementation value and vice versa. With that said,
I'm unable to construct a scenario in which this change actually helps.
However, it seems to me that dictify() output should be consistent. That
is, always produce the same output given the same input since it is
essentially a "hashing" function of sorts.
-Shawn
I think the frozenset actually hid the attrs[key_attribute] bug.. with
different keys, it would have hidden the "path" vs ["path"] bug.
Two link actions in a manifest with the same mediators should have
always been collapsed into one, but the frozenset of three items
would make this fail much less often.
- Bart
--
Bart Smaalders Solaris Core OS
bart.smaald...@oracle.com http://blogs.oracle.com/barts
"You will contribute more with Mercurial than with Thunderbird."
"Civilization advances by extending the number of important
operations which we can perform without thinking about them."
_______________________________________________
pkg-discuss mailing list
pkg-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
