t_variant:
Why the change from import p.v as v to from p import v?
Nits:
dependencies.py:
447,453: no empty comment lines please
462, 463 4 space indent please
t_variant.py:
538: extra blank line
empty comment lines: 1225, 1228, 1233, 1236, 1243,1245
t_pkgdep_resolve.py:
39: We usually put from X import Y statements after the other import lines
Other than that, LGTM.
Brock
On 05/11/10 05:24 PM, Richard Lowe wrote:
Please review:
15647 pkgdepend generates duplicate dependencies if dependency can't
be satisfied for all variant values
Webrev:
http://richlowe.net/~richlowe/pkg_15647/
(I'll put it on cr, too, when that becomes accessible again)
Please pay extra special attention, I keep feeling as if I'm missing
something (possibly something obvious). The root cause is as described
in the bug report:
http://defect.opensolaris.org/bz/show_bug.cgi?id=15647
The changes made are to limit the variants we attempt to satisfy for in
the second pass to those unsatisfied after the first. Changes to
VariantSets to return a VariantSets from get_satisfied()/get_unsatisfied()
were made to make this easier, with groups() added to preserve the style
of return the rest of the code expects (this only seems to affect code
involved in dependency calculation).
Tests were added for various cases of unsatisfied dependency in t_pkgdep
at Brock's request (I also adjusted the test for #14632 to no longer
depend on trailing whitespace, I got sick of breaking it),
get_satisfied() and get_unsatisfied() in t_variant, and this bug itself
in t_pkgdep_resolve, the test in t_variant, especially, leads me to
think I maybe doing something stupid.
-- Rich
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
