On 06/11/12 15:52, Brock Pytlik wrote:
On 06/11/12 15:45, Shawn Walker wrote:
On 06/11/12 15:10, Brock Pytlik wrote:
On 06/11/12 15:07, Shawn Walker wrote:
On 06/04/12 17:34, Brock Pytlik wrote:
Webrev:
https://cr.opensolaris.org/action/browse/pkg/bpytlik/7173699-v1/webrev/
Bug:
7173699 TestPkgInstallActions.test_legacy fails on SPARC
This passes on x86 and sparc.
src/tests/cli/t_pkg_install.py:
line 2851: only_legacy10 seems completely unused in our test suite
and if it was used, it would fail for the same reason you're fixing.
Perhaps just remove this and the references to it and check again?
Also, I wonder if we were intentionally testing usage of 'arch'.
If we were, there's not indication or documentation of it in either the
test or the bugs that the changeset that introduced the lines fixed.
Actually, test_uninstall_without_perms uses only_legacy10 (at least in
the copy of the gate I have on my system at the moment). I'm happy to
remove the variants from that package as well though, even though it's
not causing any issues at the moment.
Well, it publishes the package, but it doesn't actually *use* it as
far as I can see.
Well, on line 3659 it puts it into the list "pkg_list." As far as I can
see, it's never removed from that list. On line 3692, it iterates over
pkg_list and installs and uninstalls the packages in the list.
Presumably, it also participates in the "manually check deps" part of
the test (though I'm struggling to understand what that's actually
testing) as well as the updates on lines 3722, 3727, and 3729. While I
doubt the legacy package is a critical part of any of these tests, it
certainly seems to me like it's being *used*.
What have I missed?
Huh. Sorry, I missed that part.
If you don't killing the arch=i386 then, I'd just leave the rest of it
alone.
-Shawn
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
