On 04/25/11 05:39 PM, Shawn Walker wrote:
----- [email protected] wrote:
Here's the second round of webrev for the sysrepo work. We'd like to
land this soon, so my plan is to put this back Wednesday unless I hear
a
reason not to. If you need more time to check or look at something,
please let me know and I can wait.
Webrev:
http://cr.opensolaris.org/~bpytlik/ips-sysrepo-v2/
src/modules/client/imageconfig.py:
nit: two lines between class definitions per PEP8
Sure.
line 835: don't you mean != ?
If not a == b is the same as if a != b, right? But I'm happy to change
it to use !=.
src/sysrepo.py:
lines 209-221: So one important thing to be aware of is that when a path is
converted to a URI, characters such as '%' and so on will be encoded. This
means you can't use the path component returned from urlparse as-is. So on
line 210, you should get the path using repo_uri.get_pathname() which will
properly unescape it. You can test this by using a directory named
'/tmp/a%b'
as an example. Once you've done that, you can get rid of the urlparse
usage.
Tim knows about this and as he's on vacation, I'll let him address this
as a follow up bug fix if needed.
lines 280-294: Does this still work if no ipv6 interface is configured? It's
trivial to configure nwam as such. If not, it should at least fail
gracefully.
See above.
src/modules/client/api.py:
line 3160: This seems weird to me. d() contains a set of tuples
of (prefix, pub_obj) where names is just a set of names. Are you
sure this works as expected?
I agree this does look screwy. I'll look into it and fix it if necessary
before putback.
src/modules/client/publisher.py:
lines 290-299: nit: if you'd put these in alpha order, that'd be swell
Sure.
I tried to look at all of the files that are core to the client.
Thanks,
Brock
Thanks,
-Shawn
-Shawn
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
