[asterisk-dev] [Code Review] Resolve remaining issues with SIP object matching in trunk and 1.6.1
Russell Bryant
russell at digium.com
Fri Feb 27 15:54:23 CST 2009
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviewboard.digium.com/r/172/
-----------------------------------------------------------
(Updated 2009-02-27 15:54:23.425552)
Review request for Asterisk Developers.
Changes
-------
Add a couple of updates to the patch:
1) Remove unused ispeer parameter from build_peer()
2) Fix a problem that lmadsen found when defining a type=peer and type=user with the same name in different configuration entries.
Summary
-------
After the merge of the "kill-the-user" patch to remove the sip_user object, various problems with object matching were introduced. A number of problems were found and reported on the bug tracker and have since been fixed. This patch aims to fix the remaining issues that exist so that we can proceed with the release process for Asterisk 1.6.1.
1) find_peer() and its uses have been updated to explicitly state whether peers, users, or both object types should be matched when doing a match by name.
2) The calls to find_peer() have been audited and changed as necessary.
There is one place where the code we have here differs in what was documented in team/oej/sip-compliance/sipobjects.txt. The difference is that registrations _only_ match peers, and not users. This is to match the previous behavior. Also, registration processing uses the To header to get the name, and not the From header as documented.
This addresses bug 14505.
http://bugs.digium.com/view.php?id=14505
Diffs (updated)
-----
/trunk/channels/chan_sip.c 179167
Diff: http://reviewboard.digium.com/r/172/diff
Testing
-------
Josh and I have both tested inbound and outbound calls with peer, user, and friend configurations and have seen the expected behavior. I have also tested registrations and subscriptions, both of which appear to be working as expected.
Thanks,
Russell
More information about the asterisk-dev
mailing list