[asterisk-dev] [Code Review] 4346: Use SIPS Contact headers as prescribed by RFC 3261 (chan_sip)

[Mark Michelson]

> On Jan. 16, 2015, 8:19 a.m., wdoekes wrote:
> >

Thanks for the quick feedback!


> On Jan. 16, 2015, 8:19 a.m., wdoekes wrote:
> > /branches/11/channels/chan_sip.c, lines 13752-13755
> > <https://reviewboard.asterisk.org/r/4346/diff/1/?file=70621#file70621line13752>
> >
> >     I'd rather see this code moved/duplicated into uas_sips_contact and uac_sips_contact. Right now it looked like you missed half the criteria when looking at those functions.

I initially did that and then factored the common code out (because of DRY and all). But if it's more clear having the request URI checked in uas_sips_contact and uac_sips_contact, I can make the change.


> On Jan. 16, 2015, 8:19 a.m., wdoekes wrote:
> > /branches/11/channels/chan_sip.c, lines 14378-14380
> > <https://reviewboard.asterisk.org/r/4346/diff/1/?file=70621#file70621line14378>
> >
> >     Can this be safely removed?
> >     
> >     Does the contact get added somewhere else?
> >     Is there not a reason why it's done right after the ast_sip_ouraddrfor?

>From tracing this by hand, I could not see why this was here. build_contact() is called here in order to build the contact header on the mwi->call sip_pvt. Later, transmit_invite() is called with its init parameter set to 2, meaning that initreqprep will be called. At the end of initreqprep, the contact is rebuilt on the sip_pvt (mwi->call), overwriting what was previously created.

Between this call to build_contact() and the one in initreqprep(), I could find no references to the sip_pvt's contact field being used.

This adds up to this call to build_contact() being a no-op.

My suspicion is that this is the result of "organic" code growth in chan_sip. At one point, this may have been needed but it no longer is.


- Mark


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/4346/#review14209
-----------------------------------------------------------


On Jan. 15, 2015, 8:07 p.m., Mark Michelson wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/4346/
> -----------------------------------------------------------
> 
> (Updated Jan. 15, 2015, 8:07 p.m.)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Bugs: ASTERISK-24646
>     https://issues.asterisk.org/jira/browse/ASTERISK-24646
> 
> 
> Repository: Asterisk
> 
> 
> Description
> -------
> 
> Please see the description on /r/4345 for a description of the interoperability problem.
> 
> This patch is the chan_sip version of the patch. The chan_sip patch is a bit more weird since creation of Contact headers is not divided into dialog vs. non-dialog situations. For this patch, I have expanded the build_contact() function to base the scheme of the Contact header based on the rules of RFC 3261 sections 8.1.1.8 and 12.1.1. To do this, I have added two parameters so that it is clear whether the Contact header is being created due to a response to an incoming request or if the Contact is being added to a created outgoing request.
> 
> 
> Diffs
> -----
> 
>   /branches/11/channels/chan_sip.c 428405 
> 
> Diff: https://reviewboard.asterisk.org/r/4346/diff/
> 
> 
> Testing
> -------
> 
> This patch has been tested by the reporter of ASTERISK-24646 and has been confirmed to solve the interop issue.
> 
> 
> Thanks,
> 
> Mark Michelson
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20150116/99940ba7/attachment.html>