[asterisk-dev] [Code Review] SIP INFO DTMF non-numeric codes treated as '1'

[Matt Jordan]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/1722/
-----------------------------------------------------------

(Updated Feb. 9, 2012, 9:04 a.m.)


Review request for Asterisk Developers, Mark Michelson and rmudgett.


Changes
-------

Addressed rmudgett's findings.  Retested with test in https://reviewboard.asterisk.org/r/1723/, which now tests application/dtmf and application/dtmf-relay, since the handling of both DTMF signal delivery types was combined.  Test passes, parsing the contents of the message body correctly and raising the appropriate DTMF events.


Summary
-------

In ASTERISK-18924, SIP INFO DTMF handling was changed to account for both lowercase alphabetic DTMF events, as well as uppercase alphabetic DTMF events.  When this occurred, the comparison of the character buffer containing the event code was changed such that the buffer was first compared against '0' and '9' to determine if it was numeric.  Unfortunately, since the first character in the buffer will typically be '1' in the case of event codes 10, 11, 16, and (if non-alphabetic) 12-15 (for 'A' - 'D'), this caused those codes to be converted to a DTMF event of '1'.

Pressing '#' in voicemail resulted in hilarity.

This patch changes this to attempt to parse the DTMF codes first as '*', '#', '!', and 'A'/'a' - 'D'/'d'.  If those fail, it then converts the character buffer to an integer value; if this is invalid a 200 OK is sent and no DTMF tone is generated.

I didn't change the sending of a 200 OK for an invalid DTMF event sent from a UA, as that was the previous behavior.


This addresses bug ASTERISK-19290.
    https://issues.asterisk.org/jira/browse/ASTERISK-19290


Diffs (updated)
-----

  /branches/10/channels/chan_sip.c 354428 

Diff: https://reviewboard.asterisk.org/r/1722/diff


Testing
-------

Test written for the Asterisk Test Suite (see review https://reviewboard.asterisk.org/r/1723/).  Without this patch the test fails, with it - great success.


Thanks,

Matt

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20120209/08c95200/attachment.htm>