|Summary:||ASTERISK-02214: [patch] Add ";user=phone" when INVITE contain only phone number|
|Date Opened:||2004-08-12 14:02:47||Date Closed:||2008-01-15 15:15:28.000-0600|
|Environment:||Attachments:||( 0) asterisk-1.0-RC2-chan_sip-userphone.patch|
( 1) userphone.txt
( 2) userphone2.txt
add ;user=authname to SIP INVITEs This was required for making outbound calls to PSTN gw service running Broadworks software.
For some services, we need to indicate that the user part of the URI we're calling is a phone number by adding ;user=phone to the URI.
****** ADDITIONAL INFORMATION ******
From the specs:
" To accomodate telephone addressing, the SIP specification includes a
provision to incorporate a tel: URI  telephone-subscriber
(everything following the tel: prefix) directly into the user part of
a sip: or sips: URI, by setting the "user" parameter to "phone"."
"The tel: URI telephone-subscriber can be either a global-number or a
|Comments:||By: Olle Johansson (oej) 2004-08-12 14:56:21|
Can you explain more, give any references to why the require this? I need more information. user= is normally a hint if this a phone.
Is there any reason to do this by default, as in your patch, or make it an option. Let's do some research on how to do this right! :-)
By: fwittekind (fwittekind) 2004-08-12 16:06:16
Found on page 152, 19.1.3 Example SIP and SIPS URIs
Found on page 222, 25.1 Basic Rules
user-param = "user=" ( "phone" / "ip" / other-user)
By: Olle Johansson (oej) 2004-08-12 16:11:40
user=phone is documented in several places, but I can't find anything on user equals something else...
By: Olle Johansson (oej) 2004-08-12 16:17:26
For the record, since this may be useful another time, there's a clarification on "user=phone" in
Still no mention of user=<somethingelse> anywhere except possibly Brian Rosen's proposal for user=dialstring
By: Mark Spencer (markster) 2004-08-12 19:28:55
So what's the story here? Maybe yet another option?
By: Olle Johansson (oej) 2004-08-13 01:51:03
Markster, we are still trying to find out what the Broadworks stuff requires and the usage of the user= header. As soon as we know more, we'll come up with a proposal. Stay tuned.
My guess is that we can add ";user=phone" when the username really is a phone number or a digit-only extension (DTMF "digit", including *#). We should not add it always, as in the patch.
By: fwittekind (fwittekind) 2004-08-13 13:05:21
";user=phone" is good enough for broadworks to accept.
I uploaded a new version, that only adds ";user=phone" iff the username portion of the SIP URI matches a phone number or a digit-only extension.
Should it also be a option in sip.conf to turn it on or off?
By: Olle Johansson (oej) 2004-08-14 06:54:02
From IETF proceedings:
" agreed: the presence of the user=phone parameter implies that the user part conforms to the specification of the tel: URI."
Also, some SIP user names are numbers-only without being a tel: uri phone number.
To do this right, this would have to be an option to DIAL. We can however make it an option for a peer, saying that "everything we send to this peer is a phone number, since it's my PSTN gateway (provided that it consists of digits only, and DTMF allowed characters)"
Conclusion: I would add a peer option to your patch, enabling it only for selected peers. Otherwise we could potentially cause problems.
By: Olle Johansson (oej) 2004-08-14 07:54:34
* Allows + as first character according to Tel uri: rfc
* Adds config option for [general] and [peer]
usereqphone = yes | no
Yes means that *if* the username part is a valid tel uri, we add ;user=phone to the uri before sending it to the proxy. Use this for proxies (pstn providers and gateways) if it's required by the provider.
Fwittekind: Please test and report if this works for you.
By: fwittekind (fwittekind) 2004-08-16 14:02:47
sip debug shows ;user=phone missing from first INVITE packet, yet sip show peer shows UserEqPhone : Yes
Minor bug fix required, added r->usereqphone = p->usereqphone; to create_addr function.
By: Olle Johansson (oej) 2004-08-16 14:53:30
Great. Added userphone2.txt includes your patch to my patch to your patch, as well as the changes to sip.conf.sample.
Ready for Mark and possibly CVS integration.
Fwittekind: Thank you for working with us to solve this!
By: Mark Spencer (markster) 2004-08-31 15:19:38
Did the recent SIP contact changes help this in any way?
By: Olle Johansson (oej) 2004-08-31 15:21:32
Mark, they're not related.
By: Olle Johansson (oej) 2004-09-05 13:56:23
Mark, find me on the IRC to discuss this patch.
By: Mark Spencer (markster) 2004-10-14 10:49:23
Okay, your turn, you find me!
By: Olle Johansson (oej) 2004-10-15 02:34:17
By: twisted (twisted) 2004-10-27 17:16:43
How's the hide-and-seek game going?
By: Olle Johansson (oej) 2004-11-11 15:31:57.000-0600
Ping, Markster. If you need to discuss this more, find me on IRC
By: Olle Johansson (oej) 2004-11-14 02:56:09.000-0600
By: Olle Johansson (oej) 2004-11-21 04:17:08.000-0600
Updated patch to current CVS head
By: Mark Spencer (markster) 2004-12-02 18:31:05.000-0600
Added to CVS, thanks olle!
By: Russell Bryant (russell) 2004-12-02 19:13:39.000-0600
not in 1.0
By: Digium Subversion (svnbot) 2008-01-15 15:15:28.000-0600
r4372 | markster | 2008-01-15 15:15:28 -0600 (Tue, 15 Jan 2008) | 2 lines
Add user=phone option (bug ASTERISK-2214, thanks oej)