Closed GoogleCodeExporter closed 9 years ago
Sorry, forget adding csipsimple revision: 0.00-15-09
Original comment by karnevil...@gmail.com
on 5 Nov 2010 at 5:52
Yes unfortunately that's a back side of pjsip behavior.
As it try to get rid of NATed network detecting your public IP (as if you
activate stun), it try to always use it's public IP for all account.
The sip stack consider you have an unique public IP and as consequence as soon
as it detect a public IP for one account it consider this IP is valid for all
account. It has advantages (speed of registration in standard cases) but also
disavantages (your use case is typically one big disavantange of this
approach).
But that's the behavior of the stack CSipSimple is relying on.
I'll check if there is not an extra option in pjsip that could help to solve
this kind of issue (if so, I'll add an option in settings).
Anyway, if you have the complete hand on your network, what I can advise to you
is :
* Assign a static dhcp lease to your mobile device inside the enterprise
* Use expert wizard for your enterprise account (you can switch from existing
wizard using menu > choose wizard)
* Use "Force contact" field and fill it with display name
<sip:username@10.0.1.xxx>
where 10.0.1.xxx is the static dhcp lease.
I don't know if it will help for via field but at least for contact it will
solve your problem.
That's a temporary hack.
If your need is to be registered only with your enterprise ip-pbx on your local
network and with the public on the public network, another more user friendly
solution would be something that activate/deactivate account according to the
current android setting/configuration state.
It has been discussed in another issue here (about intent to
activate/deactivate accounts and Locale application)
So open point for me : I'll check if there is any way to bypass this behavior
in pjsip.
Else will probably linked to the issue about activation/deactivation of account
according device state (if the use case is OK for you, and if you don't
actually need both account to be registered while in your local network... but
I guess that in this case your local sip server can be a good sip proxy for
your requests to another sip server).
Original comment by r3gis...@gmail.com
on 6 Nov 2010 at 8:50
Ok, this is not big problem for us now, it's only an add-on as my main target
is using it in our enterprise ip-pbx (and I don't exclude that in future I
could publish it on internet in order that our terminals could get connected to
it also around the world).
Original comment by karnevil...@gmail.com
on 8 Nov 2010 at 1:19
Issue 536 has been merged into this issue.
Original comment by r3gis...@gmail.com
on 24 Dec 2010 at 5:38
Seems to be taken into account for pjsip-2.0.
So CSipSimple will automatically benefit it as soon as I'll move to pjsip-2.0 :
For reference ticket : http://trac.pjsip.org/repos/ticket/539
It will also improve battery life according to the ticket :)
Original comment by r3gis...@gmail.com
on 8 Jan 2011 at 1:28
That's good, but, about battery life, the main problem is the partial wake-lock.
After weeks of testing, definitively it's the big difference in standby battery
life.
Without partial wake-lock, standby time is not so less by having CSipSimple
active or not, for usual activity of the phone I don't see difference. But in
this case the problem is that Register timeout is not honoured, so looking at
asterisk log, phone will re-register with random time, most of the time when
timeout is expired (thus it doesn't answer to incoming call).
With partial wake-lock I have to recharge each day, also with few minutes of
calls. But in this case register timeout is honoured, thus always answer
incoming call.
No other functionality are affected by this settings (in all my testing).
Original comment by karnevil...@gmail.com
on 10 Jan 2011 at 5:23
Issue 481 has been merged into this issue.
Original comment by r3gis...@gmail.com
on 5 Mar 2011 at 1:17
Issue 774 has been merged into this issue.
Original comment by r3gis...@gmail.com
on 7 Mar 2011 at 10:47
Issue 820 has been merged into this issue.
Original comment by r3gis...@gmail.com
on 22 Mar 2011 at 12:20
Issue 1134 has been merged into this issue.
Original comment by r3gis...@gmail.com
on 9 Jul 2011 at 7:59
Issue 1251 has been merged into this issue.
Original comment by r3gis...@gmail.com
on 2 Sep 2011 at 10:12
If this bug tracking system allows to edit the subject of an issue, this one
should be change to contain STUN in its subject. There are duplicates filed
monthly, and I for sure know that #1251 wouldn't have been filed if I found
this issue while searching for STUN ;)
Original comment by zugsch...@gmail.com
on 2 Sep 2011 at 11:05
Done :)
But no problem with the duplicate don't worry ;)
I'm also often lost with all open issues. I should take time to treat all
issues and merge duplicates/obsolete. (But for now I've not enough times to do
so, when I got little time for csipsimple, I prefer to actually produce code ;)
).
Original comment by r3gis...@gmail.com
on 2 Sep 2011 at 12:32
Issue 1351 has been merged into this issue.
Original comment by r3gis...@gmail.com
on 29 Oct 2011 at 6:50
For reference, a new more precise ticket has been created on pjsip project
about this particular point for pjsip-2.0 :
https://trac.pjsip.org/repos/ticket/1412
BTW, I've already a working patch for this special use case, but I will wait
for pjsip implementation in their official 2.0 release.
A release of csipsimple will be done soon based on pjsip 1.x with current trunk
(which is freezed and waits for stability feedback). Once this release will be
done, the nightly builds will be based on pjsip-2.0 and it will be possible to
have either what was added by pjsip guys or my patch.
Original comment by r3gis...@gmail.com
on 5 Nov 2011 at 11:12
Issue 1447 has been merged into this issue.
Original comment by r3gis...@gmail.com
on 23 Dec 2011 at 1:30
Issue 1464 has been merged into this issue.
Original comment by r3gis...@gmail.com
on 1 Jan 2012 at 9:31
Hello,
Same behaviour here with a Freebox SIP account (public IP) and a Aastra Nexspan
(private network).
In my case it seems that if the private IP account registers first, it works.
Original comment by laurent....@gmail.com
on 8 Jan 2012 at 4:21
Yes indeed. The sip stack as only one "global" ip address. So if the first
account to register is the private one it will work for first registration.
However, re-registration will be broken then.
I really hope pjsip guys will take the point into account.
Original comment by r3gis...@gmail.com
on 8 Jan 2012 at 4:34
For those experiencing the problem, I have had success buy turning on NAT
support for the account on the *server* on the private network (or all servers).
This sounds backwards as there is no NAT going on for that link, but it makes
the server ignore the Via header and use the packet's origin so allows multiple
accounts to work until the issue is fixed!
Original comment by SpiderErrol
on 14 Mar 2012 at 1:00
The https://trac.pjsip.org/repos/ticket/1412 issue is now 5 months old. I can't
see any process in https://trac.pjsip.org/repos/log/pjproject :(
Is there a chance that it'll be solved soon?
Original comment by szims...@gmail.com
on 19 Mar 2012 at 11:21
You should ask on pjsip dev mailing list ;).
I personally have no news about this enhancement request we did.
If I get some time to propose a patch to pjsip guys I'll do, but I'm more
focused on stuff on CSipSimple/android side for now.
So we wait for pjsip developers to implement that, and then csipsimple (and all
other apps based on pjsip) will automatically have the feature. (I'll have some
low effort to do to reflect the setting in the expert wizard, but can be done
quickly once the feature is in the sip stack).
Original comment by r3gis...@gmail.com
on 19 Mar 2012 at 11:57
I sent a mail to them ;)
Original comment by szims...@gmail.com
on 19 Mar 2012 at 12:33
Good news, it's now in pjsip. (cf https://trac.pjsip.org/repos/changeset/4218)
I just started the work to integrate that in csipsimple user interface. Will
land soon :)
Original comment by r3gis...@gmail.com
on 7 Aug 2012 at 1:06
This issue was closed by revision r1763.
Original comment by r3gis...@gmail.com
on 7 Aug 2012 at 6:13
Original issue reported on code.google.com by
karnevil...@gmail.com
on 5 Nov 2010 at 4:55