[sip-comm-dev] Adapted no-registrar patch


#1

Hello all,

I've managed to manually apply an older patch made by Michael Koch to a more recent version of SC sources, the patch being intended to provide no-registrar SIP call availability. This is useful among others for testing the currently developed ZRTP integration on point-to-point communication in what I'm concerned (but probably it is also of general interest).

Some notes about the current patch version:

- it is done against the current SC encryption branch but it has nothing to do with the encryption related sources so it will be probably compatible with the main SC trunk also (anyway, the encryption branch is a bit behind the main trunk; I'll update it this week, make some patch changes, if necessary, and come back with them)

- the patch is only a manual adaptation of Michael Koch's original sources with minor modifications as said, because the older patch didn't match anymore; the main idea was to split the ProtocolProviderServiceSipImpl into two classes, derived from an Abstract one, one class for registrar mode, and one for no-registrar mode; the choice of instancing was based on an option for the AccountID, which as Michael said at that moment should be probably integrated in the account registration wizard; for now this option is hardcoded and attached to every account for testing purposes, so the patch will cause all calls to go on without a registrar (like I said I adapted the patch mainly for testing purposes for now)

- in order for the patch to work don't forget to remove the proxy fields automatically completed at the account creation ( or at least don't let it completed with the local IP address which you'll probably have used for the last part of the SIP id; I've forgot the fields set initially and took me almost three days to figure out what was wrong :slight_smile: )

- as a final note: the patch isn't yet extensively tested on the current sources so I don't know what bugs can contain - if I find any I'll come back with a mail regarding the issues; I attach the patch in two formats to this mail one is a .patch version and one is a zip archive with the modified sources; I remind again that the patch is done against the encryption branch on the SC trunk (which I'll update soon) so for applying the patch to the main trunk you should take care not to overwrite any newer source modifications

Regards,
Emanuel

no-registrar.zip (123 KB)

no-registrar.patch (187 KB)

···

--
--------------------------------------------------------------------

"Life is full of unexpected but nothing happens without a reason..."


#2

Hello Emanuel,

Thanks for sending this over.

This patch has been available for a while and there have been numerous
requests for support of a registrar-less mode of operation. It is
probably high time we integrated it into the application rather than
having to maintain a synced version.

Is anyone interested in applying it?

Cheers
Emil

Onica S. Emanuel написа:

···

Hello all,

I've managed to manually apply an older patch made by Michael Koch to a
more recent version of SC sources, the patch being intended to provide
no-registrar SIP call availability. This is useful among others for
testing the currently developed ZRTP integration on point-to-point
communication in what I'm concerned (but probably it is also of general
interest).

Some notes about the current patch version:

- it is done against the current SC encryption branch but it has
nothing to do with the encryption related sources so it will be
probably compatible with the main SC trunk also (anyway, the encryption
branch is a bit behind the main trunk; I'll update it this week, make some
patch changes, if necessary, and come back with them)

- the patch is only a manual adaptation of Michael Koch's original
sources with minor modifications as said, because the older patch didn't
match anymore; the main idea was to split the ProtocolProviderServiceSipImpl
into two classes, derived from an Abstract one, one class for registrar
mode, and one for no-registrar mode; the choice of instancing was based on
an option for the AccountID, which as Michael said at that moment should
be probably integrated in the account registration wizard; for now this
option is hardcoded and attached to every account for testing purposes, so
the patch will cause all calls to go on without a registrar (like I said I
adapted the patch mainly for testing purposes for now)

- in order for the patch to work don't forget to remove the proxy fields
automatically completed at the account creation ( or at least don't let it
completed with the local IP address which you'll probably have used for
the last part of the SIP id; I've forgot the fields set initially and took
me almost three days to figure out what was wrong :slight_smile: )

- as a final note: the patch isn't yet extensively tested on the current
sources so I don't know what bugs can contain - if I find any I'll come
back with a mail regarding the issues; I attach the patch in two formats
to this mail one is a .patch version and one is a zip archive with the
modified sources; I remind again that the patch is done against the
encryption branch on the SC trunk (which I'll update soon) so for applying
the patch to the main trunk you should take care not to overwrite any
newer source modifications

Regards,
Emanuel

------------------------------------------------------------------------

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail: dev-help@sip-communicator.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail: dev-help@sip-communicator.dev.java.net


#3

Hi Emil!

This patch has been available for a while and there have been numerous
requests for support of a registrar-less mode of operation. It is
probably high time we integrated it into the application rather than
having to maintain a synced version.

Is anyone interested in applying it?

I could do it, if noone else who is more involved in SIP Communicator
development volunteers. I won't have time to adapt the account creation
wizards, though. Emanuel, thanks for updating the patch. I'm happy that you
found it useful.

Cheers
Michael Koch

···

-----Ursprüngliche Nachricht-----
Von: Emil Ivov [mailto:emcho@sip-communicator.org]
Gesendet: Freitag, 4. Juli 2008 00:05
An: dev@sip-communicator.dev.java.net
Betreff: Re: [sip-comm-dev] Adapted no-registrar patch

Hello Emanuel,

Thanks for sending this over.

Cheers
Emil

Onica S. Emanuel написа:
> Hello all,
>
> I've managed to manually apply an older patch made by
Michael Koch to a
> more recent version of SC sources, the patch being intended
to provide
> no-registrar SIP call availability. This is useful among others for
> testing the currently developed ZRTP integration on point-to-point
> communication in what I'm concerned (but probably it is
also of general
> interest).
>
> Some notes about the current patch version:
>
> - it is done against the current SC encryption branch but it has
> nothing to do with the encryption related sources so it will be
> probably compatible with the main SC trunk also (anyway,
the encryption
> branch is a bit behind the main trunk; I'll update it this
week, make some
> patch changes, if necessary, and come back with them)
>
> - the patch is only a manual adaptation of Michael Koch's original
> sources with minor modifications as said, because the older
patch didn't
> match anymore; the main idea was to split the
ProtocolProviderServiceSipImpl
> into two classes, derived from an Abstract one, one class
for registrar
> mode, and one for no-registrar mode; the choice of
instancing was based on
> an option for the AccountID, which as Michael said at that
moment should
> be probably integrated in the account registration wizard;
for now this
> option is hardcoded and attached to every account for
testing purposes, so
> the patch will cause all calls to go on without a registrar
(like I said I
> adapted the patch mainly for testing purposes for now)
>
> - in order for the patch to work don't forget to remove the
proxy fields
> automatically completed at the account creation ( or at
least don't let it
> completed with the local IP address which you'll probably
have used for
> the last part of the SIP id; I've forgot the fields set
initially and took
> me almost three days to figure out what was wrong :slight_smile: )
>
> - as a final note: the patch isn't yet extensively tested
on the current
> sources so I don't know what bugs can contain - if I find
any I'll come
> back with a mail regarding the issues; I attach the patch
in two formats
> to this mail one is a .patch version and one is a zip
archive with the
> modified sources; I remind again that the patch is done against the
> encryption branch on the SC trunk (which I'll update soon)
so for applying
> the patch to the main trunk you should take care not to
overwrite any
> newer source modifications
>
> Regards,
> Emanuel
>
>
>
>
>
--------------------------------------------------------------
----------
>
>
---------------------------------------------------------------------
> To unsubscribe, e-mail:
dev-unsubscribe@sip-communicator.dev.java.net
> For additional commands, e-mail:
dev-help@sip-communicator.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail:
dev-help@sip-communicator.dev.java.net


#4

Hello,

First some issues of more general interest: I've merged the svn encryption branch today with the main trunk head revision, and modified the patch accordingly (it is still done against the encryption branch, but like I said in the last mail it has nothing to do with the encryption related sources so now it should be compatible with the main trunk also; I've attached it again in .patch format and as a zip archive with the modified sources).

Like last time I didn't manage yet to do extensive testing of the patch. Actually, I got a strange behaviour at some point = if the peer that responded to the call was also the one closing the call, the other peer - the call initiator - still remained connected. Anyway, I think (hope) that it doesn't anything to do with the patch (and might be related with some messed up network connections), because after a reboot of both testing peers all went fine; and I specifically tried several times to "achieve" the same behaviour and I didn't get it.

More specific issues related to ZRTP: Unfortunately at this point there is a problem. The issue I pointed out in some previous mails that the ZRTP exchange hangs in the DH phase periodically seems to appear much more often with the current patch. Actually it's a bit weird because calling from the notebook I'm using as a testing peer to my PC makes it hang, but from PC to my notebook doesn't.

So, right before writing this mail I've tried to trace a bit the issue (it is "stable" unfortunately.. for now at least - appearing every time now in the no-registrar mode). All I can tell now is that the point where the exchange hangs seems to be when the ZRTP initiator prepares to send DH part 2 after receiving DH part 1. More exactly it doesn't pass line 959 - cancelTimer(); - inside evCommitSent method from the ZrtpStateClass.

I'm writing quite late this mail :slight_smile: and I didn't get the time to investigate it deeper. I should take an extensive look on that part of Werner's sources to see if I can figure it out. At a first look it seems to be probably related with a timer that expires before the DH2 is sent, but it's just an assumption and I might be wrong. Anyway, like I said, this seems to happen now after applying the no-registrar patch, more often in this mode (I repeat: in one direction, which is weird, because the sources are the same of course on both the peers; I even remade the testing accounts). In the registrar mode I didn't get the problem today, even after merging with the main branch (but it did appear during the last weeks from time to time like I pointed out in my mails). So, I should probably take a look also at the patch I've adapted to see if I can find any connection with the issue.

Regards,
Emanuel

no-registrar-updated.zip (127 KB)

no-registrar-updated.patch (182 KB)

···

--
--------------------------------------------------------------------

"Life is full of unexpected but nothing happens without a reason..."


#5

Hey Michael,

Thanks for volunteering and res please! I guess we'll worry about the
wizard later.

Cheers
Emil

···

On Fri, Jul 4, 2008 at 11:42 AM, Koch Michael <MKoch@rowa.de> wrote:

Hi Emil!

This patch has been available for a while and there have been numerous
requests for support of a registrar-less mode of operation. It is
probably high time we integrated it into the application rather than
having to maintain a synced version.

Is anyone interested in applying it?

I could do it, if noone else who is more involved in SIP Communicator
development volunteers. I won't have time to adapt the account creation
wizards, though. Emanuel, thanks for updating the patch. I'm happy that you
found it useful.

Cheers
Michael Koch

-----Ursprüngliche Nachricht-----
Von: Emil Ivov [mailto:emcho@sip-communicator.org]
Gesendet: Freitag, 4. Juli 2008 00:05
An: dev@sip-communicator.dev.java.net
Betreff: Re: [sip-comm-dev] Adapted no-registrar patch

Hello Emanuel,

Thanks for sending this over.

Cheers
Emil

Onica S. Emanuel написа:
> Hello all,
>
> I've managed to manually apply an older patch made by
Michael Koch to a
> more recent version of SC sources, the patch being intended
to provide
> no-registrar SIP call availability. This is useful among others for
> testing the currently developed ZRTP integration on point-to-point
> communication in what I'm concerned (but probably it is
also of general
> interest).
>
> Some notes about the current patch version:
>
> - it is done against the current SC encryption branch but it has
> nothing to do with the encryption related sources so it will be
> probably compatible with the main SC trunk also (anyway,
the encryption
> branch is a bit behind the main trunk; I'll update it this
week, make some
> patch changes, if necessary, and come back with them)
>
> - the patch is only a manual adaptation of Michael Koch's original
> sources with minor modifications as said, because the older
patch didn't
> match anymore; the main idea was to split the
ProtocolProviderServiceSipImpl
> into two classes, derived from an Abstract one, one class
for registrar
> mode, and one for no-registrar mode; the choice of
instancing was based on
> an option for the AccountID, which as Michael said at that
moment should
> be probably integrated in the account registration wizard;
for now this
> option is hardcoded and attached to every account for
testing purposes, so
> the patch will cause all calls to go on without a registrar
(like I said I
> adapted the patch mainly for testing purposes for now)
>
> - in order for the patch to work don't forget to remove the
proxy fields
> automatically completed at the account creation ( or at
least don't let it
> completed with the local IP address which you'll probably
have used for
> the last part of the SIP id; I've forgot the fields set
initially and took
> me almost three days to figure out what was wrong :slight_smile: )
>
> - as a final note: the patch isn't yet extensively tested
on the current
> sources so I don't know what bugs can contain - if I find
any I'll come
> back with a mail regarding the issues; I attach the patch
in two formats
> to this mail one is a .patch version and one is a zip
archive with the
> modified sources; I remind again that the patch is done against the
> encryption branch on the SC trunk (which I'll update soon)
so for applying
> the patch to the main trunk you should take care not to
overwrite any
> newer source modifications
>
> Regards,
> Emanuel
>
>
>
>
>
--------------------------------------------------------------
----------
>
>
---------------------------------------------------------------------
> To unsubscribe, e-mail:
dev-unsubscribe@sip-communicator.dev.java.net
> For additional commands, e-mail:
dev-help@sip-communicator.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail:
dev-help@sip-communicator.dev.java.net


#6

Emanuel,

thanks for digging into the problem. Because of your description
I revised the code of the timeout class in ZRTPTransformEngine. This
is a _very_ simple timeout helper class, just good enough to support
the simple ZRTP timeout schema - however, depending on the network
delays, speed of the system, etc there may be a race condition. I
updated the code to run the timeout executor function outside the
sync lock to avoid this (did I mention that this is a really _very_
simple timeout class :slight_smile: )

I've checked in the fixes and also updated to protocol version number
in ZrtpConstants class.

Please note that the repository location changed: we now use SVN
instead of CVS. The new repository's address is:

svn co svn://svn.savannah.gnu.org/ccrtp/trunk/ccrtp/contributions/ZRTP4J

   or

svn co http://svn.savannah.gnu.org/svn/ccrtp/trunk/ccrtp/contributions/ZRTP4J

Regards,
Werner

Onica S. Emanuel schrieb:

···

Hello,

First some issues of more general interest: I've merged the svn encryption branch today with the main trunk head revision, and modified the patch accordingly (it is still done against the encryption branch, but like I said in the last mail it has nothing to do with the encryption related sources so now it should be compatible with the main trunk also; I've attached it again in .patch format and as a zip archive with the modified sources).

Like last time I didn't manage yet to do extensive testing of the patch. Actually, I got a strange behaviour at some point = if the peer that responded to the call was also the one closing the call, the other peer - the call initiator - still remained connected. Anyway, I think (hope) that it doesn't anything to do with the patch (and might be related with some messed up network connections), because after a reboot of both testing peers all went fine; and I specifically tried several times to "achieve" the same behaviour and I didn't get it.

More specific issues related to ZRTP: Unfortunately at this point there is a problem. The issue I pointed out in some previous mails that the ZRTP exchange hangs in the DH phase periodically seems to appear much more often with the current patch. Actually it's a bit weird because calling from the notebook I'm using as a testing peer to my PC makes it hang, but from PC to my notebook doesn't.

So, right before writing this mail I've tried to trace a bit the issue (it is "stable" unfortunately.. for now at least - appearing every time now in the no-registrar mode). All I can tell now is that the point where the exchange hangs seems to be when the ZRTP initiator prepares to send DH part 2 after receiving DH part 1. More exactly it doesn't pass line 959 - cancelTimer(); - inside evCommitSent method from the ZrtpStateClass.

I'm writing quite late this mail :slight_smile: and I didn't get the time to investigate it deeper. I should take an extensive look on that part of Werner's sources to see if I can figure it out. At a first look it seems to be probably related with a timer that expires before the DH2 is sent, but it's just an assumption and I might be wrong. Anyway, like I said, this seems to happen now after applying the no-registrar patch, more often in this mode (I repeat: in one direction, which is weird, because the sources are the same of course on both the peers; I even remade the testing accounts). In the registrar mode I didn't get the problem today, even after merging with the main branch (but it did appear during the last weeks from time to time like I pointed out in my mails). So, I should probably take a look also at the patch I've adapted to see if I can find any connection with the issue.

Regards,
Emanuel

------------------------------------------------------------------------

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail: dev-help@sip-communicator.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail: dev-help@sip-communicator.dev.java.net


#7

Hello,

Before talking about the subject (account wizard update), I'd like to thank Werner for the quick reply and solution to the ZRTP issue pointed out in my last mail. Indeed, moving the run function body outside the sync lock object in the timeout class seems to solve the problem (I didn't get it any more, and neither other types of ZRTP exchange hanging, in any subsequent runs until now).

Now, about the SIP account wizard update. I've took some time to look more into the no-registrar patch problem, and managed to make the modification to the account registration wizard for adding the NO_REGISTRAR option. (Essentially I've followed the same steps as for adding another checkbox based option to the account registration - the forceP2P one, except for the GUI location; hope I didn't missed something, the result is anyway the one expected = option added to account when loading it). So, now the option isn't anymore hardcoded causing all the accounts to use the no-registrar mode. This, however brought up some problems in the older patch I've adapted, as it follows.

As I said in my previous mails, I only manually added Michael Koch's patch to the latest source version, with small modifications for compatibility, without any deeper analysis or much further testing. After I've removed the hardcoded option which caused only ProtocolProviderServiceSipNoRegistrarImpl instantiations inside the ProtocolProviderFactorySipImpl, and permitted also the instantation of "registrar mode" ProtocolProviderServices, having both of the versions instantiated for two different accounts at the same time, brings up an issue I've mentioned also in a previous mail. This refers to the fact that when making a call between two "no-registrar" accounts, if the peer that answered the call is the one who hangs up, the call initiator peer remains connected. However, the initiator receives the ByeEvent, handled by a log addition inside the last branch of the if inside the update(ReceiveStreamEvent) method from the CallSessionImpl class. Anyway, after some more debugging, it seems that the respective code line isn't reached in a call which behaves/ends normally (in registrar mode or if ended by the call initiator peer), so probably it shouldn't get there neither in the case of call ended by the receiver. It's interesting however that, like I said, the problem appears only if there is a normal ("registrar mode") ProtocolProviderServiceSipImpl instantiated also for a different account. If all the accounts are using no-registrar mode (as it was before I've added the account reg wiz option) the call ends fine no matter which peer does hang up.

I didn't have yet the time to investigate further the above mentioned problems, but, as another issue, also the ProtocolProviderServiceSipImpl (registrar-mode) part of the patch seems not to be functional at all now (call fails using the patched sources in registrar mode - didn't debug it yet to see the reason, this is the last part I've just tested). So, solving this might also modify the behaviour for the call end issue in the no-registrar mode, but for now it's only an assumption. I must set up a Linux partition next days to make some ZRTP further testing (and all the hard-drive related restructuring will probably take some time), but after I'll finish with this I'll try looking further at the problem (though I'm not really a SIP expert and I think there might be some protocol issues involved also, but I'll see what I can figure out).

I attach the patch including the no-registrar mode and the account registration wizard update to this mail. I also attach a zip archive containining only the account registration wizard related sources (which adds only the addition of the NO_REGISTRAR option to the account). The modifications are done as in the last attachments against the encryption branch but as said before they have nothing to do with the encryption related sources, and the branch was updated this week from the main trunk, so the patch is probably compatible also with the main trunk too. Note however that this version of the patch wich permits registrar-mode also is, like I said, quite unstable yet ( so it's mainly for further dev-purposes, not yet for actual functionality using :slight_smile: ).

Regards,
Emanuel

Hey Michael,

Thanks for volunteering and res please! I guess we'll worry about the
wizard later.

Cheers
Emil

no-registrar-accregwiz.zip (22.3 KB)

no-registrar-with-accregwiz.patch (189 KB)

···

On Sat, 5 Jul 2008, Emil Ivov wrote:

On Fri, Jul 4, 2008 at 11:42 AM, Koch Michael <MKoch@rowa.de> wrote:

Hi Emil!

This patch has been available for a while and there have been numerous
requests for support of a registrar-less mode of operation. It is
probably high time we integrated it into the application rather than
having to maintain a synced version.

Is anyone interested in applying it?

I could do it, if noone else who is more involved in SIP Communicator
development volunteers. I won't have time to adapt the account creation
wizards, though. Emanuel, thanks for updating the patch. I'm happy that you
found it useful.

Cheers
Michael Koch

-----Urspr��ngliche Nachricht-----
Von: Emil Ivov [mailto:emcho@sip-communicator.org]
Gesendet: Freitag, 4. Juli 2008 00:05
An: dev@sip-communicator.dev.java.net
Betreff: Re: [sip-comm-dev] Adapted no-registrar patch

Hello Emanuel,

Thanks for sending this over.

Cheers
Emil

Onica S. Emanuel �ߧѧ�ڧ��:
> Hello all,
>
> I've managed to manually apply an older patch made by
Michael Koch to a
> more recent version of SC sources, the patch being intended
to provide
> no-registrar SIP call availability. This is useful among others for
> testing the currently developed ZRTP integration on point-to-point
> communication in what I'm concerned (but probably it is
also of general
> interest).
>
> Some notes about the current patch version:
>
> - it is done against the current SC encryption branch but it has
> nothing to do with the encryption related sources so it will be
> probably compatible with the main SC trunk also (anyway,
the encryption
> branch is a bit behind the main trunk; I'll update it this
week, make some
> patch changes, if necessary, and come back with them)
>
> - the patch is only a manual adaptation of Michael Koch's original
> sources with minor modifications as said, because the older
patch didn't
> match anymore; the main idea was to split the
ProtocolProviderServiceSipImpl
> into two classes, derived from an Abstract one, one class
for registrar
> mode, and one for no-registrar mode; the choice of
instancing was based on
> an option for the AccountID, which as Michael said at that
moment should
> be probably integrated in the account registration wizard;
for now this
> option is hardcoded and attached to every account for
testing purposes, so
> the patch will cause all calls to go on without a registrar
(like I said I
> adapted the patch mainly for testing purposes for now)
>
> - in order for the patch to work don't forget to remove the
proxy fields
> automatically completed at the account creation ( or at
least don't let it
> completed with the local IP address which you'll probably
have used for
> the last part of the SIP id; I've forgot the fields set
initially and took
> me almost three days to figure out what was wrong :slight_smile: )
>
> - as a final note: the patch isn't yet extensively tested
on the current
> sources so I don't know what bugs can contain - if I find
any I'll come
> back with a mail regarding the issues; I attach the patch
in two formats
> to this mail one is a .patch version and one is a zip
archive with the
> modified sources; I remind again that the patch is done against the
> encryption branch on the SC trunk (which I'll update soon)
so for applying
> the patch to the main trunk you should take care not to
overwrite any
> newer source modifications
>
> Regards,
> Emanuel
>
--------------------------------------------------------------
----------
>
---------------------------------------------------------------------
> To unsubscribe, e-mail:
dev-unsubscribe@sip-communicator.dev.java.net
> For additional commands, e-mail:
dev-help@sip-communicator.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail:
dev-help@sip-communicator.dev.java.net

--
--------------------------------------------------------------------

"Life is full of unexpected but nothing happens without a reason..."


#8

Hey there Emanuel,

Onica S. Emanuel написа:

Note however that this version of the patch wich permits registrar-mode also
is, like I said, quite unstable yet ( so it's mainly for further
dev-purposes, not yet for actual functionality using :slight_smile: ).

OK, so are you saying that it would be a better idea to let you play
with it some more before we merge it into the trunk?

I don't have anything against this, especially if you are planning on
using and potentially improving it during the rest of the summer.

Cheers
Emil

···

Regards,
Emanuel

On Sat, 5 Jul 2008, Emil Ivov wrote:

Hey Michael,

Thanks for volunteering and res please! I guess we'll worry about the
wizard later.

Cheers
Emil

On Fri, Jul 4, 2008 at 11:42 AM, Koch Michael <MKoch@rowa.de> wrote:

Hi Emil!

This patch has been available for a while and there have been numerous
requests for support of a registrar-less mode of operation. It is
probably high time we integrated it into the application rather than
having to maintain a synced version.

Is anyone interested in applying it?

I could do it, if noone else who is more involved in SIP Communicator
development volunteers. I won't have time to adapt the account creation
wizards, though. Emanuel, thanks for updating the patch. I'm happy that you
found it useful.

Cheers
Michael Koch

-----Ursprüngliche Nachricht-----
Von: Emil Ivov [mailto:emcho@sip-communicator.org]
Gesendet: Freitag, 4. Juli 2008 00:05
An: dev@sip-communicator.dev.java.net
Betreff: Re: [sip-comm-dev] Adapted no-registrar patch

Hello Emanuel,

Thanks for sending this over.

Cheers
Emil

Onica S. Emanuel написа:

Hello all,

I've managed to manually apply an older patch made by

Michael Koch to a

more recent version of SC sources, the patch being intended

to provide

no-registrar SIP call availability. This is useful among others for
testing the currently developed ZRTP integration on point-to-point
communication in what I'm concerned (but probably it is

also of general

interest).

Some notes about the current patch version:

- it is done against the current SC encryption branch but it has
nothing to do with the encryption related sources so it will be
probably compatible with the main SC trunk also (anyway,

the encryption

branch is a bit behind the main trunk; I'll update it this

week, make some

patch changes, if necessary, and come back with them)

- the patch is only a manual adaptation of Michael Koch's original
sources with minor modifications as said, because the older

patch didn't

match anymore; the main idea was to split the

ProtocolProviderServiceSipImpl

into two classes, derived from an Abstract one, one class

for registrar

mode, and one for no-registrar mode; the choice of

instancing was based on

an option for the AccountID, which as Michael said at that

moment should

be probably integrated in the account registration wizard;

for now this

option is hardcoded and attached to every account for

testing purposes, so

the patch will cause all calls to go on without a registrar

(like I said I

adapted the patch mainly for testing purposes for now)

- in order for the patch to work don't forget to remove the

proxy fields

automatically completed at the account creation ( or at

least don't let it

completed with the local IP address which you'll probably

have used for

the last part of the SIP id; I've forgot the fields set

initially and took

me almost three days to figure out what was wrong :slight_smile: )

- as a final note: the patch isn't yet extensively tested

on the current

sources so I don't know what bugs can contain - if I find

any I'll come

back with a mail regarding the issues; I attach the patch

in two formats

to this mail one is a .patch version and one is a zip

archive with the

modified sources; I remind again that the patch is done against the
encryption branch on the SC trunk (which I'll update soon)

so for applying

the patch to the main trunk you should take care not to

overwrite any

newer source modifications

Regards,
Emanuel

--------------------------------------------------------------
----------

---------------------------------------------------------------------

To unsubscribe, e-mail:

dev-unsubscribe@sip-communicator.dev.java.net

For additional commands, e-mail:

dev-help@sip-communicator.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail:
dev-help@sip-communicator.dev.java.net

------------------------------------------------------------------------

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail: dev-help@sip-communicator.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail: dev-help@sip-communicator.dev.java.net


#9

Hello,

Hey there Emanuel,

Onica S. Emanuel ������������:

Note however that this version of the patch wich permits registrar-mode also
is, like I said, quite unstable yet ( so it's mainly for further
dev-purposes, not yet for actual functionality using :slight_smile: ).

OK, so are you saying that it would be a better idea to let you play
with it some more before we merge it into the trunk?

At this moment it's certainly not a good idea for merging my adapted patch in it's current state because it's not final. The normal mode calls don't work yet as I said. The part that seems to work without any problem is the wizard related one which is quite simple after all. The rest is Michael's old patch adapted (without many modifications or analysis), which works OK for now in no-registrar mode (in some conditions).

After I'll finish with the Linux partition setup and some ZRTP related testing this week, I'll try to take some time to look deeper into the problems mentioned and hopefully find a solution for them.
Anyway, I'll probably ask Michael for advice if I'll get stuck into something :). Also if he finds some time to take a look into the patch, if not now, maybe later after I'll look a bit more into the issues, it will be probably very helpful to point out things I may have missed. (For now, like I said, except the part related with the acc reg wizard, it's pretty much identical with his older patch with minor changes).

Cheers,
Emanuel

···

On Mon, 7 Jul 2008, Emil Ivov wrote:

I don't have anything against this, especially if you are planning on
using and potentially improving it during the rest of the summer.

Cheers
Emil

Regards,
Emanuel

On Sat, 5 Jul 2008, Emil Ivov wrote:

Hey Michael,

Thanks for volunteering and res please! I guess we'll worry about the
wizard later.

Cheers
Emil

On Fri, Jul 4, 2008 at 11:42 AM, Koch Michael <MKoch@rowa.de> wrote:

Hi Emil!

This patch has been available for a while and there have been numerous
requests for support of a registrar-less mode of operation. It is
probably high time we integrated it into the application rather than
having to maintain a synced version.

Is anyone interested in applying it?

I could do it, if noone else who is more involved in SIP Communicator
development volunteers. I won't have time to adapt the account creation
wizards, though. Emanuel, thanks for updating the patch. I'm happy that you
found it useful.

Cheers
Michael Koch

-----Urspr�ngliche Nachricht-----
Von: Emil Ivov [mailto:emcho@sip-communicator.org]
Gesendet: Freitag, 4. Juli 2008 00:05
An: dev@sip-communicator.dev.java.net
Betreff: Re: [sip-comm-dev] Adapted no-registrar patch

Hello Emanuel,

Thanks for sending this over.

Cheers
Emil

Onica S. Emanuel ������������:

Hello all,

I've managed to manually apply an older patch made by

Michael Koch to a

more recent version of SC sources, the patch being intended

to provide

no-registrar SIP call availability. This is useful among others for
testing the currently developed ZRTP integration on point-to-point
communication in what I'm concerned (but probably it is

also of general

interest).

Some notes about the current patch version:

- it is done against the current SC encryption branch but it has
nothing to do with the encryption related sources so it will be
probably compatible with the main SC trunk also (anyway,

the encryption

branch is a bit behind the main trunk; I'll update it this

week, make some

patch changes, if necessary, and come back with them)

- the patch is only a manual adaptation of Michael Koch's original
sources with minor modifications as said, because the older

patch didn't

match anymore; the main idea was to split the

ProtocolProviderServiceSipImpl

into two classes, derived from an Abstract one, one class

for registrar

mode, and one for no-registrar mode; the choice of

instancing was based on

an option for the AccountID, which as Michael said at that

moment should

be probably integrated in the account registration wizard;

for now this

option is hardcoded and attached to every account for

testing purposes, so

the patch will cause all calls to go on without a registrar

(like I said I

adapted the patch mainly for testing purposes for now)

- in order for the patch to work don't forget to remove the

proxy fields

automatically completed at the account creation ( or at

least don't let it

completed with the local IP address which you'll probably

have used for

the last part of the SIP id; I've forgot the fields set

initially and took

me almost three days to figure out what was wrong :slight_smile: )

- as a final note: the patch isn't yet extensively tested

on the current

sources so I don't know what bugs can contain - if I find

any I'll come

back with a mail regarding the issues; I attach the patch

in two formats

to this mail one is a .patch version and one is a zip

archive with the

modified sources; I remind again that the patch is done against the
encryption branch on the SC trunk (which I'll update soon)

so for applying

the patch to the main trunk you should take care not to

overwrite any

newer source modifications

Regards,
Emanuel

--------------------------------------------------------------
----------

---------------------------------------------------------------------

To unsubscribe, e-mail:

dev-unsubscribe@sip-communicator.dev.java.net

For additional commands, e-mail:

dev-help@sip-communicator.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail:
dev-help@sip-communicator.dev.java.net

------------------------------------------------------------------------

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail: dev-help@sip-communicator.dev.java.net

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@sip-communicator.dev.java.net
For additional commands, e-mail: dev-help@sip-communicator.dev.java.net

--
--------------------------------------------------------------------

"Life is full of unexpected but nothing happens without a reason..."


#10

Hi!

OK, I'll keep on standby with merging the patch. I'll also try to have a
look at it soon, but it's been a while since I spent time on SIP
Communicator development (it simply works too well).

Cheers
Michael Koch

···

-----Ursprüngliche Nachricht-----
Von: Onica S. Emanuel [mailto:eonica@info.uaic.ro]
Gesendet: Montag, 7. Juli 2008 23:57
An: dev@sip-communicator.dev.java.net
Betreff: Re: [sip-comm-dev] Adapted no-registrar patch
(account reg wizard update)

Hello,

On Mon, 7 Jul 2008, Emil Ivov wrote:

> Hey there Emanuel,
>
> Onica S. Emanuel ÿÿÿÿÿÿÿÿÿÿÿÿ:
>> Note however that this version of the patch wich permits
registrar-mode also
>> is, like I said, quite unstable yet ( so it's mainly for further
>> dev-purposes, not yet for actual functionality using :slight_smile: ).
>
> OK, so are you saying that it would be a better idea to let you play
> with it some more before we merge it into the trunk?

At this moment it's certainly not a good idea for merging my
adapted patch
in it's current state because it's not final. The normal mode
calls don't
work yet as I said. The part that seems to work without any
problem is the
wizard related one which is quite simple after all. The rest
is Michael's
old patch adapted (without many modifications or analysis),
which works OK
for now in no-registrar mode (in some conditions).

After I'll finish with the Linux partition setup and some
ZRTP related
testing this week, I'll try to take some time to look deeper into the
problems mentioned and hopefully find a solution for them.
Anyway, I'll probably ask Michael for advice if I'll get stuck into
something :). Also if he finds some time to take a look into
the patch, if
not now, maybe later after I'll look a bit more into the
issues, it will
be probably very helpful to point out things I may have
missed. (For now,
like I said, except the part related with the acc reg wizard,
it's pretty
much identical with his older patch with minor changes).

Cheers,
Emanuel

> I don't have anything against this, especially if you are
planning on
> using and potentially improving it during the rest of the summer.
>
> Cheers
> Emil
>
>>
>> Regards,
>> Emanuel
>>
>> On Sat, 5 Jul 2008, Emil Ivov wrote:
>>
>>> Hey Michael,
>>
>> Thanks for volunteering and res please! I guess we'll
worry about the
>> wizard later.
>>
>> Cheers
>> Emil
>>
>> On Fri, Jul 4, 2008 at 11:42 AM, Koch Michael > <MKoch@rowa.de> wrote:
>>> Hi Emil!
>>>
>>>> This patch has been available for a while and there have
been numerous
>>>> requests for support of a registrar-less mode of operation. It is
>>>> probably high time we integrated it into the application
rather than
>>>> having to maintain a synced version.
>>>>
>>>> Is anyone interested in applying it?
>>> I could do it, if noone else who is more involved in SIP
Communicator
>>> development volunteers. I won't have time to adapt the
account creation
>>> wizards, though. Emanuel, thanks for updating the patch.
I'm happy that you
>>> found it useful.
>>>
>>> Cheers
>>> Michael Koch
>>>
>>>> -----Ursprüngliche Nachricht-----
>>>> Von: Emil Ivov [mailto:emcho@sip-communicator.org]
>>>> Gesendet: Freitag, 4. Juli 2008 00:05
>>>> An: dev@sip-communicator.dev.java.net
>>>> Betreff: Re: [sip-comm-dev] Adapted no-registrar patch
>>>>
>>>> Hello Emanuel,
>>>>
>>>> Thanks for sending this over.
>>>>
>>>>
>>>> Cheers
>>>> Emil
>>>>
>>>> Onica S. Emanuel ÿÿÿÿÿÿÿÿÿÿÿÿ:
>>>>> Hello all,
>>>>>
>>>>> I've managed to manually apply an older patch made by
>>>> Michael Koch to a
>>>>> more recent version of SC sources, the patch being intended
>>>> to provide
>>>>> no-registrar SIP call availability. This is useful
among others for
>>>>> testing the currently developed ZRTP integration on
point-to-point
>>>>> communication in what I'm concerned (but probably it is
>>>> also of general
>>>>> interest).
>>>>>
>>>>> Some notes about the current patch version:
>>>>>
>>>>> - it is done against the current SC encryption branch but it has
>>>>> nothing to do with the encryption related sources so it will be
>>>>> probably compatible with the main SC trunk also (anyway,
>>>> the encryption
>>>>> branch is a bit behind the main trunk; I'll update it this
>>>> week, make some
>>>>> patch changes, if necessary, and come back with them)
>>>>>
>>>>> - the patch is only a manual adaptation of Michael
Koch's original
>>>>> sources with minor modifications as said, because the older
>>>> patch didn't
>>>>> match anymore; the main idea was to split the
>>>> ProtocolProviderServiceSipImpl
>>>>> into two classes, derived from an Abstract one, one class
>>>> for registrar
>>>>> mode, and one for no-registrar mode; the choice of
>>>> instancing was based on
>>>>> an option for the AccountID, which as Michael said at that
>>>> moment should
>>>>> be probably integrated in the account registration wizard;
>>>> for now this
>>>>> option is hardcoded and attached to every account for
>>>> testing purposes, so
>>>>> the patch will cause all calls to go on without a registrar
>>>> (like I said I
>>>>> adapted the patch mainly for testing purposes for now)
>>>>>
>>>>> - in order for the patch to work don't forget to remove the
>>>> proxy fields
>>>>> automatically completed at the account creation ( or at
>>>> least don't let it
>>>>> completed with the local IP address which you'll probably
>>>> have used for
>>>>> the last part of the SIP id; I've forgot the fields set
>>>> initially and took
>>>>> me almost three days to figure out what was wrong :slight_smile: )
>>>>>
>>>>> - as a final note: the patch isn't yet extensively tested
>>>> on the current
>>>>> sources so I don't know what bugs can contain - if I find
>>>> any I'll come
>>>>> back with a mail regarding the issues; I attach the patch
>>>> in two formats
>>>>> to this mail one is a .patch version and one is a zip
>>>> archive with the
>>>>> modified sources; I remind again that the patch is done
against the
>>>>> encryption branch on the SC trunk (which I'll update soon)
>>>> so for applying
>>>>> the patch to the main trunk you should take care not to
>>>> overwrite any
>>>>> newer source modifications
>>>>>
>>>>> Regards,
>>>>> Emanuel