[jitsi-dev] Who becomes initiator in P2P mode?


#1

Hi,

I've not yet implemented P2P mode in Edge, but as far as Edge is the
SDP answerer P2P should work (pc.createOffer() not yet implemented in
my Edge WebRTC shim).

Said that, who becomes the initiator (SDP offerer) when in P2P? Is
there any way (or a hack I could code in my local repo) to always make
Chrome send the SDP offer to Edge?

Thanks.

···

--
Iñaki Baz Castillo
<ibc@aliax.net>


#2

Hey Inaki,

So usually it will be the one who has moderator MUC permissions. If
the "everyone's a moderator" plugin is enabled it will be random,
because it's based on MUC nickname alphabetical order (and nicknames
are random with anonymous XMPP login).

In case it's not implemented then it should just fail to establish P2P
connection, but the JVB one should still work.

Regards,
Pawel

···

On Wed, May 10, 2017 at 1:50 PM, Iñaki Baz Castillo <ibc@aliax.net> wrote:

Hi,

I've not yet implemented P2P mode in Edge, but as far as Edge is the
SDP answerer P2P should work (pc.createOffer() not yet implemented in
my Edge WebRTC shim).

Said that, who becomes the initiator (SDP offerer) when in P2P? Is
there any way (or a hack I could code in my local repo) to always make
Chrome send the SDP offer to Edge?

Thanks.

--
Iñaki Baz Castillo
<ibc@aliax.net>

_______________________________________________
dev mailing list
dev@jitsi.org
Unsubscribe instructions and other list options:
http://lists.jitsi.org/mailman/listinfo/dev


#3

So usually it will be the one who has moderator MUC permissions. If
the "everyone's a moderator" plugin is enabled it will be random,
because it's based on MUC nickname alphabetical order (and nicknames
are random with anonymous XMPP login).

Hi, honestly no idea about "moderator plugin". I just connect two
browsers into https://beta.meet.jit.si.

In case it's not implemented then it should just fail to establish P2P
connection, but the JVB one should still work.

Sure. My aim was to test P2P between Chrome and Edge forcing Chrome to
always be the SDP initiator (since pc.createOffer() is not yet
implemented in the Edge shim). This is because I want to test P2P
video (I got ugly video rendering when using the JVB [*]).

[*] https://github.com/jitsi/lib-jitsi-meet/pull/473#issuecomment-300573271

···

2017-05-11 2:56 GMT+02:00 Paweł Domas <pawel.domas@jitsi.org>:

--
Iñaki Baz Castillo
<ibc@aliax.net>


#4

So usually it will be the one who has moderator MUC permissions. If
the "everyone's a moderator" plugin is enabled it will be random,
because it's based on MUC nickname alphabetical order (and nicknames
are random with anonymous XMPP login).

Hi, honestly no idea about "moderator plugin". I just connect two
browsers into https://beta.meet.jit.si.

On that deployment all users are moderators.

In case it's not implemented then it should just fail to establish P2P
connection, but the JVB one should still work.

Sure. My aim was to test P2P between Chrome and Edge forcing Chrome to
always be the SDP initiator (since pc.createOffer() is not yet
implemented in the Edge shim). This is because I want to test P2P
video (I got ugly video rendering when using the JVB [*]).

[*] https://github.com/jitsi/lib-jitsi-meet/pull/473#issuecomment-300573271

Then you can use meet.jit.si and join with Chrome first.

Cheers,

···

On May 11, 2017, at 12:55, Iñaki Baz Castillo <ibc@aliax.net> wrote:
2017-05-11 2:56 GMT+02:00 Paweł Domas <pawel.domas@jitsi.org>:

--
Saúl


#5

I can't. Using such a URL I get JS libs directly from that server
rather than from my local webpack-dev-server.

···

2017-05-11 14:45 GMT+02:00 Saúl Ibarra Corretgé <scorretge@atlassian.com>:

Then you can use meet.jit.si and join with Chrome first.

--
Iñaki Baz Castillo
<ibc@aliax.net>


#6

You should be able to just hack in a browser type check instead of the username check:
https://github.com/jitsi/lib-jitsi-meet/blob/master/JitsiConference.js#L2195

Right, Pawel?

Boris

···

On 11/05/2017 08:49, Iñaki Baz Castillo wrote:

2017-05-11 14:45 GMT+02:00 Saúl Ibarra Corretgé <scorretge@atlassian.com>:

Then you can use meet.jit.si and join with Chrome first.

I can't. Using such a URL I get JS libs directly from that server
rather than from my local webpack-dev-server.


#7

Yes, but I don't understand the problem. Is it about not having dev
environment where you can modify anything ?

···

On Thu, May 11, 2017 at 9:58 AM, Boris Grozev <boris@jitsi.org> wrote:

On 11/05/2017 08:49, Iñaki Baz Castillo wrote:

2017-05-11 14:45 GMT+02:00 Saúl Ibarra Corretgé <scorretge@atlassian.com>:

Then you can use meet.jit.si and join with Chrome first.

I can't. Using such a URL I get JS libs directly from that server
rather than from my local webpack-dev-server.

You should be able to just hack in a browser type check instead of the
username check:
https://github.com/jitsi/lib-jitsi-meet/blob/master/JitsiConference.js#L2195

Right, Pawel?


#8

...because they are placed in a CDN so it uses absolute URLs for JS files.

···

2017-05-11 15:49 GMT+02:00 Iñaki Baz Castillo <ibc@aliax.net>:

2017-05-11 14:45 GMT+02:00 Saúl Ibarra Corretgé <scorretge@atlassian.com>:

Then you can use meet.jit.si and join with Chrome first.

I can't. Using such a URL I get JS libs directly from that server
rather than from my local webpack-dev-server.

--
Iñaki Baz Castillo
<ibc@aliax.net>


#9

I just wanted to test P2P video (Edge <-> Chrome) to confirm whether
VP8 video sent by Edge is correctly displayed by Chrome. If so, that
would confirm that this issue [1] is somehow produced due the JVB (RTP
mangling or whatever).

[1] https://github.com/jitsi/lib-jitsi-meet/pull/473#issuecomment-300573271

···

2017-05-11 17:05 GMT+02:00 Paweł Domas <pawel.domas@jitsi.org>:

Yes, but I don't understand the problem. Is it about not having dev
environment where you can modify anything ?

--
Iñaki Baz Castillo
<ibc@aliax.net>


#10

I think what Boris suggested (testing with appRTC) would be easier and achieve the same result.

Cheers,

···

On May 12, 2017, at 11:35, Iñaki Baz Castillo <ibc@aliax.net> wrote:

2017-05-11 17:05 GMT+02:00 Paweł Domas <pawel.domas@jitsi.org>:

Yes, but I don't understand the problem. Is it about not having dev
environment where you can modify anything ?

I just wanted to test P2P video (Edge <-> Chrome) to confirm whether
VP8 video sent by Edge is correctly displayed by Chrome. If so, that
would confirm that this issue [1] is somehow produced due the JVB (RTP
mangling or whatever).

[1] https://github.com/jitsi/lib-jitsi-meet/pull/473#issuecomment-300573271

--
Saúl