Commit Graph

17 Commits

Author SHA1 Message Date
buildbot@webrtc.org
dcc1f0426b (Auto)update libjingle 75852725-> 75853560
git-svn-id: http://webrtc.googlecode.com/svn/trunk@7231 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-09-18 23:14:12 +00:00
guoweis@webrtc.org
369a637ac8 Implemented Network::GetBestIP() selection logic as following.
1) return the first global temporary and non-deprecrated ones.
2) if #1 not available, return global one.
3) if #2 not available, use ULA ipv6 as last resort.

ULA stands for unique local address. They are only useful in a private
WebRTC deployment. More detail: http://en.wikipedia.org/wiki/Unique_local_address

BUG=3808

At this point, rule #3 actually won't happen at current
implementation. The reason being that ULA address starting with 0xfc 0r 0xfd will be grouped into its own Network. The result of that is WebRTC will have one extra Network to generate candidates but the lack of rule #3 shouldn't prevent turning on IPv6 since ULA should only be tried in a close deployment anyway.

R=jiayl@webrtc.org

Committed: https://code.google.com/p/webrtc/source/detail?r=7200

Committed: https://code.google.com/p/webrtc/source/detail?r=7201

Review URL: https://webrtc-codereview.appspot.com/31369004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7216 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-09-17 22:37:29 +00:00
guoweis@webrtc.org
40c2aa36f2 Implemented Network::GetBestIP() selection logic as following.
1) return the first global temporary and non-deprecrated ones.
2) if #1 not available, return global one.
3) if #2 not available, use ULA ipv6 as last resort.

ULA stands for unique local address. They are only useful in a private
WebRTC deployment. More detail: http://en.wikipedia.org/wiki/Unique_local_address

BUG=3808

At this point, rule #3 actually won't happen at current
implementation. The reason being that ULA address starting with 0xfc 0r 0xfd will be grouped into its own Network. The result of that is WebRTC will have one extra Network to generate candidates but the lack of rule #3 shouldn't prevent turning on IPv6 since ULA should only be tried in a close deployment anyway.

R=jiayl@webrtc.org

Committed: https://code.google.com/p/webrtc/source/detail?r=7200

Review URL: https://webrtc-codereview.appspot.com/31369004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7201 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-09-16 20:29:41 +00:00
guoweis@webrtc.org
f8bff762d1 Implemented Network::GetBestIP() selection logic as following.
1) return the first global temporary and non-deprecrated ones.
2) if #1 not available, return global one.
3) if #2 not available, use ULA ipv6 as last resort.

ULA stands for unique local address. They are only useful in a private
WebRTC deployment. More detail: http://en.wikipedia.org/wiki/Unique_local_address

BUG=3808

At this point, rule #3 actually won't happen at current
implementation. The reason being that ULA address starting with 0xfc 0r 0xfd will be grouped into its own Network. The result of that is WebRTC will have one extra Network to generate candidates but the lack of rule #3 shouldn't prevent turning on IPv6 since ULA should only be tried in a close deployment anyway.

R=jiayl@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/31369004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7200 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-09-16 20:17:22 +00:00
buildbot@webrtc.org
a09a99950e (Auto)update libjingle 73222930-> 73226398
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6891 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-08-13 17:26:08 +00:00
mallinath@webrtc.org
2d60c5e8bc Encoding and Decoding of TCP candidates as defined in RFC 6544.
R=juberti@chromium.org, jiayl@webrtc.org, juberti@webrtc.org
BUG=2204

Review URL: https://webrtc-codereview.appspot.com/21479004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6857 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-08-08 22:29:20 +00:00
buildbot@webrtc.org
d4e598d57a (Auto)update libjingle 72097588-> 72159069
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6799 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-07-29 17:36:52 +00:00
buildbot@webrtc.org
51c5508bf1 (Auto)update libjingle 72016417-> 72097588
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6792 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-07-28 22:26:15 +00:00
buildbot@webrtc.org
45304ff0a7 (Auto)update libjingle 71829282-> 71834788
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6773 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-07-24 16:06:35 +00:00
buildbot@webrtc.org
e2da234e27 (Auto)update libjingle 71766184-> 71775619
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6768 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-07-23 21:09:01 +00:00
jiayl@webrtc.org
a0b929b63c Revert "Reland r6707 with the fix for callclient.cc."
Breaking pulse build again.
This reverts commit 3e0bb9b5bf7f616000399e24f1d9622ad6b612f9.

TBR=wu@webrtc.org
BUG=3310

Review URL: https://webrtc-codereview.appspot.com/17979004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6740 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-07-18 22:28:36 +00:00
jiayl@webrtc.org
a6e8cf8fb7 Reland r6707 with the fix for callclient.cc.
TBR=mallinath@webrtc.org
BUG=3310

Review URL: https://webrtc-codereview.appspot.com/13039004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6737 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-07-18 21:34:11 +00:00
wu@webrtc.org
52eddec71b Revert 6707 "Add support of multiple STUN servers in UDPPort."
Reason:
Breaks the build on callclient.cc.

> Add support of multiple STUN servers in UDPPort.
> Now UDPPort signals PortComplete or PortError when the Bind requests for all STUN servers are responded or failed. If any STUN bind is successful, PortComplete is signaled; otherwise, PortError is signaled.
> 
> I discovered a bug in SocketAddress while working on this. It didn't consider two addresses unequal if they have unresolved IP and different hosts. It's fixed now.
> 
> BUG=3310
> R=mallinath@webrtc.org
> 
> Review URL: https://webrtc-codereview.appspot.com/13879004

TBR=jiayl@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14999004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6711 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-07-17 00:03:24 +00:00
jiayl@webrtc.org
46fb331bc5 Add support of multiple STUN servers in UDPPort.
Now UDPPort signals PortComplete or PortError when the Bind requests for all STUN servers are responded or failed. If any STUN bind is successful, PortComplete is signaled; otherwise, PortError is signaled.

I discovered a bug in SocketAddress while working on this. It didn't consider two addresses unequal if they have unresolved IP and different hosts. It's fixed now.

BUG=3310
R=mallinath@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/13879004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6707 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-07-16 20:55:31 +00:00
buildbot@webrtc.org
f875f15afb (Auto)update libjingle 64709629-> 64813990
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5897 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-04-14 16:06:21 +00:00
henrike@webrtc.org
f5bebd40f3 (Auto)update libjingle 64247466-> 64326665
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5845 4adac7df-926f-26a2-2b94-8c16560cd09d
2014-04-04 18:39:07 +00:00
henrike@webrtc.org
28e2075280 Adds trunk/talk folder of revision 359 from libjingles google code to
trunk/talk


git-svn-id: http://webrtc.googlecode.com/svn/trunk@4318 4adac7df-926f-26a2-2b94-8c16560cd09d
2013-07-10 00:45:36 +00:00