Intent to Deprecate: RTCPeerConnection's googIPv6 constraint

117 views
Skip to first unread message

Henrik Boström

unread,
May 10, 2022, 9:15:32 AMMay 10
to blink-dev, Harald Alvestrand
Primary eng (and PM) emails

Summary
WebRTC allows sending and receiving audio, video and arbitrary data between two endpoints using the RTCPeerConnection API.

The googIPv6 constraint was originally added to opt-in to IPv6 support while that was in development, but IPv6 support has subsequently been enabled-by-default for many years. While setting googIPv6 to true is a NO-OP, some users make use of this non-standard API to disable IPv6 by setting it to false.

Motivation
A spec-compliant WebRTC implementation supports IPv6 and has no way to disable IPv6. Non-chromium browsers does not have an ability to disable IPv6, so this constraint is not good for interop.

We want to remove googIPv6 because a) it's non-standard, and b) it resides in a dictionary argument of the RTCPeerConnection called mediaConstraints that is also non-standard. The existence of googIPv6 prevents us from removing mediaConstraints.

Interoperability and Compatibility Risk
The other browser vendors have been very supportive of removing non-standard WebRTC APIs in the past, but I have not reached out to any browser vendor specifically about googIPv6.

Alternative implementation suggestion for web developers
N/A.

Usage information from UseCounter
The LegacyConstraintGoogIPv6 use counter suggests 0.000112% of page loads use this feature by setting it to false (setting it to true is a NO-OP and not counted here).

Entry on the feature dashboard



Henrik Boström

unread,
May 10, 2022, 9:22:26 AMMay 10
to blink-dev, Henrik Boström, Harald Alvestrand
I'm already displaying a deprecation warning when "googIPv6: false" is used. The warning was added in M102 (currently Beta).
In terms of deprecation timeline, I'd like to remove this in ~6 months. M108 is released on November 29, 2022. I suggest we target that milestone.

Daniel Bratell

unread,
May 10, 2022, 1:06:48 PMMay 10
to Henrik Boström, blink-dev, Harald Alvestrand

LGTM1

I agree that usage is low enough and use-cases not compelling enough to keep it. I've gone through a couple of hundred uses of googIPv6 on GitHub and there are a couple of "false" that you may want to take a look at, but nothing that causes concerns for me (you may know better). The timeline is also conservative.

Make sure you you add the removal date to the deprecation message if it's not already there.

/Daniel

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/ac2ba002-b13a-4c17-b26c-ea38f8ae33b1n%40chromium.org.

Mike Taylor

unread,
May 10, 2022, 1:34:22 PMMay 10
to Daniel Bratell, Henrik Boström, blink-dev, Harald Alvestrand
LGTM2 - this seems pretty safe and good for interop. Thanks for working on it.

jmedley via Chromestatus

unread,
May 10, 2022, 3:27:49 PMMay 10
to blin...@chromium.org
In which version are you hoping to deprecate? remove?

Chris Harrelson

unread,
May 10, 2022, 10:41:26 PMMay 10
to Mike Taylor, Daniel Bratell, Henrik Boström, blink-dev, Harald Alvestrand

Joe Medley

unread,
May 12, 2022, 10:59:23 AMMay 12
to blink-dev, jmedley via Chromestatus
Pinging this again. In two weeks I need to write what's in 103 beta on https://blog.chromium.org (https://blog.chromium.org/search/label/beta). Will deprecation be in 103?

Daniel Bratell

unread,
May 12, 2022, 4:51:01 PMMay 12
to Joe Medley, blink-dev, jmedley via Chromestatus

(not the implementor) According to followup to the request, the deprecation message is already in 102, possibly without ever having been documented. I don't know if you want to mention it in 103 instead.

The removal we approved is for 108.

/Daniel

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.

Joe Medley

unread,
May 16, 2022, 11:01:34 AMMay 16
to Daniel Bratell, blink-dev, jmedley via Chromestatus
I really can't. That's why I have to be such a nag about when something is going out.

I have you down for 108.

Thanks.
Joe Medley | Technical Writer, Chrome DevRel | jme...@google.com | 816-678-7195
If an API's not documented it doesn't exist.

Henrik Boström

unread,
May 18, 2022, 5:14:41 AMMay 18
to blink-dev, Joe Medley, blink-dev, jmedley via Chromestatus, Daniel Bratell
Hey thanks, delay in response due to vacation but I'm back now.

> According to followup to the request, the deprecation message is already in 102, possibly without ever having been documented. I don't know if you want to mention it in 103 instead.

 The deprecation warning I added in M102 did not mention any specific timeline, only that it is deprecated and one should stop using it.
I did sent out a PSA on a webrtc specific google group but this is the first I am writing about it on blink-dev, sorry if I skipped a step there.

> I have you down for 108.

Perfect.

> Make sure you you add the removal date to the deprecation message if it's not already there.

I will update the message to say M108 as the target.


Joe Medley

unread,
May 18, 2022, 11:35:01 AMMay 18
to Henrik Boström, blink-dev, jmedley via Chromestatus, Daniel Bratell
Thanks! 

Joe Medley | Technical Writer, Chrome DevRel | jme...@google.com | 816-678-7195
If an API's not documented it doesn't exist.

Reply all
Reply to author
Forward
0 new messages