833509 - Add URI protocol handlers for STUN and TURN

Status: NEW

(Core :: WebRTC, defect, P4)

Description

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

The nsURI class does not currently have protocol handlers for "STUN:", "STUNS:", "TURN:" and "TURNS:". This is needed to centralize parsing and to give us proper validation and restriction enforcement of stun/turn urls.

Adam wrote in bug 825703, comment 43:
> ... the STUN URL scheme is defined in
> http://tools.ietf.org/html/draft-nandakumar-rtcweb-stun-uri-02
> and TURN is described in
> http://tools.ietf.org/html/draft-petithuguenin-behave-turn-uri-03
> (but note that neither of these have been adopted by an IETF working group,
> so they're still very much works in progress).
> [...]
> Jan-Ivar points out that the W3C spec gives an example that doesn't conform to
> syntax (http://dev.w3.org/2011/webrtc/editor/webrtc.html#rtcconfiguration-type
> contains a <username>) so there is still some harmonization necessary.
> ...
> Extra credit reading on the design of the URL syntax:
> http://blog.marc.petit-huguenin.org/2012/09/on-design-of-stun-and-turn-uri-formats.html

Boris wrote in bug 825703, comment 44:
> ... So looks like the relevant BNF is:
>
>      stun:<stun-host>:<stun-port>
>      stuns:<stun-host>:<stun-port>
>
>   turnURI   = scheme ":" host [ ":" port ] [ "?transport=" transport ]
>   transport = "udp" / "tcp" / transport-ext
>   transport-ext = 1*unreserved
>
> We should probably actually enforce those restrictions somewhere.

Comment 1

[Boris Zbarsky [:bzbarsky]]

We don;t need to add anything to the URI classes.  Protocol handlers are just services exposed via standardized contract IDs.  Whatever code is interested in these URIs can just implement a protocol handler.

Comment 2

[Bulk Bug Changes for mreavy's org]

Mass change P3->P4 to align with new Mozilla triage process.

Comment 3

[Randell Jesup [:jesup] (needinfo me)]

Any relevance to this bug now?

Comment 4

[Byron Campen [:bwc]]

I think this is going to be important for bug 1529588. Marking it as blocking for now.

Comment 5

[Tristan Matthews]

(In reply to Jan-Ivar Bruaroey [:jib] (needinfo? me) from comment #0)

The nsURI class does not currently have protocol handlers for "STUN:",
"STUNS:", "TURN:" and "TURNS:". This is needed to centralize parsing and to
give us proper validation and restriction enforcement of stun/turn urls.

Adam wrote in bug 825703, comment 43:

... the STUN URL scheme is defined in
http://tools.ietf.org/html/draft-nandakumar-rtcweb-stun-uri-02
and TURN is described in
http://tools.ietf.org/html/draft-petithuguenin-behave-turn-uri-03
(but note that neither of these have been adopted by an IETF working group,
so they're still very much works in progress).
[...]
Jan-Ivar points out that the W3C spec gives an example that doesn't conform to
syntax (http://dev.w3.org/2011/webrtc/editor/webrtc.html#rtcconfiguration-type
contains a <username>) so there is still some harmonization necessary.
...
Extra credit reading on the design of the URL syntax:
http://blog.marc.petit-huguenin.org/2012/09/on-design-of-stun-and-turn-uri-formats.html

This article is no longer available but here's an earlier snapshot:
https://web.archive.org/web/20160514042254/http://blog.marc.petit-huguenin.org/2012/09/on-design-of-stun-and-turn-uri-formats.html