- From: Soares Chen via GitHub <sysbot+gh@w3.org>
- Date: Fri, 21 Jul 2017 09:00:10 +0000
- To: public-webrtc-logs@w3.org
soareschen has just created a new issue for https://github.com/w3c/webrtc-pc:
== How is port number specified when verifying peer identity's assertion? ==
rtcweb-security-arch specifies that the IdP proxy script can be located at a different port number. So for example an IdP for example.com can be located at https://example.com:8443/.well-known/idp-proxy/provider.js.
Now in `setIdentityProvider()`, it seems that the port number can be specified as part of the `provider` attribute string, together with the domain name. So we can call
```javascript
pc.setIdentityProvider('example.com:8443', { protocol: 'provider.js' })
```
to load the correct IdP proxy.
But when the IdP proxy is returning assertion result, it is not clear how the port information is carried over. The WebIDL for `RTCIdentityAssertionResult` is:
```
dictionary RTCIdentityAssertionResult {
required RTCIdentityProviderDetails idp;
required DOMString assertion;
};
dictionary RTCIdentityProviderDetails {
required DOMString domain;
DOMString protocol = "default";
};
```
I am not sure if the `domain` field is the right place to specify the port number, e.g.:
```
{
idp: {
domain: "example.com:8443",
protocol: "provider.js"
},
assertion: "..."
}
```
Given that the assertion result is eventually serialized into the `a=identity` SDP line, and from rtcweb-security-arch it seems that the `domain` field is intended to hold the domain name only, I am not sure port number is allowed there.
If the port number is not specified in `domain`, how else would a remote peer know to load the IdP proxy script through custom port?
Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/1502 using your GitHub account
Received on Friday, 21 July 2017 09:00:11 UTC