Re: [apps-discuss] SRV and service discovery, draft-ietf-appsawg-webfinger-04

On Sun, Nov 25, 2012 at 8:32 AM, John Levine <johnl@taugh.com> wrote:

> >> not unlike the ones here.)  But if after 16 years, the people who
> >> maintain Javascript are unpersuaded that would be worth tneir time to
> >> add SRV lookups, it's not because they are stupid.
> >
> >Who has to do that persuasion?
>
> People who think doing SRV lookups is a good idea.  You, perhaps.
>
> >I.e. is it not part of an initiative like this, that want some service
> >discovery, to ensure JavaScript is doing The Right Thing?
>
> Presumably, but arguments that end at "add SRV to Javascript" are not
> productive.
>
> It is not clear to me even at what level it would be appropriate to do
> SRV lookups.  One possibility might be new builtin JS functions,
> although I don't understand at this point how that would interact with
> the single origin rule.
>
>
Again, it doesn't interact at all with the same origin policy. The SRV
record information would be used to construct urls that contain the
appropriate domain and port. How those urls are then put to use is a
separate matter entirely..it's only at that point where things like the
same origin policy come into play. It would be handled no differently than
any other url used in the javascript code.

- James

> Another possibility is a new URL scheme, along the lines
> of draft-jennings-http-srv, or more generally like this:
>
>   srv[s]://example.com:foo/bar
>
> does a SRV looksup of
>
>   _foo._tcp.example.com
>
> and turns the request into
>
>   http[s]://<target>:<port>/bar
>
> (Add hand waving about what to do with the priority and weight if
> there are multiple answers, along the lines of what it already does
> with multiple A or AAAA answers.)
>
> If people think this is a good idea, they sould write it up and try
> implementing it in some of the open source browsers that people use.
> Or else stop kvetching.
>
> R's,
> John
> _______________________________________________
> apps-discuss mailing list
> apps-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/apps-discuss
>