Re: ISSUE-124: rel-limits - Straw Poll for Objections

On 27.02.2011 17:53, Philip Jägenstedt wrote:
> ...
>> As far as I can tell, Philip didn't point out a "race condition".
>
> What Edward is referring to is probably this example:
>
> <link rel="stylesheet" href="foo.css">
> <!-- potential network lag here -->
> <link rel="stylesheet noreferrer" href="foo.css">
>
> Some (bad) solutions to the problem would result in a race condition,
> but all I've asked for is a solution to evaluate.
> ...

I see, thanks for the clarification.

This is a general problem when relations for specific links can be added 
in multiple places; it's not limited to <link> but would also apply to 
the Link: HTTP response header field.

It's also not specific, it would apply to any new relation that affects 
the request being made (like suppressing cookies...).

The spec will need to deal with this anyway; I think the best 
*authoring* advice is not to shoot yourself in the foot: just define the 
link relations on the link in a single place. This even could be made a 
conformance requirement.

Best regards, Julian

Received on Monday, 28 February 2011 15:51:21 UTC