RT and NT processing is not well defined

[igrigorik]

As a followup to https://github.com/w3c/server-timing/pull/36.

https://w3c.github.io/server-timing/#processing-model

When processing the response of the current document, set the serverTiming attribute on the newly created PerformanceNavigationTiming object to the return value of the server-timing header parsing algorithm

You can't run the header parsing algorithm when the object is "newly created" because we don't yet have the response back from the server.

For each resource fetched by the current browsing context, excluding resources fetched by cross-origin stylesheets fetched with no-cors policy, set the serverTiming attribute on the newly created PerformanceResourceTiming object to:

• We shouldn't have to repeat the "excluding..." bits, that's already covered by ResourceTiming, which we're extending here.
• Similar to NT, the header parsing algorithm needs to run at the end, probably right before we "queue" the object at step 20 in RT.
  • We should define a formal hook in RT processing and call into that.

[siusin]

Todd: Resource Timing 2 is the place to inject this hook, as it’s getting ready to ship. ... Server timing just points to a linkable hook in Resource Timing 2. ... Yoav: Is there a specific example we can use as prior art for this text?

@cvazac the "steps to constrain a payment handler" processing model in the payment method spec might be a good example --

Optionally, set card [" billingAddress "] to the result running the steps to create a payment address

(thanks @marcoscaceres )

[yoavweiss]

@noamr - can we close this?

[noamr]

@noamr - can we close this?

Yes!