Back/forward cache NotRestoredReasons API

[rubberyuzu]

Request for Mozilla Position on an Emerging Web Specification

• Specification Title: Back/forward cache NotRestoredReasons API
• Explainer URL (if available): https://github.com/rubberyuzu/bfcache-not-retored-reason/blob/main/NotRestoredReason.md

Other information

Chromium is doing a trial of this feature as part of PerformanceNavigationTiming API.

[zcorpan]

@petervanderbeken @smaug---- @bdekoz

I see @smaug---- gave feedback in whatwg/html#7094

The HTML standard specifies some things that make the document unsalvageable. It seems like a good idea to look into what other cases make documents unsalvageable in browsers and standardize them, and fix any interop issues.

[smaug----]

The reason strings need to be standardized. Right now even the explainer is inconsistent. Some examples use "Broadcast channel", others "broadcast channel".
And no browser internal behavior should be exposed, as we said already in whatwg/html#7094 (comment)

So, I think the proposal needs some more work.

[petervanderbeken]

I see @smaug---- gave feedback in whatwg/html#7094

That was a result of @smaug---- and me looking it over. In general we're positive on this, but only with a standardized list of reason strings.

[zcorpan]

So defer for now, and when reason strings and unsalvageable are better standardized we'd switch to positive?

[smaug----]

That sounds good to me.

[hober]

WebKit/standards-positions#154

[yoavweiss]

@zcorpan - I'm trying to assess broader support (to see if WICG/proposals#93 can get a WICG repo to enable broader collaboration). Would y'all be supportive in moving that work to WICG, even if it's not yet in the shape you'd like it to be?

[zcorpan]

Moving to WICG seems OK.

[yoavweiss]

Thanks!! :)

[rubberyuzu]

I updated the explainer around the wording, and now I am working on the spec draft here.
Can you please review this again?

Thanks!

[smaug----]

I think we can say positive here to what landed to the spec (that shouldn't leak cross origin information, unlike the initial proposal). I and peterv did both look at the spec PRs many times. whatwg/html#9360 and whatwg/html#10154