Compare the strongest integrity metadata when reusing a preload

[annevk]

A preloaded response can be reused when the consumer's strongest integrity metadata is unordered equal to that of the preload, as only the strongest metadata is used to check integrity. Also check for an empty set rather than "no metadata", which SRI no longer returns.

---

As a result, a consumer whose integrity metadata only contains unknown algorithms can reuse any preload, as SRI does not check integrity in that case.

From reading their code: Gecko already compares only the strongest metadata, ignoring order. Chromium compares the full parsed metadata, including order. WebKit instead checks the consumer's integrity metadata against the response when it is used.

• At least two implementers are interested (and none opposed):
  • …
  • …
• Tests are written and can be reviewed and commented upon at:
  • Test preload reuse based on the strongest integrity metadata web-platform-tests/wpt#63600
• Implementation bugs are filed:
  • Chromium: …
  • Gecko: …
  • WebKit: …
  • Deno (only for timers, structured clone, base64 utils, channel messaging, module resolution, web workers, and web storage): …
  • Node.js (only for timers, structured clone, base64 utils, channel messaging, and module resolution): …
• Corresponding HTML AAM & ARIA in HTML issues & PRs:
• MDN issue is filed: …
• The top of this comment includes a clear commit message to use.

(See WHATWG Working Mode: Changes for more details.)

---

/infrastructure.html ( diff )
/links.html ( diff )

[noamr]

Also @yoavweiss probably has some fresh recollection of integrity metadata stuff.

[noamr]

AI Review

This looks right to me: SRI's parse metadata now returns an (ordered) set of maps, map equality is by value, and Infra's "unordered equal" is subset+superset, which also tolerates the duplicates SRI can currently produce (it uses list append rather than set append — probably worth an SRI issue).

A few things:

• Nit: the "; or" is on the last list item; it should move to the first one now that you're rewriting both lines.
• Switching to "is empty" also means a consumer whose integrity contains only unknown algorithms (e.g. md5-…) now matches any preload. That's consistent with SRI's "empty set matches" semantics, but worth calling out and testing.
• Since SRI no longer stores options at all, is the XXX about issue Note that new commits as well as <picture> et al are licensed under CC0 #116 still accurate as written?
• Have we considered comparing "get the strongest metadata from set" of each side instead? That would allow reusing a sha384-B preload for a sha256-A sha384-B consumer with identical security.

[noamr]

Needs implementation bugs, but you can mark chromium as interested.