JavaScript ShadowRealm proposal integration

[Ms2ger]

To replace #5339.

• At least two implementers are interested (and none opposed):
  • ShadowRealm API mozilla/standards-positions#997
  • …
• [·] Tests are written and can be reviewed and commented upon at:
  • In progress across multiple pull requests; see https://github.com/web-platform-tests/wpt/pulls?q=is%3Apr+is%3Aopen+shadowrealm
• Implementation bugs are filed:
  • Chromium: https://issues.chromium.org/issues/41480387
  • Gecko: https://bugzilla.mozilla.org/show_bug.cgi?id=1797187
  • WebKit: https://bugs.webkit.org/show_bug.cgi?id=283289
  • Deno (only for timers, structured clone, base64 utils, channel messaging, module resolution, web workers, and web storage): [WIP] Implement the ShadowRealm constructor denoland/deno#16211
  • Node.js (only for timers, structured clone, base64 utils, channel messaging, and module resolution): …
• MDN issue is filed: Document ShadowRealm mdn/content#36842
• The top of this comment includes a clear commit message to use.

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

TODO:

• Prepare PRs for dependent specs to adjust to terminology changes

---

/browsing-the-web.html ( diff )
/dom.html ( diff )
/embedded-content.html ( diff )
/form-elements.html ( diff )
/index.html ( diff )
/infrastructure.html ( diff )
/media.html ( diff )
/nav-history-apis.html ( diff )
/references.html ( diff )
/scripting.html ( diff )
/structured-data.html ( diff )
/timers-and-user-prompts.html ( diff )
/webappapis.html ( diff )
/workers.html ( diff )

[nicolo-ribaudo]

The changes to the module loading algorithms look ok.

For the record, we (people working on modules in TC39) have some ideas for splitting the stateless and stateful parts of Module Records in ECMA-262. If/once that will happen, it would be great to share a module map for the stateless parts across principal realms and shadow realms, so that basically in this case:

import "/foo";
new ShadowRealm().importValue("/foo", "bar");

we only call into fetch once (to get the stateless part) and then we create the stateful part twice in both realms.

[Ms2ger]

I think this could benefit from review at this point

[littledan]

ShadowRealms reached 2.7 in TC39. Acceptance of this PR by WHATWG editors is a blocking requirement for Stage 3, though it probably shouldn't be landed until the Web API set is finalized.

@domenic @annevk @zcorpan Any thoughts/concerns?

[Ms2ger]

@domenic or @annevk - it seems to me that the we're making progress on the questions about exposure; is there anything else I can do to help move this PR forward?

[shannonbooth]

Could we still run into an issue here if there is a third layer of shadow realm nesting?

My thinking is that this should effectively be a while loop which will go up until it sees that the host defined is not a synthetic settings object. Then, principal-realm-of-realm does not need to worry about how many layers of nesting of shadow realms there are.

[Ms2ger]

Suppose a principal realm X → ShadowRealm A → ShadowRealm B → ShadowRealm C.

• Creating A: O's associated realm is X, so we set A.[[HostDefined]]'s synthetic-realm-settings-object-principal-realm to X's concept-principal-realm-of-realm which is also X
• Creating B: O's associated realm is A, and A's concept-principal-realm-of-realm is A.[[HostDefined]]'s synthetic-realm-settings-object-principal-realm, which is X, and we set B.[[HostDefined]]'s synthetic-realm-settings-object-principal-realm to X.
• Creating C: O's associated realm is B, and B's concept-principal-realm-of-realm is B.[[HostDefined]]'s synthetic-realm-settings-object-principal-realm, which is X, etc.