feat(engine): add ShadowRealm behind --unsafe-shadowrealm#795
Draft
frostney wants to merge 1 commit into
Draft
Conversation
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub. 1 Skipped Deployment
|
Contributor
|
Important Review skippedDraft detected. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Comment |
Contributor
Benchmark Results430 benchmarks Interpreted: 🟢 233 improved · 🔴 10 regressed · 187 unchanged · avg +12.0% arraybuffer.js — Interp: 🟢 6, 8 unch. · avg +11.6% · Bytecode: 🔴 14 · avg -29.7%
arrays.js — Interp: 🟢 12, 7 unch. · avg +10.6% · Bytecode: 🔴 18, 1 unch. · avg -29.5%
async-await.js — Interp: 🟢 4, 2 unch. · avg +15.6% · Bytecode: 🔴 6 · avg -33.0%
async-generators.js — Interp: 2 unch. · avg +9.0% · Bytecode: 🔴 2 · avg -32.7%
atomics.js — Interp: 🟢 3, 3 unch. · avg +14.1% · Bytecode: 🔴 6 · avg -29.4%
base64.js — Interp: 🟢 7, 3 unch. · avg +11.8% · Bytecode: 🔴 10 · avg -32.7%
classes.js — Interp: 🟢 22, 9 unch. · avg +10.3% · Bytecode: 🔴 27, 4 unch. · avg -28.4%
closures.js — Interp: 🟢 7, 4 unch. · avg +15.5% · Bytecode: 🔴 11 · avg -32.8%
collections.js — Interp: 🟢 6, 6 unch. · avg +11.0% · Bytecode: 🔴 12 · avg -35.9%
csv.js — Interp: 🟢 6, 7 unch. · avg +10.8% · Bytecode: 🔴 11, 2 unch. · avg -27.5%
destructuring.js — Interp: 🟢 17, 5 unch. · avg +16.8% · Bytecode: 🔴 21, 1 unch. · avg -29.9%
fibonacci.js — Interp: 🟢 5, 3 unch. · avg +13.8% · Bytecode: 🔴 7, 1 unch. · avg -32.2%
float16array.js — Interp: 🟢 23, 9 unch. · avg +12.7% · Bytecode: 🔴 29, 3 unch. · avg -29.7%
for-of.js — Interp: 🟢 3, 4 unch. · avg +7.5% · Bytecode: 🔴 7 · avg -32.8%
generators.js — Interp: 🟢 3, 1 unch. · avg +13.8% · Bytecode: 🔴 4 · avg -30.1%
intl.js — Interp: 🟢 6 · avg +20.0% · Bytecode: 🔴 6 · avg -34.9%
iterators.js — Interp: 🟢 15, 27 unch. · avg +6.9% · Bytecode: 🔴 41, 1 unch. · avg -26.5%
json.js — Interp: 🟢 13, 7 unch. · avg +16.0% · Bytecode: 🔴 20 · avg -30.0%
jsx.jsx — Interp: 🟢 6, 15 unch. · avg +10.8% · Bytecode: 🔴 18, 3 unch. · avg -28.8%
modules.js — Interp: 🟢 4, 5 unch. · avg +13.6% · Bytecode: 🔴 9 · avg -25.7%
numbers.js — Interp: 🟢 5, 6 unch. · avg +17.5% · Bytecode: 🔴 11 · avg -32.0%
objects.js — Interp: 🟢 2, 5 unch. · avg +14.8% · Bytecode: 🔴 7 · avg -28.7%
promises.js — Interp: 🟢 3, 9 unch. · avg +10.1% · Bytecode: 🔴 11, 1 unch. · avg -25.8%
property-access.js — Interp: 🟢 1, 4 unch. · avg +12.6% · Bytecode: 🔴 5 · avg -33.8%
regexp.js — Interp: 🟢 4, 7 unch. · avg +12.2% · Bytecode: 🔴 10, 1 unch. · avg -28.5%
strings.js — Interp: 🟢 13, 6 unch. · avg +13.9% · Bytecode: 🔴 19 · avg -31.0%
temporal.js — Interp: 🟢 4, 2 unch. · avg +15.1% · Bytecode: 🔴 6 · avg -34.5%
tsv.js — Interp: 🟢 3, 🔴 1, 5 unch. · avg +3.2% · Bytecode: 🔴 9 · avg -26.8%
typed-arrays.js — Interp: 🟢 18, 4 unch. · avg +29.5% · Bytecode: 🟢 1, 🔴 21 · avg -36.3%
uint8array-encoding.js — Interp: 🟢 11, 🔴 1, 6 unch. · avg +13.2% · Bytecode: 🟢 2, 🔴 14, 2 unch. · avg -21.3%
weak-collections.js — Interp: 🟢 1, 🔴 8, 6 unch. · avg -16.6% · Bytecode: 🟢 3, 🔴 10, 2 unch. · avg -22.8%
Deterministic profile diffDeterministic profile diff: no significant changes. Measured on ubuntu-latest x64. Benchmark ranges compare cached main-branch min/max ops/sec with the PR run; overlapping ranges are treated as unchanged noise. Percentage deltas are secondary context. |
Contributor
Suite TimingTest Runner (interpreted: 10,834 passed; bytecode: 10,834 passed)
MemoryGC rows aggregate the main thread plus all worker thread-local GCs. Test runner worker shutdown frees thread-local heaps in bulk; that shutdown reclamation is not counted as GC collections or collected objects.
Benchmarks (interpreted: 430; bytecode: 430)
MemoryGC rows aggregate the main thread plus all worker thread-local GCs. Benchmark runner performs explicit between-file collections, so collection and collected-object counts can be much higher than the test runner.
Measured on ubuntu-latest x64. |
Contributor
test262 Conformance
Areas closest to 100%
Per-test deltas (+83 / -5)Newly failing (5):
Newly passing (83):
Steady-state failures are non-blocking; regressions vs the cached main baseline (lower total pass count, or any PASS → non-PASS transition) fail the conformance gate. Measured on ubuntu-latest x64, bytecode mode. Areas grouped by the first two test262 path components; minimum 25 attempted tests, areas already at 100% excluded. Δ vs main compares against the most recent cached |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Implements the TC39 ShadowRealm proposal (stage 2.7) as an opt-in core language built-in, gated by a new
--unsafe-shadowrealmflag /"unsafe-shadowrealm": trueconfig key. It is off by default becauseShadowRealm.prototype.evaluateperforms dynamic source evaluation, which GocciaScript keeps out of normal runtimes — the same explicit-capability pattern as--unsafe-function-constructor.What's implemented
globalThis.ShadowRealmconstructor — engine-owned, installed only when the flag is set (genuinely absent otherwise).newrequired (ShadowRealm()throwsTypeError);length0,name"ShadowRealm",[[Prototype]]=Function.prototype;ShadowRealm.prototype[Symbol.toStringTag]="ShadowRealm".ShadowRealm.prototype.evaluate(sourceText)— runs source in a fresh child realm (its own intrinsics + global object) using eval-style declaration instantiation:var/functionpersist on the realm'sglobalThis,let/constare per-call. Parse failures throw a caller-realmSyntaxError; runtime errors throw a caller-realmTypeError. Nested realms are supported.TypeError, callables are wrapped (name/length copied, arguments/results marshaled across realms, thrown errors re-wrapped into the caller realm).ShadowRealm/realms-testsfeatures map to--unsafe-shadowrealmin the conformance runner.Key implementation notes
TGocciaEnginefreed at the creating engine's teardown (a GC sweep can't safely free a child engine from a collected instance). See ADR 0070.InvokeCallablenow routes callable non-function/class values (e.g. callable proxies) through the proxy apply hook so wrapped callable proxies work — previously they raised "object is not a function".Deferred / out of scope (follow-ups)
ShadowRealm.prototype.importValue(module import in the child realm) — a separate proposal method, not in this issue's expected behavior (~12 test262 cases).eval, consistent with the no-eval default.Symbol.forregistry sharing — a pre-existing per-realm limitation also visible via$262.createRealm.arguments) and native-functionlengthdeletability are pre-existing engine limitations surfaced by a few test262 cases.FWrappedHostsretains wrapped-function state for the realm's lifetime (memory bounded by engine lifetime; candidate for a future GC-owned wrapper).Closes #777
Testing
tests/built-ins/ShadowRealm/**tests. test262built-ins/ShadowRealmat the pinned SHA: 48/64 (was 0/64), 0 wrapper-infra failures; the remaining 16 are the deferredimportValue(12) and the pre-existing-limitation edge cases (4) noted above../format.pas --checkclean.docs/built-ins.md,docs/language-tables.md,docs/build-system.md, and ADR 0070.🤖 Generated with Claude Code