Implement ECMAScript realm execution contexts#703
Merged
Conversation
Add explicit realm and execution-context plumbing across interpreter and bytecode execution, move tagged template caching into realm-owned template maps, and cover repeated engine execution in CLI and native tests. Verification: - ./build/Goccia.Engine.Realm.Test - ./build/GocciaTestRunner tests --no-progress - ./build/GocciaTestRunner tests --no-progress --mode=bytecode - bun run scripts/test-cli-apps.ts - bun run scripts/test-cli.ts - ./format.pas --check
|
The latest updates on your projects. Learn more about Vercel for GitHub. 1 Skipped Deployment
|
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (5)
🚧 Files skipped from review as they are similar to previous changes (3)
📝 WalkthroughWalkthroughThis PR centralizes realm ownership and ECMA-262-style execution contexts: introduces TGocciaExecutionContext/stack and RAII scopes, expands TGocciaRealm to own intrinsics/global/env/loaded-module/template maps, adds template-site IDs and realm-scoped template caching, wires realm propagation through engine/interpreter/executor/VM/frame lifecycles, updates tests/docs, and lowers default stack depth. ChangesRealm Isolation and Execution Context Integration
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Possibly related PRs
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Comment |
Contributor
Suite TimingTest Runner (interpreted: 9,985 passed; bytecode: 9,985 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: 407; bytecode: 407)
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
Benchmark Results407 benchmarks Interpreted: 🟢 25 improved · 🔴 43 regressed · 339 unchanged · avg -0.1% arraybuffer.js — Interp: 🔴 5, 9 unch. · avg -5.1% · Bytecode: 🔴 9, 5 unch. · avg -9.4%
arrays.js — Interp: 🟢 4, 15 unch. · avg +2.0% · Bytecode: 🔴 12, 7 unch. · avg -9.9%
async-await.js — Interp: 🟢 1, 5 unch. · avg +2.6% · Bytecode: 🔴 1, 5 unch. · avg -6.0%
async-generators.js — Interp: 2 unch. · avg +7.5% · Bytecode: 🔴 1, 1 unch. · avg -4.1%
base64.js — Interp: 10 unch. · avg -1.1% · Bytecode: 🔴 2, 8 unch. · avg -1.4%
classes.js — Interp: 🟢 1, 30 unch. · avg -0.5% · Bytecode: 🔴 12, 19 unch. · avg -4.8%
closures.js — Interp: 🔴 2, 9 unch. · avg -3.3% · Bytecode: 🔴 11 · avg -20.0%
collections.js — Interp: 12 unch. · avg -1.5% · Bytecode: 🔴 1, 11 unch. · avg -3.7%
csv.js — Interp: 13 unch. · avg -1.3% · Bytecode: 🔴 3, 10 unch. · avg -2.2%
destructuring.js — Interp: 🔴 1, 21 unch. · avg -2.4% · Bytecode: 🔴 8, 14 unch. · avg -7.7%
fibonacci.js — Interp: 8 unch. · avg -2.6% · Bytecode: 🔴 6, 2 unch. · avg -18.2%
float16array.js — Interp: 🔴 4, 28 unch. · avg -1.7% · Bytecode: 🟢 4, 🔴 11, 17 unch. · avg -3.7%
for-of.js — Interp: 7 unch. · avg -0.3% · Bytecode: 🔴 2, 5 unch. · avg -6.3%
generators.js — Interp: 🟢 1, 🔴 1, 2 unch. · avg -0.3% · Bytecode: 🔴 1, 3 unch. · avg -6.2%
iterators.js — Interp: 🟢 1, 🔴 1, 40 unch. · avg +1.0% · Bytecode: 🟢 4, 🔴 11, 27 unch. · avg -2.8%
json.js — Interp: 🟢 2, 18 unch. · avg +0.8% · Bytecode: 🔴 8, 12 unch. · avg -5.8%
jsx.jsx — Interp: 🔴 1, 20 unch. · avg -0.7% · Bytecode: 🔴 9, 12 unch. · avg -7.4%
modules.js — Interp: 🔴 1, 8 unch. · avg -5.8% · Bytecode: 🔴 7, 2 unch. · avg -34.9%
numbers.js — Interp: 11 unch. · avg -2.6% · Bytecode: 🔴 6, 5 unch. · avg -7.1%
objects.js — Interp: 7 unch. · avg +2.9% · Bytecode: 🔴 4, 3 unch. · avg -13.5%
promises.js — Interp: 🔴 1, 11 unch. · avg -2.5% · Bytecode: 🔴 5, 7 unch. · avg -8.9%
regexp.js — Interp: 11 unch. · avg +0.3% · Bytecode: 🔴 2, 9 unch. · avg -4.8%
strings.js — Interp: 🔴 5, 14 unch. · avg -12.7% · Bytecode: 🔴 11, 8 unch. · avg -17.9%
tsv.js — Interp: 🔴 3, 6 unch. · avg -0.6% · Bytecode: 🔴 3, 6 unch. · avg -9.0%
typed-arrays.js — Interp: 🔴 15, 7 unch. · avg -24.8% · Bytecode: 🟢 3, 🔴 8, 11 unch. · avg -1.8%
uint8array-encoding.js — Interp: 🟢 10, 8 unch. · avg +42.7% · Bytecode: 🟢 7, 🔴 9, 2 unch. · avg +15.2%
weak-collections.js — Interp: 🟢 5, 🔴 3, 7 unch. · avg +16.3% · Bytecode: 🟢 4, 🔴 4, 7 unch. · avg +11.9%
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
test262 Conformance
Areas closest to 100%
Per-test deltas (+1 / -0)Newly passing (1):
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 |
coderabbitai
Bot
added
documentation
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 4
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@source/units/Goccia.ExecutionContext.pas`:
- Around line 123-127: In TGocciaExecutionContextStack.Pop the code currently
zeroes the popped entry with FillChar on
GExecutionContextStack[GExecutionContextStackCount], which bypasses
managed-string finalization for the embedded TGocciaExecutionContext (which has
SourcePath: string); change the cleanup to a managed-safe operation by assigning
GExecutionContextStack[GExecutionContextStackCount] :=
Default(TGocciaExecutionContextStackEntry) (or call Finalize(...) then
Default(...)) instead of FillChar so SourcePath and other managed fields are
finalized properly; update the Pop implementation that references
GExecutionContextStack and GExecutionContextStackCount accordingly.
In `@source/units/Goccia.Executor.Bytecode.pas`:
- Around line 120-128: The code currently overwrites the VM realm on exit by
resetting FVM.Realm to FRealm; instead capture the prior realm before changing
it and restore that saved realm in the finally block. Specifically, save the
existing realm (e.g., SavedRealm := FVM.Realm) before assigning FVM.Realm :=
AContext.Realm in the try section, and in the finally block restore FVM.Realm :=
SavedRealm (leaving the existing FVM.GlobalScope save/restore using
SavedGlobalScope and the AContext.Scope handling unchanged). This ensures nested
ExecuteModule calls (SavedGlobalScope, AContext.Scope, AContext.Realm,
FVM.Realm) correctly restore the outer realm rather than forcing FRealm.
In `@source/units/Goccia.Realm.Test.pas`:
- Around line 378-405: The test resets the global CurrentRealm to nil in the
finally block which can clobber an existing thread realm; modify
TTestRealm.TestExecutionContextStackSetsCurrentRealm to capture the prior realm
into a local variable (e.g., PrevRealm) before calling
SetCurrentRealm(OuterRealm) and then in the finally restore it via
SetCurrentRealm(PrevRealm) instead of SetCurrentRealm(nil); keep the rest of the
teardown (freeing Guard, InnerRealm, OuterRealm) intact and ensure you reference
SetCurrentRealm, CurrentRealm, CreateExecutionContext and
RunningExecutionContext when making the change.
In `@source/units/Goccia.VM.pas`:
- Around line 7715-7720: The code pushes a new execution context using
FCurrentModuleSourcePath (caller module) causing source metadata to attribute to
the caller; change the source-file argument passed into CreateExecutionContext
to prefer the callee template's debug-info source: use
ATemplate.DebugInfo.SourceFile or AClosure.Template.DebugInfo.SourceFile when
available and only fall back to FCurrentModuleSourcePath if debug info is
missing; update the TGocciaExecutionContextStack.Push call in the
APushExecutionContext branch (and the similar block at the OP_IMPORT_META
location) to pass that resolved source-file string instead of
FCurrentModuleSourcePath so contexts reflect the defining module.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: df75f0cd-2eaa-47a4-a89d-a96777344fff
📒 Files selected for processing (28)
CONTEXT.mddocs/architecture.mddocs/build-system.mddocs/bytecode-vm.mddocs/core-patterns.mddocs/embedding.mdscripts/test-cli-apps.tssource/units/Goccia.AST.Expressions.passource/units/Goccia.Bytecode.Chunk.passource/units/Goccia.Engine.Realm.Test.passource/units/Goccia.Engine.passource/units/Goccia.Evaluator.Context.passource/units/Goccia.Evaluator.passource/units/Goccia.ExecutionContext.passource/units/Goccia.Executor.Bytecode.passource/units/Goccia.Executor.Interpreter.passource/units/Goccia.Executor.passource/units/Goccia.Interpreter.passource/units/Goccia.Modules.Loader.passource/units/Goccia.Realm.Test.passource/units/Goccia.Realm.passource/units/Goccia.StackLimit.passource/units/Goccia.VM.CallFrame.passource/units/Goccia.VM.Closure.passource/units/Goccia.VM.passource/units/Goccia.Values.FunctionValue.passource/units/Goccia.Values.GeneratorValue.pastests/language/functions/stack-depth-limit.js
- clear execution-context stack records without FillChar on managed fields - restore the previous bytecode VM realm after nested module execution - restore prior CurrentRealm in realm stack tests - attribute bytecode function contexts to callee debug source paths
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
RangeErrorinstead of native stack failure.Testing
Commands run:
./build.pas clean testrunner loader repl bundler benchmarkrunner loaderbare tests./build.pas clean tests testrunner./build/Goccia.Realm.Test./build/Goccia.Engine.Realm.Test./build/GocciaTestRunner tests --no-progress./build/GocciaTestRunner tests --no-progress --mode=bytecodebun run scripts/test-cli-apps.tsbun run scripts/test-cli.ts./format.pas --checkgit diff --check