[devin-ai-integration]

What and why

Since #1276 (an object is typed by every type its feature inherits through subsetting), Context.effectiveTypes is asked for the same feature symbol on every object made as that feature: extractType, directTypes, conjoinedTypes and withConjoinedTypes each call it, and each call re-resolved the feature's typing relationships (declaredTypes → RelationshipsOf → resolveQualified), re-ran FeatureTypeSet over the subsetted features and rebuilt the conjunction. In the 0.11.0 release-gate measurement this was the largest avoidable part of the tests/stressmodel Satisfy/FleetSatisfy and tests/perf Instantiate/FeaturesOf/BatchSatisfy allocation growth: a 300-iteration allocation profile of Satisfy/satellites=32 at #1276's merge commit puts 2.1M of the 2.4M new allocations under declaredTypes/statedTypes and the slices.Clone in effectiveTypes.

effectiveTypes is now memoized per feature symbol on the runtime Model, beside its other per-symbol tables (features, chainHosts, denotedFeatures); the computation moved unchanged to computeEffectiveTypes. The memoized slice is shared, so conjoinedTypes clones before slices.DeleteFunc (the other callers only read). No behaviour changes: the same types in the same order for every call.

How it was verified

• Failing-before differential test TestEffectiveTypesAreComputedOncePerFeature (internal/exec/runtime/effective_types_memo_test.go): a repeated effectiveTypes query for test::Rig::front allocates 755 times on develop d8b60ed and 0 with this change.

• Six interleaved rounds, benchstat, same machine (Intel Xeon, 8 vCPU, Go 1.25.11); pre-fix(runtime): type an object by every type its feature inherits through subsetting #1276 bb293df | develop d8b60ed + fix(semantics): derive a declaration's implied generals once per symbol #1350 | this change + fix(semantics): derive a declaration's implied generals once per symbol #1350. Allocations and bytes are exact (±0%, p=0.002 throughout); sec/op was taken on a loaded machine (the release-gate run was sharing it) and is not reported here.

  row	pre-fix(runtime): type an object by every type its feature inherits through subsetting #1276	develop + fix(semantics): derive a declaration's implied generals once per symbol #1350	this PR + fix(semantics): derive a declaration's implied generals once per symbol #1350
  tests/perf Instantiate allocs/op	1.995M	3.333M (+67%)	2.633M (+32%)
  tests/perf FeaturesOf allocs/op	4.189M	6.725M (+61%)	5.177M (+24%)
  tests/perf BatchSatisfy allocs/op	2.789M	4.889M (+75%)	3.736M (+34%)
  tests/perf Instantiate B/op	187.8Mi	224.3Mi (+19%)	207.4Mi (+10%)
  tests/stressmodel Satisfy/satellites=32/assertions=96 allocs/op	33.77k	54.48k (+61%)	41.07k (+22%)
  tests/stressmodel FleetSatisfy/satellites=32/assertions=24 allocs/op	9.069k	15.365k (+69%)	11.367k (+25%)
  tests/stressmodel FleetInstantiate/satellites=32/elements=1179 allocs/op	14.49k	16.69k (+15%)	15.45k (+7%)
  tests/stressmodel Satisfy/satellites=32/assertions=96 B/op	1.866Mi	2.193Mi (+17%)	1.976Mi (+6%)

  The residual over pre-fix(runtime): type an object by every type its feature inherits through subsetting #1276 is the inherent cost of fix(runtime): type an object by every type its feature inherits through subsetting #1276 and feat(semantics): usages subset the base library usage of their kind #1319 (FeatureTypeSet over the subsetted features, the w9c inherited-name check over the wider type set), measured and attributed in the 0.11.0 performance record (docs(perf): measure the 0.11.0 candidate against 0.10.1 #1337).

• go build ./..., go vet ./..., gofmt -l . empty, go test ./internal/exec/runtime (219 s, all pass), python3 scripts/changelog.py check.

Checklist

• make test and make lint pass locally (runtime package run locally; the full suite runs in CI)
• Tests added or updated for the change
• Documentation extended where it already covers the surface (none covers this internal memo)
• Changelog entry added as changes/unreleased/<slug>.<section>.md, not as an edit to CHANGELOG.md
• baselines regenerated and make docs-counts run if a gate count moved (no gate count moved)
• No internal work-item labels (waves, slices, F4, K5) in the body, docs, or changelog

Link to Devin session: https://nasa-jpl-demo.devinenterprise.com/sessions/cf22d7a746814d20b2f0eaf11682fef9
Open in Devin Desktop: https://nasa-jpl-demo.devinenterprise.com/desktop/session/cf22d7a746814d20b2f0eaf11682fef9?variant=devin
Requested by: @HuiJun

---

[devin-ai-integration]

I'll fix CI failures and address comments from users with write access. I'll skip comments containing "(aside)".

• Disable automatic comment, CI, and merge conflict monitoring

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no bugs or issues to report.