v1.1.15 — inliner: acyclic CF (#226) + multi-site small leaves & whole-function DCE (#228)#229

Merged

avrabe merged 16 commits into

mainfrom

release/v1.1.15-acyclic-inline

Jun 22, 2026

Commits on Jun 22, 2026

test(219): add seam-SROA grounding fixture (pack/unpack round-trip)
avrabe
committed
test(219): add gale's verified real repro (sem.loom.wasm) + notes
avrabe
committed
feat(219): ISLE term scaffolding for seam-SROA (i32.wrap_i64 / i64.extend_i32_u)
avrabe
committed
feat(219): live seam-SROA rewrite — wrap_i64(extend_i32_u(x)) → x
avrabe
committed
feat(219): step-3 masked pack/unpack SROA rules (the &mask and >>k halves)
avrabe
committed
docs(219): PR-C design — precise acyclic CF symbolic execution (verifier)
avrabe
committed
feat(219): PR-C M2 — is_noreturn_callee classifier (divergent-call detection)
avrabe
committed
feat(219): PR-C M1a — precise acyclic CF executor (isolated, Z3-tested)

avrabe
and
claude
committed
feat(219): PR-C M1b — Call by-body modeling + no-return diverge in the acyclic executor

avrabe
and
claude
committed
feat(219): PR-C M3.1 — verify_acyclic_equivalence precise fast-path

avrabe
and
claude
committed
feat(219): PR-C M3.2 — wire the precise acyclic fast-path into the inliner

avrabe
and
claude
committed
feat(219): PR-C M3.3 — admit acyclic-CF callees as inline candidates + Unreachable trap

avrabe
and
claude
committed
feat(219): PR-C Phase 2a — thread memory Array through the acyclic executor

avrabe
and
claude
committed
feat(219): PR-C Phase 2b — impure-call havoc in the acyclic executor

avrabe
and
claude
committed
fix(228): inline modelable multi-site small leaves + expose whole-function DCE

avrabe
and
claude
committed
chore(release): v1.1.15 — inliner: acyclic CF (#226) + multi-site leaves & whole-function DCE (#228)

avrabe
and
claude
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

v1.1.15 — inliner: acyclic CF (#226) + multi-site small leaves & whole-function DCE (#228)#229

v1.1.15 — inliner: acyclic CF (#226) + multi-site small leaves & whole-function DCE (#228)#229
avrabe merged 16 commits into
mainfrom
release/v1.1.15-acyclic-inline

Commits on Jun 22, 2026

test(219): add seam-SROA grounding fixture (pack/unpack round-trip)

test(219): add gale's verified real repro (sem.loom.wasm) + notes

feat(219): ISLE term scaffolding for seam-SROA (i32.wrap_i64 / i64.extend_i32_u)

feat(219): live seam-SROA rewrite — wrap_i64(extend_i32_u(x)) → x

feat(219): step-3 masked pack/unpack SROA rules (the &mask and >>k halves)

docs(219): PR-C design — precise acyclic CF symbolic execution (verifier)

feat(219): PR-C M2 — is_noreturn_callee classifier (divergent-call detection)

feat(219): PR-C M1a — precise acyclic CF executor (isolated, Z3-tested)

feat(219): PR-C M1b — Call by-body modeling + no-return diverge in the acyclic executor

feat(219): PR-C M3.1 — verify_acyclic_equivalence precise fast-path

feat(219): PR-C M3.2 — wire the precise acyclic fast-path into the inliner

feat(219): PR-C M3.3 — admit acyclic-CF callees as inline candidates + Unreachable trap

feat(219): PR-C Phase 2a — thread memory Array through the acyclic executor

feat(219): PR-C Phase 2b — impure-call havoc in the acyclic executor

fix(228): inline modelable multi-site small leaves + expose whole-function DCE

chore(release): v1.1.15 — inliner: acyclic CF (#226) + multi-site leaves & whole-function DCE (#228)