release: v0.7.0#104
Merged
Merged
Conversation
Bump workspace version 0.6.0 → 0.7.0. Infrastructure-and-cleanup release. Closes the trivial pipeline-order gap in v0.6.0's vacuum, lands function-summary interprocedural analysis (IPA) as the foundation for cross-call optimization, and adds verification-aware canonicalization as the IR-normalization substrate that downstream passes (and the Z3 verifier) compose better against. Merged PRs: #101 fix(pipeline): run vacuum after dead-stores/dead-locals (-1.97% on gale) #102 feat(opt): function-summary IPA + pure-zero-arg call/drop folding #103 feat(opt): canonicalize pass — if/else→select + local.tee normalization Measured impact on gale_ffi (1.9 KB kernel FFI): baseline: code section 811 bytes v0.5.0 (regression): code section 862 bytes (+6.3%) v0.6.0: code section 804 bytes (-0.86%) v0.7.0 (this): code section 795 bytes (-1.97%) Test count: 317 → 288+ in loom-core lib alone (cross-crate total higher; new tests dominate from the three PRs). The v0.7.0 release is infrastructure-heavy: PR-F and PR-G ship the analysis substrate (function summaries + canonical IR forms) that PR-H+ (v0.8.0) needs to compound on. PR-E was the direct byte win.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Bump workspace version 0.6.0 → 0.7.0.
Infrastructure-and-cleanup release. Closes the trivial pipeline-order gap in v0.6.0's vacuum, lands function-summary interprocedural analysis (IPA) as the foundation for cross-call optimization, and adds verification-aware canonicalization as the IR-normalization substrate that downstream passes (and the Z3 verifier) compose better against.
Merged PRs
Measured impact on gale_ffi
A further +1.1 percentage points from the pipeline-order fix (30 LOC). PR-F and PR-G are infrastructure — measurable wins from them will compound in v0.8.0 (CSE cross-call dedup, broader if/select, arg-aware call-drop fold).
Why infrastructure-heavy
PR-F builds function summaries — a function-purity / no-trap analysis that every cross-call optimization needs. Without it, every
Callis an opaque side-effecting wall. With it: CSE can dedupe pure calls, DCE can drop pure calls with unused results, vacuum can fold pure call+drop pairs. The v0.7.0 consumer is conservative (zero-arg only), but the analysis is reusable.PR-G builds canonical IR forms —
if/else → select(path-INSENSITIVE) andlocal.teenormalization. The Z3 verifier and every downstream pass now sees a canonical shape, which is a force-multiplier for verification confidence and rewrite quality.PR-E was the direct byte win.
Research artifacts
Four v0.7.0-planning research documents under
docs/research/v0.7.0/:issue-triage-68-75.md— verdicts on Cross-component optimization passes for meld-fused modules #68-Optimize P3 async callback trampolines and stream buffer access patterns #75optimization-methods-survey.md— 13-family survey, v0.7.0/v0.8.0 plangale-deep-scan.md— gale-specific opportunities + Verus axiom shortlistalgorithmic-solver-feasibility.md— Souper-shaped SMT peephole synthesis verdictCHANGELOG
Detailed v0.7.0 entry already in
CHANGELOG.md(this PR).🤖 Generated with Claude Code