[sunerpy]

Ledger entry for the round that shipped v0.43.2 → v0.45.0.

Why this round existed at all

git ls-remote --tags still reports v1.5.0 as upstream's newest tag, so a version-delta sync had nothing to pull. But main had accumulated 104 unreleased commits past that tag (43 source files, +5737/−368), and issues #1496–#1566 / PRs #1402–#1564 had never been triaged.

"No new version" is not "no new content". The header now instructs the next sync to diff v1.5.0..origin/main too, so this is not rediscovered from scratch.

What the entry records

release	PR	what landed
v0.43.2	#209	one shared file-classification signal across every surface
v0.43.3	#211	stop binding calls, imports and supertypes to coincidences
v0.44.0	#213	unions, unit structs and COM interface as containers
v0.44.1	#215	imported members, cross-module Go calls, declaration initializers
v0.45.0	#217	Tauri commands bound to their invoke call sites

Plus the dispositions that matter more than the ports:

• Two agent findings refuted by measurement. Candidate ordering and the explore staleness banner were both already correct — the first because both production resolution contexts sort themselves with a stricter key than upstream's, the second because the banner is prepended after truncation and a cut section's citation is dropped first. Recorded so the next sync does not re-investigate.
• One upstream premise upstream itself rejected (CG-27), with its numbers: delivered chars 7,539 → 397, inner closures 7/11 → 0/11.
• Two permanent divergences — self-hosted telemetry and the #1509 vector-RAG request — which scripts/guardrail.sh makes structural rather than a matter of preference.
• Four issues that do not reproduce here, each with the measurement that closed it, and eight that are TypeScript/Node-only.
• Where we are ahead: upstream is only now adding GDScript, which we shipped with a full framework resolver and a byte-stable golden corpus.

Verification

Every number in the entry was re-derived before committing: 104 commits and 43 files changed, 5737 insertions(+), 368 deletions(-) from the upstream clone, upstream's tag still v1.5.0, and each cited code location re-checked — two line references were corrected in the process (context.rs:111,125 and snapshot_context.rs:118-124, not the numbers I first wrote).

Docs only: one file, oxfmt idempotent. Tracked parity stays 1.5.0 — none of these releases ports an upstream release range.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

@@           Coverage Diff           @@
##             main     #219   +/-   ##
=======================================
  Coverage   95.28%   95.28%           
=======================================
  Files         140      140           
  Lines       77287    77287           
=======================================
+ Hits        73643    73644    +1     
+ Misses       3644     3643    -1     

see 1 file with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.