chore(release): bump to v0.18.0

[avrabe]

v0.18.0 — the network-calculus foundation

The runnable spine that makes spar a tsnkit / TTTech Slate / RTaW-Pegase competitor end-to-end: AADL model → network-wide NC analysis → per-stream delay bounds. Three coherent, merged-to-main requirements:

Req	What	PR
REQ-NC-VALIDATION-001	NC-bound cross-validation against the panco reference (Bouillard)	#279
REQ-NC-PLP-001	PLP (polynomial-LP) network-calculus bounds, arXiv:2010.09263	#280
REQ-NC-BRIDGE-001	AADL instance → network-wide TFA/PLP solver bridge (--pmoo arm)	#283

All three are implemented with passing feature tests and a closed V-trace (TEST-NC-VALIDATION / TEST-NC-PLP / TEST-NC-BRIDGE satisfies; DEC-TSN-OSS-001 the strategic decision) — matching the all-implemented release convention of v0.14–v0.17. All three commits are on main with green CI.

Falsification statement: on a shared multi-hop FIFO TSN line, PLP yields a strictly tighter per-flow delay bound than TFA (measured ~2.7× on a 3-hop line: 39 µs vs 105.5 µs), and never looser on a feed-forward sink-tree — the analysis-dominance wedge a real competitor needs. If PLP ever exceeded TFA on a feed-forward topology, the cross-validation gate (sandwich EXACT ≤ PLP ≤ TFA) would fail.

Release-scope hygiene (deliberate, logged)

v0.18.0 was overloaded with 7 unrelated carry-overs. Moved so the scope is exactly the three NC spine requirements:

• REQ-INGEST-ARXML-001, REQ-INGEST-DBC-FLOWS-001 → v0.19.0 (ingest track, alongside the already-slated REQ-NC-PLP-002 converging-PLP work)
• REQ-PROOF-SCHED-002 → v0.20.0 (proofs track, with REQ-PROOF-NC-MINPLUS-001)
• REQ-SECURITY-001 → backlog (no security theme on the roadmap yet — unscheduled rather than a fake commitment)
• REQ-INST-003, REQ-ANALYSIS-007, COVERAGE-CH12 → backlog (perpetual AS5506-coverage trackers, partial-by-design — were re-blocking every release's readiness query)

Version bump

Workspace 0.17.0 → 0.18.0 (Cargo.toml, Cargo.lock ×23 crates, vscode-spar/package.json). Plus *.lp gitignore — HiGHS/good_lp solver dumps the NC test runs write into the repo root.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[github-actions]

Rivet verification gate

✅ 20/20 passed

	count
Passed	20
Failed	0
Skipped (no steps)	0

Filter: (and (= type "feature") (or (has-tag "v093") (has-tag "v0100")))

Failed artifacts

(none)

_{Updated automatically by tools/post_verification_comment.py. Source of truth: artifacts/verification.yaml.}