Nightly research PR triage (2026-06-25): merge #603/#602/#604, scope #601, pass #600
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 4.5k
- Forks
- 603
- Avg merge
- 23h 32m
- Merged PRs (30d)
- 59
Description
Nightly research PR triage (2026-06-25) — merge the green, fix the trivial, prune the rest
Triage of the latest nightly research PRs (#600–#604) with merge decisions. Tracking the three actionable merges plus the cross-cutting CI bug.
Merge decisions
| PR | Topic | Verdict | Action |
|---|---|---|---|
| #603 | TimesFM 1.0 200M candle port | ✅ Merge now | Real transformer impl + RevIN bug fix + numerical regression tests; only known-broken shards red. Publish after PyTorch weight-parity check. |
| #602 | SPANN partition-spill | 🟡 Fix → merge | One-line bug: crate Cargo.toml declares its own [workspace] → two workspace roots → all CI red. Remove it + drop nested Cargo.lock. Result is real: +57% recall@nprobe=4 vs hard IVF. |
| #604 | Capability-gated ANN | 🟡 Fix → merge | Only clippy/fmt cleanup (2 warnings + hex-grouping nit). 7.9× speedup @ 12.5% access, 100% recall. Novel: symmetric complement to ADR-227 proof-gated writes. |
| #601 | Semantic cache | 🟠 Scope down | Headline 13.49× is cherry-picked (1.55× at realistic 50% hit); Adaptive variant is slower than no-cache. Not merged here — needs adaptive variant dropped + honest reframing. |
| #600 | Diversity rerank | 🔴 Pass | Reimplements MMR (1998); advertised diversity gains hide recall@10 → 0.100. Not merged. |
Cross-cutting CI bug (fix separately)
Tests (research-nightly) shard fails on every nightly PR with package ID specification 'ruvector-maxsim' did not match any packages — the test matrix hardcodes a crate that no longer exists. This is CI-matrix rot, not a real per-PR failure. Should be fixed in the workflow so nightly PRs get real signal.
Checklist
- Merge #603 (TimesFM)
- Fix + merge #602 (SPANN workspace-root)
- Fix + merge #604 (capgated clippy/fmt)
- (follow-up) Fix
research-nightlyCI matrixruvector-maxsimreference
🤖 Generated with claude-flow
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Review PRs #600–#604 and the research-nightly CI workflow first. Verify the listed merge, scope, and pass decisions, then address the SPANN workspace-root issue, capgated clippy/fmt cleanup, and stale ruvector-maxsim matrix reference. Done means the appropriate PRs are merged or scoped as stated and nightly CI reports real results.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- github-actions, rust
- Domain
- build-system, ci-cd
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100