Repository navigation
chore(register): C-44 — the release gate has moved; record which half - #144
Merged
Merged
Conversation
C-44 holds the pipeline-core 3.0.0 bump on two conditions. Their status has diverged and nobody had noticed, so the entry read as "both still closed." Condition 2 (platform running smoothly on development across consumer repos) now has substantial evidence: run-0 on 2026-07-27 exercised the whole FAO delivery chain on development — views-models config -> pipeline-core frame-native fetch and Track A publish -> this repo's contract delivery (108 shards, 64,742 cells, 28.4M rows, 5.6 GB peak) -> faoapi ingest, confirmed serving live on 2026-07-31 (smoke ALL PASS, IDN present, i.e. outside the old Africa-only region). Scoped honestly in the entry: that covers the FAO delivery chain across five repos, NOT consumers outside it (views-reporting's report path, views-baseline, other model families). Whether "all consumer repos" means the chain or literally everything is a maintainer judgement the register cannot make. The condition moved from unevidenced to substantially evidenced for the path that matters most — which is a real change, and less than a clearance. Condition 1 (3.0.0 on PyPI) remains unmet and is not blocked on engineering: pipeline-core's development already carries version = "3.0.0" with the heavy deps removed, and views-evaluation's to_metric_frame — the one release-train edge in their runbook #313 — exists in code, 8 commits past v0.4.0, needing a version bump and a publish. Net: one release signal, not two open questions. Explicitly does NOT authorise the bump; the signal is the maintainer's and the scoping caveat is a real reason it might still be withheld. Cross-referenced to C-62, which records what continuing to wait costs. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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.
C-44 holds the pipeline-core 3.0.0 bump on two conditions. Their status has diverged and nobody noticed, so the entry still reads as though both are closed.
Condition 2 — now substantially evidenced
"the platform is confirmed running smoothly on
developmentacross all consumer repos"Run-0, 2026-07-27 exercised the whole FAO delivery chain on
development:Confirmed serving live 2026-07-31 —
scripts/smoke.pyALL PASS at v1.3.11,IDNpresent on both forecast and historical surfaces. Indonesia is outside the oldafrica_me_legacyregion, so that's proof of global serving, not just liveness.Scoped honestly in the entry: this covers the FAO delivery chain across five repos. It is not evidence about consumers outside it — views-reporting's report path, views-baseline, other model families. Whether "all consumer repos" means the chain or literally everything is a maintainer judgement the register can't make.
The condition moved from unevidenced to substantially evidenced for the path that matters most. That's a real change, and deliberately less than a clearance.
Condition 1 — still unmet, and not blocked on engineering
"views-pipeline-core 3.0.0 is published to PyPI"
Verified today:
developmentalready carriesversion = "3.0.0", heavy dependency set already removedto_metric_frame— the one real release-train edge in their runbook (ADR-013 requires a per-run sidecar that is run-independent by construction — and that is an ontology question, not a packaging one #313) — exists in code, 8 commits pastv0.4.0, needing a version bump and a publishTheir runbook fixes the chain: views-frames ✅ → views-evaluation 0.5.0 → views-pipeline-core 3.0.0 → views-reporting 0.3.0 → models/postprocessing — and states it "executes only on Simon's platform-wide signal."
Net
The entry is no longer waiting on two open questions. It's waiting on one release signal, with all the underlying code already written.
This does not authorise the bump. The signal is the maintainer's alone, and the scoping caveat above is a real reason it might still be withheld. It's recorded so the next reader sees a gate that has moved rather than assuming it's still shut.
C-62 records what continuing to wait costs: 3.4 GB of unused dependencies, 2.7 GB of it CUDA libraries for a package that isn't even installed.
Register integrity guard green. No source touched.
🤖 Generated with Claude Code