Skip to content

ci: restart the nightly size history after the libmoq rename - #9071

Open
kixelated wants to merge 2 commits into
mainfrom
quest/m2/libmoq-size-history
Open

kixelated wants to merge 2 commits into
mainfrom
quest/m2/libmoq-size-history

Conversation

@kixelated

@kixelated kixelated commented Nov 18, 2026 •

Copy link
Copy Markdown
Collaborator

Restarts Nightly's size history now that moq-dev/moq#8941 renamed the hand-written C archives from moq-c-<target> to libmoq-<target> and stopped matching moq-c-generated-<target>. Finishes quest/m2/libmoq-size-history.md.

Changes

The restart drops every series' 30-day baseline: for the next 30 days each night compares only against the previous night.

Decisions

  • Deleted the artifact before tonight's Nightly ran instead of waiting for a red run to prove the refusal. A local simulation showed the same thing (below), and waiting would only have cost one failed night.
  • Didn't dispatch a Nightly by hand to confirm: it rebuilds every release target. The next scheduled run is the confirmation. If its Collect and compare sizes step fails, the cause is something else, not the rename.

Validation

  • Downloaded the deleted artifact first (kept locally) and confirmed its history.json keys name release/moq-c-<target>/… and release/moq-c-generated-<target>/….
  • Rewrote that night's sizes.json into the post-feat(c): the hand-written C API ships as libmoq, its final line #8941 shape (moq-c-* became libmoq-*, moq-c-generated-* dropped) and ran sh/size/report.py on it:
    • with --history set to the retained history: exit 1, artifacts missing or renamed: release/moq-c-… (all 8 old keys), which is the failure the quest predicted.
    • without history: exit 0, "a new artifact history starts with one report", with the four release/libmoq-<target>/… keys recorded.
  • gh api …/actions/artifacts?name=nightly-size-history now returns total_count: 0.
  • Upload names on main match: libmoq.yml uploads libmoq-${{ matrix.target }} and sh/size/release.py collects libmoq-*.
  • Not yet observed: the next scheduled Nightly passing Collect and compare sizes.

Impact

No public API or wire change. CI state only: Nightly's size baseline restarts.

Caused by

  • moq-dev/moq#8941 - renamed the release artifacts, which the retained history can't follow by design (found by person, planned in the quest before merge)

Follow-ups

None.

Review

Skipped on 6ffcc68: the diff only deletes the finished quest file, and the work itself (deleting one CI artifact) has no code to review; its evidence is under Validation.

Models

Size: XS
Implementer: Claude Code, claude-opus-5-5, unknown effort (arm B)

(Written by Opus 5.5)

kixelated and others added 2 commits November 17, 2026 20:01
Deleted the one retained nightly-size-history artifact (run 44127032469,
cd9a8e9), which still named the moq-c-<target> and
moq-c-generated-<target> keys, so the next Nightly starts a fresh history.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@kixelated kixelated changed the title Restart the nightly size history after the libmoq rename ci: restart the nightly size history after the libmoq rename Nov 18, 2026
@kixelated
kixelated marked this pull request as ready for review November 18, 2026 04:03
@coderabbitai

coderabbitai Bot commented Nov 18, 2026

Copy link
Copy Markdown
Contributor

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 3 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used all 4 included reviews currently available.

Learn how review limits work.

Review configuration:

Run configuration

  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: Organization UI
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: 377f473d-45d3-4cf8-aa39-c5457587432e
Commits

Reviewing files that changed from the base of the PR and between f1eecd3 and 6ffcc68.

Files selected for processing: 1

  • quest/m2/libmoq-size-history.md
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

@kixelated
kixelated enabled auto-merge November 18, 2026 04:04

@pullfrog pullfrog Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Important

The restart also permanently stops measuring the four moq-c-generated-<target> archives, which still ship. #8941 replaced the moq-c-* glob in sh/size/release.py with libmoq-*, and that old glob was what matched the generated C package's artifacts that release-c.yml keeps publishing on c-v* tags. Fresh history means report.py's "missing or renamed" refusal can never name them again, so this is the cheapest moment to restore the pattern.

Approach: sound — the history can't follow a rename by design, and deleting the retained artifact is the documented restart in sh/size/README.md, backed by a local simulation instead of a wasted red run. The gap is that the same restart silently retires a series that is still shipping.

Reviewed changes

  • Quest file deleted — quest/m2/libmoq-size-history.md recorded the merged plan and its waiting state; nothing else in the repo links to it, so removing it finishes the quest.
  • CI artifact deleted (outside the diff, verified) — actions/artifacts?name=nightly-size-history now returns total_count: 0, and sh/size/nightly.sh therefore passes no --history, so the next Nightly writes a one-row history as described.
  • Rename coverage checked — libmoq.yml uploads libmoq-${{ matrix.target }} and sh/size/release.py collects libmoq-*, matching the post-#8941 upload names.

⚠️ The restart drops the moq-c-generated-* series, which is still a shipping artifact

Last night's report (run 44127032469, bench-size artifact) carries 8 C keys: four size/release/moq-c-<target>/… and four size/release/moq-c-generated-<target>/…. The second four were only ever collected because moq-c-* matched them, and no pattern in ARTIFACTS matches them now, while release-c.yml still publishes the c/moq package on c-v* tags. Restarting the history is what makes that loss unobservable: compare() refuses removed keys, but only keys the retained history knows about.

Technical details
# Generated C package's size series disappears with the restart

## Affected sites
- `sh/size/release.py:13` — `ARTIFACTS` has no pattern matching `moq-c-generated-*`; `libmoq-*` replaced the `moq-c-*` glob that used to match both the hand-written and generated uploads.
- `.github/workflows/release-c.yml:124` — still uploads `moq-c-generated-${{ matrix.target }}`, and its `release` job publishes those archives on `c-v*` tags.
- `.github/workflows/nightly.yml:778` — the `sizes` job downloads artifacts from `release-c` but does not list it in `needs`.
- `sh/size/report.py:40` — the "artifacts missing or renamed" refusal only compares against keys the history contains; with an empty history it cannot report a dropped series.

## Required outcome
- Either the four `moq-c-generated-<target>` archives come back under comparison, or the PR records that the generated package is deliberately no longer size-tracked (including that it still ships).
- The choice must be made before the fresh history is written, because a pattern added after tonight's run starts the series from that later night.

## Suggested approach (optional)
- Add `"moq-c-generated-*"` to `ARTIFACTS` so the first fresh report records it alongside `libmoq-*`; `sizes()` already keys on the artifact directory, so there is no duplicate-key clash with `libmoq-*`.
- Adding `release-c` to the `sizes` job's `needs` would remove the reliance on that job happening to finish before the download, which currently works only by timing.

## Open questions for the human (optional)
- Is dropping the generated package's measurement intended, given `quest/m2/README.md` keeps "the generated C line" for after the cut?

Pullfrog  | Fix it ➔ | View workflow run | Using deepseek-v4.1-flash | 𝕏

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant