ci: per-app vanilla-vs-Pilot A/B report on publish PRs (reopen of #33 against main)#49
Merged
Conversation
Adds .github/workflows/ab-report.yml: when a PR touches submissions/<id>/, it
extracts the app's bundle, runs the adapter in socket mode (no daemon — it
stages artifacts from the R2 registry on startup), runs the equivalent commands
vanilla vs Pilot via cmd/ipc-call, and posts an HTML A/B report (artifact +
upserted PR comment with a summary table).
- scripts/ab_report.py: add --mode {pilotctl,socket}; socket mode drives the
adapter over cmd/ipc-call. Load the command set from
submissions/<id>/ab-commands.json (else a version+help default). Pure stdlib.
- submissions/io.pilot.smolvm/ab-commands.json: CI-safe set (version + top-level
and subcommand --help) — proves the adapter forwards the whole smolvm command
tree (machine/pack/serve/config) without booting a VM.
- docs/CI-AB-REPORT.md: workflow, the no-KVM limitation, per-app command sets.
Validated locally: socket mode against the running smolvm adapter produced a
valid 6-pair report, all exit-matched, ~8-21ms adapter overhead; YAML + JSON +
PR-comment regex checked.
Co-Authored-By: Claude Opus 4.8 (1M context) <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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Reopens #33 against
main. The original PR was merged into the deprecated dev branchfeat/r2-artifacts-on-cli, so its changes never reachedmain. This re-applies the feature as it was last modified before that merge (headci/ab-report-workflow@dc0fbf46c), rebased cleanly onto currentmain.What's included
.github/workflows/ab-report.yml— the per-app vanilla-vs-Pilot A/B report workflow.scripts/ab_report.py— externalized command loading (load_commands) so each app supplies its own command pairs.submissions/io.pilot.smolvm/ab-commands.json— sample command set.docs/CI-AB-REPORT.md— docs.dc0fbf46cfollow-up).Notes on the rebase onto main
scripts/ab_report.pydiverged onmain(via docs: clarify required vs optional + publish-path parity #37). Resolved as a 3-way merge: keptmain's non-conflicting changes and took ci: per-app vanilla-vs-Pilot A/B report on publish PRs #33'sload_commandsrefactor for the one conflicting hunk (it's the basis of the per-app design and the downstreamsrcplumbing depends on it).2ae1dc304(dropkeyFromURL,http.StatusServiceUnavailable) is intentionally omitted: it only touched the oldinternal/publish/r2.go/artifactProxyAPI, whichmainreplaced withartifact.go(R2Config) in Native cli artifacts: stage binaries from the R2 registry + close the verify false-pass #42/App-store UPDATE mechanism: versioned updates, same-key gate, R2 presign, admin key rotation #46. Those Go changes are superseded and would not compile onmain. This PR carries no Go changes — only the workflow/script/docs/sample.Build +
python3 -m py_compile scripts/ab_report.pypass locally.🤖 Generated with Claude Code