Repository navigation
ci: compare release candidates with the previous release - #879
Merged
Merged
Conversation
Release Please now dispatches a same-runner comparison of the candidate wheel with the previous published release and posts the required Verify status only after it passes. Wall time, CPU time and peak RSS medians of compile, plan, lineage, dag, scope and build on the generated benchmark fail above +25% and the absolute floors.
A release tag only counts as a baseline while its version is not yet on PyPI, so a yanked release is never compared against.
Release Please passes a per-attempt request id to each trusted check and waits only on the run carrying it, never with a baseline override. The checks run in their own job so re-running it re-dispatches all three and repaints Verify on the current head, and the delivery token is minted just before enabling auto-merge.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
sqb planregressed twice (0.118.0, 0.121.0) without any release failing: fixed ceilings miss gradual creep and doublings that stay under a loose limit. Release PRs now compare the candidate with the previous published release on the same runner.Changes
scripts/release_performance/: installs candidate and baseline in separate venvs, generates the existing benchmark projects once, and times compile (cold/warm),plan --json, lineage (hub downstream, rollup upstream, one column trace),dag --json,scope --jsonand a 1,000-modelbuild, interleaved 3 runs each; fails when a median is >25% worse and above a 0.5 s / 32 MiB floor. Markdown table in the job summary, JSON artifact.release-please.yml: a rerunnablerelease-checksjob dispatches metadata, version and performance checks with a per-attempt request id, matches exactly those runs, and postsVerifysuccess only when all pass (failure otherwise); a fresh token enables auto-merge.Verification
make test,make check-ci; one review plus follow-up.