test: disable the string-building micro-benchmark - #29
Merged
Conversation
The test asserts that buildString beats joinToString at the 90th percentile of measured wall-clock time. On a shared CI runner that comparison flips, and because unit tests gate publishing, it fails releases that are otherwise sound -- one such release failed today and passed on a rerun of the same commit. It measures a performance claim rather than correctness, so it should not hold a veto over publishing. Disable it and run it locally when the claim needs revisiting.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
JointToStringVsStringBuilderPerformanceTestasserts thatbuildStringbeatsjoinToStringat the 90th percentile of measured wall-clock time, over 10,000repetitions of joining 10,000 items.
On a shared CI runner that comparison flips. Because
testReleaseUnitTestgatespublishing, the flip fails a release that is otherwise sound: a release failed
on this test today, and the very same commit passed on a rerun with no code
change. Upstream has already tried to stabilise it more than once (moving to a
percentile comparison, dropping the seed).
Change
Disable the test.
It measures a performance claim, not correctness, so it should not hold a veto
over publishing. The annotation records why and points at running it locally
when the claim needs revisiting.
Alternatives considered
only makes the flip rarer.
and configuration for a single known case; worth revisiting if more tests
turn flaky.
repository runs them nowhere else.