Skip to content

Conversation

@MichaelSims
Copy link
Owner

@MichaelSims MichaelSims commented Dec 28, 2025

Mitigate "push adds commit IDs" test flakiness

I noticed the last dynamic test in this function titled
"only commits in the middle missing IDs" was failing fairly consistently
during functional test runs. It seems there's a delay in GitHub closing
test PRs created by the previous dynamic test. Since this test reused
the same titles and the titles are used as keys, it was causing a "leak"
from one test into the next. I'm fixing this by ensuring the titles are
unique, but I could also mitigate it by adding a delay after "remote
rollback"

Stack:

⚠️ Part of a stack created by jaspr. Do not merge manually using the UI - doing so may have unexpected results.

In commit 43379f I configured junit to run tests in parallel to speed
up verifications, but this unfortunately breaks assumptions that the
functional tests make and additionally increases the odds that I'll
hit GitHub's API rate limits, so I'm configuring the functional tests
to run serially while leaving the global default to parallel.

commit-id: I22786ee4
This is to help with avoiding hitting API rate limits

commit-id: Ie7089fdb
commit-id: I48126df2
@github-actions
Copy link

github-actions bot commented Dec 28, 2025

Test Results

164 tests   164 ✅  54s ⏱️
  9 suites    0 💤
  9 files      0 ❌

Results for commit 75c01f9.

♻️ This comment has been updated with latest results.

This was failing during the functional test because it assumed there
was a "fake remote" (local git repo acting as the "remote" for the test)
. The functional test was using a real remote, so we need to make sure
branches are deleted properly based on which mode we're in.

commit-id: I740f8375
commit-id: Ib9d4f25e
I noticed the last dynamic test in this function titled
"only commits in the middle missing IDs" was failing fairly consistently
during functional test runs. It seems there's a delay in GitHub closing
test PRs created by the previous dynamic test. Since this test reused
the same titles and the titles are used as keys, it was causing a "leak"
from one test into the next. I'm fixing this by ensuring the titles are
unique, but I could also mitigate it by adding a delay after "remote
rollback"

commit-id: Ia649cdf8
@MichaelSims MichaelSims changed the base branch from jaspr/main/I8be43d76 to main December 28, 2025 01:02
@MichaelSims MichaelSims merged commit 75c01f9 into main Dec 28, 2025
7 checks passed
@MichaelSims MichaelSims deleted the jaspr/main/Ia649cdf8 branch December 28, 2025 01:02
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.

3 participants