Say what a pulled lever leaves on the commit the branch was at (#90) - #283
Merged
Conversation
The header of `.github/workflows/branch-health-lever.yml` says how to take the reading and what the lever costs the registry, and it did not say what a run of it leaves behind. A check-run is reported against whatever commit the default branch pointed at when the workflow was dispatched, and one is not replaced by a later run of the same name, so the pull that asks for `failure` marks that commit permanently and the pull that asks for `success` adds a second check-run beside the first rather than clearing it. That was found by reading the merge commit this workflow landed on after both pulls, and the commit carries both conclusions under one name. The failure it prevents is somebody pulling the lever while the default branch is at a commit they care about. The report itself reads each workflow's run listing and takes the newest conclusion, so it saw the recovery and its verdict is unaffected; anything that asks a COMMIT for its check-runs and treats a failure among them as the verdict reads it the other way, which is how a merge button and most dashboards read one. The paragraph names both readings rather than leaving the second to be discovered. Nothing else changes. The workflow's triggers, its input, its default and its step are untouched, and no run was taken for this change. Signed-off-by: Nils Lehnen <30603423+iderex@users.noreply.github.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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The issue this belongs to
Refs #90, which is closed. This is a correction to a comment block that landed
under it, not a change to any condition on it.
What changed
One paragraph in the header of
.github/workflows/branch-health-lever.yml,saying what a run of that workflow leaves on the commit the default branch was
at when it was dispatched.
What failure it prevents
Somebody pulling the lever while the default branch is at a commit they care
about.
A run reports a check-run against whatever commit the branch pointed at, and a
check-run is not replaced by a later one of the same name. So the pull that asks
for
failuremarks that commit permanently, and the pull that asks forsuccessadds a second check-run beside the first rather than clearing it. The header said
how to take the reading and what the lever costs the registry, and said nothing
about this.
This has already happened rather than being a possibility: the merge commit that
landed the workflow carries both conclusions under one name, and it is the commit
the default branch still points at.
Evidence
Run at
a35f441cbff92e13fc414a06d854e273af098aa3, the head being pushed. Theparagraph pastes this command and this output:
The report's own verdict is unaffected, and that is a reading rather than an
assumption. Run
33616999340, taken after the second pull:The document check still resolves every path it reads:
The changed set:
What this does not cover
Nothing is repaired. A check-run cannot be removed, so the commit the branch
points at keeps the failure conclusion it was given. This adds a sentence and
takes nothing back.
No run was taken for this change, and no workflow was dispatched. The numbers
above are re-reads of runs that already existed.
Nothing measures how anything else reads that commit. The sentence about a
merge button and a dashboard treating a failure among a commit's check-runs as
the verdict is a claim about how such tools generally behave, not a reading of
one, and it is written as a claim.
No trigger, input, default or step changed, so the lever behaves exactly as
it did.
No scope comparison stands behind this. #90 declares no
Scope:line atcolumn zero, so the hygiene check's changed-paths comparison is NOT MADE rather
than made and passed.
Who has read it
Nobody other than the author. There is no second reader on this board tonight,
and the evidence above stands in place of one rather than the question being left
open.