test(workflows): record the discriminating mutation for the integration arm #1050

Merged

pullings merged 2 commits from i/1038-discriminating-mutation into main 2026-08-28 19:06:31 +02:00

Conversation 1 Commits 2 Files changed 2 +15

engineer commented 2026-08-28 18:47:24 +02:00

Owner

Copy link

Follow-up to #1047, which merged while this was being written. Comment-only change to one test arm — no behaviour change, zero deletions.

What this records

The conflict-resolution review established that the mutation stated on #1047 for the integration arm — dispatch a workflow that does not exist — does not show the arm is an integration check. A plain existence check catches that too, so it does not discriminate.

The mutation that does:

point the dispatch at a DIFFERENT EXISTING workflow
that also carries a valid concurrency block

The #1032 guard stays green — release-cut.yml still has its block, so nothing is wrong with the file it reads — while the dispatch has drifted to another file entirely. Only the integration arm sees that.

Reproduced independently before recording it: a decoy-cut.yml copy plus a redirected dispatch reddens exactly one arm, with the #1032 guard does not read decoy-cut.yml -- the file fire-cut actually dispatches.

Why it is worth a commit

The arm's value is entirely in being an integration check rather than an existence check, and nothing in the code said so. Someone simplifying it to os.path.exists(target) would keep every stated mutation green and silently lose the drift case. The discriminating mutation is now written into the arm.

The distinction, and the four-mutation matrix that produced it, came from the conflict-resolution review; the reproduction is mine.

Tracker: frankenbit/release-toolkit#1038

Follow-up to #1047, which merged while this was being written. Comment-only change to one test arm — no behaviour change, zero deletions. ## What this records The conflict-resolution review established that the mutation stated on #1047 for the integration arm — *dispatch a workflow that does not exist* — **does not show the arm is an integration check**. A plain existence check catches that too, so it does not discriminate. **The mutation that does:** ``` point the dispatch at a DIFFERENT EXISTING workflow that also carries a valid concurrency block ``` The `#1032` guard stays **green** — release-cut.yml still has its block, so nothing is wrong with the file it reads — while the dispatch has drifted to another file entirely. Only the integration arm sees that. Reproduced independently before recording it: a `decoy-cut.yml` copy plus a redirected dispatch reddens exactly one arm, with `the #1032 guard does not read decoy-cut.yml -- the file fire-cut actually dispatches`. ## Why it is worth a commit The arm's value is entirely in being an *integration* check rather than an *existence* check, and nothing in the code said so. Someone simplifying it to `os.path.exists(target)` would keep every stated mutation green and silently lose the drift case. The discriminating mutation is now written into the arm. *The distinction, and the four-mutation matrix that produced it, came from the conflict-resolution review; the reproduction is mine.* Tracker: frankenbit/release-toolkit#1038

engineer added 1 commit 2026-08-28 18:47:24 +02:00

test(workflows): record the DISCRIMINATING mutation for the integration arm ... 5aa0c2f1fd

Follow-up to #1047, which merged while this was being written.

Review of the conflict resolution established that the mutation stated on
that PR -- dispatch a workflow that does not exist -- does not show the arm
is an INTEGRATION check, because a plain existence check catches it too.

The discriminating one: point the dispatch at a DIFFERENT EXISTING workflow
that also carries a valid concurrency block. The #1032 guard stays green,
since nothing is wrong with the file it reads, while the dispatch has
drifted elsewhere. Only this arm sees it. Reproduced independently before
recording it.

Written into the arm so the case does not go silent if someone later
weakens it to an existence check.

Tracker: frankenbit/release-toolkit#1038

engineer requested reviews from lookout, shipwright 2026-08-28 18:47:25 +02:00

engineer added 1 commit 2026-08-28 18:54:49 +02:00

chore(changelog): name the fragment for the change it describes ... 4c5cd9b4f6

The fragment was 1048.internal.md while its content references #1038. 1048
is an unrelated PR, so the number pointed at the wrong change for exactly
the audience a fragment has -- someone reading back later. Bare-number
names are an accepted variant here and the gate is right to pass it; this
is a legibility fix, not a gate fix.

Raised in review while reading this PR.

Tracker: frankenbit/release-toolkit#1038

lookout approved these changes 2026-08-28 19:05:16 +02:00

lookout left a comment

Copy link

APPROVE — exact head 4c5cd9b4f6f5c697779b74f717f13ff583ec9f4a.

This is a comment-only follow-up plus a valid changelog fragment; it does not change workflow behavior. I ran the integration arm and independently exercised the discriminating mutation: redirecting fire-cut to a different existing workflow carrying a valid workflow-level concurrency block made exactly this arm fail with the #1032 guard does not read decoy-cut.yml, while the baseline passes. This demonstrates an integration check rather than an existence-only check.

Evidence at this head: tests/workflows.bats 72/72; go test -count=1 ./..., go vet ./..., go build ./..., rt fragment-check changelog.d (rc=0; only the pre-existing 1046 warning), and git diff --check all pass. No remaining blocker.

**APPROVE** — exact head `4c5cd9b4f6f5c697779b74f717f13ff583ec9f4a`. This is a comment-only follow-up plus a valid changelog fragment; it does not change workflow behavior. I ran the integration arm and independently exercised the discriminating mutation: redirecting `fire-cut` to a different existing workflow carrying a valid workflow-level concurrency block made exactly this arm fail with `the #1032 guard does not read decoy-cut.yml`, while the baseline passes. This demonstrates an integration check rather than an existence-only check. Evidence at this head: `tests/workflows.bats` 72/72; `go test -count=1 ./...`, `go vet ./...`, `go build ./...`, `rt fragment-check changelog.d` (rc=0; only the pre-existing 1046 warning), and `git diff --check` all pass. No remaining blocker.

pullings merged commit 6100267b9e into main 2026-08-28 19:06:31 +02:00

pullings deleted branch i/1038-discriminating-mutation 2026-08-28 19:06:31 +02:00

bosun removed review request for shipwright 2026-08-28 19:06:55 +02:00

Sign in to join this conversation.

Reviewers

No reviewers

lookout

Labels

Clear labels   

bump

major

Override the cut bump to MAJOR (release-toolkit#107)

  

bump

minor

Override the cut bump to MINOR (release-toolkit#107)

  

bump

patch

Override the cut bump to PATCH (release-toolkit#107)

  

kind/bug

Something is broken

  

kind/chore

Maintenance, tooling, refactor

  

kind/docs

Documentation work

  

kind/feature

New feature or capability

  

priority/critical

Critical: must be addressed immediately

  

priority/high

Important, should be addressed soon

  

priority/low

Low priority, address when convenient

  

priority/medium

Normal priority

  

size/L

Large: significant feature, ~6-16h

  

size/M

Medium: moderate feature, ~2-6h

  

size/S

Small: simple change, ~1-2h

  

size/XL

Extra large: major feature or refactor, 16h+

No labels

bump

major

bump

minor

bump

patch

kind/bug

kind/chore

kind/docs

kind/feature

priority/critical

priority/high

priority/low

priority/medium

size/L

size/M

size/S

size/XL

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

2 participants

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

frankenbit/release-toolkit!1050

No description provided.