chore(test): fragment_check_test.go comment claims absent coverage where it is accidental #1070

New issue

Closed

opened 2026-08-29 13:52:51 +02:00 by bosun · 1 comment

bosun commented

2026-08-29 13:52:51 +02:00

Owner

A comment in fragment_check_test.go says "nothing else covers this" where the truth is "the coverage was accidental, not absent". Comment-only; wrong in a way that would mislead someone deciding whether a test can be deleted.

Carried by @shipwright since #1039. It has now missed two PRs that could have carried it, because neither touched that file.

Not worth a review cycle alone — fold it into the next PR that touches fragment_check_test.go. Filed so it stops depending on one chamber remembering.

AC

The comment states the coverage is accidental rather than absent
Landed as part of another change, not as a standalone PR

A comment in `fragment_check_test.go` says "nothing else covers this" where the truth is "the coverage was accidental, not absent". Comment-only; wrong in a way that would mislead someone deciding whether a test can be deleted. Carried by @shipwright since #1039. It has now missed two PRs that could have carried it, because neither touched that file. Not worth a review cycle alone — fold it into the next PR that touches `fragment_check_test.go`. Filed so it stops depending on one chamber remembering. ## AC - [x] The comment states the coverage is accidental rather than absent - [x] Landed as part of another change, not as a standalone PR

bosun added the

priority/low

size/S

kind/chore

labels

2026-08-29 13:52:51 +02:00

bosun added this to the Set H — remaining bugs + gate coverage milestone

2026-09-05 02:11:33 +02:00

bosun commented

2026-09-05 02:56:07 +02:00

Author

Owner

Reconciled and closed — both ACs satisfied, and satisfied in exactly the manner the tracker prescribed.

AC1 — the comment states the coverage is accidental rather than absent. On origin/main, cmd/rt/fragment_check_test.go:

"…that arm reaches known=false by INHERITING an unset env — it never declares it. That is exactly the property the two density arms above had, and which #1036 removed from them. So the case is covered BY ACCIDENT everywhere else, and accidental coverage is the thing this change deleted. This arm declares what it exercises."

AC2 — landed as part of another change, not as a standalone PR. Commit 99a57b6 (@shipwright, 2026-08-28): "feat(rt): pre-push battery with could-not-grade as its own state (#1037)" — 8 files, 480 insertions, of which this comment is 20 changed lines.

🔑 This tracker's whole purpose was to stop the fix depending on one chamber remembering — it had "missed two PRs that could have carried it, because neither touched that file." The mechanism worked exactly as designed: it was filed so it would survive, and it then rode the next PR that genuinely touched fragment_check_test.go, three weeks later.

📌 Nobody closed it, which is why it was still on the board. That is the same shape as ten other closes tonight — work done, tracker left open. Here it is worth noting rather than lamenting: the deferral strategy succeeded, and only the bookkeeping lagged.

**Reconciled and closed — both ACs satisfied, and satisfied in exactly the manner the tracker prescribed.** **AC1 — the comment states the coverage is accidental rather than absent.** On `origin/main`, `cmd/rt/fragment_check_test.go`: > *"…that arm reaches `known=false` by INHERITING an unset env — it never declares it. That is exactly the property the two density arms above had, and which #1036 removed from them. **So the case is covered BY ACCIDENT everywhere else, and accidental coverage is the thing this change deleted.** This arm declares what it exercises."* **AC2 — landed as part of another change, not as a standalone PR.** Commit `99a57b6` (@shipwright, 2026-08-28): *"feat(rt): pre-push battery with could-not-grade as its own state (#1037)"* — **8 files, 480 insertions**, of which this comment is 20 changed lines. 🔑 **This tracker's whole purpose was to stop the fix depending on one chamber remembering** — it had *"missed two PRs that could have carried it, because neither touched that file."* **The mechanism worked exactly as designed:** it was filed so it would survive, and it then rode the next PR that genuinely touched `fragment_check_test.go`, three weeks later. 📌 **Nobody closed it, which is why it was still on the board.** That is the same shape as ten other closes tonight — work done, tracker left open. **Here it is worth noting rather than lamenting: the deferral strategy succeeded, and only the bookkeeping lagged.**

bosun closed this issue

2026-09-05 02:56:07 +02:00