publish-snapshot.sh:32 points at the wrong suite's scenario 8 #6

New Issue

2026-08-24T17:03:36Z

claude commented

2026-08-24 17:03:36 +00:00

Pre-existing; surfaced during the review of
#4 and correctly left out
of it.

Problem

scripts/publish-snapshot.sh:32 reads, of a torn-clone failure mode:

the failure mode this script's own selftest (scenario 8) reproduces

That is wrong twice over, and each half points somewhere real but different:

"this script's own selftest" is publish-snapshot-selftest.sh. Its
scenario 8 exists — it is the abandoned-marker case — and it is not the
failure mode this comment is describing.
The scenario that reproduces a truncated clone lives in a different
suite, seed-target-dir-selftest.sh, and since #4 it is 8b rather than
plain 8 (#4 split the old scenario 8 into 8a, source replaced under the
clone, and 8b, subtree lost from the walk).

So a reader following the pointer lands on a real scenario with a plausible
number that tests something else — the worst kind of stale cross-reference,
because it does not look broken.

Ask

Repoint it at seed-target-dir-selftest.sh scenario 8b, and say which
suite it is in rather than "this script's own", since it is not.

While there: git grep -n 'scenario [0-9]' across scripts/ to catch any
other cross-suite reference that names a bare number without naming its suite.
#4 fixed publish-snapshot-selftest.sh:44 for exactly this reason; this is the
sibling it could not reach.

Acceptance criteria

publish-snapshot.sh:32 names the correct suite and scenario.
Every remaining scenario <n> reference in scripts/ either names its
suite or is unambiguous because it is in-file.

Pre-existing; surfaced during the review of [#4](https://gitdan.com/daniel/gitdan-actions/pulls/4) and correctly left out of it. ## Problem `scripts/publish-snapshot.sh:32` reads, of a torn-clone failure mode: > the failure mode this script's own selftest (scenario 8) reproduces That is wrong twice over, and each half points somewhere real but different: - **"this script's own selftest"** is `publish-snapshot-selftest.sh`. Its scenario 8 exists — it is the abandoned-marker case — and it is *not* the failure mode this comment is describing. - The scenario that reproduces a truncated clone lives in a **different** suite, `seed-target-dir-selftest.sh`, and since #4 it is **8b** rather than plain 8 (#4 split the old scenario 8 into 8a, source replaced under the clone, and 8b, subtree lost from the walk). So a reader following the pointer lands on a real scenario with a plausible number that tests something else — the worst kind of stale cross-reference, because it does not look broken. ## Ask Repoint it at `seed-target-dir-selftest.sh` scenario **8b**, and say which suite it is in rather than "this script's own", since it is not. While there: `git grep -n 'scenario [0-9]'` across `scripts/` to catch any other cross-suite reference that names a bare number without naming its suite. #4 fixed `publish-snapshot-selftest.sh:44` for exactly this reason; this is the sibling it could not reach. ## Acceptance criteria - [ ] `publish-snapshot.sh:32` names the correct suite and scenario. - [ ] Every remaining `scenario <n>` reference in `scripts/` either names its suite or is unambiguous because it is in-file.

claude added the bug label 2026-08-24 17:03:36 +00:00

claude referenced this issue from a commit

2026-08-24 17:50:43 +00:00

docs(publish): point the torn-clone note at the suite that reproduces it

claude referenced a pull request that will close this issue

2026-08-24 17:51:30 +00:00

docs(cache): write down the producer half of the leftover naming contract #8

claude-reviewer referenced this issue

2026-08-24 18:04:21 +00:00

docs(cache): write down the producer half of the leftover naming contract #8

claude referenced this issue from a commit

2026-08-24 18:06:51 +00:00