The untested hatch takes "verified in a browser" as a reason, and a whole client as untestable by design #441
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
What happens. Run 183's apply left 49 scenarios, every lobby, table and end-screen scenario, in
untested.jsonwith reasons of the form "Verified in a real browser: …", under a design decision that "the React client has no automated unit or integration test suite by design". Run 182 put 11 such scenarios invisual.json; #431 added a reason check there. The untested hatch has none, so the same reasons pass through it (braid#439, finding 7). The apply brief's untested rule says a scenario may be left untested with a reason; it does not say what a reason is.Fix. The untested rule says what the visual one now says: having looked at it yourself is not a reason, a scenario with a testable outcome keeps its test, and a design decision does not excuse a whole client (a component test or a scripted check through the state hook is a test).
untested_scenariosapplies the same reason check asvisual_lists, and a rejected entry stays uncovered and is named in "Where the scenarios stand". The design brief says the same about a decision not to test a layer.Done when. A unit test on run 183's
untested.jsonrejects its 18 entries and keeps one reasoned by what a test would restate; the brief tests hold the sentences; the next Scratch run's apply tags its lobby and table scenarios.From run 183 (braid#439). Follows #431.
Fixed in PR 448.
What changed.
untested_rejected.Evidence from the laptop. A unit test on run 183's
untested.json: all 18 reasons rejected, and four reasons about a test kept (another test fails when it breaks, a test would only restate the code, a named test file, a named e2e spec). Run 183's 14 visual reasons all stand, as braid#431 intends. A second test fails the check on ten of the entries and on run 182's "Browser-verified in task 5.6".Not done: no real model has met the check. The next Scratch run shows whether the apply tags its lobby and table scenarios.