Fixed in a356441: test_an_open_reply_to_an_open_hint_is_no_second_hint, red without the author filter.
Checked on git.javagil.de just now: PATCH /issues/comments/1406 with the unchanged body of line comment 1406 answered HTTP 200 with the same id, so Gitea 1.27 edits line comments under this…
Fixed in 3d0fbbc: requests listed only as REQUEST_REVIEW review are withdrawn too.
Answer to the review of 308b44e (ERROR):
- ERROR
POST /pulls/33/reviewsHTTP 500: the finding attests/test_line_comments.py:464-477lies beyond the end of the file (224 lines), and Gitea…
Answer to the review of 08ecc13 (ORANGE):
- ORANGE
werkrevisor/cli.py:532-536(request withdrawn although the merge may not follow): its text is lost again, since 0.20.1 reused my reply 1440,…
Fixed in 308b44e: the two defects found in the reviews of this PR moved from the solution into the problem. Scenario#33.03 keeps its table: both causes lead to the same outcome, which the template…
Answer to the review of dfd5f69 (ORANGE):
- ORANGE
werkrevisor/cli.py:530-533(withdrawal without a scheduled merge): fixed in 08ecc13, thread resolved. - YELLOW risky path `werkrevisor/gitea.…
Not changed: the count stays right, as the finding says, since the open line comment counts. The author still replies to that line comment and resolves it; the result comment only carries the…
Fixed in 08ecc13: the request is withdrawn only after merge scheduled or merge already scheduled; test_a_merge_not_scheduled_keeps_the_review_request covers a missing branch protection (red…
Fixed in 08ecc13: one Given (the open request), one When (the GREEN review schedules the merge), one Then; the So-that states the goal.
Answer to the review of f572e65 (GREEN, merge held for 10 open hints):
- hint
werkrevisor/cli.py:453-454(edit while publishing can fail): not changed, as in the thread atcli.py:715. -…
Not changed: the same point as the thread at werkrevisor/cli.py:715; the result comment must know the outcome of the edit to carry the text after a failed edit, and the next review writes the…
Fixed in f572e65: the reply case is split into Scenario#33.04 (not reused) and Scenario#33.05 (no hold), one test each.
Not changed: werkrevisor never replies in a thread; it posts line comments only as part of its own review. Should it ever reply, its reply would be bot text in an open thread, and reusing or…
Answer to the review of c2043de (YELLOW):
- YELLOW spec Scenario#33.01
given_when_then_misused: reworded in f572e65 with concrete texts: the Given holds the earlier comment, the When the new…
Answer to the review of 704fd0b (YELLOW):
- YELLOW spec Scenario#33.01
given_when_then_misused: the When step held state; fixed in c2043de, the different wording is now a Given. - YELLOW…
Not changed: only a model text that ends with exactly this note line, with a hex SHA, is affected, and the only consequence is one skipped edit of wording; the finding itself stays open and…
Not changed, on purpose. The edit only brings the wording of the open finding up to date and names the head that wrote it; if the result comment fails afterwards, the next run of that head or of a…
Answer to the review of a0bffc1 (YELLOW):
- YELLOW
werkrevisor/cli.py:207-215(CRLF/trailing whitespace defeats the note regex): fixed in 704fd0b, thread resolved. - hint `werkrevisor/cli.py:4…
Fixed in 704fd0b: _UPDATE_NOTE accepts any blanks and CRLF around the note and upper-case hex; the comparison normalizes line endings. Stacking cannot happen anyway: an edit replaces the whole…