Repository navigation
test(switch): regen alt-l picker snapshot for the new Remote⇅ column - #3227
Conversation
The picker list gained a Remote⇅ column and dropped Age between when test_switch_picker_alt_l_does_not_hscroll's snapshot was generated and when #3226 merged, so the committed frame no longer matched the rendered list (the test failed on main). Regenerate it — the gutter sigils stay intact after alt-l/alt-h, which is what the test asserts. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
Approval stands — the one red check is an unrelated Windows flake, not this PR.
This PR only edits a snapshot data file and changes no production code, so it can't affect the interactive picker's runtime behavior. The interactive picker tests were only just enabled on Windows in #3221, so this is a first-seen timing flake on a loaded Windows runner — the preview pane hadn't repainted |
…rees Merge: pulls in #3231, #3223, #3229, #3227, #3157. #3223 (unify preview tabs + dim empty diff tabs) added `comments_fetched` / `local_content_slots` to `PickerHandler`; threaded both through `PipelineFactory::spawn`'s handler construction and combined the `items` import with this branch's shortcut types. Regenerated the `empty_all_local_diffs` preview snapshot (#3223's new snapshot captured the pre-shortcut controls line). Fix (review feedback from worktrunk-bot on #3233): `alt-y` on a detached worktree row copied the literal `"(detached)"` — the `branch_name()` fallback the side table stored. `RowShortcutData.branch` is now `Option<String>`, carried from the row's raw `ListItem.branch` (and `Some(head_branch)` for `--prs` rows); the `alt-y` callback `.and_then`s it, so a detached row no-ops just as `alt-o` no-ops on a row with no PR. Covered by `shortcut_table_branch_is_none_for_detached_worktree`. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Follow-up to #3226. The picker list gained a
Remote⇅column and droppedAge(from concurrent picker work) in the window between whentest_switch_picker_alt_l_does_not_hscroll's snapshot was generated and when #3226 merged. Both PRs were green independently, but main ended up with the new columns plus the old-column snapshot, so the test fails on main.This regenerates the snapshot for the current layout. The assertion is unchanged in spirit — the worktree-status gutter sigils (
@/+) stay intact afteralt-l/alt-h, proving the keys still don't horizontally scroll the list; only the surrounding column header/data moved.