Worktrunk is a CLI for Git worktree management, designed for parallel AI agent workflows
test(picker): restore auto-refresh test with cursor-nav selection (#3277)
Maximilian Roos0862c8e7c60b
diff --git a/tests/integration_tests/switch_picker.rs b/tests/integration_tests/switch_picker.rsindex c65fc4a99..2a9b9fe9a 100644--- a/tests/integration_tests/switch_picker.rs+++ b/tests/integration_tests/switch_picker.rs@@ -594,10 +594,17 @@ fn wait_for_cursor_on_row(rx: &mpsc::Receiver<Vec<u8>>, parser: &mut vt100::Pars/// rely on this navigate by cursor and never type, so the query stays empty —/// only the selected row both starts with `>` and carries a worktree `name`,/// which uniquely picks it out.+///+/// The match is scoped to the list pane (cols `0..LIST_WIDTH`). The preview pane+/// shares each physical row to the right of the border, so `name` is sought only+/// in the row's own list text — otherwise a token that also renders in the+/// preview (e.g. a PR title carrying a branch word) could satisfy the check from+/// the wrong row.fn cursor_points_at(screen: &str, name: &str) -> bool {- screen- .lines()- .any(|line| line.starts_with('>') && line.contains(name))+ screen.lines().any(|line| {+ let list: String = line.chars().take(LIST_WIDTH as usize).collect();+ list.starts_with('>') && list.contains(name)+ })}/// Drive the PTY reader until the screen satisfies `ready` and then settles, or@@ -1569,14 +1576,15 @@ fn test_switch_picker_prs_rows_survive_alt_x_removal(mut repo: TestRepo) {/// would strand until the next keystroke — the gap the picker's test harness/// used to paper over by re-issuing the tab key.///-/// Ignored: flakes under CI contention. Filtering to a single visible row-/// (`#`, then `!main`) doesn't control which row is *selected* — when the-/// filter applies before the async `--prs` row has streamed in, the result-/// set empties, the cursor is lost, and skim doesn't reselect the lone row-/// once it arrives, so the `"Not checked out"` gate times out. Tracked in-/// https://github.com/max-sixty/worktrunk/issues/3269.+/// The PR row is selected by driving the cursor with a re-issued Down, not by an+/// `!main` filter: a filter applied before the async `--prs` row streams in+/// empties the result set, and skim doesn't reselect the lone row when it arrives+/// — so the gate timed out under contention (#3269). Down is idempotent and+/// clamps on the bottom row, where the streamed PR row sits (below the worktree+/// row), so the cursor-arrow wait re-issues it until the `>` pointer holds on+/// `flaky`, outlasting both the stream-in and the cursor reset that the+/// item-list refresh triggers.#[rstest]-#[ignore = "flaky under CI contention; restore tracked in #3269"]fn test_switch_picker_preview_auto_refreshes_when_compute_lands(mut repo: TestRepo) {repo.remove_fixture_worktrees();repo.run_git(&[@@ -1614,22 +1622,13 @@ fn test_switch_picker_preview_auto_refreshes_when_compute_lands(mut repo: TestRerepo.root_path(),&env_vars,&[- // Isolate the PR row with skim's inverse-match operator `!main`,- // which excludes the worktree row (uniquely carrying its branch name- // `main` in its matcher text) and leaves only the `--prs` row. This- // puts the cursor on the PR row deterministically. Two reasons a- // simpler approach won't do: (1) a `--prs` row's order against the- // worktree row isn't fixed — it streams in on its own thread and the- // worktree row's `PathName` tiebreak reads the random temp path, so- // either order can win and a single Down lands on either row; (2) the- // mock returns this PR for *every* `gh pr list` (including the- // worktree row's CI-status lookup), folding the PR's number/title/- // author into the worktree row's matcher text too — so no PR-content- // query (`#`, the title, the author) can single out the PR row. The- // worktree branch name is the one token only the worktree row has.- // The PR row is then the sole, selected row; its preview is the "not- // checked out locally" pane.- ("!main", Some("Not checked out")),+ // Drive the cursor onto the PR row. Its narrow (preview-shown) list+ // line shows the short head branch `flaky`; the worktree row shows+ // `main`. A re-issued Down lands the `>` pointer on `flaky` (the+ // bottom row) and re-drives it there if a streamed-row refresh resets+ // the cursor — see the docstring for why this replaces the `!main`+ // filter.+ ("\x1b[B", Some("flaky")),// alt-7: comments tab. The fetch is still in flight, so the pane shows// "Loading comments…"; the comment appears with NO further input once// the delayed fetch lands and the picker repaints on its own.