status: nest a spawned lane under the lane that spawned it

[JulienMartel]

What

bench status's lane table draws a lane spawned from a pane — scruff child reaching into a second repo, or ⌘↵ pressed inside another lane — indented under the lane that made it, marked └.

  workshop     worktree-omit-child-lanes        dirty  ~/.cache/scruff/workshop/omit-child-lanes
  scruff       └ worktree-omit-child-lanes      ·      ~/.cache/scruff/scruff/omit-child-lanes
  haus         └ worktree-omit-child-lanes      ·      ~/.cache/scruff/haus/omit-child-lanes

Why

A workshop lane and the three child repos it opened read as four unrelated lanes in registry order. They are one piece of work.

They are indented, never dropped. The branch is that lane's own, it gets its own PR, and closing the parent's pane does not reap it — this table and scruff are the only places it surfaces once that pane is gone. Two rules keep that honest: a child whose parent bench does not list (someone else's repo) stays at the top level, and a child printed in a different block from its parent (live child, parked parent — exactly what survives a closed pane) is demoted rather than drawn pointing at a row that isn't there.

branch_cell pads by columns for the reason dirty_cell does: └ is three bytes and one column. It computes the full-cell case rather than measuring it, because ${#b} counts bytes under LC_ALL=C and … is three of them — measured, the pad went negative and printf padded the other way.

Verify

nix shell nixpkgs#bats -c bats test/bench.bats     # 190/190
./bench status
LC_ALL=C ./bench status                            # columns hold in either locale

Watch out

• The doc line added to docs/workflows.md says the palette's Lanes picker is the one surface that drops these. That is true only with hausfold/haus#… (branch worktree-omit-child-lanes, same lane) merged — land them together, or drop that parenthetical.
• scruff nests the same way via list: nest spawned lanes, and publish where each lane's chat lives scruff#93. Independent: neither PR needs the other, and nothing here reads scruff's output.
• lane_rows grew a sixth field (parent); lane_table is its only consumer.

🤖 Generated with Claude Code

https://claude.ai/code/session_011T9Lb6mUXWEudUcs2E1B9X