Funnel run history + bottleneck — the idiot guide
What is it? The funnel pane draws a graph (the "plan-DAG") that grows as you funnel ideas. Two new helpers:
-
Run history — a little dropdown of past versions of the graph. Pick one and the pane shows you what the graph looked like back then. Pick
◷ latest (live)to snap back to now. Like the "undo history" thumbnails in a photo editor, but for the funnel graph. -
Bottleneck hint — an orange line that says
clear next · <name>. It points at the one task that's blocking the most other tasks. Do that one next and the most work frees up.
Why? As the graph grows you can (a) scrub back to see how it got here, and (b) always know where the jam is.
How does it know the bottleneck? It doesn't compute it itself — it asks the graph component (facett-plandag), which already knows how to rank nodes by "how much stuff is stuck behind me". We just show its answer. (That's the "don't build the same thing twice" rule.)
Where does the history come from? Every time the pane refreshes and the graph actually changed, it quietly saves a snapshot. Identical refreshes don't pile up. It keeps the last 32. If the server is offline it simply saves nothing — no crash, empty list.
Can I poke it without clicking? Yes — send it little messages:
{"select_run": 2}→ show past run #2{"select_run": null}→ back to live{"clear_runs": true}→ forget the history
Is it tested? Yes: one test drives the dropdown and checks the right old graph + its bottleneck show up (and breaks on purpose if the dropdown handler is disabled), plus a real on-screen render test that the orange hint and the summary actually paint.