The archive fix that landed on main at 13:09 (“Looking up an archived card no longer builds the whole archive each time”) already removes the stall this card is about. No further code change is needed.
| Startup sharing path 665 archived cards, 199 shared | Time | |
|---|---|---|
| Before the fix (8d4ed733) | 4,089 ms | |
| Now (5d49f738) | 15 ms |
On this machine: Obeya restarted at 13:10:08 with the fix. ~/.obeya/stalls.log has had no new entry since; its last entry, at 13:09:58, came from the old code. Before, the stall came about 12 s after each start.
What is left: after any change to a card, the next lookup rebuilds the archive once (about 5 ms in the test). That only adds up in loops that write on every pass, and those are rare. Left as it is.