Run 4: memory discipline verified (1833 KB flat, no OOM) but 0 hits
with the small maps demonstrably retained — window-repetition theory
refuted, sub-second bar failed (splice 2.83 s cold / ~2 s steady).
Decision: wire the splice in anyway (~9-10 s cold real-repo :sync vs
611 s/OOM for every alternative). Run 5: cache removal confirmed
free; the ~1.85 MB "resident" was mwindow's live window set, which
makes the mwindow opts in shipping git_sync.rs load-bearing.
The design record for Typoena — the decisions, specs, and bench write-ups
behind the writing appliance. Start with the ADRs for the
load-bearing choices, or the v0.1 specs for what the
first release actually does.
9309f3f239