v0.1.52

OmniTerm v0.1.52

Cursor jump: stronger coalescing

0.1.51 helped (you had to type faster to trigger it) but the drain only merged bytes that were already buffered - if the line reprint arrived a millisecond after the carriage-return, it still split into two renders. Now the reader waits a few ms for more output after each read (idle-gap, capped at 20ms) so the whole redraw merges into a single write and renders atomically. Latency stays imperceptible.

Upgrade

python -m pip install --upgrade --no-cache-dir omniterm