01ef5e8a

docs: the transports are dumb; it is the frames that changed

a73x 2026-09-04 18:04

Commit message

docs: the transports are dumb; it is the frames that changed

The README said "the wire protocol has survived all three without changing",
which the cells branch made untrue: the replay frames were renumbered and
their payloads replaced. The sentence's point was about the transports, so it
says that instead — none of the three has ever needed a frame of its own, and
the renumbering went past all of them unnoticed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CsWfuJFQbTfGtKZLS5qw4q

README.md

Old		New
@@ -844,8 +844,9 @@ a network tear, or a laptop sleep — tells the daemon what sequence number it
844	already holds, and the daemon answers with a delta instead of a repaint	844	already holds, and the daemon answers with a delta instead of a repaint
845	whenever it can. Transport is deliberately dumb: the same frames ride a unix	845	whenever it can. Transport is deliberately dumb: the same frames ride a unix
846	socket, an ssh pipe (`mux d proxy` is a byte pump with zero protocol	846	socket, an ssh pipe (`mux d proxy` is a byte pump with zero protocol
847	knowledge), or a QUIC stream, and the wire protocol has survived all three	847	knowledge), or a QUIC stream, and none of the three has ever needed a frame of
848	without changing.	848	its own. The frames themselves do change: the move to cells renumbered the
		849	replay frames and replaced their payloads, and no transport noticed.
849		850
850	Every decision and measurement: `docs/decisions.md`. What's next: `git-collab	851	Every decision and measurement: `docs/decisions.md`. What's next: `git-collab
851	issue list` (`docs/roadmap.md` is the older ranked queue).	852	issue list` (`docs/roadmap.md` is the older ranked queue).