steid

docs: fix milestone cross-references after the reorder

Git-over-HTTP became milestone 4, so the streaming and body_limit notes in
current.md and the runbook's transport section pointed at the wrong number.

The SSH channel-bridging notes in progress.md said "re-read before milestone
3", which was true under the old ladder and is now misleading twice over --
SSH is not the transport at all after decision 0001, and would be milestone
8+ if it ever returns.

Runbook now distinguishes the two reasons git is a runtime dependency:
git init --bare from milestone 3, git http-backend from milestone 4.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

JamesPatrickGill authored 24 days agoparent2eb8681Browse filesca76e1b7564bbece14603e1ed7bb629f9139b226

3 files changed+7 −5

plans/current.md+2 −2View file

		@@ -91,7 +91,7 @@ Ordered. Pull from the top.
91	91
92	92	- Topcoat is early (v0.5.0, first released 2026-07-22, breaking changes expected
93	93	by its own authors). Expect churn that isn't feature work.
94		−- Body size limits will reject large pushes at Milestone 5 — `topcoat-router` has a
	94	+- Body size limits will reject large pushes at Milestone 4 — `topcoat-router` has a
95	95	`body_limit` layer that needs raising on the git routes. Recorded here because it
96	96	will surface as a confusing failure rather than a clear one.
97	97	- Topcoat ships Tailwind without Node, which reopens the design system attempt #1
		@@ -116,4 +116,4 @@ Ordered. Pull from the top.
116	116	- `HOST` / `PORT` configure the bind address, so `STEID_LISTEN_ADDR` is gone.
117	117	- `Body` is a boxed `http_body::Body` used for both requests and responses, with
118	118	`into_data_stream()` to read and `Body::new()` to wrap a stream — pack data can
119		− stream both directions without buffering. This is what makes Milestone 5 viable.
	119	+ stream both directions without buffering. This is what makes Milestone 4 viable.

plans/progress.md+2 −1View file

		@@ -139,7 +139,8 @@ application layer.
139	139
140	140	#### SSH channel bridging
141	141
142		−The fiddly part, and worth re-reading before Milestone 3:
	142	+The fiddly part, and worth re-reading if SSH ever returns as a transport
	143	+(milestone 8+ — [0001](decisions/0001-git-over-http-not-ssh.md) chose HTTP):
143	144
144	145	- Store `Channel<Msg>` per `ChannelId` in the handler's map on `channel_open_session`
145	146	- On `exec_request`, take the channel, split it with `into_stream()` +

plans/runbook.md+3 −2View file

		@@ -14,7 +14,8 @@ attempt and has not been re-verified against Topcoat — treat it as a sketc
14	14	- rustc ≥ 1.95 — Topcoat 0.5 requires it, and on an older toolchain `cargo add
15	15	topcoat` silently resolves to an empty `topcoat v0.0.0` placeholder rather than
16	16	failing. Verified on 1.97.1.
17		−- `git` on `PATH` (from Milestone 3 — the protocol server shells out to it)
	17	+- `git` on `PATH` (from Milestone 3 — `git init --bare` creates repos, and from
	18	+ Milestone 4 `git http-backend` serves the protocol)
18	19
19	20	## Dev setup
20	21
		@@ -97,7 +98,7 @@ Sign in at `/login` with the email, not the handle.
97	98
98	99	Bare repos at `{STEID_DATA_DIR}/{org}/{repo}.git`. Created empty — no initial commit.
99	100
100		−## Git transport (Milestone 5)
	101	+## Git transport (Milestone 4)
101	102
102	103	Smart HTTP, delegated to `git http-backend`, authenticated with personal access tokens
103	104	over HTTP Basic — see [0001](decisions/0001-git-over-http-not-ssh.md). No SSH, no host