[fogydes]

What and why

Handles incoming browser sign-in connections concurrently in \crates/net/src/browser.rs.

• Previously, \serve()\ sequentially awaited \handle()\ on accepted streams inside the accept loop. A preconnecting, slow, or stalled local connection could hold the loop for up to 5 seconds, delaying the actual redirect callback.
• Connections are now spawned concurrently via \ okio::select!\ and an \mpsc\ completion channel, allowing stalled or extraneous requests to time out independently while the callback is delivered immediately.
• Adds test \�_stalled_connection_does_not_block_the_callback\ verifying that a stalled connection on the loopback port does not delay the sign-in redirect.

Checklist

• \cargo fmt --all -- --check, \cargo clippy --workspace --all-targets -- -D warnings\ and \cargo test --workspace\ pass.
• New behaviour has a hermetic test (own servers on loopback, no internet, no game).
• A change to a specification or a decision comes with an ADR in \docs/adr/.
• A change to the bridge protocol (\spec/bridge-protocol.md, \crates/bridge-protocol) states its wire effect: the version, and whether the AnyPS5 fork has to change with it.
• It works for every game. No title ID in code; what is known per title is in specifications and tests.
• It stays on the right side of ADR 0021, and no game data is attached: no dumps, keys, firmware, saves, tokens, or logs with game memory.

---

^{Need help on this PR? Tag @codesmith-bot with what you need.}

[AlxcBTW]

Thanks fago. Moving each connection to its own task fixes the stalled-connection block, and the test for it is good. Two things before merge:

1. There is no bound on concurrent connections. Each task lives at most 5 seconds, but a local program can still open thousands in that time. Please add a small cap (a tokio::sync::Semaphore with about 16 permits, try_acquire_owned before spawning; over the cap, just drop the stream) and a test that the callback still works while the cap is full of stalled connections.
2. When the sign-in finishes or times out, serve returns but spawned handlers keep running up to 5 s with the stream. Please hold the tasks in a JoinSet (or abort them on return) so they end with the sign-in.

The accept-error backoff and the loopback check are kept, good. Minor: put use std::sync::Arc; with the other std import at the top.

[fogydes]

Updated per review in commit 72875a3:

1. Bounded concurrency (\MAX_CONNECTIONS = 16): Added \ okio::sync::Semaphore\ with 16 permits; new connections over the cap are immediately dropped (\ ry_acquire_owned).
2. **Lifecycle management via \AbortOnDrop(JoinSet)**: Connection tasks are held in a \JoinSet\ wrapped with \AbortOnDrop\ and joined incrementally in the select loop, and aborted upon returning the callback so handlers do not outlive sign-in or timeout.
3. Tests:
   • Updated \�_stalled_connection_does_not_block_the_callback\ to hold \MAX_CONNECTIONS - 1\ stalled streams and verify the callback connects and finishes cleanly.
   • Added \connections_over_the_cap_are_dropped\ to assert incoming connections beyond 16 are dropped immediately and that releasing permits re-enables accepts.
4. Cleaned up imports: Grouped \use std::sync::Arc;\ with std imports at the top.

[AlxcBTW]

Thanks fago, the cap, the abort on return and the test with the cap nearly full all look right. Merging once CI is green.