fix(files): keep serve connection alive until fetcher has the bytes
cargo-deny / cargo-deny (push) Has been cancelled
windows-build / windows-build (push) Has been cancelled

Chat file fetches failed 100% of the time with "file fetch: read failed:
read error: connection lost" (both images and arbitrary files, both
directions). Root cause: the FileRouter serve handler called send.finish()
and immediately returned Ok(()), which dropped the Connection. In QUIC,
finish() only marks the stream's EOF -- it does not wait for the written
bytes to be delivered and acknowledged -- so the connection's
CONNECTION_CLOSE raced ahead of the still-in-flight stream data and the
fetcher's read_to_end aborted.

Fix: after finishing, wait on connection.closed() (bounded by
FILE_FETCH_TIMEOUT) so the link stays up until the fetcher has read
everything and closed the connection itself, which is the signal the
transfer landed.

Wire-compatible (no protocol change), so version stays 0.3.0; both peers
just need the rebuilt binary since either side can be the file server.

Adds tests/file_transfer_loopback.rs: a real two-endpoint serve->fetch
round-trip over FILES_ALPN with a 2 MiB multi-packet blob (deterministic
A/B: 0/20 pass without the fix, 20/20 with it) plus an unknown-id "gone"
case.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
2026-06-21 00:07:23 -04:00
co-authored by Claude Opus 4.8
parent 2d067a2e41
commit efadc228eb
2 changed files with 150 additions and 0 deletions
+9
View File
@@ -489,6 +489,15 @@ impl iroh::protocol::ProtocolHandler for FileRouter {
// Finish either way: an unknown id closes with an empty body, which
// the fetcher reads as a zero-length result and treats as "gone".
let _ = send.finish();
// CRITICAL: `finish()` only marks the stream's EOF — it does NOT wait
// for the written bytes to be delivered and acknowledged. If we return
// here the `connection` drops, and its CONNECTION_CLOSE can race ahead
// of the still-in-flight stream data, so the fetcher's read aborts with
// "connection lost". Wait for the fetcher to receive everything and
// close the connection itself (it drops `conn` right after read_to_end);
// that close is our signal the transfer landed. Bounded so a fetcher
// that vanishes can't pin this task forever.
let _ = tokio::time::timeout(FILE_FETCH_TIMEOUT, connection.closed()).await;
Ok(())
}
}