If a localhost connection is closed by the peer, all other GOOS seem to allow writes to happen without error, while js/wasm returns write tcp 127.0.0.1:XXX->127.0.0.1:1: Socket is not connected with XXX changing.
I admit it does feel like something that should return an error, and I don't know if there is a precise requirement to do otherwise, but js being the only GOOS to do it is very inconsistent, and not even NaCl on the playground with its fictitious network returns an error.
Unfortunately, the fake implementation of TCP-like byte-sequence full-duplex pipe on JS is broken from the beginning. Also, it doubles the maintenance cost by having each implementation on NaCl and JS. I have a CL that may solve the issues: https://go-review.googlesource.com/c/go/+/120958, but it will probably land in Go 1.13 or above.
Can you please skip the roadblock test cases for now?
Just for congestion control; the plan for Go 1.12 had AIX and Fuchsia ports that need to churn up the runtime, os and net packages. Like other communication protocols, congestion control is reviewee/sender side's responsibility (and flow control is reviewer/receiver side's responsibility.)
Added some assertions to testHandshake, but avoided checking the error
of one of the Close() because the one that would lose the race would
write the closeNotify to a connection closed on the other side which is
broken on js/wasm (#28650). Moved that Close() after the chan sync to
ensure it happens second.
Accepting a ticket with client certificates when NoClientCert is
configured is probably not a problem, and we could hide them to avoid
confusing the application, but the current behavior is to skip the
ticket, and I'd rather keep behavior changes to a minimum.
Reviewed-by: Adam Langley <email@example.com>