TCP always times out

[jmeaster30]

It seems to finish resolving exactly when the connection times out

[jmeaster30]

I was testing this last night a bit and the behavior is strange.

Steps:

1. Start simpledns server with 1 thread, tcp on, udp off, some other listening port if needed, and debug logging on (base off Split the cache table from the record table #29 for the better feature flag thing)
2. Make dig request like dig @0.0.0.0 -p 12345 google.com +tcp

Notice:

1. You get a log saying a request was made
2. It stops at the remote lookup and hangs It was hanging before the remote lookup call
3. Once the dig command times out, the remote lookup request goes through, we get the response properly, and seemingly we return everything properly
4. All of the logs after the hang have the same timestamp showing that they were just not flushed to stdout when they ran This thought was based off 2 which was wrong
5. Failed shutdown error could be because we try to shut down both ends of the tcp connection but the client already disconnected
6. Dig retries immediately but there is a 10 second gap in logs between the previous attempt and the retried attempt
7. Remote lookup is always in udp. Not sure if mixing those is an issue but with the log timestamps showing the remote lookup happening in less than a second I am more inclined to think that isnt the issue.

[jmeaster30]

I have considered switching all networking/async stuff to use tokio but I would rather figure this out before switching anyway

[jmeaster30]

I definitely read the logs wrong LOL but I fixed this with #29

What was wrong:

1. used stream.read_to_end but the TCP stream doesn't have an EOF indicator after each packet so this just blocked the thread
2. not writing the response packet's size to the stream prior to sending the response packet