connection to smsc with double tcp sesion?

[Slavix72]

Python 3.7 Naz 0.7.4 socket_timeout = 30.0

Client performs connection to smsc... Are packets 1 and 4 establish 2 tcp-sessions? And bind_transceiver is sent across second session on 30 seconds after starting...

[komuw]

What is happening here is;

• when the naz command line app starts, we shedule about 4 different coroutines in the python eventloop; https://github.com/komuw/naz/blob/cc6ce6472f9622e7c317fb4309cbf66bf60d15e1/cli/cli.py#L107-L112

• Of note, is the connect() and tranceiver_bind() coroutines. Since those coroutines are scheduled to run concurrently, it so happens that the tranceiver_bind() coroutine can run before the connect() coroutine has finished.

• If that happens, tranceiver_bind() will fail. If it fails, then naz needs to schedule another attempt to bind to the SMSC. Since the previous bind had failed, it makes sense to wait for a while before issuing another bind request.

• How long does naz wait before sending another bind request? currently naz waits for a duration equal to naz.Client.socket_timeout. By default socket_timeout is 30seconds. ; https://github.com/komuw/naz/blob/cc6ce6472f9622e7c317fb4309cbf66bf60d15e1/naz/client.py#L1803-L1816

• When issuing the bind request, naz also does a tcp connection because the previous bind may have failed because of tcp failure.

So, it would be good to have tranceiver_bind() coroutine wait for the connect() coroutine to finish first before running. But the implementation to do that would be overkill and I do not think the effort is worth the gain

[komuw]

How long does naz wait before sending another bind request? currently naz waits for a duration equal to naz.Client.socket_timeout. By default socket_timeout is 30seconds.

Do note that this is an implementation detail and is likely to change. ie, users should not rely on this behaviour.

[Slavix72]

1. Why can not to use the first TCP-session (after 30 sec)?
2. Why is impoissible to kill unused TCP-session ?

[Slavix72]

The first (simplest, but not beautiful) step is remove client.connect() and client.tranceiver_bind() tasks = asyncio.gather(

client.connect(),

#client.tranceiver_bind(),
client.dequeue_messages(),
client.receive_data(),
client.enquire_link()

) re_establish_conn_bind() can perform tcp-connection and bind_transciever......

[komuw]

I do not consider this to be a major problem that needs to be solved. It is an annoyance, but an annoyance that I'm happy to live with.

Unless it turns out that there is a serious bug hiding somewhere underneath, I'm going to close this issue.