Program hangs where token is missing or is invaild without any feedback

[swalx]

When starting an http endpoint, the program hangs (seems indefinitely) if the token is invalid or does not exists.

At first I thought simply checking if the environment variable exists and that it has a value would resolve the issue. However, this was not the case.

I understand that the Connect func blocks until success but shouldn't it also return an error? In this case, the ngrok service should at least return an authentication error.

// Connect begins a new ngrok [Session] by connecting to the ngrok service,
// retrying transient failures if they occur.
//
// Connect blocks until the session is successfully established or fails with
// an error that will not be retried. Customize session connection behavior
// with [ConnectOption] arguments.
func Connect(ctx context.Context, opts ...ConnectOption) (Session, error) {

Whether or not this is expected behaviour (for it to hang), I want to believe that where an attempt to start an http endpoint fails due to a missing or invalid token. The most appropriate one of the following errors should be returned to the user: ERR_NGROK_105, ERR_NGROK_106 or ERR_NGROK_107. This would give the user enough feedback for them to make any necessary changes.

This can be reproduced by following the instructions exactly as seen in the Quickstart example found in the README with or without the example token.

This library is really awesome and useful. I just decided to test it out and realise this problem. If I am missing something, it would be great if I could be pointed in the right direction. Otherwise, I will continue to familiarise myself with the code base to potentially open a PR with a fix (assuming no one would have already done so).

Similar issue but includes the error: #107

[CK-Ward]

Hi @swalx, thank you for bringing this to our attention. Are you still experiencing this issue? We'll have our team take a look.

[jrobsonchase]

It's not quite that there's no feedback, it's just somewhat hidden. Each time a connection is attempted and fails, or the session is disconnected, the DisconnectHandler is called with the error.

	sess, err := ngrok.Connect(ctx,
		ngrok.WithAuthtokenFromEnv(),
		ngrok.WithDisconnectHandler(func(ctx context.Context, sess ngrok.Session, err error) {
			if err != nil {
				fmt.Println("error: ", err)
			}
		}),
	)

Additionally, ngrok.Connect will give up and return an error if the reconnect loop stops. Currently, this only occurs when the session is closed, which you can also do via the disconnect handler:

	sess, err := ngrok.Connect(ctx,
		ngrok.WithAuthtokenFromEnv(),
		ngrok.WithDisconnectHandler(func(ctx context.Context, sess ngrok.Session, err error) {
			if err != nil {
				sess.Close()
			}
		}),
	)
	if err != nil {
		fmt.Println("ngrok.Connect failed!")
		return err
	}

We could definitely improve on the out-of-the-box experience with a default disconnect handler which makes an attempt to classify errors as retryable or not, and kill the session accordingly.

[swalx]

@CK-Ward, thank you for your response and for looking into this. My usage have been somewhat limited as of recent. However, I had identified (2928705) what I believed to have been the problem. Which was a missing early return statement. Upon added said statement, the issue was resolved and the expected error messages were being returned.

A PR was opened with the potential fix for review: #177

I wanted to add a reviewer or assign someone to it but none of those options were available and I completely forgot I could just mention it here.

I did an upgrade to v1.10.0 just now in the same quickstart example and the issue seems to still be present.