Library errors out when auth token has whitespace / newlines

[nijikokun]

What's happening:

Library rejects authentication tokens with additional newlines or whitespace before / after the token (e.g. <authtoken>\n):

ERRO[0000] failed to reconnect session                   
err="authentication failed: The authtoken you specified does not look like a proper ngrok tunnel authtoken.\n
Your authtoken: <authtoken>\n\n
Instructions to install your authtoken are on your ngrok dashboard:\nhttps://dashboard.ngrok.com/get-started/your-authtoken\r\n\r\nERR_NGROK_105\r\n" id=974c591a9879 obj=csess

Expectation:

• User sets authentication token with whitespace / newlines (e.g. <authtoken>\n)
• Library trims authentication token for whitespace / newlines (becomes <authtoken>)
• Library works as expected

[euank]

This feels like it should be the caller's responsibility, not the library's.

If we had something like WithAuthTokenFromFile("/some/path"), chomping newlines off that would make sense, but the methods we provide are WithAuthtoken(string) and WithAuthtokenFromEnv().

For WithAuthtokenFromEnv(), trimming makes some sense to me. For WithAuthtoken(string), it really feels like it should be the caller's responsibility to do WithAuthtoken(strings.TrimSpace(authToken)) if they've gotten it from a source where it might have whitespace (such as from reading a file).

Are you expecting to just trim the environment variable one (which makes sense to me!), or are you suggesting trimming for both methods?