Details for different port/address added #3124
Conversation
listenbrainz/config.py.sample
Outdated
| @@ -22,6 +22,9 @@ SQLALCHEMY_TIMESCALE_URI = "postgresql://listenbrainz_ts:listenbrainz_ts@lb_db/l | |||
| TIMESCALE_ADMIN_URI = "postgresql://postgres:postgres@lb_db/postgres" | |||
| TIMESCALE_ADMIN_LB_URI = "postgresql://postgres:postgres@lb_db/listenbrainz_ts" | |||
|
|
|||
| # Server address | |||
| SERVER_NAME = "localhost:8100" | |||
There was a problem hiding this comment.
Can you explain why setting this SERVER_NAME is required?
I only see it use in the testing framework so i am wondering.
As mentioned in my previous comment, this could potentially use the value of SERVER_ROOT_URL?
There was a problem hiding this comment.
In situations where generating absolute external URLs is required, such as during OAuth redirection, Flask relies on the SERVER_NAME configuration. Without setting SERVER_NAME, Flask defaults to localhost, which can lead to incorrect or invalid redirect URLs. To ensure OAuth and similar features point to the correct local or production URL, it is necessary to configure SERVER_NAME with the actual base URL of the application.
While SERVER_ROOT_URL serves as a full external-facing URL, it is not directly usable for Flask’s purposes since it includes the protocol (http:// or https://), which Flask does not interpret in SERVER_NAME. Therefore, the two values, while related, cannot be interchanged directly.
You can refer to the attached Flask Documentation.
There was a problem hiding this comment.
Can you share a case where Flask generated invalid urls with the existing config?
There was a problem hiding this comment.
@amCap1712 I tried it out, set my local web port to 8103 in docker-compose and API_URL in config.py.
MetaBrainz Oauth fails, I thought because the redirect_uri now has port 8103 which is not what my existing oauth app expects (and how we instruct users to create in the devdocs).
I assumed the issue would be resolved if I created a new application on MB with the correct port in the callback URL, but I get the same error.
invalid_client is not the error I expected though...
So I went back to using port 8100 and my existing MB application for oauth... and it does not work !
Something looks like it is wrong with OAuth.
Try logging out in your local LB instance and logging back in, see if it works for you?
…enbrainz-server into config-docs-update
Problem
The documentation lacked information about necessary configuration changes for users working with a port other than
8100or a host other thanlocalhost.Solution
Updated the configuration file to include details for working with different ports or hosts. Added a note in the server documentation for user assistance regarding the same:
8100, update theSERVER_ROOT_URLfield with the correct port number.SERVER_NAMEandSERVER_ROOT_URLfields with the appropriate host details.Action