Fallback to polling does not work #3998
-
We have a couple of customers who are unable to connect with the server due to the websocket being blocked by their network. Our client transports are configured as:
I would suspect the client to fallback to polling however this does not work in their situation. When the following transports order is used the polling connection IS working for them:
See the image of what the customer is seeing in the browser console when the websocket is specified as first option. Why is the fallback not working in this case? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 4 replies
-
Hi! I think the word "fallback" may be a bit misleading here, as the order of the transports in the
Related: https://github.com/socketio/engine.io-client/issues/575 |
Beta Was this translation helpful? Give feedback.
Hi! I think the word "fallback" may be a bit misleading here, as the order of the transports in the
transports
option matters:["websocket", "polling"]
means: establish a WebSocket connection first (or use HTTP long-polling if WebSocket is not available, for example in older browsers)["polling", "websocket"]
means: establish a HTTP long-polling connection first, and then upgrade to WebSocketRelated: https://github.com/socketio/engine.io-client/issues/575