You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Mon Jan 30 12:17:49 GMT+100 2023 com.vaadin.client.communication.MessageHandler
WARNING: Gave up waiting for message 51 from the server
Mon Jan 30 12:17:49 GMT+100 2023 com.vaadin.client.communication.MessageSender
INFO: Resynchronize from server requested
but no request is send, which can be seen from the network log of the browser.
Decreasing the sleep time makes the issue appear more often.
But increasing it to does not make it vanish completely, so that
I think that the browser is able to handle all responses in time.
We also experience infrequent freezes in our production code, even without refresh. But in that case there is proxy in between that might interrupt push long polling requests with gateway timeout responses, which might be related to the freezes.
I was hoping that the client side code is able to resynchronize when messages get lost.
Until this is fixed we need to revert to polling, which breaks session timeouts and automatic logout unfortunately.
This method is called when the UI is reinitialized, and closing the old connection here already (rather than later, when the new connection is ready to be registered) stops the Push from attempting to send more requests through the old connection and increasing the counter with every call while the new connection is still being initialized. I haven't run this solution through our validation builds yet, but at least my test project seems to behave better with it.
Using the attached example code, refreshing the browser leads to infrequent freezes of the UI. The progress bar is shown and does not vanish.
The browser log shows:
but no request is send, which can be seen from the network log of the browser.
Decreasing the sleep time makes the issue appear more often.
But increasing it to does not make it vanish completely, so that
I think that the browser is able to handle all responses in time.
We also experience infrequent freezes in our production code, even without refresh. But in that case there is proxy in between that might interrupt push long polling requests with gateway timeout responses, which might be related to the freezes.
I was hoping that the client side code is able to resynchronize when messages get lost.
Until this is fixed we need to revert to polling, which breaks session timeouts and automatic logout unfortunately.
Probably related: #7719
I can also upload the full source code for reference if you like.
The text was updated successfully, but these errors were encountered: