Leif Åstrand

From what I can see in the code, the server-side sync id value is incremented once every time `ServerRpcHandler::handleRpc` is run for a given UI instance. Since this is consistently...

Right, I was thinking if nginx when I wrote "load balancer". You wouldn't need to log the request contents to verify my hypothesis about request multiplication between nginx and the...

Would `GridRowDragger` that was introduced in 8.2 also help in this case? (ping @tsuoanttila)

I suspect `DomEventFilterIT.debounce` is no longer a valid test if we change the definition of debouncing to always trigger immediately in case some other event occurs. It might be necessary...

We might also want to make sure `ON_CHANGE` also works in a sensible way when shortcut keys are triggered. My assumption is that it would just automatically do that based...

I realized that there might be an alternative approach to this issue that would also help make `ON_CHANGE` work as expected (with some adjustments to how `ON_CHANGE` is implemented internally)....

I was actually not proposing the mode that only fields with explicit listeners would trigger round trips, but that's indeed a potential incremental step that we could take if we...

I see two ways of solving this. One would be to include something like your `SpringComposite` in the Spring integration. The drawbacks here are that you'd need to do something...

I think we can take that idea one step further. What if we wouldn't ignore the module from the returned promise, but instead do something sensible with the default export...

We will also have the same situation if you create a component purely in `action`. There will be routes for which no name is available unless we also provide a...