[fornwall]

• Tested on all platforms changed
• Added an entry to CHANGELOG.md if knowledge of this change could be valuable to users
• Updated documentation to reflect any user-facing changes, including notes of platform-specific behavior
• Created or updated an example program if it would help users understand this functionality
• Updated feature matrix, if new features were added or implemented

In #2261, the web backend was changed to use the pagehide event instead of unload, to allow the bfcache to be used.

See https://web.dev/bfcache/ for more information about the bfcache (back/forward cache), and https://developer.chrome.com/docs/devtools/application/back-forward-cache/ for how to test it in chrome. Summary from the first resource:

bfcache is an in-memory cache that stores a complete snapshot of a page (including the JavaScript heap) as the user is navigating away

The problem with winit and the bfcache is that the event loop is destroyed in pagehide - which happens before the snapshot is put in the bfcache. So when a user navigates back to the page running a winit program it's no longer responding.

This can be seen on e.g. the web example program (cargo run-wasm --example web) - after restoring the page from the bfcache the event log is no longer updated.

This PR changes pagehide handling to only destroy the event loop if persisted is false. If persisted is true, we now emit a Suspended event.

It also adds a corresponding listener for pageshow with persisted being true, to emit a Resumed event in that case.

[daxpedda]

I think what we should do here is look at PageTransitionEvent.persisted.

If it's true we should send Suspended and also hook up the pageshow event with a Resumed event.
If it's false we can resume to destroy the event loop.

Thank you for looking into this!

[fornwall]

I think what we should do here is look at PageTransitionEvent.persisted.

If it's true we should send Suspended and also hook up the pageshow event with a Resumed event. If it's false we can resume to destroy the event loop.

Thanks @daxpedda for the helpful feedback! I updated the PR according to this, could you have a look?

Specifically, is it correct to just send_event(Event::Suspended) and send_event(Event::Resumed), or is something else required for proper lifecycle management?

[daxpedda]

Great stuff, thank you!
Couple of nits.

Lifecycle management seems fine. Though I'm trying to figure out if it's necessary for us to actively ignore a request_redraw() after a Suspend event.

[fornwall]

@daxpedda Thanks again for the quick review!

I applied your suggested code change, and answered your comments about pageshow events happening initially with persisted == false. Let me know what you think!