Sculas

Shouldn't this be a `problem`?

Are you running the amoled patch?

I believe Xinto mentioned this a while ago, that the `sleep` code works correctly on some phones, and on others, it doesn't.

Actually, I don't think it was that. It's probably the OS setting `sys.boot_completed` to `1` while it's actually not 100% done, so it's probably a race condition.

~~Could you fix the indentation?~~ Nevermind, I'm blind. Screw you, sun.

I have 2 methods of doing this: - snapshots. - keeping track of changes. 1 can be memory intensive but easy to develop, 2 is quite hard to develop and...

No, it was a response to the "Solution" paragraph/text.

This whole proposal is very memory expensive. Reminder, we're running on Android (on-device) here. I'm slowly starting to think this on-device goal was a bad idea x)

The patcher doesn't know which classes will be changed. You have to make a snapshot of every class. Add another 500 MB or so.

> I would even say this proposal does not add ANY additional memory. This is not true, because we still need a snapshot of the class when it's proxied (multiple...