Basically the result of the Delombok function, except we use IntelliJ's equals,
hashCode and toString and don't re-organize imports (cause that breaks the
preprocessor) and a bunch of manual cleanup was necessary (and half the classes
weren't even converted).
This fixes both, replay-exclusive keybindings blocking vanilla ones as well as
vanilla keybindings blocking replay-exclusive ones.
The magic happens in Mixin_ContextualKeyBindings.
The fetching of the ticks variable was accidentally moved to before the timer
update while updating to 1.16, this caused the interpolation to be off by one
tick each frame in which the timer advances to the next tick.
For one, the forge handshake takes place *after* login phase (but before the
JoinGame packet), so we must ensure that the SquashFilter does not reorder
those.
Secondly, we were ignoring the initial packet (timestamp = 0) in marker
processor when it should have been passed to the SquashFilter. This was not an
issue before cause the SquashFilter did not rely on it. Now however the
SquashFilter needs to be able to see a login phase packet, so it knows that the
following packets till the JoinGame packet are part of the forge handshake.
In ReplayStudio:
20b5183 Fix forge handshake being reordered to after the JoinGame packet
Also need to update ModMenu because the version we depended on still bundled the
snapshot FAPI and that was only fine because we overwrote it with our version.
MC stores server light data and only applies it during the next frame.
Therefore, if there is no frame between receiving the chunk and block changes in
it (like when jumping in replays), then the light update from the block change
gets lost.
This wasn't much of on issue because no tick events are fired normally while
paused. When seeking we do however manually run a tick for better result, and
this manual tick would cause this async sending code to run and seek wildly into
the future.
