Missing:
* More testing and adjusting.
* Other cases found, but not yet done. E.g. multiple 0-ydistchange.
* Jumping with oddd gain (2x 3.8 instead 0.42 + 0.33333).
* Lost ground and all the cases apply here (use only sf or refactor).
* Do or start to distinguish ascend vs. descend, if gravity is used, if
modifiers are used, similar, based on the model in use (the vertical
check part had never accounted for descending anyway). Shuffle chunks of
code.
* Reflect new options in the configuration, sort in more to vertical and
horizontal sections.
* Alter set-back handling, use ordinary set-back for maxheight,
prioritie other violations over maxheight.
* fly-nofly transition: Clear active horizontal velocity, allow less
horizontal friction.
* Increase horizontal speed to 420 for spectator mode.
* Add more tags with specific cases applying.
* Adjust logging format of hdist/vdist.
This probably isn't a usable state, more changes and testing will
follow.
If false positives with players receiving knockback from pvp hits
increase, setting checks.fight.pvp.knockbackvelocity to true instead of
default will force-activate the workaround.
Debug logging would reveal if this works or not. If the velocity events
fire correctly, the horizontal components will be clearly greater than
zero most of the time with pvp hits. [Can't test this right now.]
Very coarse modeling, players likely are able to abuse this and there
are verly likely more false positives.
Especially elytra will have issues:
* Players can fly very fast.
* Elytra will make players fly even faster than the set limit, resulting
in false positives at some point. Setting the allowed speed that high
will yield the problem of players being able to abuse even worse. Thus
limiting to the speed of spectator mode. Modeling will be changed to
accomodate for gain vs. max. distance and other.
This won't cover all effects of internal changes, some blocks might be
interpreted wrongly still, some shapes may have changed in an
incompatible way (e.g. skulls).
