Improvements to RJ water consistency

[zike1017]

What feature is your improvement idea related to? Please describe.
While sitting on the surface of water with +jump, the player bobs giving you a varying vertical speed of about 0-100 u/s. When you do jumps similar to this: https://youtu.be/RNmCL2JnaTo?t=66 where you're shooting yourself out of a pool of water you get quite a bit of a difference in vertical speed/height.
Just slowing down the game and floating in a pool of water and shooting straight down you can see the variance in vertical speed. This is seen here: https://streamable.com/h7cyd9. Here's the frames of said vid when the rocket explodes: https://imgur.com/a/cU9ngJF. Bottom speedometer is vertical speed and just from 3 quick tests you can see that noticeable speed variance.
If you just sit in a pool of water shooting straight down and then marking your approximate height against a wall you get a spread like this: https://i.imgur.com/AofYJBr.png.
This test here isn't perfect but there's still a pretty big gap in height which matters quite a bit when you're going for records. It functionally just ends up being RNG whether or not you get the desired height (99% of the time you want the max).

Describe the solution you'd like
Ideally there is little to no RNG involved in the amount of height from jumping out of water; when a player jumps out of water while bobbing at the surface, they will get the same height every time. A reasonable base line for improvement is to mimic the amount of bobbing from using +moveup at the surface. The comparison between the two can be seen here: https://streamable.com/fiogql.
+moveup level of consistency would still have some variance in height, but this can hopefully be improved, or will be minor enough to not matter. Additionally, the maximum amount of height that can be gained from +moveup is less than +jump. This may or may not end up being problematic.

Describe alternatives you've considered, if any
Another method Gocnak considered earlier was the following:
"I wonder if we check IN_JUMP and not apply gravity would that feel better
And only if the water level is at feet"
This may or may not be better than a +moveup styled solution.
Another idea is to just outright remove any bobbing on entering water, but this would likely feel/look jarring to the player.
In a similar vein, snap the player's hitbox to a consistent height so there is no bobbing, but have the player's view bob up and down like it would normally.
Regardless this will require a large amount of testing and feedback.

Additional context
Comparison of heights from sitting at the water surface, holding crouch, and shooting straight down. Left uses +moveup to stay afloat, while right uses +jump.
https://i.imgur.com/HUZRTz6.png
It is also worth mentioning getting consistent boosts out of water is desirable in conc and sj too. However, these types of jumps are significantly less common in these game modes compared to rj.

[Mac-tf2]

There's a few issues with mom_mv_tf_flat_sharking 1 currently.

• Water friction isn't applied immediately upon touching water, it takes until the next bob cycle for speed to decrease. Speed should decrease instantly when hitting the water.
• You can hit the bottom of shallow water (~64u) if you fall into it with high vertical speed, causing inconsistent deceleration caused by being grounded for a few frames under the water.
• The command doesn't do anything on jump_sauciix_v4 for some reason.