Refactor trig functions into LUT

[jonvmey]

Replaced the use of the standard library trig functions with a LUT-based
implementation instead. The standard library implementations produce
more accurate results but the usage here doesn't need that. This ends up
saving nearly 7kB of binary size.

[jonvmey]

Final tally for binary size change with this diff is exactly 7000 bytes less. I still need to load this on my watch tomorrow to verify that everything's good but I think it's all correct.

[Riksu9000]

There already seems to exist this in LVGL. Can you use this instead?
https://github.com/joaquimorg/lvgl/blob/23430cf20e32294549fff9b2879a9466dacc19bb/src/lv_misc/lv_math.c

[Riksu9000]

There seem to be some issues with this.

VID_20210710_120654.mp4

[jonvmey]

Yeah, that definitely looks wrong, I'll take another look to see what I messed up.

The implementation in LVGL doesn't appear to be part of its exposed API so I wasn't sure if we wanted to use that directly.

[Avamander]

The implementation in LVGL doesn't appear to be part of its exposed API so I wasn't sure if we wanted to use that directly.

If they remove it I guess we can think of a solution. For now I suspect it'd be fine. Quite a bit less to maintain.

[jonvmey]

Updated the changes to use LVGL's _lv_trigo_sin function instead. The issue @Riksu9000 mentioned should also be fixed now.

[Riksu9000]

Seems to work fine now.

text goes down 7160 and data goes up 4 compared to develop.

[JF002]

Looks good to me, thanks for this PR, this nice review and these 7160 free bytes in memory ;)