Implement adjustable shield assist cost

[lL1l1]

Issue

Cybran shields have terrible buildtime stats due to previously high shield assist costs, as written in #2703.
UEF T3 shields cost significantly more than other shields to assist. This is just a problem with RegenAssistMult never being touched, alongside the faction having poor shield stats.

Current assist cost stats:

Shield	AssistRegen/BP	AssistRegen/Mass
Sera T3	2.80	6.06
UEF T3	2.18	4.40
Aeon T3	2.50	5.69
Cybran ED5	2.33	5.19
Cybran T3 ED4	2.17	4.13

Buildtime stats:

Shield	BT
Sera T3	5841
UEF T3	4988
Aeon T3	4097
Cybran ED5	7100
Cybran T3 ED4	3515

Description of the proposed changes

Implements new shield blueprint fields that make it easier/possible to balance shield assist.

• RegenPerBuildRate: Makes it easier to think about RegenAssistMult by disconnecting the desired balance effect from the regen rate of the shield.
• AssistCostEnergyPerBuildRate and AssistCostMassPerBuildRate: Makes it possible to balance shield assist cost by disconnecting the assist cost from the shield unit's cost stats.

I am interested in thoughts on how user-friendly (UI information display and game intuition), how developer-friendly, and how performant the implementation is.

Testing done on the proposed changes

I read the decompilation that mentions the string "RegenRate", which is around where the RegenAssistMult is used (.text:005F602E).

Re-apply the commit with the Aeon T2 shield.
Spawn in the test units:

   CreateUnitAtMouse('ual0301_rambo', 0,    5.07,    8.48, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,   -0.93,   -4.00, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,   -2.07,   -4.26, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,   -1.21,   -2.60, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,   -1.93,   -6.00, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,   -3.21,   -6.60, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,   -2.93,   -8.00, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,   -4.27,   -8.86, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,    3.93,    7.74, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,    4.07,    6.00, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,    2.79,    5.40, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,    1.93,    3.74, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,    3.07,    4.00, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,    2.07,    2.00, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,    1.07,   -0.00, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,    0.79,    1.40, -0.00000)
   CreateUnitAtMouse('ual0301_rambo', 0,    0.07,   -2.00, -0.00000)
   CreateUnitAtMouse('uab4202', 0,   -4.93,    2.48,  0.00000)
   CreateUnitAtMouse('ual0105', 0,   -9.74,    2.05,  0.35222)
   CreateUnitAtMouse('ual0301_rambo', 0,   -0.71,    1.58, -0.22020)
   CreateUnitAtMouse('xab1401', 0,    7.07,   -2.52,  0.00000)

Damage the shield by groundfiring with the SACU and assist the shield with the T1 engineer. It should repair 50 HP per tick and consume 50m/s and 50e/s, which matches up with the test stats.
Let the shield regen to full, then reclaim the shield with an SACU to damage the unit's HP and see that the engineer consumes an additional 1.89 mass/s and 22.7 energy/s.

Damage the shield, assist, and then reclaim the unit. The engi should update its consumption correctly after the unit is damaged (50% of both the assist and repair costs). This is also an opportunity to see the engine behavior of the repair effects being split evenly when both the shield and unit are damaged: the 50 HP per tick regen will go down to 25 HP per tick after the unit is damaged.

Reclaim the unit, assist, and then damage the shield. The engi should again update the consumption correctly.

Checklist

• Changes are annotated, including comments where useful
• Changes are documented in a changelog snippet according to the guidelines.
• Request 2-3 reviewers from the list of reviewers and their areas of knowledge.

[Garanas]

I like the idea of it. I'm not a huge fan of the amount of code that requires it to implement it. In particular, the complexity that is introduced to UpdateConsumptionValues which was already a difficult function to reason about before these additions.

[lL1l1]

I found an engine bug that you can repair shield HP of disabled shields if you trigger the repair by repairing the unit HP, but I didn't remove the cost logic (in OnShieldDisabled, and focusShield:IsUp checks) hoping the bug gets fixed/so that consumption logically makes sense visually (UI isn't displaying updated shield HP when this repair occurs).

[lL1l1]

I like the idea of it. I'm not a huge fan of the amount of code that requires it to implement it. In particular, the complexity that is introduced to UpdateConsumptionValues which was already a difficult function to reason about before these additions.

I refactored it to try to make it easier to reason about the function by putting the condition branches in order of increasing complexity.

[lL1l1]

I assume the usage format of the blueprint values is ok, so I marked the PR as ready for review.