Unit tests for character stamina functions

[wapcaplet]

Summary

SUMMARY: Infrastructure "Unit tests for character stamina functions"

Purpose of change

To ensure the behavior of the lower-level character stamina functions are well-understood, documented, and automatically tested.

Describe the solution

Adds test cases exercising behaviors of these functions:

• Character::stamina_move_cost_modifier
• Character::burn_move_stamina
• Character::mod_stamina
• Character::update_stamina

Test cases cover the following:

• stamina movement cost (stamina_move_cost_modifier)
  • running cost is double walking cost for the same stamina level
  • walking cost is double crouching cost for the same stamina level
  • running cost goes from 2.0 to 1.0 as stamina goes to zero
  • walking cost goes from 1.0 to 0.5 as stamina goes to zero
  • crouching cost goes from 0.5 to 0.25 as stamina goes to zero
• modify character stamina (mod_stamina)
  • regain partial stamina
  • regain all stamina
  • regain more than all stamina (ceiling at max stamina)
  • lose partial stamina
  • lose all remaining stamina (not winded)
  • lose more than all remaining stamina (floored at 0; become winded)
• stamina burn rate (burn_move_stamina)
  • carrying nothing: walking burns 1x, running 14x, crouching 1/2x normal stamina rate
  • carrying max weight: walking burns 1x, running 14x, crouching 1/2x normal stamina rate
  • overburdened: walking burns +1x, running +14x, crouching +1/2x per percent overburdened
• moving while overburdened causing pain (burn_move_stamina)
  • with 0 stamina, or with "bad back" trait
• stamina regeneration rate (update_stamina)
  • normal baseline when character is not winded
  • 10% of normal when character is winded
• stamina regen in different movement modes (update_stamina)
  • run/walk/crouch all give the same stamina regen per turn
  • (this is a generic regression test for the bug described by Dissecting corpses when running drains stamina #38877 - see notes in additional context)
• stamina regen with mouth encumbrance (update_stamina)
  • wearing a scarf reduces stamina regen
  • wearing another scarf reduces it further

Depends (only a tiny bit) on #38926
Depends on #38945 for test data

Describe alternatives you've considered

Leaving the spoilers unspoiled

Testing

make && tests/cata_test [stamina]

Additional context

Total of 7 new test cases with 343 assertions, running in 0.002 seconds:

Regarding the regression test for #38877 , to be perfectly honest this entire PR was driven by my desire to test that aspect of stamina regeneration better - from my first commit testing the curious scaling function stamina_move_cost_modifier to the last, which aimed to expose the discrepancy between stamina regeneration rates in different movement modes.

To that end, I tested by reverting the fix made in #38891 , and running the same tests. Quite a few of them failed (as expected), including those relating to movement mode, but a couple others indicate that stamina regen rate overall is decreased now, particularly at lower stamina levels, since the stamina_move_cost_modifier scaling function gives bigger buffs at lower stamina (and the test case checks regen rate starting with stamina only 10% full).

For posterity, here's a condensed version of what those failures looked like before #38891

=====================
Given: character is not winded
Then: they regain stamina at the normal rate per turn
--------
  CHECK( actual_regen_rate( dummy ) == normal_regen_rate )
  37.0f == 20.0f

=====================
Given: character is winded
Then: they regain stamina at only 10%% the normal rate per turn
--------
  CHECK( actual_regen_rate( dummy ) == 0.1 * normal_regen_rate )
  4.0f == 2.0

=====================
Then: run and walk mode give the same stamina regen per turn
--------
  CHECK( run_regen_rate == walk_regen_rate )
  18.0f == 36.0f

=====================
Then: walk and crouch mode give the same stamina regen per turn
--------
  CHECK( walk_regen_rate == crouch_regen_rate )
  36.0f == 73.0f

=====================
Then: crouch and run mode give the same stamina regen per turn
--------
  CHECK( crouch_regen_rate == run_regen_rate )
  73.0f == 18.0f

=====================
Given: character has mouth encumbrance
Then: stamina regen is reduced
--------
  CHECK( actual_regen_rate( dummy ) == normal_regen_rate - 2 )
  33.0f == 18.0f

When: they have even more mouth encumbrance
Then: stamina regen is reduced further
--------
  CHECK( actual_regen_rate( dummy ) == normal_regen_rate - 6 )
  25.0f == 14.0f

This confirms the overall stamina-nerfing that resulted from the PR, especially for crouching. I hope by having these tests outlined now, we will be in a better position to understand how stamina burn and regen works now, and make any necessary rebalances.