Add base test for incentive

[SidestreamStrongStrawberry]

implement test for https://github.com/orgs/makerdao/projects/15/views/3?filterQuery=&sortedBy%5Bdirection%5D=desc&sortedBy%5BcolumnId%5D=Status&pane=issue&itemId=69698126

Context:

• Ensure liquidation penalty is always bigger than combined keeper incentives
  https://hackmd.io/HsJAhxaoRn-tWOK82pggQg?comment=a9e760be-2557-4f65-acb2-f17d2c4ae4d7
• Ensure there is enough time for the governance to unwind SBE LP tokens
  https://hackmd.io/HsJAhxaoRn-tWOK82pggQg?comment=09a103a7-2ee5-4e10-940e-83278c42ea72

[SidestreamColdMelon]

Performed sanity checks

Numbers that are fetched for ETH-A from etherscan

• vat.ilks(ilk).dust = 7500.000000000000000000000000000000000000000000000 [RAD]
• dog.chop(ilk) = 1.130000000000000000 [WAD]
• clip.tip = 250.000000000000000000000000000000000000000000000 [RAD]
• clip.chip = 0.001000000000000000 [WAD]

Values calculated in this test (logged during the review)

• tab = 8475.000000000000000000000000000000000000000000000 [RAD]
  • ✅ Sounds reasonable (7500 * 1.13)
• penaltyAmount = 975.000000000000000000000000000000000000000000000 [RAD]
  • ✅ Sounds reasonable (7500 * 0.13)
• incentiveAmount = 258.475000000000000000000000000000000000000000000 [RAD]
  • ✅ Sounds reasonable (250 + 7500 * 1.13 * 0.001)

Overall

• ✅ Both penaltyAmount and incentiveAmount values are in the same RAD denomination
• ✅ Asserting that penaltyAmount is greater or equal to incentiveAmount makes sense to prevent "infinite money" draining attack
• ✅ Skipping the assert when line == 0 makes sense, since no new vaults can be opened, therefore draining is not possible
• ✅ Skipping the assert when clip.stopped() != 0 makes sense, since no new auctions can be started, therefore draining is not possible
• ✅ Asserting that flop delay is longer than GSM delay makes sense to allow potential (theoretical) unwinding of USDS/SKY LP tokens via out-of-schedule spell