Certora Audit Fixes

[jalextowle]

This PR addresses the remaining Certora issues reported before June 12th, 2023. Specifically this PR addresses the following questions:

Week 1 Question E

Question calculateSharesInGivenBondsOut and calculateBondsInGivenSharesOut in YieldSpaceMath.sol : for non-zero amount to receive, the functions can return a zero amount to pay. Can this be problematic?

Mitigation The mitigation of question E from Week 1 was to write a test that tests trades at various orders of magnitude to verify that the YieldSpaceMath function in question (calculateSharesInGivenBondsOut) doesn't return zero when reasonable values are provided. (The only pool in which this would be conceivable is a USDC pool, and even then, the input amount would have to be on the order of 1 wei which has a tiny value relative to gas costs).

Week 4 Question A.1

Question The inputs calculated in HyperdriveLP.sol for the average times have a great loss of accuracy since there is a transition between 18 decimals to seconds and then back to 18 decimals. An error of 1e18 seconds in longAverageMaturityTime is then multiplied by a factor of 1e18/ _positionDuration ~ 1e10.

Mitigation The mitigation of question A.1 from Week 4 was to accept the suggestion by adding _calculateTimeRemainingScaled to avoid descaling and rescaling the average time remaining values in HyperdriveLP.

Week 4 Question A.2

Question The netCurveTrade has units of [bonds] but maxCurveTrade has units of [shares] or [bonds/price]. So the ‘else’ branch where the net curve trade is negative, there seems to be a mixture of units, and the value that is passed to the calculateSharesInGivenBondsOut should be in bonds, not shares.

Mitigation The mitigation of question A.2 from Week 4 was to update calculatePresentValue to make use of the new YieldSpaceMath.calculateMaxBuy. After thinking about the problem, we realized that we didn't need to use HyperdriveMath.calculateMaxLong because when we apply the net trade we can assume that the amount of longs outstanding is zero (we close them all and still have outstanding shorts). With this in mind, all we need to do is calculate the amount of shorts that need to be closed to get to a spot price approximating one. We use the change in bonds from calculateMaxBuy, which addresses the units mismatch. The logic itself was also updated based on the recent work on calculateMaxLong.

[github-actions]

Hyperdrive Gas Benchmark

Benchmark suite	Current: 39968bc	Previous: 51af3a2	Deviation	Status
addLiquidity: min	711 gas	711 gas	0%	🟰
addLiquidity: avg	48293 gas	46022 gas	4.9346%	🚨
addLiquidity: max	82677 gas	82666 gas	0.0133%	🚨
checkpoint: min	514 gas	492 gas	4.4715%	🚨
checkpoint: avg	23022 gas	23006 gas	0.0695%	🚨
checkpoint: max	33418 gas	33405 gas	0.0389%	🚨
closeLong: min	659 gas	637 gas	3.4537%	🚨
closeLong: avg	45964 gas	45760 gas	0.4458%	🚨
closeLong: max	81328 gas	80396 gas	1.1593%	🚨
closeShort: min	616 gas	616 gas	0%	🟰
closeShort: avg	40762 gas	40932 gas	-0.4153%	✅
closeShort: max	81627 gas	80864 gas	0.9436%	🚨
initialize: min	645 gas	667 gas	-3.2984%	✅
initialize: avg	139425 gas	139985 gas	-0.4000%	✅
initialize: max	140665 gas	141275 gas	-0.4318%	✅
openLong: min	685 gas	685 gas	0%	🟰
openLong: avg	104937 gas	106144 gas	-1.1371%	✅
openLong: max	147691 gas	147242 gas	0.3049%	🚨
openShort: min	711 gas	711 gas	0%	🟰
openShort: avg	139284 gas	144037 gas	-3.2998%	✅
openShort: max	206577 gas	206572 gas	0.0024%	🚨
removeLiquidity: min	547 gas	547 gas	0%	🟰
removeLiquidity: avg	60959 gas	61634 gas	-1.0952%	✅
removeLiquidity: max	116479 gas	113267 gas	2.8358%	🚨

This comment was automatically generated by workflow using github-action-benchmark.

[jrhea]

just some nits to fix