fix: Fixes the problems that result from empty share reserves

[jalextowle]

Fixes: #385.

To add some context to the issue, this PR adds a minimum share reserves that should prevent a host of critical and high severity issues from being possible. During the development of this feature, we noticed another DOS vector that could result from the share reserves being reduced to a small non-zero value. As part of this PR, we found sane defaults for the major yield sources and added tests demonstrating that these default values are sufficient to prevent the numerical issues that have been found with _updateLiquidity.

This minimum share reserves change consists of two parts:

• Making sure that the share reserves never falls below a minimum threshold.
• Minting the zero address the minimum share reserves of LP shares during intialization.

With this in mind, we have minimumShareReserves shares set aside in the share reserves that sets a hard lower bound on what the value can be. Similarly, we have minimumShareReserves LP shares set aside to serve as a hard lower bound for the active LP total supply.

[coveralls]

coverage: 96.372% (-0.9%) from 97.26% when pulling 190807c on jalextowle/bug/empty-share-reserves-2 into 3ccbe22 on main.

[github-actions]

Hyperdrive Gas Benchmark

Benchmark suite	Current: 190807c	Previous: 271650c	Deviation	Status
addLiquidity: min	709 gas	709 gas	0%	🟰
addLiquidity: avg	45882 gas	45450 gas	0.9505%	🚨
addLiquidity: max	77728 gas	83400 gas	-6.8010%	✅
checkpoint: min	514 gas	514 gas	0%	🟰
checkpoint: avg	23019 gas	23022 gas	-0.0130%	✅
checkpoint: max	33407 gas	33418 gas	-0.0329%	✅
closeLong: min	659 gas	659 gas	0%	🟰
closeLong: avg	47495 gas	46621 gas	1.8747%	🚨
closeLong: max	83240 gas	81764 gas	1.8052%	🚨
closeShort: min	616 gas	616 gas	0%	🟰
closeShort: avg	42212 gas	41352 gas	2.0797%	🚨
closeShort: max	68589 gas	81627 gas	-15.9727%	✅
initialize: min	532 gas	529 gas	0.5671%	🚨
initialize: avg	160281 gas	139316 gas	15.0485%	🚨
initialize: max	161781 gas	140816 gas	14.8882%	🚨
openLong: min	661 gas	661 gas	0%	🟰
openLong: avg	116401 gas	116166 gas	0.2023%	🚨
openLong: max	175911 gas	175672 gas	0.1360%	🚨
openShort: min	709 gas	709 gas	0%	🟰
openShort: avg	157459 gas	159418 gas	-1.2288%	✅
openShort: max	218810 gas	218571 gas	0.1093%	🚨
removeLiquidity: min	569 gas	569 gas	0%	🟰
removeLiquidity: avg	62904 gas	59523 gas	5.6802%	🚨
removeLiquidity: max	116558 gas	116368 gas	0.1633%	🚨

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

[jrhea]

Okay i made an initial pass, but i still haven't reviewed the tests

[ControlCplusControlV]

Minor nits from me + Jonny's suggestions, can re-review once those are patched

[ControlCplusControlV]

Have you looked at fixing CI? Would prefer if CI failing was not the norm

[ControlCplusControlV]

Some minor nits, looks really good but I still have some outstanding questions and still a lot of commented out code

[jalextowle]

Have you looked at fixing CI? Would prefer if CI failing was not the norm

It's just a spelling issue. I fixed most of the spelling mistakes before I left, but I missed this one. My style is generally to review when the linter is broken or a unit tests or two is broken, but to make sure that everything is fixed prior to the PR being merged. I'm happy to do something different, but in this particular case, it would have slowed down our development cycle over a typo.

[ControlCplusControlV]

Hmm, looks like ERC4626HyperdriveDeployer has broken the codesize limit by 105.0

[ControlCplusControlV]

[gtowle03]

Excellent work. I am very proud