invariant_reserves failure when take / bucketTake

[grandizzy]

Description

reserves invariant failure

Add logic in invariant test harness to allow reserves to increase up to ~1e-18th of the current inflator.

Purpose

Because debt is stores in t0 terms and reserves include current debt, due to roundoff error the reserves will fluctuate up to 1 WAD times the current inflator value. This change allows for this expected error in reserves changes, enabling deeper invariant test runs. In particular, it fixes regression sequences test_regression_failure_reserves_bucketTake_18precision and test_regression_failure_reserves_regularTake_18precision.

Impact

Core contracts are untouched by this change, so no impact on size or gas.

Tasks

• Changes to protocol contracts are covered by unit tests executed by CI.
• Protocol contract size limits have not been exceeded.
• Gas consumption for impacted transactions have been compared with the target branch, and nontrivial changes cited in the Impact section above.
• Scope labels have been assigned as appropriate.
• Invariant tests have been manually executed as appropriate for the nature of the change.

[ith-harvey]

Could be missing something but is this accurate? -> https://github.com/ajna-finance/contracts/blob/6da58b4cdd2f75decac124502da0c55f6c70e9d2/tests/INVARIANTS.md?plain=1#L67

also see -> https://github.com/ajna-finance/contracts/blob/dcae802c991caab094af74fe7a47776117d3b882/tests/forge/invariants/base/ReserveInvariants.t.sol#L20-L21

I think the reservesErrorMargin value has a maximum of 1e13 value not 1e15. Suggest invariants.md is updated to match the code accordingly.

[EdNoepel]

Fine with this. Note I was not able to reproduce the failure at 600 depth. Also, merge conflict needs resolution.

[ith-harvey]

Fine with this. Note I was not able to reproduce the failure at 600 depth. Also, merge conflict needs resolution.

Fixed. -> dcae802

[ith-harvey]

Could be missing something but is this accurate? ->

https://github.com/ajna-finance/contracts/blob/6da58b4cdd2f75decac124502da0c55f6c70e9d2/tests/INVARIANTS.md?plain=1#L67

also see ->

https://github.com/ajna-finance/contracts/blob/dcae802c991caab094af74fe7a47776117d3b882/tests/forge/invariants/base/ReserveInvariants.t.sol#L20-L21

I think the reservesErrorMargin value has a maximum of 1e13 value not 1e15. Suggest invariants.md is updated to match the code accordingly.

Changed 1e15 -> 1e13 to reflect code -> e8abacf

[ith-harvey]

LGTM

[MikeHathaway]

LGTM