[R&D review] Wrong hypothetical balance state when entering a market

[QGarchery]

https://github.com/morpho-dao/morpho-v1/blob/b45de2424c6c1a89671af890fc2d6c7f4a8a7ee7/contracts/aave-v2/lens/UsersLens.sol#L218

This check affects whether or not the hypothetical withdrawborrow is taken into account. If the user is not part of the poolToken market, it means that hypothetical borrows are not taken into account at all, potentially giving the false impression that the user is able to borrow

[MerlinEgalite]

Clearly yes

[Rubilmax]

This just made me realize that Morpho-Compound's _isLiquidatable and MorphoAaveV2's _liquidityData are working as expected only because the entered markets of the user are updated before calling the function. Perhaps this should be added in the NatSpec

[Rubilmax]

I also realized that underflows are not happening when subtracting liquidity data due to an hypothetical withdraw because we the hypothetical withdraw amount passed in this function is at most the supply amount of the user on the market
Lots of edge cases were not covered here, this is not an easy one