Return early in PosBase::transfer if an attempt is made to transfer zero tokens

[james-chf]

Relates to #603

This potential PR would add a helper fn for checking if an Amount::is_zero, and use it to short circuit PosBase::transfer if somehow a zero amount was attempted to be transferred.

[james-chf]

pls update wasm

[james-chf]

@tzemanovic not sure if this is worth adding, to resolve audit issue #603, if not this PR can be close

[tzemanovic]

@tzemanovic not sure if this is worth adding, to resolve audit issue #603, if not this PR can be close

If I understand the issue correctly, it is that we're not checking that slash amount is greater than 0, which is probably a good thing to verify with cubic slashing. I think we should somehow ensure this from slashing logic, maybe having some lower bound param or something.

The transfer with 0 is almost no-op (it might write 0 into target balance, if there was none before), but I think that might be besides the point. I'm happy to merge this as the early return is better, but I don't think it solves the issue (if I grokked it right). Let's please the leave the issue open for now as we still have to change the slashing logic substantially.

[james-chf]

If I understand the issue correctly, it is that we're not checking that slash amount is greater than 0

yep, it is almost a noop like you say, if we were to slash a validator by 0 (which should not really be occurring in any case) - we may end up initializing a balance to be explicitly 0 where there was none before (but no balance and 0 balance should be being treated functionally equivalently)

[tzemanovic]

I rebased this into v0.14.3 - the PosBase trait is gone and we're just using token::transfer in PoS now, so I moved the short-circuiting there

[tzemanovic]

pls update wasm