📈assertRevert with message checking.

[MoMannn]

Provide assertRevert with the ability to check the revert message (added in solidity 0.4.22).

[nventuro]

Not 100% sure on what you have in mind with this: do you want to provide an expected error message and have the assert fail on a mismatch?

[MoMannn]

Bingo, that is exactly what I meant.

[nventuro]

With #326 on the horizon, it makes sense to add the capability to the helpers, but I don't know if we want to commit to error messages in OZ itself yet. @frangio WDYT?

[frangio]

There's #888 for discussing revert reasons (error messages) in OpenZeppelin itself.

It makes sense to move forward with this helper though even before we introduce reasons in OpenZeppelin.

[justuswilhelm]

Hi @MoMannn , I was looking into this issue and wanted to extend assertRevert as per your suggestion to check for revert() error messages. Now I realize, that I actually don't know how to access the error message in a Truffle test. Do you have any pointers on how to approach this?

[MoMannn]

@justuswilhelm Sorry I have never went into the details of truffle so I can't really help you here it is just a feature I would like to have.

[nventuro]

Sadly, it looks like this isn't yet possible in web3: web3/web3.js#1707.

[frangio]

web3.js is not the culprit. There is actually no way to retrieve the revert reason from a node. EIP 758 is a proposal to solve that, and there's Geth (ethereum/go-ethereum#16424) and Ganache (trufflesuite/ganache#116) issues to implement it.

I guess we'll have to wait and see how that plays out.

[nventuro]

@justuswilhelm looks like truffle 5.0 will allow for this to be possible: see the relevant beta release notes.

[remon-nashid]

It is worth noting that checking revert reason is currently possible with Ganache CLI, as demoed in this repo https://github.com/remon-nashid/assertRevert.

[frangio]

@remon-nashid Nice! Are you interested in incorporating your code into the OpenZeppelin helper?

[remon-nashid]

@frangio I have since found an equivalent to assertRevert and that is shoudFail.withMessage. I opened an issue earlier and submitted a PR to export that function. See #1502.

[remon-nashid]

IMO this issue can be closed as 'assertRevert' is superseded by shouldFail.withMessage.

[MoMannn]

@remon-nashid agreed. Closing.