Fixes false positives in expectTransactionFailedAsync

[hysz]

Description

The function expectTransactionFailedAsync(transactionReceipt, expectedRevertReason) returns a false positive when the expected revert reason is a substring of the exception message.

For example, if the expected revert reason is INVALID_DATA and the actual reason is INVALID_DATA_OFFSET then there will be a false positive.

To go one step further, the full exception message is VM Exception while processing transaction: revert INVALID_DATA_OFFSET. Any substring within that message will also return a false positive, for example, processing trans.

Testing instructions

Types of changes

Checklist:

• Prefix PR title with [WIP] if necessary.
• Add tests to cover changes as needed.
• Update documentation as needed.
• Add new entries to the relevant CHANGELOG.jsons.

[coveralls]

Coverage decreased (-0.02%) to 83.353% when pulling 56bc294 on fix/contracts/expectTransactionFailedAsyncFalsePositives into b2cf701 on development.

[albrow]

I don't think this will work in Geth. We have disabled Geth tests for now (see #1754 and #1775). Going forward, instead of running all of our unit tests through Geth and relying on features that don't exist upstream, we want to have integration tests that operate at a higher level and don't rely on any extra features.

Anyways, I'm okay with this for now but if we write integration tests for Geth in the future we'll have to update it.

[albrow]

To go one step further, the full exception message is VM Exception while processing transaction: revert INVALID_DATA_OFFSET. Any substring within that message will also return a false positive, for example, processing trans.

Oof. If this is true we should report the issue to Mocha/Chai. I can't imagine this was the intended behavior.

[albrow]

Like I mentioned we might have to change this in the future to support revert reasons from Geth (and maybe Parity too), but I'm okay with leaving it as is for now.

[hysz]

Oof. If this is true we should report the issue to Mocha/Chai. I can't imagine this was the intended behavior.

@albrow I had the same thought and checked out their source code (here). Funny enough it does appear to be intentional ¯\_(ツ)_/¯...