need to fix the test failure of the `TestNonce` method about iota.go

[knarz]

The test used an old magnitude (18), so I changed it to use the DefaultMinWeightMagnitude constant, and the trytes were for another mwm.
Also the HasValidNonce was counting trailing 9 trytes, which works fine for any mwm mod 3 == 0 but would fail for the current mwm of 14.

I think that's everything.

[jeffwillette]

I am looking into fixing that too. It is caused by the Hash function. It was introduced in 5ef65e7 and if I change the numberOfRounds in curl.go back to 27 (the value before this commit), it passes.

This is my first time looking into this code so I am unsure about how everything is working and I haven't been able to come up with a fix so far...

[jeffwillette]

Reading over the spec and the javascript implementation and tests (https://github.com/iotaledger/iota.lib.js/blob/433f5d37a684abf875aaa6f1fb97082ab06218f0/lib/utils/inputValidator.js#L94) I think this is just a mistake in the test...

It looks like Curl-P-27 ends with 6 9's, but Curl-P-81 just needs to be 81 valid trytes...

[knarz]

This should be fixed with the latest commit.

We need to count the trailing zeros of the transaction hash and have >= mwm zeroes at the end.

Please re-open if there are any more problems.

[jeffwillette]

@knarz it looks like the problem was the input trytes into the test?

I am trying to understand the code here, so if you have time for a short explanation of why the input trytes were wrong, I would appreciate it!