crypto-square wrong tests

[celavek]

Hi,

I think that the last 2 tests for the crypto-square problem are wrong. You can verify that by
stacking the segment in output of the normalized cipher text which should yield the plaintext in
normalized form:

E.g.

vrela 
epems     ==>     veeorptieeprlmapasoo (far away from truth) 
etpao 
oirpo

Correct

vrel 
aepe 
mset       ==>  vampiresarepeopletoo
paoo 
irpo

Also the following result

msemoa anindn inndla etltsh ui
does not satisfy the property stated in the README

Imperfect squares will have n empty spaces. Those spaces should be distributed evenly across the last n rows.

[arcuru]

Yea, you're correct on that. It appears everybody in this track just followed the tests and didn't pay much attention to the readme.

Would you like to fix it?

[LegalizeAdulthood]

I shamelessly imitated the JavaScript track while implementing the C++ track 😺

[celavek]

@patricksjackson OK, I could give it a try. I'm not sure about the time frame as I came across it while implementing the D track and my main focus is right now on that.

[kytrinyx]

There's a ginormous discussion about this over in the metadata repo: exercism/problem-specifications#190

We have a canonical data set here, I think: https://github.com/exercism/x-common/blob/master/crypto-square.json

[arcuru]

Oops, this is still an issue. I'm looking into it.

The correct test suite is now located here: https://github.com/exercism/problem-specifications/blob/master/exercises/crypto-square/canonical-data.json