perfect-numbers: add canonical data

[stkent]

Closes #575. See that issue for background information. Please contribute any new comments in this PR. Thanks!

I attempted to follow the proposed canonical data spec from #602. If you spot any deviations from the proposal, please let me know and I can correct them. cc @rbasso for his expertise and validation 😄 [One thing I was unsure of: whether it is expected that inputs/outputs always be strings, or whether other data types (as used for inputs here) are permitted by that spec.]

[rbasso]

I didn't followed the discussion about the tests' contents, so I'm reviewing just the format.

I attempted to follow the proposed canonical data spec from #602.

Nice to see the schema being used. 😄

If you spot any deviations from the proposal, please let me know and I can correct them. cc @rbasso for his expertise and validation 😄

I just validated the canonical-data.json file here, so it is formally correct.

...[One thing I was unsure of: whether it is expected that inputs/outputs always be strings, or whether other data types (as used for inputs here) are permitted by that spec.]

description, property and error need to be string, other values can be anything. In this case, representing yourinput values as numbers seems a perfect choice. 👍

[stkent]

@rbasso thanks for the checks and confirmations; I had not seen the schema lint site before, that's really neat!

[ErikSchierboom]

Great work @stkent! 🎉

[stkent]

Thanks!