Add support for CBOR encoding

[cmitsakis]

I needed support for CBOR encoding, so I created this.

It is enabled only if you use the --encoding-cbor flag.
Ideally it shouldn't need a flag. I should detect if the value is valid CBOR, but the function cbor.Valid() from https://github.com/fxamacker/cbor recognizes as CBOR values that are not CBOR.

I can make changes if you want.

[x448]

@cmitsakis The cbor.Valid() in v2.4.0 ignores extraneous data after the valid CBOR data in order to be consistent with Decoder which can continue validating and decoding extra data.

Next release of fxamacker/cbor will add an option to treat the extraneous data differently in Valid().

Source:

Comment by @fxamacker at fxamacker/cbor,

Unmarshal() and Valid() work this way to be consistent with decoder, which can continue decoding extra data. But I understand your point because Unmarshal() and Valid() only decodes one piece of data.

Rejecting extraneous data would be a breaking change. So maybe it can be added as a non-default DecOption. This new option would only apply to Unmarshal() and Valid().

• bug: Unmarshal and Valid should check following extraneous data fxamacker/cbor#359
• Check following extraneous data (#359) fxamacker/cbor#360

[x448]

@cmitsakis fxamacker/cbor#380 fixed the issue you mentioned about Valid() after classifying it as a bug:

After rereading relevant sections in RFC 8949 and comparing to how this is handled in encoding/json, I changed this from enhancement to a bug.

[boreq]

As they say "the mills of free software grind slowly". I created a new pull request #6 to make everything more robust, avoid config flags, make adding more formats easy etc. Let me know if it works for you or not, I rarely work with CBOR so its possible I missed something.