Don't compact arrays

[reinkrul]

This makes it hard to predict what the marshalled JSON object will look like, meaning things like creating presentation submissions will be harder. E.g., input descriptor mapping path can be indexed or not, depending on the number of VCs in the VP.

We also incorrectly applied the principal to serviceEndpoint, fixed that as well.

We should still support parsing compacted arrays, of course.

[gerardsn]

Does this break validation, signature checking, etc. on existing docs/VC/VP?

[gerardsn]

Does this break validation, signature checking, etc. on existing docs/VC/VP?

is that a no?

[reinkrul]

Does this break validation, signature checking, etc. on existing docs/VC/VP?

is that a no?

Now I think of it, we need to check whether JSON-LD signature validation now fails. If existing VCs are unmarshalled/marshalled again using the MarshalJSON methods, the signature might change (since a compacted array became uncompacted).

[woutslakhorst]

Now I think of it, we need to check whether JSON-LD signature validation now fails. If existing VCs are unmarshalled/marshalled again using the MarshalJSON methods, the signature might change (since a compacted array became uncompacted).

The signature is computed over quads. The quads for an object or array of size 1 are the same. So "should" be ok.

[reinkrul]

Tested JSON-LD signatures, feeding the signature checking an uncompacted array does not influence the outcome.