Add helpers to convert to DescriptorPublicKey and Descriptor<DefinitePublicKey> data structures

[philipr-za]

closes #817

I added the common cases of converting from Descriptor<PublicKey> and Descriptor<XOnlyPublicKey> to Descriptor<DefinitePublicKey>. Not sure if this is considered too specific but it feels like a use case that will be very common.

[apoelstra]

In 5937090:

I think this is too specific. You're the first person I've encountered who wound up with XOnlyPublicKeys (which is a low level secp256k1 type) and wanted to use it in a PSBT like this.

The From impls on the keys themselves are fine.

I agree that it would be nice if we had a .map_pk function or something that was a simpler version of .translate_pk that assumed the hash types were all the same and that the mapping function were infallible. Then you could do descriptor.map_pk(DefiniteDescriptorKey::from).

[philipr-za]

In 5937090:

I think this is too specific. You're the first person I've encountered who wound up with XOnlyPublicKeys (which is a low level secp256k1 type) and wanted to use it in a PSBT like this.

The From impls on the keys themselves are fine.

I agree that it would be nice if we had a .map_pk function or something that was a simpler version of .translate_pk that assumed the hash types were all the same and that the mapping function were infallible. Then you could do descriptor.map_pk(DefiniteDescriptorKey::from).

Fair enough, I will drop those impl's and run rustfmt

[philipr-za]

Removed the Descriptor impl's. Had to make the new constructor public on DefiniteDescriptorKey to allow for the translate_pk to work I left in the unit test to prove the application of these From impl's to the use case in the issue.

[apoelstra]

Looks like the formatter is still not happy. But concept ACK.

[apoelstra]

ACK a00c993; successfully ran local tests

[apoelstra]

BTW if you want to use this today, can you do a backport PR to the 12.x branch (along with a minor version bump)?