address: Address type and encoding

[jharveyb]

Split from #43. Adds tests.

Currently have a test failure for the TLV encode + decode round trip, with the ScriptKey and InternalKey fields. I think it's related to those fields being pubkey values and not pointers, but not sure.

=== CONT  TestAddressEncoding
    /home/jhb/taro/taro/address/address_test.go:226:
                Error Trace:    address_test.go:78
                                                        address_test.go:172
                                                        address_test.go:226
                Error:          Not equal:
                                expected: secp256k1.PublicKey{x:secp256k1.FieldVal{n:[10]uint32{0x30571be, 0x1322145, 0x1d0fa14, 0x7d489a, 0x946188, 0x239116, 0x3bd77a4, 0x13588c5, 0x2d52456, 0x2d6238}}, y:secp256k1.FieldVal{n:[10]uint32{0x2eb9e12, 0xa5e12e, 0x6c6750, 0x3e06e8e, 0x5e90f7, 0x3306d49, 0x369c176, 0x25d7b1, 0x21c75d3, 0x23ce57}}}
                                actual  : secp256k1.PublicKey{x:secp256k1.FieldVal{n:[10]uint32{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}}, y:secp256k1.FieldVal{n:[10]uint32{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}}}

                                Diff:
                                --- Expected
                                +++ Actual
                                @@ -3,12 +3,12 @@
                                   n: ([10]uint32) (len=10) {
                                -   (uint32) 50688446,
                                -   (uint32) 20062533,
                                -   (uint32) 30472724,
                                -   (uint32) 8210586,
                                -   (uint32) 9724296,
                                -   (uint32) 2330902,
                                -   (uint32) 62748580,
                                -   (uint32) 20285637,
                                -   (uint32) 47522902,
                                -   (uint32) 2974264
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0
                                   }
                                @@ -17,12 +17,12 @@
                                   n: ([10]uint32) (len=10) {
                                -   (uint32) 48995858,
                                -   (uint32) 10871086,
                                -   (uint32) 7104336,
                                -   (uint32) 65040014,
                                -   (uint32) 6197495,
                                -   (uint32) 53505353,
                                -   (uint32) 57262454,
                                -   (uint32) 2480049,
                                -   (uint32) 35419603,
                                -   (uint32) 2346583
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0,
                                +   (uint32) 0
                                   }
                Test:           TestAddressEncoding

Fixes #15

[Roasbeef]

                            actual  : secp256k1.PublicKey{x:secp256k1.FieldVal{n:[10]uint32{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}}, y:secp256k1.FieldVal{n:[10]uint32{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}}}

Looks like the issue is that this one is empty? There's also a slight quirk w/ how things are: with the family key we store the raw key, but only serialize the tweaked key. We'll likely need to make a custom comparison function.

[jharveyb]

Looks like the issue is that this one is empty? There's also a slight quirk w/ how things are: with the family key we store the raw key, but only serialize the tweaked key. We'll likely need to make a custom comparison function.

Seems like addresses should be including only the tweaked key? That's the only part of the FamilyKey struct i imagine would be public or relevant for receiver addresses.

[Roasbeef]

So FWIW, we already have some encode/decode methods for things like these in the existing asset package. I think the only reason not to use those, is if we end up w/ some weird import cycle down the line.

[Roasbeef]

Here's the fix I needed to throw in the proof PR to get the tests passing again: 74cc870#diff-0444dc75ab5f91d776dfc97ff7a8de031c33a2a5c25c0a488b4896ba617b53aaR159

The issue is we have a field in the FamilyKey struct that actually isn't ever encoded on the wire, so when you compare the two versions the in-memory one has it set but the other doesn't.

[dstadulis]

Updating coming later today then JHB needs review

[jharveyb]

Updated to address feedback, use a Taro-specific params struct for handling address HRP, and fix the PubKey encode/decode bug.

Unsure if having this private schnorrPubKeyEncode/Decode is the best option; I imagine we'd want to use this elsewhere. Could adjust naming to better distinguish from the version in the asset package.

[Roasbeef]

Unsure if having this private schnorrPubKeyEncode/Decode is the best option; I imagine we'd want to use this elsewhere. Could adjust naming to better distinguish from the version in the asset package.

Yeah I think ultimately we'll aggregate these into some internal encoding package.

[Roasbeef]

LGTM 🍬