shakedex: use derive path for private key

[chikeichan]

this is harder than i thought - mostly because of all the edge around rescanning auctions.

This probably take some time for review and won't make it into 0.9.0-rc.1 tomorrow. let's see if we can include this before 0.9.0 turn stable.

Includes:

• [x] Generate script address using standardized branch path
• [x] Add rescan listing
• [x] Add rescan fills by name
• [x] Add loading state in UI
• [x] Add HIP-0009 for deriving key for script address

requesting review from @kurumiimari @pinheadmz

[pinheadmz]

I plan on reviewing this in depth but the first comment I have is this: I think we should try to limit Bob just to UI functions and keep wallet / node features exclusively in hsd (this applies to rescan stuff / batch TXs, too).

Especially now that you have proposed HIP-0009 formally we can begin working on using the wallet software we already have to recover HIP-9 addresses, etc. I understand that may slow down progress, but I think that is ok, it will certainly get better review and test coverage that way and opens the door for other HIP-1 implementations to build on hsd. Are there any other concerns with opening this kind of PR to hsd? We could discuss in an issue or something.

[pinheadmz]

You might want to follow / review https://github.com/handshake-org/hsd/pull/639 as well. @anunayj is also interested in a HIP-9 standard custom script and we are working on adding custom script signing to hsd wallet.

[chikeichan]

I plan on reviewing this in depth but the first comment I have is this: I think we should try to limit Bob just to UI functions and keep wallet / node features exclusively in hsd (this applies to rescan stuff / batch TXs, too).

Especially now that you have proposed HIP-0009 formally we can begin working on using the wallet software we already have to recover HIP-9 addresses, etc. I understand that may slow down progress, but I think that is ok, it will certainly get better review and test coverage that way and opens the door for other HIP-1 implementations to build on hsd. Are there any other concerns with opening this kind of PR to hsd? We could discuss in an issue or something.

yea i am open to it. so i think we would need...

• add support for custom branch to the bloomfilter in wallet.rescan, maybe enable by a hip9 (uint32[]) config to WalletDB?
• (unsure?) return hip9 attributes on tx and domain object from API. this help UI with identifying non-default TX and Domains. In the case of shakedex, it is used to show listings and fills. we could maintain a reverse index of addresses to hip9 for more efficient look up

[pinheadmz]

Yeah so if you look through wallet/account.js in hsd theres a bunch of functions like deriveChange() deriveReceive() and syncDepth() which calls both of those. Then when you get an account object it has properties like:

  "receiveAddress": "rs1qy9uplxpt5cur32rw3zmyf8e7tp87w8slly2fms",
  "changeAddress": "rs1q30ppv5gyrwpy4wyk0v6uzawxygdtvrpux8yrg2",

So the naive way to go forward with this would be duplicate every single one of those and change it to deriveSakedex() and shakedexAddress: ... and all that... but we'll have to do that for every single HIP9 proposal forever.

What we may want to do is add a new wallet type or account type similar to Bitcoin's descriptor wallets where arbitrary scripts with public-key placeholders can be imported. The wallet will then proceed as normal: generate the lookahead amount of addresses and add them to the bloom filter for re-scanning, etc.

I think the hsd wallet should not necessarily be required to know how to spend from those outputs, I'm not sure about that yet. As long as the wallet has good API for signing arbitrary scripts we should be ok with minimal additions.