Show Posts
Any news on 0.97? (features, not ETA 
)
)Bunch of stuff going on in the dev branch.
A lot of these changes are a result of a partnership I hold with an upcoming exchange. The most significant parts are:
1) Optimizing the DB to operate a c10k capable supernode (basically a chain explorer service).
2) In that same vain, swapping the server interface from FCGI to websockets.
3) Since I don't like the CA model SSL hinges upon, I'm having droark implement BIP151 for the client/server encryption layer. This is done in consultation with Core, and we're hoping the code will be accepted into Core when they're ready to work on this layer.
From the user perspective, these changes mostly result in better stability. BIP151 and websockets will improve privacy and quality of life for the few users who use remote DBs. For the stuff that's more immediately relevant to individual users there's:
1) new wallet format
2) BIP32 support
3) bech32 support
For .1-2, I also intent to swap out cryptopp in full for Jonas Schnelli's libbtc, which operates as a wrapper around the Core crypto libraries in Armory's use case.
There will be some work needed to accomodate the new wallet features in the GUI. I'd also like to use this opportunity to introduce an advanced filter feature for ledgers (some sort of pseudo regex), so that people can customize their ledgers without me having to hardcode a filter GUI for each scenario.
I would also like to leverage the new wallet code and its flexibility with meta data to introduce a new privacy feature, in which you can create and manage a list of vendors and voluntarily flag your payments to that vendor, to improve coin selection privacy.
As an example, assume you use BitPay vendors. You would add BitPay as a vendor to your wallet, and flag the addresses of theirs that you use while construction payment tx. This would allow the coin selection to prioritize change addresses you used with BitPay in the past, and warn you when you have to introduce an unrelated UTXO to meet your payment. I've noticed through my own usage that I often end up moving coins to the same handful of services, and the code is lacking the meta data that would otherwise minimize my coins exposure to these services.
There are other big items on the list, like hardware wallet support, blocks over p2p, common export formats like BIP39, py3/qt5, but these will most likely be for 0.98.
I also have a couple contacts at Lightning Labs and am looking to interface Armory with their daemon so that people can at least create, close and revoke LN channels in the future, with the understanding that most of the heavy lifting will be offloaded to their code. This would be either 0.98 or 0.99, and obviously optional.
