TABconf was great! I spent the two workshop days hacking on hashpool with Gary. We don't get to meet in person much, even though we live relatively close. We didn't produce a code branch that compiles but Gary's knowledge of Sv2 was very useful in navigating the codebase and understanding the flaws of my previous design. When I returned home and continued work I was able to reference the two branches we created at TAB to push forward. I was previously trying to send the keyset_id downstream to the proxy on the SetupConnection message, but this is a common message that is shared between many Sv2 roles which explains my issues with the job declarator. A better design is to include the keyset_id on the extended channel success message which does not involve the job declarator.
I gave my talk at 9am on the last day of the conference. Honestly, I was not feeling 100% so it was kind of low energy. I feel like I did a better job at btc++ so I'm glad it's the one with a recording. After this I hosted the Lightning Socratic Seminar and sat on a Mining Socratic Panel which I feel went really well. All of my events were in the socratic village which doesn't allow recordings.
In the two weeks since then I have successully implemented passing the keyset_id downstream from the pool to the proxy. The proxy now uses the keyset_id to generate blinded secrets. Passing the blinded secret around is more complicated than the keyset_id. keyset_id can be represented as a u64, which is a primitive type. The blinded secret is composed of multiple types, including a cryptgraphic key. I implemented a Sv2BlindedMessage struct using Sv2's custom data types to enable it to be serialized and sent back upstream. I had a hell of a time getting it to derive the Serialize and Deserialize trait for this struct. I think I was just pushing myself too hard under false time constraints. After a night of rest I was able to find out the problem. I needed to import Decodable and Encodable from both the binary_codec_sv2 and the derive_codec_sv2 crates. These traits are aliased to Serialize and Deserialize everywhere in the codebase. This confusing arrangement threw me for a loop. I should find out why it's written this way.
Another stumbling block I ran into was converting the 33 byte cdk pubkey (actually a secp256k1::PublicKey) into a 32 byte Sv2 PubKey. I just dropped the first byte that encodes the parity of the y coordinate on the elliptic curve and left a TODO comment to resolve later. Now that I think about it, I should probably just add a parity byte to the struct. I will work on this tomorrow.
Aside from dev work I also submitted a grant application to HRF. I went a little crazy and wrote 2500 words to describe my project. The web form asked for a detailed description! When I tried to paste my answer in I realized they only give you 2000 characters. Oops. I threw up a secret gist and linked to it in my application. Just like the talk I gave, this was a good exercise in clarifying my thinking. I can probably reuse this text elsewhere in the future.
I need to apply to Spiral and OpenSats also. Part of the reason I haven't done this yet is that I set a goal to do at least a little dev work every day. And I mean every day. Today marks 13 consecutive days with a github contribution. I'm going to keep the streak going as long as I can; I'm calling it Operation Greenwash. My buddy Johnny pointed out that I need to pay attention to my github profile since grant committees will be looking at it. The green squares in github are far from a perfect metric but if it's the metric I have to game to get a grant then that's just what I will do.
After my experience with Serialize and Deserialize last weekend I realize I need to get better at stopping work and committing what I have rather than thrashing in pursuit of a minor goal. This is an unhealthy approach that will eventually lead to burnout. It's a marathon, not a sprint. I should also get better at queueing up small tasks to knock out on the weekend. I think that contributing to hashpool and/or cdk daily will help me more quickly climb the skill curve and enable me to tackle bigger tasks, which will surely be needed to get this thing across the finish line.