Matheus Degiovani

> But I was also suggesting that the more common golang idiom for this would be [][]byte as they're basically equivalent; Yep, I got that. The current implementation is just...

Ah, nice find! Looks like it's implemented for both windows and unix `*conn` objects. I'll have to reassess some of my plans based on this, though I should probably at...

Sure, but some workflows may _want_ to copy (because they're building an object with well-known size, not rewriting fields, and want to reuse the object reference on another slice that...

> > they're building an object with well-known size, not rewriting fields, and want to reuse the object reference on another slice that will aliased-upon). > > I've always just...

Not sure about "better", but the only other place where there are discussions on go-capnp is the matrix channel: https://matrix.to/#/!pLcnVUHHRZrUPscloW:matrix.org?via=matrix.org

FYI, the existing Arena implementations now zero out their buffers during `Release()`([single](https://github.com/capnproto/go-capnp/blob/main/arena.go#L145), [multi](https://github.com/capnproto/go-capnp/blob/main/arena.go#L195)). This doesn't particularly address the issue at hand, but at least avoids keeping secrets around in memory...

Do you have a list of specific metrics that would be interesting to you?

Not sure what the issue with different platforms would be. A standard directive would be something like `//go:generate protoc --go_out=. structdef.proto` from inside a package, so a `go generate ./...`...

Some ideas, in no particular order. - [ ] Flag mix transactions as such (ideally this would be done by dcrwallet itself since it knows which transactions were done in...

> I could be wrong here, but I suspect that due to the variable size of records (aka may or may not have the full block hash due to a...