This PR cleans up the schnorrkel code a little to prepare for a new release.
Switched back to curve25519-dalek from curve25519-dalek-ng (it turns out using curve25519-dalek-ng was a bad decision as that fork is effectively dead now)
Updated to Rust 2021
Removed old cruft and dead code
Fixed all clippy lints
Updated all of the dependencies
Cleaned up the feature flags a little
Marked musig as deprecated (as requested by @burdges)
Bumped version to 0.11.0
This should not yet be released on crates.io. Before we cut a new release I want to do two more things:
1) Wait until my PR adding runtime SIMD autodetection is merged in curve25519-dalek and a release is cut, and then bump schnorrkel to use that. This should provide massive speedups in the range of ~30%-~50%.
2) Do more testing. I'm going to run a full substrate burn-in to make sure nothing's accidentally broken. I might also do some fuzz testing to make sure the behavior's equivalent to the 0.9.1 version currently used by substrate.
This PR cleans up the
schnorrkel
code a little to prepare for a new release.curve25519-dalek
fromcurve25519-dalek-ng
(it turns out usingcurve25519-dalek-ng
was a bad decision as that fork is effectively dead now)clippy
lintsmusig
as deprecated (as requested by @burdges)This should not yet be released on crates.io. Before we cut a new release I want to do two more things:
1) Wait until my PR adding runtime SIMD autodetection is merged in
curve25519-dalek
and a release is cut, and then bumpschnorrkel
to use that. This should provide massive speedups in the range of ~30%-~50%. 2) Do more testing. I'm going to run a fullsubstrate
burn-in to make sure nothing's accidentally broken. I might also do some fuzz testing to make sure the behavior's equivalent to the 0.9.1 version currently used bysubstrate
.