Change Details

**Package Changes:** - Enable assembly compilation to be able to use processor specific implementations (docs says that this should only be enabled when building for WebAssembly). **Summarized Changelog:** - Generic hashing got a 10% speedup on AVX2. - New functions to map a hash to an edwards25519 point or get a random point: `core_ed25519_from_hash()` and `core_ed25519_random()`. - `crypto_core_ed25519_scalar_mul()` has been implemented for scalar*scalar (mod L) multiplication. - Support for the Ristretto group has been implemented for interoperability with wasm-crypto. - Portability improvements have been made. - getentropy() is now used on systems providing this system call. - `randombytes_salsa20` has been renamed to `randombytes_internal`. - Support for NativeClient has been removed. - Most `((nonnull))` attributes have been relaxed to allow 0-length inputs to be NULL. - The `-ftree-vectorize` and `-ftree-slp-vectorize` compiler switches are now used, if available, for optimized builds.

**Package Changes:** - Enable assembly compilation to be able to use processor specific implementations (docs says that this should only be disabled when building for WebAssembly and we are not doing that). **Summarized Changelog:** - Generic hashing got a 10% speedup on AVX2. - New functions to map a hash to an edwards25519 point or get a random point: `core_ed25519_from_hash()` and `core_ed25519_random()`. - `crypto_core_ed25519_scalar_mul()` has been implemented for scalar*scalar (mod L) multiplication. - Support for the Ristretto group has been implemented for interoperability with wasm-crypto. - Portability improvements have been made. - getentropy() is now used on systems providing this system call. - `randombytes_salsa20` has been renamed to `randombytes_internal`. - Support for NativeClient has been removed. - Most `((nonnull))` attributes have been relaxed to allow 0-length inputs to be NULL. - The `-ftree-vectorize` and `-ftree-slp-vectorize` compiler switches are now used, if available, for optimized builds.

**Package Changes:** - Enable assembly compilation to be able to use processor specific implementations (docs says that this should only be endisabled when building for WebAssembly and we are not doing that). **Summarized Changelog:** - Generic hashing got a 10% speedup on AVX2. - New functions to map a hash to an edwards25519 point or get a random point: `core_ed25519_from_hash()` and `core_ed25519_random()`. - `crypto_core_ed25519_scalar_mul()` has been implemented for scalar*scalar (mod L) multiplication. - Support for the Ristretto group has been implemented for interoperability with wasm-crypto. - Portability improvements have been made. - getentropy() is now used on systems providing this system call. - `randombytes_salsa20` has been renamed to `randombytes_internal`. - Support for NativeClient has been removed. - Most `((nonnull))` attributes have been relaxed to allow 0-length inputs to be NULL. - The `-ftree-vectorize` and `-ftree-slp-vectorize` compiler switches are now used, if available, for optimized builds.