package zero_test import ( "testing" . "github.com/p9c/p9/pkg/util/zero" ) var ( bytes32 = make([]byte, 32) // typical key size bytes64 = make([]byte, 64) // passphrase hash size bytea32 = new([32]byte) bytea64 = new([64]byte) ) // xor is the "slow" byte zeroing implementation which this package originally replaced. If this function benchmarks // faster than the functions exported by this package in a future Go version (perhaps by calling runtime.memclr), // replace the "optimized" versions with this. func xor(b []byte) { for i := range b { b[i] ^= b[i] } } // zrange is an alternative zero implementation that, while currently slower than the functions provided by this // package, may be faster in a future Go release. Switch to this or the xor implementation if they ever become faster. func zrange(b []byte) { for i := range b { b[i] = 0 } } func BenchmarkXor32(b *testing.B) { for i := 0; i < b.N; i++ { xor(bytes32) } } func BenchmarkXor64(b *testing.B) { for i := 0; i < b.N; i++ { xor(bytes64) } } func BenchmarkRange32(b *testing.B) { for i := 0; i < b.N; i++ { zrange(bytes32) } } func BenchmarkRange64(b *testing.B) { for i := 0; i < b.N; i++ { zrange(bytes64) } } func BenchmarkBytes32(b *testing.B) { for i := 0; i < b.N; i++ { Bytes(bytes32) } } func BenchmarkBytes64(b *testing.B) { for i := 0; i < b.N; i++ { Bytes(bytes64) } } func BenchmarkBytea32(b *testing.B) { for i := 0; i < b.N; i++ { Bytea32(bytea32) } } func BenchmarkBytea64(b *testing.B) { for i := 0; i < b.N; i++ { Bytea64(bytea64) } }