Use of weak random number generator (math/rand instead of crypto/rand)
90func NegativeUniformDistributedFloat32VectorGenerator(n, dim int) (vecs [][]float32) {
91 left, right := dim/2, dim-dim/2
92 lvs := genF32Slice(n, left, func() float32 {
93 return -rand.Float32() 94 })
95 rvs := UniformDistributedFloat32VectorGenerator(n, right)
96 vecs = make([][]float32, 0, n)
Description
math/rand
is much faster for applications that don’t need crypto-level or security-related random data generation. crypto/rand
is suited for secure and crypto-ready usage, but it’s slower. But in most cases, crypto/rand
is likely to be more suitable, unless the performance is critical but the application's security is not (which is rare).
It is highly recommended to use crypto/rand
when needing to be secure with random numbers such as generating session ID in a web application.
- crypto/rand package
- math/rand package
Bad practice
package main
import "math/rand"
func main() {
bad := rand.Int()
println(bad)
}
Recommended
package main
import "crypto/rand"
func main() {
good, _ := rand.Read(nil)
println(good)
}