rand.Rand.Seed() vs rand.NewSource()

[Liam Breck]

Could anyone clarify the difference between these two?

r := rand.New(rand.NewSource(...))

var r rand.Rand; r.Seed(...)

[Ian Lance Taylor]

Well, it's not a complete program, and maybe I'm missing something,
but it seems to me that the difference is that the first one will give
you a random number generator and the second one will panic at run
time.

Ian

[jake...@gmail.com]

Following up on Ian's post. It is generally a good idea to go to the playground and make a minimal working "program" to demonstrate your question before posting. In doing so you would have discovered that the zero value of a rand.Rand can not be used, and will panic.

Also, a tip. You can go to the docs for NewSource at https://golang.org/pkg/math/rand/#NewSource, then click on the "NewSource"  link, and get to the source code at https://golang.org/src/math/rand/rand.go?s=1614:1647#L34, which may be informative.

[dja...@gmail.com]

Hi,

I think OP question actually is what is difference  between:

r := rand.New(rand.NewSource(VALUE))

and

r := rand.New(rand.NewSource(ANYTHING))

r.Seed(VALUE)

https://play.golang.org/p/afOfMjcNs1q

[Ian Lance Taylor]

I don't think there is any meaningful difference between those two.

Ian

[prrat...@gmail.com]

Seed initialise default resource once for random method and using NewSource create different resources for random method   

[anderson...@blacklane.com]

If you use the same seed the source will output the same pseudo-random output. Which is quite useful when you need to be able to replay/predict the random behaviour