It looks the results of RNG are not distributed well.
If I change the order of the cases, the final print result will become even
larger than 1.02.
package main
import "fmt"
func main() {
foo, bar := make(chan struct{}), make(chan struct{})
close(foo)
close(bar)
x, y := 0.0, 0.0
f := func(){x++}
g := func(){y++}
for i := 0; i < 1000000; i++ {
select {
case <-foo: f()
case <-foo: f()
case <-foo: g()
case <-foo: f()
case <-foo: g()
case <-foo: g()
}
}
fmt.Println(x/y)
}
This one constantly prints 1.04+
On Friday, October 27, 2017 at 4:08:39 PM UTC-4, T L wrote:
>
> The following program constantly prints a value larger than 1.02
>
> package main
>
> import "fmt"
>
> func main() {
> foo, bar := make(chan struct{}), make(chan struct{})
> close(foo)
> close(bar)
> x, y := 0.0, 0.0
> f := func(){x++}
> g := func(){y++}
> for i := 0; i < 1000000; i++ {
> select {
> case <-bar: g()
> case <-foo: f()
> case <-bar: g()
> case <-foo: f()
> case <-bar: g()
> case <-foo: f()
> }
> }
> fmt.Println(x/y)
> }
>
> if the last two cases are removed, then it becomes quit fair.
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
