All above discussion is talking about cpu excution order or memory order,
lets make it easy and clear.
If I use a mutex when create a temporary map to guarantee that this
temporary map has been successully created, modified and set to gMap1.
Is it safe setting gMap1 pointer to gMap pointer when reading data from
gMap?
In my opinion, It is safe to set a pointer while another goroutine is
reading this pointer.
// example.go
package main
var gMap = map[int]int{1:100, 2:200}
var gMap1 = make(map[int]int)
var gMap1Lock sync.Mutex
func create(){
gMap1Lock.Lock()//Use a mutex to guarantee that gMap1 has been created
and modified successfully in memory
gMap1 := make(map[int]int)
temp[1] = 100
temp[2] = 200
gMap1Lock.Unlock()
}
func reset() {
for{
create()//Use a mutex to guarantee that gMap1 has been created and
modified successfully in memory
gMap = gMap1//is this map pointer set operation atomic?
}
}
func read() {
for{
local := gMap
println(local[1], local[2])
delete(local,1)
delete(local,2)
}
}
func main() {
go w()
go r()
// ...
}
>
>
--
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.
