I'm trying to port a simple synchronous bit of PHP to Go, but am having a hard time getting my head around how concurrency works with regards to channels. The PHP script makes a request to get a list of media library sections, then makes requests to get the items within each of these sections. If the section is a list of TV Shows, it then makes a request for each show to get all the seasons and then another to get the episodes within each season.
I've trying writing in pidgeon-go what I expected to work, but I'm not having any luck. I've tried various channel guides online, but normally end up with deadlock warnings. Currently this example warns about item := <-ch used as value and doesn't look like it's waiting on the goroutines to return. Does anyone have any ideas what I can do?
package main
import (
"fmt"
"time"
)
// Get all items for all sections
func main() {
ch := make(chan string)
sections := getSections()
for _, section := range sections {
go getItemsInSection(section, ch)
}
items := make([]string, 0)
for item := <- ch {
items = append(items, item)
}
fmt.Println(items)
}
// Return a list of the various library sections
func getSections() []string {
return []string{"HD Movies", "Movies", "TV Shows"}
}
// Get items within the given section, note that some items may spawn sub-items
func getItemsInSection(name string, ch chan string) {
time.Sleep(1 * time.Second)
switch name {
case "HD Movies":
ch <- "Avatar"
ch <- "Avengers"
case "Movies":
ch <- "Aliens"
ch <- "Abyss"
case "TV Shows":
go getSubItemsForItem("24", ch)
go getSubItemsForItem("Breaking Bad", ch)
}
}
// Get sub-items for a given parent
func getSubItemsForItem(name string, ch chan string) {
time.Sleep(1 * time.Second)
ch <- name + ": S01E01"
ch <- name + ": S01E02"
}