Pool::collect

(PECL pthreads >= 2.0.0)

Pool::collect回收已完成任务的引用

说明

public int Pool::collect ([ Callable $collector ] )

对于视为垃圾的引用,使用给定的垃圾收集器进行收集

参数

collector

垃圾收集器,它返回一个布尔值表示这个任务是否可以被进行垃圾收集。 仅在极少的情况下需要一个自定义的垃圾收集器。

返回值

池中剩余的待收集的任务数量。

更新日志

版本 说明
v3 collector 参数变为可选参数, 并且返回值改为整数。

范例

Example #1 Pool::collect() 基本用法示例

<?php
$pool 
= new Pool(4);

for (
$i 0$i 15; ++$i) {
    
$pool->submit(new class extends Threaded {});
}

while (
$pool->collect()); // 直到全部的任务都完成执行之后才会继续下面的代码

$pool->shutdown();

User Contributed Notes

meadowsjared at gmail dot com 15-Mar-2016 04:13
Please note, when using the collect function, it's important that you extend the pool class so you can keep checking for finished threads until they're all done.

<?php
class TestWork extends Threaded {
    protected
$complete;
   
//$pData is the data sent to your worker thread to do it's job.
   
public function __construct($pData){
       
//transfer all the variables to local variables
       
$this->complete = false;
       
$this->testData = $pData;
    }
   
//This is where all of your work will be done.
   
public function run(){
       
usleep(2000000); //sleep 2 seconds to simulate a large job
       
$this->complete = true;
    }
    public function
isGarbage() {
        return
$this->complete;
    }
}
class
ExamplePool extends Pool
{
    public
$data = array();
    public function
process()
    {
       
// Run this loop as long as we have
        // jobs in the pool
       
while (count($this->work)) {
           
$this->collect(function (TestWork $task) {
               
// If a task was marked as done
                // collect its results
               
if ($task->isGarbage()) {
                   
$tmpObj = new stdclass();
                   
$tmpObj->complete = $task->complete;
                   
//this is how you get your completed data back out [accessed by $pool->process()]
                   
$this->data[] = $tmpObj;
                }
                return
$task->isGarbage();
            });
        }
       
// All jobs are done
        // we can shutdown the pool
       
$this->shutdown();
        return
$this->data;
    }
}
$pool = new ExamplePool(3);
$testData = 'asdf';
for(
$i=0;$i<5;$i++) {
   
$pool->submit(new TestWork($testData));
}
$retArr = $pool->process(); //get all of the results
echo '<pre>';
print_r($retArr); //return the array of results (and maybe errors)
echo '</pre>';
?>
your dot brother dot t at hotmail dot com 29-Dec-2014 06:33
The example code crashes and made me waste 2 working days
First of all, `Stackable` has no attribute named $worker or it's access method made it inaccessible.

Secondly, `Stackable` also doesn't have `getThreadId()` . It's best practice to use `Thread` class for realization of a thread since it has more control functions. It's better to use `Stackable` for object storage and use it's `run()` as its initialization.

The working example is

<?php
   
class MyWork extends Thread {
        protected
$complete;

        public function
__construct() {
           
$this->complete = false;
        }

        public function
run() {
           
printf(
               
"Hello from %s in Thread #%lu\n",
               
__CLASS__, $this->getThreadId());
           
$this->complete = true;
        }

        public function
isComplete() {
            return
$this->complete;
        }
    }

    class
Something {}

    class
MyWorker extends Worker {

        public function
__construct(Something $something) {
           
$this->something = $something;
        }

        public function
run() {
           
/** ... **/
       
}
    }

   
$pool = new Pool(8, \MyWorker::class, [new Something()]);
   
$pool->submit(new MyWork());

   
usleep(1000);

   
$pool->collect(function($work){
        return
$work->isComplete();
    });
   
var_dump($pool);
?>