PHP object references and object optimization strategies

When object-oriented thinking appeared in PHP5, we had a new summary of the program when constructing PHP programs, encapsulating object characteristics into classes, especially when the PHP framework is applied to actual projects, constructing class objects and calling classes The objects appear to be very particular.

We often find that when we write some programs, we expect that as long as the object is constructed for the first time, subsequent operations will be to directly perform other operations on the object. In fact, every class will It can realize the calling of methods or properties by objects, which is the most basic object-oriented approach. But if you study some framework collections, especially when applied to framework development programs in the MVC mode, you will find that there are many actions in a Controller, which is what we call many methods in the controller. Of course, you You will also find that when a controller controls the operation of a transaction object, there will be many actions to construct the object and operate the data collection. However, you will also find that if you rely on the NEW method to construct the object every time, it will undoubtedly be one time. This time increases the burden on your own memory space. Because constructing a new object undoubtedly means allocating a new memory space to it, but if you actually just require the continuation of the same object action, then it is undoubtedly a very good idea to try to find an already constructed object and call it. The so-called continuation of the actions of the same object is actually easy to explain. For example, if I construct a controller class as Controller A, which contains operations such as adding and deleting data Model A, then our control actions will It is assumed that action Add and actionDelete are present. According to convention, we usually use the single point entry principle of the frameset to find Controller A through the route distributor program, but there will be a separate call to actionAdd or actionDelete, and both actions are For data operations on Model A, each operation must construct a data object according to the convention, and then call SQL to perform the database operation. Assume that when constructing two actions to operate the same data object, the method called is using the same method. In the form of a statement $a=new A(), two operations actually produce two objects, which also occupy twice the memory space at the same time. Just imagine that it is just asking the same person to perform two actions, and two operations are necessary. Do individuals do two things separately? Obviously, under this condition, it is a waste of resources, and when there are too many actions, a large number of object resources will occupy memory, seriously causing system performance to decline, and eventually leading to serious crashes.

In fact, when we write many programs, we have repeatedly emphasized that technology to complete functions is only the most basic part. For a good program or website, it is not only functional issues that need to be studied, but more importantly, overall performance, especially Performance is very important for large utility programs. Frame collections are a good example of this, as they are good at using address references wisely. If something can be handled by one object, let one object handle it, leaving enough space for other objects of different categories.

The best way for framework collections to solve whether object resources are reused is to use global variables. When your object exists, usually when it is created for the first time, load it into a global variable. We can also call it Register the object, so when you consider calling methods on the object in the future, you can directly check whether the object of this class is in the global variable, that is to say, see if it has registered the object. If it exists, directly return the object of this class and inform the address for reference. In this way, it is easy to implement object address reference. When all object variables are created to point to the same memory space, we say that they only have different names at this moment, but they are actually the same object.

So, we can create such a program code as follows:

<?php
$GLOBALS[&#39;objects&#39;][&#39;classname&#39;]=null; //是否注册类，类名变量
$GLOBALS[&#39;objects&#39;][&#39;obj&#39;]=null; //对象变量
function & getSingle($classname){
if($GLOBALS[&#39;objects&#39;][&#39;classname&#39;]==$classname){
return $GLOBALS[&#39;objects&#39;][&#39;obj&#39;];
}else
{
$object= new $classname();
$GLOBALS[&#39;objects&#39;][&#39;classname&#39;]=$classname;
$GLOBALS[&#39;objects&#39;][&#39;obj&#39;]=&$object;
return $GLOBALS[&#39;objects&#39;][&#39;obj&#39;];
}
}
class Test{
var $p1;
function Test(){
$this->p1=1;
}
function add()
{
$this->p1++;
}
function show()
{
return $this->p1;
}
}
$test1=&getSingle(&#39;Test&#39;);
$test1->add();
$test2=&getSingle(&#39;Test&#39;);
echo $test2->show();
$test2->add();
$test3=&getSingle(&#39;Test&#39;);
echo $test3->show();
if($test1===$test2)
echo &#39;yes&#39;;
else
echo &#39;no&#39;;
?>

According to the fact that after the object is created for the first time, the second call to read the data must be based on the original object, which can be done by the subsequent program Seeing that when $test2 removes data, the result is 2, and then after the add operation is performed, when $test3 reads it again, it is found that the result is 3, which is obviously due to the completion of action 2. From this, we see the benefit of using the same object for step-by-step operations. This is because we put the code together. Of course, when we use the framework, we often have this situation, which is to replace the above $test1, $ test2, $test3 and their actions are placed in different actions of the same controller. Because the action blocks are independent of each other, therefore, who of the three object variables is triggered first and who is the first to create the object memory space is random and Determined. The first time is to check whether the object is registered in the global variable and check whether the object address is stored in the global variable. After that, the object address is compared and accessed.