First of all, there are many declaration methods, and there are direct and indirect methods, only two modes are listed in general:
Suppose the code we want to load is:
var foo=function(){
("I am content foo() function by never-online");
};
one. Inserting src directly is simple and direct, but has limitations.
1）
<script>
var x=("SCRIPT");
=""; =true;
("HEAD")[0].appendChild(x);
foo();
</script>
Put the above code in the head tag, and in most cases an error occurs when executing, the message is: Error: Missing object
This is because when the object script is created dynamically, it is not fully loaded. Execute the following code and you can find out the reason.
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=gb2312">
<title>never-online dynamic code test page</title>
</head>
<body>
<pre>
The meaning of readyState
- uninitialized: The script object has just been created and the script code has not been loaded;
- loading: The script code is loading;
- loaded: The script code has been read in, but the explanation and execution have not yet begun;
- Interactive: Explain the execution process;
- Complete : The script has been executed.
</pre>
<div ></div>
<script type="text/javascript">
=function(msg,url,line){
("viewer").innerHTML+='<p style="color:red">Error:'+msg+'line:'+line+'</p>';
return true;
}
function bar(u) {
var x=("SCRIPT");
=u;
=true;
("HEAD")[0].appendChild(x);
}
bar("");
(function getReadyState(){
var e=("viewer")
var x=true;
var a = ("SCRIPT");
for (var i=0; i<; i++) {
if (a[i].readyState=='complete' && x!=false) x=true; else x=false
+=(a[i].src?a[i].src+':':'noname:')+a[i].readyState+"<br />";
}
+="<hr/>";
if (x) (); else
=('getReadyState()',1000);
}());
foo();
</script>
<script type="text/javascript">
//<![CDATA[
foo();
//]]>
</script>
</body>
</html>
The initial value is:
:loading
noname:interactive
We can know that it is of course wrong to execute foo() when still in the loading state. But in the next script tag execution, the readyState is complete, so you can execute the foo() function. Therefore, I recommend that you can simply use dynamically to generate script tags to add js urls.
The solution is as follows
1) Use methods to execute, and the functions in it are estimated that they have been loaded before they are executed. This method is still not safe
<script>
var x=("SCRIPT");
=""; =true;
("HEAD")[0].appendChild(x);
('foo()',1000);
</script>
2) Add an extra script tag to place the code to be executed
<script>
var x=("SCRIPT");
=""; =true;
("HEAD")[0].appendChild(x);
</script>
<script>
//One more script tag to place
//The readyState here has been completed.
foo();
</script>
2. Use XMLHttpRequest and dynamic execution, The advantages of this method are quite obvious, but the efficiency may decrease. There is no test. Friends who are interested can test the speed by themselves.
The code is as follows:
<script language="javascript">
function bar(u) {
var x=?new ActiveXObject(""):new XMLHttpRequest();
("GET",u,false);
(null);
s=;
try {(s)}catch(ex){(s)};//Mozilla is roughly the same as IE's method
}
bar("");
foo();
</script>
But this method still has its disadvantages, that is, the code in the script has Chinese, how to deal with it? Then we need to decode it frequently, and decoding is precisely the weakness of js. If vbs is used to decode, then compatibility will be gone. It depends on your specific application. When I load the js package in neverModules, I use the method to parse the code, which can be used to match the js namespace application
Add decoding
<script type="text/javascript">
//<![CDATA[
function bar(u) {
var x=?new ActiveXObject(""):new XMLHttpRequest();
("GET",u,false);
(null);
s=parseScript();
try {(s)}catch(ex){(s)};
}
function parseScript(jscode) {
// -- toCurrentCharset(), by aimingoo Decoding
(''+
'Function Asc2Unicode(n) \n'+
' Asc2Unicode = Chr(n) \n'+
'End Function \n'+
'Function SafeArray2Str(body) \n'+
' SafeArray2Str = CStr(body)\n'+
'End Function','VBScript');
var r1 = /%u(..)(..)/g, r2 = /%([8,9,A-F].)%(..)/g;
var toUnicode = function($0, $1, $2) {return Asc2Unicode(parseInt($1+$2, 16))}
toCurrentCharset = function(body) {
return unescape(escape(SafeArray2Str(body)).replace(r1, "%$2%$1").replace(r2, toUnicode));
}; jscode=toCurrentCharset(jscode);
(jscode, 'JavaScript'); //IE is valid, vbs decoding
return jscode;
}
bar('');
foo();
//]]>
</script>
However, in most cases, the second method should be handled without any problems. If it is strictly executed, the first method still has improved code, such as loading content, parsing the script again and executing it, but the complexity is increased, so there is a very perfect solution that needs to be discussed further.
I won’t write so much, just for a reminder and a role in throwing bricks and attracting jade.
Suppose the code we want to load is:
var foo=function(){
("I am content foo() function by never-online");
};
one. Inserting src directly is simple and direct, but has limitations.
1）
<script>
var x=("SCRIPT");
=""; =true;
("HEAD")[0].appendChild(x);
foo();
</script>
Put the above code in the head tag, and in most cases an error occurs when executing, the message is: Error: Missing object
This is because when the object script is created dynamically, it is not fully loaded. Execute the following code and you can find out the reason.
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=gb2312">
<title>never-online dynamic code test page</title>
</head>
<body>
<pre>
The meaning of readyState
- uninitialized: The script object has just been created and the script code has not been loaded;
- loading: The script code is loading;
- loaded: The script code has been read in, but the explanation and execution have not yet begun;
- Interactive: Explain the execution process;
- Complete : The script has been executed.
</pre>
<div ></div>
<script type="text/javascript">
=function(msg,url,line){
("viewer").innerHTML+='<p style="color:red">Error:'+msg+'line:'+line+'</p>';
return true;
}
function bar(u) {
var x=("SCRIPT");
=u;
=true;
("HEAD")[0].appendChild(x);
}
bar("");
(function getReadyState(){
var e=("viewer")
var x=true;
var a = ("SCRIPT");
for (var i=0; i<; i++) {
if (a[i].readyState=='complete' && x!=false) x=true; else x=false
+=(a[i].src?a[i].src+':':'noname:')+a[i].readyState+"<br />";
}
+="<hr/>";
if (x) (); else
=('getReadyState()',1000);
}());
foo();
</script>
<script type="text/javascript">
//<![CDATA[
foo();
//]]>
</script>
</body>
</html>
The initial value is:
:loading
noname:interactive
We can know that it is of course wrong to execute foo() when still in the loading state. But in the next script tag execution, the readyState is complete, so you can execute the foo() function. Therefore, I recommend that you can simply use dynamically to generate script tags to add js urls.
The solution is as follows
1) Use methods to execute, and the functions in it are estimated that they have been loaded before they are executed. This method is still not safe
<script>
var x=("SCRIPT");
=""; =true;
("HEAD")[0].appendChild(x);
('foo()',1000);
</script>
2) Add an extra script tag to place the code to be executed
<script>
var x=("SCRIPT");
=""; =true;
("HEAD")[0].appendChild(x);
</script>
<script>
//One more script tag to place
//The readyState here has been completed.
foo();
</script>
2. Use XMLHttpRequest and dynamic execution, The advantages of this method are quite obvious, but the efficiency may decrease. There is no test. Friends who are interested can test the speed by themselves.
The code is as follows:
<script language="javascript">
function bar(u) {
var x=?new ActiveXObject(""):new XMLHttpRequest();
("GET",u,false);
(null);
s=;
try {(s)}catch(ex){(s)};//Mozilla is roughly the same as IE's method
}
bar("");
foo();
</script>
But this method still has its disadvantages, that is, the code in the script has Chinese, how to deal with it? Then we need to decode it frequently, and decoding is precisely the weakness of js. If vbs is used to decode, then compatibility will be gone. It depends on your specific application. When I load the js package in neverModules, I use the method to parse the code, which can be used to match the js namespace application
Add decoding
<script type="text/javascript">
//<![CDATA[
function bar(u) {
var x=?new ActiveXObject(""):new XMLHttpRequest();
("GET",u,false);
(null);
s=parseScript();
try {(s)}catch(ex){(s)};
}
function parseScript(jscode) {
// -- toCurrentCharset(), by aimingoo Decoding
(''+
'Function Asc2Unicode(n) \n'+
' Asc2Unicode = Chr(n) \n'+
'End Function \n'+
'Function SafeArray2Str(body) \n'+
' SafeArray2Str = CStr(body)\n'+
'End Function','VBScript');
var r1 = /%u(..)(..)/g, r2 = /%([8,9,A-F].)%(..)/g;
var toUnicode = function($0, $1, $2) {return Asc2Unicode(parseInt($1+$2, 16))}
toCurrentCharset = function(body) {
return unescape(escape(SafeArray2Str(body)).replace(r1, "%$2%$1").replace(r2, toUnicode));
}; jscode=toCurrentCharset(jscode);
(jscode, 'JavaScript'); //IE is valid, vbs decoding
return jscode;
}
bar('');
foo();
//]]>
</script>
However, in most cases, the second method should be handled without any problems. If it is strictly executed, the first method still has improved code, such as loading content, parsing the script again and executing it, but the complexity is increased, so there is a very perfect solution that needs to be discussed further.
I won’t write so much, just for a reminder and a role in throwing bricks and attracting jade.