Two questions:
First question:
The responseText of xmlhttp is always encoded in utf-8 by default. Some time ago, in order to solve this problem at the minimum cost, the entire project was simply encoded in utf-8. There are not many solutions online, and they are all quite messy, which I don’t like. We can’t use utf-8 for everything in the future.
Second question:
The javascript script of the xmlhttp loading page cannot be executed. The event driver is still available. In order to avoid this thing some time ago, I simply wrote the entire script together. It's disgusting. I always do this. Every time I have to load several hundred K of js files, it's disgusting.
So, now that I have time, the most orthodox way is to solve it.
xmlhttp is originally used to operate xml. The things returned by responseText are changed to UTF-8 using binary hard coding on the Internet. The method of encoding into GB2312 is naturally unreliable, and there is nothing you can do if you deal with other encodings. The IXMLHTTPRequest object returned by responseXML relies on the encoding declaration of xml, so it is naturally impossible to be garbled. There's no reason not to use it.
Here will be the html text I need,
]]>
The script can be like this Write a line;
returnValue = xmldom.documentElement.text;
returnValue is the html text I need. It is relatively rigid to use the byte stream to change the encoding. Why not?
The first question is basically solved. No matter what encoding you use, utf-8, gb2315.gbk, or 8859-1, just change the document statement.
Since the first problem uses xml, the second problem is also easy to solve. Simply analyze our needs and execute the script of a page, which is generally on the loaded page. It is in two places, the declaration before loading the HTML and the call when loading is complete. As for the use in the page, it can be driven by DHTML events, then change the structure of XML.
// The script here is equivalent to being written in the head;
]]>
HtmlCode
]]>
// The script here is equivalent to being written after the body;
]]>
To process the script, try writing these lines;
BeginScript = xmldom.documentElement.selectSingleNode("onStart").text;
htmlCode = xmldom.documentElement.selectSingleNode("body").text;
endScript = xmldom.documentElement.selectSingleNode("onEnd").text;
There seems to be no problem under IE, but an error is reported under firefox. The hateful firefox turns out to be textContent, and the selectSingleNode method does not work. The method of browser identification , just copy it anywhere. Now I only care about IE and Firefox, which are the only two on my machine, and then change the code;
The first three lines below are copied from the NetEase blog. If you have any questions, please contact NetEase. I’m too lazy to write it myself, let alone ask me to search it online.
var isIE=(document.all&&document.getElementById&&!window.opera)?true:false;
var isMozilla=(!document.all&&document.getElementById&&!window.opera)?true:false;
var isOpera =(window.opera)?true:false;
if(isIE){
BeginScript = xmldom.documentElement.getElementsByTagName("onStart").item(0).text;
htmlCode = xmldom.documentElement.getElementsByTagName("body").item(0).text;
endScript = xmldom.documentElement.getElementsByTagName("onEnd").item(0).text;
}else{
BeginScript = xmldom.documentElement.getElementsByTagName("onStart").item(0).textContent;
htmlCode = xmldom.documentElement.getElementsByTagName("body").item(0).textContent;
endScript = xmldom. documentElement.getElementsByTagName("onEnd").item(0).textContent;
}
Complete. The rest is how to execute it.
If you need to dynamically execute a script, there are three methods.
a) eval() method of Global object in JavaScript;
b) execScript() method in DHTML window;
c) new Function(); object in JavaScript;
three Each method has its own advantages and disadvantages.
The first method uses the context of the call as the context of the script. There is an obvious scope problem. The scope after execution is only within the body of the called function or method. Disgusting problem;
The second method is executed on the top-level object window of DHTML, so there is no scope problem, but the execScript() method has browser compatibility problems. Just a proprietary method for IE. It cannot be used on Firefox;
The third method, in addition to being inconvenient to use, must be used to declare a method. If you only need to declare a variable, it is also a very annoying problem.
If there are no browser compatibility issues, then using the second option is the best choice.
The third option will not be considered.
The first method is the best if it can solve the scope problem.
The method I thought of is as follows;
We declare a variable in the web page,
that is, var author = "Xi De San Ren";
or declare a function.
That is:
function getBlogUrl(){
return author "The blog address: http://shizhong8841.blog.163.com";
}
At this time, we are equivalent to
window.author = "Xi De Sanren";
window.getMyBlogUrl = function(){
return this.author "The blog address: http://shizhong8841.blog.163.com";
}
So, we just need to expand the window object a little bit.
window.runScript = function(str){
eval(str);
}
Magic this Next, just use this.a=0; or this.funName = function(arg0) {}; Writing scripts in this way and calling the runScript(str); method is equivalent to defining global properties and methods in the page, which can eliminate scope problems and achieve our purpose.
Now, all problems are solved and all tests are passed under InternetElpxerer 6.0 and Firefax 2.0. The code I wrote is too messy, so I won’t post it to avoid embarrassment, but the idea should still be correct.
More questions will be discussed later.