i get some HTML it a as ajax response, and i need to get just the body contents. So i made this regex:
/(<body>|<\/body>)/ig
works well in all browser but for some reason IE gives me an other array when i use split:
data.split(/(<body>|<\/body>)/ig)
In all normal browsers the content of the body is split(/(<body>|<\/body>)/ig)[2] but in ie its in split(/(<body>|<\/body>)/ig)[1]. (tested in IE7 & 8)
Why is this? And how could i modify it, in order to get the same array in all browsers?
edit just to clarify. I alrady have a solution as mentioned by tobyodavies. I want to understandy, why it behaves differently.
this is the HTML from the response: (the string in data)
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "">.dtd">
<html xmlns="" xml:lang="de" lang="de" dir="ltr">
<head>
blablabla...
</head>
<body>
<div class="iframe">
<div id="block-menu-menu-primary-links-user" class="block-menu">
<h3>Primary Links - User</h3> <div class="content"><ul class="menu"><li class="leaf first"><a target="content" href="#someurl" title="">Login</a></li>
<li class="leaf last"><a target="content" href="#someurl" title="">Register</a></li>
</ul></div>
</div>
</div>
</body>
</html>
PS: i know that parsing HTML with regex is bad, but its not my code, i just need to fix it.
i get some HTML it a as ajax response, and i need to get just the body contents. So i made this regex:
/(<body>|<\/body>)/ig
works well in all browser but for some reason IE gives me an other array when i use split:
data.split(/(<body>|<\/body>)/ig)
In all normal browsers the content of the body is split(/(<body>|<\/body>)/ig)[2] but in ie its in split(/(<body>|<\/body>)/ig)[1]. (tested in IE7 & 8)
Why is this? And how could i modify it, in order to get the same array in all browsers?
edit just to clarify. I alrady have a solution as mentioned by tobyodavies. I want to understandy, why it behaves differently.
this is the HTML from the response: (the string in data)
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "">http://www.w3/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3/1999/xhtml" xml:lang="de" lang="de" dir="ltr">
<head>
blablabla...
</head>
<body>
<div class="iframe">
<div id="block-menu-menu-primary-links-user" class="block-menu">
<h3>Primary Links - User</h3> <div class="content"><ul class="menu"><li class="leaf first"><a target="content" href="#someurl" title="">Login</a></li>
<li class="leaf last"><a target="content" href="#someurl" title="">Register</a></li>
</ul></div>
</div>
</div>
</body>
</html>
PS: i know that parsing HTML with regex is bad, but its not my code, i just need to fix it.
Share Improve this question edited Apr 4, 2011 at 9:32 meo asked Apr 4, 2011 at 9:10 meomeo 31.3k19 gold badges89 silver badges123 bronze badges 11- don't use regexes to parse HTML... the <center> cannot hold, it is too late! stackoverflow./questions/1732348/… – tobyodavies Commented Apr 4, 2011 at 9:13
- i know that its bad. Its not my code. I just need to fix it but thank you :P I wonder why the result is different – meo Commented Apr 4, 2011 at 9:14
- In your situation an XML parser will be more appropriate than a regex. – Stephan Commented Apr 4, 2011 at 9:16
- Is it because IE is using a 0 based array and the rest 1? – BugFinder Commented Apr 4, 2011 at 9:16
- 1 The following page lists differences in the 'split' implementation between browsers: blog.stevenlevithan./archives/cross-browser-split - not sure if any of the items listed there apply here. – Matthew Wilson Commented Apr 4, 2011 at 9:43
4 Answers
Reset to default 9The reason it behaves differently is because of the subexpression capture you have using parenthesis. Other browsers add the match inside these captures to the resulting array, IE 8 and lower do not. To get a more consistent result, you'd have to make the group non-capturing:
/(?:<body>|<\/body>)/ig
This is the reason other browsers have the content in [2] rather than [1] — [1] will, in theory, contain the string "<body>". The other browsers have it right on this one and Internet Explorer 9 fixed the problem by implementing the method as outlined by the ECMAScript 5th Edition specification.
There are more inconsistencies than this, though. ECMAScript 5 pliance in all browsers will resolve these differences, but you might want to take a look at Steven Levithan's blog, where he outlines the differing implementations and even provides a custom split() method as a solution to the problem.
Have you considered just using xhr.responseXML.body.innerHTML the DOM is a lot better at parsing HTML than regexes
The following page lists differences in the 'split' implementation between browsers: http://blog.stevenlevithan./archives/cross-browser-split
You can do something like this :
var body_content;
var isIE = ( (ua.indexOf("msie") != -1) && (ua.indexOf("opera") == -1) && (ua.indexOf("webtv") == -1) );
var results = data.split(/(<body>|<\/body>)/ig);
if (isIE) {
body_content = results[1];
} else {
body_content = results[2];
}