I'm trying to integrate analytics into my GWT application. To do this, I'm calling a service that returns a String of HTML that needs to be parsed and eval'ed.
I need a regex that looks for and grabs either 1) the body of the tag or 2) the contents of the "src" attribute. I want to eval both of these with JavaScript. I'm happy with assuming that if a "src" attribute exists, the body can be ignored.
Thanks,
Matt
To match the body of the tag, you can try something like
<script[^>]*?>(.*?)</script>
which you want to match case-insensitively. Works assuming there is no "" appearing in the actual script body and no ">" in the attributes for the tag. You can add whitespace globbers to the regexp to make it more robust. Note the use of .*? to make sure the scanning stops at the first closing tag.
To add the src attribute, you can try
<script[^>]*?(src="([^"]*)")?[^>]*?>(.*?)</script>
and use the second submatch to get 'src', and third to get the body. Again, you might want to add whitespace globbers.
But would be best off by running the thing through a proper HTML/XML/SGML parser, because regexps can blow up in special cases.
How about
<script>(.*)</script>|<script src="(.*)">.*</script>
to start with. You may need to customize it a bit to
You also must use the DOTALL mode to ensure the . captures newlines.
Must it be a regex? You can use the DOM to obtain such information, here is a trivial example of getting the contents of the BODY tag, you could apply it to whatever you like:
function test(){
var body = document.getElementsByTagName("body")[0];
alert(body.innerHTML);
}
This seems to do what you want:
final String srcOne = "<html>\r\n<head>\r\n<script src=\"http://test.com/some.js\"/>\r\n</head></html>";
final String srcTwo = "<html>\r\n<head>\r\n<script src=\"http://test.com/some.js\"></script>\r\n</head></html>";
final String tag = "<html>\r\n<head>\r\n<script>\r\nfunction() {\r\n\talert('hi');\r\n}\r\n</script>\r\n</head></html>";
final String tagAndSrc = "<html>\r\n<head>\r\n<script src=\"http://test.com/some.js\">\r\nfunction() {\r\n\talert('hi');\r\n}\r\n</script>\r\n</head></html>";
final String[] tests = new String[] {srcOne, srcTwo, tag, tagAndSrc, srcOne + srcTwo, tag + srcOne + tagAndSrc};
final String regex = "<script(?:[^>]*src=['\"]([^'\"]*)['\"][^>]*>|[^>]*>([^<]*)</script>)";
final Pattern pattern = Pattern.compile(regex, Pattern.CASE_INSENSITIVE | Pattern.DOTALL);
for (int testNumber = 0; testNumber < tests.length; ++testNumber) {
final String test = tests[testNumber];
final Matcher matcher = pattern.matcher(test);
System.out.println("--------------------------------");
System.out.println("TEST " + testNumber + ": " + test);
while (matcher.find()) {
System.out.println("GROUP 1: " + matcher.group(1));
System.out.println("GROUP 2: " + matcher.group(2));
}
System.out.println("--------------------------------");
System.out.println();
}
That being said, you would probably be better off using something like Tag Soup if it is at all possible.
Thanks for all the great suggestions everyone. I quickly discovered it's not possible to use Java's Regex API in GWT and was able to do what I wanted with JSNI.
public static native String evalJS(Element e) /*-{
var scripts = e.getElementsByTagName("script");
for (i=0; i < scripts.length; i++) {
// if src, eval it, otherwise eval the body
if (scripts[i].hasAttribute("src")) {
eval(scripts[i].getAttribute("src")); // silently fails here
} else {
eval(scripts[i].innerHTML); // this works
}
}
}-*/;
Unfortunately, I ran into additional issues as documented in the following thread:
http://groups.google.com/group/Google-Web-Toolkit/browse_thread/thread/ac2589369ddec8a3