regex pattern and matcher with runtime input and some exception cases

Question

In java file I read jsp file and trying to finding out number of css clasess used, "class=" and its value by using regex below.

Pattern p = Pattern.compile("class=\"([^\"]*)\"");
Set set = new HashSet();
Iterator iterator;
while ((strLine = br.readLine()) != null)
{
    Matcher m = p.matcher(strLine);
}
while (m.find())
{
    String classValue = m.group(1);
    set.add(classValue);
}

it gives me class name means if jsp contents class="List" or class="listItem".

Output is { List listItem } like this. My problem is as follows if my JSP contents

1. then it will shows com.metaparadigm.jsonrpc.JSONRPCBridge that i dont want
2. "> in this it will give me output = "<%=w_canEdit?" but i want only one class either IconSpacing or IconDisable how can do this

Answer 1

Question

From your encrypted description assuming I have deciphered it correctly !

Looks to me that your jsp page contains following line

<img src="a.jpeg" class="<%=w_canEdit?"IconSpacing":"IconDisable"%>"/>

And your regular expression matches <%=w_canEdit?\

@Test
public void testRegex() {

    Pattern p = Pattern.compile("class=\"([^\"]*)\"");
    Set set = new HashSet();


    //<img class="<%=w_canEdit?"IconSpacing":"IconDisable"%>" src="a.jpeg"/>
    String str="<img src=\"a.jpeg\" class=\"<%=w_canEdit?\"IconSpacing\":\"IconDisable\"%>\"/>";
    System.out.println(str);


    Matcher m = p.matcher(str);
    while (m.find())
    {
        String classValue = m.group(1);
        set.add(classValue);
    }
            System.out.println("Result:");
    System.out.println(set);
}

Output

Input:
<img src="a.jpeg" class="<%=w_canEdit?"IconSpacing":"IconDisable"%>"/>
Result:
[<%=w_canEdit?]

what you expect in the result

[IconSpacing,IconDisable]

---

Answer

Short answer :

you can not do it with regex

Long answer :

you can not do it with regex, even if with lookahead hacks you might be able to resolve it as <%=w_canEdit?"IconSpacing":"IconDisable"%> like for ex using following pattern

Pattern p = Pattern.compile("class=\"(<%=(.(?<!%>\"))*)\"");
// [<%=w_canEdit?"IconSpacing":"IconDisable"%>]

you will still never be able to identify runtime value of a class [as either IconSpacing or IconDisable ] by parsing jsp file anyways.

Easiest way to do this is to to do it manually

1. grep class= *.jsp
2. identify css classes that have jsp scriptlets in them
3. extract the required details from the result

If you can raise a separate Question with exact details of your requirement people here would be definitely happy to help

---

Also See this post RegEx match open tags except XHTML self-contained tags to understand why using regex to parse html/jsp pages is not that great idea !