I have a Java String that contains javascript code, and i need to extract all the javascript vars' names.
So, for the following javasctipt:
var x;
var a,b,c,d;
var y = "wow";
var z = y + 'x';
I need to get "x,a,b,c,d,y,z" as a result.
I dont need to get their values, just their names.
Well you can try and get the bindings that the execution of the script creates:
ScriptEngineManager mgr = new ScriptEngineManager();
ScriptEngine se = mgr.getEngineByName("JavaScript");
try {
se.eval("var x;var a,b,c,d;var y = \"wow\";var z = y+'x';");
Bindings bindings = se.getBindings(ScriptContext.ENGINE_SCOPE);
System.out.println(bindings.keySet());
}
catch (ScriptException e) {
//e.printStackTrace();
}
this prints [d, b, c, println, a, context, z, y, print, x]
as you see some extra bindings are defined: context, print and println
and here we filter them out
Set<String> keySet = bindings.keySet();
keySet.removeAll(Arrays.asList("context", "print", "println"));
System.out.println(keySet);
this prints [d, b, c, a, z, y, x]
Something like the following:
List<String> vars = new ArrayList<String>();
Pattern p = Pattern.compile("^\\s*var\\s+(.*?)\\s*;?$");
BifferedReader reader = .... // create the reader
String line = null;
while ((line = reader.readLine()) != null) {
Matcher m = p.matcher(line);
if (m.find()) {
vars.addAll(Arrays.asList(m.group(1).split("\\s*,\\s*")));
}
}
Please note that I wrote this code without IDE and have never compiled and ran it. So, sorry for possible bugs but I thing it is readable enough and can be a good starting point for you.
I'm not sure whether this would be too convoluted, but I would probably get the BNF for Javascript (see SO: Repository of BNF Grammars?) and then use something like ANTLR to create a parser for Javascript, which you can then use to extract all the variables based on the Javascript source code. See also SO: ANTLR Tutorials.
