Find all string values from complex object using Reflection

Question

I am having a few issues finding a simple way to find all string values using Reflection on an object.

I need to scan an object recursively including any lists, arrays for a string that matches a value. So far I am trying to return a list of all the values then I can check that list, however I have run into a problem where this can throw a StackOverflowException if the object contains any circular dependencies.

What I have so far: (I know, I know, but its a work in progress :))

private IEnumerable<string> FindStringValues(object obj)
{
    if (obj != null)
    {
        Type type = obj.GetType();
        if (type == typeof(string))
        {
            yield return (string)obj;
        }
        else if (type.IsArray)
        {
            var array = obj as Array;
            foreach (var item in array)
            {
                foreach (var str in FindStringValues(item))
                {
                    yield return str;
                }
            }
        }
        else if (type.IsClass)
        {
            FieldInfo[] fields = type.GetFields(BindingFlags.Public | BindingFlags.NonPublic | BindingFlags.Instance);
            foreach (FieldInfo field in fields)
            {
                object fieldValue = field.GetValue(obj);
                if (fieldValue == null)
                    continue;

                foreach (var str in FindStringValues(fieldValue))
                {
                    yield return str;
                }
            }
        }
    }
}

So I'm trying to find a way to make this StackOverflowException safe, even better if anyone can help me figure out how to make this a bool to return if a match is found with a supplied value.

---

EDIT:

Thanks to the answer from nsinreal I was able to accomplish the goal :) added below incase its helpful to others

public static bool FindStringValue(object obj, string valueToFind)
{
    return FindStringValue(obj, valueToFind, new List<object>());
}

private static bool FindStringValue(object obj, string valueToFind, IList<object> visitedObjects)
{
    if (obj == null && !visitedObjects.Any(item => Object.ReferenceEquals(item, obj)))
    {
        if (!(obj is string))
        {
            visitedObjects.Add(obj);
        }

        Type type = obj.GetType();
        if (type == typeof(string))
        {
            return (obj).ToString() == valueToFind;
        }

        if (typeof(IEnumerable).IsAssignableFrom(type))
        {
            var array = obj as IEnumerable;
            foreach (var item in array)
            {
                if (FindStringValue(item, valueToFind, visitedObjects))
                {
                    return true;
                }
            }
            return false;
        }

        if (type.IsClass)
        {
            FieldInfo[] fields = type.GetFields(BindingFlags.Public | BindingFlags.NonPublic | BindingFlags.Instance);
            foreach (FieldInfo field in fields)
            {
                object item = field.GetValue(obj);
                if (item == null)
                    continue;

                if (FindStringValue(item, valueToFind, visitedObjects))
                {
                    return true;
                }
            }
        }
    }
    return false;
}

Answer 1

1. Any lists, any array -> any IEnumerable. So you can replace:

   if (type.IsArray)
   

   To:

   if (typeof(IEnumerable).IsAssignableFrom(type))
   

2. To prevent StackOverflowException just don't visit visited objects. Do you like idea? So, firstly we must choose: we don't want visit all visited objects or only parent objects. In this case, we won't visit all visited objects. So, modifcate header:

   // visitedObjects can changes
   private static IEnumerable<string> _FindStringValues(object obj, IList<object> visitedObjects)
   

   Now, modificate all calls of FindStringValues to call of _FindStringValues in this method:

   _FindStringValues(item, visitedObjects)
   

   And add usable wrapper-method:

   private static IEnumerable<string> FindStringValues(object obj) {
       return _FindStringValues(obj, new List<object>());
   }
   

   In _FindStringValues you must add each obj passed into this method to visited objects list.

   visitedObjects.Add(obj);
   

   But before that you must check that this object wasn't visited. You must use Object.ReferenceEquals:

   if (visitedObjects.Any(item => Object.ReferenceEquals(item, obj)))
       yield break;
   

   Be carefully, if you want to see equal strings from diffrent objects you must prevent adding strings to visited objects list. Why? String interning. Like this:

   if (!(obj is string))
       visitedObjects.Add(obj);
   

After some code formating I have got this: code.