an object by reference, won’t allow me to swap the reference to a new object

Question

Ok, I will try to explain my problem as best as I can, from the code snippet below I’m passing tempURLTestedVsCaptured, to my method CheckForDuplicates_capturedUrls.

This method will check if there are any duplicates, if there are not, it will add the URL (contained within my URLs object) to a new URLs object. Then once done, it will set the original tempURLs to the reference of the new object.

The problem I have tempURLTestedVsCaptured is not getting the new reference. If I watch the tempURLs, it has the correct value at the end of the method, when it jumps out back to the Crawl method, the tempURLTestedVsCaptured has returned to the original value.

If I change tempURLs, for example adding a URL to it, the changes are reflected.

If I do:

tempURLs = new URLs();
tempURLs = processedURLs;

It won’t pick up the change. I’m clearly missing something very fundermental here in my learning, but I can’t put my finger on it.

private void CheckForDuplicates_capturedUrls(URLs tempURLs)
        {
            URLs unprocessedURLs = (URLs)tempURLs;
            URLs processedURLs = new URLs();

                foreach (URL url in unprocessedURLs)
                {
                    if (!crawlContext.capturedUrls.ContainsURL(url))
                    {
                        processedURLs.AddURL(url);
                    }


                }
                    tempURLs = new URLs();
                    tempURLs = processedURLs;  
        }


        private void Crawl(WebScraper_Context crawlContext)
        {
            URLs tempURLTestedVsVisited = new URLs();
            URLs tempURLTestedVsCaptured = new URLs();

            while (crawlContext.unVistedURLs.Count() != 0) //While we have URLS we not visited, continue
            {
                foreach (URL url in crawlContext.unVistedURLs)
                {
                    //  If we not visted the page yet
                    if (!crawlContext.vistedURLs.ContainsURL(url))      //  Vist the URL if there is one
                    {
                        crawlContext.vistedURLs.AddURL(url);
                        LoadPage(url.url);                  
                        doc = GetSubSetXPath(doc, crawlContext.xPath);         
                    }                                                                       
                    if (doc != null)
                    {
                        crawlContext.scrapedUrls = ScrapeURLS();                                        
                        crawlContext.scrapedUrls = GetLocalUrls(crawlContext.scrapedUrls);              

                        // Cache the URLS into, so we can check if we seen them before
                        foreach (URL newURL in crawlContext.scrapedUrls)
                        {
                            if (!tempURLTestedVsVisited.ContainsURL(newURL))
                            {
                                tempURLTestedVsVisited.AddURL(newURL);
                                tempURLTestedVsCaptured.AddURL(newURL);
                            }
                            else
                            {
                                System.Windows.Forms.MessageBox.Show("Duplicate URL found in scraped URLS");
                            }
                        }


                        **this.CheckForDuplicates_capturedUrls(tempURLTestedVsCaptured);**



                        foreach (URL newURL in crawlContext.scrapedUrls)
                        {
                            if (tempURLTestedVsVisited.ContainsURL(newURL) && tempURLTestedVsCaptured.ContainsURL(newURL))
                            {
                                crawlContext.newURLs.AddURL(newURL);
                                crawlContext.capturedUrls.AddURL(newURL);

                            }
                        }



                    }
                }

                crawlContext.unVistedURLs = new URLs();                                                  crawlContext.unVistedURLs = crawlContext.newURLs;
                crawlContext.newURLs = new URLs();

            }
            if (RequestStop == true)
            {
                RequestStop = false;
            }
            System.Windows.Forms.MessageBox.Show("Complete");
        }

Ok T. Kiley completely explains my problem and why I’m getting it. The reason I’m not returning URLS, and I’m doing a pointless cast, is because the method signature is planned to be:

private void CheckForDuplicates_capturedUrls(object tempURLs).

The method is going to be used as a thread start “DuplicateCheckerB = new Thread(this.CheckForDuplicates_capturedUrls);” and “DuplicateCheckerA.Start(tempURLTestedVsVisited);” I originally thought my problem was down to threading, so I stripped it in in process of debugging.

Now, would I be right in thinking that I have to modify the actual object to remove the URLs if I am going to pass it to thread?

I'd check out by reference & by value parameters so you understand *why* this happens. — Arran, Dec 03 '13 at 22:12
Your are not pass by reference http://stackoverflow.com/questions/8708632/passing-objects-by-reference-or-value-in-c-sharp — Bit, Dec 03 '13 at 22:15

Sam Axe · Answer 1 · 2013-12-04T03:38:27.043

3

You are not passing your method parameter by reference:

private void CheckForDuplicates_capturedUrls(ref URLs tempURLs) {...}

and call it thusly:

CheckForDuplicates_capturedUrls(ref tempURLTestedVsCaptured);

Alternatively, and preferably, just return a new list:

private URLs CheckForDuplicates_capturedUrls(URLs tempURLs) {
    URLs result = new URLs();
    // process tempURLs, storing the result
    return result;
}

and use it like so:

tempURLTestedVsCaptured = CheckForDuplicates_capturedUrls(tempURLTestedVsCaptured);

edited Dec 04 '13 at 03:38

answered Dec 03 '13 at 22:11

Sam Axe

33,313
9
55
89

1

This will work, but a better solution is to return the new list. – Joel Coehoorn Dec 03 '13 at 22:17
@JoelCoehoorn: agreed, but that wasn't the question. – Sam Axe Dec 03 '13 at 22:18
2

@Dan-o The question wasn't asking how to pass a parameter by reference though either; technically it doesn't ask a question at all. It just describes a desired result and some non-working code, both solutions are technically answers, although yours is not particularly idiomatic in C#, unlike Joel's. – Servy Dec 03 '13 at 22:20
@Servy: I'll buy that. – Sam Axe Dec 04 '13 at 03:35

score 2 · Accepted Answer · answered Dec 03 '13 at 22:16

There are a few things wrong with your code, but taking your issue with your first method you first need to understand the difference between pass by reference and pass by value. This concept is more confusing in languages like C# and Java (than C++ [imo]) since it looks like it is passing by reference when in fact, it (almost) always passes by value.

Basically, when a function passes by value, the value of the parameter is copied to the function. Clearly, if you do this, any modification made will be lost.

Pass by reference, on the other hand, simply tells the function where to look, so modifications will be kept.

C# is pass by value, but the value you are passing is a pointer

What this means in practise is, you pass in your pointer to your tempUrls. Were you to modify the thing this pointer is pointing at, you would modify the thing you passed in.

However, what you do in your method:

tempURLs = processedURLs;

Points tempUrls to a new thing. We passed the address in by value, so changing the address won't do anything. The caller is still looking in the old location.

You can get around this using the ref keyword but in this method it would probably be better to return the new tempUrls

Other issues Your code as a few other problems:

No need to cast if the types match
When you call new() on tempUrls, you point it to some new memory. The line below, you point it to processedUrls, so the new is completely redundant. You only need to use new if you actually want a new object

Ok this completely explains my problem and why I’m getting it. — Gregory William Bryant, Dec 03 '13 at 22:33

Ufuk Hacıoğulları · Answer 3 · 2013-12-03T22:20:59.147

1

In C# parameters are by passed by value. That means you get a copy of the reference in the method you are calling. Any changes you make to that copy won't effect the other reference which lives outside of that method call. By using ref keyword for the parameter you can make changes to the original reference but don't do that. Returning new value from the method is the better way.

private URLs CheckForDuplicates_capturedUrls(URLs tempURLs) {...}

edited Dec 03 '13 at 22:20

answered Dec 03 '13 at 22:15

Ufuk Hacıoğulları

37,978
12
114
156

+1 for mentioning ownership of the reference and returning a new object from the method instead. – Simon Whitehead Dec 03 '13 at 22:17

Joel Coehoorn · Answer 4 · 2013-12-03T22:23:52.593

You could fix this with a reference parameter, but a better design is to return your new list, following this pattern:

tempUrls = CheckForDuplicates_capturedUrls(URLs tempURLs)

.

private URLs CheckForDuplicates_capturedUrls(URLs tempURLs)
{
    //It's not clear what options your URLs type has, but based on
    // your example use, it looks like you might be inheriting List<url>
    // or implementing IList. This code makes that assumption

    return tempURLs.Where(url => !crawlContext.capturedUrls.ContainsURL(url)).ToList();
}

We even get the code down to a one-liner :)

The original code did not work because .Net passes references to functions by value. This means the function gets a reference that refers to the same object, but is still just a copy. The function can make changes to properties in an object, or call it's methods, and those changes will be evident to the caller of the function. However, if the function tries to make an assignment directly to the variable itself, it will only change the copy.

score -3 · Answer 5 · answered Dec 03 '13 at 22:14

-3

In C# objects are passed by reference. However when you provide a method an object it makes a copy of the pointer (a reference is a safe pointer) then provides it to the function.

Add 'ref' to the method call and you will pass the original reference.

private void CheckForDuplicates_capturedUrls(ref URLs tempURLs)

answered Dec 03 '13 at 22:14

Shravan Sunder

31
4
7

2

"In C# objects are passed by reference". No they aren't. The reference is copied.. which makes it pass-by-value.. not pass-by-reference. – Simon Whitehead Dec 03 '13 at 22:15

an object by reference, won’t allow me to swap the reference to a new object

5 Answers5