code sexiness question

[honey the codewitch]

The "extra" variable doesn't bother me. I don't think there's much difference in the cognitive load required to understand each function. Given that the first one passed code review, it's "sexier", IMO. Edit: I have to add, is the extra time you're spending trying to decide which version is more readable adding the requisite amount of value? *hides*

Real programmers use butterflies

[Daniele Rota Nodari]

Storing the value of the property (SelectedObject) into a local variable (objects) is like a snapshot. It ensures that any subsequent instruction refers to the same object. This is important when you need consistency throughout the method, the method can be relatively time consuming and some other thread can concurrently change the property value. You can always rely on compiler optimizations and hope that the compiled machine code will take care of such thing, or you can do it by yourself with the local variable. Another reason could be that you know in advance that someone will add code inside the method to purposely change the object value; such change will require to put the object into a variable, like the collegue told you to do. Doing it later will require to replace any reference to the property with references to the variable: those changes would be spread along the method polluting versioning differences. It can even be of help if copy-paste is performed to some other method where the logic must be kept but the property to process has a different name. These are indeed pre-optimizations, and we can argue about their usefullness and their development cost. Additionally, the name of the property is not totally meaningful, due to the fact it is singular but it can store multiple objects (like the name of the variable unfolds): this could be an important hint for the future-you maintaining the code.

[HobbyProggy]

If you can't change the rules, break the rules XD Nah srsly, i tend to go on change it, love it or leave it mentality. This solves many frustration related issues for me so far.

Rules for the FOSW ![^]

MessageBox.Show(!string.IsNullOrWhiteSpace(_signature)
? "This is my signature: " + Environment.NewLine + _signature
: "404-Signature not found");

[Richard Deeming]

Just for giggles, how about:

private IMultipleComponentHandler SelectionHandler => m_selectionHandler ??= SelectedObject switch
{
null => new InspectorMultipleComponentHandler(Array.Empty<object>()),
IMultipleComponentHandler handler => handler,
object[] e when !e.GetAttributes<IgnoreIEnumerableAttribute>.Any() => e,
IEnumerable e when !e.GetAttributes<IgnoreIEnumerableAttribute>.Any() => e.Cast<object>().ToArray(),
var e => new InspectorMultipleComponentHandler(new[] { e }),
};

As with your first example, this only accesses SelectedObject once.

---

"These people looked deep within my soul and assigned me a number based on the order in which I joined." - Homer

[Super Lloyd]

we're still stuck on .NET 4.7.2 at the moment.... :(( not for long I heard .NET 6 is coming, like the winter!

A new .NET Serializer All in one Menu-Ribbon Bar Taking over the world since 1371!

[Daniel Pfeffer]

Whiskey. Tango. Foxtrot. :omg:

Freedom is the freedom to say that two plus two make four. If that is granted, all else follows. -- 6079 Smith W.

[Super Lloyd]

I am just fed up with all those rubbing me wrong micro management useless comments... I try to just shrug it off... But it annoys me every time some (of those particular) guys reviews... but on the other hand getting any review at all is also hard work, so bloody annoying...

A new .NET Serializer All in one Menu-Ribbon Bar Taking over the world since 1371!

[Richard Deeming]

If you're using VS2019 or 2022, you can still use that construct in .NET 4.7.2; you just need to manually edit your project file to enable C# 9. :) If you already have a <LangVersion> element in the file, change it to <LangVersion>9.0</LangVersion>. Otherwise, add that element next to the <TargetFramework> element. Quite a few C# 8/9/10 features will work in .NET Framework projects: Using C# 9 outside .NET 5 · Discussion #47701 · dotnet/roslyn · GitHub[^]

---

"These people looked deep within my soul and assigned me a number based on the order in which I joined." - Homer

[Richard Deeming]

switch expression - C# reference | Microsoft Docs[^] Pattern matching overview - C# guide | Microsoft Docs[^] :)

---

"These people looked deep within my soul and assigned me a number based on the order in which I joined." - Homer

[OriginalGriff]

I'd use the first version - that way if SelectedObject is changed (by another thread for example) the non-null value is preserved and the app doesn't crash. It's the way I handle event raising - my standard template code is:

    /// /// Event to indicate Description
    /// 
    public event EventHandler Name;
    /// /// Called to signal to subscribers that Description
    /// 
    /// 
    protected virtual void OnName(EventArgs e)
        {
        EventHandler eh = Name;
        if (eh != null)
            {
            eh(this, e);
            }
        }

That way, in the (unlikely) event that the last handler is removed from the c=hain, the app doesn't crash and does something sensible.

"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony "Common sense is so rare these days, it should be classified as a super power" - Random T-shirt AntiTwitter: @DalekDave is now a follower!

[Super Lloyd]

Interesting... Though in our case there is some sort of build system, which is still mysterious to me, that generate the .csproj files.. so I would need to get familiar with that first! :laugh: In fact... I really ought to become more familiar with this particular system.... :sigh:

A new .NET Serializer All in one Menu-Ribbon Bar Taking over the world since 1371!

[Richard Deeming]

Two obvious alternatives to that:

public event EventHandler Name = delegate { };

protected virtual void OnName(EventArgs e)
{
Name(this, e); // Name can never be null
}

public event EventHandler Name;

protected virtual void OnName(EventArgs e)
{
Name?.Invoke(this, e);
}

---

"These people looked deep within my soul and assigned me a number based on the order in which I joined." - Homer

[honey the codewitch]

Understandable. I can't stand code reviews myself, because I have some different philosophies about how code needs to be written than a lot of people I have worked with. However, when I am in a position where I am in charge of code reviews, I tend to go easy and stick to enforcing in-shop style guidelines more than anything. I don't care about fast for bizdev unless something is slow enough you want to get out and push. I would have accepted either version of your code. I think both are readable *enough* - and this is one of the areas where I differ with a lot of people. I don't spend as much time chasing readability as other coders. I like to look at cognitive load more than readability, because I feel like readability can be had by reading the complicated parts of a function more than once. The trick is in *understanding* what you've read. That's the part where I care, but also the part I'm not great at. One of the reasons I write here is to try to improve my skillset in terms of making my code understandable. My functions are too long, but that's due to some cognitive issues I have myself, and it's part of how I've adapted to them.

Real programmers use butterflies

[OriginalGriff]

The first relies on the empty delegate: if someone else sees the code and removes it as it clearly does nothing then you are back to a potential failure. Unlikely, yes - but I don't like app failures. :D The second is .NET version dependant: the null conditional operator was introduced at C# 6, and some of my code predates that.

"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony "Common sense is so rare these days, it should be classified as a super power" - Random T-shirt AntiTwitter: @DalekDave is now a follower!

[11917640 Member]

Very good good code review result. Absolutely meaningless and harmless change. You and code reviewer, and of course, managers, are happy. BTW, after native optimizations both versions may be absolutely identical. But don't tell about this to your manager.

[snorkie]

Are we coding in C# or Perl? I always felt like C# was a nice balance between COBOL and Perl. But the recent changes have it trending towards Perl. I hear the excuse "it saves typing" as if we aren't on a message board or slack typing all day long.

Hogan

[Lost User]

I have problems with both. "var" is a coding shortcut when you're too lazy to type, and should be replaced with the actual type when it resolves (before release); so the next guy doesn't have to "intelli-sense" it. "Object" seems more appropriate in this case. I dislike "long" if's and would have "if not null return x" instead of (if null etc.) Even though it works, I don't declare properties "after" the method that references them. The return "assign" is "different". Then there are the if's with bracketed blocks and some without. Saving keystrokes? etc.

It was only in wine that he laid down no limit for himself, but he did not allow himself to be confused by it. ― Confucian Analects: Rules of Confucius about his food

[Gary R Wheeler]

I would have done things this way:

private IMultipleComponentHandler SelectionHandler
{
get
{
if (m_selectionHandler == null)
{
if (SelectedObject is IMultipleComponentHandler handler)
{
m_selectionHandler = handler;
}

        else
        {
            object\[\] collection;
            
            if (SelectedObject is IEnumerable e
                && !SelectedObject.GetAttributes().Any())
            {
                collection = e as object\[\] ?? e.Cast().ToArray();
            }
            else if (SelectedObject != null)
            {
                collection = new\[\] { SelectedObject };
            }
            else
            {
                collection = Array.Empty();
            }

            m\_selectionHandler = new InspectorMultipleComponentHandler(collection);
        }
    }
    
    return m\_selectionHandler;
}

}
private IMultipleComponentHandler m_selectionHandler;I'm an old fart who prefers single-exit :-D .

Software Zen: delete this;

[Fueled By Decaff]

I would be tempted to extract the contents of the if statement as a function, like this:

    public IMultipleComponentHandler SelectionHandler
    {
        get
        {
            if (m\_selectionHandler == null)
            {
                m\_selectionHandler = InitialiseSelectionHandler(SelectedObject);
            }
            return m\_selectionHandler;
        }
    }

    private IMultipleComponentHandler InitialiseSelectionHandler(object selectedObject)
    {
        if (selectedObject is IMultipleComponentHandler handler)
            return m\_selectionHandler = handler;

        object\[\] collection;
        if (selectedObject is IEnumerable e
            && !selectedObject.GetAttributes().Any())
        {
            collection = e as object\[\] ?? e.Cast().ToArray();
        }
        else if (selectedObject != null)
        {
            collection = new\[\] { selectedObject };
        }
        else
        {
            collection = Array.Empty();
        }
        return new InspectorMultipleComponentHandler(collection);
    }

    private IMultipleComponentHandler m\_selectionHandler;

This hides the issue of using a temporary variable, snapshots the SelectedObject so the value being used can not change while updating the handler and the getter property for SelectionHandler becomes easier to read.

An added bonus is if SelectedObject changes you can update the handler by calling InitialiseSelectionHandler.

[Daniele Rota Nodari]

This is way cleaner.. It is a sort of lazy initialization.. and usually it is split exactly that way: a backing field, a property that lazily initializes the field, and a method that performs the initialization. This also highlight a possible race condition after the very first null check, that can be avoided by a double-checked locking inside the property get accessor; and the InitialiseSelectionHandler method would remain untouched. I'd just keep the backing field close to the property (I usually keep the field immediately before the property).