This is quirk of the decompiler you use to look at a .NET assembly.  It is the behavior of ildasm.exe, other ones like Reflector or ILSpy get this right.  The Microsoft programmer who wrote it took a shortcut, he generates a string from the IL that just displays the type argument the way it is encoded, without writing the *extra* code to lookup the type argument name in the metadata.  

You need to read `!n` as the n-th type argument of the generic type.  Where !0 means &quot;first type argument&quot;, !1 means &quot;second type argument&quot;, etcetera.  For Nullable&lt;&gt;, you know that &#39;!0` means &#39;T&#39; from the MSDN article.

You may also encounter something like `!!T`.  Two exclamation marks indicate a type argument for a generic *method*.  This time, ildasm.exe *does* look up the type argument name instead of using `!!0`.  Why the programmer took the shortcut on generic types but not on generic methods is hard to reverse-engineer.  Ildasm is a pretty quirky program and is written in a C++ coding style that&#39;s very different from other C++ code in .NET.  Not as disciplined, non-zero odds that this was an intern&#39;s assignment :)

The \`1 suffix on &quot;Nullable&quot; is a normal encoding for generic type names, it indicates the generic type has one type argument.  In other words, for Nullable&lt;&gt; you&#39;ll never see !1 being used.

So simply read `!0` as &quot;T&quot;.  Or use a better decompiler.

That is a generic type parameter. 

They are positional.

Decompile some generic code to see how they are used (compare IL vs C#).

I am following this example

http://www.androidhive.info/2015/04/android-getting-started-with-material-design/

and in this example it is showing hamburger icon white,i want to customize it and make it black,but i am not able to find anything to how to change it,can any one tell how to customize it?



Manifest

    &lt;manifest xmlns:android=&quot;http://schemas.android.com/apk/res/android&quot;
        package=&quot;info.androidhive.materialdesign&quot; &gt;
    
    
        &lt;uses-permission android:name=&quot;android.permission.INTERNET&quot;&gt;&lt;/uses-permission&gt;
    
        &lt;uses-permission android:name=&quot;android.permission.ACCESS_NETWORK_STATE&quot;&gt;&lt;/uses-permission&gt;
        &lt;application
            android:allowBackup=&quot;true&quot;
            android:icon=&quot;@mipmap/ic_launcher&quot;
            android:label=&quot;@string/app_name&quot;
            android:theme=&quot;@style/MyMaterialTheme&quot; &gt;
            &lt;activity
                android:name=&quot;.activity.MainActivity&quot;
                android:label=&quot;@string/app_name&quot; &gt;
                &lt;intent-filter&gt;
                    &lt;action android:name=&quot;android.intent.action.MAIN&quot; /&gt;
    
                    &lt;category android:name=&quot;android.intent.category.LAUNCHER&quot; /&gt;
                &lt;/intent-filter&gt;
            &lt;/activity&gt;
        &lt;/application&gt;
    
    &lt;/manifest&gt;

style

    &lt;resources&gt;
    
        &lt;style name=&quot;MyMaterialTheme&quot; parent=&quot;MyMaterialTheme.Base&quot;&gt;
    
        &lt;/style&gt;
    
        &lt;style name=&quot;MyMaterialTheme.Base&quot; parent=&quot;Theme.AppCompat.Light.DarkActionBar&quot;&gt;
            &lt;item name=&quot;windowNoTitle&quot;&gt;true&lt;/item&gt;
            &lt;item name=&quot;windowActionBar&quot;&gt;false&lt;/item&gt;
            &lt;item name=&quot;colorPrimary&quot;&gt;@color/colorPrimary&lt;/item&gt;
            &lt;item name=&quot;colorPrimaryDark&quot;&gt;@color/colorPrimaryDark&lt;/item&gt;
            &lt;item name=&quot;colorAccent&quot;&gt;@color/colorAccent&lt;/item&gt;
            &lt;item name=&quot;homeAsUpIndicator&quot;&gt;@drawable/hamburger&lt;/item&gt;
        &lt;/style&gt;
    
    &lt;/resources&gt;

MainActivity

    public class MainActivity extends AppCompatActivity implements FragmentDrawer.FragmentDrawerListener {
    
        private static String TAG = MainActivity.class.getSimpleName();
    
        private Toolbar mToolbar;
        private FragmentDrawer drawerFragment;
    
        @Override
        protected void onCreate(Bundle savedInstanceState) {
            super.onCreate(savedInstanceState);
            setContentView(R.layout.activity_main);
           mToolbar = (Toolbar) findViewById(R.id.toolbar);
    
            setSupportActionBar(mToolbar);
            getSupportActionBar().setDisplayShowHomeEnabled(true);
    
            drawerFragment = (FragmentDrawer)
                    getSupportFragmentManager().findFragmentById(R.id.fragment_navigation_drawer);
            drawerFragment.setUp(R.id.fragment_navigation_drawer, (DrawerLayout) findViewById(R.id.drawer_layout), mToolbar);
            drawerFragment.setDrawerListener(this);
    
            // display the first navigation drawer view on app launch
            displayView(0);
        }
    
    
        @Override
        public boolean onCreateOptionsMenu(Menu menu) {
            // Inflate the menu; this adds items to the action bar if it is present.
            getMenuInflater().inflate(R.menu.menu_main, menu);
            return true;
        }
    
        @Override
        public boolean onOptionsItemSelected(MenuItem item) {
            // Handle action bar item clicks here. The action bar will
            // automatically handle clicks on the Home/Up button, so long
            // as you specify a parent activity in AndroidManifest.xml.
            int id = item.getItemId();
    
            //noinspection SimplifiableIfStatement
            if (id == R.id.action_settings) {
                return true;
            }
    
            if(id == R.id.action_search){
                Toast.makeText(getApplicationContext(), &quot;Search action is selected!&quot;, Toast.LENGTH_SHORT).show();
                return true;
            }
    
            return super.onOptionsItemSelected(item);
        }
    
        @Override
        public void onDrawerItemSelected(View view, int position) {
            displayView(position);
        }
    
        private void displayView(int position) {
            Fragment fragment = null;
            String title = getString(R.string.app_name);
            switch (position) {
                case 0:
                    fragment = new HomeFragment();
                    title = getString(R.string.title_home);
                    break;
                case 1:
                    fragment = new FriendsFragment();
                    title = getString(R.string.title_friends);
                    break;
                case 2:
                    fragment = new MessagesFragment();
                    title = getString(R.string.title_messages);
                    break;
                case 3:
                    fragment = new ContactUsFragment();
                    title = getString(R.string.title_contactus);
                    break;
                case 4:
                    fragment = new AboutUsFragment();
                    title = getString(R.string.title_aboutus);
                    break;
                default:
                    break;
            }
    
            if (fragment != null) {
                FragmentManager fragmentManager = getSupportFragmentManager();
                FragmentTransaction fragmentTransaction = fragmentManager.beginTransaction();
                fragmentTransaction.replace(R.id.container_body, fragment);
                fragmentTransaction.commit();
    
                // set the toolbar title
                getSupportActionBar().setTitle(title);
            }
        }



How to change color of hamburger icon in material design navigation drawer

I&#39;m practicing/studying both JavaScript and Python.  I&#39;m wondering if Javascript has the equivalence to this type of coding.

I&#39;m basically trying to get an array from each individual integer from the string for practice purposes.  I&#39;m more proficient in Python than JavaScript

Python:

    string = &#39;1234-5&#39;

    forbidden = &#39;-&#39;

    print([int(i) for i in str(string) if i not in forbidden])

Does Javascript have something similar for me to do above?

Does JavaScript support array/list comprehensions like Python?

In many MSIL listings, I have observed the following:

    System.Nullable`1&lt;!0&gt; etc ...
or

    class !0 etc ...

What&#39;s the meaning of `!0` in these circumstances?

Why is !0 a type in Microsoft Intermediate Language (MSIL)?

In many MSIL listings, I have observed the following:
<pre><code class="hljs">System.Nullable`1&#x3C;!0> etc ...
</code></pre>
or
<pre><code class="hljs language-arduino">class !0 etc ...
</code></pre>
What's the meaning of <code>!0</code> in these circumstances?

Environment: Visual Studio 2015 RTM. (I haven&#39;t tried older versions.)

Recently, I&#39;ve been debugging some of my [Noda Time][1] code, and I&#39;ve noticed that when I&#39;ve got a local variable of type `NodaTime.Instant` (one of the central `struct` types in Noda Time), the &quot;Locals&quot; and &quot;Watch&quot; windows don&#39;t appear to call its `ToString()` override. If I call `ToString()` explicitly in the watch window, I see the appropriate representation, but otherwise I just see:

    variableName       {NodaTime.Instant}

which isn&#39;t very useful.

If I change the override to return a constant string, the string *is* displayed in the debugger, so it&#39;s clearly able to pick up that it&#39;s there - it just doesn&#39;t want to use it in its &quot;normal&quot; state.

I decided to reproduce this locally in a little demo app, and here&#39;s what I&#39;ve come up with. (Note that in an early version of this post, `DemoStruct` was a class and `DemoClass` didn&#39;t exist at all - my fault, but it explains some comments which look odd now...)

    using System;
    using System.Diagnostics;
    using System.Threading;
    
    public struct DemoStruct
    {
        public string Name { get; }
    
        public DemoStruct(string name)
        {
            Name = name;
        }
    
        public override string ToString()
        {
            Thread.Sleep(1000); // Vary this to see different results
            return $&quot;Struct: {Name}&quot;;
        }
    }

    public class DemoClass
    {
        public string Name { get; }
    
        public DemoClass(string name)
        {
            Name = name;
        }
    
        public override string ToString()
        {
            Thread.Sleep(1000); // Vary this to see different results
            return $&quot;Class: {Name}&quot;;
        }
    }
    
    public class Program
    {
        static void Main()
        {
            var demoClass = new DemoClass(&quot;Foo&quot;);
            var demoStruct = new DemoStruct(&quot;Bar&quot;);
            Debugger.Break();
        }
    }

In the debugger, I now see:

    demoClass    {DemoClass}
    demoStruct   {Struct: Bar}

However, if I reduce the `Thread.Sleep` call down from 1 second to 900ms, there&#39;s still a short pause, but then I see `Class: Foo` as the value. It doesn&#39;t seem to matter how long the `Thread.Sleep` call is in `DemoStruct.ToString()`, it&#39;s always displayed properly - and the debugger displays the value before the sleep would have completed. (It&#39;s as if `Thread.Sleep` is disabled.)

Now `Instant.ToString()` in Noda Time does a fair amount of work, but it certainly doesn&#39;t take a whole second - so presumably there are more conditions that cause the debugger to give up evaluating a `ToString()` call. And of course it&#39;s a struct anyway.

I&#39;ve tried recursing to see whether it&#39;s a stack limit, but that appears not to be the case.

So, how can I work out what&#39;s stopping VS from fully evaluating `Instant.ToString()`? As noted below, `DebuggerDisplayAttribute` appears to help, but without knowing *why*, I&#39;m never going to be entirely confident in when I need it and when I don&#39;t.

**Update**

If I use [`DebuggerDisplayAttribute`][2], things change:

    // For the sample code in the question...
    [DebuggerDisplay(&quot;{ToString()}&quot;)]
    public class DemoClass

gives me:

    demoClass      Evaluation timed out

Whereas when I apply it in Noda Time:

    [DebuggerDisplay(&quot;{ToString()}&quot;)]
    public struct Instant
    
a simple test app shows me the right result:

    instant    &quot;1970-01-01T00:00:00Z&quot;

So presumably the problem in Noda Time is some condition that `DebuggerDisplayAttribute` *does* force through - even though it doesn&#39;t force through timeouts. (This would be in line with my expectation that `Instant.ToString` is easily fast enough to avoid a timeout.)

This *may* be a good enough solution - but I&#39;d still like to know what&#39;s going on, and whether I can change the code simply to avoid having to put the attribute on all the various value types in Noda Time.

**Curiouser and curiouser**

Whatever is confusing the debugger only confuses it sometimes. Let&#39;s create a class which *holds* an `Instant` and uses it for its own `ToString()` method:

    using NodaTime;
    using System.Diagnostics;

    public class InstantWrapper
    {
        private readonly Instant instant;

        public InstantWrapper(Instant instant)
        {
            this.instant = instant;
        }

        public override string ToString() =&gt; instant.ToString();
    }

    public class Program
    {
        static void Main()
        {
            var instant = NodaConstants.UnixEpoch;
            var wrapper = new InstantWrapper(instant);

            Debugger.Break();
        }
    }

Now I end up seeing:

    instant    {NodaTime.Instant}
    wrapper    {1970-01-01T00:00:00Z}

However, at the suggestion of Eren in comments, if I change `InstantWrapper` to be a struct, I get:

    instant    {NodaTime.Instant}
    wrapper    {InstantWrapper}

So it *can* evaluate `Instant.ToString()` - so long as that&#39;s invoked by another `ToString` method... which is within a class. The class/struct part seems to be important based on the type of the variable being displayed, not what code needs 
to be executed in order to get the result.

As another example of this, if we use:

    object boxed = NodaConstants.UnixEpoch;

... then it works fine, displaying the right value. Colour me confused.

  [1]: http://nodatime.org
  [2]: https://msdn.microsoft.com/en-us/library/system.diagnostics.debuggerdisplayattribute.debuggerdisplayattribute


What makes the Visual Studio debugger stop evaluating a ToString override?

I have installed VS 2015 RTM (nothing else) and I&#39;m unable to debug any solution, not matter if it&#39;s an existing one or a brand new one (created with VS 2015 and compiled against .Net Framework 4.6), it only opens a new tab in VS which is called Break Mode with the following text:
The application is in break mode
Your app has entered a break state, but no code is executing that is supported by the selected debug  engine  (for e.g. only native runtime code is executing).
And if I check the Debug --&gt; Module Window:
VS2015Test.vshost.exe no symbols loaded (even if I click load symbol it does not work)
VS2015Test.exe symbols loaded

And it also doesn&#39;t show the output on the console(it&#39;s a console application that just has the following lines of code:

    class Program
    {
            static void Main(string[] args)
            {
                Console.WriteLine(&quot;TEST&quot;);
                Console.ReadKey();
            }
        }


I tried to reinstall VS 2015, restarted the computer, deleted all files in %temp%/AppData/Microsoft/Visual Studio/14, started VS in Admin Mode but nothing seems to work.

One thing which makes debugging working is this option:
Tools --&gt; Options --&gt; Debugging --&gt; Use Managed Compability Mode

^^But that can&#39;t be the solution to use an old/legacy mode.

BTW: Debugging in VS 2013 is working fine.

Any help would be appreciated.

Visual Studio 2015 RTM - Debugging not working

After installing and then repairing my **VS2015** instance I still cannot get intellisense (server side) to work in my MVC views.
I get alerted by message prompt as soon as I open for the first time in a session an *.cshtml* file and get addressed to the Activitylog file.
Error message retrieved in ActivityLog.xml (short version): 

&gt; System.ArgumentException: Item has already been added. Key in
&gt; dictionary: &#39;RazorSupportedRuntimeVersion&#39;

Here&#39;s the full version:

&gt; System.Reflection.TargetInvocationException: Exception has been thrown
&gt; by the target of an invocation. ---&gt; **System.ArgumentException: Item
&gt; has already been added. Key in dictionary:
&gt; &#39;RazorSupportedRuntimeVersion&#39;** Key being added:
&gt; &#39;RazorSupportedRuntimeVersion&#39; at
&gt; System.Collections.Hashtable.Insert(Object key, Object nvalue, Boolean
&gt; add) at System.Collections.Hashtable.Add(Object key, Object value) at
&gt; System.Collections.Specialized.HybridDictionary.Add(Object key, Object
&gt; value) at
&gt; Microsoft.VisualStudio.Utilities.PropertyCollection.AddProperty(Object
&gt; key, Object property) at
&gt; Microsoft.VisualStudio.Html.Package.Razor.RazorVersionDetector.Microsoft.Html.Editor.ContainedLanguage.Razor.Def.IRazorVersionDetector.GetVersion(ITextBuffer
&gt; textBuffer) at
&gt; Microsoft.Html.Editor.ContainedLanguage.Razor.RazorUtility.TryGetRazorVersion(ITextBuffer
&gt; textBuffer, Version&amp; razorVersion) at
&gt; Microsoft.Html.Editor.ContainedLanguage.Razor.RazorErrorTagger..ctor(ITextBuffer
&gt; textBuffer) --- End of inner exception stack trace --- at
&gt; System.RuntimeMethodHandle.InvokeMethod(Object target, Object[]
&gt; arguments, Signature sig, Boolean constructor) at
&gt; System.Reflection.RuntimeConstructorInfo.Invoke(BindingFlags
&gt; invokeAttr, Binder binder, Object[] parameters, CultureInfo culture)
&gt; at System.RuntimeType.CreateInstanceImpl(BindingFlags bindingAttr,
&gt; Binder binder, Object[] args, CultureInfo culture, Object[]
&gt; activationAttributes, StackCrawlMark&amp; stackMark) at
&gt; System.Activator.CreateInstance(Type type, BindingFlags bindingAttr,
&gt; Binder binder, Object[] args, CultureInfo culture, Object[]
&gt; activationAttributes) at System.Activator.CreateInstance(Type type,
&gt; Object[] args) at
&gt; Microsoft.Html.Editor.ContainedLanguage.Common.ContainedCodeErrorTaggerProvider&#39;1.CreateTagger[T](ITextBuffer
&gt; textBuffer) at
&gt; Microsoft.VisualStudio.Text.Tagging.Implementation.TagAggregator&#39;1.GatherTaggers(ITextBuffer
&gt; textBuffer)

If it&#39;s any help, I have some font-coloring customizations, and have the Web Essentials, Bundler &amp; Minifier and Web Compiler extensions installed.

**UPDATE**: it happens either when I open an existing web application or when I create a brand new one from scratch.
There&#39;s also an instance of **VS2013** installed and running on the same machine.

[![No intellisense at all][1]][1] 

As you can see from the pic above I get no intellisense at all, nor any kind of code coloring.

Any clue about how to fix it?
Thanks in advance.


  [1]: http://i.stack.imgur.com/W1Waz.jpg

Visual Studio 2015 Broken Razor Intellisense

I&#39;m developing an ASP.NET MVC 5 application. I have an existing DB, from which I created my ADO.NET Entity Data Model.
I have a table in that DB which contains &quot;username&quot; and &quot;password&quot; column, and I want to use them to implement authentication and authorization in my Webapp; I cannot create any other database or table or column and I cannot use the standard Identity authentication, because of customer&#39;s requirements.
I don&#39;t need to manage signup, password changing or other stuffs: just login with password and username.
How can I do that? 

How to implement custom authentication in ASP.NET MVC 5

I know where to find the [_C# 5 Language Specification_](https://stackoverflow.com/questions/13467103/where-can-i-find-the-c-sharp-5-language-specification) but I cannot find the C# 6 Language Specification anywhere. 

_Where is the **C# 6** Language Specification?_

Where Can I Find the C# Language Specification 6.0?

Version 6.0 got a new feature of `nameof`, but I can&#39;t understand the purpose of it, as it just takes the variable name and changes it to a string on compilation.

I thought it might have some purpose when using `&lt;T&gt;` but when I try to `nameof(T)` it just prints me a `T` instead of the used type.

Any idea on the purpose?

What is the purpose of nameof?

.NET 4.6 introduces the `AsyncLocal&lt;T&gt;` class for flowing ambient data along the asynchronous flow of control. I&#39;ve previously used `CallContext.LogicalGet/SetData` for this purpose, and I&#39;m wondering if and in what ways the two are semantically different (beyond the obvious API differences like strong typing and lack of reliance on string keys).

How do the semantics of AsyncLocal differ from the logical call context?

I just recently downloaded and installed Visual Studio Professional 2015 (14.0.23107.0). The first time I opened up our solution (28 projects) and performed a Build -&gt; Rebuild Solution, my development machine came to an absolute crawl. The CPU was maxed out at 100% and the build never completed - even after &gt; 10 minutes.

I opened up Windows Task Manager and noticed: &gt; 10 VBCSCompiler.exe tasks running. When combined, these tasks sent the CPU &gt; 90%.

Any idea why there are so many of these tasks running? Any way to stop this from happening?

This is the closest thing I can find to someone else experiencing the same issue: https://github.com/dotnet/roslyn/issues/2790

**Update (8/7)**

-Hans Passant, great thought. My manager provided me with this release (14.0.23107.0). Is this the correct version for the &quot;official release&quot;?? I did not knowingly install any of the per-release versions of Visual Studio 2015. I don&#39;t think there are any beta bits lying around.

-Kyle Trauberman, I&#39;m not that familiar with environment variables in the context of Visual Studio; however, I naively ran `set DisableRosyln=true` in a VS (and MSBuild) Command Prompt Window. This did not seem to have any impact. VBCSCompiler.exe showed right back up even after restarting VS2015.

I repaired my VS2015 install and performed a reboot. This did not help.

**Update Part 2 (8/7)**
-Hans Passant, very impressive write up!! Although, the problem didn&#39;t happen this time around, I took a look at the things you described:

As far as the modules loaded with the VBCSCompiler.exe, here&#39;s what I have:

[![enter image description here][1]][1]

It&#39;s interesting that our .NET core assemblies are at different versions. You&#39;re at 4.06.79 while, I&#39;m at 4.06.81.

My &quot;client side dlls&quot; (located at C:\Program Files (x86)\MSBuild\14.0\Bin\Microsoft.Build.Tasks.CodeAnalysis.dll) are at the same version and time stamp as yours:
[![enter image description here][2]][2]

Oddly enough, when I look at the code in ILSpy, I see something slightly different - optimization perhaps?

        private static NamedPipeClientStream TryAllProcesses(string pipeName, int timeoutMs, CancellationToken cancellationToken, out string newPipeName)
    {
    	string str = pipeName;
    	int num = 1;
    	while (File.Exists(string.Format(&quot;\\\\.\\pipe\\{0}&quot;, pipeName)))
    	{
    		NamedPipeClientStream result;
    		if ((result = BuildClient.TryConnectToProcess(pipeName, timeoutMs, cancellationToken)) != null)
    		{
    			newPipeName = pipeName;
    			return result;
    		}
    		pipeName = str + &quot;.&quot; + num.ToString(CultureInfo.InvariantCulture);
    		num++;
    	}
    	newPipeName = pipeName;
    	return null;
    }

**Let me get back with you on the specific pipname argument passed to the instances of VBCSCompiler.exe. I will have to wait until it happens again.

  [1]: http://i.stack.imgur.com/rPd61.png
  [2]: http://i.stack.imgur.com/kcXmI.png

Numerous instances of VBCSCompiler.exe

I&#39;m trying to set the 4.6 .NET framework for my project and in the settings, as it wasn&#39;t listed, I chose the last option - to install more frameworks. That directed me to [a page](http://getdotnet.azurewebsites.net/target-dotnet-platforms.html#) where I selected VS2013, clicked on .NET 4.6 and [downloaded the software](http://www.microsoft.com/en-us/download/details.aspx?id=48130) but despite the installation being successful and restart of the computer, I still don&#39;t get to see the 4.6 .NET in the list of available options.

I&#39;ve googled around for it and [found mentioned](https://msdn.microsoft.com/en-us/library/5a4x27ek%28v=vs.110%29.aspx) that 4.6 is available on Win 8.1 (which is my system). However it doesn&#39;t say explicitly that it&#39;s available for VS 2013 (it doesn&#39;t say that it isn&#39;t, neither, though).

It&#39;s being said that .NET 4.6 is **included** in VS 2015 but I&#39;m not clear if it&#39;s **available** for VS 2013. Is it? And how to get it to work?

# Edit

After having installed the TP (a.k.a. Targeting Pack), I still get nothing new in the options, as the image depicts. Suggestions?

[![enter image description here][1]][1]


  [1]: http://i.stack.imgur.com/cuhAb.png

Visual Studio 2013 and .NET 4.6

Yesterday I found an [article by Christoph Nahr titled &quot;.NET Struct Performance&quot;][1]  which benchmarked several languages (C++, C#, Java, JavaScript) for a method which adds two point structs (`double` tuples).

As it turned out, C++ version takes about 1000ms to execute (1e9 iterations), while C# cannot get under ~3000ms on the same machine (and performs even worse in x64).

To test it myself, I took the C# code (and simplified slightly to call only the method where parameters are passed by value), and ran it on a i7-3610QM machine (3.1Ghz boost for single core), 8GB RAM, Win8.1, using .NET 4.5.2, RELEASE build 32-bit (x86 WoW64 since my OS is 64-bit). This is the simplified version:

    public static class CSharpTest
    {
        private const int ITERATIONS = 1000000000;
    
        [MethodImpl(MethodImplOptions.AggressiveInlining)]
        private static Point AddByVal(Point a, Point b)
        {
            return new Point(a.X + b.Y, a.Y + b.X);
        }
    
        public static void Main()
        {
            Point a = new Point(1, 1), b = new Point(1, 1);

            Stopwatch sw = Stopwatch.StartNew();
            for (int i = 0; i &lt; ITERATIONS; i++)
                a = AddByVal(a, b);
            sw.Stop();

            Console.WriteLine(&quot;Result: x={0} y={1}, Time elapsed: {2} ms&quot;, 
                a.X, a.Y, sw.ElapsedMilliseconds);
        }
    }

With `Point` defined as simply:

    public struct Point 
    {
        private readonly double _x, _y;

        public Point(double x, double y) { _x = x; _y = y; }
        
        public double X { get { return _x; } }
        
        public double Y { get { return _y; } }
    }
    
Running it produces the results similar to those in the article:
    
    Result: x=1000000001 y=1000000001, Time elapsed: 3159 ms
    
**First strange observation**

Since the method should be inlined, I wondered how the code would perform if I removed structs altogether and simply inlined the whole thing together:

    public static class CSharpTest
    {
        private const int ITERATIONS = 1000000000;

        public static void Main()
        {
            // not using structs at all here
            double ax = 1, ay = 1, bx = 1, by = 1;

            Stopwatch sw = Stopwatch.StartNew();
            for (int i = 0; i &lt; ITERATIONS; i++)
            {
                ax = ax + by;
                ay = ay + bx;
            }
            sw.Stop();

            Console.WriteLine(&quot;Result: x={0} y={1}, Time elapsed: {2} ms&quot;, 
                ax, ay, sw.ElapsedMilliseconds);
        }
    }
    
And got practically the same result (actually 1% slower after several retries), meaning that JIT-ter seems to be doing a good job optimizing all function calls:

    Result: x=1000000001 y=1000000001, Time elapsed: 3200 ms

It also means that the benchmark doesn&#39;t seem to measure any `struct` performance and actually only seem to measure basic `double` arithmetic (after everything else gets optimized away).

**The weird stuff**

Now comes the weird part. If I merely add **another stopwatch outside the loop** (yes, I narrowed it down to this crazy step after several retries), the code runs **three times faster**:

    public static void Main()
    {
        var outerSw = Stopwatch.StartNew();     // &lt;-- added

        {
            Point a = new Point(1, 1), b = new Point(1, 1);

            var sw = Stopwatch.StartNew();
            for (int i = 0; i &lt; ITERATIONS; i++)
                a = AddByVal(a, b);
            sw.Stop();

            Console.WriteLine(&quot;Result: x={0} y={1}, Time elapsed: {2} ms&quot;,
                a.X, a.Y, sw.ElapsedMilliseconds);
        }

        outerSw.Stop();                         // &lt;-- added
    }
    
    Result: x=1000000001 y=1000000001, Time elapsed: 961 ms
    
That&#39;s ridiculous! And it&#39;s not like `Stopwatch` is giving me wrong results because I can clearly see that it ends after a single second.

**Can anyone tell me what might be happening here?**

**(Update)**

Here are two methods in the same program, which shows that the reason is not JITting:

    public static class CSharpTest
    {
        private const int ITERATIONS = 1000000000;

        [MethodImpl(MethodImplOptions.AggressiveInlining)]
        private static Point AddByVal(Point a, Point b)
        {
            return new Point(a.X + b.Y, a.Y + b.X);
        }

        public static void Main()
        {
            Test1();
            Test2();

            Console.WriteLine();

            Test1();
            Test2();
        }

        private static void Test1()
        {
            Point a = new Point(1, 1), b = new Point(1, 1);

            var sw = Stopwatch.StartNew();
            for (int i = 0; i &lt; ITERATIONS; i++)
                a = AddByVal(a, b);
            sw.Stop();

            Console.WriteLine(&quot;Test1: x={0} y={1}, Time elapsed: {2} ms&quot;, 
                a.X, a.Y, sw.ElapsedMilliseconds);
        }

        private static void Test2()
        {
            var swOuter = Stopwatch.StartNew();

            Point a = new Point(1, 1), b = new Point(1, 1);

            var sw = Stopwatch.StartNew();
            for (int i = 0; i &lt; ITERATIONS; i++)
                a = AddByVal(a, b);
            sw.Stop();

            Console.WriteLine(&quot;Test2: x={0} y={1}, Time elapsed: {2} ms&quot;, 
                a.X, a.Y, sw.ElapsedMilliseconds);

            swOuter.Stop();
        }
    }

Output:

    Test1: x=1000000001 y=1000000001, Time elapsed: 3242 ms
    Test2: x=1000000001 y=1000000001, Time elapsed: 974 ms

    Test1: x=1000000001 y=1000000001, Time elapsed: 3251 ms
    Test2: x=1000000001 y=1000000001, Time elapsed: 972 ms

**[Here is a pastebin.][2]** You need to run it as a 32-bit release on .NET 4.x (there are a couple of checks in the code to ensure this).

**(Update 4)**

Following @usr&#39;s comments on @Hans&#39; answer, I checked the optimized disassembly for both methods, and they are rather different:

[![Test1 on the left, Test2 on the right][4]][4]

This seems to show that the difference might be due to compiler acting funny in the first case, rather than double field alignment?

Also, if I add **two** variables (total offset of 8 bytes), I still get the same speed boost - and it no longer seems it&#39;s related to field alignment mention by Hans Passant:

    // this is still fast?
    private static void Test3()
    {
        var magical_speed_booster_1 = &quot;whatever&quot;;
        var magical_speed_booster_2 = &quot;whatever&quot;;

        {
            Point a = new Point(1, 1), b = new Point(1, 1);

            var sw = Stopwatch.StartNew();
            for (int i = 0; i &lt; ITERATIONS; i++)
                a = AddByVal(a, b);
            sw.Stop();

            Console.WriteLine(&quot;Test2: x={0} y={1}, Time elapsed: {2} ms&quot;,
                a.X, a.Y, sw.ElapsedMilliseconds);
        }

        GC.KeepAlive(magical_speed_booster_1);
        GC.KeepAlive(magical_speed_booster_2);
    }

  [1]: http://kynosarges.org/StructPerformance.html
  [2]: http://pastebin.com/9RykzS6g
  [3]: http://stackoverflow.com/a/32117038/69809
  [4]: http://i.stack.imgur.com/Bg3Fc.png

Weird performance increase in simple benchmark

What is the explanation for the result from the following operation?

    k += c += k += c;

I was trying to understand the output result from the following code:

    int k = 10;
    int c = 30;
    k += c += k += c;
    //k=80 instead of 110
    //c=70

and currently I am struggling with understanding why the result for &quot;k&quot; is 80. Why is assigning k=40 not working (actually Visual Studio tells me that that value is not being used elsewhere)?

Why is k 80 and not 110?

If I split the operation to:

    k+=c;
    c+=k;
    k+=c;

the result is k=110.

I was trying to look through the [CIL][1], but I am not so profound in interpreting generated CIL and can not get a few details:

     // [11 13 - 11 24]
    IL_0001: ldc.i4.s     10
    IL_0003: stloc.0      // k

    // [12 13 - 12 24]
    IL_0004: ldc.i4.s     30
    IL_0006: stloc.1      // c

    // [13 13 - 13 30]
    IL_0007: ldloc.0      // k expect to be 10
    IL_0008: ldloc.1      // c
    IL_0009: ldloc.0      // k why do we need the second load?
    IL_000a: ldloc.1      // c
    IL_000b: add          // I expect it to be 40
    IL_000c: dup          // What for?
    IL_000d: stloc.0      // k - expected to be 40
    IL_000e: add
    IL_000f: dup          // I presume the &quot;magic&quot; happens here
    IL_0010: stloc.1      // c = 70
    IL_0011: add
    IL_0012: stloc.0      // k = 80??????

  [1]: http://en.wikipedia.org/wiki/Common_Intermediate_Language


Content Type	Original Author	Original Content on Stackoverflow
Question	Dragno	View Question on Stackoverflow
Solution 1 - C#	Hans Passant	View Answer on Stackoverflow
Solution 2 - C#	leppie	View Answer on Stackoverflow

Why is !0 a type in Microsoft Intermediate Language (MSIL)?

C# Problem Overview

C# Solutions

Solution 1 - C#

Solution 2 - C#

Does JavaScript support array/list comprehensions like Python?

How to change color of hamburger icon in material design navigation drawer

Attributions