Performance - Replace use of Activator in Is<T>

[rcollette]

Activator is notably slow.
Since you already have put a zero arg constructor constraint on T, you should be able to just do

return new T().FeatureEnabled;

If for some reason that doesn't work then you can look at a generic version of DynamicMethod, but I don't think you will need it. https://stackoverflow.com/a/22651558

[mikanyg]

As far as I can see, then current implementation does what you suggest https://github.com/jason-roberts/FeatureToggle/blob/master/src/FeatureToggle.Common/Is.cs

[rcollette]

That's interesting. When I decompile in Jetbrains Rider I get a different set of source. I wonder if new T() actually compiles to using Activator.

// Decompiled with JetBrains decompiler
// Type: FeatureToggle.Is`1
// Assembly: FeatureToggle.Common, Version=4.0.2.0, Culture=neutral, PublicKeyToken=null
// MVID: 56948869-D82D-4DE1-93B1-08121C96F0D7
// Assembly location: /Users/richardcollette/.nuget/packages/featuretoggle.common/4.0.2/lib/netstandard1.4/FeatureToggle.Common.dll

using System;

namespace FeatureToggle
{
  public static class Is<T> where T : IFeatureToggle, new()
  {
    public static bool Enabled
    {
      get
      {
        return Activator.CreateInstance<T>().FeatureEnabled;
      }
    }

    public static bool Disabled
    {
      get
      {
        return !Activator.CreateInstance<T>().FeatureEnabled;
      }
    }
  }
}

[mikanyg]

Decompiling is not nearly as good as looking at actual source code. That is my experience when decompiling .NET framework BCL code and comparing with the open source code.

[jason-roberts]

Is this something that needs to be looked at in v5 or can we close @rcollette @mikanyg ?

[mikanyg]

I don’t know any other way to instantiate generic types than what is already implemented. So I would say close this issue.

[rcollette]

Close please.