In Ninject 1.0 I had following binding definitions:
Bind<ITarget>().To<Target1>().Only(When.Context.Variable("variable").EqualTo(true));
Bind<ITarget>().To<Target2>();
Given such bindings I had calls:
ITarget target = kernel.Get<ITarget>(With.Parameters.ContextVariable("variable", true));
ITarget target = kernel.Get<ITarget>(With.Parameters.ContextVariable("variable", false));
First call was resolved to instance of Target1, second call was resolved to instance of Target2.
How to translate this into Ninject 2.0?
You can use metadata,
[Fact]
public void MetadataBindingExample()
{
string metaDataKey = "key";
kernel.Bind<IWeapon>().To<Shuriken>().WithMetadata(metaDataKey, true);
kernel.Bind<IWeapon>().To<Sword>().WithMetadata(metaDataKey, false);
kernel.Bind<IWeapon>().To<Knife>();
var weapon = kernel.Get<IWeapon>(metadata => metadata.Has(metaDataKey) && metadata.Get<bool>(metaDataKey));
Assert.IsType<Shuriken>( weapon );
weapon = kernel.Get<IWeapon>(metadata => metadata.Has(metaDataKey) && !metadata.Get<bool>(metaDataKey));
Assert.IsType<Sword>(weapon);
weapon = kernel.Get<IWeapon>(metadata => !metadata.Has(metaDataKey));
Assert.IsType<Knife>(weapon);
}