This is an internal package used to make work with reflection cleaner and easier than ever before.
$ dotnet add package TryAtSoftware.Extensions.Reflection
TryAtSoftware.Extensions is a library containing extension methods and utility components that should simplify (and optimize) some common operations with reflection.
Try At Software is a software development company based in Bulgaria. We are mainly using dotnet technologies (C#, ASP.NET Core, Entity Framework Core, etc.) and our main idea is to provide a set of tools that can simplify the majority of work a developer does on a daily basis.
TypeNamesThis is an utility component that should construct and cache a beautified name for a type. There are two ways to use it:
TypeNames<T> class and the Value property;TypeNames class and the Get(type) method.The main idea behind this component is to provide efficiently a meaningful and readable type name (including open generics).
Here is a comparison between the type.ToString() and TypeNames.Get(type):
| type | type.ToString() | TypeNames.Get(type) |
|---|---|---|
Task<int> | System.Threading.Tasks.Task`1[System.Int32] | Task<Int32> |
Task<List<long>> | System.Collections.Generic.List`1[System.Threading.Tasks.Task`1[System.Int64]] | List<Task<Person>> |
Task<> | System.Threading.Tasks.Task`1[TResult] | Task<TResult> |
List<> | System.Collections.Generic.List`1[T] | List<T> |
IMembersBinderThis is an interface defining the structure of an utility component exposing information about a specific subset of the members for a given type.
There are two classes implementing this interface - a non-generic MembersBinder and a generic MembersBinder<T>.
They accept the following parameters throughout their constructors:
isValid: A filtering function that should determine whether or not a member should be included within the final result set. If no value is provided for this parameter, every member will be included.keySelector: A function mapping each each member to unique key. If no value is provided to this parameter, the name of the member will be used (which means that in case of overrides or members with the same name there will be issues)bindingFlags: The binding flags used to control the member search throughout reflection.Example:
IMembersBinder binder = new MembersBinder<TEntity>(IsValidMember, BindingFlags.Public | BindingFlags.Instance);
// Equivalent to:
// IMembersBinder binder = new MembersBinder(typeof(TEntity), IsValidMember, BindingFlags.Public | BindingFlags.Instance);
static bool IsValidMember(MemberInfo memberInfo)
=> memberInfo switch
{
PropertyInfo pi => pi.CanWrite,
FieldInfo _ => true,
_ => false
};
// Now every discovered member for the `TEntity` type will be mapped against its name throughout the `binder.MemberInfos` dictionary.It is known that whenever the .GetMembers() method is invoked for an interface type, none of the members defined in extended interfaces will be returned.
The default implementations of the IMembersBinder method overcome this and will include members from all of the extended interfaces by recursively retrieving all of them.
That being said, it is obvious that the ReflectedType in this case may not equal the Type of the IMembersBinder instance.
IHierarchyScannerThis is an interface defining the structure of a component responsible for scanning type hierarchies. A common application of this interface and its corresponding implementing types is to scan a type hierarchy for a given attribute.
It is a common practice to decorate methods or properties with attributes in order to model some kind of behavior related to them.
But it often becomes inconvenient to use the same attribute(s) over and over again for all methods or properties within a given class or assembly.
A better solution would be to define the attribute(s) at a higher level.
And this is where the IHierarchyScanner may be used - it can scan for attributes throughout the type hierarchy, allowing for effortless implementation, management and control.
The default implementation would scan the requested member and its type but this can be easily extend to include the assembly as well (if the attribute is inheritable, base types and overriden methods will be included as well). The result set should always be ordered hierarchically!
Example:
IHierarchyScanner hierarchyScanner = new HierarchyScanner();
IReadOnlyCollection<MyAttribute> attributes = hierarchyScanner.ScanForAttribute<MyAttribute>(memberInfo);ConstructPropertyAccessorThis is an extension method that should construct an expression for accessing the value of a specific property.
Usually, it is a good practice to minimize the reflection calls in code. One way of achieving this is throughout expressions (that are constructed and compiled only once for the lifetime of a program).
This expression method can be easily used with the IMembersBinder we described in the previous chapter.
Conversions are also handled, e.g. a common use case is to retrieve the values of all properties by boxing them as object instances - this can be achieved without any additional configurations as long as the required conversion can be executed.
Example:
IMembersBinder binder = new MembersBinder<TEntity>(x => x is PropertyInfo { CanRead: true }, BindingFlags.Public | BindingFlags.Instance);
List<Expression<Func<TEntity, object>>> valueAccessors = new List<Expression<Func<TEntity, object>>>();
foreach (var (_, memberInfo) in binder.MemberInfos)
{
var propertyInfo = memberInfo as PropertyInfo;
var accessor = propertyInfo.ConstructPropertyAccessor<TEntity, object>();
valueAccessors.Add(accessor);
}ConstructPropertySetterThis is an extension method that should construct an expression for setting the value of a specific property.
Usually, it is a good practice to minimize the reflection calls in code. One way of achieving this is throughout expressions (that are constructed and compiled only once for the lifetime of a program).
This expression method can be easily used with the IMembersBinder we described in the previous chapter.
Conversions are also handled, e.g. a common use case is to set the values of all properties after unboxing object instances - this can be achieved without any additional configurations as long as the required conversion can be executed.
Example:
IMembersBinder binder = new MembersBinder<TEntity>(x => x is PropertyInfo { CanWrite: true}, BindingFlags.Public | BindingFlags.Instance);
List<Expression<Action<TEntity, object>>> valueSetters = new List<Expression<Action<TEntity, object>>>();
foreach (var (_, memberInfo) in binder.MemberInfos)
{
var propertyInfo = memberInfo as PropertyInfo;
var setter = propertyInfo.ConstructPropertySetter<TEntity, object>();
valueSetters.Add(setter);
}ConstructObjectInitializerThis is an extension method that should construct an expression for instantiating an object using a specific constructor.
Usually, it is a good practice to minimize the reflection calls in code. One way of achieving this is throughout expressions (that are constructed and compiled only once for the lifetime of a program).
This expression method can be easily used with the IMembersBinder we described in the previous chapter.
This methods accept one optional parameter called includeParametersCountValidation. It indicates whether or not the subsequently constructed expression should include a check to validate the correct count of provided values.
An expression constructed by this extension method can be compiled to a function that accepts an array of values that correspond to the parameters of the extended ConstructorInfo instance.
If any of the parameters is optional and its default value should be used, the corresponding element (from the provided array) must be null.
Example:
IMembersBinder binder = new MembersBinder<TEntity>(x => x is ConstructorInfo, x => PrepareConstructorKey((ConstructorInfo)x), BindingFlags.Public | BindingFlags.Instance);
List<Expression<Func<object?[], TEntity>>> objectInitializers = new List<Expression<Func<object?[], TEntity>>>();
foreach (var (_, memberInfo) in binder.MemberInfos)
{
var constructorInfo = memberInfo as ConstructorInfo;
var objectInitializer = constructorInfo.ConstructObjectInitializer<TEntity>();
objectInitializers.Add(objectInitializer);
}
static string PrepareConstructorKey(ConstructorInfo constructorInfo)
{
var parameterTypeNames = constructorInfo.GetParameters().Select(p => TypeNames.Get(p.ParameterType));
return $"Constructor[{string.Join(", ", parameterTypeNames)}]";
}ExtractGenericParametersSetupThis method will produce a dictionary where against the name of a generic parameter is stored the actual type associated with that parameter according to some external configuration. This method should be used whenever each generic parameter is uniquely identified with a certain attribute. All entries within the external configuration map should have as a key the attribute type and as a value the type that should be substituted for a generic parameter decorated with the corresponding attribute.
If there are none or multiple attributes for a generic parameter, this method will throw an exception. If the external configuration is missing some attribute type, an exception will be thrown as well.
Example:
public class MyType<[KeyType] TKey, [ValueType] TValue> {}
IDictionary<Type, Type> typesMap = new Dictionary<Type, Type> { { typeof(KeyTypeAttribute), typeof(int) }, { typeof(ValueTypeAttribute), typeof(string) } };
/// should return { "TKey": typeof(int), "TValue": typeof(string) }
IDictionary<string, Type> genericParametersSetup = ExtractGenericParametersSetup(typeof(MyType<,>), typesMap);MakeGenericTypeUse this method to make the extended type generic using a parameters setup.
Example:
public class MyType<[KeyType] TKey, [ValueType] TValue> {}
IDictionary<Type, Type> typesMap = new Dictionary<Type, Type> { { typeof(KeyTypeAttribute), typeof(int) }, { typeof(ValueTypeAttribute), typeof(string) } };
IDictionary<string, Type> genericParametersSetup = ExtractGenericParametersSetup(typeof(MyType<,>), typesMap);
Type genericType = typeof(MyType<,>).MakeGenericType(genericParametersSetup);