Reflective programming

In computer science, reflective programming or reflection is the ability of a process to examine, introspect, and modify its own structure and behavior.

Historical background

The earliest computers were programmed in their native assembly languages, which were inherently reflective, as these original architectures could be programmed by defining instructions as data and using self-modifying code. As the bulk of programming moved to higher-level compiled languages such as ALGOL, COBOL, Fortran, Pascal, and C, this reflective ability largely disappeared until new programming languages with reflection built into their type systems appeared.

Brian Cantwell Smith's 1982 doctoral dissertation introduced the notion of computational reflection in procedural programming languages and the notion of the meta-circular interpreter as a component of 3-Lisp.

Uses

Reflection helps programmers make generic software libraries to display data, process different formats of data, perform serialization and deserialization of data for communication, or do bundling and unbundling of data for containers or bursts of communication.

Effective use of reflection almost always requires a plan: A design framework, encoding description, object library, a map of a database or entity relations.

Reflection makes a language more suited to network-oriented code. For example, it assists languages such as Java to operate well in networks by enabling libraries for serialization, bundling and varying data formats. Languages without reflection such as C are required to use auxiliary compilers for tasks like Abstract Syntax Notation to produce code for serialization and bundling.

Reflection can be used for observing and modifying program execution at runtime. A reflection-oriented program component can monitor the execution of an enclosure of code and can modify itself according to a desired goal of that enclosure. This is typically accomplished by dynamically assigning program code at runtime.

In object-oriented programming languages such as Java, reflection allows inspection of classes, interfaces, fields and methods at runtime without knowing the names of the interfaces, fields, methods at compile time. It also allows instantiation of new objects and invocation of methods.

Reflection is often used as part of software testing, such as for the runtime creation/instantiation of mock objects.

Reflection is also a key strategy for metaprogramming.

In some object-oriented programming languages such as C# and Java, reflection can be used to bypass member accessibility rules. For C#-properties this can be achieved by writing directly onto the (usually invisible) backing field of a non-public property. It is also possible to find non-public methods of classes and types and manually invoke them. This works for project-internal files as well as external libraries such as .NET's assemblies and Java's archives.

Implementation

A language that supports reflection provides a number of features available at runtime that would otherwise be difficult to accomplish in a lower-level language. Some of these features are the abilities to:

Discover and modify source-code constructions (such as code blocks, classes, methods, protocols, etc.) as first-class objects at runtime.
Convert a string matching the symbolic name of a class or function into a reference to or invocation of that class or function.
Evaluate a string as if it were a source-code statement at runtime.
Create a new interpreter for the language's bytecode to give a new meaning or purpose for a programming construct.

These features can be implemented in different ways. In MOO, reflection forms a natural part of everyday programming idiom. When verbs (methods) are called, various variables such as <code>verb</code> (the name of the verb being called) and <code>this</code> (the object on which the verb is called) are populated to give the context of the call. Security is typically managed by accessing the caller stack programmatically: Since <code>callers()</code> is a list of the methods by which the current verb was eventually called, performing tests on <code>callers()[0]</code> (the command invoked by the original user) allows the verb to protect itself against unauthorised use.

Compiled languages rely on their runtime system to provide information about the source code. A compiled Objective-C executable, for example, records the names of all methods in a block of the executable, providing a table to correspond these with the underlying methods (or selectors for these methods) compiled into the program. In a compiled language that supports runtime creation of functions, such as Common Lisp, the runtime environment must include a compiler or an interpreter.

Reflection can be implemented for languages without built-in reflection by using a program transformation system to define automated source-code changes.

Security considerations

Reflection may allow a user to create unexpected control flow paths through an application, potentially bypassing security measures. This may be exploited by attackers. Historical vulnerabilities in Java caused by unsafe reflection allowed code retrieved from potentially untrusted remote machines to break out of the Java sandbox security mechanism. A large scale study of 120 Java vulnerabilities in 2013 concluded that unsafe reflection is the most common vulnerability in Java, though not the most exploited.

Runtime Performance

Reflective programming commonly introduces a non-negligible runtime performance overhead. Because reflective operations are resolved dynamically at execution time, many Java compiler and JVM optimizations—such as method inlining, static binding, and aggressive just-in-time specialization—cannot be fully applied. As a consequence, reflective calls are typically slower than their statically resolved counterparts. Microbenchmark and application-level studies on Java have shown that reflective operations can incur substantial runtime overhead, especially for method invocation and dynamic object creation, with slowdowns ranging from around 20–40% in moderate cases to more than 300× in heavily reflective dispatch scenarios. In sequential applications, reflection can significantly increase execution time and memory consumption when used in performance-critical code paths. In multithreaded applications, reflective implementations generally preserve scalability, but still exhibit noticeably higher absolute execution times—commonly between 1.5x and 10× slower—than equivalent non-reflective code.

import std;

using std::string_view;

using std::views::filter;

using namespace std::meta;

consteval bool isNonstaticMethod(info mem) noexcept {

return is_class_member(mem) && !is_static_member(mem) && is_function(mem);

}

consteval info findMethod(info type, string_view name) {

for (info member : members_of(type, access_context::current())

| filter(isNonstaticMethod)) {

if (has_identifier(member) && identifier_of(member) == name) {

return member;

}

// Note: this is std::meta::exception, not std::exception

throw exception(std::format("Failed to retrieve method {} from type {}", name, identifier_of(type)), ^^findMethod);

}

template <info T, const char* Name>

constexpr auto createInvokerImpl = []() -> auto {

static constexpr info M = findMethod(T, Name);

contract_assert(

parameters_of(M).size() == 0 &&

return_type_of(M) == ^^void

);

return []([:T:]& instance) -> void { instance.[:M:](); };

}();

consteval info createInvoker(info type, string_view name) {

return substitute(^^createInvokerImpl, { reflect_constant(type), reflect_constant_string(name) });

}

class Foo {

private:

// ...

public:

Foo() = default;

void printHello() const {

std::println("Hello, world!");

}

};

int main(int argc, char* argv[]) {

Foo foo;

// Without reflection

foo.printHello();

// With reflection

auto invokePrint = [:createInvoker(^^Foo, "printHello"):];

invokePrint(foo);

return 0;

}

</syntaxhighlight>

C#

The following is an example in C#:

namespace Wikipedia.Examples;

using System;

using System.Reflection;

class Foo

{

// ...

public Foo() {}

public void PrintHello()

{

Console.WriteLine("Hello, world!");

}

public class InvokeFooExample

{

static void Main(string[] args)

{

// Without reflection

Foo foo = new();

foo.PrintHello();

// With reflection

Object reflectedFoo = Activator.CreateInstance(typeof(Foo));

MethodInfo method = reflectedFoo.GetType()

.GetMethod("PrintHello");

method.Invoke(foo, null);

}

</syntaxhighlight>

Delphi, Object Pascal

This Delphi and Object Pascal example assumes that a class has been declared in a unit called :

uses RTTI, Unit1;

procedure WithoutReflection;

var

Foo: TFoo;

begin

Foo := TFoo.Create;

try

Foo.Hello;

finally

Foo.Free;

end;

procedure WithReflection;

var

RttiContext: TRttiContext;

RttiType: TRttiInstanceType;

Foo: TObject;

begin

RttiType := RttiContext.FindType('Unit1.TFoo') as TRttiInstanceType;

Foo := RttiType.GetMethod('Create').Invoke(RttiType.MetaclassType, []).AsObject;

try

RttiType.GetMethod('Hello').Invoke(Foo, []);

finally

Foo.Free;

end;

</syntaxhighlight>

eC

The following is an example in eC:

// Without reflection

Foo foo{};

foo.hello();

// With reflection

Class fooClass = eSystem_FindClass(__thisModule, "Foo");

Instance foo = eInstance_New(fooClass);

Method m = eClass_FindMethod(fooClass, "hello", fooClass.module);

((void(*)())(void*)m.function)(foo);

</syntaxhighlight>

Go

The following is an example in Go:

import (

"fmt"

"reflect"

)

type Foo struct{}

func (f Foo) Hello() {

fmt.Println("Hello, world!")

}

func main() {

// Without reflection

var f Foo

f.Hello()

// With reflection

var fT reflect.Type = reflect.TypeOf(Foo{})

var fV reflect.Value = reflect.New(fT)

var m reflect.Value = fV.MethodByName("Hello")

if m.IsValid() {

m.Call(nil)

} else {

fmt.Println("Method not found")

}

</syntaxhighlight>

Java

The following is an example in Java:

package org.wikipedia.examples;

import java.lang.reflect.Method;

class Foo {

// ...

public Foo() {}

public void printHello() {

System.out.println("Hello, world!");

}

public class InvokeFooExample {

public static void main(String[] args) {

// Without reflection

Foo foo = new Foo();

foo.printHello();

// With reflection

try {

Foo reflectedFoo = Foo.class

.getDeclaredConstructor()

.newInstance();

Method m = reflectedFoo.getClass()

.getDeclaredMethod("printHello", new Class<?>[0]);

m.invoke(reflectedFoo);

} catch (ReflectiveOperationException e) {

System.err.printf("An error occurred: %s%n", e.getMessage());

}

</syntaxhighlight>

Java also provides an internal class (not officially in the Java Class Library) in module <code>jdk.unsupported</code>, <code>sun.reflect.Reflection</code> which is used by <code>sun.misc.Unsafe</code>. It contains one method, for obtaining the class making a call at a specified depth. This is now superseded by using the class <code>java.lang.StackWalker.StackFrame</code> and its method .

JavaScript/TypeScript

The following is an example in JavaScript:

import 'reflect-metadata';

// Without reflection

const foo = new Foo();

foo.hello();

// With reflection

const foo = Reflect.construct(Foo);

const hello = Reflect.get(foo, 'hello');

Reflect.apply(hello, foo, []);

// With eval

eval('new Foo().hello()');

</syntaxhighlight>

The following is the same example in TypeScript:

import 'reflect-metadata';

// Without reflection

const foo: Foo = new Foo();

foo.hello();

// With reflection

const foo: Foo = Reflect.construct(Foo);

const hello: (this: Foo) => void = Reflect.get(foo, 'hello') as (this: Foo) => void;

Reflect.apply(hello, foo, []);

// With eval

eval('new Foo().hello()');

</syntaxhighlight>

Julia

The following is an example in Julia:

julia> struct Point

x::Int

end

Inspection with reflection

julia> fieldnames(Point)

(:x, :y)

julia> fieldtypes(Point)

(Int64, Any)

julia> p = Point(3,4)

Access with reflection

julia> getfield(p, :x)

</syntaxhighlight>

Kotlin

Using Java reflection:

package org.wikipedia.examples

import java.lang.reflect.Method

class Foo {

// ...

constructor()

fun printHello() {

println("Hello, world!")

}

fun main(args: Array<String>) {

// Without reflection

val foo = Foo()

foo.printHello()

// With reflection

try {

// Foo::class.java retrieves a java.lang.Class<Foo>

val reflectedFoo = Foo::class.java

.getDeclaredConstructor()

.newInstance()

val m: Method = reflectedFoo.javaClass

.getDeclaredMethod("printHello")

m.invoke(reflectedFoo)

} catch (e: ReflectiveOperationException) {

System.err.printf("An error occurred: %s%n", e.message)

}

</syntaxhighlight>

Using pure Kotlin:

package org.wikipedia.examples

import kotlin.reflect.full.createInstance

import kotlin.reflect.full.functions

class Foo {

// ...

fun printHello() {

println("Hello, world!")

}

fun main(args: Array<String>) {

// Without reflection

val foo = Foo()

foo.printHello()

// With reflection

try {

val kClass = Foo::class

val reflectedFoo = kClass.createInstance()

val function = kClass.functions.first { it.name == "printHello" }

function.call(reflectedFoo)

} catch (e: Exception) {

System.err.printf("An error occurred: %s%n", e.message)

}

</syntaxhighlight>

Objective-C

The following is an example in Objective-C, implying either the OpenStep or Foundation Kit framework is used:

// Foo class.

@interface Foo : NSObject

- (void)hello;

@end

// Sending "hello" to a Foo instance without reflection.

Foo* obj = [[Foo alloc] init];

[obj hello];

// Sending "hello" to a Foo instance with reflection.

id obj = [[NSClassFromString(@"Foo") alloc] init];

[obj performSelector: @selector(hello)];

</syntaxhighlight>

Perl

The following is an example in Perl:

Without reflection

my $foo = Foo->new;

$foo->hello;

Foo->new->hello;

With reflection

my $class = "Foo"

my $constructor = "new";

my $method = "hello";

my $f = $class->$constructor;

$f->$method;

$class->$constructor->$method;

with eval

eval "new Foo->hello;";

</syntaxhighlight>

PHP

The following is an example in PHP:

// Without reflection

$foo = new Foo();

$foo->hello();

// With reflection, using Reflections API

$reflector = new ReflectionClass("Foo");

$foo = $reflector->newInstance();

$hello = $reflector->getMethod("hello");

$hello->invoke($foo);

</syntaxhighlight>

Python

The following is an example in Python:

from typing import Any

class Foo:

def print_hello(self) -> None:

print("Hello, world!")

if __name__ == "__main__":

Without reflection

obj: Foo = Foo()

obj.print_hello()

With reflection

obj: Foo = globals()["Foo"]()

_: Any = getattr(obj, "print_hello")()

With eval

eval("Foo().print_hello()")

</syntaxhighlight>

R

The following is an example in R:

Without reflection, assuming foo() returns an S3-type object that has method "hello"

obj <- foo()

hello(obj)

With reflection

class_name <- "foo"

generic_having_foo_method <- "hello"

obj <- do.call(class_name, list())

do.call(generic_having_foo_method, alist(obj))

</syntaxhighlight>

Ruby

The following is an example in Ruby:

Without reflection

obj = Foo.new

obj.hello

With reflection

obj = Object.const_get("Foo").new

obj.send :hello

With eval

eval "Foo.new.hello"

</syntaxhighlight>

Rust

Rust does not have compile-time reflection in the standard library, but it is possible using some third-party libraries such as "".

use std::any::TypeId;

use bevy_reflect::prelude::*;

use bevy_reflect::{

FunctionRegistry,

GetTypeRegistration,

Reflect,

ReflectFunction,

ReflectFunctionRegistry,

ReflectMut,

ReflectRef,

TypeRegistry

};

[derive(Reflect)]
[reflect(DoFoo)]

struct Foo {

// ...

}

impl Foo {

fn new() -> Self {

Foo {}

}

fn print_hello(&self) {

println!("Hello, world!");

}

[reflect_trait]

trait DoFoo {

fn print_hello(&self);

}

impl DoFoo for Foo {

fn print_hello(&self) {

self.print_hello();

}

fn main() {

// Without reflection

let foo: Foo = Foo::new();

foo.print_hello();

// With reflection

let mut registry: TypeRegistry = TypeRegistry::default();

registry.register::<Foo>();

registry.register_type_data::<Foo, ReflectFunctionRegistry>();

registry.register_type_data::<Foo, ReflectDoFoo>();

let foo: Foo = Foo;

let reflect_foo: Box<dyn Reflect> = Box::new(foo);

// Version 1: call hello by trait

let trait_registration: &ReflectDoFoo = registry

.get_type_data::