Is it possible to extend or modify the code of a C# class at runtime?
My question specifically revolves around Monkey Patching / Duck Punching or Meta Object Programming (MOP), as it happens in scripting languages such as Groovy, Ruby etc.
For those still stumbling on this question in the present day, there is indeed a present-day library called Harmony that relatively-straightforwardly enables such monkey-patching at runtime. Its focus is on video game modding (particularly games built with Unity), but there ain't much stopping folks from using it outside of that use case.
Copying the example from their introduction, if you have an existing class like so:
public class SomeGameClass
{
public bool isRunning;
public int counter;
private int DoSomething()
{
if (isRunning)
{
counter++;
}
return counter * 10;
}
}
Then Harmony can patch it like so:
using HarmonyLib;
using Intro_SomeGame;
public class MyPatcher
{
// make sure DoPatching() is called at start either by
// the mod loader or by your injector
public static void DoPatching()
{
var harmony = new Harmony("com.example.patch");
harmony.PatchAll();
}
}
[HarmonyPatch(typeof(SomeGameClass))]
[HarmonyPatch("DoSomething")]
class Patch01
{
static AccessTools.FieldRef<SomeGameClass, bool> isRunningRef =
AccessTools.FieldRefAccess<SomeGameClass, bool>("isRunning");
static bool Prefix(SomeGameClass __instance, ref int ___counter)
{
isRunningRef(__instance) = true;
if (___counter > 100)
return false;
___counter = 0;
return true;
}
static void Postfix(ref int __result)
{
__result *= 2;
}
}
Here, we have a "prefix" patch which gets inserted before the original method runs, allowing us to set variables within the method, set fields on the method's class, or even skip the original method entirely. We also have a "postfix" patch which gets inserted after the original method runs, and can manipulate things like the return value.
Obviously this ain't quite as nice as the sorts of monkey-patching you can do in e.g. Ruby, and there are a lot of caveats that might hinder its usefulness depending on your use case, but in those situations where you really do need to alter methods, Harmony's a pretty proven approach to doing so.
Is it possible to extend or modify the code of a C# class at run-time?
No it is not possible to do this in .NET. You could write derived classes and override methods (if they are virtual) but you cannot modify an existing class. Just imagine if what you were asking was possible: you could modify the behavior of some existing system classes like System.String.
You may also take a look at Extension methods to add functionality to an existing class.
You can add functionality, but you cannot change or remove functionality.
You can extend classes by adding extra methods, but you cannot override them because added methods have always lower priority than existing ones.
For more info, check Extension Methods in C# Programming Guide.
Related
Suppose the next class
interface Thing {
void doSomething();
}
public class Test {
public void doWork() {
//Do smart things here
...
doSomethingToThing(index);
// calls to doSomethingToThing might happen in various places across the class.
}
private Thing getThing(int index) {
//find the correct thing
...
return new ThingImpl();
}
private void doSomethingToThing(int index) {
getThing(index).doSomething();
}
}
Intelli-J is telling me that I'm breaking the law of demeter since DoSomethingToThing is using the result of a function and supposedly you can only invoke methods of fields, parameters or the object itself.
Do I really have to do something like this:
public class Test {
//Previous methods
...
private void doSomething(Thing thing) {
thing.doSomething();
}
private void doSomethingToThing(int index) {
doSomething(getThing(index));
}
}
I find that cumbersome. I think the law of demeter is so that one class doesn't know the interior of ANOTHER class, but getThing() is of the same class!
Is this really breaking the law of demeter? is this really improving design?
Thank you.
Technically, this is breaking the Demeter's laws. Though I would contest that private functions should be considered for LoD-F, as supposedly they are not accessible from outside. At the same time, it's not really breaking Demeter's laws, if 'thing' is owned by Test. But in Java, the only way to get to thing may be through a getter, which takes this back to that technicality (no clear separation between getter and action methods).
I would say, do this:
public class Test {
private Thing getThing(int index) {
//find the thing
return thing;
}
private void DoSomethingToThing(Thing thing) {
thing.doSomething();
}
private void DoSomethingToThing(int index) {
DoSomethingToThing(getThing(index));
}
}
Or, probably better, have the caller use thing directly. Which is possible if Test's function is to produce or expose things, rather than being the intermediary to manipulate thing.
IntelliJ is not detecting object instantiation correctly.
Wikipedia (what IDEA links to) describes that you can call objects created in the current context.
That's what I do, but still I get the warning on getMajor():
Version version = Loader.readVersion(inputStream); // Instantiates a new Version
if (version.getMajor() != 2)
throw new IOException("Only major version 2 is supported");
IDEA's inspection offers an option to ignore calls to 'library' methods. In my case, Loader.readVersion() is a library method, however it's located inside the current project (the project must be self-supporting). IDEA thinks it's not a library method.
Since the mechanism of this inspection is inadequate/incomplete/naive (like with MANY of IDEA's inspections btw), the only solution is disabling it and attempt to avoid these situations manually.
I have a base class where all common functions are written. I many classes which override this functions by virtual keyword. Like,
public class Base
{
public virtual void sample()
{
..............
}
}
public class a : Base
{
public override sample()
{
}
}
public class implement
{
public void ToSample()
{
Base baseclass = new Base();
Switch(test)
{
case a: baseclass = a();
break;
case b: baseclass = b();
break;
}
baseclass.sample();
}
}
This perfect code for current situation but now I have more class to be assign in switch case. It is not good practice for adding huge amount of cases so I want something that automatically assign child class.
Is anybody know something to be implement ?
As stated in the comment, you can decouple the implementation by using dependency injection. Note however, that in some cases you have no choice but doing that kind of switch (e.g. when you need to create a class based on a text received in a socket). In such cases the important thing is to always keep the switch statement encapsulated in one method and make your objects rely on it (or, in other words, don't copy-and-paste it everywhere :)). The idea here is too keep your system isolated from a potentially harmful code. Of course that if you add a new class you will have to go and modify that method, however you will only have to do it in one time and in one specific place.
Another approach that I have seen (and sometimes used) is to build a mapping between values an classes. So, if your class-creation switch depends on an integer code, you basically create a mapping between codes and classes. What you are doing here is turning a "static" switch into a dynamic behavior, since you can change the mappings contents at any time and thus alter the way your program behaves. A typical implementation would be something like (sorry for the pseudocode, I'm not familiar with C#):
public class implement
{
public void ToSample()
{
class = this.mapping.valueForKey(test);
Base baseclass = new class();
baseclass.sample();
}
}
Note however that for this example to work you need reflection support, which varies according to the language you are using (again, sorry but I don't know the C# specifics).
Finally, you can also check the creational family of patterns for inspiration regarding object creation issues and some well known forms of solving them.
HTH
I am developing a game, the game have different mode. Easy, Normal, and Difficult.
So, I'm thinking about how to store the game mode. My first idea is using number to represent the difficulty.
Easy = 0 Normal = 1 Difficult = 2
So, my code will have something like this:
switch(gameMode){
case 0:
//easy
break;
case 1:
//normal
break;
case 3:
//difficult
break;
}
But I think it have some problems, if I add a new mode, for example, "Extreme", I need to add case 4... ... it seems not a gd design.
So, I am thinking making a gameMode object, and different gameMode is sub class of the super class gameMode.
The gameMode object is something like this:
class GameMode{
int maxEnemyNumber;
int maxWeaponNumber;
public static GameMode init(){
GameMode gm = GameMode();
gm.maxEnemyNumber = 0;
gm.maxWeaponNumber = 0;
return gm;
}
}
class EasyMode extends GameMode{
public static GameMode init(){
GameMode gm = super.init();
gm.maxEnemyNumber = 10;
gm.maxWeaponNumber = 100;
return gm;
}
}
class NormalMode extends GameMode{
public static GameMode init(){
GameMode gm = super.init();
gm.maxEnemyNumber = 20;
gm.maxWeaponNumber = 80;
return gm;
}
}
But I think it seems too "bulky" to create an object to store gameMode, my "gameMode" only store different variables for game settings.... Is that any simple way to store data only instead of making an Object? thz u.
I think you are trying to represent a table of configuration data. Either put this in a configuration file if you're using a language that supports that, or use literal data in your code.
For instance, you might write this in C:
typedef enum difficulties {
DIFFICULTY_EASY,
DIFFICULTY_MEDIUM,
DIFFICULTY_HARD
} difficulties;
struct {
int max_enemies;
int max_weapons;
} difficulty_settings[] = {
{10, 4},
{20, 5},
{30, 6}
};
And when you want to read a particular setting, for example max_enemies for the easy level, then you can writedifficulty_settings[DIFFICULTY_EASY].max_enemies
It's easy to add more configuration (either more parameters, or more difficulty levels) by extending the table.
The overriding goal you should have here is to centralize the logic for retrieving the values related to different levels. By providing one place where these values are stored, you minimize the number of places within the code you need to change if you add another level, add other values, etc.
A class interface is a good choice for this solution. However, if you have a limited number of configuration options represented by the class, there is no reason you need to use inheritance. You can start out with a single class that encapsulates the logic. If the rest of your code retrieves its settings via the class interface you can later introduce a more complex design, such as subclasses for each mode, if it becomes necessary with limited modifications to the rest of your game.
For example, a first implementation may be something like
enum mode {
MODE_EASY = 0,
MODE_NORMAL = 1,
MODE_DIFFICULT = 2,
};
class gameSettings {
public gameSettings(mode GameMode) {
m_mode = GameMode;
}
public int getMaxWeaponNumber() {
int maxWeaponNumber;
switch(m_mode) {
case EASY_MODE:
maxWeaponNumber = 100;
break;
// Other mode settings.
}
return maxWeaponNumber;
}
// Other game settings....
private mode m_mode;
}
This combines the straightforwardness of a switch() statement with the benefits of a class interface. You can also swap out your switch() statement with a lookup table, as suggested by another poster, or some other mechanism as appropriate for your application.
I don't know java (which is what your examples look like), so I present my ideas in some simple C#.
Here is an idea. Use your game mode as a flag instead. If you start with:
[Flags]
enum GameModes
{
Unknown = 0,
ModeA = 1,
ModeB = 2,
ModeC = 4,
}
Now you have levels 1-7 available.
GameModes Difficulty = GameModes.ModeA | GameModes.ModeB; // difficulty = 3
GameModes Difficulty = GameModes.ModeB; // difficulty = 2
In addition, either method you showed will require you to add more options should levels (modes) change, get added, etc. Have your mode templates read in from XML (or other source of your choice), save the mode data into a serializable class. I don't think you should need base class extended by anything.
Use the switch approach in the constructor of your GameMode class.
Besides some syntax issues, I think you're on the right track. I don't think you have to worry about memory, considering there is probably only one mode at once. This is a form of the strategy pattern. You could extend it so the modes do more. For instance, instead of basically just holding constants, perhaps there could be a generateEnemies method that actually creates a set or list of enemies. This moves more of the strategy into the mode object. Sane defaults in the superclass can help avoid redundant code.
Its difficult to say what kind of refactoring could be done here, as there is too less information about other classes. But you could check the State pattern which encapsulates different behaviours in different state objects. Your approach of extending a base GameMode class is very similar to the state pattern. I think it's better than a switch-case-block... and patterns are reliable ways of doing things, if well applied.
Why do you think the switch is harder to mantain? If you add another mode you will have to add code, no matter what solution you employ.
The only case I can think of where you don't have to add code if you add another mode is if you generate the parameters of the game from the value of gameMode.
For instance: maxenemy = 5 * gameMode;
I think that unless you have very complicated initialisation to perform a switch is more than sufficient. I know, I know, objects and classes are nice and all that jazz, but if you just have to define a few vars and the thing works, investing time in developing a complex game mode class may not be a rewarding solution after all (I mean, how many game modes are you planning to add?).
Make use of the strategy pattern.
In Java terms:
public interface Strategy {
void execute();
}
public class SomeStrategy implements Strategy {
public void execute() {
System.out.println("Some logic.");
}
}
which you use as follows:
Map<String, Strategy> strategies = new HashMap<String, Strategy>();
strategies.put("strategyName1", new SomeStrategy1());
strategies.put("strategyName2", new SomeStrategy2());
strategies.put("strategyName3", new SomeStrategy3());
// ...
strategies.get(s).execute();
To show an example what is this question about:
I have currently a dilemma in PHP project I'm working on. I have in mind a method that will be used by multiple classes (UIs in this case - MVC model), but I'm not sure how to represent such methods in OO design. The first thing that came into my mind was to create a class with static functions that I'd call whenever I need them. However I'm not sure if it's the right thing to do.
To be more precise, I want to work, for example, with time. So I'll need several methods that handle time. I was thinking about creating a Time class where I'd be functions that check whether the time is in correct format etc.
Some might say that I shouldn't use class for this at all, since in PHP I can still use procedural code. But I'm more interested in answer that would enlighten me how to approach such situations in OOP / OOD.
So the actual questions are: How to represent such methods? Is static function approach good enough or should I reconsider anything else?
I would recommend creating a normal class the contains this behavior, and then let that class implement an interface extracted from the class' members.
Whenever you need to call those methods, you inject the interface (not the concrete class) into the consumer. This lets you vary the two independently of each other.
This may sound like more work, but is simply the Strategy design pattern applied.
This will also make it much easier to unit test the code, because the code is more loosely coupled.
Here's an example in C#.
Interface:
public interface ITimeMachine
{
IStopwatch CreateStopwatch();
DateTimeOffset GetNow();
}
Production implementation:
public class RealTimeMachine : ITimeMachine
{
#region ITimeMachine Members
public IStopwatch CreateStopwatch()
{
return new StopwatchAdapter();
}
public DateTimeOffset GetNow()
{
return DateTimeOffset.Now;
}
#endregion
}
and here's a consumer of the interface:
public abstract class PerformanceRecordingSession : IDisposable
{
private readonly IStopwatch watch;
protected PerformanceRecordingSession(ITimeMachine timeMachine)
{
if (timeMachine == null)
{
throw new ArgumentNullException("timeMachine");
}
this.watch = timeMachine.CreateStopwatch();
this.watch.Start();
}
public abstract void Record(long elapsedTicks);
public virtual void StopRecording()
{
this.watch.Stop();
this.Record(this.watch.ElapsedTicks);
}
}
Although you say you want a structure for arbitrary, unrelated functions, you have given an example of a Time class, which has many related functions. So from an OO point of view you would create a Time class and have a static function getCurrentTime(), for example, which returns an instance of this class. Or you could define that the constuctors default behaviour is to return the current time, whichever you like more. Or both.
class DateTime {
public static function getNow() {
return new self();
}
public function __construct() {
$this->setDateTime('now');
}
public function setDateTime($value) {
#...
}
}
But apart from that, there is already a builtin DateTime class in PHP.
Use a class as a namespace. So yes, have a static class.
class Time {
public static function getCurrentTime() {
return time() + 42;
}
}
I don't do PHP, but from an OO point of view, placing these sorts of utility methods as static methods is fine. If they are completely reusable in nature, consider placing them in a utils class.
For example, do you use accessors and mutators within your method definitions or just access the data directly? Sometimes, all the time or when in Rome?
Always try to use accessors, even inside the class. The only time you would want to access state directly and not through the public interface is if for some reason you needed to bypass the validation or other logic contained in the accessor method.
Now if you find yourself in the situation where you do need to bypass that logic you ought to step back and ask yourself whether or not this need betrays a flaw in your design.
Edit: Read Automatic vs Explicit Properties by Eric Lippert in which he delves into this very issue and explains things very clearly. It is about C# specifically but the OOP theory is universal and solid.
Here is an excerpt:
If the reason that motivated the
change from automatically implemented
property to explicitly implemented
property was to change the semantics
of the property then you should
evaluate whether the desired semantics
when accessing the property from
within the class are identical to or
different from the desired semantics
when accessing the property from
outside the class.
If the result of that investigation is
“from within the class, the desired
semantics of accessing this property
are different from the desired
semantics of accessing the property
from the outside”, then your edit has
introduced a bug. You should fix the
bug. If they are the same, then your
edit has not introduced a bug; keep
the implementation the same.
In general, I prefer accessors/mutators. That way, I can change the internal implementation of a class, while the class functions in the same way to an external user (or preexisting code that I dont want to break).
The accessors are designed so that you can add property specific logic. Such as
int Degrees
{
set
{
_degrees = value % 360;
}
}
So, you would always want to access that field through the getter and setter, and that way you can always be certain that the value will never get greater than 360.
If, as Andrew mentioned, you need to skip the validation, then it's quite possible that there is a flaw in the design of the function, or in the design of the validation.
Accessors and Mutators are designed to ensure consistency of your data, so even within your class you should always strive to make sure that there's no possible way of injecting unvalidated data into those fields.
EDIT
See this question as well:
OO Design: Do you use public properties or private fields internally?
I don't tend to share with the outside world the 'innards' of my classes and so my internal needs for data (the private method stuff) tends to not do the same sort of stuff that my public interface does, typically.
It is pretty uncommon that I'll write an accessor/mutator that a private method will call, but I suspect I'm in the minority here. Maybe I should do more of this, but I don't tend to.
Anyway, that's my [patina covered] two cents.
I will often start with private auto properties, then refactor if necessary. I'll refactor to a property with a backing field, then replace the backing field with the "real" store, for instance Session or ViewState for an ASP.NET application.
From:
private int[] Property { get; set; }
to
private int[] _property;
private int[] Property
{
get { return _property; }
set { _property = value; }
}
to
private int[] _property;
private int[] Property
{
get
{
if (_property == null)
{
_property = new int[8];
}
return _property;
}
set { _property = value; }
}
to
private int[] Property
{
get
{
if (ViewState["PropertyKey"] == null)
{
ViewState["PropertyKey"] = new int[8];
}
return (int[]) ViewState["PropertyKey"];
}
set { ViewState["PropertyKey"] = value; }
}
Of course, I use ReSharper, so this takes less time to do than to post about.