Recently I've found my self to implement code following a pattern like:
public class SomeClass()
{
private T Execute<T>(Func<T> function)
{
// Do some common stuff for every function like logging and try-catch
function();
}
public Type1 Command1()
{
Execute<Type1>(() => funcForCommand1);
}
public Type2 Command2()
{
Execute<Type2>(() => funcForCommand2);
}
}
Is this just a functio开发者_JAVA百科nal approach on the CommandPattern? Depending on the situation I've different version of this? You could probably achieve exactly the same thing by letting funcForCommandX
inherit from somekind of ICommand
that defines the Execute
function, but I like my way in many situations better since most of the time the commands are only used in one location in the code and not needs to be exposed to the rest of the code. Of course you should implement the real command pattern if it is used in more locations in the code.
It is up to your needs. That is all I can say.
But I would like to state that this is not a Command Pattern
but may be a method delegation.
Command Pattern
focuses mainly on execution of a method/task. Following are standard behaviors that can be expected from a Command
in this pattern:
- Undo/Redo
- Transactions
- Composite Command Execution
- Macros
When you wrap a method/task implementation in a Command
, you can provide implementations on what to do to reverse/undo what has been done. Provide default implementation on attaching current execution to transactions, macro recording, thread-safe execution, etc.
With your approach you don't have that. And it's not easily doable until you wrap each task/method in a Command
wrapper and provide above mentioned behaviors.
Take a look at Wikipedia
article for further details on a Command Pattern
.
精彩评论