Is it possible to intercept all the methods called in a application? I'd like to do something with them, and then let them execute. I tried to override this behaviour in Object.metaClass.invokeMethod, but it doesn't seem to work.
Is this doable?
Have you looked at Groovy AOP? There's very little documentation, but it allows you to define pointcuts and advice in a conceptually similar way as for AspectJ. Have a look at the unit tests for some more examples
The example below will match all calls to all woven types and apply the advice before proceeding:
// aspect MyAspect
class MyAspect {
static aspect = {
//match all calls to all calls to all types in all packages
def pc = pcall("*.*.*")
//apply around advice to the matched calls
around(pc) { ctx ->
println ctx.args[0]
println ctx.args.length
return proceed(ctx.args)
}
}
}
// class T
class T {
def test() {
println "hello"
}
}
// Script starts here
weave MyAspect.class
new T().test()
unweave MyAspect.class
First of all, overriding Object.metaClass.invokeMethod doesn't work because when Groovy tries to resolve a method call for a type X, it checks the metaClass of X, but not the metaClass of its parent class(es). For example, the following code will print "method intValue intercepted"
Integer.metaClass.invokeMethod = {def name, def args ->
System.out.println("method $name intercepted")
}
6.intValue()
// Reset the metaClass
Integer.metaClass = null
But this code will not:
Object.metaClass.invokeMethod = {def name, def args ->
System.out.println("method $name intercepted")
}
6.intValue()
// Reset the metaClass
Object.metaClass = null
Your question was "Is it possible to intercept all the methods called in a application?", but could you be a bit more precise about whether you want to:
Intercept calls to Groovy methods, Java methods, or both
Intercept calls to only your Groovy/Java methods or also intercept calls to Groovy/Java library classes
For example, if you only want to intercept calls to your Groovy classes, you could change your classes to implement GroovyInterceptable. This ensures that invokeMethod() is invoked for every method called on those classes. If the nature of the interception (i.e. the stuff you want to do before/after invoking the called method) is the same for all classes, you could define invokeMethod() in a separate class and use #Mixin to apply it to all your classes.
Alternatively, if you also want to intercept calls to Java classes, you should check out the DelegatingMetaClass.
Related
I am using groovy 2.3.8
I am trying to figure out how method calls work in groovy. Specifically if we have a Java class hierarchy each having a metaClass like below
class A {
}
A.metaClass.hello = {
"hello superclass"
}
class B extends A {
}
B.metaClass.hello = {
"hello subclass"
}
If I use new B().hello() I get hello subclass. If I remove meta class of B then I get hello superclass.
Based on changing the above example I think groovy goes in the below sequence to find which method to call
method-in-subclass's-metaclass ?: subclass-metho ?: method-in-superclass's metaclass ?: method-in-superclass
So how does groovy lookup which method to call?
Well, the hierarchy is the expected object oriented programming method overloading, which is what you witnessed. What differs is the dispatching. Instead of starting with a method lookup in instance's class, it begins with the MOP (meta object protocol).
In layman's terms, because the MOP is programmable, so is the way methods are invoked :)
How it works
The following diagram from Groovy's documentation shows how methods are looked up.
What's not clear in the diagram is that there's an instance metaclass as well, and it comes before the class's metaclass.
Something that may help is looking at an object's or class's .metaClass.methods Methods added through inheritance, traits, metaclass, etc are listed in a flat list. The inheritance hierarchy is flattened. .metaClass.metaMethods on the other hand seems to contain methods added via the GDK. From the list I could not tell method precedence :(
Based on observation, the rule seems to be this: the last MetaClass standing wins.
class A { }
class B extends A { }
A.metaClass.hello = {
"hello superclass"
}
B.metaClass.hello = {
"hello subclass"
}
def b = new B()
assert b.hello() == "hello subclass"
b.metaClass = A.metaClass
assert b.hello() == "hello superclass"
I have a lot of code using and expecting java.util.Date which I would like to migrate to org.joda.time.LocalDate. The problem is that most of the code is dynamically typed.
So I wonder if there is any groovy way to intercept the ClassCastException, do the conversion at runtime (instead of letting the exception bubble up) and log the operation (so I could fix the code).
Example:
import org.joda.time.LocalDate
def someMethod(date) {
println date.year()
}
// this call is ok
someMethod(new LocalDate())
// this call raises an exception
someMethod(new Date())
I don't want to modify the code above, like surrounding the second call with a try-catch and recalling with the right type. I wanted a way to do this globally.
Idea 1:
You can benefit from Groovy multimethods: you write a method overload for the JodaTime's and let the java.util.Date method be chosen when it is a dynamic method parameter:
def someMethod(date) {
println date.year()
}
My advice is to write in this class a overload of these methods doing the conversion:
def someMethod(LocalDate date) {
someMethod date.convertToJavaDate()
}
If you don't have access to that code, you can use metaprogramming.
Idea 2:
If you have a huge library, my guess is that you are better mixing JodaTime class. You will get a power benefit from duck typing. You will need to know which members from java.util.Date are called and reroute them to Joda's own methods:
class LocalDateMixin {
def year() { this[Calendar.YEAR] }
}
LocalDate.metaClass.mixin LocalDateMixin
You can also apply the metaprogramming using extension methods.
UPDATE
I have to apologize for confusing the readers. After I got totally lost in the code, I reverted all my changes from Mercurial repo, carefully applied the same logic as before -- and it worked. The answers below helped me understand the (new to me) concept better, and for that I gave them upvotes.
Bottom line: if a call to a missing method happens within a closure, and resolution set to DELEGATE_FIRST, methodMissing() will be called on the delegate. If it doesn't -- check you own code, there is a typo somewhere.
Thanks a lot!
Edit:
OK, now that you've clarified what your are doing (somewhat ;--))
Another approach (one that I use for DSLs) is to parse your closure group to map via a ClosureToMap utility like this:
// converts given closure to map method => value pairs (1-d, if you need nested, ask)
class ClosureToMap {
Map map = [:]
ClosureToMap(Closure c) {
c.delegate = this
c.resolveStrategy = Closure.DELEGATE_FIRST
c.each{"$it"()}
}
def methodMissing(String name, args) {
if(!args.size()) return
map[name] = args[0]
}
def propertyMissing(String name) { name }
}
// Pass your closure to the utility and access the generated map
Map map = new ClosureToMap(your-closure-here)?.map
Now you can iterate through the map, perhaps adding methods to applicable MCL instance. For example, some of my domains have dynamic finders like:
def finders = {
userStatusPaid = { Boolean active = true->
eq {
active "$active"
paid true
}
}
}
I create a map using the ClosureToMap utility, and then iterate through, adding map keys (methods, like "userStatus") and values (in this case, closure "eq") to domain instance MCL, delegating the closure to our ORM, like so:
def injectFinders(Object instance) {
if(instance.hasProperty('finders')) {
Map m = ClosureToMap.new(instance.finders).map
m?.each{ String method, Closure cl->
cl.delegate = instance.orm
cl.resolveStrategy = Closure.DELEGATE_FIRST
instance.orm.metaClass."$method" = cl
}
}
}
In this way in controller scope I can do, say:
def actives = Orders.userStatusPaid()
and "eq" closure will delegate to the ORM and not domain Orders where an MME would occur.
Play around with it, hopefully I've given you some ideas for how to solve the problem. In Groovy, if you can't do it one way, try another ;--)
Good luck!
Original:
Your missingMethod is defined on string metaclass; in order for it to be invoked, you need "someString".foo()
If you simply call foo() by itself within your closure it will fail, regardless of delegation strategy used; i.e. if you don't use the (String) delegate, good luck. Case in point, do "".foo() and it works.
I don't fully understand the issue either, why will you not have access to the closure's delegate? You are setting the closure's delegate and will invoke the closure, which means you will have access to the delegate within the closure itself (and can just delegate.foo())
nope, you will not catch a missing method and redirect it to the delegate with metaclass magic.
the closure delegate is the chance to capture those calls and adapt them to the backing domain.
that means...
you should create your own delegate with the methods required by the dsl.
do not try to force a class to do delegate work if it's not designed for the task, or the code will get really messy in not time.
keep everything dsl related in a set of specially designed delegate classes and everything will suddenly become ridiculously simple and clear.
I need to mock a static method. I'm using the EMC approach described at Mocking static methods using groovy. Like this
TestDaemon.metaClass.'static'.newDownloadManager = {downloadManager}
The method newDownloadManager has no parameters and for some reason it is not replaced. The original code is called. In debug mode I can see that the closure that I define has a parameter. May be that's the reason? How can I define a closure without parameters? Or how can I mock a static method with no parameters?
Meta class changes aren't visible to Java code. Groovy can't help you to mock a static method that gets called from Java code. You will have to use something like JMockit instead (or refactor the code under test).
A closure written like that has an implicit parameter. Write the closure with { -> } syntax. Example:
x = { println "foo" }
y = { -> println "foo" }
assert x.parameterTypes as List == [Object]
assert y.parameterTypes as List == []
I have a Groovy class similar to
class MyClass {
Foo foo
}
Under certain circumstances I don't want to initialize foo and want to stub out all the calls to it. Any methods that return a value should do nothing. I could do it like this:
Foo.metaClass.method1 = {param -> }
Foo.metaClass.method2 = { -> }
Foo.metaClass.method3 = {param1, param2 -> }
While this will work, it has a couple of problems
Tedious and long-winded, particularly if Foo has a lot of methods
This will stub out calls to any instance of Foo (not just foo)
Although Groovy provides a StubFor class, if I do this:
this.foo = new groovy.mock.interceptor.StubFor(Foo)
I get a ClassCastException at runtime. Although this would work if I could redefine foo as:
def foo
But for reasons I won't go into here, I can't do that.
Thanks,
Don
I found the solution:
this.foo = {Object[] args -> println "I don't do anything"} as Foo
You need to pass the Foo instance to your MyClass object. In the test, pass in a stub implementation. In the real program pass in a real implementation.
You don't need any special framework to write the stub. If you're only stubbing queries to Foo and are not interested in expecting commands to it, then it's easier to write a stub implementation by hand. A mock object library is overkill and will confuse later readers of the code who will expect the test to include expectations.
You could make a stub with a map as well, making it possible to stub multiple methods, like so:
def stubbedList = [empty : { false }, get: {index -> "always return this"}] as ArrayList
If you stub out a class (like in the ArrayList example above), then the methods that you do not override are kept as they are implemented in the class. You can also stub out interfaces obviously.
If you are in need of mocking capabilities (i.e verifying behaviour in a test, like counting the number of calls to a given method), I would strongly recommend checking out the Spock test framework. It has really great support for more advanced stubbing and mocking.
I'm guessing this is for test purposes only.
The line:
Foo foo
creates a getter / setter pair, so you can inject a mock in your test cases by doing this:
new MyClass().foo = new MockFoo()
If you don't want to create the mock for yourself, try a mocking library. I recommend using mockito. With this small library you can do something like this:
import static org.mockito.Mockito.*;
new MyClass().foo = mock(Foo.class);