All Downloads are FREE. Search and download functionalities are using the official Maven repository.

groovy.lang.Closure Maven / Gradle / Ivy

There is a newer version: 1.0.1457
Show newest version
/*
 *  Licensed to the Apache Software Foundation (ASF) under one
 *  or more contributor license agreements.  See the NOTICE file
 *  distributed with this work for additional information
 *  regarding copyright ownership.  The ASF licenses this file
 *  to you under the Apache License, Version 2.0 (the
 *  "License"); you may not use this file except in compliance
 *  with the License.  You may obtain a copy of the License at
 *
 *    http://www.apache.org/licenses/LICENSE-2.0
 *
 *  Unless required by applicable law or agreed to in writing,
 *  software distributed under the License is distributed on an
 *  "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
 *  KIND, either express or implied.  See the License for the
 *  specific language governing permissions and limitations
 *  under the License.
 */
package groovy.lang;

import org.codehaus.groovy.reflection.ReflectionCache;
import org.codehaus.groovy.reflection.stdclasses.CachedClosureClass;
import org.codehaus.groovy.runtime.ComposedClosure;
import org.codehaus.groovy.runtime.CurriedClosure;
import org.codehaus.groovy.runtime.ExceptionUtils;
import org.codehaus.groovy.runtime.InvokerHelper;
import org.codehaus.groovy.runtime.InvokerInvocationException;
import org.codehaus.groovy.runtime.callsite.BooleanClosureWrapper;
import org.codehaus.groovy.runtime.memoize.LRUCache;
import org.codehaus.groovy.runtime.memoize.Memoize;
import org.codehaus.groovy.runtime.memoize.UnlimitedConcurrentCache;

import java.io.IOException;
import java.io.Serializable;
import java.io.StringWriter;
import java.io.Writer;

/**
 * Represents any closure object in Groovy.
 * 

* Groovy allows instances of Closures to be called in a * short form. For example: *

 * def a = 1
 * def c = { a }
 * assert c() == 1
 * 
* To be able to use a Closure in this way with your own * subclass, you need to provide a doCall method with any * signature you want to. This ensures that * {@link #getMaximumNumberOfParameters()} and * {@link #getParameterTypes()} will work too without any * additional code. If no doCall method is provided a * closure must be used in its long form like *
 * def a = 1
 * def c = {a}
 * assert c.call() == 1
 * 
* * @author James Strachan * @author John Wilson * @author Jochen Theodorou * @author Graeme Rocher * @author Paul King */ public abstract class Closure extends GroovyObjectSupport implements Cloneable, Runnable, GroovyCallable, Serializable { /** * With this resolveStrategy set the closure will attempt to resolve property references and methods to the * owner first, then the delegate (this is the default strategy). * * For example the following code: *
     * class Test {
     *     def x = 30
     *     def y = 40
     *
     *     def run() {
     *         def data = [ x: 10, y: 20 ]
     *         def cl = { y = x + y }
     *         cl.delegate = data
     *         cl()
     *         assert x == 30
     *         assert y == 70
     *         assert data == [x:10, y:20]
     *     }
     * }
     *
     * new Test().run()
     * 
* Will succeed, because the x and y fields declared in the Test class shadow the variables in the delegate.

* Note that local variables are always looked up first, independently of the resolution strategy. */ public static final int OWNER_FIRST = 0; /** * With this resolveStrategy set the closure will attempt to resolve property references and methods to the * delegate first then the owner. * * For example the following code: *

     * class Test {
     *     def x = 30
     *     def y = 40
     *
     *     def run() {
     *         def data = [ x: 10, y: 20 ]
     *         def cl = { y = x + y }
     *         cl.delegate = data
     *         cl.resolveStrategy = Closure.DELEGATE_FIRST
     *         cl()
     *         assert x == 30
     *         assert y == 40
     *         assert data == [x:10, y:30]
     *     }
     * }
     *
     * new Test().run()
     * 
* This will succeed, because the x and y variables declared in the delegate shadow the fields in the owner class.

* Note that local variables are always looked up first, independently of the resolution strategy. */ public static final int DELEGATE_FIRST = 1; /** * With this resolveStrategy set the closure will resolve property references and methods to the owner only * and not call the delegate at all. For example the following code : * *

     * class Test {
     *     def x = 30
     *     def y = 40
     *
     *     def run() {
     *         def data = [ x: 10, y: 20, z: 30 ]
     *         def cl = { y = x + y + z }
     *         cl.delegate = data
     *         cl.resolveStrategy = Closure.OWNER_ONLY
     *         cl()
     *         println x
     *         println y
     *         println data
     *     }
     * }
     *
     * new Test().run()
     * 
* * will throw "No such property: z" error because even if the z variable is declared in the delegate, no * lookup is made.

* Note that local variables are always looked up first, independently of the resolution strategy. */ public static final int OWNER_ONLY = 2; /** * With this resolveStrategy set the closure will resolve property references and methods to the delegate * only and entirely bypass the owner. For example the following code : * *

     * class Test {
     *     def x = 30
     *     def y = 40
     *     def z = 50
     *
     *     def run() {
     *         def data = [ x: 10, y: 20 ]
     *         def cl = { y = x + y + z }
     *         cl.delegate = data
     *         cl.resolveStrategy = Closure.DELEGATE_ONLY
     *         cl()
     *         println x
     *         println y
     *         println data
     *     }
     * }
     *
     * new Test().run()
     * 
* * will throw an error because even if the owner declares a "z" field, the resolution strategy will bypass * lookup in the owner.

* Note that local variables are always looked up first, independently of the resolution strategy. */ public static final int DELEGATE_ONLY = 3; /** * With this resolveStrategy set the closure will resolve property references to itself and go * through the usual MetaClass look-up process. This means that properties and methods are neither resolved * from the owner nor the delegate, but only on the closure object itself. This allows the developer to * override getProperty using ExpandoMetaClass of the closure itself.

* Note that local variables are always looked up first, independently of the resolution strategy. */ public static final int TO_SELF = 4; public static final int DONE = 1, SKIP = 2; private static final Object[] EMPTY_OBJECT_ARRAY = {}; public static final Closure IDENTITY = new Closure(null) { private static final long serialVersionUID = 730973623329943963L; public Object doCall(Object args) { return args; } }; private Object delegate; private Object owner; private Object thisObject; private int resolveStrategy = OWNER_FIRST; private int directive; protected Class[] parameterTypes; protected int maximumNumberOfParameters; private static final long serialVersionUID = 4368710879820278874L; private BooleanClosureWrapper bcw; public Closure(Object owner, Object thisObject) { this.owner = owner; this.delegate = owner; this.thisObject = thisObject; final CachedClosureClass cachedClass = (CachedClosureClass) ReflectionCache.getCachedClass(getClass()); parameterTypes = cachedClass.getParameterTypes(); maximumNumberOfParameters = cachedClass.getMaximumNumberOfParameters(); } /** * Constructor used when the "this" object for the Closure is null. * This is rarely the case in normal Groovy usage. * * @param owner the Closure owner */ public Closure(Object owner) { this(owner, null); } /** * Sets the strategy which the closure uses to resolve property references and methods. * The default is Closure.OWNER_FIRST * * @param resolveStrategy The resolve strategy to set * * @see groovy.lang.Closure#DELEGATE_FIRST * @see groovy.lang.Closure#DELEGATE_ONLY * @see groovy.lang.Closure#OWNER_FIRST * @see groovy.lang.Closure#OWNER_ONLY * @see groovy.lang.Closure#TO_SELF */ public void setResolveStrategy(int resolveStrategy) { this.resolveStrategy = resolveStrategy; } /** * Gets the strategy which the closure uses to resolve methods and properties * * @return The resolve strategy * * @see groovy.lang.Closure#DELEGATE_FIRST * @see groovy.lang.Closure#DELEGATE_ONLY * @see groovy.lang.Closure#OWNER_FIRST * @see groovy.lang.Closure#OWNER_ONLY * @see groovy.lang.Closure#TO_SELF */ public int getResolveStrategy() { return resolveStrategy; } public Object getThisObject(){ return thisObject; } public Object getProperty(final String property) { if ("delegate".equals(property)) { return getDelegate(); } else if ("owner".equals(property)) { return getOwner(); } else if ("maximumNumberOfParameters".equals(property)) { return getMaximumNumberOfParameters(); } else if ("parameterTypes".equals(property)) { return getParameterTypes(); } else if ("metaClass".equals(property)) { return getMetaClass(); } else if ("class".equals(property)) { return getClass(); } else if ("directive".equals(property)) { return getDirective(); } else if ("resolveStrategy".equals(property)) { return getResolveStrategy(); } else if ("thisObject".equals(property)) { return getThisObject(); } else { switch(resolveStrategy) { case DELEGATE_FIRST: return getPropertyDelegateFirst(property); case DELEGATE_ONLY: return InvokerHelper.getProperty(this.delegate, property); case OWNER_ONLY: return InvokerHelper.getProperty(this.owner, property); case TO_SELF: return super.getProperty(property); default: return getPropertyOwnerFirst(property); } } } private Object getPropertyDelegateFirst(String property) { if (delegate == null) return getPropertyOwnerFirst(property); return getPropertyTryThese(property, this.delegate, this.owner); } private Object getPropertyOwnerFirst(String property) { return getPropertyTryThese(property, this.owner, this.delegate); } private Object getPropertyTryThese(String property, Object firstTry, Object secondTry) { try { // let's try getting the property on the first object return InvokerHelper.getProperty(firstTry, property); } catch (MissingPropertyException e1) { if (secondTry != null && firstTry != this && firstTry != secondTry) { try { // let's try getting the property on the second object return InvokerHelper.getProperty(secondTry, property); } catch (GroovyRuntimeException e2) { // ignore, we'll throw e1 } } throw e1; } catch (MissingFieldException e2) { // see GROOVY-5875 if (secondTry != null && firstTry != this && firstTry != secondTry) { try { // let's try getting the property on the second object return InvokerHelper.getProperty(secondTry, property); } catch (GroovyRuntimeException e3) { // ignore, we'll throw e2 } } throw e2; } } public void setProperty(String property, Object newValue) { if ("delegate".equals(property)) { setDelegate(newValue); } else if ("metaClass".equals(property)) { setMetaClass((MetaClass) newValue); } else if ("resolveStrategy".equals(property)) { setResolveStrategy(((Number) newValue).intValue()); } else if ("directive".equals(property)) { setDirective(((Number) newValue).intValue()); } else { switch(resolveStrategy) { case DELEGATE_FIRST: setPropertyDelegateFirst(property, newValue); break; case DELEGATE_ONLY: InvokerHelper.setProperty(this.delegate, property, newValue); break; case OWNER_ONLY: InvokerHelper.setProperty(this.owner, property, newValue); break; case TO_SELF: super.setProperty(property, newValue); break; default: setPropertyOwnerFirst(property, newValue); } } } private void setPropertyDelegateFirst(String property, Object newValue) { if (delegate == null) setPropertyOwnerFirst(property, newValue); else setPropertyTryThese(property, newValue, this.delegate, this.owner); } private void setPropertyOwnerFirst(String property, Object newValue) { setPropertyTryThese(property, newValue, this.owner, this.delegate); } private void setPropertyTryThese(String property, Object newValue, Object firstTry, Object secondTry) { try { // let's try setting the property on the first object InvokerHelper.setProperty(firstTry, property, newValue); } catch (GroovyRuntimeException e1) { if (firstTry != null && firstTry != this && firstTry != secondTry) { try { // let's try setting the property on the second object InvokerHelper.setProperty(secondTry, property, newValue); return; } catch (GroovyRuntimeException e2) { // ignore, we'll throw e1 } } throw e1; } } public boolean isCase(Object candidate){ if (bcw==null) { bcw = new BooleanClosureWrapper(this); } return bcw.call(candidate); } /** * Invokes the closure without any parameters, returning any value if applicable. * * @return the value if applicable or null if there is no return statement in the closure */ public V call() { final Object[] NOARGS = EMPTY_OBJECT_ARRAY; return call(NOARGS); } @SuppressWarnings("unchecked") public V call(Object... args) { try { return (V) getMetaClass().invokeMethod(this,"doCall",args); } catch (InvokerInvocationException e) { ExceptionUtils.sneakyThrow(e.getCause()); return null; // unreachable statement } catch (Exception e) { return (V) throwRuntimeException(e); } } /** * Invokes the closure, returning any value if applicable. * * @param arguments could be a single value or a List of values * @return the value if applicable or null if there is no return statement in the closure */ public V call(final Object arguments) { return call(new Object[]{arguments}); } protected static Object throwRuntimeException(Throwable throwable) { if (throwable instanceof RuntimeException) { throw (RuntimeException) throwable; } else { throw new GroovyRuntimeException(throwable.getMessage(), throwable); } } /** * @return the owner Object to which method calls will go which is * typically the outer class when the closure is constructed */ public Object getOwner() { return this.owner; } /** * @return the delegate Object to which method calls will go which is * typically the outer class when the closure is constructed */ public Object getDelegate() { return this.delegate; } /** * Allows the delegate to be changed such as when performing markup building * * @param delegate the new delegate */ public void setDelegate(Object delegate) { this.delegate = delegate; } /** * @return the parameter types of the longest doCall method * of this closure */ public Class[] getParameterTypes() { return parameterTypes; } /** * @return the maximum number of parameters a doCall method * of this closure can take */ public int getMaximumNumberOfParameters() { return maximumNumberOfParameters; } /** * @return a version of this closure which implements Writable. Note that * the returned Writable also overrides {@link #toString()} in order * to allow rendering the result directly to a String. */ public Closure asWritable() { return new WritableClosure(); } /* (non-Javadoc) * @see java.lang.Runnable#run() */ public void run() { call(); } /** * Support for Closure currying. *

* Typical usage: *

     * def multiply = { a, b -> a * b }
     * def doubler = multiply.curry(2)
     * assert doubler(4) == 8
     * 
* Note: special treatment is given to Closure vararg-style capability. * If you curry a vararg parameter, you don't consume the entire vararg array * but instead the first parameter of the vararg array as the following example shows: *
     * def a = { one, two, Object[] others -> one + two + others.sum() }
     * assert a.parameterTypes.name == ['java.lang.Object', 'java.lang.Object', '[Ljava.lang.Object;']
     * assert a(1,2,3,4) == 10
     * def b = a.curry(1)
     * assert b.parameterTypes.name == ['java.lang.Object', '[Ljava.lang.Object;']
     * assert b(2,3,4) == 10
     * def c = b.curry(2)
     * assert c.parameterTypes.name == ['[Ljava.lang.Object;']
     * assert c(3,4) == 10
     * def d = c.curry(3)
     * assert d.parameterTypes.name == ['[Ljava.lang.Object;']
     * assert d(4) == 10
     * def e = d.curry(4)
     * assert e.parameterTypes.name == ['[Ljava.lang.Object;']
     * assert e() == 10
     * assert e(5) == 15
     * 
* * * @param arguments the arguments to bind * @return the new closure with its arguments bound */ public Closure curry(final Object... arguments) { return new CurriedClosure(this, arguments); } /** * Support for Closure currying. * * @param argument the argument to bind * @return the new closure with the argument bound * @see #curry(Object...) */ public Closure curry(final Object argument) { return curry(new Object[]{argument}); } /** * Support for Closure "right" currying. * Parameters are supplied on the right rather than left as per the normal curry() method. * Typical usage: *
     * def divide = { a, b -> a / b }
     * def halver = divide.rcurry(2)
     * assert halver(8) == 4
     * 
* * The position of the curried parameters will be calculated lazily, for example, * if two overloaded doCall methods are available, the supplied arguments plus the * curried arguments will be concatenated and the result used for method selection. * * @param arguments the arguments to bind * @return the new closure with its arguments bound * @see #curry(Object...) */ public Closure rcurry(final Object... arguments) { return new CurriedClosure(-arguments.length, this, arguments); } /** * Support for Closure "right" currying. * * @param argument the argument to bind * @return the new closure with the argument bound * @see #rcurry(Object...) */ public Closure rcurry(final Object argument) { return rcurry(new Object[]{argument}); } /** * Support for Closure currying at a given index. * Parameters are supplied from index position "n". * Typical usage: *
     * def caseInsensitive = { a, b -> a.toLowerCase() <=> b.toLowerCase() } as Comparator
     * def caseSensitive = { a, b -> a <=> b } as Comparator
     * def animals1 = ['ant', 'dog', 'BEE']
     * def animals2 = animals1 + ['Cat']
     * // curry middle param of this utility method:
     * // Collections#binarySearch(List list, Object key, Comparator c)
     * def catSearcher = Collections.&binarySearch.ncurry(1, "cat")
     * [[animals1, animals2], [caseInsensitive, caseSensitive]].combinations().each{ a, c ->
     *   def idx = catSearcher(a.sort(c), c)
     *   print a.sort(c).toString().padRight(22)
     *   if (idx < 0) println "Not found but would belong in position ${-idx - 1}"
     *   else println "Found at index $idx"
     * }
     * // =>
     * // [ant, BEE, dog]       Not found but would belong in position 2
     * // [ant, BEE, Cat, dog]  Found at index 2
     * // [BEE, ant, dog]       Not found but would belong in position 2
     * // [BEE, Cat, ant, dog]  Not found but would belong in position 3
     * 
* * The position of the curried parameters will be calculated eagerly * and implies all arguments prior to the specified n index are supplied. * Default parameter values prior to the n index will not be available. * * @param n the index from which to bind parameters (may be -ve in which case it will be normalized) * @param arguments the arguments to bind * @return the new closure with its arguments bound * @see #curry(Object...) */ public Closure ncurry(int n, final Object... arguments) { return new CurriedClosure(n, this, arguments); } /** * Support for Closure currying at a given index. * * @param argument the argument to bind * @return the new closure with the argument bound * @see #ncurry(int, Object...) */ public Closure ncurry(int n, final Object argument) { return ncurry(n, new Object[]{argument}); } /** * Support for Closure forward composition. *

* Typical usage: *

     * def times2 = { a -> a * 2 }
     * def add3 = { a -> a + 3 }
     * def timesThenAdd = times2 >> add3
     * // equivalent: timesThenAdd = { a -> add3(times2(a)) }
     * assert timesThenAdd(3) == 9
     * 
* * @param other the Closure to compose with the current Closure * @return the new composed Closure */ public Closure rightShift(final Closure other) { return new ComposedClosure(this, other); } /** * Support for Closure reverse composition. *

* Typical usage: *

     * def times2 = { a -> a * 2 }
     * def add3 = { a -> a + 3 }
     * def addThenTimes = times2 << add3
     * // equivalent: addThenTimes = { a -> times2(add3(a)) }
     * assert addThenTimes(3) == 12
     * 
* * @param other the Closure to compose with the current Closure * @return the new composed Closure */ public Closure leftShift(final Closure other) { return new ComposedClosure(other, this); } /* * * Alias for calling a Closure for non-closure arguments. *

* Typical usage: *

     * def times2 = { a -> a * 2 }
     * def add3 = { a -> a * 3 }
     * assert add3 << times2 << 3 == 9
     * 
* * @param arg the argument to call the closure with * @return the result of calling the Closure */ public V leftShift(final Object arg) { return call(arg); } /** * Creates a caching variant of the closure. * Whenever the closure is called, the mapping between the parameters and the return value is preserved in cache * making subsequent calls with the same arguments fast. * This variant will keep all cached values forever, i.e. till the closure gets garbage-collected. * The returned function can be safely used concurrently from multiple threads, however, the implementation * values high average-scenario performance and so concurrent calls on the memoized function with identical argument values * may not necessarily be able to benefit from each other's cached return value. With this having been mentioned, * the performance trade-off still makes concurrent use of memoized functions safe and highly recommended. * * The cache gets garbage-collected together with the memoized closure. * * @return A new closure forwarding to the original one while caching the results */ public Closure memoize() { return Memoize.buildMemoizeFunction(new UnlimitedConcurrentCache(), this); } /** * Creates a caching variant of the closure with upper limit on the cache size. * Whenever the closure is called, the mapping between the parameters and the return value is preserved in cache * making subsequent calls with the same arguments fast. * This variant will keep all values until the upper size limit is reached. Then the values in the cache start rotating * using the LRU (Last Recently Used) strategy. * The returned function can be safely used concurrently from multiple threads, however, the implementation * values high average-scenario performance and so concurrent calls on the memoized function with identical argument values * may not necessarily be able to benefit from each other's cached return value. With this having been mentioned, * the performance trade-off still makes concurrent use of memoized functions safe and highly recommended. * * The cache gets garbage-collected together with the memoized closure. * * @param maxCacheSize The maximum size the cache can grow to * @return A new function forwarding to the original one while caching the results */ public Closure memoizeAtMost(final int maxCacheSize) { if (maxCacheSize < 0) throw new IllegalArgumentException("A non-negative number is required as the maxCacheSize parameter for memoizeAtMost."); return Memoize.buildMemoizeFunction(new LRUCache(maxCacheSize), this); } /** * Creates a caching variant of the closure with automatic cache size adjustment and lower limit * on the cache size. * Whenever the closure is called, the mapping between the parameters and the return value is preserved in cache * making subsequent calls with the same arguments fast. * This variant allows the garbage collector to release entries from the cache and at the same time allows * the user to specify how many entries should be protected from the eventual gc-initiated eviction. * Cached entries exceeding the specified preservation threshold are made available for eviction based on * the LRU (Last Recently Used) strategy. * Given the non-deterministic nature of garbage collector, the actual cache size may grow well beyond the limits * set by the user if memory is plentiful. * The returned function can be safely used concurrently from multiple threads, however, the implementation * values high average-scenario performance and so concurrent calls on the memoized function with identical argument values * may not necessarily be able to benefit from each other's cached return value. Also the protectedCacheSize parameter * might not be respected accurately in such scenarios for some periods of time. With this having been mentioned, * the performance trade-off still makes concurrent use of memoized functions safe and highly recommended. * * The cache gets garbage-collected together with the memoized closure. * @param protectedCacheSize Number of cached return values to protect from garbage collection * @return A new function forwarding to the original one while caching the results */ public Closure memoizeAtLeast(final int protectedCacheSize) { if (protectedCacheSize < 0) throw new IllegalArgumentException("A non-negative number is required as the protectedCacheSize parameter for memoizeAtLeast."); return Memoize.buildSoftReferenceMemoizeFunction(protectedCacheSize, new UnlimitedConcurrentCache(), this); } /** * Creates a caching variant of the closure with automatic cache size adjustment and lower and upper limits * on the cache size. * Whenever the closure is called, the mapping between the parameters and the return value is preserved in cache * making subsequent calls with the same arguments fast. * This variant allows the garbage collector to release entries from the cache and at the same time allows * the user to specify how many entries should be protected from the eventual gc-initiated eviction. * Cached entries exceeding the specified preservation threshold are made available for eviction based on * the LRU (Last Recently Used) strategy. * Given the non-deterministic nature of garbage collector, the actual cache size may grow well beyond the protected * size limits set by the user, if memory is plentiful. * Also, this variant will never exceed in size the upper size limit. Once the upper size limit has been reached, * the values in the cache start rotating using the LRU (Last Recently Used) strategy. * The returned function can be safely used concurrently from multiple threads, however, the implementation * values high average-scenario performance and so concurrent calls on the memoized function with identical argument values * may not necessarily be able to benefit from each other's cached return value. Also the protectedCacheSize parameter * might not be respected accurately in such scenarios for some periods of time. With this having been mentioned, * the performance trade-off still makes concurrent use of memoized functions safe and highly recommended. * * The cache gets garbage-collected together with the memoized closure. * @param protectedCacheSize Number of cached return values to protect from garbage collection * @param maxCacheSize The maximum size the cache can grow to * @return A new function forwarding to the original one while caching the results */ public Closure memoizeBetween(final int protectedCacheSize, final int maxCacheSize) { if (protectedCacheSize < 0) throw new IllegalArgumentException("A non-negative number is required as the protectedCacheSize parameter for memoizeBetween."); if (maxCacheSize < 0) throw new IllegalArgumentException("A non-negative number is required as the maxCacheSize parameter for memoizeBetween."); if (protectedCacheSize > maxCacheSize) throw new IllegalArgumentException("The maxCacheSize parameter to memoizeBetween is required to be greater or equal to the protectedCacheSize parameter."); return Memoize.buildSoftReferenceMemoizeFunction(protectedCacheSize, new LRUCache(maxCacheSize), this); } /** * Builds a trampolined variant of the current closure. * To prevent stack overflow due to deep recursion, functions can instead leverage the trampoline mechanism * and avoid recursive calls altogether. Under trampoline, the function is supposed to perform one step of * the calculation and, instead of a recursive call to itself or another function, it return back a new closure, * which will be executed by the trampoline as the next step. * Once a non-closure value is returned, the trampoline stops and returns the value as the final result. * Here is an example: *
     * def fact
     * fact = { n, total ->
     *     n == 0 ? total : fact.trampoline(n - 1, n * total)
     * }.trampoline()
     * def factorial = { n -> fact(n, 1G)}
     * println factorial(20) // => 2432902008176640000
     * 
* * @param args Parameters to the closure, so as the trampoline mechanism can call it * @return A closure, which will execute the original closure on a trampoline. */ public Closure trampoline(final Object... args) { return new TrampolineClosure(this.curry(args)); } /** * Builds a trampolined variant of the current closure. * To prevent stack overflow due to deep recursion, functions can instead leverage the trampoline mechanism * and avoid recursive calls altogether. Under trampoline, the function is supposed to perform one step of * the calculation and, instead of a recursive call to itself or another function, it return back a new closure, * which will be executed by the trampoline as the next step. * Once a non-closure value is returned, the trampoline stops and returns the value as the final result. * @return A closure, which will execute the original closure on a trampoline. * @see #trampoline(Object...) */ public Closure trampoline() { return new TrampolineClosure(this); } /* (non-Javadoc) * @see java.lang.Object#clone() */ public Object clone() { try { return super.clone(); } catch (final CloneNotSupportedException e) { return null; } } /* * Implementation note: * This has to be an inner class! * * Reason: * Closure.this.call will call the outer call method, but * with the inner class as executing object. This means any * invokeMethod or getProperty call will be called on this * inner class instead of the outer! */ private class WritableClosure extends Closure implements Writable { private static final long serialVersionUID = -5749205698681690370L; public WritableClosure() { super(Closure.this); } /* (non-Javadoc) * @see groovy.lang.Writable#writeTo(java.io.Writer) */ public Writer writeTo(Writer out) throws IOException { Closure.this.call(new Object[]{out}); return out; } /* (non-Javadoc) * @see groovy.lang.GroovyObject#invokeMethod(java.lang.String, java.lang.Object) */ public Object invokeMethod(String method, Object arguments) { if ("clone".equals(method)) { return clone(); } else if ("curry".equals(method)) { return curry((Object[]) arguments); } else if ("asWritable".equals(method)) { return asWritable(); } else { return Closure.this.invokeMethod(method, arguments); } } /* (non-Javadoc) * @see groovy.lang.GroovyObject#getProperty(java.lang.String) */ public Object getProperty(String property) { return Closure.this.getProperty(property); } /* (non-Javadoc) * @see groovy.lang.GroovyObject#setProperty(java.lang.String, java.lang.Object) */ public void setProperty(String property, Object newValue) { Closure.this.setProperty(property, newValue); } /* (non-Javadoc) * @see groovy.lang.Closure#call() */ public Object call() { return ((Closure) getOwner()).call(); } /* (non-Javadoc) * @see groovy.lang.Closure#call(java.lang.Object) */ public Object call(Object arguments) { return ((Closure) getOwner()).call(arguments); } public Object call(Object... args) { return ((Closure) getOwner()).call(args); } public Object doCall(Object... args) { return call(args); } /* (non-Javadoc) * @see groovy.lang.Closure#getDelegate() */ public Object getDelegate() { return Closure.this.getDelegate(); } /* (non-Javadoc) * @see groovy.lang.Closure#setDelegate(java.lang.Object) */ public void setDelegate(Object delegate) { Closure.this.setDelegate(delegate); } /* (non-Javadoc) * @see groovy.lang.Closure#getParameterTypes() */ public Class[] getParameterTypes() { return Closure.this.getParameterTypes(); } /* (non-Javadoc) * @see groovy.lang.Closure#getParameterTypes() */ public int getMaximumNumberOfParameters() { return Closure.this.getMaximumNumberOfParameters(); } /* (non-Javadoc) * @see groovy.lang.Closure#asWritable() */ public Closure asWritable() { return this; } /* (non-Javadoc) * @see java.lang.Runnable#run() */ public void run() { Closure.this.run(); } /* (non-Javadoc) * @see java.lang.Object#clone() */ public Object clone() { return ((Closure) Closure.this.clone()).asWritable(); } /* (non-Javadoc) * @see java.lang.Object#hashCode() */ public int hashCode() { return Closure.this.hashCode(); } /* (non-Javadoc) * @see java.lang.Object#equals(java.lang.Object) */ public boolean equals(Object arg0) { return Closure.this.equals(arg0); } /* (non-Javadoc) * @see java.lang.Object#toString() */ public String toString() { final StringWriter writer = new StringWriter(); try { writeTo(writer); } catch (IOException e) { return null; } return writer.toString(); } public Closure curry(final Object... arguments) { return (new CurriedClosure(this, arguments)).asWritable(); } public void setResolveStrategy(int resolveStrategy) { Closure.this.setResolveStrategy(resolveStrategy); } public int getResolveStrategy() { return Closure.this.getResolveStrategy(); } } /** * @return Returns the directive. */ public int getDirective() { return directive; } /** * @param directive The directive to set. */ public void setDirective(int directive) { this.directive = directive; } /** * Returns a copy of this closure where the "owner", "delegate" and "thisObject" * fields are null, allowing proper serialization when one of them is not serializable. * * @return a serializable closure. * * @since 1.8.5 */ @SuppressWarnings("unchecked") public Closure dehydrate() { Closure result = (Closure) this.clone(); result.delegate = null; result.owner = null; result.thisObject = null; return result; } /** * Returns a copy of this closure for which the delegate, owner and thisObject are * replaced with the supplied parameters. Use this when you want to rehydrate a * closure which has been made serializable thanks to the {@link #dehydrate()} * method. * @param delegate the closure delegate * @param owner the closure owner * @param thisObject the closure "this" object * @return a copy of this closure where owner, delegate and thisObject are replaced * * @since 1.8.5 */ @SuppressWarnings("unchecked") public Closure rehydrate(Object delegate, Object owner, Object thisObject) { Closure result = (Closure) this.clone(); result.delegate = delegate; result.owner = owner; result.thisObject = thisObject; return result; } }