Interoperation with Java

Ceylon is designed to execute in a virtual machine environment (loosely speaking, in an environment with built-in garbage collection), but it doesn't have its own native virtual machine. Instead, it "borrows" a virtual machine designed for some other language. For now the options are:

  • the Java Virtual Machine, or
  • a JavaScript virtual machine.

In this chapter, we're going to learn about how to interoperate with native code running on the Java Virtual Machine, and focus on all the nitty-gritty details and corner cases involved in inter-language interop. In the next chapter we'll discuss JavaScript.

Defining a native Java module

Many Ceylon modules that call native Java code are designed to execute only on the JVM. In this case, we declare the whole Ceylon module as a native JVM module using the native annotation.

native ("jvm")
module ceylon.formatter "1.3.1" {
    shared import java.base "7";
    shared import com.redhat.ceylon.typechecker "1.3.1";
    import ceylon.collection "1.3.1";
    ...
}

A cross-platform module may call native Java code too, but in this case we need to apply the native import to import statements that declare dependencies on native Java jar archives.

module ceylon.formatter "1.3.1" {
    native ("jvm") shared import java.base "7";
    native ("jvm") shared import com.redhat.ceylon.typechecker "1.3.1";
    import ceylon.collection "1.3.1";
    ...
}

Furthermore, in this case, we must annotate declarations which make use of the Java classes in these archives.

native ("jvm")
void hello() {
    import java.lang { System }
    System.out.println("Hello, world!");
}

It's not necessary to annotate individual declarations in a module that is already declared as a native JVM module.

Importing Java modules containing native code

A Ceylon program can only use native code belonging to a module. Ceylon doesn't have the notion of a global class path, and not every Ceylon program can be assumed to be executing on the JVM, so dependencies to native Java code must always be expressed explicitly in the module descriptor.

The native Java modules your Ceylon program depends on might come from a Ceylon module repository, from a Maven repository, or they might be part of the JDK or Android SDK.

Depending on the JDK

The Java SE Development Kit (JDK) is represented in Ceylon as a set of modules, according to the modularization proposed by the Jigsaw project. So to make use of the JDK we need to import one or more of these modules. For example:

  • the module java.base contains core packages including java.lang, java.util, and javax.security,
  • the module java.desktop contains the AWT and Swing desktop UI frameworks, and
  • java.jdbc contains the JDBC API.

Thus, if we need to use the Java collections framework in our Ceylon program, we need to create a Ceylon module that depends on java.base.

native ("jvm")
module org.jboss.example "1.0.0" {
    import java.base "7";
}

Now, we can simply import the Java class we're interested in and use it like any ordinary Ceylon class:

import java.util { HashMap }

void hashyFun() {
    value hashMap = HashMap<String,Object>();
}

Tip: using JavaFX

JavaFX is part of the JDK, but it's split into several modules, javafx.base, javafx.graphics, javafx.media, etc, so you'll probably need to import several of them, for example:

native ("jvm")
module example.javafx.circles "1.0.0" {
    shared import java.base "8";
    shared import javafx.base "8";
    shared import javafx.graphics "8";
    shared import javafx.controls "8";
}

Depending on the Android SDK

For Android development, it's necessary to use the Ceylon Android plugin for Gradle to create a Ceylon module repository containing the bits of the Android SDK that are needed to compile an app written in Ceylon. This repository will be created in the subdirectory build/intermediates/ceylon-android/repository of your Ceylon Android project. This repository must be specified as the source of all Android-related Java modules using --rep, and as the provider of the Java SDK using the command-line argument --jdk-provider of ceylon compile.

However, when compiling using Gradle, or from inside Android Studio, the repository is created by the build, and the compiler option is set automatically, so you don't need to mess with it explicitly. You can find it in .ceylon/config, if you're interested:

[compiler]
jdkprovider=android/24

[repositories]
lookup=./build/intermediates/ceylon-android/repository

Note: Android itself has an immensely complicated toolchain, and Android apps cannot reasonably be compiled using only commmand line tools. In practice, you'll always use Gradle to build an Android app.

A basic boilerplate module descriptor for Android development looks like this:

native ("jvm")
module com.acme.example.androidapp "1.0" {
    import java.base "7";
    import android "24";
    import "com.android.support.appcompat-v7" "24.2.1";
    import "com.android.support.design" "24.2.1";
}

You can get started with Ceylon on Android by following this getting started guide.

Depending on a Java archive

To make use of native code belonging to a packaged .jar archive, you have two options:

To add a Java .jar to a Ceylon module repository, you need to provide some metadata describing its dependencies. A module.xml or module.properties file specifies dependency information for a .jar.

  • The format of the Ceylon module.properties file is documented here, and
  • the JBoss Modules module.xml descriptor format is defined here.

The command line tool ceylon import-jar can help make this task easier.

If you're using Ceylon IDE for Eclipse, and you don't want to write the module.xml descriptor by hand, go to File > Export ... > Ceylon > Java Archive to Module Repository.

Depending on a Maven module

Alternatively, the Ceylon module architecture interoperates with Maven via Aether. You can import a module from maven by specifying the maven: repository type:

import maven:org.hibernate:"hibernate-core" "5.0.4.Final";

The module name is composed from the Maven group id and artifact id, and the artifact id must be quoted.

You can find more information here.

Depending on Java EE

The Java EE APIs are available in Ceylon Herd, so you can import them like this:

import javax.javaeeapi "7.0";

Alternatively, you can get them from Maven:

import maven:javax:"javaee-api" "7.0";

The second approach might work better if you're also importing other related Java modules from Maven.

Interoperation with Java types

Calling and extending Java types from Ceylon is mostly completely transparent. You don't need to do anything "special" to use or extend a Java class or interface in Ceylon, or to call its methods.

There's a handful of things to be aware of when writing Ceylon code that calls a Java class or interface, arising out of the differences between the type systems of the two languages.

Certain very abstract Java supertypes are mapped to Ceylon types

There are three Java classes and one interface which simply can't be used at all in Ceylon code, because Ceylon provides types which are exactly equivalent in the package ceylon.language. When these types occur in the signature of an operation in Java, they're always represented by the equivalent Ceylon type:

  • java.lang.Object is represented by Ceylon's Object class,
  • java.lang.Exception is represented by Ceylon's Exception class,
  • java.lang.Throwable is represented by Ceylon's Throwable class, and
  • java.lang.annotation.Annotation is represented by the interface Annotation.

It's an error to attempt to import one of these Java types in Ceylon code.

Java primitive types are mapped to Ceylon types

You're never exposed to Java primitive types when calling a Java method or field from Ceylon. Instead:

  • boolean is represented by Ceylon's Boolean class,
  • char is represented by Ceylon's Character class,
  • long, int, and short are represented by Ceylon's Integer class,
  • byte is represented by Ceylon's Byte class,
  • double and float are represented by Ceylon's Float class, and
  • java.lang.String is represented by Ceylon's String class.

Almost all of the time, this behavior is completely intuitive. But there's two wrinkles to be aware of...

Gotcha!

According to these rules, all conversions from a Java primitive to a Ceylon type are widening conversions, and are guaranteed to succeed at runtime. However, conversion from a Ceylon type to a Java primitive type might involve an implicit narrowing conversion. For example, if:

  • a Ceylon Integer is assigned to a Java int or short,
  • a Ceylon Float is assigned to a Java float, or if
  • a Ceylon UTF-32 Character is assigned to a Java 16-bit char,

the assignment can result in silent overflow or loss of precision at runtime.

Note: it is not a goal of Ceylon's type system to warn about operations which might result in numeric overflow. In general, almost any operation on a numeric type, including + or *, can result in numeric overflow.

Gotcha again!

There's no mapping between Java's wrapper classes like java.lang.Integer or java.lang.Boolean and Ceylon basic types, so these conversions must be performed explicitly by calling, for example, longValue() or booleanValue(), or by explicitly instantiating the wrapper class, just like you would do in Java when converting between a Java primitive type and its wrapper class.

This is mainly important when working with Java collection types. For example, a Java List<Integer> doesn't contain Ceylon Integers.

import java.lang { JInteger=Integer }
import java.util { JList=List } 

JList<JInteger> integers = ... ;
for (integer in integers) { //integer is a JInteger!
    Integer int = integer.longValue(); //convert to Ceylon Integer
    ...
}

(This isn't really much worse than Java, by the way: a Java List<Integer> doesn't hold primitive Java ints either!)

Worse, a Java List<String> doesn't contain Ceylon Strings.

import java.lang { JString=String }
import java.util { JList=List } 

JList<JString> strings = ... ;
for (string in strings) { //string is a JString!
    String str = string.string; //convert to Ceylon String
    ...
}

Watch out for this!

Tip: converting Java strings

Explicitly converting between String and Java's java.lang.String is easy:

  • the .string attribute of a Java string returns a Ceylon string, and
  • one of the constructors of java.lang.String accepts a Ceylon String, or, alternatively,
  • the function javaString in the module ceylon.interop.java converts a Ceylon string to a Java string without requiring an object instantiation.

Tip: converting Java primitive wrapper types

Likewise, conversions between Ceylon types and Java primitive wrapper types are just as trivial, for example:

  • the .longValue() methods of java.lang.Long and java.lang.Integer return a Ceylon Integer, and
  • the constructors of java.lang.Integer and java.lang.Long accept a Ceylon Integer, as do the static valueOf() methods of these types.

Likewise:

  • the .doubleValue() methods of java.lang.Double and java.lang.Float return a Ceylon Float, and
  • the constructors of java.lang.Double and java.lang.Float accept a Ceylon Float, as do the static valueOf() methods of these types.

The story is similar for other primitive wrapper types.

Tip: using the small annotation

If, for some strange reason, you really need a 32-bit int or float at the bytecode level, instead of the 64-bit long or double that the Ceylon compiler uses by default, you can use the small annotation.

small Integer int = string.hash; //a 32-bit int

You can also use small to represent a Character as a 16-bit char at the bytecode level, instead of as a 32-bit int.

small Character char = charArray.get(0); //a 16-bit char

It's important to understand that small Integer isn't a different type to Integer. So any Integer is directly assignable to a declaration of type small Integer, and the compiler will silently produce a narrowing conversion, which could result in silent overflow or loss of precision at runtime.

Note also that small is defined as a hint, and may be completely ignored by the compiler. (And, indeed, it is always ignored when compiling to JavaScript.)

Java array types are represented by special Ceylon classes

Since there are no primitively-defined array types in Ceylon, arrays are represented by special classes. These classes are considered to belong to the package java.lang. (Which belongs to the module java.base.) So these classes must be explicitly imported!

  • boolean[] is represented by the class BooleanArray,
  • char[] is represented by the class CharArray,
  • long[] is represented by the class LongArray,
  • int[] is represented by the class IntArray,
  • short[] is represented by the class ShortArray,
  • byte[] is represented by the class ByteArray,
  • double[] is represented by the class DoubleArray,
  • float[] is represented by the class FloatArray, and, finally,
  • T[] for any object type T is represented by the class ObjectArray<T>.

We can obtain a Ceylon Array without losing the identity of the underlying Java array.

import java.lang { ByteArray }

ByteArray javaByteArray = ByteArray(10);
Array<Byte> byteArray = javaByteArray.byteArray;

You can think of the ByteArray as the actual underlying byte[] instance, and the Array<Byte> as an instance of the Ceylon class Array that wraps the byte[] instance.

The module ceylon.interop.java contains a raft of additional methods for working with these Java array types.

Java object types and null values

Java's primitive types do not hold null values, so any primitive type is mapped to a non-null Ceylon type. The situation is a bit more complicated for Java object and array types, which are treated differently depending upon whether they occur:

  • as a method return type or field type, or
  • as a method parameter type.

Non-primitive parameter types are treated as nullable

Java method signatures offer no information about whether a method parameter accepts a null value, except in the very special case of a parameter with primitive type. Therefore, Ceylon assigns an optional type to every parameter of non-primitive type.

Null return values are checked at runtime

Likewise, Java field or method types offer no information about whether a field or method can produce a null value, except, again, in the special case of a primitive type. But it would be unacceptably intrusive for Ceylon to treat almost every Java method or field as returning an optional type, forcing you to explicitly check for null every time you call Java! So Ceylon doesn't do that. Instead, it treats Java methods and fields as having non-optional type.

But that's, of course, unsound from the point of view of the type system of Ceylon. A method can still return null at runtime!

Therefore, the compiler must insert runtime null value checks wherever Ceylon code calls a Java function that returns an object type, or evaluates a Java field of object type, and assigns the result to an non-optional Ceylon type.

In this example, no runtime null value check is performed, since the return value of System.getProperty() is assigned to an optional type:

import java.lang { System }

void printUserHome() {
    String? home  //optional type
            = System.getProperty("user.home");
    print(home);
}

In this example, however, a runtime type check occurs when the return value of System.getProperty() is assigned to the non-optional type String:

import java.lang { System }

void printUserHome() {
    String home  //non-optional type, possible runtime exception
            = System.getProperty("user.home");
    print("home: " + home);
}

However, no runtime check occurs upon assignment to a value or function with inferred type. Thus, there is no runtime check at all in this code:

import java.lang { System }

void printUserHome() {
    value home  //inferred non-optional type, no runtime check!
            = System.getProperty("user.home");
    print(home);
}

Instead, a runtime check is inserted when the value or function with inferred type is used in a way which indicates it must be non-null:

import java.lang { System }

void printUserHome() {
    value home  //no runtime check
            = System.getProperty("user.home");
    print("home: " + home); //runtime check here instead
}

These runtime checks ensure that null can never unsoundly propagate from native Java code with unchecked null values into Ceylon code with checked null values, resulting in an eventual NullPointerException in Ceylon code far from the original call to Java.

Java types annotated @Nullable are exposed as optional types

There are now a number of Java frameworks that provide annotations (@Nullable and @Nonnull/@NonNull/@NotNull) for indicating whether a Java type can contain null values. The Ceylon compiler understands these annotations.

So, as an exception to the above discussion, when one of these annotations is present, Java null values are checked at compile time, and no runtime checks are necessary.

Java methods, properties, and constants

Ceylon makes use of the JavaBeans conventions when representing a Java class.

Java properties are exposed as Ceylon attributes

A getter or getter/setter pair belonging to a Java class will appear to a Ceylon program as an attribute. For example:

import java.util { Calendar, TimeZone } 

void calendaryFun() {
    Calendar calendar = Calendar.instance;
    TimeZone timeZone = calendar.timeZone;
    Integer timeInMillis = calendar.timeInMillis;
}

If you want to call the Java setter method, assign a value to the attribute using =, the assignment operator:

calendar.timeInMillis = system.milliseconds;

However, if there is no exactly-matching getter method for a setter, the setter is treated as a regular method.

Gotcha!

Note that there are certain corner cases here which might be confusing. For example, consider this Java class:

public class Foo {
    public String getBar() { ... }
    public void setBar(String bar) { ... }
    public void setBar(String bar, String baz) { ... }
}

From Ceylon, this will appear as if it were defined like this:

shared class Foo {
    shared String bar { ... }
    assign bar { ... }
    shared void setBar(String bar, String baz) { ... }
}

Therefore:

  • to call the single-argument setter, use an assignment statement like foo.bar = bar, but
  • to call the two-argument setter, use a method call like foo.setBar(bar,baz).

Other Java methods are exposed as Ceylon methods

An Java method that doesn't follow the JavaBeans property conventions is represented to the Ceylon program as a regular method.

Gotcha: calling Java methods from Ceylon

There's two minor limitations to be aware of when calling Java methods from Ceylon.

  • You can't call Java methods using the named argument syntax, since Java 7 doesn't expose the names of parameters at runtime (except for code compiled on Java 8 with a special compiler switch explicitly enabled).
  • You can't obtain a method reference, nor a static method reference, to an overloaded method.

Methods accepting a SAM interface

Java has no true function types, so there's no equivalent to Ceylon's Callable interface in Java. Instead, Java features SAM (Single Abstract Method) conversion where an anonymous function is converted by the compiler to an instance of an interface type like java.util.function.Predicate which declares only one abstract method.

Thus, there are many operations in the Java SDK and other Java libraries which accept a SAM interface. For example, Stream.filter() accepts a Predicate. Such methods are represented in Ceylon as an overloaded method:

  • one overload accepts the SAM, like Stream<T> filter(Predicate<in T> predicate), and
  • the second overload accepts a Ceylon function type, like Stream<T> filter(Boolean(T) predicate).

Thus, we can pass anonymous functions and function references to such Java APIs without needing to explicitly implement the SAM interface type.

import java.util {
    Arrays
}
import java.util.stream {
    Collectors { toList }
    Stream { with=\iof }
}

value list
    = Stream.with("hello", "world", "goodbye")
        .filter((s) => s.longerThan(2))
        .map(String.uppercased)
        .collect(toList<String>());

As you can see, use of the Java streams API is completely natural in Ceylon.

Tip: getting a function from a SAM

There's no inverse conversion from a SAM type to a Ceylon function type. A Predicate<T> can't be passed directly to an operation that expects a Boolean(T). But this is perfectly fine, because it's completely trivial to perform the conversion explicitly, just by taking a reference to the test method of Predicate:

Predicate<String> javaPredicate = ... ;
Boolean(String) ceylonPredicate = javaPredicate.test;

Remember, the Ceylon language defines no implicit type conversions anywhere!

Java constants and enum values

Java constants like Integer.MAX_VALUE and Java enum values, like RetentionPolicy.RUNTIME, follow an all-uppercase naming convention. Since this looks rather alien in Ceylon code, it's acceptable to refer to them using camel case. This:

Integer maxInteger = JInteger.maxValue;
RetentionPolicy policy = RetentionPolicy.runtime;

is preferred to this:

Integer maxInteger = JInteger.\iMAX_VALUE;
RetentionPolicy policy = RetentionPolicy.\iRUNTIME;

However, both options are accepted by the compiler.

Tip: switching over Java enum types

A Java enum type is treated as as enumerated class. Each enumerated value of the enum type is represented as a value constructor of the class. Thus, it's possible to switch over the members of a Java enum, just like you can in Java.

RetentionPolicy policy = ... ;
switch (policy)
case (RetentionPolicy.source) {
    ...
}
case (RetentionPolicy.runtime) {
    ...
}
case (RetentionPolicy.\iclass) {
    ...
}

Java generic types

A generic instantiation of a Java type like java.util.List<E> is representable without any special mapping rules. The Java type List<String> may be written as List<String> in Ceylon, after importing List and String from the module java.base.

import java.lang { String }
import java.util { List, ArrayList }

List<String> strings = ArrayList<String>();

The only problem here is that it's really easy to mix up Java's String, List, and ArrayList with the Ceylon types with the same names!

Tip: disambiguating Java types using aliases

It's usually a good idea to distinguish a Java type with the same name as a common Ceylon type using an import alias. So we would rewrite the code fragment above like this:

import java.lang { JString = String }
import java.util { JList = List, JArrayList = ArrayList }

JList<JString> strings = JArrayList<JString>();

That's much less likely to cause confusion!

Wildcards and raw types are represented using use-site variance

In pure Ceylon code, we almost always use declaration-site variance. However, this approach doesn't work when we interoperate with Java generic types, which are by nature all invariant. In Java, covariance or contravariance is represented at the point of use of the generic type, using wildcards. Therefore, Ceylon also supports use-site variance (wildcards).

Use-site variance is indicated using the keywords out and in:

  • List<out Object> has a covariant wildcard, and is equivalent to List<? extends Object> in Java, and
  • Topic<in Object> has a contravariant wildcard, and is equivalent to Topic<? super Object> in Java.

Java raw types are also represented with a covariant wildcard. The raw type List is represented as List<out Object> in Ceylon.

Note that for any invariant generic type Type<X>, the instantiations Type<out Anything> and Type<in Nothing> are exactly equivalent, and are supertypes of every other instantiation of Type.

Wildcard types are unavoidable when interoperating with Java, and perhaps occasionally useful in pure Ceylon code. But we recommend avoiding them, except where there's a really good reason.

Gotcha!

Instances of Java generic types don't carry information about generic type arguments at runtime. So certain operations which are legal in Ceylon are rejected by the Ceylon compiler when a Java generic type is involved. For example, the following code is illegal, since JList is a Java generic type, with erased type arguments:

import java.lang { JString = String }
import java.util { JList = List, JArrayList = ArrayList }

Object something = JArrayList<JString>();
if (is JList<JString> something) { //error: type condition cannot be fully checked at runtime
    ... 
}

This isn't a limitation of Ceylon; the equivalent code is also illegal in Java!

Tip: dealing with Java type erasure

When you encounter the need to narrow to an instantiation of a Java generic type, you can use an unchecked type assertion:

import java.lang { JString = String }
import java.util { JList = List, JArrayList = ArrayList }

Object something = JArrayList<JString>();
if (is JList<out Anything> something) { //ok
    assert (is JList<JString> something); //warning: type condition might not be fully checked at runtime
    ...
}

This code is essentially the same as what you would do in Java:

  1. test against the raw type List using instanceof, and then
  2. narrow to List<String> using an unchecked typecast.

Just like in Java, the above code produces a warning, since the type arguments in the assertion cannot be checked at runtime.

Inheriting Java types and refining Java methods

A Ceylon class may extend a Java class and/or implement Java interfaces, even if the Java types are generic.

As we saw above, Java method parameter types are treated as optional. However, when refining a Java method in a Ceylon class, we're allowed to redeclare the parameter as non-optional, for example:

//java interface
public interface Stringifier<T> {
    public String stringify(T thing);
}
//ceylon implementation
class EntryStringifier<Key,Item>()
        satisfies Stringifier<Key->Item> {
    stringify(Key->Item entry) 
            => "``entry.key``->``entry.item``";
}

Here, we did not have to declare entry as being of type <Key->Item>?, though we could have, if we expected our method to be called with a null argument.

There's nothing else particularly interesting to say about this topic, except to point out one small wrinkle. In Java, the following inheritance hierarchy is legal:

interface Foo {
    public boolean test();
}

interface Bar {
    public boolean test();
}

class Baz implements Foo, Bar {
    @Override
    public boolean test() {
        return true;
    }
}

However, a Ceylon class which implements the Java types Foo and Bar is not legal, since a Ceylon class may not define an operation that refines two completely different but same-named operations unless they both descend from a common supertype definition:

class Baz() satisfies Foo & Bar {
    //error: May not inherit two declarations with the 
    //same name that do not share a common supertype
    test() => true;
}

In this case, it's necessary to either:

  • split Baz into two classes,
  • define an intermediate Java interface that inherits Foo and Bar, or
  • define a common superinterface of Foo and Bar that declares test().

Java annotations

Java annotations may be applied to Ceylon program elements. An initial lowercase identifier must be used to refer to the Java annotation type.

For example:

import javax.persistence {
    id,
    generatedValue,
    entity,
    column,
    manyToOne,
}

entity
class Person(firstName, lastName, city) {

    id generatedValue
    column { name = "pid"; }
    value _id = 0;

    shared String firstName;
    shared String lastName;

    manyToOne { optional=false; }
    shared City city;

    string => "Person(``_id``, ``firstName`` ``lastName``, ``city.name``)";
}

Note that id here refers to the Java annotation javax.persistence.Id.

Tip: specifying arguments to anotation parameters of type Class

Annotation values of type java.util.Class may be specified by passing the corresponding Ceylon ClassDeclaration. For example, you would use type = `class Person` where you would have used type = Person.class in Java.

Java language modifiers are represented as annotations

The following Java language modifiers do not naturally correspond to any annotation in ceylon.language:

  • transient,
  • volatile,
  • synchronized,
  • native, and
  • strictfp.

Ceylon treats these modifiers as annotations belonging to the package java.lang (in the module java.base).

Therefore, if you want to use one of these JVM-specific modifiers in Ceylon, you must explicitly import it:

import java.lang { volatile }

volatile variable value counter = 0;

Of these annotations, transient and volatile are the most commonly-used in Ceylon code. We discourage direct use of synchronized, which is extremely vulnerable to deadlocks, preferring the use of java.util.concurrent.

Syntax sugar that applies to Java types

Certain syntactic constructs that are defined by the language specification in terms of types defined in ceylon.language are also supported for similar Java types. These constructs are:

  • the for loop and comprehensions,
  • resource expressions in try,
  • the element lookup and in operators, and
  • the "spread" operators *. and *.

Java Iterable or array in for

It's possible to use a java.lang.Iterable or Java array in a Ceylon for statement or comprehension.

import java.lang { JIterable=Iterable }

JIterable<Object> objects = ... ;

//for statement
for (obj in objects) {
    ...
}

//comprehension
print(", ".join { for (obj in objects) if (exists obj) obj.string.trimmed });

Imagine how great it would be to be able to write a comprehension involving a Java collection in Java itself!

Wait, a quick reminder...

Gotcha! (redux)

As we already noted above, for does nothing special to the type of the iterated elements of a Java array or Iterable: if the elements are Java Strings, they're not magically transformed to Ceylon Strings:

import java.lang { ObjectArray, JString=String }

ObjectArray<JString> strings = ...;
for (s in strings) {  // s is a JString
    String string = s.string;  //string is a Ceylon String
    ...
}

Just don't forget to use s.string to get the Ceylon String if that's what you need!

Gotcha again!

Also note that java.lang.Iterable is not supported together with entry or tuple destructuring:

JIterable<String->String> stringPairs = ...;
for (key->item in stringPairs) {
    // not supported
}

In practice it's unusual to have a Java Iterable containing Ceylon Entrys or Tuples.

Java AutoCloseable in try

Similarly, it's possible to use a Java AutoCloseable in a Ceylon try statement.

import java.io { File, FileInputStream }

File file = ... ;
try (inputStream = FileInputStream(file)) {
     ...
}

The semantics, naturally, are identical to what you get in Java.

Java collections and operators

Two of Ceylon's built-in operators may be applied to Java types:

  • the element lookup operator (list[index]) may be used with Java arrays, java.util.List, and java.util.Map, and
  • the containment operator (element in container) may be used with the type java.util.Collection.

This is syntax sugar you don't even have in Java itself!

import java.util { JArrayList=ArrayList }

value list = JArrayList<String>();
list.add("hello");
list.add("world");

list[0] = "goodbye";
assert ("world" in list);
printAll { for (word in list) word };

There's almost no friction here; if you prefer to use Java's collection types instead of the mutable collection types defined by the module ceylon.collection, go right ahead!

Tip: creating Java lists

The following idioms are very useful for instantiating Java Lists:

import java.util { Arrays }

value words = Arrays.asList("hello", "world");
value squares = Arrays.asList(for (x in 0..100) x^2);
value args = Arrays.asList(*process.arguments);

(Note that these code examples work because Arrays.asList() has a variadic parameter.)

Tip: copying Java collection elements into Ceylon collections

You can use the spread operator to obtain a Ceylon collection with the same elements as a Java collection:

import java.util { JList=List }

JList<Object> objects = .... ;
value sequenceOfObjects = [*objects];
value setOfObjects = set { *objects };

A comprehension gives you even more power:

import java.util { JList=List }
import java.lang { JString=String }

JList<JString> strings = .... ;
value sequenceOfStrings = [ for (str in strings) str.string ];

However, copying collections by nature involves memory allocation and this can be slow. A more efficient approach is to wrap the Java collection. Fortunately, Ceylon has a library for that.

Utility functions and classes

In the module ceylon.interop.java you'll find a suite of useful utility functions and classes for Java interoperation. For example, there are classes that adapt between Ceylon collection types and Java collection types.

Tip: converting between Iterables

An especially useful adaptor is CeylonIterable, which lets you apply any of the usual operations of a Ceylon stream to a Java Iterable.

import java.util { JList=List, JArrayList=ArrayList }
import ceylon.interop.java { CeylonIterable }

...

JList<String> strings = JArrayList<String>();
strings.add("hello");
strings.add("world");

CeylonIterable(strings).each(print);

(Alternatively, we could have used CeylonList in this example.)

Similarly there are CeylonStringIterable, CeylonIntegerIterable, CeylonFloatIterable,CeylonByteIterable and CeylonBooleanIterable classes which as well as converting the iterable type also convert the elements from their Java types to the corresponding Ceylon type.

Tip: getting a java.util.Class

Another especially useful function is javaClass, which obtains an instance of java.util.Class for a given type.

import ceylon.interop.java { javaClass }
import java.lang { JClass=Class }

JClass<Integer> jc = javaClass<Integer>();
print(jc.protectionDomain);

The functions javaClassFromInstance and javaClassFromDeclaration are also useful.

Additional support for Java interoperation

Ceylon provides the following additional features to support interoperation with the Java platform.

META-INF and WEB-INF

Some Java frameworks and environments require metadata packaged in the META-INF or WEB-INF directory of the module archive, or sometimes even in the root directory of the module archive. We've already seen how to package resources in Ceylon module archives by placing the resource files in the module-specific subdirectory of the resource directory, named resource by default.

Then, given a module named net.example.foo:

  • resources to be packaged in the root directory of the module archive should be placed in resource/net/example/foo/ROOT/,
  • resources to be packaged in the META-INF directory of the module archive should be placed in resource/net/example/foo/ROOT/META-INF/, and
  • resources to be packaged in the WEB-INF directory of the module archive should be placed in resource/net/example/foo/ROOT/WEB-INF/.

Interoperation with Java serialization

Ceylon classes are implicitly serializable by Java's built-in object serialization facility.

When compiling a Ceylon class, the compiler adds the supertype java.io.Serializable to the generated Java class, along with a package-private default constructor. Neither the supertype, nor the constructor, are visible to other Ceylon code. But they're enough to make the Ceylon class serializable via Java's built-in binary serialization APIs.

Of course, if your Ceylon object holds a reference to some other object that's not serializable, you still won't be able to serialize the Ceylon object!

Interoperation with Java's ServiceLoader

Ceylon services and service providers work transparently with Java's service loader architecture, having been designed and implemented as a simple abstraction of Java's ServiceLoader.

  • Annotating a Ceylon class with the service annotation makes the class available to Java's ServiceLoader.
  • Similarly, a Ceylon module may gain access to Java services just by calling Module.findServiceProviders().

The service annotation and Module.findServiceProviders() work portably across the JVM and JavaScript environments.

Java EE and other annotation-driven frameworks

There are a number of widely-used Java frameworks that depend upon direct reflection-based access to the fields of annotated classes.

For example, the following libraries and frameworks use this approach extensively:

When using this sort of framework in Ceylon, you'll probably need to enable a special compiler mode which slightly changes the way the compiled Ceylon class represents its internal state using Java fields.

With EE mode activated, Ceylon classes should work transparently with these frameworks and libraries. (Note: we've already tested all the above listed technologies with Ceylon, to be sure they work correctly!)

In many cases EE mode is activated implicitly when you import a Java module or use a Java EE annotation.

You'll often need to use the late annotation in conjunction with EE mode.

Alternative module runtimes

When you execute a Ceylon program using ceylon run, it executes on a lightweight module container that is based on JBoss Modules. But it's also possible to execute a Ceylon module in certain other module containers, or without any module container at all.

Tip: running your program without a Ceylon installation

Don't forget that you can run a Ceylon program assembled using ceylon assemble --include-runtime on a machine with no Ceylon installation. In this case, the JBoss Modules-based module container is included in the .cas archive itself.

Deploying Ceylon on OSGi

Ceylon is fully interoperable with OSGi, so that Ceylon modules:

  • can be deployed as pure OSGi bundles in an OSGi container out-of-the-box, without any modification of the module archive file,
  • can embed additional OSGi metadata, to declare services for example, and
  • can easily use OSGi standard services.

This provides a great and straightforward way to run Ceylon modules inside Java EE application servers or enterprise containers that support or are are built around OSGi.

You can learn more about Ceylon and OSGi from the reference documentation.

Deploying Ceylon on Jigsaw

The command line tool ceylon jigsaw deploys a module and all its dependencies to a Jigsaw-style mlib/ directory. Since Jigsaw is not yet released, and is still undergoing changes, this functionality is considered experimental.

Deploying Ceylon on Java EE or WildFly Swarm

These command line tools make it easy to assemble a Ceylon module that makes use of Java EE APis in a Java EE environment:

  • ceylon war repackages a module and its dependencies as a Java EE .war web archive, and
  • ceylon swarm repackages a module, its dependencies, and the Wildfly Swarm environment as a .jar archive.

Tip: including static web content

To include a directory containing static web resources like HTML, image, and JavaScript files in an archive generated by ceylon war or ceylon swarm, use --resource-root or -R, for example:

ceylon war -R=web com.acme.todolist

Deploying Ceylon as a fat jar

To run a Ceylon program on a machine with no Ceylon distribution available, and without the runtime module isolation provided by JBoss Modules, the command line tool ceylon fat-jar is indispensable. The command simply assembles a Java .jar archive that contains a Ceylon module and everything it depends on at runtime.

Publishing Ceylon via Maven

Finally, the ceylon maven-export command assembles a Maven repository containing a Ceylon module and its dependencies. This is most useful if we want to make a Ceylon module available to a Java project that is built using Maven.

However, this isn't the only way to integrate Ceylon modules into a Maven build.

Ceylon and Maven

Every compiled Ceylon module archive includes generated Maven metadata, in META-INF/maven/groupId/artifactId/pom.xml and META-INF/maven/groupId/artifactId/pom.properties.

You can specify the Maven group id and artifact id in your module descriptor:

module org.hibernate.core                    //Ceylon module name
       maven:org.hibernate:"hibernate-core"  //Maven group id + artifact id
       "2.1.1" {                             //module version
   ...
} 

If this information is missing, the group id and artifact id will be inferred from the module name.

There are three ways to publish a Ceylon module to a Maven repository, allowing use of the Ceylon module directly from Java via Maven:

  1. simply publish the Ceylon module to Ceylon Herd, which automatically makes it available in the Herd Maven repository,
  2. use the ceylon maven-export command to assemble a Maven repository containing the Ceylon module archive and its dependencies, or
  3. build the Ceylon modules using Maven.

If you build your Ceylon modules using Maven, or if you use ceylon maven-export, you can easily publish the resulting module archives to a public Maven repository such as Maven Central.

All Ceylon SDK platform modules are already available in Maven Central, under the group id org.ceylon-lang. (They're also available in under the same group id in the Herd Maven repository.) For example, ceylon.collection has the group id org.ceylon-lang and artifact id ceylon.collection.

There's more ...

You can learn more about native declarations from the reference documentation.

Ceylon's experimental support for Jigsaw is covered in greater detail in this blog post.

In a mixed Java/Ceylon project, you'll probably need to use a build system like Gradle, Maven, or Apache ant. Ceylon has plugins for each of these build systems.

Finally, we're going to learn about interoperation with languages like JavaScript with dynamic typing.