Version 2.0.0.M1

© 2017 Pivotal Software, Inc.

Copies of this document may be made for your own use and for distribution to others, provided that you do not charge any fee for such copies and further provided that each copy contains this Copyright Notice, whether distributed in print or electronically.

What is Spring Shell?

Not all applications need a fancy web user interface! Sometimes, interacting with an application using an interactive terminal is the most appropriate way to get things done.

Spring Shell allows one to easily create such a runnable application, where the user will enter textual commands that will get executed until the program terminates. The Spring Shell project provides the infrastructure to create such a REPL (Read, Eval, Print Loop), allowing the developer to concentrate on the commands implementation, using the familiar Spring programming model.

Advanced features such as parsing, TAB completion, colorization of output, fancy ascii-art table display, input conversion and validation all come for free, with the developer only having to focus on core command logic.

Using Spring Shell

Getting Started

To see what Spring Shell has to offer, let’s write a trivial shell application that has a simple command to add two numbers together.

Let’s Write a Simple Boot App

Starting with version 2, Spring Shell has been rewritten from the ground up with various enhancements in mind, one of which is easy integration with Spring Boot, although it is not a strong requirement. For the purpose of this tutorial, let’s create a simple Boot application, for example using start.spring.io. This minimal application only depends on spring-boot-starter and configures the spring-boot-maven-plugin, generating an executable ├╝ber-jar:

...
<dependencies>
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter</artifactId>
    </dependency>
    ...

Adding a Dependency on Spring Shell

The easiest way to get going with Spring Shell is to depend on the spring-shell-starter artifact. This comes with everything one needs to use Spring Shell and plays nicely with Boot, configuring only the necessary beans as needed:

...
<dependency>
    <groupId>org.springframework.shell</groupId>
    <artifactId>spring-shell-starter</artifactId>
    <version>2.0.0.M1</version>
</dependency>
...

Given that Spring Shell will kick in and start the REPL by virtue of this dependency being present, you’ll need to either build skipping tests (-DskipTests) throughout this tutorial or remove the sample integration test that was generated by start.spring.io. If you don’t do so, the integration test will create the Spring ApplicationContext and stay stuck in the eval loop.

Your first command

It’s time to add our first command. Create a new class (name it however you want) and annotate it with @ShellComponent (a variation of @Component that is used to restrict the set of classes that are scanned for candidate commands).

Then, create an add method that takes two ints (a and b) and returns their sum. Annotate it with @ShellMethod and provide a description of the command in the annotation (the only piece of information that is required):

package com.example.demo;

import org.springframework.shell.standard.ShellMethod;
import org.springframework.shell.standard.ShellComponent;

@ShellComponent
public class MyCommands {

    @ShellMethod("Add two integers together.")
    public int add(int a, int b) {
        return a + b;
    }
}

Let’s Give It a Ride!

Build the application and run the generated jar, like so;

./mvnw clean install -DskipTests
[...]

java -jar target/demo-0.0.1-SNAPSHOT.jar

You’ll be greeted by the following screen (the banner comes from Spring Boot, and can be customized as usual):

  .   ____          _            __ _ _
 /\\ / ___'_ __ _ _(_)_ __  __ _ \ \ \ \
( ( )\___ | '_ | '_| | '_ \/ _` | \ \ \ \
 \\/  ___)| |_)| | | | | || (_| |  ) ) ) )
  '  |____| .__|_| |_|_| |_\__, | / / / /
 =========|_|==============|___/=/_/_/_/
 :: Spring Boot ::        (v1.5.6.RELEASE)

shell:>

Below is a yellow shell:> prompt that invites you to type commands. Type add 1 2 then ENTER and admire the magic!

shell:>add 1 2
3

Try to play with the shell (hint: there is a help command) and when you’re done, type exit ENTER.

The rest of this document delves deeper into the whole Spring Shell programming model.

Writing your own Commands

The way Spring Shell decides to turn a method into an actual shell command is entirely pluggable (see Extending Spring Shell), but as of Spring Shell 2.x, the recommended way to write commands is to use the new API described in this section (the so-called standard API).

Using the standard API, methods on beans will be turned into executable commands provided that

  • the bean class bears the @ShellComponent annotation. This is used to restrict the set of beans that are considered.

  • the method bears the @ShellMethod annotation.

The @ShellComponent is a stereotype annotation itself meta-annotated with @Component. As such, it can be used in addition to the filtering mechanism to also declare beans (e.g. using @ComponentScan).

The name of the created bean can be customized using the value attribute of the annotation.

It’s all about Documentation!

The only required attribute of the @ShellMethod annotation is its value attribute, which should be used to write a short, one-sentence, description of what the command does. This is important so that your users can get consistent help about your commands without having to leave the shell (see Integrated Documentation with the help Command).

The description of your command should be short, one or two sentences only. For better consistency, it is recommended that it starts with a capital letter and ends with a dot.

Customizing the Command Name(s)

By default, there is no need to specify the key for your command (i.e. the word(s) that should be used to invoke it in the shell). The name of the method will be used as the command key, turning camelCase names into dashed, gnu-style, names (that is, sayHello() will become say-hello).

It is possible, however, to explicitly set the command key, using the key attribute of the annotation, like so:

        @ShellMethod(value = "Add numbers.", key = "sum")
        public int add(int a, int b) {
                return a + b;
        }

The key attribute accepts multiple values. If you set multiple keys for a single method, then the command will be registered using those different aliases.

The command key can contain pretty much any character, including spaces. When coming up with names though, keep in mind that consistency is often appreciated by users (i.e. avoid mixing dashed-names with spaced names, etc.)

Invoking your Commands

By Name vs. Positional Parameters

As seen above, decorating a method with @ShellMethod is the sole requirement for creating a command. When doing so, the user can set the value of all method parameters in two possible ways:

  • using a parameter key (e.g. --arg value). This approach is called "by name" parameters

  • or without a key, simply setting parameter values in the same order they appear in the method signature ("positional" parameters).

These two approaches can be mixed and matched, with named parameters always taking precedence (as they are less prone to ambiguity). As such, given the following command

        @ShellMethod("Display stuff.")
        public String echo(int a, int b, int c) {
                return String.format("You said a=%d, b=%d, c=%d", a, b, c);
        }

then the following invocations are all equivalent, as witnessed by the output:

shell:>echo 1 2 3               (1)
You said a=1, b=2, c=3
shell:>echo --a 1 --b 2 --c 3   (2)
You said a=1, b=2, c=3
shell:>echo --b 2 --c 3 --a 1   (3)
You said a=1, b=2, c=3
shell:>echo --a 1 2 3           (4)
You said a=1, b=2, c=3
shell:>echo 1 --c 3 2           (5)
You said a=1, b=2, c=3
1 This uses positional parameters
2 This is an example of full by-name parameters
3 By-name parameters can be reordered as desired
4 You can use a mix of the two approaches
5 The non by-name parameters are resolved in the order they appear
Customizing the Named Parameter Key(s)

As seen above, the default strategy for deriving the key for a named parameter is to use the java name of the method signature and prefixing it with two dashes (--). This can be customized in two ways:

  1. to change the default prefix for the whole method, use the prefix() attribute of the @ShellMethod annotation

  2. to override the whole key on a per-parameter fashion, annotate the parameter with the @ShellOption annotation.

Have a look at the following example:

        @ShellMethod(value = "Display stuff.", prefix="-")
        public String echo(int a, int b, @ShellOption("--third") int c) {
                return String.format("You said a=%d, b=%d, c=%d", a, b, c);
        }

For such a setup, the possible parameter keys will be -a, -b and --third.

It is possible to specify several keys for a single parameter. If so, these will be mutually exclusive ways to specify the same parameter (so only one of them can be used). Here is an example:

        @ShellMethod("Describe a command.")
        public String help(@ShellOption({"-C", "--command"} String command) {
                ...
        }

Optional Parameters and Default Values

Spring Shell provides the ability to give parameters default values, which will allow the user to omit those parameters:

        @ShellMethod("Say hello.")
        public String greet(@ShellOption(defaultValue="World"} String who) {
                return "Hello " + who;
        }

Now, the greet command can still be invoked as greet Mother (or greet --who Mother), but the following is also possible:

shell:>greet
Hello World

Parameter Arity

Up to now, it has always been assumed that each parameter mapped to a single word entered by the user. Situations may arise though, when a parameter value should be multi valued. This is driven by the arity() attribute of the @ShellOption annotation. Simply use a collection or array for the parameter type, and specify how many values are expected:

        @ShellMethod("Add Numbers.")
        public float add(@ShellOption(arity=3) float[] numbers) {
                return numbers[0] + numbers[1] + numbers[2];
        }

The command may then be invoked using any of the following syntax:

shell:>add 1 2 3.3
6.3
shell:>add --numbers 1 2 3.3
6.3

When using the by-name parameter approach, the key should not be repeated. The following does not work:

shell:>add --numbers 1 --numbers 2 --numbers 3.3
Infinite Arity

TO BE IMPLEMENTED

Special Handling of Boolean Parameters

When it comes to parameter arity, there is a kind of parameters that receives a special treatment by default, as is often the case in command-line utilities. Boolean (that is, boolean as well as java.lang.Boolean) parameters behave like they have an arity() of 0 by default, allowing users to set their values using a "flag" approach. Take a look at the following:

        @ShellMethod("Terminate the system.")
        public String shutdown(boolean force) {
                return "You said " + force;
        }

This allows the following invocations:

shell:>shutdown
You said false
shell:>shutdown --force
You said true

This special treatment plays well with the default value specification. Although the default for boolean parameters is to have their default value be false, you can specify otherwise (i.e. @ShellOption(defaultValue="true")) and the behavior will be inverted (that is, not specifying the parameter will result in the value being true, and specifying the flag will result in the value being false)

Having this behavior of implicit arity()=0 prevents the user from specifying a value (e.g. shutdown --force true). If you would like to allow this behavior (and forego the flag approach), then force an arity of 1 using the annotation:

        @ShellMethod("Terminate the system.")
        public String shutdown(@ShellOption(arity=1, defaultValue="false") boolean force) {
                return "You said " + force;
        }

Quotes Handling

Spring Shell takes user input and tokenizes it in words, splitting on space characters. If the user wants to provide a parameter value that contains spaces, that value needs to be quoted. Both single (') and double (") quotes are supported, and those quotes will not be part of the value:

        @ShellMethod("Prints what has been entered.")
        public String echo(String what) {
                return "You said " + what;
        }
shell:>echo Hello
You said Hello
shell:>echo 'Hello'
You said Hello
shell:>echo 'Hello World'
You said Hello World
shell:>echo "Hello World"
You said Hello World

Supporting both single and double quotes allows the user to easily embed one type of quotes into a value:

shell:>echo "I'm here!"
You said I'm here!
shell:>echo 'He said "Hi!"'
You said He said "Hi!"

Should the user need to embed the same kind of quote that was used to quote the whole parameter, the escape sequence uses the backslash (\) character:

shell:>echo 'I\'m here!'
You said I'm here!
shell:>echo "He said \"Hi!\""
You said He said "Hi!"
shell:>echo I\'m here!
You said I'm here!

It is also possible to escape space characters when not using enclosing quotes, as such:

shell:>echo This\ is\ a\ single\ value
You said This is a single value

Interacting with the Shell

The Spring Shell project builds on top of the JLine library, and as such brings a lot of nice interactive features, some of which are detailed in this section.

First and foremost, Spring Shell supports TAB completion almost everywhere possible. So if there is an echo command and the user presses e, c, TAB then echo will appear. Should there be several commands that start with ec, then the user will be prompted to choose (using TAB or Shift+TAB to navigate, and ENTER for selection.)

But completion does not stop at command keys. It also works for parameter keys (--arg) and even parameter values, if the application developer registered the appropriate beans (see Providing TAB Completion Proposals).

Another nice feature of Spring Shell apps is support for line continuation. If a command and its parameters is too long and does not fit nicely on screen, a user may chunk it and terminate a line with a backslash (\) character then hit ENTER and continue on the next line. Uppon submission of the whole command, this will be parsed as if the user entered a single space on line breaks.

shell:>register module --type source --name foo  \ (1)
> --uri file:///tmp/bar
Successfully registered module 'source:foo'
1 command continues on next line

Line continuation also automatically triggers if the user has opened a quote (see Quotes Handling) and hits ENTER while still in the quotes:

shell:>echo "Hello (1)
dquote> World"
You said Hello World
1 user presses ENTER here

Lastly, Spring Shell apps benefit from a lot of keyboard shortcuts you may already be familiar with when working with your regular OS Shell, borrowed from Emacs. Notable shortcuts include Ctrl+r to perform a reverse search, Ctrl+a and Ctrl+e to move to beginning and end of line respectively or Esc f and Esc b to move forward (resp. backward) one word at a time.

Providing TAB Completion Proposals

TBD

Validating Command Arguments

Spring Shell integrates with the Bean Validation API to support automatic and self documenting constraints on command parameters.

Annotations found on command parameters as well as annotations at the method level will be honored and trigger validation prior to the command executing. Given the following command:

        @ShellMethod("Change password.")
        public String changePassword(@Size(min = 8, max = 40) String password) {
                return "Password successfully set to " + password;
        }

You’ll get this behavior, for free:

shell:>change-password hello
The following constraints were not met:
	--password string : size must be between 8 and 40 (You passed 'hello')
Applies to All Command Implementations

It is important to note that bean validation applies to all command implementations, whether they use the "standard" API or any other API, through the use of an adapter (see Supporting Other APIs)

Dynamic Command Availability

There may be times when registered commands don’t make sense, due to internal state of the application. For example, maybe there is a download command, but it only works once the user has used connect on a remote server. Now, if the user tries to use the download command, the shell should gracefully explain that the command does exist, but that it is not available at the time. Spring Shell lets the developer do that, even providing a short explanation of the reason for the command not being available.

There are three possible ways for a command to indicate availability. They all leverage a no-arg method that returns an instance of Availability. Let’s start with a simple example:

@ShellComponent
public class MyCommands {

    private boolean connected;

    @ShellMethod("Connect to the server.")
    public void connect(String user, String password) {
        [...]
        connected = true;
    }

    @ShellMethod("Download the nuclear codes.")
    public void download() {
        [...]
    }

    public Availability downloadAvailability() {
        return connected
            ? Availability.available()
            : Availability.unavailable("you are not connected");
    }
}

Here you see the connect method is used to connect to the server (details omitted), altering state of the command through the connected boolean when done. The download command will be marked as unavailable till the user has connected, thanks to the presence of a method named exactly as the command method with the Availability prefix in its name. The method returns an instance of Availability, constructed with one of the two factory methods. In case of the command not being available, an explanation has to be provided. Now, if the user tries to invoke the command while not being connected, here is what happens:

shell:>download
Command 'download' exists but is not currently available because you are not connected.
Details of the error have been omitted. You can use the stacktrace command to print the full stacktrace.

Information about currently unavailable commands is also leveraged in the integrated help. See Integrated Documentation with the help Command.

The reason provided when the command is not available should read nicely if appended after "Because …​"

It’s best not to start the sentence with a capital and not add a final dot.

If for some reason naming the availability method after the name of the command method does not suit you, you can provide an explicit name using the @ShellMethodAvailability, like so:

    @ShellMethod("Download the nuclear codes.")
    @ShellMethodAvailability("availabilityCheck") (1)
    public void download() {
        [...]
    }

    public Availability availabilityCheck() { (1)
        return connected
            ? Availability.available()
            : Availability.unavailable("you are not connected");
    }
1 the names have to match

Lastly, it is often the case that several commands in the same class share the same internal state and thus should all be available or unavailable all at one. Instead of having to stick the @ShellMethodAvailability on all command methods, Spring Shell allows the user to flip things around and put the @ShellMethodAvailabilty annotation on the availability method, specifying the names of the commands that it controls:

    @ShellMethod("Download the nuclear codes.")
    public void download() {
        [...]
    }

    @ShellMethod("Disconnect from the server.")
    public void disconnect() {
        [...]
    }

    @ShellMethodAvailability({"download", "disconnect"})
    public Availability availabilityCheck() {
        return connected
            ? Availability.available()
            : Availability.unavailable("you are not connected");
    }

The default value for the @ShellMethodAvailability.value() attribute is "*" and this serves as a special wildcard that matches all command names. It’s thus easy to turn all commands of a single class on or off with a single availability method.

Spring Shell does not impose much constraints on how to write commands and how to organize classes. But it’s often good practice to put related commands in the same class, and the availability indicators can benefit from that.

Built-In Commands

Any application built using the spring-shell-starter artifact (or, to be more precise, the spring-shell-standard-commands dependency) comes with a set of built-in commands. These commands can be overridden or disabled individually (see Overriding or Disabling Built-In Commands), but if they’re not, this section describes their behavior.

Integrated Documentation with the help Command

Running a shell application often implies that the user is in a graphically limited environment. And although, in the era of mobile phones we’re always connected, accessing a web browser or any other rich UI application such as a pdf viewer may not always be possible. This is why it is important that the shell commands are correctly self documented, and this is where the help command comes in.

Typing help + ENTER will list all the known commands to the shell (including unavailable commands) and a short description of what they do:

shell:>help
AVAILABLE COMMANDS
        add: Add numbers together.
      * authenticate: Authenticate with the system.
      * blow-up: Blow Everything up.
        clear: Clear the shell screen.
        connect: Connect to the system
        disconnect: Disconnect from the system.
        exit, quit: Exit the shell.
        help: Display help about available commands.
        register module: Register a new module.
        script: Read and execute commands from a file.
        stacktrace: Display the full stacktrace of the last error.

Commands marked with (*) are currently unavailable.
Type `help <command>` to learn more.

Typing help <command> will display more detailed information about a command, including the available parameters, their type and whether they are mandatory or not, etc.

Here is the help command applied to itself:

shell:>help help


NAME
	help - Display help about available commands.

SYNOPSYS
	help [[-C] string]

OPTIONS
	-C or --command  string
		The command to obtain help for.  [Optional, default = <none>]

Clearing the Screen

The clear command does what you would expect and clears the screen, resetting the prompt in the top left corner.

Exitting the Shell

The quit command (also aliased as exit) simply requests the shell to quit, gracefully closing the Spring application context. If not overridden, a JLine History bean will write a history of all commands executed to disk, so that they are available again (see Interacting with the Shell) on next launch.

Displaying Details about an Error

When an exception occurs inside command code, it is caught by the shell and a simple, one-line message is displayed so as not to overflow the user with too much information. There are cases though when understanding what exactly happened is important (especially if the exception has a nested cause).

To this purpose, Spring Shell remembers the last exception that occurred and the user can later use the stacktrace command to print all the gory details on the console.

Running a Batch of Commands

The script command accepts a local file as an argument and will replay commands found there, one at a time.

Reading from the file behaves exactly like inside the interactive shell, so lines starting with // will be considered as comments and ignored, while lines ending with \ will trigger line continuation.

Customizing the Shell

Overriding or Disabling Built-In Commands

Built-in commands are provided with Spring Shell to achieve everyday tasks that many if not all shell applications need. If you’re not happy with the way they behave though, you can disable or override them, as explained in this section.

Disabling all Built-in Commands

If you don’t need built-in commands at all, then there is an easy way to "disable" them: just don’t include them! Either use a maven exclusion on spring-shell-standard-commands or, if you’re selectively including Spring Shell dependencies, don’t bring that one in!

<dependency>
    <groupId>org.springframework.shell</groupId>
    <artifactId>{starter-artifactId}</artifactId>
    <version>{project-version}</version>
    <exclusions>
        <exclusion>
            <groupId>org.springframework.shell</groupId>
            <artifactId>spring-shell-standard-commands</artifactId>
        </exclusion>
    </exclusion>
</dependency>
Disabling Specific Commands

To disable a single built-in command, simply set the spring.shell.command.<command>.enabled property to false in the app Environment. One easy way to do this is to pass extra args to the Boot application in your main() entry point:

        public static void main(String[] args) throws Exception {
                String[] disabledCommands = {"--spring.shell.command.help.enabled=false"}; (1)
                String[] fullArgs = StringUtils.concatenateStringArrays(args, disabledCommands);
                SpringApplication.run(MyApp.class, fullArgs);
        }
1 This disables the integrated help command
Overriding Specific Commands

If, instead of disabling a command you’d rather provide your own implementation, then you can either

  • disable the command like explained above and have your implementation registered with the same name

  • have your implementing class implement the <Command>.Command interface. As an example, here is how to override the clear command:

    public class MyClear implements Clear.Command {
    
        @ShellCommand("Clear the screen, only better.")
        public void clear() {
            // ...
        }
    }
Please Consider Contributing your Changes

If you feel like your implementation of a standard command could be valuable to the community, please consider opening a pull-request at github.com/spring-projects/spring-shell.

Alternatively, before making any changes on your own, you can open an issue with the project. Feedback is always welcome!

ResultHandlers

PromptProvider

Customizing Command Line Options Behavior

ConversionService

Extending Spring Shell

Support for Spring Shell 1 and JCommander

Discovering Methods that Can Act as Commands

Resolving Parameter Values

Supporting TAB Completion