The Sling Launchpad is a ready-to-run Sling configuration, providing an embedded content repository, a web server, a selection of Sling modules, documentation and examples. The Launchpad makes it easy to get started with Sling and to develop script-based applications.
This page will help you get started with the Launchpad. Fifteen minutes should be enough to get an overview of what Sling does.
While simple to run and understand, the Launchpad is a full-featured instance of Sling, an example configuration that we have created with the most common modules and configurations. The full functionality of Sling is available by loading additional Sling modules as needed.
More Sling samples can be found under https://github.com/apache/sling-samples/
Once you grok the basic examples of this page, we recommend studying the slingbucks and espblog samples. Both have README files with more info.
You'll need a running Sling application. If you haven't started Sling yet, have a look at the Getting Started page on how to run Sling. If you follow the instructions on that page, you should have Sling running on port 8080. If you choose a different port, replace the port 8080 in the examples below with your port.
To show the simplicity of the REST-style approach taken by Sling the examples below will be using cURL. Any HTTP client would do, but cURL is the easiest to document in a reproducible way.
Once started, look at http://localhost:8080/system/console/bundles with your browser. Use admin with password admin if Sling asks you for a login. Sling then displays the Felix Web Management Console page.
On the bundles page, all bundles should be marked Active. They're all OSGi bundles powered by Apache Felix, but that doesn't really matter to us right now.
Log files: If things go wrong, have a look at the sling/logs/error.log
log file - that's where Sling writes any error messages.
Until we have ready-to-test forms, you can create content with cURL, or you can create an HTML form that posts to the specified URL.
To create a content node (nodes are a JCR concept, a unit of storage) with cURL, use:
curl -u admin:admin -F"sling:resourceType=foo/bar" -F"title=some title" http://localhost:8080/content/mynode
The resulting node can be seen at http://localhost:8080/content/mynode.html, or as json format under http://localhost:8080/content/mynode.json. Lets try with cURL:
$ curl http://localhost:8080/content/mynode.json
This returns the properties of the /content/mynode
in JSON format as we have created it above.
{"title":"some title","sling:resourceType":"foo/bar","jcr:primaryType":"nt:unstructured"}
The additional property jcr:primaryType
is a special JCR property added by the content repository, indicating the JCR primary node type.
Monitoring requests: Sling provides a simple tool (an OSGi console plugin) to monitor HTTP requests, which helps understand how things work internally. See the Monitoring Requests page for details.
Sling uses scripts or servlets to render and process content.
Several scripting languages are available as additional Sling modules (packaged as OSGi bundles that can be installed via the Sling management console), but the launchpad currently includes the ESP (server-side ECMAscript), JSP (Java Server Pages), and Groovy language modules by default.
To select a script, Sling uses the node's sling:resourceType property, if it is set.
That is the case in our example, so the following script will be used by Sling to render the node in HTML, if the script is found at /apps/foo/bar/html.esp in the repository.
<html>
<body>
<h1><%= currentNode.title %></h1>
</body>
</html>
To select the script, Sling:
Store this script under /apps/foo/bar/html.esp, either using a WebDAV client (connected to http://admin:admin@localhost:8080/), or using cURL as shown here, after creating the html.esp script in the current directory on your system:
curl -X MKCOL -u admin:admin http://localhost:8080/apps/foo
curl -X MKCOL -u admin:admin http://localhost:8080/apps/foo/bar
create a local file html.esp and copy above content.
curl -u admin:admin -T html.esp http://localhost:8080/apps/foo/bar/html.esp
The HTML rendering of your node, at http://localhost:8080/content/mynode.html, is now created by this ESP script. You should see the node's title alone as an <h1> element in that page.
A script named POST.esp instead of html.esp would be called for a POST request, DELETE.esp for DELETE, xml.esp for a GET request with a .xml extension, etc. See URL to Script Resolution on the Sling wiki for more info.
Servlets can also be easily "wired" to handle specific resource types, extensions, etc., in the simplest case by using SCR annotations in the servlet source code. Servlets and scripts are interchangeable when it comes to processing Sling requests.
These simple examples show how Sling uses scripts to work with JCR data, based on sling:resourceType or node types.
There's much more to Sling of course - you'll find some additional simple examples below, as well as above in the see also section.
To create a node with a unique path at a given location, end the URL of the POST request with /.
In this case, the Sling response redirects to the URL of the created node.
Start by creating a new /blog folder:
curl -X POST -u admin:admin "http://localhost:8080/content/blog"
And create a node with a Sling-generated name under it:
curl -D - -u admin:admin -F"title=Adventures with Sling" "http://localhost:8080/content/blog/"
Using cURL's -D option shows the full HTTP response, which includes a Location header to indicate where the new node was created:
Location: /blog/adventures_with_slin
The actual node name might not be adventures_with_slin - depending on existing content in your repository, Sling will find a unique name for this new node, based on several well-know property values like title, description, etc. which are used for this if provided.
So, in our case, our new node can be displayed in HTML via the http://localhost:8080/blog/adventures_with_slin.html URL.
Note that we didn't set a sling:resourceType property on our node, so if you want to render that node with a script, you'll have to store the script under /apps/nt/unstructured/html.esp.
The sling.include function can be called from scripts to include the rendered result of another node.
In this example, we create a node at /content/header, rendered with a logo using an html.esp script, then use that header at the top of the html.esp script that we created previously for the foo/bar resource type.
Start by checking that http://localhost:8080/content/mynode.html is rendered using the html.esp script created above.
Create this script and name it header.esp:
<div>
<p style="color:blue;">
<img src="/images/sling.jpg" align="right"/>
<%= currentNode.headline %>
</p>
</div>
Upload it so that it is used to render resources having sling:resourceType=foo/header:
curl -X MKCOL -u admin:admin http://localhost:8080/apps/foo/header/
curl -u admin:admin -T header.esp http://localhost:8080/apps/foo/header/html.esp
Create the header node:
curl -u admin:admin -F"sling:resourceType=foo/header" -F"headline=Hello, Sling world" http://localhost:8080/content/header
Upload the logo that the script uses (using sling.jpg or another logo in the current directory):
curl -X MKCOL -u admin:admin http://localhost:8080/images/
curl -u admin:admin -T sling.jpg http://localhost:8080/images/sling.jpg
And check that the header is rendered with the logo at http://localhost:8080/content/header.html.
Now, update the html.esp script that we created for our first example above, to include the header:
<html>
<body>
<div id="header">
<% sling.include("/content/header"); %>
</div>
<h1><%= currentNode.title %></h1>
</body>
</html>
And upload it again to replace the previous version:
curl -u admin:admin -T html.esp http://localhost:8080/apps/foo/bar/html.esp
The http://localhost:8080/content/mynode.html, once refreshed, now shows the blue headline and logo, and this layout also applies to any node created with sling:resourceType=foo/bar.