Packaging, Extracting and transferring
To understand how Courier works, and to use its API, you must know and understand 3 central concepts Courier uses:
These 3 concepts are what enables Courier to perform deployments in a way that can decouple the Courier client from the sites/destinations it's deploying to/from.
The descriptions below are from an API point of view and provide examples on how to use each concept.
Packaging collects data and files from a given target repository that supports packaging. Out of the box, any Courier enabled website supports this, however, you cannot perform a packaging on a network share or subversion repository.
When an item has been packaged its data is serialized to an xml file and stored at a given destination. For storage you can use the local storage on any repository the client is connected to.
So to take packaging step by step:
- A Courier client is started and given a manifest of items to package
- The client connects to a given source site and starts retrieving data
- During packaging, the client will find dependencies and resources and add these to its queue
- When each item is packaged, it's saved as a xml file at a given destination.
To translate this into code, we need a couple of things defined:
- The client runs on "Machine A"
- The source repository is "devsite"
- The destination repository is "qasite"
So the code we are executing is not on either the devsite or qasite, but it could be. If no source or destination is set, Courier will try to use the local machine for storage.
For this configuration the following two repositories have been set up in the
<repositories> <repository name="QA site" alias="qasite" type="CourierWebserviceRepositoryProvider" visible="true"> <url>http://qa.local</url> <login>admin</login> <password>1234</password> <passwordEncoding>Hashed</passwordEncoding> </repository> <repository name="DEV site" alias="devsite" type="CourierWebserviceRepositoryProvider" visible="true"> <url>http://dev.local</url> <login>admin</login> <password>1234</password> <passwordEncoding>Hashed</passwordEncoding> </repository> </repositories>
Notice the Webservice repository type is used for both. However, you can use any supported repository as your destination.
The following namespaces are required to work with the sample. Only Courier Core should be needed.
using Umbraco.Courier.Core; using Umbraco.Courier.Core.Storage; using Umbraco.Courier.Core.Packaging; using Umbraco.Courier.Core.Collections.Manifests;
Connecting to destination and source
RepositoryStorage to retrieve repositories from the
var rs = new RepositoryStorage(); Repository destination = rs.GetByAlias("qasite"); Repository source = rs.GetByAlias("devsite");
Creating a new RevisionPackaging
To create a new Packaging operation you need to specify a name of the data you are working with (this defines where the files are stored) and connect it with the two repositories.
var engine = new RevisionPackaging(Revision); engine.Source = source; engine.Destination = destination;
If you want to allow Courier to perform comparison checking against a destination, you can enable this. This means that Courier will do a hashed comparison of all items to determine if they are needed in the revision. This can save time, but it should only be used if you know the destination won't change before the extraction happens.