Jump to: navigation, search

Equinox/p2/p2 index

What is the p2.index file?

The p2.index file provides a description of the kind of repository available at a particular location; simple or composite.

The presence of this file in a repository will cause p2 to replace its default lookup order for repository to match the one advised in the file. This only happens the first time a repository is accessed, as p2 then caches "what worked before" and uses that on subsequent accesses. See bug 347448 for more detail, links, and history behind this advisory file. The latest version of the b3 aggregator adds an appropriate p2.index file as it publishes its p2 repository (not sure about other p2 repository publishers).

For example the example p2.index contents below will cause p2 to look for compositeContent.jar, and if not found, then compositeContent.xml and would skip looking for content.jar or content.xml as it normally would look for them first. (And, then, naturally, if it needed artifacts from that site, would look for compositeArtifacts.jar and if not found compositeArtifacts.xml and would skip looking for artifacts.jar and artifacts.xml).

 version = 1
 metadata.repository.factory.order = compositeContent.xml,\!
 artifact.repository.factory.order = compositeArtifacts.xml,\!

How does it help me?

As a provider of a repository, it does not help you directly, but it helps your users to get a better experience and also slightly reduce the number of hits your server will be subject to.

What happens if I get it wrong ?

Should you get the content of this file wrong, the repository will fail loading. For example specifying compositeContent.xml where the repository is a content.xml will cause p2 to only look for compositeContent.xml and never look for the content.xml.

How many files for composite repositories?

Given that a composite repository is just a repository that refers to other repositories, the full benefit of p2.index can only be achieved if every child repo has the file (with only a few exceptions to this general rule) the but benefit is greatest, or most important, for composite sites as that is where the "default rules" are changed the most.

Why say 'jar' instead of 'xml' in factory name?

The content of p2.index does not reflect whether the files are zipped or not.

To quote the output of the b3 aggregator:

Please note that the values in this file denotes repository factories and
not files. The factory '<name>.xml' will look for both the <name>.jar
and the <name>.xml file, in that order

For example for a simple repository that contains artifacts.jar and content.jar, the p2.index would still name: "artifacts.xml" and "content.xml" as the factories.

As an aside, jarred versions of p2 content and artifacts files should always be provided for production sites, when the files are of any substantial size.

Example for composite repository

 version = 1
 metadata.repository.factory.order = compositeContent.xml,\!
 artifact.repository.factory.order = compositeArtifacts.xml,\!

Example for simple repository

 version = 1
 metadata.repository.factory.order = content.xml,\!
 artifact.repository.factory.order = artifacts.xml,\!