Fetching from Repositories
The infrastructure offered by PDE Build provides steps to fetch the
source code and pre-built plug-ins from a repository as part of the build process (see the eclipse.fetch task
and the properties to control fetching).
Two possibilities are available: let PDE build automatically fetch the
source, or fetch the source code manually.
First step to automation, the map file
Given the identifier of a feature or a plug-in that needs to be built, the build
infrastructure is able to fetch from a repository all the included features and
plug-ins.
In order to know where to get things from, PDE uses map files. A map
file is a java property file which maps feature and plug-in ids to a
location and a tag in a repository. The format of map files is open and but
the key part is fixed by PDE Build:
<elementType>@<elementID>[,<elementVersion>] = <repo specific content>
where elementType is one of bundle, feature,
plugin or fragment, and elementId and elementVersion give the specifics of the element.
Map file entry for CVS
The format of a map file entry to fetch content from CVS is the following:
where args is a comma-separated list of key/value pairs described below:
- tag, the CVS tag to check out;
- cvsRoot, a CVS connection string;
- path, an optional CVS module and path where to get the folder containing the element. By default the element name is used as module name;
- prebuilt, an optional boolean indicating if the element downloaded is prebuilt. By default this is false;
- cvsPassFile, an option path to a cvs pass file;
- password, an optional password to use to connect to CVS;
Example:
plugin@com.ibm.icu,3.6.1=CVS,tag=v200704191630,cvsRoot=:pserver:anonymous@dev.eclipse.org:/cvsroot/tools,path=org.eclipse.orbit/com.ibm.icu/bin,prebuilt=true
Old style Map Entries |
In Eclipse 3.2 and earlier, the map file format for CVS was:
<elementType>@<elementID> = CVS, <TAG>, <CVSROOT>[,<PASSWORD>[,<PATH>[,<CVSPASSFILE>]]]
If any of the optional arguments aren't specified, then the commas are still required to delimit any following arguments. This format should be considered deprecated and the key/value format is prefered.
|
Map file entry for Ant GET
The format of a map file entry to fetch content from any URL supported by Ant GET is the following:
<key> = GET, <url> [,args]
where url is the url to retrieve the data from and args is an optional comma-separated list of key/value pairs described below:
- unpack, set to true the indicate that the downloaded element should be unzipped;
- verbose, set to true to show verbose progress information when downloading. The default value is false;
- ignoreerrors, set to true to log errors during the transport. The default value is false;
- usetimestamp, conditionally download a file based on the timestamp of the local copy (for http URLs only) when set to true. The default value is false;
- username, the username for "BASIC" http authentication;
- password, the password to perform the download.
Example:
plugin@com.ibm.icu,3.4.5=GET,https://download.eclipse.org/tools/orbit/downloads/drops/S200705301823/bundles/com.ibm.icu_3.4.5.jar, unpack=true
A new extension has been added to PDE/Build that enables users to fetch artifacts from p2 repositories. You can now add an
entry to your map file for build input and PDE/Build will retrieve the artifact and include it in the build. The map file entry needs to include the
id and version of the installable unit (IU), and the URI of the repository where the IU can be found. The metadata and artifact repositories are
assumed to be co-located.
The format of a map file entry to fetch content via p2 is as follows:
<elementType>@<elementID> = p2IU, id=<id>, version=<version>, repository=<repository>
- id, the identifier of the Installable Unit;
- version, the version of the Installable (optional). If not specified, the highest version available will be fetched.
- repository, the location (URI) of the repository where the IU will be found.
Fetching features with p2
In the Helios release, features can be fetched with p2 by specifying the Installable Unit representing the feature jar. See the feature metadata page
for details on the structure of p2 metadata for features.
Examples:
plugin@my.bundle.id,1.0.0=p2IU,id=my.bundle.id,version=1.0.0,repository=http://example.eclipse.org/repo
feature@my.feature.id = p2IU,id=my.feature.id.feature.jar, version=1.0.0,repository=http://example.eclipse.org/repo
Map file entry for other repositories
PDE Build provides an extension point
where fetch script generators for different repositories can be plugged in. When your eclipse
install contains a bundle that provides a script generator extension
for a different type of repository, you can use that repository by
specifying it in your map file entries and adhering to the format specified by the extension provider.
Setting up for CVS source fetching
To set up fetching your source from CVS, you should do the following:
- Make sure you have a cvs.exe on your system search path.
(Download CVS from here).
- Create a folder called maps in the build directory.
- Create a file with extension .map in the maps folder.
- Fill in this map file with all the elements that need to be
fetched from a repository.
- Edit the build.properties file from the configuration folder and
comment out the line skipFetch=true. If that property
is set, then PDE build will skip the entire fetch phase. If you
want the tag specified in the map files to be ignored you can set the
property fetchTag to a specific value. This is useful when
doing a nightly build from HEAD instead of from the tagged versions.
Getting the map files from CVS
Map files are usually stored in a repository. PDE Build offers default
infrastructure to get those map files from a CVS repository.
To enable this function change the following properties in your
configuration's build.properties from your configuration directory:
- skipMaps: Comment
out this property. If this property is set, then PDE build will
not fetch the map files.
- mapsRepo =
:pserver:anonymous@example.com/path/to/repo. The CVS repository
from which to get the map files.
- mapsRoot =
/path/to/maps. The path in the CVS repository to the directory
containing the map files.
- mapsCheckoutTag =
HEAD. The CVS tag to use to checkout the map files.
Fetching the map files from CVS occurs during the Pre-Build phase of
the build. Scripts to
fetch all the features and plug-ins included in the feature you are
building will be generated and
run during the fetch phase of the build.
Automatic fetching from other repositories
If you are automatically fetching your map files from the
repository, you will need to copy the customTarget.xml file from
org.eclipse.pde.build/templates/headless-build into your configuration
directory. The target getMapFiles is used to fetch the map
files, this should be modified to fetch from your repository.
Getting source code manually
You can write custom fetch targets to retrieve your plug-ins and
features which can be invoked from
the preSetup or postSetup targets in the customTargets.xml file.
Features and plug-ins should
be fetched to
${buildDirectory}/features
and
${buildDirectory}/plugins
respectively.