|The build always produces jars for features.|
Maintaining the features in jar format once installed is not possible and even if it was it would be equivalent to not have any features there at all since as Katya mentioned several pieces of code expect to find the feature.xml on disk (e.g. PDE, Update configurator, I believe intro support and other obscure branding related things).
Note that it is also possible for you to ship without the features at all.
On 2011-10-05, at 12:19 PM, Todorova, Katya wrote:
Eclipse unzips every feature that contains touchpoint instruction “zipped” in its metadata. So theoretically speaking, if you publish your feature without this instruction, after installation you’ll have a jar instead of folder. However, it seems that having jars in the “features” folder leads to some issues with Eclipse update configurator – it expects to look for a feature.xml in a folder. Have in mind that even if update configurator presence is not critical in your environment, there could be other Eclipse components that may rely on feature being unzipped. In short, it’s possible but could lead to strange issues. From: p2-dev-bounces@xxxxxxxxxxx [mailto:p2-dev-bounces@xxxxxxxxxxx] On Behalf Of James Perry
Sent: mercredi 5 octobre 2011 16:25
Subject: [p2-dev] Installing features using P2 and maintaining JAR output
The Rational UA Infrastructure team is working with P2 installs as part of our builds, and one thing we noticed is that when P2 installs a feature (even if the feature is a .jar), once the feature is installed into the Eclipse environment, it is installed as a folder - not a JAR. Is there a switch or other way to maintain the feature as a .JAR once it is installed? The reason I am asking is we are building a site.xml file based on the contents of our builds, and we want the site.xml to have JARs in it, not folders.
Any help is greatly appreciated.
p2-dev mailing listp2-dev@xxxxxxxxxxxhttps://dev.eclipse.org/mailman/listinfo/p2-dev