I have no memories about this and fully defer to you guys to figure it out.

IIRC, the content.xml is tweaked in some further steps, and since p2 preferes xml.xz or jar, if those files are still present and only the content.xml is tweaked, then the tweaks will not be found by consumers (as .xml.xz without trick will be processed instead)

IIRC, the content.xml is tweaked in some further steps, and since p2 preferes xml.xz or jar, if those files are still present and only the content.xml is tweaked, then the tweaks will not be found by consumers (as .xml.xz without trick will be processed instead)

do you mean that still we need to retain content.xml with jar and xz ?

I'm traveling so have no time for reviews. I assume that nothing (no script) is altering the contents of content.jar after it is produced by p2. A "normal" simple p2 repository has just these files:

https://download.eclipse.org/releases/2024-09/202409111000/

The very next mojo execution execution in pom.xml with id generate-extra-iu tweaks the content.xml file after it's produced. As long as we use it, we need to clear the .jar and .xml.xz (and optionally recreate them later from the altered content.xml).
There might be better ways to do that, such as using p2 instructions in a p2.inf file. If we can leverage such more standard approach, then we may be more easily able to get rid of some executions.

The very next mojo execution execution in pom.xml with id generate-extra-iu tweaks the content.xml file after it's produced. As long as we use it, we need to clear the .jar and .xml.xz (and optionally recreate them later from the altered content.xml). There might be better ways to do that, such as using p2 instructions in a p2.inf file. If we can leverage such more standard approach, then we may be more easily able to get rid of some executions.

Shall we then try to extract the content.jar to another target ? and retain .jar and .xz

IMO you should investigate why the jar has to be extracted and it's content manipulated after that and remove the need for that.

IMO you should investigate why the jar has to be extracted and it's content manipulated after that and remove the need for that.

Sure 👍

I'm closing this one as the real change will have to be quite different.