Is a jextracted jar supposed to be published?

[Sebastian Stenzel]

> On 28. Jan 2021, at 13:25, Maurizio Cimadamore <maurizio.cimadamore at oracle.com> wrote:
> 
> Yes, the source generation is eager, while the class generation is as
> lazy as possible, using every inch of VM features we have available
> there. We could do more on the source side, to make things lazier
> there, but we're at the fence - kind of hoping for this:
> 
> https://openjdk.java.net/jeps/8209964 <https://openjdk.java.net/jeps/8209964>
> 
> If we had this, then we could just use same strategy for both class AND
> source and they would be both "naturally" lazy.
> 
> We do not have plans to remove the --source flag.

Above is a quote from the "Feedback / query on jextract for Windows 10" thread, which I didn't want to hijack. So I thought to create a new one instead. So far I have always been working with `--source`, however after reading the aforementioned, I'm reconsidering.

However, I really don't understand how jextracted .jar files are intended to be used. Without further artifacts such as a source jar, we can not just straightaway publish them to public repos like Maven Central, which effectively disqualifies them to be used as a dependency as is.

This brings me to the question: Is the jextracted jar file even intended to be "released" or should I consider it just an intermediate build artifact that _requires_ post-processing such as re-packaging (i.e. using maven-shade-plugin)?

I know that this question may be a bit off-topic, but since there seem to be significant differences between generated sources and classes, I thought I should bring up these "downstream problems" before it is too late.