Today, I uploaded the release version of 3.0.0 of the XPages Jakarta EE Support project. It's been proving stable in my use since the last beta, and so I think this is as good a time as any to release it properly.

Changes

The big-ticket change remains the move to Jakarta EE 10 as the baseline, which brings a handful of new features as well as a new Java version requirement. That means that this release also requires at least Domino 14. Domino 12.x served us well, but its time has passed.

Jakarta EE 10, for its part, is mostly about solving a lot of old business in the JEE community: it continues the gradual deprecation of EJB in favor of CDI, it removes some old stuff like applet requirements, and then also brings in a couple "scratch an itch" features.

Of particular note is the addition of the EntityPart type for REST services. Though it's a small feature, it's a real "finally" one, in that there hadn't been a proper way to deal with multipart/form-data MIME body parts individually, and so each implementation of Jakarta REST would bring in its own, or you'd have to fall back to taking an InputStream and parsing the MIME body yourself. Now, you can do so in a spec-based way:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11

@POST
@Consumes(MediaType.MULTIPART_FORM_DATA)
public String post(@FormParam("part") EntityPart part) throws IOException {
	String mediaType = part.getMediaType();
	String name = part.getName();
	String fileName = part.getFileName();
	MultivaluedMap<String, String> headers = part.getHeaders();
	byte[] data = part.getContent().readAllBytes();
		
	// ...
}

There's also the split of Jakarta NoSQL into that spec plus the new Jakarta Data. In this release of XPages JEE, I mostly aimed to keep the same level of functionality while accounting for the renaming of packages and types, but I'll be interested in building on this in the future.

Finally, there's the project-specific change of condensing the many, many XPages libraries just down to Core, UI, and MicroProfile. That didn't impact functionality as such, but it sure is nice only having three (or six, with the source features) features to say "yes, install this" to when updating it in Designer and only three to check in Xsp Properties. It also allowed me to delete a lot of weird shim and conditional code, and it'll make maintenance of it much easier in the future without having to worry about every permutation of what libraries you have enabled in an NSF.

The Future

Speaking of which, that brings me to some of the next things in the docket. I imagine that the immediate work will be cleaning up any loose ends from the move. For example, Jakarta Concurrency 3.0 brought a bunch of new features, but I haven't actually checked to see if they work or if I need to do more adapting.

Additionally, Jakarta Data is intended to go beyond just NoSQL, and can also layer on top of Jakarta Persistence (née JPA, the API for working with relational DBs) and arbitrary services. I don't know yet if there's a usable implementation beyond the one in Open Liberty, so that may have to wait, but it'll be interesting to tinker with.

There are also a bunch of features I'd like to get cracking on now that this hurdle is done. For example, I'd like to move the NoSQL driver to use JNX, which would let me do a couple things that the Notes.jar classes just can't. Along with that, I'd like to add an option to publish MP Metrics to the Domino statistics store, which adopting JNX would allow easily.

Fortunately, I don't expect that there will be any other breaking-change discontinuities in the future. Jakarta EE 11 has some deprecations and removals, but it's mostly similar to JEE 10 in that it's about classes and idioms that are much, much older than any code built using this project. That should give this 3.x series a good, long period to be a comfortable baseline, even though the next major version of JEE.