Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [] Why was there never another Logging JSR?

Yes, this was the time I mentioned, Markus was involved but he may have responded to what Antonio raised.

There were also brief discussions about a MP feature but quite frankly now that the JDK offered at least a very basic abstraction layer I think it makes more sense to use the new Java 9 feature where possible. As soon as a Jakarta EE version has the minimal JDK to be 9 (which could be already the case for Jakarta EE 9) it should be changed to something similar as we did in the JSR 385 API in a Multi-Release-JAR:  so the API is still backward compatible with Java 8.


On Mon, Apr 13, 2020 at 10:00 PM arjan tijms <arjan.tijms@xxxxxxxxx> wrote:

There were some more discussions for a Logging JSR. I think Antonio Goncalves started multiple discussions about it as well, and wrote some blogs. See e.g.

I remember the requirements for a potential Logging JSR were twofold: the ability to set the logging level of logical APIs (like set Faces logging, independent of whether it's implemented by MyFaces or Mojarra and whatever they are using for logging), and an API for applications running on Java EE to use for logging.

Kind regards,

On Mon, Apr 13, 2020 at 9:48 PM Werner Keil <werner.keil@xxxxxxxxx> wrote:

There was an interesting question on Twitter the other day, why beside configuration there had never been a JSR for logging which is also a fundamental aspect of many enterprise applications.

There was in the very early days of the JCP:, but java.util.logging never was vendor-neutral or extendable and unlike the Collections Framework which has many different implementations also at Eclipse I'm not aware any alternate implementation outside the JDK ever existed.

Markus Eisele thought about this when he worked at Red Hat for the first time, several years ago about a decade or so. 
However, the idea never got beyond initial discussions with involved vendors. And from what I heard back then it was a similar kind of vanity and pride by authors of several frameworks like Log4J, LogBack/SLF4J or Apache Commons Logging and of course the JDK team behind JUL. 
Every committer or team insisted, only their project could be the RI for such JSR, so eventually Markus gave up the idea.
Interestingly the JDK picked up the idea mostly to make itself a little more modular and simplify the dependency graph with Java 9:
I am not sure, what exact TCK support exists for this API, but it aims to be easily adoptable by applications which use external logging framework, such as SLF4J or Log4J, so that'll have to do for now, and it does not seem like a Jakarta standard or MP feature is really worth the effort in this case.


_______________________________________________ mailing list
To unsubscribe from this list, visit
_______________________________________________ mailing list
To unsubscribe from this list, visit

Back to the top