Given that OpenNLP 2.x (J17) might maintained and around for some more months 
(n = ~6 once 3.x is out?), Tika has an option to get a compatible, yet update 
release.

As mentioned by Jeff, other ASF projects such as Lucene and Solr went forward 
already.

In view of the opinions and feedback received (2x +1, 1x +0, zero negative 
comments), and given that we don’t have a (real) show stopper / blocker, 
I’d like to conclude we have a consensus to migrate OpenNLP 3.x (*main*) to a 
baseline of Java 21.

Let’s move on towards 3.0.0-M2.

Best
Martin | mawiesne

> Am 05.03.2026 um 14:28 schrieb Jeff Zemerick <[email protected]>:
> 
> I'm +1 as long as we don't "out run" projects that depend on OpenNLP.
> I think Lucene 10 requires Java 21, and Solr 10 will also require Java
> 21. I believe Tika is still on Java 17...? Are there any other notable
> projects we should consider?
> 
> Thanks,
> Jeff
> 
> On Thu, Mar 5, 2026 at 1:25 AM Martin Wiesner <[email protected]> wrote:
>> 
>> Hi all,
>> 
>> I’m +1 for Java 21 and getting the new language features into the OpenNLP 
>> 3.x release line on our main branch.
>> 
>> For OpenNLP 2.x, I’m fine if we stick with Java 17 for users that depend on 
>> that.
>> 
>> Best
>> Martin | mawiesne
>> 
>>> Am 02.03.2026 um 20:46 schrieb Richard Zowalla <[email protected]>:
>>> 
>>> I am +0 for Java 21
>>> 
>>> Am Montag, dem 02.03.2026 um 20:44 +0100 schrieb Richard Zowalla:
>>>> Hi all,
>>>> 
>>>> We have an open issue that may be relevant for 3.0.0-M2:
>>>> https://issues.apache.org/jira/browse/OPENNLP-1735 (Java 21)
>>>> 
>>>> Martin and I discussed this offline. While I personally favor keeping
>>>> Java 17 as the baseline (since I am a EE folk and EE11 is still Java
>>>> 17
>>>> baseline), I understand the concerns: more and more tools are moving
>>>> toward Java 21 (recently checkstyle), and we may want to stay aligned
>>>> with that ecosystem and potentially take advantage of the latest
>>>> language features available in Java 21.
>>>> 
>>>> Although we could use multi-release JARs to leverage newer bytecode
>>>> versions, this would likely add some complexity to our build process.
>>>> 
>>>> For that reason, I’d like to start a discussion on whether we should
>>>> move to Java 21 for the 3.x release line and drop Java 17 support.
>>>> 
>>>> Copying users@ for information. Feel free to provide your opinion.
>>>> 
>>>> What do you think?
>>>> 
>>>> Gruß
>>>> Richard
>>> 
>> 

Reply via email to