Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [ee4j-community] EE4J Project for annotation scanning?

Arjan,

indeed I think that prescanning and including an index in META-INF would be an efficient implementation.   It's a big pity that this feature was not implemented in java 9 with the original modules, as such an index could have been required from day 1 and all tools that create modules would create the index.   However, that horse has already bolted and there are now modules in the wild that do not have such an index.   So any solution is going to need to handle those module - perhaps lazily generating an index if there is not one that is not baked into the module (or jar - hey let's fix it for non modules as well).

But I think discussing specific implementations is a bit premature as I think there first needs to be some agreement for a common effort and a common API.  If all the eventual outcome was just a document that described how modules/update/classpaths/jlinks/multi-releases are going to be resolved - then that would be a big step, as the java module system really lacks documentation and specification.   If we can agree on a resolution algorithm, then we can perhaps agree on an API.  Once we have an API and a spec, I don't really care if there is just one implementation or many.  Implementations might be based on jandex or asm or be part of java 10.

cheers


 

On 30 November 2017 at 13:54, arjan tijms <arjan.tijms@xxxxxxxxx> wrote:
Hi,

Incidentally, I've recently been thinking about the exact same thing.

What about generating an index like Jandex (https://github.com/wildfly/jandex) by the java compiler and store that in module-info.class?

Any Java environment needing annotations could then just ask the java runtime for it, with an option of manually scanning and/or using the JDK scanner algorithm at run time for those modules that don't have the index.

This would need to become a JEP then I guess.

Thoughts?

Kind regards,
Arjan Tijms





On Thu, Nov 30, 2017 at 1:33 PM, Markus KARG <markus@xxxxxxxxxxxxxxx> wrote:

If your intention is to simply provide some kind of commonly used library, Github is a good place to start. Once your project gains attention and you feel the need for some kind of official organization around it, start a project on Eclipse: https://www.eclipse.org/projects/handbook/#starting. If you found an existing Eclipse top level project that looks like what you want to be part of, ask them for adoption. :-)

-Markus

 

 

 

From: ee4j-community-bounces@eclipse.org [mailto:ee4j-community-bounces@eclipse.org] On Behalf Of Greg Wilkins
Sent: Donnerstag, 30. November 2017 13:28
To: EE4J community discussions
Subject: Re: [ee4j-community] EE4J Project for annotation scanning?

 

 

On 30 November 2017 at 13:26, Markus KARG <markus@xxxxxxxxxxxxxxx> wrote:

I do not see that annotation scanning is limited to enterprise use, so I would say no.

 

So where would you suggest such an effort is coordinated?

 


 

--


_______________________________________________
ee4j-community mailing list
ee4j-community@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/ee4j-community



_______________________________________________
ee4j-community mailing list
ee4j-community@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/ee4j-community




--

Back to the top