You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@commons.apache.org by henrib <he...@apache.org> on 2015/12/04 09:59:35 UTC
Re: [jexl] 3.0 release date?
Hi;
It still seems no member is interested in publishing an official Apache
release.
I gave up trying to release after a failed attempt at 2.1.2 where the only
help available was antagonistic :-).
That said, the 3.0 trunk is stable and used in production.
If you're allowed to use internal builds, pull the source, 'mvn site' and
you're up.
You can also 'fork' the source code, change the package name (from
apache.org to your.own.xyz) and build your own, flanking the issue of using
a not-released external component. I've got an ant-run skeleton you can add
to the pom.xml that automates the renaming process if you're interested. It
would actually be fairly easy to publish a non-apache version of jexl3 to
sonatype using this procedure.
Let me know if you need more info on the workarounds.
Regards,
Henrib
--
View this message in context: http://apache-commons.680414.n4.nabble.com/jexl-3-0-release-date-tp4680703p4680972.html
Sent from the Commons - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@commons.apache.org
For additional commands, e-mail: user-help@commons.apache.org
Re: [jexl] 3.0 release date?
Posted by Balázs Zsoldos <ba...@everit.biz>.
>
> AST nodes also do cache getters/setters/methods Executors.
Thanks for the info. I will have a look. I used MVEL before that is about 3
times faster with its ReflectionOptimizer (based on my stress tests).
However, its source code is impenetrable so I decided to go with JEXL and
check if the same optimizations could be applied here.
On Fri, Dec 4, 2015 at 11:06 AM, henrib <he...@apache.org> wrote:
> Hi again;
> On the optimization topic, I'm not sure you'll gain as much as you expect
> since there is already caching performed during classes introspection and
> the AST nodes also do cache getters/setters/methods Executors.
> Cheers,
> Henrib
>
>
>
> --
> View this message in context:
> http://apache-commons.680414.n4.nabble.com/jexl-3-0-release-date-tp4680703p4680974.html
> Sent from the Commons - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@commons.apache.org
> For additional commands, e-mail: user-help@commons.apache.org
>
>
Re: [jexl] 3.0 release date?
Posted by henrib <he...@apache.org>.
Hi again;
On the optimization topic, I'm not sure you'll gain as much as you expect
since there is already caching performed during classes introspection and
the AST nodes also do cache getters/setters/methods Executors.
Cheers,
Henrib
--
View this message in context: http://apache-commons.680414.n4.nabble.com/jexl-3-0-release-date-tp4680703p4680974.html
Sent from the Commons - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@commons.apache.org
For additional commands, e-mail: user-help@commons.apache.org
Re: [jexl] 3.0 release date?
Posted by Balázs Zsoldos <ba...@everit.biz>.
Hi,
thanks for the answer!
I think I will do the fork and rename the groupId and the package to
"org.everit" and do a release to maven-central.
Also, I would be interested in some optimization that needs a bit of
refactor. Probably JEXL guys will take that back later when JEXL project is
more active and that could be JEXL 4.0.
The optimization:
- If the types of variables are specified in advance, method and field
instances of global and local variables could be cached. My guess is that
this would bring 2-3X performance increase
- I would also return the type of the newly created global variables
somehow so another expressions that use them can also work with this
"strict-type" mode
- Question: Why could not we find out the type by just calling
getClass() on the variable object? Answer: The top type should be used that
is visible for the classloader, otherwise problems would occur in OSGi
environments. Also, there could be issues if some magical technology
created the variable objects that do runtime class inheritance.
Kind regards,
Balazs
On Fri, Dec 4, 2015 at 9:59 AM, henrib <he...@apache.org> wrote:
> Hi;
> It still seems no member is interested in publishing an official Apache
> release.
> I gave up trying to release after a failed attempt at 2.1.2 where the only
> help available was antagonistic :-).
>
> That said, the 3.0 trunk is stable and used in production.
> If you're allowed to use internal builds, pull the source, 'mvn site' and
> you're up.
> You can also 'fork' the source code, change the package name (from
> apache.org to your.own.xyz) and build your own, flanking the issue of
> using
> a not-released external component. I've got an ant-run skeleton you can add
> to the pom.xml that automates the renaming process if you're interested. It
> would actually be fairly easy to publish a non-apache version of jexl3 to
> sonatype using this procedure.
>
> Let me know if you need more info on the workarounds.
> Regards,
> Henrib
>
>
>
>
>
>
> --
> View this message in context:
> http://apache-commons.680414.n4.nabble.com/jexl-3-0-release-date-tp4680703p4680972.html
> Sent from the Commons - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@commons.apache.org
> For additional commands, e-mail: user-help@commons.apache.org
>
>