| Re: [jetty-users] ExpressionFactory performance issue with service discovery in javax.el 2.2.0.v201108011116 |
Hi Jan,Tests so far have been ok on a limited scope, we're putting it through its pace in the next couple few minutes and should have a canary in production following that. I should have some more concrete news in the next few hours and will let you know (and will update jira accordingly)-- stephaneOn Tue, Mar 5, 2013 at 12:55 AM, Jan Bartel <janb@xxxxxxxxxxx> wrote:
Stephane,
Any luck with testing the newer jars?? We're coming up to a release
and I'd like to get this sorted out as soon as possible ...
> Hi Jesse,
>
> Thanks for the reply.
>
> I went digging a little bit around, found that the javax.el-api 2.2.4 source
> code seems to have that particular problem fixed in the sense they don't
> cache the classes but rather they just lazily instantiate the factory
> statically only once in BeanELResolver. So different fix than the one in
> Tomcat land but that should fix our problem. We'll going to test that
> version. Will let you know.
>
> Available at
> http://search.maven.org/#artifactdetails%7Cjavax.el%7Cjavax.el-api%7C2.2.4%7Cjar
> It's that same version that is shipped with Glassfish 3.1.2.2 as far as I
> can see.
>
> -- stephane
>
>
> On Wed, Feb 27, 2013 at 11:40 AM, Jesse McConnell
> <jesse.mcconnell@xxxxxxxxx> wrote:
>>
>> we take the javax.el stuff from the jasper setup that glassfish uses,
>> think there is a jsp project for it at java.net now a days for it (this one
>> I think: http://jsp.java.net/)
>>
>> as for quick fix...we can look at updating the versions but between
>> getting it into the CQ process and then orbitifying it is anywhere from a 1
>> week to 4 month process, which is why are are noticeably slow updating that
>> :/
>>
>> If updating to the tomcat one works for you then that is definitely the
>> best short term solution. In the meantime the correct process here it to
>> open a bug in bugzilla under RT/Jetty component with details so we can start
>> the update process
>>
>> I'll note that you should be able to test updated jsp artifacts from the
>> glassfish jsp setup by replacing them and seeing how they work...giving us
>> an idea of what version resolves your problem will help
>>
>> cheers,
>> jesse
>>
>> --
>> jesse mcconnell
>> jesse.mcconnell@xxxxxxxxx
>>
>>
>> On Wed, Feb 27, 2013 at 10:31 AM, Stephane Bailliez <sbailliez@xxxxxxxxx>
>> wrote:
>>>
>>> I opened a few minutes ago:
>>>
>>> https://bugs.eclipse.org/bugs/show_bug.cgi?id=401916
>>>
>>> This is obviously causing significant performance impact and affect
>>> scaling, the only quick alternative I see would be to replace temporarily
>>> the orbit javax.el version with the one from tomcat.
>>> Short of that working would be to patch the orbit el code.
>>>
>>> This might seem a bit of a stretch as I'm not sure what is the
>>> significant difference without doing a diff between the 2 source codes. I'll
>>> have a look at that soonish, in the meantime I was wondering:
>>>
>>> Is there any significant difference in the codebase that you know of ?
>>> From where is coming the repackaged orbit: javax.el 2.2.0.v201108011116 ?
>>>
>>> Do you see a possible release with updated version of it "soonish" ?
>>>
>>> Thanks !
>>>
>>> -- stephane
>>>
>>> _______________________________________________
>>> jetty-users mailing list
>>> jetty-users@xxxxxxxxxxx
>>> https://dev.eclipse.org/mailman/listinfo/jetty-users
>>>
>>
>>
>> _______________________________________________
>> jetty-users mailing list
>> jetty-users@xxxxxxxxxxx
>> https://dev.eclipse.org/mailman/listinfo/jetty-users
>>
>
>
> _______________________________________________
> jetty-users mailing list
> jetty-users@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/jetty-users
>
--
Jan Bartel <janb@xxxxxxxxxxx>
www.webtide.com – Developer advice, services and support
from the Jetty & CometD experts.