On 4 avr. 2012, at 09:17, "Brook, James" <
email@hidden> wrote:
>
>
> Sent from my iPhone
>
>>> - is the injection mechanism efficient (in our case, a new object is created each time there is an http request)?
>>
>> I still don't have any benchmarks comparing an application running with WOInject and without it. Our empirical analysis didn't capture any discrepancies over the application performance after the addition of WOInject. I'll try to produce meaningful numbers when I finish a refactoring in the WOInject initialization code. I'll probably replace the Javassist library by ASM, which can lead to smaller performance gains.
>
> We have been using Guice to inject per request in a live app for over a year. The requests are all ERRest with an in memory database behind. Even at load we see sub 10ms response times in WOStats.
>
>>
>> Anyway, if you want to check it out, you can configure one application to use WOInject in a few minutes. Even if you don't configure any Guice module, components, EOs, DirectActions and Sessions will be created by Guice. Other internal WebObjects classes will be instantiated via reflection, and I'll be injected after that. If you feel your application is slow after the change, let me know.
>>
>> [1]
https://github.com/hprange/woinject/issues/1
>>
>> Cheers,
>>
>> Henrique
>> _______________________________________________
>> Do not post admin requests to the list. They will be ignored.
>> Webobjects-dev mailing list (
email@hidden)
>> Help/Unsubscribe/Update your Subscription:
>>
>> This email sent to
email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (
email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to
email@hidden