Benchmark

[tandraschko]

i just have seen your project and i hope you can provide a benachmark soon

in the past i have used JUEL, which provided better performance OOTB compared to glassfish or tomcat EL.

i also implemented a own ELResolver to use MethodHandles in MF: https://github.com/apache/myfaces/blob/master/api/src/main/java/org/apache/myfaces/core/api/shared/lang/PropertyDescriptorUtils.java https://github.com/apache/myfaces/blob/master/impl/src/main/java/org/apache/myfaces/el/resolver/LambdaBeanELResolver.java

The performance is indeed better but almost not possible to measure the difference in a JSF application, even on a very big view

[Thihup]

Hi @tandraschko!

Sure. This is something I'd like to do. I'm not sure exactly how to measure it, and having some Jakarta Faces application would be a greater integration test (to see if it impacts).

However, only using the TCK is not sufficient to provide the real use cases for a Jakarta Expression Language implementation, so if you could provide some examples of uses I'd really appreciate it.

Thanks for taking a look in the project 😉

[tandraschko]

i think a good way for testing is the PrimeFaces showcase but we are still on javax for now

i did performance testing by a view with many many EL resolutions in a repeat, see: https://github.com/tandraschko/quarkus-myfaces/blob/master/quarkus-myfaces-showcase/src/main/resources/META-INF/resources/methodHandleELResolver.xhtml https://github.com/tandraschko/quarkus-myfaces/blob/master/quarkus-myfaces-showcase/src/main/java/io/quarkus/myfaces/showcase/view/MethodHandleELResolverBean.java

the project itself uses quarkus, so better just take the view and bean as example