Optional elresolver updates

[markt-asf]

Fixes Issue
No issue. The EL spec has been updated. One change and some additional tests are required.

Related Issue(s)
None.

Describe the change
Updates the existing OptionalELResolver tests for property resolution.
Adds new OptionalELResolver tests for method invocation.
Cleans up the code for the OptionalELResolver tests.

Additional context
None.

CC @alwin-joseph @anajosep @arjantijms @cesarhernandezgt @dblevins @m0mus @edbratt @gurunrao @jansupol @jgallimore @kazumura @kwsutter @LanceAndersen @bhatpmk @RohitKumarJain @shighbar @gthoman @brideck @OndroMih @dmatej
@starksm64 @scottmarlow

[alwin-joseph]

Will verify when new API jar is released after jakartaee/expression-language#208 is merged.

[markt-asf]

Tx. 6.0.0-RC1 has been published that includes that change.

[alwin-joseph]

I encountered 1 test failure , using GlassfishM2 (and copying el api 6.0.0-RC1 jar). Other new tests passed.

Errors:
[ERROR] ELClientIT.optionalELResolverObjectInvoke:227->testOptionalELResolverInvoke:270
[ERROR] Tests run: 361, Failures: 0, Errors: 1, Skipped: 0

java.lang.NullPointerException: Cannot read the array length because "<local7>" is null
        at com.sun.ts.tests.el.api.jakarta_el.optionalelresolver.ELClientIT.testOptionalELResolverInvoke(ELClientIT.java:270)
        at com.sun.ts.tests.el.api.jakarta_el.optionalelresolver.ELClientIT.optionalELResolverObjectInvoke(ELClientIT.java:227)

[markt-asf]

A fix will be required in Glassfish (ExpressLy?) for that. The expected behaviour has changed.

[alwin-joseph]

A fix will be required in Glassfish (ExpressLy?) for that. The expected behaviour has changed.

Okey. I will merge this and update the bundle at https://www.eclipse.org/downloads/download.php?file=/ee4j/jakartaee-tck/jakartaee11/staged/eftl/jakarta-expression-language-tck-6.0.0.zip. Will retest when expressly has the required changes integrated.

[arjantijms]

A fix will be required in Glassfish (ExpressLy?) for that. The expected behaviour has changed.

I'll take a look at that soon (should be in M3 then). Thanks!

[arjantijms]

A fix will be required in Glassfish (ExpressLy?) for that. The expected behaviour has changed.

As it appeared this issue was entirely in the API jar. I do wonder how Tomcat was able to pass this, unless a different API jar was used.

[markt-asf]

Tomcat has always provided its own API JARs for all the specs it implements.

[arjantijms]

Tomcat has always provided its own API JARs for all the specs it implements.

Thanks, that clears it up.