IMO the goal should be to NOT revert this changes

@Werner Could you fix/refactor the tests maybe? I really dont like to
revert because TCK uses a technology which doesnt support long time
existing JS/HTML features....

Am Mo., 20. Feb. 2023 um 17:16 Uhr schrieb Volodymyr Siedlecki <
[email protected]>:

> Hello,
>
> I regret bringing this topic up, but the new JavaScript code in RC5 is 
> encountering more HTMLUnit trouble.
>
> I've pulled the RC5 jars into our application server and run our test suite 
> against it, and we saw about 5-10% of tests fail.
>
> We have yet to run against the full TCK, but initial runs also show failures 
> and errors (see list 1). The majority of the failures/errors are caused by 
> either "missing formal parameter" or "syntax error".
>
> This issue can be traced back to Rhino not supporting "rest parameters" 
> (Found https://github.com/mozilla/rhino/issues/652 via  
> https://github.com/HtmlUnit/htmlunit/issues/232 ).
>
>
> [ERROR]
> ee.jakarta.tck.faces.test.javaee6.viewParamNullValueAjax.Issue4550IT.testViewParamNullValueAjax
> Time elapsed: 2.33 s <<< ERROR! ======= EXCEPTION START ======== Exception
> class=[net.sourceforge.htmlunit.corejs.javascript.EvaluatorException]
> com.gargoylesoftware.htmlunit.ScriptException: missing formal parameter (
> http://localhost:9080/viewParamNullValueAjax/faces/jakarta.faces.resource/faces.js?ln=jakarta.faces#2)
>
>
> Back to the topic of the TCK, the ajax tests (faces22/ajax, faces23/ajax, 
> etc) were ported over to use the selenium driver, but other tests, which also 
> use ajax, were not ported (faces23/exactmapping, faces40/inputfile, etc).
> Another portion of the TCK that has yet to be tested is the old-tck (which 
> has thousands of tests, some of which will be hit by these script 
> exceptions). *We'd be losing testing coverage with the new scripts.*
>
> If we remove the rest parameters, then our problems might be resolved. It's 
> not a guarantee and I'm not sure how feasible it would be.
>
> Unfortunately, though, a revert is the best course of action as I don't think 
> challenging all of these previously working tests is a sound approach.
>
> My hope is to have a 4.0.0 release with the previous scripts. This would then 
> give us time to automate MyFaces to run against the TCK and also move away 
> from HTMLUnit.
>
> The new scripts could then be merged at a later time (with more confidence).
>
> 1) List of New TCK Failures:
> ee.jakarta.tck.faces.test.javaee6.viewParamNullValueAjax.Issue4550IT.testViewParamNullValueAjax
> ee.jakarta.tck.faces.test.javaee8.cdi.Spec1351IT.testInjectValidator
> ee.jakarta.tck.faces.test.javaee8.cdi.Spec1386IT.testInjectFlowMap
> ee.jakarta.tck.faces.test.javaee8.commandScript.Spec613IT.test
> ee.jakarta.tck.faces.test.servlet40.exactmapping.Spec1260IT.testLinkToNonExactMappedView
> ee.jakarta.tck.faces.test.servlet40.exactmapping.Spec1260IT.testAjaxFromExactMappedView
> ee.jakarta.tck.faces.test.servlet40.exactmapping.Spec1260IT.testExactMappedViewLoads
> ee.jakarta.tck.faces.test.servlet40.exactmapping.Spec1260IT.testPostBackOnLinkedNonExactMappedView
> ee.jakarta.tck.faces.test.servlet40.exactmapping.Spec1260IT.testPostBackToExactMappedView
> ee.jakarta.tck.faces.test.servlet40.exactmapping.Spec1260IT.testResourceReferenceFromExactMappedView
> ee.jakarta.tck.faces.test.servlet40.facelets.Issue4830IT.testUIRepeatResetValues
> ee.jakarta.tck.faces.test.servlet40.facelets.Issue5078IT.testUIRepeatVisitTreeDuringInvokeApplication
> ee.jakarta.tck.faces.test.javaee8.searchExpression.Issue4331IT.test
> ee.jakarta.tck.faces.test.javaee8.searchExpression.Spec1238IT.test
> ee.jakarta.tck.faces.test.javaee8.uiinput.Issue5081IT.testIssue4734
> ee.jakarta.tck.faces.test.servlet50.inputfile.Spec1555IT.testMultipleSelectionNonAjax
> ee.jakarta.tck.faces.test.servlet50.inputfile.Spec1555IT.testSingleSelectionNonAjax
> ee.jakarta.tck.faces.test.servlet50.inputfile.Spec1555IT.testMultipleSelectionAjax
> ee.jakarta.tck.faces.test.servlet50.inputfile.Spec1555IT.testSingleSelectionAjax
>
> Thanks,
>
> Volodymyr
>
>

Reply via email to