jsr372-experts@javaserverfaces-spec-public.java.net

[jsr372-experts] Re: [SPEC-1238] request for review

From: Leonardo Uribe <leonardo.uribe_at_irian.at>
Date: Tue, 17 Jan 2017 18:23:50 -0500

Hi

I just wanted to comment that the idea is allow use keywords in "targets"
attribute. In MyFaces the change worked well.

But this is a minor feature of the search expression API, so in the worst
case, the changes related to AttachedObjectTargetImpl can be reverted
without significant impact. The two main use cases where this API is used
are f:ajax and h:outputLabel.

regards,

Leonardo Uribe


2017-01-17 16:57 GMT-05:00 Edward Burns <edward.burns_at_oracle.com>:

> >>>>> On Tue, 17 Jan 2017 21:25:27 +0100, Thomas Andraschko <
> andraschko.thomas_at_gmail.com> said:
>
> TA> All right, so lets wait!
> TA> If i don't sleep, i will try to fix every unit tests, which is related
> to
> TA> my change in few hours.
>
> Ok, I have a genuine failure here. I've replicated it on Solaris and
> Mac OS X in the build and on Hudson. Here are the details.
>
> The test app is test/servlet30/systest.
>
> The test app must be build with the following properties
>
> webapp.projectStage=Development
> webapp.stateSavingMethod=server
> webapp.partialStateSaving=false
> webapp.serializeServerState=false
>
> I have attached a pre-built WAR to the JIRA for your convenience. I
> modified the WAR slightly from HEAD to comment out all but the
> offending component from the XHTML.
>
> Once you deploy that app, visit
>
> http://localhost:8080/jsf-systest/faces/composite/attachedvalidator.xhtml
>
> With the change in place you will see a StackOverflow.
>
> I'm pretty sure the failure is related to your mods in
> AttachedObjectTargetImpl because if I revert just your mods to that file
> and visit the same page, I do not see the StackOverflow. Indeed, there
> are some comments in that file that prophecy that messing with this code
> will likely produce StackOverflow.
>
> Obviously, this change needs to run without problem.
>
> Please give this urgent attention.
>
> It is the same set of nine tests that fail whenever they are run.
> Coincidentally, they are only run when
>
> webapp.stateSavingMethod=server
> serializeServerState=false
>
> are set.
>
> The set of failed tests is
>
> >>> com.sun.faces.test.servlet30.systest.CompositeComponentsIT.
> testValidator1
> >>> com.sun.faces.test.servlet30.systest.CompositeComponentsIT.
> testValidator2
> >>> com.sun.faces.test.servlet30.systest.CompositeComponentsIT.
> testValidator3
> >>> com.sun.faces.test.servlet30.systest.CompositeComponentsIT.
> testValidator4
> >>> com.sun.faces.test.servlet30.systest.CompositeComponentsIT.
> testActionSource1
> >>> com.sun.faces.test.servlet30.systest.CompositeComponentsIT.
> testActionSource2
> >>> com.sun.faces.test.servlet30.systest.CompositeComponentsIT.
> testActionSource3
> >>> com.sun.faces.test.servlet30.systest.CompositeComponentsIT.
> testActionSource4
> >>> com.sun.faces.test.servlet30.systest.CompositeComponentsIT.
> testConverters
>
> I'm going to mark these tests as @Ignore for now, but we have to make
> them pass.
>
> Thanks,
>
> Ed
>
> --
> | edward.burns_at_oracle.com | office: +1 407 458 0017
> | 5 business days until planned start of JSF 2.3 Public Review
> | 25 business days until DevNexus 2017
> | 50 business days until JavaLand 2017
>