Hi I think the behavior presented is because ExceptionHandler api is not complete (some exceptions are published but since there is no code that consume and show this exceptions simply get lost or empty page is shown). Maybe it could be better to check this fact and complete this point first before apply other changes on current code. I remember that a temporal change in some catch of LifecycleImpl to get back the old behavior (throw instead publish) make the error appear.
regards Leonardo Uribe 2009/8/10 Michael Concini <[email protected]> > Jan-Kees van Andel wrote: > >> Sounds like a good idea, if it is not used anymore. >> >> The only reference I see, is FaceletViewHandler.NullWriter. >> JspViewHandlerImpl is only used in tests. (at least, IDEA tells me so) >> >> I've commented both classes and the test class, since there can still >> be some valuable source code in there. Once properly tested, we can >> easily remove the commented code. I've moved NullWriter to >> FaceletViewDeclarationLanguage. >> >> The only issue I see are people who explicitly specify >> JSPViewHandlerImpl in their faces-config.xml files, but I don't think >> many people do that. >> >> /JK >> >> >> > Would it be worthwhile to output a warning for any users who are explicitly > specifying either of the old viewhander classes and deprecate them instead > of deleting them outright? I know they're internal API but they're internal > API that someone might have been explicitly specifying, even if that group > of people is small. Could be a good thing to do, and fairly painless for us > to ipmplement, to help customers migrating from the 1.2 to 2.0 runtime in > their apps. > >> 2009/8/10 Simon Lessard <[email protected]>: >> >> >>> Hi, >>> >>> Since it's an internal API I think we should just flush the class down >>> the >>> drain. >>> >>> >>> ~ Simon >>> >>> On Mon, Aug 10, 2009 at 12:47 PM, Jan-Kees van Andel >>> <[email protected]> wrote: >>> >>> >>>> Removing the <view-handler /> element from my faces-config.xml did the >>>> trick. >>>> >>>> Funny that everything worked properly, except the rerendering. Should >>>> we add some code to prevent people from making the same mistake? >>>> >>>> Thanks Mike, >>>> JK >>>> >>>> >>>> 2009/8/10 Michael Concini <[email protected]>: >>>> >>>> >>>>> Jan-Kees van Andel wrote: >>>>> >>>>> >>>>>> The default in the current 2.0 runtime should be using >>>>>>> org.apache.mfyaces.application.ViewHandlerImpl, which is a single >>>>>>> class >>>>>>> implementation for both JSP and facelets. It uses the FactoryFinder >>>>>>> to >>>>>>> get >>>>>>> the right ViewDeclarationLanguage implementation depending on whether >>>>>>> you're >>>>>>> using JSP or facelets since nearly all of the actual work is done in >>>>>>> the >>>>>>> VDL >>>>>>> classes now. The old JSP and facelets ViewHandler impls shouldn't >>>>>>> really >>>>>>> be >>>>>>> used anymore and as far as I know haven't been updated to the 2.0 >>>>>>> spec. >>>>>>> >>>>>>> >>>>>>> >>>>>> I'm gonna check it tonight (@work now), but the last time I checked, I >>>>>> saw the ViewHandlerImpl being returned, but I have to check to be >>>>>> sure. >>>>>> >>>>>> But you're saying that specifying the <view-handler /> element in >>>>>> faces-config.xml is redundant and probably incorect? In that case, I'm >>>>>> gonna remove it and check if it works. I've added it a while ago >>>>>> because I got the JSP ViewHandler back then. Maybe things have changed >>>>>> in the meantime. >>>>>> >>>>>> >>>>>> >>>>>> >>>>> That makes sense if you had originally set things up more than a few >>>>> weeks >>>>> ago. The new ViewHandlerImpl class was in place but we hadn't changed >>>>> the >>>>> default config to use it until sometime in July once most of the VDL >>>>> impls >>>>> were ready and we were trying to get some apps working. >>>>> >>>>> >>>>>> If you're getting the JSP VDL back from ViewHandlerImpl, then there is >>>>>>> something not configured right either. Maybe in our default config >>>>>>> files >>>>>>> or >>>>>>> the factory definition? >>>>>>> Let me know if you need me to check into it and I should be able to >>>>>>> later >>>>>>> this week. I'm on vacation visiting family right now though so it >>>>>>> might >>>>>>> be >>>>>>> a few days. >>>>>>> >>>>>>> >>>>>>> >>>>>> Thanks for the offer. I'm gonna debug a bit first, and then I'll come >>>>>> back to you. >>>>>> >>>>>> Regards, >>>>>> Jan-Kees >>>>>> >>>>>> >>>>>> >>>>>> >>>>> Sounds good...shoot an email to let me know if you need me to check on >>>>> it. >>>>> Thanks, >>>>> Mike >>>>> >>>>> >>>>> >>>> >>> >> >> >> > >
