Aggregation: +7 (+8 including myself) for the idea with the ProjectStage. Because it really is not that much of work, I will include the cyclic reference detection in my patch for MYFACES-2375.
Thank you all for your contribution and a special thanks to Michael Concini for the idea with the project stage! Regards, Jakob 2009/10/13 Jan-Kees van Andel <jankeesvanan...@gmail.com> > +1, but it might be a lot of work. Maybe we should first get a > spec-compliant release out and plan this one for a second release. > > /JK > > 2009/10/13 Gerhard Petracek <gerhard.petra...@gmail.com>: > > +1 for the project stage idea > > > > regards, > > gerhard > > > > http://www.irian.at > > > > Your JSF powerhouse - > > JSF Consulting, Development and > > Courses in English and German > > > > Professional Support for Apache MyFaces > > > > > > 2009/10/13 Werner Punz <werner.p...@gmail.com> > >> > >> +1 as well for the project stage idea, the dev stage is definitely > >> the one which should track this but for production we need optimal > >> performance. > >> > >> Werner > >> > >> > >> Kito Mann schrieb: > >>> > >>> +1 > >>> > >>> On Tue, Oct 13, 2009 at 10:52 AM, Bruno Aranda <brunoara...@gmail.com > >>> <mailto:brunoara...@gmail.com>> wrote: > >>> > >>> +1 with the project stage sounds good to me! > >>> > >>> 2009/10/13 Simon Lessard <simon.lessar...@gmail.com > >>> <mailto:simon.lessar...@gmail.com>> > >>> > >>> +1 for the project stage idea. > >>> > >>> > >>> On Tue, Oct 13, 2009 at 10:28 AM, Michael Concini > >>> <mconc...@gmail.com <mailto:mconc...@gmail.com>> wrote: > >>> > >>> What about using project stage to determine which behavior > >>> to follow? If we're in production stage we don't check for > >>> best performance, but in development/test stages we perform > >>> the check. > >>> Alternatively, could we at least make it configurable > >>> through an org.apache.myfaces param in web.xml so apps that > >>> have been fully tested can disable the check? > >>> > >>> Thanks, > >>> Mike > >>> > >>> > >>> Jakob Korherr wrote: > >>> > >>> Hi everbody. > >>> > >>> While working on MYFACES-2375, I got stuck at the > >>> following scenario: > >>> > >>> Managed bean m1 has a custom scope #{m2.scope} and > >>> managed bean m2 has a custom scope #{m1.scope}. > >>> In this scenario you will get a StackOverflowException > >>> when trying to create one of the two managed beans. > >>> > >>> RI really ends in a StackOverflowException, should > >>> MyFaces end in such a Exception too or detect the cyclic > >>> reference and throw an ELException? > >>> > >>> Mike Kienenberger told me the following: "We have a > >>> precedent set on making MyFaces proactive on detecting > >>> error conditions in the configuration." > >>> > >>> The only problem is, that checking the cyclic references > >>> would not happen once at MyFaces startup, but every time > >>> a managed bean will be created, which means it slows > >>> down the application. > >>> > >>> What is your opinion on this question? > >>> Vote +1, if you think MyFaces should detect cyclic > >>> references in the managed bean scope. > >>> > >>> > >>> > >>> > >>> > >> > > > > >