> Yes I'd drop the whole package. Unless we want to introduce/perpetuate > yet another service framework. > I prefer we do not do that at this time, and continue with the Turbine > service framework for now
I agree, it will then be more Turbine-esque that way. Plus it will be easier to manage/understand. I currently have some personal issues with fulcrum, so I want to start this out as a coupled service. The change shouldn't be to invasive as it appears the only classes using the ServiceFactory are AbstractPortlet and CustomizeAction. Does this sound right, or should I look deeper? New package name suggestion: org.apache.jetspeed.services.persistence Is this kewl? I'm going to start out simple at first, probably just passing in a portlet, attribute name, and RunData to getter and setter methods and returning the attribute in question. Scott > -----Original Message----- > From: David Sean Taylor [mailto:[EMAIL PROTECTED]] > Sent: Thursday, June 27, 2002 1:21 PM > To: 'Jetspeed Developers List' > Subject: RE: Question on Muliple portlet instance behavior > > Yes I'd drop the whole package. Unless we want to introduce/perpetuate > yet another service framework. > I prefer we do not do that at this time, and continue with the Turbine > service framework for now > > > -----Original Message----- > > From: Weaver, Scott [mailto:[EMAIL PROTECTED]] > > Sent: Thursday, June 27, 2002 11:18 AM > > To: 'Jetspeed Developers List' > > Subject: RE: Question on Muliple portlet instance behavior > > > > > > David, > > > > If we are implementing it as a TurbineService, stop me if I'm > > wrong here, there is no need for the ServiceFactory, correct? > > > > Scott > > > > > -----Original Message----- > > > From: David Sean Taylor [mailto:[EMAIL PROTECTED]] > > > Sent: Thursday, June 27, 2002 1:06 PM > > > To: 'Jetspeed Developers List' > > > Subject: RE: Question on Muliple portlet instance behavior > > > > > > > You still want to keep the PersistenceService interface as is > > > > correct? > > > > > > Im not tied to it at all. The interface looks overly > > complicated with > > > the inner Page interface Please write a new interface if > > you think its > > > necessary. > > > > > > I would rewrite it as a Turbine service, like every other > > service in > > > Jetspeed (except this one) > > > > > > > I will start looking at gutting/rewriting the service. > > > > > > Great! > > > > > > > -----Original Message----- > > > > From: Weaver, Scott [mailto:[EMAIL PROTECTED]] > > > > Sent: Thursday, June 27, 2002 10:49 AM > > > > To: 'Jetspeed Developers List' > > > > Subject: RE: Question on Muliple portlet instance behavior > > > > > > > > > > > > You still want to keep the PersistenceService interface as is > > > > correct? > > > > > > > > One the first suggestion I have would be to replace all > > the instance > > > > of Vector with ArrayList, to improve performance. They > > did this with > > > > Torque and supposedly it garnered them a nice increase in > > > > performance. The only issue I see with this would be if you are > > > > betting on those Lists to be threadsafe. > > > > > > > > I will start looking at gutting/rewriting the service. > > > > > > > > Scott > > > > > > > > > > > > > -----Original Message----- > > > > > From: David Sean Taylor [mailto:[EMAIL PROTECTED]] > > > > > Sent: Thursday, June 27, 2002 12:29 PM > > > > > To: 'Jetspeed Developers List' > > > > > Subject: RE: Question on Muliple portlet instance behavior > > > > > > > > > > I wouldn't mind if you rewrote the whole class and > > threw away the > > > > > existing. That is what I was planning on doing. I think > > > > > JetspeedPersistenceService is pretty bad > > > > > > > > > > > -----Original Message----- > > > > > > From: Weaver, Scott [mailto:[EMAIL PROTECTED]] > > > > > > Sent: Thursday, June 27, 2002 10:29 AM > > > > > > To: 'Jetspeed Developers List' > > > > > > Subject: RE: Question on Muliple portlet instance behavior > > > > > > > > > > > > > > > > > > Thanks for the response, David. I am VERY interested in > > > > getting it > > > > > > work. I'm currently looking at the PersistenceService and > > > > its inner > > > > > > class, PersistenceService.Page, right now. Let me now if > > > > I'm on the > > > > > > right track and I will continue to work on it. Hopefully, I > > > > > > will be able to come up with a suitable patch. > > > > > > > > > > > > I welcome any pointers you may have. > > > > > > > > > > > > Thanks, > > > > > > Scott > > > > > > > > > > > > > -----Original Message----- > > > > > > > From: David Sean Taylor [mailto:[EMAIL PROTECTED]] > > > > > > > Sent: Thursday, June 27, 2002 12:11 PM > > > > > > > To: 'Jetspeed Developers List' > > > > > > > Subject: RE: Question on Muliple portlet instance behavior > > > > > > > > > > > > > > The JetspeedPersistenceService still needs to be updated to > > > > > > > support portlet instances. It doesn't consider portlet ids > > > > > > > (instances) or non-default pages. Its high on my todo > > > > list, unless > > > > > > > someone > > > > > > else fixes > > > > > > > it first... > > > > > > > > > > > > > > > -----Original Message----- > > > > > > > > From: Weaver, Scott [mailto:[EMAIL PROTECTED]] > > > > > > > > Sent: Thursday, June 27, 2002 7:57 AM > > > > > > > > To: 'Jetspeed Developers List' > > > > > > > > Subject: Question on Muliple portlet instance behavior > > > > > > > > > > > > > > > > > > > > > > > > I have noticed some "odd" behavior when it comes > > to multiple > > > > > > > > instance of the same portlet. > > > > > > > > > > > > > > > > Example: > > > > > > > > > > > > > > > > Let's say I have a portlet, "MyPortlet", which extends > > > > > > > > VelocityPortlet. It has a single attribute called > > > > "test." Now, > > > > > > > > I add this portlet to 2 separate panes, pane_1 and > > > > pane_2. My > > > > > > > > PSML file reflects that MyPortlet is in fact in > > both pane_1 > > > > > > and pane_2, > > > > > > > > each instance of MyPortlet has a unique ID value. > > So far so > > > > > > > > good. > > > > > > > > > > > > > > > > Now, this is the part that does not seem right, to me at > > > > > > least. If > > > > > > > > I have an action that sets the value of the attribute > > > > > > > > "test", regardless of which instance of MyPortlet > > calls the > > > > > > action the value > > > > > > > > is both stored and retrieved from the first (as > > it appears > > > > > > > > in the PSML file) instance of MyPortlet in pane_1. The > > > > > > > > second > > > > > > instance of > > > > > > > > MyPortlet in pane_2 never gets or sets its own attribute > > > > > > "test", and > > > > > > > > it always refers to the attribute "test" in the first > > > > instance > > > > > > > > of MyPortlet. > > > > > > > > > > > > > > > > Before I dig any deeper I need to ask, "Is this correct > > > > > > behavior for > > > > > > > > multiple instances of the same portlet?" > > > > > > > > > > > > > > > > I'm using the most current CVS HEAD for 1.3a3. > > > > > > > > > > > > > > > > Thanks, > > > > > > > > Scott > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > To unsubscribe, e-mail: <mailto:jetspeed-dev- > > > > > > > [EMAIL PROTECTED]> > > > > > > > For additional commands, e-mail: <mailto:jetspeed-dev- > > > > > > > [EMAIL PROTECTED]> > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > To unsubscribe, e-mail: <mailto:jetspeed-dev- > > > > > [EMAIL PROTECTED]> > > > > > For additional commands, e-mail: <mailto:jetspeed-dev- > > > > > [EMAIL PROTECTED]> > > > > > > > > > > > > > > > > -- > > > To unsubscribe, e-mail: <mailto:jetspeed-dev- > > > [EMAIL PROTECTED]> > > > For additional commands, e-mail: <mailto:jetspeed-dev- > > > [EMAIL PROTECTED]> > > > > > > -- > To unsubscribe, e-mail: <mailto:jetspeed-dev- > [EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:jetspeed-dev- > [EMAIL PROTECTED]>
