As I mentioned, I didn't want to use Shale as it doesn't not have the built-in ability to redirect from the methods (prerender for example), at least not when I last looked. I was also afraid of incompatibilities between Shale and JBoss-Seam.
-Andrew On 5/1/06, Craig McClanahan <[EMAIL PROTECTED]> wrote:
On 5/1/06, Andrew Robinson <[EMAIL PROTECTED] > wrote: > I agree, I have not given up in trying to find a "built in" way of > doing it in JSF, but still have not found a way. The problem is that > JSF is all about actions when execution Java code. The problem is that > actions are never run on the first time a page is hit, only on POST > back. I guess it is simply a major gap when decoupling the business > code from the UI. This kind of issue was a key motivation for the ViewController functionality in Shale[1]. In particular, the prerender() method is a good place to deal with this kind of issue, because it: * Is called on both GET and POST requests. * Is called *before* rendering takes place, so you can effect the response (include error messages, etc.) * Allows you to set up resources that are required for rendering (allocate a JDBC connection, open a Hibernate session, whatever) that is required for rendering, then clean up after rendering in the destroy() method. You can do this kind of thing yourself with a phase listener, but you're going to end up basically replicating what Shale already provides. Craig [1] http://struts.apache.org/struts-shale/ > I use the on-load component mostly for bookmarks. Say, somone > bookmarks "edit.jsf", I use the on-load component to redirect them > back to "index.jsf" if there is no property to edit (the backing bean > is not in an editing state). > > The other thing I use it for is to load data that may fail. Say > hitting a web service for example. In a getter method, the response is > probably already committed and therefore too late to redirect the user > or forward to a new page. Also, it is hard to add FacesMessages in a > getter as the message(s) component(s) may have already been rendered, > and once again it would be too late. The on-load allows me to > gracefully handle errors in backing-bean methods during the loading of > a new page. > > If there is a better way that you find or a way I can improve it, let > me know. I would have much preferred to see this in the specification > than have to make a one-off open source component. > > -Andrew > > On 4/30/06, benjamin van der veen <[EMAIL PROTECTED]> wrote: > > Andrew, > > > > Very informative message. Thanks a lot. I'm still a bit troubled by > > the fact that JSF didn't have this built in. It seems like such an > > integral piece of a web framework that is used to accomplish a wide > > range of tasks, so much so that it seems to me that the spec writers > > couldn't have simply forgotten it, and it makes me wonder if the tasks > > I would accomplish with it were accounted for somewhere else in the > > specification. Could you give me an example of what you use the onload > > method for? I know what I would do with it, but I'd just like to see > > what you have to say without trying to take into account what I write. > > > > Thanks, > > benjamin van der veen > > > > On 4/30/06, Andrew Robinson <[EMAIL PROTECTED]> wrote: > > > If you find any issues with the onload with jsf-comp please let me > > > know (I am the author). I am using it at home and at work without > > > issues except for JBoss-Seam EJB interceptors. Seam depends on the > > > action method unfortunately. > > > > > > I looked fairly closely at the spec before writing the onload and made > > > sure that the code I wrote conforms to what you are allowed to do. It > > > behaves similarly to an action that decides to forward/include a > > > non-JSF page. It uses the correct redirection and ensures the faces > > > response is marked complete. So in these respects, it should not > > > interfere with the specification at all. The only issue (like Seam) is > > > when code may not expect action methods to be executed outside of the > > > execute action JSF phase. > > > > > > I couldn't believe it either, before I wrote it, that JSF was missing > > > such a huge piece of functionality. All other languages (PHP, Perl, > > > ASP, ASP.NET, JSP) allow you to perform code when loading a page so I > > > don't know why it wouldn't be in JSF. > > > > > > As a side note, Shale has something similar with the page controller. > > > What I didn't like about it was that it had "void" as the return type > > > and therefore you could not as easily interact with the navigation > > > handling. The other problem I had with the JSF spec in these regards > > > was the lack of externalized functionality in the NavigationHandler > > > interface. I would love to get the view for a result without actually > > > navigating there, or be able to extend the navigation handler and be > > > able to modify the view after the method is fired but before the new > > > view is created. > > > > > > -Andrew > > > > > > On 4/30/06, benjamin van der veen < [EMAIL PROTECTED]> wrote: > > > > Xie Min, > > > > > > > > This is a great find, very close to what I'm talking about. However, > > > > in the following line: > > > > > > > > <to-view-id>/view.jsp?objectId=#{view.objectId}</to-view-id> > > > > > > > > what object does 'view' in the EL expression reference? Is this an > > > > implicit object, or one defined in a configuration somewhere? If it is > > > > implicit, how do I set it's properties from a backing bean? > > > > > > > > Thanks a lot, > > > > benjamin > > > > > > > > > > > > On 4/29/06, xie min < [EMAIL PROTECTED]> wrote: > > > > > > > > > > > > > > > in myfaces > > > > > wiki( http://wiki.apache.org/myfaces/Custom_Navigation_Handler), > > > > > threr are some suggestion on url parameters passing: > > > > > > > > > > Dynamic Url parameters passing: This allows passing parameters between pages > > > > > even when using redirection method. For exmaple, in order to navigate from a > > > > > pages name: view.jsp to a page named: edit.jsp, with the parameter: > > > > > objectId, the following thing can be done in navigation rule code: > > > > > <navigation-rule> > > > > > <from-view-id>/view.jsp</from-view-id> > > > > > <navigation-case> > > > > > <from-outcome>edit</from-outcome> > > > > > <to-view-id>/view.jsp?objectId=#{view.objectId}</to-view-id> > > > > > <redirect /> > > > > > </navigation-case> > > > > > </navigation-rule> > > > > > > > > > > > > > > > > > > > > > > > > > 2006/4/30, benjamin van der veen < [EMAIL PROTECTED]>: > > > > > > I think that jsf-comp onload thing might be what I'm looking for. I > > > > > > really surprised that something like this isn't built in to JSF. Is > > > > > > using this extension violating some design principle fundamental to > > > > > > JSF? > > > > > > > > > > > > benjamin > > > > > > > > > > > > On 4/28/06, Andrew Robinson < [EMAIL PROTECTED]> wrote: > > > > > > > Have a look at: > > > > > > > > > > > > > > > > > > > http://wiki.apache.org/myfaces/InvokingJsfPagesWithStandardUrls > > > > > > > > > > > > http://jsf-comp.sourceforge.net/components/onload/index.html > > > > > > > > > > > > > > -Andrew > > > > > > > > > > > > > > On 4/28/06, benjamin van der veen < [EMAIL PROTECTED]> wrote: > > > > > > > > Hello, > > > > > > > > > > > > > > > > I sent the message quoted below to this list a few days ago, but > > > > > > > > received no replies. I'm just wondering if perhaps it just got > > > > > > > > overlooked; I would appreciate any insight on this issue. > > > > > > > > > > > > > > > > Thanks a lot, > > > > > > > > benjamin > > > > > > > > > > > > > > > > On 4/26/06, benjamin van der veen < [EMAIL PROTECTED]> wrote: > > > > > > > > > Hello, > > > > > > > > > > > > > > > > > > I'm coming from a PHP background, and I've run into a wall with JSF. > > > > > I > > > > > > > > > know how I'd get around it with PHP or ASP.NET, but for JSF I'm > > > > > > > > > stumped. > > > > > > > > > > > > > > > > > > Here's the scenario: > > > > > > > > > > > > > > > > > > I'm making a message board. It has multiple forum each of which has > > > > > > > > > many topics, each of which has many posts. Very standard thing here. > > > > > I > > > > > > > > > am implementing the page which allows a user to create a new topic. > > > > > > > > > The URI is of the form "createTopic.jsf?forumId=<n>" where <n> is an > > > > > > > > > integer corresponding to the forum's primary key in the database. I > > > > > > > > > have a backing bean mapped in faces-config to "CreateTopicBean", and > > > > > > > > > it is request scoped. Now, when the user submits the form to create > > > > > a > > > > > > > > > topic, what is the best way to let the bean know which forum this > > > > > > > > > topic should be created under? I tried having a managed property > > > > > like > > > > > > > > > this: > > > > > > > > > > > > > > > > > > <managed-bean> > > > > > > > > > > > > > > <managed-bean-name>CreateTopicBean</managed-bean-name> > > > > > > > > > <managed-bean-class> > > > > > > > > > > > > > > com.wepwnsnails.faces.managedbean.CreateTopicBean > > > > > > > > > </managed-bean-class> > > > > > > > > > > > > > > <managed-bean-scope>request</managed-bean-scope> > > > > > > > > > <managed-property> > > > > > > > > > > > > > > <property-name>forumId</property-name> > > > > > > > > > <value>#{ param.forumId}</value> > > > > > > > > > </managed-property> > > > > > > > > > </managed-bean> > > > > > > > > > > > > > > > > > > but when the "action" attribute of the html <form> element rendered > > > > > by > > > > > > > > > JSF is simply "createTopic.jsf", so an EL exception is thrown when > > > > > JSF > > > > > > > > > fails to set the managed property. > > > > > > > > > > > > > > > > > > I've also tried adding a <h:inputHidden> tag set to the forumId on > > > > > the > > > > > > > > > backing bean in a variety of ways, but they leave much to be desired > > > > > > > > > as far as cleanliness. > > > > > > > > > > > > > > > > > > I'm sure that this is an issue every newbie JSF developer has > > > > > > > > > scratched his head over; so what's the best way to go about this? > > > > > I've > > > > > > > > > had a hard time finding anything in Google, mostly because I don't > > > > > > > > > know how to tersely phrase the problem for a seach query. > > > > > > > > > > > > > > > > > > Thanks a lot, > > > > > > > > > benjamin van der veen > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
