I have a Resource Handler that needs to access lower level beans that work in conversation scope. Our application allows multiple login sessions on different browser tabs working within the same browser session by keeping the login details at this scope - so a database access (my Resource Handler) needs the login for the Conversation that referenced it.
I've tried just expecting it to work, even using ExternalContext.encodeRedirectUrl() to try to make Faces include whatever it needs.
I've tried explicitly putting the ?cid parameter onto my request path so I am accessing
/myapp/javax.faces.resource/thumbnail_3963075518712822225212162218.png.xhtml?ln=com.myapp.attachment&cid=1
No Conversation appears in the Resource Handler.
Further if I try to test for this using Conversation.isTransient() I get a
javax.enterprise.context.ContextNotActiveException: Conversation Context not active when method called on conversation Transient conversation
(Yet I thought Conversation is always meant to be active, and the stack trace for the above does include the org.jboss.weld.servlet.ConversationPropagationFilter!
I'm hoping this is not to do with the bug https://issues.jboss.org/browse/JBSEAM-3689 - for which the workaround is "Do not use conversations!"
Is there better?
Thanks
- Richard
Take a look at Seam Faces and see if you can get things to work the way you want. JSF has next to no compatibility with CDI in the JSF 2.0 and 2.1 specs. We're working hard with the lead for JSF to make it better for JSF 2.2, sorry for any inconveniences.
Mad idea - write an XHTML page with a Restore View event handler that manually renders the resource and calls Complete() on the FacesContext. Ugh! But XHTML pages do get conversations.
I have done what is perhaps suggested in the bug report. I've found a way of solving the problem not using Conversation Scope. In my case I have a ThreadLocal that, if present, provides alternate information to the component that access Conversation Scoped data.
Take a look at Apache Myfaces CODI,
Conversation management is superb with their extensions
Related
Here is my scenario. I have a component which searches some records and while searching, those records are marked locked. I am using JSF-2 and primefaces. My MBean is ViewScoped. Now i have a requirement, while closing the browser, the lock on those records need to be released. I was searching and my best bet appears to be #PostDestroy. So can someone help, whether using PostDestroy is correct approach. Some of the threads on StackOverFlow suggest, PostDestroy doesn't get called on browser close events. I I am not able to find much on PostDestroy. Can someone provide pointers on this approach?
Thanks,
Ben
Your #ViewScoped Bean does not know if the user is closing the browser. For the Bean it does not even matter if the client is a browser or some other HTTP-client.
The lifetime of a #ViewScoped Bean ends normally if your application sends a postback to the next view (see How and when is a view scope bean destroyed in JSF for an exact answer).
You can try to detect a closing browser with the JavaScript window.onunload event but I would not recommend that. Some browsers fire this event on browser close, others not, some fire it on reload etc. (see DOM Window unload event, is it reliable? for more details).
The #PreDestroy annotated Method (not #PostDestroy btw.) may also not be called immediately in some cases. You can only rely that it will be called on a session timeout. Which may be a long time for locking records...
You could release the locked records after a shorter period on your own using some kind of scheduler but I'd recommend to switch to optimistic locking (Don't lock and check if data has changed before you write).
When I open two browsers and enter details parallely, the values are getting mixed. The values in one browser are populated in the other browser... We are using JSF framework and the managed beans are in session scope (having them in request scope is not possible in our case)... In a nutshell, the values are shared across the browsers... How to avoid this? Any clue?
If you "open two browsers" means you open two windows of same browser then everything is correct - you can get just one session per browser (for firefox there is plug-in which allow to avoid this limitation).
But if you use really two different browsers than in this case seams you save all data in application scope or you have a problem with you container.
You need to put the bean in request or view scope instead. Storing request/view based data in a session scoped bean is a bad idea, as you've encountered yourself.
I think the view scope would help a lot, given the fact that you mentioned that the request scope is "not possible" (it actually is, it only requires preloading the right data so that JSF can take the right actions accordingly).
Hello stackoverflow world.
(This is the first time I actually post a question here. Exciting)
A while ago I inherited a 2 year old MVC website from one of the teams within my corporation. I know most the ins and outs of this solution now but there is something strange cropping up in my error logs which I do not understand.
Every now and then I will get an error messages like this one:
A public action method 'xyz.gif' could not be found on controller MyNamespace.MyController
What I don't understand is WHY is this action (a gif image) being called in the first place?
I've seen 2 different gifs in the error logs pageerror.gif and refresh.gif
As this is an inherited solution I double checked everything and made sure that there are in fact no images like that in the project and no reference anywhere even to those words in the controllers, views, style-sheets or even in the source of pages within the same controller.
I seriously doubt that the users are playing around with the URLs and adding random gif names to them to see what happens.
I'm all out of ideas. Anyone out there who can suggest more places to look for the culprit?
Ta!
As Tchami pointed out in a comment on the original question, this is related to Internet Explorer's default error pages.
As I have set up custom error pages I believe this is either due to an internal server error or possibly somehow an action cancel error from the client side, i.e. client side error. I can't be 100% at this point.
The question is not fully answered but I mostly know what the cause is now.
From my point of view I've identified that I need to improve this ASP.NET MVC application so that
1) it doesn't report/log errors when someone tries to navigate to a non-existing controller action (e.g. these refresh.gif actions or any other)
and 2) handle it better for the client so that they don't end up clicking from one error page (default IE error page) to another (my custom error page when clicking the refresh icon on the IE page)
Another stackoverflow thread on a related topic:
Significance of 'pagerror.gif'?
(i can't post more links as I'm a new user)
CHEERS!
Solveig
Can you get the error to show up in the logs when you use the site yourself? If so, an add-in such as HttpWatch might help you see those .gif requests. If you can understand more about when they happen you might be able to figure out what's going on.
Pagerror.gif and refresh.gif are the default images from the IE browser/IIS server. Normally these images were shown when the browser is not able to retrieve the content.
If you see these errors in logs, then check the iis log to get more information.
For example,
IIS Log : look for this feild ,
cs(User-Agent)
compatible;+MSIE+6.0;+Windows
Here few things to be noted,
1)IE 6.0 apparently makes these requests on its own. I am not sure if any other IE 6+ Browser would show similar behavior.
2)All this will do is generates a 'bogus" event log entry because a null reference exception could happens when you request a non-existing GIF and that request goes through MVC pipeline.
3)Technically this simply can be ignored.
4) Optionally we could check if through routing we could stop “.gif” files from being processed by MVC Pipeline
All i would suggest is to handle it gracefully.
It may be difficult to pinpoint this, but my thought would be: check the javascript. If the image name is being dynamically generated somewhere, and then requested, a simple "find and replace" may miss the reference.
I have this idea for a project. Associated with any web page, i want to create notes that will be saved locally in a database, the notes will be reloaded automatically from that database the next time i visit the same page.
Creating the note is easy, but i'm looking for how to link the notes to the web page url and how to keep aware of the active web page. Any idea?
(Note: i have come to this searching on the internet: http://webkit.org/demos/sticky-notes/ - this is part of WebKit Open source projects) - this is about what i'm looking for.
Thank.
Browserdependent probably. You'll have to have a plugin for every browser type.
IE might be doable via the COM interface, but that probably would require starting IE via a way you control. So that probably will have to be a plugin too.
For browser independence, there are quite a few challenges in this one. One way would be to implement a proxy server and watch for text/html content....this will work for most of the general cases, but not every case. Handling frames for instance... which resource is the "parent" and which is the "child"? Which one contains the sticky note? I think you would have to inject some client side javascript to keep track of things, and that might break some websites.
protonotes.com is a web service version of this. Not sure how they do it though.
Actually, Daniel H hit the nail on the head mate: http://www.protonotes.com
It does exactly as you want, in fact it gives you two options to store your data, the first is hosted, the second is your own mySQL db - protonotes pipes the data from the tack-on style notes to your own db, if you prefer. This means that you're not the only person who can see the notes - access is granted by a unique 'group' key.
I've just deployed protonotes as our main online review tool for two reasons, we can save our own data, and it lacks some features which I generally label "dubious" anyway.
It's simplicity is great, the only thing I'm aware of that could cause a prob is that it dumps a bunch of stuff in the global namespace - if that's a potential problem for you.
d
I'm trying to develop an application which modifies a couple of tasks of the famous Online-TODO List RememberTheMilk (rememberthemilk.com) using the REST API.
Unfortunately the modifying takes a lot of time, so I want to give a feedback to the users.
My idea was just to display a couple of text lines (e.g. modifying task 1 of n...).
Therefore I used the periodically_call_remote on my page and called a which reads a Singleton.
In the request I store the text that should be displayed in the same singleton. But I found out, that once I set up a request, the periodically_call_remote does not update the specified div.
My question to this:
1. is this a good way to implement this behaviour?
2. if it is, how do get the periodically_call_remote to work during a submit?
Using a Singleton is most definitely a bad idea. In an advanced production setup it isn't guaranteed that subsequent requests will go to the same process or to the same machine (and subsequently will have a different Singleton). Plus, if you have many users, I don't even want to think about what'll happen to those poor Singletons.
Does any of this stuff actually need to go through your Rails app? It seems like you can call the RTM API via Javascript from the page the user is on and then update the page when the XHR request is complete.