I have a rule that really should not stop the saving process, like using the addbrokenrule call would do.  How would you suggest taking care of this type of issue?  How about a warnings collection and an addwarning call?

Just a thought.

Are you thinking of more functionality than just storing off the message? Like displaying a warning to the end user before they save?

I guess there are a couple of way it could be done.  I was thinking of something like the error messages with a "warning" or "information" icon that would provide the user with the message.

I would think that simply showing these warning icons after the save would require the least amout of changes to the framework as aposed to messaging the user before the save takes place.

Do you have a need for something like this in the applications that you are developing?

Scott

We don't currently have a need for this in our applications. We've had the functionality in the past, but we never used it, and anything we didn't use, we pulled, because it was unneeded overhead. If you can find a convincing reason to add it, we might put it in Currently, if you wanted to implement some type of warning, you could add a collection to the business object you want the warnings on. Then, in the BeforeSaveEvent, you could show a popup to the end-user asking them if they are sure they want to save the record. As for using the ErrorProvider with a different icon, you could add that in the AfterSaveEvent handler.



Let me know what you think...

Sounds good to me.  That should be a good learning experience for a .Net novice like myself.

I'd like to add my 2 cents to this topic. In the type of applications our organization (Social Services) develops, there is a definite need for warning messages that don't prevent an object from being saved. There are a number of data elements that we would like to capture for reporting purposes, but if the client refuses to provide that information, we are still required by law to provide the service. Using a warning message we can prompt the user to try to obtain the information, even though it's not 'required'. This has been an issue with our current framework, CSLA.

We are now planning on implementing a warning option through the business objects when checking the Broken Rules.  There will be an event that is raised and will contain all of the "Warnings."  We will have an auto-show option or you will be able to manage them yourself.  Instead of using an error provider like the broken rules implement, a window displaying all of the warnings will be presented to the user.  Obviously you will be able to turn this off and display the warnings however you see fit. 

Since there is obviously a need for this type of interaction in the field we will definitely be adding this type of functionality.  Thanks for your input and we will let you know when this new functionality is available.

In fact, we are planning on using this functionality ourselves now when we checking drug indications on our medical application.  Thanks again guys!

Well, I told someone to come up with a good reason to add warnings, and you did, so we're adding it in...

Are warning messages still on the "To do" list?

-Larry

Hehe, yes, they are, as a matter of fact.  Might not make the release today or tomorrow, but they will be included with the next version that includes full, role-based security.