Hi, sometimes I'm very disappointed from the Microsoft AJAX team, there are good ideas, but implementation... It's very interesting - did they test a new release before put it for the downloading?!

I have a project, and all are working good with previous version, but  I installed new version... On the first home page with a simple MultiView control in the UpdatePanel and two ValidatorCalloutExtender controls, there are different errors after clicks - "'_events' is null or not an object", "'originalWidth' is null or not an object", also if both textboxes related to the ValidatorCalloutExtender controls are empty and I click on the submit button - the error message for first textbox is displayed correctly, but I enter some text in the first textbox, click submit button again - an error message for second (empty!) textbox isn't displayed.

So what is it? Release version?!!

P.S. I modified and checked changes in the web.config few times... I didn't miss something. Also I updated AjaxControlToolkit.dll...

The custom validators were removed.  The original validators have been made compatible and will be on WindowsUpdate soon.  In the meantime, there is a solution here.

This is a mess. The final release is put up with no notice about the fact that this requires some other update which has not been released. Would have been much better off just holding the release until the patch was available IMHO. I'm rolling back to RC. Thanks for the warning!

 ...Scott

Will this update be released before next the monthly updates for February?

The changes that go out on Windows Update get extensive testing.  I hope it is complete in time for the February update but it may be after that.

 In the meantime, I've posted the sources and the DLL.  You can go forward with the 1.0 release by just including the dll in the /bin directory and including the <tagMapping> that references the temporary asembly.

We are going to stay with the RC for now.  We just switched most of our live sites over to it from the July CTP during the last week, and we have a lot of changes that we are working on now, so we are going to wait until nothing will change for a while before we upgrade again.

hello guys.

just letting some steam out: my opinion is that this was a bad call.

I understand the frustration.  It was a tough call but one I still think was the right call for the future and for the biggest number of people.  By eliminating the compatibility validators and relying on the core validators we make it so that many derived validator controls will just work with Atlas.  They won't need to re-work, re-compile, and re-release in order to be part of partial page updates.  The confusion over the Compatibility validators would go on for years.  Sure, we could have waited to release until the update is out and this change would hardly have been noticed but many have been asking for supported RTM bits in order to adopt AJAX for corporate projects.

hello Matt.

Speaking only by myself, i'd say that the only validator I've built till today was a callback validator that would perform a callback through xmlhttprequest to validate the value of a control.

btw, what about the other errors I've found? most of them are on the future bits, but they're simply incredible. I'll put them here:

* the previewscript embedded file is different from the one you get in the installation folder. the onle in the installation forlder has at least 2 double classes/interfaces registration (which produces an error that says that then name you're using is already in use). Currently, i'm using the file that I extracted from the dll, but this file is missing several behaviors that are defined on the file that you get in the installation folder. what's the story here? I know that we're talking about future bits, but if you don't intend to maintain them and upgrade their functionalities, why have you kept them for so long? the transition for beta was a good opportunity to get rid of them (though I'd be really sad if this happens because i do like most of the things that's defined there)

* you've changed the json web services calling "convention" and didn't even put  a note about it. why do I care? well because i'm using the datasource class to call a web service method (not a data service) and in previous releases getting away with this was really easy (it still is, but i did loose some type investigating the dll to see how the handler factory decides on which handle to use)

* the listview control selected index bug. this is one of those bugs that are there for centuries! I've been complaining about it for several months and it still persists...how hard can it be to solve it?

* propertyKey property defined as Object (instead of string) in the animation files

* error on the descriptor of the darggablelistitem: this one is interesting because the error message you get tells you exactly why you're getting that error!

* there's also a bug on floatingbehavior (in the file you're using undraggableTagNames.indexOf instead of Array.indexOf( undraggableTagNames, ... )

* dont know why, but the invokemethod of the profile class in the future bits insists in calling this.load.apply( this, parameters) instead of simply calling this.load(parameters) - the 1st one crashes the app

yep, all of these are on the future bits...only one is new; all the others were there in previews releases. they weren't corrected. does that mean that these features will be discarded? If that is the case, I'd really appreciate getting some info on that...

I downloaded the validators DLL and added the tagmapping and for most part my validator callout extenders are working again.  One exception is that I have a textbox with a calender popup and a require field validator and a compare validator (making sure its a date).  Both have callout extenders.  The calender popup works as do the callouts, but after after a postback attempt and the callout for the rfv shows, the ability for the popup calender panel to update the textbox errors with "originalWidth is null or not an object".

I created this post and didn't expect that discussion will be so intensive... I think it is a right name of the post ;)

To Jason Wilson: I did the same, and on the default home page (there are few required validators with extenders) - I have the same javascript error (!) - "originalWidth is null or not an object", or sometimes "this.get_element().style is null or not an object".

Honestly - I'm tired to search fixes for problems and my project don't allow it...

It is fustrating that I've spent all day fixing things that were working yesterday.  Just goes to show that you don't mess with development environment mid way through a project (but when am I not mid way through a project?) -- I only decided to do so because of the "78 bug fixes since RC1".

I guess it was a good thing that I didn't get much of a chance to start on this yesterday because I would have done more than just installing the RTM before I found out about the problem.

The question is whether I have done too much to go back...  At this point I am seriously considering it...  What a pain

The question is whether I have done too much to go back...  At this point I am seriously considering it...  What a pain

 

Never again on day 1