Wildfly 8 and IE 8 exception

[jpappalardo]

SmartClient Version: v9.1p_2014-04-15/EVAL Deployment



I have tested with IE 8 and FF 28 and only occurs with IE 8.

[Isomorphic]

We haven't seen undertow's logs before, but this is likely the same things that Tomcat and others report as a "SocketError": the client dropped the connection mid-download. It's quite normal, and the reason you'd see this error in IE8 only is that only IE8 needs that file (for a workaround).

[jpappalardo]

http://grepcode.com/file/repo1.maven.org/maven2/io.undertow/undertow-core/1.0.0.Beta13/io/undertow/server/TruncatedResponseException.java

An exception indicating that the response channel was prematurely closed. The response channel must be shut down and flushed successfully after all requests, even those which do not send a response body.

Author(s):
    David M. Lloyd

Why would the client drop the connection?

[Isomorphic]

Because it navigated away, or the DOM element that triggered the request was removed before the request completed.

Simple example: you render an Img, and destroy() it immediately.

Something analogous is happening here, but it's related to hidden frames or RTEs or on of the other cases where we need to use empty.html for IE8.

Again, a totally normal condition, and undertow is being quite hyperbolic by reporting at ERROR level.

[jpappalardo]

Ok, but I am left with an empty web page following this exception, none of the UI is loaded.

[Isomorphic]

Start by telling us that next time please!

It's unlikely that whatever problem you're having is actually related to this server log. Run through the basics:

1. are there JS errors or anything in the Developer Console / Eclipse console aside from this error?

2. what changed such that IE8 suddenly stopped functioning?

[jpappalardo]

The dev console will not open, in the script console all I find is 'isc' is undefined as if this error stopped the client from importing your scripts?

[Isomorphic]

If the scripts seemingly aren't loading, the next step would be to look in IE8's built-in dev tools to see if there are requests going to the server for ISC_Core.js et and what happens.

Best to ignore this error (which is almost certainly a red herring) and follow the usual debug procedures.

[jpappalardo]

When trying to run in dev mode I do get the following:

[Isomorphic]

This suggests corruption or misconfiguration in your project, separate from SmartGWT.

Again if you've changed something recently, you should look into whatever you changed.

If nothing was changed and this suddenly started happening, you should clear out your GWT unitCache, restart your IDE, run a fresh GWT compile, restart your browser and clear its cache before visiting any SmartGWT page.

[jpappalardo]

That error is produced after trying to implement using SuperDevMode for debugging since FF no longer supports the dev mode plugin. However, I have a version prior to these changes running on a remote machine, and using dev mode in IE I get the following error:


As for when IE 8 broke, I cannot say for certain, I only discovered this issue when trying to ensure that I hadn't broken anything by making changes to use SuperDevMode, I always develop and test in FF.

[Isomorphic]

This error looks similar to the last - nonsense errors from GWT indicating project corruption or similar, unrelated to SmartGWT.

[jpappalardo]

It appears to work fine when I set IE to use IE 8 Standards Mode

[Isomorphic]

That seems to suggest that you have GWT configured in a way that will only run in IE8 standards. SmartGWT runs in either standards or quirks mode in IE8 (but not that HTML5 DOCTYPE is required for IE9 and up due to intractable browser bugs).

[jpappalardo]

Is that true for GWT 2.6?

Also, is there a cause you might know of off hand that would cause trying to open the dev console in IE 8 to redirect you to the login screen?