Announcement

**Isomorphic** · 10 Dec 2015, 07:52

Both of these are crashes somewhere in your application code, but you are looking at stack traces for obfuscated code. See the Debugging overview to see how to set up either SuperDevMode or classic Development Mode, so you can debug what's going wrong in your code.

At that point, let us know if you think there is actually a framework issue at fault.

**Emilien** · 10 Dec 2015, 08:11

The stacktrace is from our code because it's handle by our UncaughtExceptionHandler but the breakpoint is trigger by "grid.startEditingNew();"

Here is the stacktrace while using the SuperDevMode .. I'm not sure how I can help you understand this issue, it's really bugging me for 2 days now.

May I emphasises the fact that everything works fine with the 4.1p-2014-12-05, so I guess something changed in the startEditingNew() between thoses 2 versions :x

**Isomorphic** · 10 Dec 2015, 08:17

With both the original trace and the SDM screenshot, it appears that you've got a click handler on some kind of button, and your click handler is crashing. Once your code in the click event crashes, *then* the UncaughtExceptionHandler is called.

Note that if you look at the line when you've set a breakpoint, that line of code cannot actually crash. So that's definitely not the problem.

You can easily verify this by just logging something right when your click handler is called, before you call any APIs that might crash.

**Emilien** · 10 Dec 2015, 08:26

Yup my click handler is crashing, here is the code of the click handler :

Code:

  @Override     public void onClick(ClickEvent event) {
             LOG.fine("on click");
             grid.startEditingNew();  
             LOG.fine("after click");  
         }

And if I remove the dataSource of the Grid, so new line is created on the grid and the edition is possible; this function ("StartEditingNew") appears to crash only when the grid is using a RestDatasource.

Is there any way to debug inside the StartEditingNew to understand what is really happening and where is this null property ?

**Isomorphic** · 10 Dec 2015, 08:31

startEditingNew() doesn't appear in the stack trace, so it appears your code crashes before ever calling that API. Perhaps "grid" is null.

**Emilien** · 10 Dec 2015, 08:38

I can assure you that grid is not null .. If I comment the "grid.setDataSource(ds);" everything works fine (except record from the database are not loaded in the grid, but the grid is editable fine) but when setting everything the startEditingNew don't seems to be able to work at all :
Before executing startEditingNew:

After :

**Isomorphic** · 10 Dec 2015, 08:47

The stack trace makes it clear that startEditingNew() was never entered. One way or another, this indicates your environment is broken - either the trace is wrong, or some kind of GWT issue is occurring where there is a crash before startEditingNew() is actually invoked.

At this point, you should try to shake out whatever is wrong with your environment: clear the GWT unitcache, rerun a GWT compile, restart the IDE, possibly reinstall GWT, possibly reboot. You could also try another developer's machine, and/or a fresh install of the whole toolchain.

Note that, while there are instructions in the Debugging guide for how to run with debug modules so you can see the source of startEditingNew(), again all signs point to this method never being entered, so this probably a waste of time.

Announcement

RestDataSource and startEditingNew

Comment

Comment

Comment

Comment

Comment

Comment

Comment