Do you use NotNull instrumentation?

I am using defaults and did not know that by default @NotNull assertion checking is turned on. So, yes, I use NotNull instrumentation, because it is ON by default.

Does turning them off help?

I have the same issue. Turning off NotNull assertions while keeping code coverage reporting on doesn't help.
Turning off code coverage gets rid of the issue.

This must be emma tool issue (IDEA uses it as a code coverage tool). Recently we've applied some fixes to it. The updated emma version is used in IDEA 6.0.2 (currendly EAP version is available). I was not able to reproduce the problem with it.
Please try IDEA 6.0.2 to check whether the problem still exist.

I've upgraded to 6.02 and I am encountering the same behavior. I can provide further details if need be.

A small project that illustrates the behaviour would be great.

Complete project demonstrating this problem - simply run in debug mode, enabling code coverage for CoverageClass if necessary.

I've also added a screenshot illustrating this bug in action. Let me know if you can't duplicate it.

Thanks a lot, repeatable. Looks like indeed emma issue (Error 35 is in fact "Invalid Slot" JDWP error). Will investigate whether it will be possible to patch it.

I have this error even when I don't have code coverage enabled, is it a known issue?

This may also occur when you have @NotNull assertions generation turned on.

I am using version 6.0.4 and having the same issue. The only way to stop it is to turn off code coverage. Will this be fixed in a future 6.0 version patch?

I'm still seeing this in 6.0.5.

The problem has been fixed completely for @NotNull compiler.
As for the code coverage, we had to disable the feature if coverage-enabled configuration is launched by debugger. The problem is that Emma coverage tool used by IDEA leaves "LocalVariableTable" attribute unchanged for instrumented methods. Thus the essential part of debug information gets screwed up, so that debugger simply does not get on well with the instrumented code.

Will this fix make it into a v6 point release?

yep, I'll backport the changes, though can't give you an estimate now when 6.0.6 is going to be available.

It looks like this is fixed in build 6189 (6.0.6 RC2), Eugene can you confirm that it has been back-ported, and maybe set the correct "Fix Version" flag so that it shows up in the 6.0.6 release notes?

Yes, I indeed backported this one recently :-)

Thanks!

Note, however, my comment above:
"As for the code coverage, we had to disable the feature if coverage-enabled configuration is launched by debugger"

I have IDEA 6.0.6 and bug is still reproduced :(

I've tried to check/uncheck "Add @NotNull assertions" but it doesn't help. More information is in the Declined Issue
http://www.jetbrains.net/jira/browse/IDEA-10670