Here is one more exception trace I got.  It seems that by varying things
a little, I can get all kinds of different exception traces.  It
definitely looks like something is getting corrupted.

-------------------------------------------------------------------------------
ARRAY               item @1                Segmentation violation:      
<000000000E87212C>                         Operating system signal.      Fail
-------------------------------------------------------------------------------

In this case, I didn't change the .ecf file.  I simply finalized the
system (using ec -loop and without assertions or inlining), then changed
the root class so that it didn't reference any other classes except
ARRAY and STRING.  I re-finalized in same ec session.  Then I changed
the root class back to its original contents and re-finalized again.
This resulted in the attached trace (stack17.txt).

To reproduce the crash:

Unpack attached tar file, creating a directory "bug".
Change bug/bug.ecf line referencing "mosis_source" to the correct path.
Run "ec -loop -config bug/bug.ecf" (config file is in "bug" directory).
  (Note: I didn't try this with the -experiment flag yet, but I suspect
   that will crash as well)
Select I for Compile menu.
Type Z to finalize (discard assertions).
When compilation finishes, comment out the two lines in class TEST
that reference the "real" root class DELTA.
Finalize again (same ec -loop session), discarding assertions.
When compilation finishes, add back the two lines commented out above,
   so that TEST has its original contents.
Finalize a 3rd time (same ec -loop session), discarding assertions.
Compiler crashes with a seg fault in degree -3.

If you add the -verbose flag to the "ec -loop" commnand line, it may
crash in a different place.

In my opinion, this is a critical bug that needs to be fixed for 6.6.
Please add it to my list of requested bug fixes for 6.6.
This bug may explain why we sometimes get estudio crashes with exception
traces that don't look like any of the known bugs in the eweasel test suite.

I cannot add an eweasel test for this bug, because so far it has only
occurred when all the compilations where in one estudio or "ec -loop" session.

I originally encountered this bug when running estudio, so I think it occurs
there as well.