[Bug tree-optimization/46987] [4.6 Regression] g++.dg/torture/covariant-1.C ICE: double free or corruption with -fno-inline

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46987

--- Comment #3 from Martin Jambor <jamborm at gcc dot gnu.org> 2010-12-17 
16:12:25 UTC ---
I've nailed this down to

  gimple_call_set_arg (call_stmt, 0, tmp);

in gimple_adjust_this_by_delta.  When I comment it out, the corruption
goes away (though the produced code is of course wrong).  Otherwise it
all blows up later in tree-ssa-ccp.  (The new parameter is a result of
a new PTR_PLUS with a thunk delta which is a constant).

But I don't see anything wrong with gimple_adjust_this_by_delta, at
least not at the moment...