Wrong ciConstant type for arrays from ConstantPool::_resolved_reference

Staring form 8u40-b06 following crash may happen:

#
# A fatal error has been detected by the Java Runtime Environment:
#
# Internal Error (/tmp/jprt/T1/140126.ctornqvi/s/src/share/vm/c1/c1_GraphBuilder.cpp:890), pid=9025, tid=18
# assert(obj->is_instance()) failed: must be java_mirror of klass
#
# JRE version: Java(TM) SE Runtime Environment (8.0_40-b06) (build 1.8.0_40-ea-fastdebug-b06)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.40-b11-internal-201409171401.ctornqvi.hs-8u40-fastdebug compiled mode solaris-sparc compressed oops)
# Core dump written. Default location: /export/local/aurora/sandbox/results/ResultDir/compiledTargetInStackTrace_copy_1/core or core.9025
#
# If you would like to submit a bug report, please visit:
# http://bugreport.sun.com/bugreport/crash.jsp
#

--------------- T H R E A D ---------------

Current thread (0x0000000100491800): JavaThread "C1 CompilerThread3" daemon [_thread_in_native, id=18, stack(0xffffffff5ab00000,0xffffffff5ac00000)]

Stack: [0xffffffff5ab00000,0xffffffff5ac00000], sp=0xffffffff5abfdaf0, free space=1014k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
V [libjvm.so+0x12e5020] void VMError::report_and_die()+0x588
V [libjvm.so+0x760cc0] void report_vm_error(const char*,int,const char*,const char*)+0x78
V [libjvm.so+0x448774] void GraphBuilder::load_constant()+0x56c
V [libjvm.so+0x45b8c0] BlockEnd*GraphBuilder::iterate_bytecodes_for_block(int)+0x14c8
V [libjvm.so+0x45dba4] void GraphBuilder::iterate_all_blocks(bool)+0xfc
V [libjvm.so+0x461400] GraphBuilder::GraphBuilder #Nvariant 1(Compilation*,IRScope*)+0x828
V [libjvm.so+0x473c60] BlockBegin*IRScope::build_graph(Compilation*,int)+0x30
V [libjvm.so+0x473fa8] IRScope::IRScope #Nvariant 1(Compilation*,IRScope*,int,ciMethod*,int,bool)+0x2d8
V [libjvm.so+0x4746e0] IR::IR(Compilation*,ciMethod*,int)+0xb8
V [libjvm.so+0x434774] void Compilation::build_hir()+0x16c
V [libjvm.so+0x436140] int Compilation::compile_java_method()+0x2d8
V [libjvm.so+0x4368b4] void Compilation::compile_method()+0xac
V [libjvm.so+0x437454] Compilation::Compilation(AbstractCompiler*,ciEnv*,ciMethod*,int,BufferBlob*)+0x2ec
V [libjvm.so+0x43a38c] void Compiler::compile_method(ciEnv*,ciMethod*,int)+0x12c
V [libjvm.so+0x6d0618] void CompileBroker::invoke_compiler_on_method(CompileTask*)+0x590
V [libjvm.so+0x6cfa74] void CompileBroker::compiler_thread_loop()+0x4cc
V [libjvm.so+0x12271b0] void JavaThread::thread_main_inner()+0x268
V [libjvm.so+0x1226ecc] void JavaThread::run()+0x414
V [libjvm.so+0xf90da8] java_start+0x258
C [libc.so.1+0xe25fc] _lwp_start+0x8


Current CompileTask:
C1: 68040 6101 b 3 java.lang.invoke.LambdaForm$MH/863831416::collect (19 bytes)


I was not able to reproduce issue with 8u40-b05 and previous 8u releases.
With product bits nothing bad happens, so it could be just an incorrect assert.