Machine: VM-X86-3.sfbay
OS: Win 2000

Externaly problem looks like the one described in CR 6354582.
After some investigation turned out that there is a problem in printf function from native code. According to the test it have to print one of the following messages
"WeakRef gc'ed - Test PASSED"
or
"WeakRef still alive"
But in first case it prints only half of message "WeakRef gc'ed". Moreover if standard output is redirected to the file nothing will be printed.

According to the CR 6354582 evaluation Kaith build native library of this test locally and was not able to reproduce the problem. Therefore one of the possible solution is to rebuild this dll in the workspace.
Posted Date : 2007-08-01 14:00:37.0

N/A

Looking at disassembly I see incorrect message there.
After recompiling the dll the message is correct.
The fix should be as simple as replacing libraries with proper ones.

 Attributes: bp-based frame

; __stdcall Java_CheckedIsSameObjectTest_checkWeakRef(x, x)
		public _Java_CheckedIsSameObjectTest_checkWeakRef@8
_Java_CheckedIsSameObjectTest_checkWeakRef@8 proc near

arg_0		= dword	ptr  8

		push	ebp
		mov	ebp, esp
		push	0
		mov	eax, dword_10003034
		push	eax
		mov	ecx, [ebp+arg_0]
		push	ecx
		mov	edx, [ebp+arg_0]
		mov	eax, [edx]
		call	dword ptr [eax+60h]
		and	eax, 0FFh
		test	eax, eax
		jz	short loc_1000104F
		push	offset aWeakrefGcEd ; "WeakRef gc'ed"
		call	ds:printf
		add	esp, 4
		jmp	short loc_1000105D
Posted Date : 2007-08-23 15:40:13.0