|
Evaluation
|
This problem is as a result of the fix for bugId 4784641 in 1.3.1_10.
It is not a regression in the true sense of the word, but from a end users point
of view it will look like a regression. The fix for bugId 4784641 simply
exposed the problem.
All 5 tests are failing with a similar stack:
=>[1] jni_FatalError(0x289d4, 0xff111764, 0x24, 0xffbfc, 0xff156174, 0xffbfcd7c), at 0xfefff7e8
[2] checked_jni_IsSameObject(0xff16bcc0, 0x28948, 0xb8800, 0x289d4, 0xff156174, 0xde92c), at 0xff014f98
[3] awt_currentThreadIsPrivileged(0x289d4, 0xf5b8381c, 0x262, 0x4, 0x20, 0x2), at 0xf5a9d3a4
[4] awt_util_getCurrentServerTime(0xf5b8b4f8, 0xf5b85d64, 0x289d4, 0xf5b8381c, 0x1d0460, 0xff156174), at 0xf5abccac
[5] Java_sun_awt_motif_X11Selection_pGetSelectionOwnership(0x289d4, 0xffbfd074, 0x289d4, 0xf5b8381c, 0x2, 0x0), at 0xf5ab8670
[6] 0x4e980(0xf60f8e48, 0xffbfd3b8, 0x28948, 0xff156174, 0xb6, 0xfa21e6c8), at 0x4e97f
.......
As a result of 4784641 checked_jni_IsSameObject now validates jni handles. One
of the handles being passed to checked_jni_IsSameObject, awt_MainThread is not
a valid handle. Hence the failure.
The fix is to make awt_MainThread a Global Reference ( see suggested fix ).
This is the Solaris part of 4789983.
This is not an issue in 1.4.2_0x
xxxxx@xxxxx 2004-03-05
---------------------------------------
|