suppress test is failing on Windows since r485

Question

From timurrrr.com on September 01, 2011 10:08:21

Since r485 , suppress test is failing on Win XP 32-bit http://build.chromium.org/p/client.drmemory/builders/win-xp-drm/builds/1450 [taken from build 1460] 16/23 Test: suppress Command: "C:/Program Files/CMake 2.8/bin/cmake.exe" "-D" "cmd:STRING=C:/Perl/bin/perl.exe@../bin/drmemory.pl@-debug@-dr_debug@-dr@{DRMEMORY_CTEST_DR_DIR}/..@-batch@-dr_ops@-dumpcore_mask@@@-stderr_mask@@@-msgbox_mask@@-callstack_style@-suppress@{DRMEMORY_CTEST_SRC_DIR}/suppress.win.suppress@-suppress@{DRMEMORY_CTEST_SRC_DIR}/suppressB.win.suppress@--@./suppress.exe" "-D" "TOOL_DR_HEAPSTAT:BOOL=OFF" "-D" "outpat:STRING={DRMEMORY_CTEST_SRC_DIR}/suppress.out" "-D" "respat:STRING={DRMEMORY_CTEST_SRC_DIR}/suppress.res" "-D" "nudge:STRING=C:/Perl/bin/perl.exe@../bin/drmemory.pl@-debug@-dr_debug@-dr@{DRMEMORY_CTEST_DR_DIR}/..@-batch" "-D" "VMKERNEL:BOOL=OFF" "-D" "USE_DRSYMS:BOOL=OFF" "-D" "toolbindir:STRING={DRMEMORY_CTEST_DR_DIR}/../bin32" "-D" "DRMEMORY_CTEST_SRC_DIR:STRING=E:/b/build/slave/win_xp_drm_checkout/drmemory/tests" "-D" "DRMEMORY_CTEST_DR_DIR:STRING=E:/b/build/slave/win_xp_drm_checkout/build/build_drmemory-cyg-dbg-32/dynamorio/cmake" "-D" "CMAKE_SYSTEM_VERSION:STRING=5.1" "-D" "postcmd:STRING=" "-P" "./runtest.cmake" Directory: E:/b/build/slave/win_xp_drm_checkout/build/build_drmemory-cyg-dbg-32/tests "suppress" start time: Sep 01 05:42 Pacific Daylight Time Output:

CMake Error at runtest.cmake:352 (message): stderr failed to match ":::Dr.Memory::: 0 unique, 0 total uninitialized access(es) ?", found ":::Dr.Memory::: 1 unique, 1 total uninitialized access(es)" instead

(related to issue #576 ?)

Other Windows bots are failing too but looks like they were failing earlier.

Original issue: http://code.google.com/p/drmemory/issues/detail?id=577

该提问来源于开源项目：DynamoRIO/drmemory

Answer 1

From bruen....com on September 01, 2011 07:29:54

(related to issue #576 ?)

no, no relation whatsoever, issue #576 is a DR memory area issue w/ execution on the stack, which cannot occur on windows

this is likely mod+offs not quite matching. I relaxed the +offs later. r490 only has cygwin failing. cygwin passes on my machine but perhaps there's some slightly different build bits on the bot. need to run with -callstack_style 0x10f on the bot to know for sure.