https://bugs.winehq.org/show_bug.cgi?id=52987
Bug ID: 52987 Summary: comctl32:tooltips - test_customdraw() fails systematically on some Windows 10 machines Product: Wine Version: unspecified Hardware: x86-64 OS: Windows Status: NEW Severity: normal Priority: P2 Component: comctl32 Assignee: wine-bugs@winehq.org Reporter: fgouget@codeweavers.com
comctl32:tooltips - test_customdraw() fails systematically on some Windows machines:
tooltips.c:251: Test failed: 0: Failed to get current tool 0. tooltips.c:258: Test failed: 0: Unexpected hwnd CCCCCCCCCCCCCCCC. tooltips.c:259: Test failed: 0: Unexpected hinst CCCCCCCCCCCCCCCC. tooltips.c:260: Test failed: 0: Unexpected uId cccccccccccccccc. tooltips.c:261: Test failed: 0: Unexpected lParam cccccccccccccccc. tooltips.c:251: Test failed: 1: Failed to get current tool 0. tooltips.c:258: Test failed: 1: Unexpected hwnd CCCCCCCCCCCCCCCC. tooltips.c:259: Test failed: 1: Unexpected hinst CCCCCCCCCCCCCCCC. tooltips.c:260: Test failed: 1: Unexpected uId cccccccccccccccc. tooltips.c:261: Test failed: 1: Unexpected lParam cccccccccccccccc. [... all 6 iterations fail ...]
https://test.winehq.org/data/patterns.html#comctl32:tooltips
Also this impacts both test_customdraw() calls: the one after test_create_tooltip(FALSE) and the one after test_create_tooltip(TRUE) though the latter is more random. When the latter fails there is an extra couple of failures in test_TTN_SHOW():
tooltips.c:261: Test failed: 5: Unexpected lParam cccccccccccccccc. tooltips.c:1216: Test failed: TTN_SHOW parent seq: the msg sequence is not complete: expected 004e - actual 0000 tooltips.c:1216: Failed sequence TTN_SHOW parent seq: tooltips.c:1216: 0: expected: msg 004e - actual: nothing v6util.h:92: created cc6.manifest tooltips.c:251: Test failed: 0: Failed to get current tool 0. [...] tooltips.c:1216: Test failed: TTN_SHOW parent seq: the msg sequence is not complete: expected 004e - actual 0000
However the failure pattern is strange: * The failures mostly happen on the cw-gtx560 and cw-rx460 machines. But w1064v1709-64 has the same failures so it's not a real hardware vs. VM issue. * The failures happen on Windows 10 1709, 21h1 and 21h2, but not on 1507, 1607, 1809 or 2004. 1909 had failures on March 21 and 22 but not in the month since then. So it does not really seem related to the Windows version. * cw-gtx560-21H1 and cw-rx460-21H1 never had failures in the 32-bit tests but the failures happen in 32-bit on cw-gtx560-1709 and cw-rx460-1709. So it's not a bitness issue and it should not be a machine configuration issue either.
The failure on line 251 indicates that there is no tooltip at all. Maybe there is a window that has the focus and prevented the tooltip from opening? (that window if any is not visible in the final screenshot)
If it is a focus issue, skipping when the windows does not have the focus could avoid the failures.