[Bug 36558] New: valgrind shows a definite leak in comctl32/tests/treeview.c

https://bugs.winehq.org/show_bug.cgi?id=36558

Bug ID: 36558
           Summary: valgrind shows a definite leak in
                    comctl32/tests/treeview.c
           Product: Wine
           Version: 1.7.19
          Hardware: x86
                OS: Linux
            Status: NEW
          Keywords: download, source, testcase
          Severity: normal
          Priority: P2
         Component: comctl32
          Assignee: [e-mail]
          Reporter: [e-mail]

==26061== 140 bytes in 1 blocks are definitely lost in loss record 620 of 921
==26061==    at 0x7BC4C6DF: notify_alloc (heap.c:255)
==26061==    by 0x7BC50F23: RtlAllocateHeap (heap.c:1716)
==26061==    by 0x4FE3841: ImageListImpl_CreateInstance (imagelist.c:3697)
==26061==    by 0x4FDC1EA: ImageList_Create (imagelist.c:772)
==26061==    by 0x505A37F: TREEVIEW_InitCheckboxes (treeview.c:2954)
==26061==    by 0x505A6A5: TREEVIEW_Paint (treeview.c:3013)
==26061==    by 0x5060D32: TREEVIEW_WindowProc (treeview.c:5817)
==26061==    by 0x538D4F5: ??? (winproc.c:173)
==26061==    by 0x538D66A: call_window_proc (winproc.c:244)
==26061==    by 0x538E858: WINPROC_CallProcAtoW (winproc.c:603)
==26061==    by 0x538FA49: CallWindowProcA (winproc.c:961)
==26061==    by 0x4D435DF: TreeviewWndProc (treeview.c:315)
==26061==    by 0x538D4F5: ??? (winproc.c:173)
==26061==    by 0x538D66A: call_window_proc (winproc.c:244)
==26061==    by 0x538F8B9: WINPROC_call_window (winproc.c:909)
==26061==    by 0x5350859: DispatchMessageA (message.c:3948)
==26061==    by 0x4D4B261: test_TVS_CHECKBOXES (treeview.c:1943)
==26061==    by 0x4D4C651: func_treeview (treeview.c:2228)
==26061==    by 0x4D51DB9: run_test (test.h:584)
==26061==    by 0x4D521A8: main (test.h:654)
==26061==

[Bug 36558] valgrind shows a definite leak in comctl32/tests/treeview.c

https://bugs.winehq.org/show_bug.cgi?id=36558

--- Comment #1 from Austin English  ---
A second one:
==26061== 560 bytes in 4 blocks are definitely lost in loss record 756 of 921
==26061==    at 0x7BC4C6DF: notify_alloc (heap.c:255)
==26061==    by 0x7BC50F23: RtlAllocateHeap (heap.c:1716)
==26061==    by 0x4FE3841: ImageListImpl_CreateInstance (imagelist.c:3697)
==26061==    by 0x4FDC1EA: ImageList_Create (imagelist.c:772)
==26061==    by 0x505A37F: TREEVIEW_InitCheckboxes (treeview.c:2954)
==26061==    by 0x505FD23: TREEVIEW_StyleChanged (treeview.c:5495)
==26061==    by 0x5060DE8: TREEVIEW_WindowProc (treeview.c:5838)
==26061==    by 0x538D4F5: ??? (winproc.c:173)
==26061==    by 0x538D66A: call_window_proc (winproc.c:244)
==26061==    by 0x538E858: WINPROC_CallProcAtoW (winproc.c:603)
==26061==    by 0x538FA49: CallWindowProcA (winproc.c:961)
==26061==    by 0x4D435DF: TreeviewWndProc (treeview.c:315)
==26061==    by 0x538D4F5: ??? (winproc.c:173)
==26061==    by 0x538D66A: call_window_proc (winproc.c:244)
==26061==    by 0x538F697: WINPROC_CallProcWtoA (winproc.c:858)
==26061==    by 0x538F842: WINPROC_call_window (winproc.c:902)
==26061==    by 0x534C11C: call_window_proc (message.c:2223)
==26061==    by 0x534EFBA: send_message (message.c:3260)
==26061==    by 0x534F5D8: SendMessageW (message.c:3454)
==26061==    by 0x5381439: WIN_SetWindowLong (win.c:2515)
==26061==

[Bug 36558] valgrind shows a definite leak in comctl32/tests/treeview.c

https://bugs.winehq.org/show_bug.cgi?id=36558

--- Comment #2 from Austin English  ---
The second leak also shows up in explorerframe/tests/nstc.c:
==27764== 140 bytes in 1 blocks are definitely lost in loss record 568 of 813
==27764==    at 0x7BC4C6DF: notify_alloc (heap.c:255)
==27764==    by 0x7BC50F23: RtlAllocateHeap (heap.c:1716)
==27764==    by 0x6725841: ImageListImpl_CreateInstance (imagelist.c:3697)
==27764==    by 0x671E1EA: ImageList_Create (imagelist.c:772)
==27764==    by 0x679C37F: TREEVIEW_InitCheckboxes (treeview.c:2954)
==27764==    by 0x67A1D23: TREEVIEW_StyleChanged (treeview.c:5495)
==27764==    by 0x67A2DE8: TREEVIEW_WindowProc (treeview.c:5838)
==27764==    by 0x550E4F5: ??? (winproc.c:173)
==27764==    by 0x550E66A: call_window_proc (winproc.c:244)
==27764==    by 0x5510ABE: CallWindowProcW (winproc.c:981)
==27764==    by 0x66CB9CE: ???
==27764==    by 0x550E4F5: ??? (winproc.c:173)
==27764==    by 0x550E66A: call_window_proc (winproc.c:244)
==27764==    by 0x5510803: WINPROC_call_window (winproc.c:900)
==27764==    by 0x54CD11C: call_window_proc (message.c:2223)
==27764==    by 0x54CFFBA: send_message (message.c:3260)
==27764==    by 0x54D05D8: SendMessageW (message.c:3454)
==27764==    by 0x5502439: WIN_SetWindowLong (win.c:2515)
==27764==    by 0x55026BC: SetWindowLongW (win.c:2670)
==27764==    by 0x66CDB2E: ???
==27764==

[Bug 36558] valgrind shows a definite leak in comctl32/tests/treeview.c

https://bugs.winehq.org/show_bug.cgi?id=36558

--- Comment #3 from Nikolay Sivov  ---
I remember this thing. I tried to fix it once and it turned out it's not
possible to properly maintain state imagelist because some applications rely on
it not being destroyed (as i remember some app used imagelist from one treeview
in another and it broke once i started to release it).

[Bug 36558] valgrind shows a definite leak in comctl32/tests/treeview.c

https://bugs.winehq.org/show_bug.cgi?id=36558

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |WONTFIX

--- Comment #4 from Austin English  ---
(In reply to Nikolay Sivov from comment #3)
> I remember this thing. I tried to fix it once and it turned out it's not
> possible to properly maintain state imagelist because some applications rely
> on it not being destroyed (as i remember some app used imagelist from one
> treeview in another and it broke once i started to release it).

I've marked it as intentional, thanks.

[Bug 36558] valgrind shows a definite leak in comctl32/tests/treeview.c

https://bugs.winehq.org/show_bug.cgi?id=36558

Nikolay Sivov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|WONTFIX                     |---

--- Comment #5 from Nikolay Sivov  ---
For now unmark it please. I have to look at it again, probably I can use
refcount to make it work.

[Bug 36558] valgrind shows a definite leak in comctl32/tests/treeview.c

https://bugs.winehq.org/show_bug.cgi?id=36558

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |valgrind

[Bug 36558] valgrind shows a definite leak in comctl32/tests/treeview.c

https://bugs.winehq.org/show_bug.cgi?id=36558

--- Comment #6 from Austin English  ---
(In reply to Nikolay Sivov from comment #5)
> For now unmark it please. I have to look at it again, probably I can use
> refcount to make it work.

Sure.