|
2026-10-29
| ||
| 15:59 | • Closed ticket [546607ff]: problem with cursor change plus 6 other changes artifact: c6f4de57 user: nemethi | |
| 15:51 | Fix [546607ffff]: Do not reset the cursor over a window not managed by Tk check-in: 3602edd9 user: csaba tags: core-8-6-branch | |
| 15:43 | Fix [546607ffff]: Do not reset the cursor over a window not managed by Tk check-in: fef2f0d0 user: csaba tags: core-9-0-branch | |
|
2026-10-23
| ||
| 14:29 | • Pending ticket [546607ff]: problem with cursor change plus 8 other changes artifact: 5cd77522 user: jan.nijtmans | |
| 14:28 | Fix [546607ffff]: Do not reset the cursor over a window not managed by Tk check-in: e836917d user: jan.nijtmans tags: trunk, main | |
|
2026-10-22
| ||
| 22:49 | • Ticket [546607ff] problem with cursor change status still Open with 4 other changes artifact: 4f3a0b1d user: serhiy.storchaka | |
|
2026-09-23
| ||
| 21:58 | • Ticket [546607ff]: 3 changes artifact: d690c959 user: serhiy.storchaka | |
| 21:55 |
Fix [546607ffff]: Do not reset the cursor over a window not managed by Tk.
Tk_HWNDToWindow() returns NULL when the pointer is over a window which Tk does not own, for example a child window created by an extension. Tk_UpdatePointer() then called UpdateCursor(NULL), which fell through to TkpSetCursor(None) and loaded the default arrow cursor, so the 250 ms mouse timer kept overwriting the cursor set by the owner of that window. closed check-in: a283eb2a user: serhiy.storchaka tags: win-foreign-cursor | |
|
2002-07-06
| ||
| 05:51 | • Ticket [546607ff] problem with cursor change status still Open with 2 other changes artifact: 13f3c7ad user: dkf | |
|
2002-04-20
| ||
| 21:58 | • New ticket [546607ff]. artifact: 8e94f994 user: nobody | |
| Ticket UUID: | 546607 | |||
| Title: | problem with cursor change | |||
| Type: | Bug | Version: | 9.1.0 | |
| Submitter: | nobody | Created on: | 2002-04-20 21:58:02 | |
| Subsystem: | 68. Win Window Operations | Assigned To: | jan.nijtmans | |
| Priority: | 5 Medium | Severity: | Minor | |
| Status: | Closed | Last Modified: | 2026-10-29 15:59:04 | |
| Resolution: | Fixed | Closed By: | nemethi | |
| Closed on: | 2026-10-29 15:59:04 | |||
| Description: |
I have a problem with cursor change using Tkogl (opengl extension) under Windows.
Tkogl creates Tk_Window-window (A) and opengl-window (B) as a child of A.
B is not Tk_Window. B just posts some messages to its parent (A) so I have no
problem with bindings and with cursor change. But I can see the desired cursor (over B)
only when I move the mouse or press the button. As soon as I stop moving or release the button
- the cursor changes to IDC_ARROW.
It happens because of "mouseTimer" that is created in Tk_PointerEvent(hwnd, x, y)
(tkWinPointer.c line 149).
MouseTimerProc calls Tk_PointerEvent(NULL, pos.x, pos.y) (tkWinPointer.c line 239).
Note that hwnd is NULL, so Tk_PointerEvent "must" find hwnd.
hwnd = WindowFromPoint(pos); (tkWinPointer.c line 139).
WindowFromPoint returns hwnd of B (rightly) which is not Tk_Window.
When
tkwin = Tk_HWNDToWindow(hwnd); (tkWinPointer.c line 141)
so tkwin is NULL now.
After this
Tk_UpdatePointer(tkwin, pos.x, pos.y, state); (tkWinPointer.c line 145)
with tkwin==NULL will change cursor to "None".
I see two simple ways to avoid this problem -
1. Not to change cursor to "None" in UpdateCursor(winPtr) (TkPointer.c line 546) when
winPtr==NULL.
2. Replace
tkwin = Tk_HWNDToWindow(hwnd); (tkWinPointer.c line 141)
with
tkwin = NULL;
while (tkwin==NULL && hwnd!=NULL) {
tkwin = Tk_HWNDToWindow(hwnd);
hwnd = GetParent(hwnd);
}
| |||
| User Comments: |
nemethi (claiming to be Csaba Nemethi) added on 2026-10-29 15:59:04:
Backported to core-9-0-branch and core-8-6-branch by commits [fef2f0d0] and [3602edd9]. Closing the ticket. jan.nijtmans added on 2026-10-23 14:29:40: Fixed on [e836917d|trunk] If CI build is clean, can be merged further. serhiy.storchaka added on 2026-10-22 22:49:16: Confirmed on Windows 11 with current main: over a window which Tk does not own, inside a frame with -cursor watch, the cursor turns into the arrow as soon as the mouse stops. With branch win-foreign-cursor the frame cursor stays. serhiy.storchaka added on 2026-09-23 21:58:01: Still reproducible in 9.1b1 on Windows 11, and the analysis in the report is still accurate.
A fix is proposed in branch win-foreign-cursor, the first of the two suggestions in the report: leave the cursor alone when there is no Tk window under the pointer. | |||
