[Bug]: macOS ARM64 seamless mode leaves stale window images after moving or closing Linux X11 windows
Maintainers usually reply within 6 days
Nobody has claimed this yet.
Assessment
- Difficulty
- 5/5
- Estimated time
- Over a week
- Newbie friendliness
- 35/100
Research direction
Start by comparing the guest seamless window visible-region updates with the host transparency and repaint handling; the report has not established which boundary fails. Reproduce on macOS ARM64 with Ubuntu ARM64/X11 and Guest Additions 7.2.22, noting that the reproduction uses a private libX11 workaround whose interaction is unknown. Identify the failing boundary and verify that moving or closing a guest window clears its old desktop region.
Written by the indexing model from the issue text.
Description
Version
7.2.22 r175245 (official macOS ARM64 package and matching Linux Guest Additions)
Host OS Type
macOS
Host OS name + version
macOS 27.2.1, build 26B131; Apple M4 Pro; Retina display
Host Architecture
ARM
Guest OS Type
Linux
Guest Architecture
ARM
Guest OS name + version
Ubuntu 26.04.1 LTS ARM64, kernel 7.0.0-44-generic; XFCE / X11 / LightDM
Component
GUI / seamless windows
What happened?
In seamless mode, moving or closing a guest application window leaves persistent copies of its old image on the host desktop. The live window can still accept input, and its actual X11 geometry changes, but the previous displayed area is not cleared. This was observed with Mousepad and a terminal, independently of any web browser.
The person at the host confirmed that the artifacts are visible on the actual desktop, not just in automated screenshots. Inspection inside the guest showed that the stale images did not correspond to additional application processes/windows. Ordinary windowed desktop mode displays the applications; the problematic integration is seamless mode.
Expected: after a guest window moves or closes, its old region becomes transparent and reveals the host desktop.
VM configuration: VMSVGA, 128 MiB VRAM, 3D acceleration disabled, one virtual monitor, 4 vCPUs, 8 GiB RAM. No Extension Pack is involved.
How can we reproduce this?
- Run an Ubuntu ARM64 VM with XFCE/X11 and Guest Additions 7.2.22 on a macOS ARM64 host.
- Address the separate XShape symbol-loading failure described below so that
VBoxClient --seamlessremains running and guest application windows appear. - Open Mousepad and/or a terminal in the guest.
- Enter View → Seamless Mode (Host+L).
- Drag a guest window by its title bar and release it at a different location. Repeat, or close the guest window.
- Observe old window images remaining on the host desktop while the real X11 window has moved or been closed.
The report concerns step 6, not the initial invisible-window failure in #319. The exact failing boundary (guest visible-region tracking versus host transparency/repainting) has not yet been established.
Important prerequisite and local modifications
With the unmodified 7.2.22 Linux client, enabling seamless mode first reproduces the separate failure reported in #319:
"Visible region" mode requested
Lazy loader failed to load "XShapeQueryExtension"
Trace/breakpoint trap (core dumped)
To get past that blocker, I used a private copy of the system libX11.so.6, added a DT_NEEDED dependency on libXext.so.6 with patchelf, and set LD_LIBRARY_PATH only for VBoxClient --seamless. Resolving XShapeQueryExtension, XShapeGetRectangles, and XShapeSelectInput through that copy was verified. The Oracle client binary and system libraries were not overwritten. This is a local workaround, not a stock configuration, and I cannot yet exclude an interaction with it.
The client no longer crashes on entering seamless mode after this workaround; windows become visible, but leave the stale images described above.
Checks already performed
- XFCE compositing enabled versus disabled: stale images still observed.
- VirtualBox display scale 100% versus 200%: stale images still observed.
- Host frontend restarted with
QT_SCREEN_SCALE_FACTORS=1: did not eliminate the problem. - Guest initially did not follow host size hints. Applying the preferred XRandR mode manually, and then using the official legacy X11
VBoxClient --vmsvgapath instead ofVBoxDRMClient/--vmsvga-session, made resizing follow the host hints. This did not eliminate stale window images. See also the distinct resizing report #936. - This is not an enabled-3D configuration, unlike #71.
- #55 concerns fractional-scale pixel warping/resize loops; this report concerns entire stale window regions. I have not established that its rounding-policy patch applies here.
- #317 concerns the whole guest background remaining visible on a Linux/Wayland host, which differs from this macOS ARM64 case.
Diagnostic excerpts
Matching guest client log:
VBoxClient 7.2.22 r175245 (verbosity: 2) linux.arm64
OS Release: 7.0.0-44-generic
Detected via connection: VBGHDISPLAYSERVERTYPE_X11
Service: Seamless Mode Support
Service started
"Visible region" mode requested
The same client log also contains X11 errors, whose relevance to the stale images is not yet established:
BadWindow (invalid Window parameter) (error code 3). Request code: 20, minor code: 0
BadWindow (invalid Window parameter) (error code 3). Request code: 15, minor code: 0
BadWindow (invalid Window parameter) (error code 3). Request code: 3, minor code: 0
Error: Failed to get the window attributes for window 29360294
BadWindow (invalid Window parameter) (error code 3). Request code: 129, minor code: 6
Host log confirms seamless capability and applied guest resolutions:
VMMDev: Guest Additions capability report: (0x1 -> 0x5) seamless: yes, hostWindowMapping: no, graphics: yes
VMMDev: SetVideoModeHint: Got a video mode hint (5120x2820x32)@(0x0),(1;0) at 0
Display::i_handleDisplayResize: uScreenId=0 [...] w=5120 h=2820 bpp=32 cbLine=0x5000 flags=0x1 origin=0,0
These are excerpts, not a full VBox.log attachment. Full logs and screenshots have not been uploaded with this initial report.
- Dominant language
- C
- Stars
- 1.4k
- Forks
- 219
- PR merge metrics
- No merged PRs in 30d
Getting set up
- No Dockerfile or Docker Compose file
- No pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from VirtualBox/virtualbox
-
bug
Difficulty 1/5 Under an hour Newbie friendliness 84/100
VirtualBox/virtualbox#996 ·
Maintainers usually reply within 6 days
-
enhancement
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
VirtualBox/virtualbox#979 ·
Maintainers usually reply within 6 days
-
[Bug]: Guest Additions 7.2.20 vboxvideo fails to build on Linux kernel 6.12.113 (drm_fb_helper_restore_fbdev_mode_unlocked: too few arguments)Possibly taken @Berbe claimed this 18 days ago. Openbug
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
VirtualBox/virtualbox#918 · 9 comments ·
Maintainers usually reply within 6 days
-
bug
Difficulty 4/5 3-5 days Newbie friendliness 25/100
VirtualBox/virtualbox#995 ·
Maintainers usually reply within 6 days
-
bug
Difficulty 3/5 1-2 days Newbie friendliness 45/100
VirtualBox/virtualbox#994 ·
Maintainers usually reply within 6 days
All issues in VirtualBox/virtualbox
Similar issues
-
Difficulty 2/5 1-3 hours Newbie friendliness 80/100
ESPresense/ESPresense#2680 · 2 comments ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 74/100
ImageMagick/ImageMagick#9134 ·
Maintainers usually reply within 1 day
-
Builder: a structure opened without calculation data shows the Diffraction (fibre) view by defaultOpenbug
Difficulty 2/5 1-3 hours Newbie friendliness 65/100
FriendsofECCE/ECCE#1245 · 1 comment ·
-
Arch: ARM/AArch64 Doc
Difficulty 1/5 Under an hour Newbie friendliness 90/100
RT-Thread/rt-thread#12082 · 1 comment ·
Maintainers usually reply within 2 days
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
apache/nuttx-apps#4108 · 4 comments ·
Maintainers usually reply within 1 day