Qt World Summit: Submit your Presentation

Qt5 WindowsVista widget style, broken QTreeWidget controls hover graphics

  • We just ported our large application from Qt 4.8.5 to Qt 5.5.1. We need to continue using the "WindowsVista" widget style. But we are seeing some graphical anomalies in QTreeWidget open/close controls.

    Sometimes some drawing artifacts are left over on items from the current or selected ornamentation.

    But more prominently the MOUSE HOVER ornamentation on the tree item Open/Closed control is a SQUARE FRAME (which also sometimes leaves drawing artifacts behind).

    See these image sequences (with the mouse hovering in different places):

    alt text

    We're not seeing this problem with the new (Qt5) "Fusion" widget style.

    We're seeing this on both Windows 8.1 and Windows 7. (But it doesn't happen in Windows 7 if the "Windows Classic" Personalization Theme is selected).

    We're building with Visual Studio 2010 SP1.

    Any ideas about what might be causing this QTreeWidget problem in Qt 5.5.1 with the "WindowsVista" widget style?

    Thank you in advance.

  • Moderators

    there were some known issues with the treeview expand/collapse icons.
    Try upgrading to Qt 5.6.1 at least and check again.
    See QTBUG-52230 and QTBUG-53310

  • Qt commercial support was able to provide to us a patch for Qt 5.5.1 which incorporated the necessary fixes from Qt 5.6 and Qt 5.6.1. For those who can see this (having a commercial license), this was incident INC-1070436. I'm attaching the patch file. It is applied from the 'qtbase' directory, with this patch command:

    • patch -i 1070436.diff -p 1

    It modifies these files, with no changes to public Qt interfaces:

    Usual disclaimers apply. This is, of course, has not undergone the normal release testing regimen. We are building this with Visual Studio 2010 SP1 on Windows 8.1.

    See attachment: 1070436.diff
    How do I attach a file to a forum post? I'm not seeing it.

  • Lifetime Qt Champion

    Hi please use external site and post link here.
    Thank you for reporting back.

  • Here is the a link to 1070436.diff. Usual disclaimers apply. This, of course, has not undergone the normal release testing regimen. See my prior messages in this thread:

Log in to reply