RFR: 8346281: [Windows] RenderScale doesn't update to HiDPI changes

Jose Pereda jpereda at openjdk.org
Tue Jan 7 22:21:41 UTC 2025


On Tue, 7 Jan 2025 21:29:17 GMT, Kevin Rushforth <kcr at openjdk.org> wrote:

>> This PR adds the missing native implementation for Windows `GlassWindow::HandleDPIEvent`, to notify the (Java) window when there is a DPI change event, which can happen when the user changes the resolution of the screen (via Settings -> System -> Display -> scale), while the JavaFX application is running.
>> 
>> When such `WM_DPICHANGED` event happens, `GlassWindow::HandleDPIEvent` notifies the (Java) window, which changes its platform scale via `Window::notifyScaleChanged`, and `GlassScreen::HandleDisplayChange();` is needed too, to update the platform scale of the screen where the window is at as well.
>> 
>> There are no tests added to this PR, since these would require manual intervention to change the display. In any case, the test case added to the [issue](https://bugs.openjdk.org/browse/JDK-8346281) runs fine now when the app runs on a given screen and the user changes its resolution. 
>> 
>> There is a follow-up issue after this PR, that might require a more complex fix, for the case where the user changes the resolution of a different screen that is placed before the one that has the application (as location coordinates of the latter depend on the former), because there is no `WM_DPICHANGED` event in this case.
>
> As a quick note, when the multi-monitor support was added for HiDPI, the code to check `WM_DPICHANGED` was intentionally removed. See [this comment in JDK-8146920](https://bugs.openjdk.org/browse/JDK-8146920?focusedId=14034147&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14034147). I'll look closer when I review it.

Thanks for the link @kevinrushforth , it helps a lot clarifying a few things. 

I didn't find issues with `WM_DPICHANGED` (yet), but I didn't test those edge cases that are mentioned.

I see several different issues related to DPI changes (with and without my patch), and I can reproduce the issue that @andy-goryachev-oracle  mentioned, so definitely this PR needs more work, and probably some follow-ups.

-------------

PR Comment: https://git.openjdk.org/jfx/pull/1668#issuecomment-2576337387


More information about the openjfx-dev mailing list