April 2024 - Wine-gitlab - list.winehq.org

[PATCH v4 0/2] MR5269: Add NtQueryInformationProcess(ProcessQuotaLimits) stub
by Gijs Vermeulen (＠gverm) 28 May '24

28 May '24

Based on a patch by Qian Hong. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=44812 -- v4: ntdll: Add NtQueryInformationProcess(ProcessQuotaLimits) tests. ntdll: Add NtQueryInformationProcess(ProcessQuotaLimits) stub. https://gitlab.winehq.org/wine/wine/-/merge_requests/5269

5 8

[PATCH v5 0/2] MR5497: ntdll: Implement NtQueueApcThreadEx()
by Isaac Marovitz (＠IsaacMarovitz) 28 May '24

28 May '24

Should resolve a crash in `libogg` in the game "Cabal Online" -- v5: wow64: Add NtQueueApcThreadEx() https://gitlab.winehq.org/wine/wine/-/merge_requests/5497

2 3

[PATCH v3 0/1] MR5497: ntdll: Implement NtQueueApcThreadEx()
by Isaac Marovitz (＠IsaacMarovitz) 28 May '24

28 May '24

Should resolve a crash in `libogg` in the game "Cabal Online" -- v3: ntdll: Implement NtQueueApcThreadEx() https://gitlab.winehq.org/wine/wine/-/merge_requests/5497

3 2

[PATCH 0/10] MR5444: jscript: Do not allocate dispex_prop_t for indexed props.
by Gabriel Ivăncescu 28 May '24

28 May '24

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/5444

3 26

[PATCH v2 0/4] MR5488: win32u: Add display device cache modes all at once.
by Rémi Bernon 28 May '24

28 May '24

To make https://gitlab.winehq.org/wine/wine/-/merge_requests/5422 simpler. This will also make it easier to write the mode list to the registry as a single binary blob, although it is a larger change that could be done later. -- v2: win32u: Add all display device source modes at once. win32u: Keep the primary current mode in the device manager context. win32u: Remove fake source creation when adding display mode. win32u: Close device manager source key in write_source_to_registry. https://gitlab.winehq.org/wine/wine/-/merge_requests/5488

2 6

[PATCH 0/4] MR5488: win32u: Add display device cache modes all at once.
by Rémi Bernon 28 May '24

28 May '24

To make https://gitlab.winehq.org/wine/wine/-/merge_requests/5422 simpler. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/5488

2 10

[PATCH 0/1] MR5489: win32u: Also reset the unnamed values in prepare_devices.
by Alex Henrie (＠alexhenrie) 28 May '24

28 May '24

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=56551 -- https://gitlab.winehq.org/wine/wine/-/merge_requests/5489

3 2

[PATCH 0/2] MR5487: winemac.drv: Update OpenGL context immediately after the window content view is visible.
by Zhiyi Zhang (＠zhiyi) 28 May '24

28 May '24

Fix errors such as GL_FRAMEBUFFER_UNDEFINED and GL_INVALID_FRAMEBUFFER_OPERATION for OpenGL functions when the default framebuffer 0 is bound on macOS. These errors happen because the NSOpenGLContext doesn't have a view bound at the time of an OpenGL call. The view could not be set for the NSOpenGLContext because the window or view can still be invisible. Setting an invisible view for the NSOpenGLContext can generate invalid drawable messages as 682ed910 has shown. Thus setting the view could be deferred because the NSOpenGLContext needs a visible view. However, right after the window becomes visible, an OpenGL function that involves the default framebuffer can be called. And when the view is still not set at the time of the OpenGL call, errors such as GL_FRAMEBUFFER_UNDEFINED and GL_INVALID_FRAMEBUFFER_OPERATION could happen and result in rendering errors such as black screen. So we need to set the view for the NSOpenGLContext as soon as the view becomes visible. It's possible that the window and view are still invisible when OpenGL functions involving the default framebuffer get called. In such cases, I think errors like GL_FRAMEBUFFER_UNDEFINED are justified on macOS. Fix Active Trader Pro black screen at launch on macOS. The application creates a d3d9 device with an invisible window. And then it shows the window before calling d3d9_swapchain_Present(). So all the [NSOpenGLContext setView] opportunities are missed and no view is set. Then when glBlitFramebuffer() gets called, GL_FRAMEBUFFER_UNDEFINED happens and so nothing is rendered. The application only renders one frame before login so the window remains black. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/5487

3 3

[PATCH v4 0/7] MR5482: winevulkan: Move swapchain wrappers out of the user drivers.
by Rémi Bernon 28 May '24

28 May '24

This introduces a custom `VkPresentSurfacesInfoWine` struct to the driver queue present. It could very well be a separate vkQueuePresent parameter, now that the driver interface is manually written. For now it provides the swapchains HWND, for vulkan_surface_presented, which is mostly just for winewayland purposes. Later I intend to change it to pass win32u surfaces wrappers. -- v4: winevulkan: Remove now unnecessary vkCreateSwapchainKHR driver entry. winevulkan: Remove now unnecessary vkDestroySwapchain driver entry. win32u: Move vkQueuePresent implementation out of the user drivers. winewayland: Remove now unnecessary swapchain wrapper. winewayland: Remove now unnecessary swapchain extents checks. win32u: Introduce a new vulkan_surface_presented driver entry. winevulkan: Pass surface info for each vkQueuePresent swapchain to win32u. https://gitlab.winehq.org/wine/wine/-/merge_requests/5482

3 10

[PATCH 0/7] MR5482: winevulkan: Move swapchain wrappers out of the user drivers.
by Rémi Bernon 28 May '24

28 May '24

This introduces a custom `VkPresentSurfacesInfoWine` struct to the driver queue present. It could very well be a separate vkQueuePresent parameter, now that the driver interface is manually written. For now it provides the swapchains HWND, for vulkan_surface_presented, which is mostly just for winewayland purposes. Later I intend to change it to pass win32u surfaces wrappers. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/5482

1 9