On Mon Jul 21 18:34:10 2025 +0000, Nikolay Sivov wrote:
> Windows SDK always defines it to either 5, or 6 for XP and newer.
> Nothing in headers depends on it. So it's not the same as with d3dx.
But it is the same as d3dx. Windows headers don't check D3DX_SDK_VERSION and conditionally define things the same way we do. They just ship multiple versions of the same file with different contents and different values of D3DX_SDK_VERSION.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/8595#note_110541
--
v2: server: Set IOSB for cancelled asyncs in async_terminate().
ntdll: Make wait in wait_async() always non-alertable.
ntdll: Factor out test_alert_with_status().
ntdll/tests: Add more tests for canceling synchronous IO.
https://gitlab.winehq.org/wine/wine/-/merge_requests/8592
On Mon Jul 21 18:05:49 2025 +0000, Elizabeth Figura wrote:
> Wait, why, though? This is the approach we use for e.g. d3dx.
Windows SDK always defines it to either 5, or 6 for XP and newer. Nothing in headers depends on it. So it's not the same as with d3dx.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/8595#note_110536
On Mon Jul 21 06:09:39 2025 +0000, Zhiyi Zhang wrote:
> changed this line in [version 6 of the diff](/wine/wine/-/merge_requests/8595/diffs?diff_id=194446&start_sha=4ae32c680b3c52bdfa972a955a18abeeb0532ec1#515e569bebc6ce6f3ec15b1e7bd470691646060a_4_4)
Wait, why, though? This is the approach we use for e.g. d3dx.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/8595#note_110534
If bluetooth_radio_set_properties gets called before the PnP manager has enumerated the radio device, IoSetDevicePropertyData will fail as the registry entries for the radio's instance ID have not been created yet. This would result in the following error messages whenever a new prefix was created:
```
011c:err:plugplay:IoSetDevicePropertyData Failed to open device, error 0xe000020b.
011c:err:plugplay:IoSetDevicePropertyData Failed to open device, error 0xe000020b.
011c:err:plugplay:IoSetDevicePropertyData Failed to open device, error 0xe000020b.
011c:err:plugplay:IoSetDevicePropertyData Failed to open device, error 0xe000020b.
```
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/8611
--
v2: bluetoothapis/tests: Implement tests for BluetoothGATTGetCharacteristics.
bluetoothapis: Implement BluetoothGATTGetCharacteristics.
winebth.sys: Remove GATT characteristic entries when they are removed from the Unix Bluetooth service.
winebth.sys: Implement IOCTL_WINEBTH_LE_DEVICE_GET_GATT_CHARACTERISTICS.
https://gitlab.winehq.org/wine/wine/-/merge_requests/8556
Loïc Rebmeister (@Fox2Code) commented about server/fd.c:
>
> if (fd->unix_fd == -1)
> {
> + if (stat( name, &st ))
> + {
> + file_set_error();
> + goto error;
> + }
I think the `stat` call should be moved after the "check for trailing slash on file path" code.
As if the stat check is done here, the "check for trailing slash on file path" code path would no longer be called for `C:\NonExistingFolder\`.
Everything else seems fine to me, I took my time reading to make sure I fully understood the code.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/8578#note_110527
Also, add todo'd conformance tests for converting to and from `VT_ARRAY | VT_UI1` values.
--
v2: propsys/tests: Add test for PropVariantToGUID with VT_ARRAY | VT_UI1 values.
https://gitlab.winehq.org/wine/wine/-/merge_requests/8604