`WINE_UNICODE_NATIVE` handles non-PE targets where appropriate, but on PE targets, we should always use `wchar_t`. This is particularly important for C++, where `wchar_t` is a distinct built-in type.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/7488
On Tue Mar 4 12:15:46 2025 +0000, Bernhard Übelacker wrote:
> Sorry, this was not my intention. This line should make sure
> `check_PropVariantToBSTR2` is never called with anything else than
> `VT_R4`, but I have not considered this to be inside a `todo_wine` altogether.
> I created !7487 to fix this.
No worries, thanks for the fix :)
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/7460#note_96759
On Tue Mar 4 12:15:46 2025 +0000, Rémi Bernon wrote:
> This now causes a "Test succeeded inside todo block" test failure on linux.
Sorry, this was not my intention. This line should make sure `check_PropVariantToBSTR2` is never called with anything else than `VT_R4`, but I have not considered this to be inside a `todo_wine` altogether.
I created !7487 to fix this.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/7460#note_96755
Cannot use always succeeding ok statement,
because check_PropVariantToBSTR2 is used inside a todo_wine block.
Followup to a1637b167f.
This was intended to make sure check_PropVariantToBSTR2 is just called with `VT_R4`, but makes now more problems than good.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/7487
This is an adjustment of 7324.
d3d9:visual already has a test that's relatively close to what a simplified
version of the tests in 7324 looked like, so I made the few changes to expand
that test to match and then ported it to ddraw.
Implementation-wise, this removes the SD/HD difference (which only exists on
NVidia and is not necessary to improve the mentioned application).
It also removes the clamping of YUV values, which as the tests show is not
correct.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/7416