December 2024 - wine-gitlab

[PATCH 0/1] MR7075: comctl32/tests: Add missing broken condition.
by Bernhard Übelacker (＠bernhardu) 01 Jan '25

01 Jan '25

Followup on 82f7c0cc3b. Seems above commit introduced a failure with [32-bit Windows 7](https://test.winehq.org/data/tests/comctl32:tab.html). -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7075

4 3

[PATCH 0/1] MR7076: d2d1/tests: Remove a duplicate test.
by Zhiyi Zhang (＠zhiyi) 01 Jan '25

01 Jan '25

test_effect_custom_pixel_shader is already queued two lines before. The duplicate was added by eb7199cb. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7076

3 2

[PATCH 0/1] MR7077: mscoree: load libmono as a system library
by Zephyr Lykos (＠self) 01 Jan '25

01 Jan '25

This allows libmono to load extra libraries (notably win-iconv) from system_dllpath. For example, Fedora's wine-mono is built against a shared win-iconv installed in system_dllpath, but wine will not search there unless it's explicitly instructed to do so since bfbccf1a038e (ntdll: Prevent loading Wine system dependencies in place of identically named application DLLs.). Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=54573 -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7077

4 8

[PATCH 0/1] MR7082: ntdll: Return an error if count is zero in NtRemoveIoCompletionEx.
by Alex Henrie (＠alexhenrie) 01 Jan '25

01 Jan '25

Without this check, the return value is undefined. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7082

4 3

[PATCH 0/2] MR6983: wined3d: Add B4G4R4A4 and B5G5R5A1 support in Vulkan.
by Stefan Dösinger (＠stefan) 31 Dec '24

31 Dec '24

While this is arguably a feature addition, it is a feature that the GL backend had since forever and I'm surprised we got away without it in d3d11 so far. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6983

4 7

[PATCH v3 0/4] MR7056: ddraw: Correctly apply state on multiple devices.
by Paul Gofman (＠gofman) 31 Dec '24

31 Dec '24

-- v3: ddraw/tests: Release device earlier in test_d3d_state_reset(). ddraw: Correctly apply state when multiple devices are used. https://gitlab.winehq.org/wine/wine/-/merge_requests/7056

4 6

[PATCH 0/1] MR7071: dxgi: Support more feature levels in debug_feature_level().
by Zhiyi Zhang (＠zhiyi) 31 Dec '24

31 Dec '24

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/7071

3 2

[PATCH 0/2] MR7079: ddraw: Invalidate the window when the primary surface is released.
by Anton Baskanov (＠baskanov) 31 Dec '24

31 Dec '24

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=57599 -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7079

4 4

Re: [PATCH v6 0/2] MR7060: scrrun/tests: Add drive_get_path test
by Quinn Rafferty (＠ramleejo) 30 Dec '24

30 Dec '24

On Mon Dec 30 17:30:16 2024 +0000, Alex Henrie wrote: > Thanks for the patches. The implementation looks good and it allows the > Windows Search Index Analyzer to display its initial dialog. It's > possible that drive_get_Path should really call GetVolumePathNameW and > strip the trailing backslash, but Wine's implementation of the Drive > class currently only supports letter drives, so a simpler implementation > is fine for now. > I'm not sure that all the tests are necessary. If it were me, I would > just do this: > ```diff > diff --git a/dlls/scrrun/tests/filesystem.c b/dlls/scrrun/tests/filesystem.c > index e53bd092ce9..aabec74b3f9 100644 > --- a/dlls/scrrun/tests/filesystem.c > +++ b/dlls/scrrun/tests/filesystem.c > @@ -1471,7 +1471,7 @@ static void test_DriveCollection(void) > while (IEnumVARIANT_Next(enumvar, 1, &var, &fetched) == S_OK) { > IDrive *drive = (IDrive*)V_DISPATCH(&var); > DriveTypeConst type; > - BSTR str; > + BSTR str, path; > > hr = IDrive_get_DriveType(drive, &type); > ok(hr == S_OK, "Unexpected hr %#lx.\n", hr); > @@ -1482,7 +1482,17 @@ static void test_DriveCollection(void) > hr = IDrive_get_DriveLetter(drive, &str); > ok(hr == S_OK, "Unexpected hr %#lx.\n", hr); > ok(SysStringLen(str) == 1, "got string %s\n", wine_dbgstr_w(str)); > + > + hr = IDrive_get_Path(drive, NULL); > + ok(hr == E_POINTER, "Unexpected hr %#lx.\n", hr); > + > + hr = IDrive_get_Path(drive, &path); > + ok(hr == S_OK, "Unexpected hr %#lx.\n", hr); > + ok(SysStringLen(path) == 2 && path[0] == str[0] && path[1] == ':', > + "got string %s\n", wine_dbgstr_w(path)); > + > SysFreeString(str); > + SysFreeString(path); > > hr = IDrive_get_IsReady(drive, NULL); > ok(hr == E_POINTER, "Unexpected hr %#lx.\n", hr); > ``` > However, when submitting tests I am often told that I have added either > too few tests or too many tests, and I'm not sure what the preferred > amount of tests is here. Thanks for the feedback; I'm happy I could help fix the issue for you. I had thought about adding my test to a previous test like in your example, but I wasn't sure what would be considered best practice so I went with the safe route. I'm open to changing the tests to what you suggested, it's much more concise than my solution. Also, thanks for adding this to the Bugzilla report. I had meant to but had forgotten to as I was sick all last week. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7060#note_91195

1 0

Re: [PATCH v6 0/2] MR7060: scrrun/tests: Add drive_get_path test
by Alex Henrie (＠alexhenrie) 29 Dec '24

29 Dec '24

Thanks for the patches. The implementation looks good and it allows the Windows Search Index Analyzer to display its initial dialog. It's possible that drive_get_Path should really call GetVolumePathNameW and strip the trailing backslash, but Wine's implementation of the Drive class currently only supports letter drives, so a simpler implementation is fine for now. I'm not sure that all the tests are necessary. If it were me, I would just do this: ```diff diff --git a/dlls/scrrun/tests/filesystem.c b/dlls/scrrun/tests/filesystem.c index e53bd092ce9..aabec74b3f9 100644 --- a/dlls/scrrun/tests/filesystem.c +++ b/dlls/scrrun/tests/filesystem.c @@ -1471,7 +1471,7 @@ static void test_DriveCollection(void) while (IEnumVARIANT_Next(enumvar, 1, &var, &fetched) == S_OK) { IDrive *drive = (IDrive*)V_DISPATCH(&var); DriveTypeConst type; - BSTR str; + BSTR str, path; hr = IDrive_get_DriveType(drive, &type); ok(hr == S_OK, "Unexpected hr %#lx.\n", hr); @@ -1482,7 +1482,17 @@ static void test_DriveCollection(void) hr = IDrive_get_DriveLetter(drive, &str); ok(hr == S_OK, "Unexpected hr %#lx.\n", hr); ok(SysStringLen(str) == 1, "got string %s\n", wine_dbgstr_w(str)); + + hr = IDrive_get_Path(drive, NULL); + ok(hr == E_POINTER, "Unexpected hr %#lx.\n", hr); + + hr = IDrive_get_Path(drive, &path); + ok(hr == S_OK, "Unexpected hr %#lx.\n", hr); + ok(SysStringLen(path) == 2 && path[0] == str[0] && path[1] == ':', + "got string %s\n", wine_dbgstr_w(path)); + SysFreeString(str); + SysFreeString(path); hr = IDrive_get_IsReady(drive, NULL); ok(hr == E_POINTER, "Unexpected hr %#lx.\n", hr); ``` However, when submitting tests I am often told that I have added either too few tests or too many tests, and I'm not sure what the preferred amount of tests is here. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/7060#note_91192

2 1

2025

2024

2023

2022

wine-gitlab December 2024