- Wine-GitLab - list.winehq.org

Re: [PATCH v2 0/1] MR9279: msxml3: Fix use-after-free of dom attributes.
by Nikolay Sivov (＠nsivov) April 9, 2026

April 9, 2026

I'm going to close this one, fix like that is no longer applicable after !10599, and is also not needed since all nodes are tracked the same way now. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/9279#note_135634

1 0

[PATCH 0/4] MR10608: msxml3/tests: Add some more appendChild() tests with fragments.
by Nikolay Sivov (＠nsivov) April 9, 2026

April 9, 2026

Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10608

2 4

[PATCH 0/4] MR10607: winhttp: Stub WinHttpQueryOption(WINHTTP_OPTION_CONNECT_RETRIES).
by Hans Leidekker (＠hans) April 9, 2026

April 9, 2026

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/10607

2 4

Re: [PATCH v2 0/7] MR10582: opengl32: Misc simplifications to make_opengl and tests.
by Rémi Bernon (＠rbernon) April 9, 2026

April 9, 2026

Anything wrong with this? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10582#note_135630

1 0

[PATCH v16 0/1] MR10464: vbscript: Return "Name redefined" error for duplicate declarations.
by Francis De Brabandere (＠francisdb) April 9, 2026

April 9, 2026

When a name is defined twice (Const/Const, Dim/Const, Const/Dim, Dim/Dim, Sub vs Const/Dim, Class vs Const/Dim/Function, duplicate class members), Windows returns VBScript error 1041 ("Name redefined"). Wine was returning E_FAIL. Store source locations in const_decl_t, dim_decl_t, function_decl_t, and class_decl_t so errors point to the duplicate name. Also fix duplicate Sub/Function/Property/Dim detection inside classes, including the case where a Property collides with an existing Sub/Function. -- v16: vbscript: Return "Name redefined" error for duplicate declarations. https://gitlab.winehq.org/wine/wine/-/merge_requests/10464

2 1

[PATCH v15 0/1] MR10464: vbscript: Return "Name redefined" error for duplicate declarations.
by Francis De Brabandere (＠francisdb) April 9, 2026

April 9, 2026

When a name is defined twice (Const/Const, Dim/Const, Const/Dim, Dim/Dim, Sub vs Const/Dim, Class vs Const/Dim/Function, duplicate class members), Windows returns VBScript error 1041 ("Name redefined"). Wine was returning E_FAIL. Store source locations in const_decl_t, dim_decl_t, function_decl_t, and class_decl_t so errors point to the duplicate name. Also fix duplicate Sub/Function/Property/Dim detection inside classes, including the case where a Property collides with an existing Sub/Function. -- v15: vbscript: Return "Name redefined" error for duplicate declarations. https://gitlab.winehq.org/wine/wine/-/merge_requests/10464

2 1

[PATCH v14 0/1] MR10464: vbscript: Return "Name redefined" error for duplicate declarations.
by Francis De Brabandere (＠francisdb) April 9, 2026

April 9, 2026

When a name is defined twice (Const/Const, Dim/Const, Const/Dim, Dim/Dim, Sub vs Const/Dim, Class vs Const/Dim/Function, duplicate class members), Windows returns VBScript error 1041 ("Name redefined"). Wine was returning E_FAIL. Store source locations in const_decl_t, dim_decl_t, function_decl_t, and class_decl_t so errors point to the duplicate name. Also fix duplicate Sub/Function/Property/Dim detection inside classes, including the case where a Property collides with an existing Sub/Function. -- v14: vbscript: Return "Name redefined" error for duplicate declarations. https://gitlab.winehq.org/wine/wine/-/merge_requests/10464

2 1

[PATCH v10 0/11] MR9628: Implement pointer input APIs for multi-touch support
by Hecheng Yu (＠kde-yyds) April 9, 2026

April 9, 2026

These patches implemeted core pointer input APIs required by modern applications with touchscreen support, particularly Chromium and Qt-based applications. Implemented APIs: - GetPointerType: Query pointer input type (touch, pen, mouse) - GetPointerTouchInfo: Retrieve detailed touch pointer information - GetPointerFrameTouchInfo: Get all active touch pointers in current frame Added stubs: - SkipPointerFrameMessages: Used by applications to skip pointer frame processing like Qt - GetPointerFrameTouchInfoHistory: Historical touch frame data (not yet implemented) - GetPointerPenInfoHistory: Historical pen data (not yet implemented) System metrics changes: - SM_DIGITIZER: Reports touch capabilities (NID_INTEGRATED_TOUCH | NID_MULTI_INPUT | NID_READY) - SM_MAXIMUMTOUCHES: Returns maximum simultaneous touch points (10) Limitations: - We explictly set pointer type to PT_TOUCH when storing pointer info to queue, so GetPointerType will always return PT_TOUCH - GetPointerFrameTouchInfo returns 0 pointers after finger lift, while Windows returns 1 pointer with UP flags. And SkipPointerFrameMessages is currently a stub returning TRUE. Qt's pointer message processing uses both of them, but it's just working well Application requirements: - Chromium: Requires GetPointerType and GetPointerTouchInfo for touch input - Qt applications: Need GetPointerFrameTouchInfo and proper system metrics for multi-touch support - Qt5: Additionally checks for presence of all pointer APIs at runtime, requiring stubs to enable pointer messages processing -- v10: user32: Implement GetPointerFrameTouchInfo win32u: Implement NtUserGetPointerFrameTouchInfo user32: Implement GetPointerTouchInfo win32u: Implement NtUserGetPointerTouchInfo user32: Implement GetPointerType win32u: Implement NtUserGetPointerType win32u: Add pointer information cache and storage user32: Stub GetPointerPenInfoHistory user32: Stub GetPointerFrameTouchInfoHistory user32: Stub SkipPointerFrameMessages win32u: Add SM_DIGITIZER and SM_MAXIMUMTOUCHES system metrics https://gitlab.winehq.org/wine/wine/-/merge_requests/9628

4 15

Re: [PATCH v2 0/3] MR10298: mfreadwrite: Implement sample encoding.
by Ziqing Hui (＠zhui) April 9, 2026

April 9, 2026

@nsivov This needs reviewing, thanks. :p -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10298#note_135608

1 0

Re: [PATCH v2 0/4] MR10604: setupapi, ntoskrnl: Implement CM_Get_Parent with device tree properties
by Jan Robert Gerigk (＠RgSg86) April 9, 2026

April 9, 2026

On Thu Apr 9 01:29:48 2026 +0000, Elizabeth Figura wrote: > It should be possible to test a lot more of this. 1/4 can be tested in > setupapi. 2/4 can be tested in ntoskrnl, not just the CM APIs but also > retrieving the devpkeys directly. > From 2/4: > ``` > > + /* Store parent device relationship for CM_Get_Parent. */ > ``` > This comment is unnecessary. > ``` > + if (parent_device) > ``` > Why are we checking this? When can it be NULL? > ``` > + { > + WCHAR parent_id[MAX_DEVICE_ID_LEN]; > + if (!get_device_instance_id( parent_device, parent_id )) > + { > + if (!SetupDiSetDevicePropertyW( set, &sp_device, > &DEVPKEY_Device_Parent, DEVPROP_TYPE_STRING, > + (BYTE *)parent_id, (lstrlenW( parent_id ) + 1) > * sizeof(WCHAR), 0 )) > ``` > wcslen() > ``` > + WARN("Failed to set parent device property.\n"); > ``` > Precedent notwithstanding, I would make this an ERR since it should > never legitimately happen. Or just delete it. > ``` > + DEVICE_OBJECT *original_parent = parent; > ``` > FDOs are supposed to pass through IRP_MN_QUERY_ID. You'll note that > get_device_id() calls IoGetAttachedDevice() anyway, so this isn't > actually accomplishing anything. > ``` > + /* Store DEVPKEY_Device_Children on the parent and > DEVPKEY_Device_Siblings on each child. */ > ``` > That's more than what the commit message says. Mentioning it in the > commit message would be an improvement, but splitting the > DEVPKEY_Device_Children and DEVPKEY_Device_Siblings parts to their own > commits would be even better. > As elsewhere the comment isn't really adding anything. > ``` > + if (relations->Count > 0) > ``` > That's unnecessary; the for loop takes care of it. > ``` > + if (!get_device_instance_id( relations->Objects[i], > child_ids[valid_count] )) > + { > + multi_len += lstrlenW( child_ids[valid_count] ) + 1; > + valid_count++; > + if (valid_count >= 64) break; > + } > ``` > Error handling here doesn't really make sense. We don't expect > get_device_instance_id() to fail other than memory exhaustion, and if it > does there's no point proceeding. > As elsewhere I would avoid lstr*() functions in favor of simple wcs*(). > ``` > + children_multi = HeapAlloc( GetProcessHeap(), > HEAP_ZERO_MEMORY, multi_len * sizeof(WCHAR) ); > + if (children_multi) > ``` > You can use malloc(), which is simpler. I also wouldn't bother with > checking for memory exhaustion for small allocations in new code. > ``` > + /* Set DEVPKEY_Device_Siblings on each child (all other > children). */ > + for (i = 0; i < valid_count; ++i) > ``` > You can move this out of the "if (valid_count > 0)" block and save a > level of indentation. > 3/4 should be implemented in cfgmgr32 on top of cfgmgr32 APIs, and > setupapi should forward to it. We're in the middle of moving that right now. > 4/4: > ``` > + ok(ret == CR_SUCCESS || ret == CR_FAILURE || ret == CR_NO_SUCH_DEVNODE > + || ret == CR_INVALID_DEVNODE, > ``` > When can all of these happen? Same with the check later. > ``` > + ok(ret == CR_SUCCESS, "CM_Get_Device_IDA on parent: got > %#lx\n", ret); > + if (ret == CR_SUCCESS) > + trace("Parent of ROOT\\LEGACY_BOGUS\\0000: %s\n", buffer); > ``` > You don't need that if(). What's the parent ID? Can it be checked? > ``` > + SetupDiRemoveDevice(set, &device); > ``` > I would check that this succeeds. Great detailed feedback, thanks! That is a lot to work through, I think it will take me a while. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/10604#note_135606

1 0