December 2024 - Wine-GitLab

[PATCH v5 0/3] MR6904: ntdll.dll: Update NtQueryDirectoryFile to align with current Windows behaviour
by Eugene McArdle (＠eugenemcardle) Dec. 6, 2024

Dec. 6, 2024

This PR updates the behaviour of `NtQueryDirectoryFile`, bringing it in line with current Windows behaviour. The need for this update was discovered when attempting to build the Unreal Engine with MSVC under Wine. In certain cases conditional include statements do not behave as expected, due to MSVC depending on undocumented behaviour of `NtQueryDirectoryFile`. We ran tests on multiple versions of Windows, and discovered that the behaviour has changed since the original Wine implementation, but the documentation has not. The source code for our test tool, and a set of results can be found [here](https://github.com/TensorWorks/NtQueryDirectoryFile-Test). As of Windows 8, calling `NtQueryDirectoryFile` with a re-used handle, a new mask, and setting the `RestartScan` flag to True, causes the cached results to be erased and a new scan to be performed with the updated mask. Currently, Wine performs as did earlier versions of Windows, where the changed mask is ignored, and the cache is reused. This can cause `NtQueryDirectoryFile` under Wine to falsely report that files exist, when they do not. This PR corrects this behaviour, invalidating the cache when required. Implementing this exposed further undocumented behaviour of `NtQueryDirectoryFile`, where a search for a non-existent file will return either `STATUS_NO_MORE_FILES` or `STATUS_NO_SUCH_FILE`, depending on whether or not the handle had been previously used regardless of the value of `RestartScan`. This was reflected in a `winetest` which allowed for the response to be either `STATUS_SUCCESS` or `STATUS_NO_MORE_FILES`. This test has been updated to only allow current Windows behaviour, and `NtQueryDirectoryFile` will return either `STATUS_NO_MORE_FILES` or `STATUS_NO_SUCH_FILE` as appropriate. This patch also adds unit tests for the new behaviour of `NtQueryDirectoryFile`. These tests pass when running `winetest` under Windows, and under Wine with these changes in place, but they will fail under older versions of Wine. -- v5: ntdll: Add tests to confirm filemasks are tied to handles and not directories https://gitlab.winehq.org/wine/wine/-/merge_requests/6904

4 6

[PATCH 0/1] MR6976: ucrtbase: Always use CP_ACP when converting environment block.
by Piotr Caban (＠piotr) Dec. 6, 2024

Dec. 6, 2024

-- https://gitlab.winehq.org/wine/wine/-/merge_requests/6976

2 1

[PATCH 0/1] MR6975: winedbg: Add support for dynamic debug channel.
by eric pouech (＠epo) Dec. 6, 2024

Dec. 6, 2024

Updating also strategy to get back to Wine's ntdll settings. Signed-off-by: Eric Pouech <epouech(a)codeweavers.com> -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6975

3 2

[PATCH v3 0/5] MR6936: winebth.sys part 2: Add support for Bluetooth radio properties.
by Vibhav Pant (＠vibhavp) Dec. 6, 2024

Dec. 6, 2024

This MR adds support for getting a Bluetooth adapter's properties from its corresponding `org.bluez.Adapter1` object, and making them available to userspace via device properties and the `IOCTL_BTH_GET_LOCAL_INFO` ioctl, updating these properties whenever a `PropertiesChanged` signal is received for the adapter. It also adds code for creating and removing radio PDOs on receiving `InterafacesAdded` and `InterfacesRemoved` signals from BlueZ, respectively. -- v3: winebth.sys: Implement IOCTL_BTH_GET_LOCAL_INFO. winebth.sys: Update radio PDO properties on receiving PropertiesChanged for an org.bluez.Adapter1 object. winebth.sys: Remove the corresponding radio PDO on receiving InterfacesRemoved for a org.bluez.Adapter1 object. winebth.sys: Create new radio PDOs on receiving InterfacesAdded for objects that implement org.bluez.Adapter1. winebth.sys: Set radio PDO properties from the device's corresponding org.bluez.Adapter1 object properties. https://gitlab.winehq.org/wine/wine/-/merge_requests/6936

5 15

Re: [PATCH v8 0/1] MR4174: shell32: Add Internet Explorer toolbar icons.
by Jacek Caban (＠jacek) Dec. 6, 2024

Dec. 6, 2024

Sorry for the delay, I'm not sure what to think about it. If it's dropped in modern Windows, it's not clear if we want it. I guess it wouldn't hurt... We usually add SVGs so that maintainer mode can generate binary blobs. It seems that Tango theme repo contains SVGs, could we use them? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/4174#note_89955

1 0

Re: [PATCH v2 0/2] MR6970: jscript: Add support for deleting host properties.
by Jacek Caban (＠jacek) Dec. 6, 2024

Dec. 6, 2024

On Fri Dec 6 10:42:56 2024 +0000, Gabriel Ivăncescu wrote: > This looks fine for now, pretty neat way to fix it. > But I still don't think we can escape a revamp at some point. The 2 > "largest" volatile objects I tested before were the storage, and > doc/window, and they function differently. On doc and window, the > volatile props (stuff like accessing DOM elements via props) are looked > up after the prototypes, while on storage, they are looked up first, > before even the prop on the object. > For example: > - localStorage.getItem("test") returns null (no volatile item). > - Object.defineProperty(localStorage, "test", ...) creates normal prop > on it. > - localStorage.setItem("test", "blah") now it is `blah` (only way to > retrieve the defined prop underneat is getOwnPropertyDescriptor) > - localStorage.removeItem("test") now reveals the original defined prop > Worse is that it also hijacks deletion, so delete acts like removeItem, > but if you call delete now, it won't delete the defineProperty prop > (since there's no setItem so nothing to remove). > For doc and window it's the opposite and if anything in the prototype > chain has the name of an element id, it will be returned instead. Thanks. This intentionally doesn't touch `PROP_EXTERN`, as deleting those requires interaction with the host object. For all jscript knows, the host may choose to ignore the delete. Fixing accessor and function properties should cover everything prototypes currently need, so those should be in good shape for the code freeze. It could be useful to allow external properties to opt into this delete mechanism for cases where host-owned properties should always behave like regular ones. Examples include `constructor`, `prototype`, and possibly constants like `Node.TEXT_NODE`. For cases like document and window, we might introduce a flag in `property_info` to modify name lookup. Jscript could continue searching the prototype chain and use returned info only if the search fails. Storage properties seem more complicated. None of the solutions that come to mind feel ideal, but this is a very niche corner case. I doubt any reasonable code depends on it, so it seems safe to ignore for now. It also doesn't seem like something Proton needs to worry about. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6970#note_89951

1 0

[PATCH 0/2] MR6974: winex11: Listen to root window _NET_SUPPORTED property changes.
by Rémi Bernon Dec. 6, 2024

Dec. 6, 2024

This might happen if the window manager is restarted or changed while a Wine window is opened. We should support that use case properly. Split from https://gitlab.winehq.org/wine/wine/-/merge_requests/6944 -- https://gitlab.winehq.org/wine/wine/-/merge_requests/6974

1 2

[PATCH v9 0/8] MR6800: mf: Handle unexpected media source shutdown in a media session.
by Conor McCarthy (＠cmccarthy) Dec. 6, 2024

Dec. 6, 2024

-- v9: winegstreamer: Send media source shutdown notification via IMFMediaShutdownNotify. mf: Introduce IMFMediaShutdownNotify for notification of media source shutdown. https://gitlab.winehq.org/wine/wine/-/merge_requests/6800

4 11

[PATCH v12 0/7] MR4826: Mach vm operations improvements
by Marc-Aurel Zent (＠mzent) Dec. 5, 2024

Dec. 5, 2024

For all mach vm operations this removes the task suspend and resume, which are not needed. This uses `mach_vm_read_overwrite` to read into a caller-specified buffer, saving the `mach_vm_deallocate` call (bringing all read operations down to 1 syscall from 4). The only alignment restriction on `mach_vm_write` according to the original CMU documentation is that data is > [pointer to page aligned in array of bytes] An array of data to be written. (In practice it also works with arbitrary addresses on macOS, but it probably doesn't hurt to follow the original specifications here). The only other reference that these read/writes should be page-aligned is from the GNU Hurd documentation > The current implementation requires that address, data and data_count all be page-aligned. Otherwise, KERN_INVALID_ARGUMENT is returned. which I assume was the reason why this was originally done (plus it sounds to me like they will fix that in the future and 4fe19777 already broke GNU Hurd support anyways, if that was supposed to be working). Also this includes the missing mach part of 5b1f3b14, which was only applied to the ptrace backend, and together with the `write_process_memory` rework, this gets rid of all fixmes in mach.c -- v12: server: Work around macOS W^X limitations in write_process_memory. https://gitlab.winehq.org/wine/wine/-/merge_requests/4826

2 9

[PATCH v11 0/7] MR4826: Mach vm operations improvements
by Marc-Aurel Zent (＠mzent) Dec. 5, 2024

Dec. 5, 2024

For all mach vm operations this removes the task suspend and resume, which are not needed. This uses `mach_vm_read_overwrite` to read into a caller-specified buffer, saving the `mach_vm_deallocate` call (bringing all read operations down to 1 syscall from 4). The only alignment restriction on `mach_vm_write` according to the original CMU documentation is that data is > [pointer to page aligned in array of bytes] An array of data to be written. (In practice it also works with arbitrary addresses on macOS, but it probably doesn't hurt to follow the original specifications here). The only other reference that these read/writes should be page-aligned is from the GNU Hurd documentation > The current implementation requires that address, data and data_count all be page-aligned. Otherwise, KERN_INVALID_ARGUMENT is returned. which I assume was the reason why this was originally done (plus it sounds to me like they will fix that in the future and 4fe19777 already broke GNU Hurd support anyways, if that was supposed to be working). Also this includes the missing mach part of 5b1f3b14, which was only applied to the ptrace backend, and together with the `write_process_memory` rework, this gets rid of all fixmes in mach.c -- v11: server: Work around macOS W^X limitations in write_process_memory. server: Do not page-align address in write_process_memory. server: Do not suspend mach task in write_process_memory. server: Use mach_vm_read_overwrite in get_selector_entry. server: Do not suspend mach task in get_selector_entry. server: Use mach_vm_read_overwrite in read_process_memory. server: Do not suspend mach task in read_process_memory. https://gitlab.winehq.org/wine/wine/-/merge_requests/4826

3 12