- [PSDK] Introduce our own Microsoft compatible and MIT-licensed shlwapi.h and fix breaking changes
- [INCLUDE/WINE] Import shlwapi.h from wine-10.0 for use with Wine compatible code
- [CMAKELISTS.TXT] Bump up IE version from IE60 to IE60SP2. This is required because we are already using IE60SP2 guarded functionality; the new header unhides this.
- [SHLWAPI] Fix breaking changes, including a typo for 'SHGVSPB_GLOBALDEAFAULTS' (should be 'SHGVSPB_GLOBALDEFAULTS')
- [MSI] Use wine's shlwapi.h instead of our own in automation.c, don't precompile shlwapi.h.
- [URLMON_WINETEST] Primarily use wineheaders for this test
- [IPHLPAPI_WINETEST] Sync to wine-10.0
- [IPHLPAPI] Add stubs for Vista+ functions so these functions can link properly when DLL_EXPORT_VERSION >= 0x600
- [kmtest_drv] Increase timeout period from 10 seconds to 30 seconds. This was needed on my machine because of log spam in the kmtest:IoDeviceInterface test. I also reduced the log spam in that test, making this not necessary. However, I left it in as I figured if I ran into this issue, others may too in the future with other tests.
- [kmtest:ExFastMutex] Skip concurrent fast mutex tests on Vista+. It appears that this was also causing hangs on WS03 x64 so I moved the skip logic there and combined it with the vista+ skip.
- [kmtest:IoDeviceInterface] Change some traces to DPRINTs since it was creating excessive log spam that slowed down the test enough to trigger the previous timeout.
- [kmtest:KeEvent] Skip concurrent event tests on Vista+.
- [kmtest:KeGuardedMutex] Skip on Vista+.
- [kmtest:ObType] Disable making custom object types as this is invalid on Vista+.
- [kmtest:PoIrp] Skip on Vista+ (already skipped on WS03 x64).
- [kmtest:SeInheritance] Skip on Vista+.
- [KMTEST:IOFILESYSTEM] Add support for NTFS
- [KMTEST:IODEVICEOBJECT] Work around a Windows 7 bug, account for struct size changes on x64, accept Vista+ behavior
- [KMTEST:EXPOOLS] Accept pool being promoted to NonPaged on Vista+, accept 2GB allocation on x64, guard an undocumented behavior
- [KMTEST:CMSECURITY] Add tests for Vista-Win8.1
[COM_APITEST]
- Add NTDDI version gating to classes and interfaces in tests
- Remove rigid offset tests. Instead, offsets to known interfaces can be logged by defining LOG_COM_INTERFACE_OFFSETS.
- Replace the noreg member of the known interfaces array with a function pointer that can be used to determine if a com test should be globally registered depending on the current NTDDI version. I could probably replace this with a min and max registered NTDDI value, but I thought this was confusing for interfaces that are never globally registered.
[INCLUDE/REACTOS]
- Rename the existing interface definition for IShellBrowserService to IShellBrowserService_XP at {1307EE17-EA83-49EB-96B2-3A28E2D7048A}. This only exists on XP and Server 2003, no earlier or later versions.
- Rename IID_IShellBrowserServce to IID_IShellBrowserService since this definition is correct for Windows 7+ at {DFBC7E30-F9E5-455F-88F8-FA98C1E494CA}
- Add an interface definition for IShellBrowserService4 at {F5A24314-5B8B-44FA-BC2E-31285544B520}. This definition for IShellBrowserService exists for Vista only and no other version.
[PSDK]
- Add a define for NTDDI_WIN7SP1. While this isn't defined in the public Windows headers, it clearly exists in private headers because it is referenced in the public WinBase.h header.
- [INCLUDE/WINE] Add synced Wine headers for the iphlpapi winetest to use. This helps us slowly decouple our headers and wine's headers.
- [PSDK] Add definitions we are missing for other winetest syncs.
- [WINETESTS][APITESTS] Fix breaking changes, remove duplicate definitions for debugstr_guid.
- Add a tristate option to accept either true or false for the bStdOutput and bStdError tests
- Use this new option on the four tests that are inconsistent.
This is more explicit than defining it for all wine tests and it directly shows us which tests need to be synced to get rid of the hack.
Addendum to 381e224 (PR #7576)
The LOGICAL (32-bit boolean) type is already defined for kernel-mode
in the XDK/DDK ntdef.h header. Add the type in the NDK for usage in
user-mode NT code also.
The definition present in kmt_platform.h can now be removed, since
it can now use the one from the NDK.
- Introduce GetNTDDIVersion() to get NTDDI version at runtime for apitests.
- Introduce GetMajorNTVersion() and GetMinorNTVersion() macros and use them for GetNTVersion()
- Introduce GENERATE_NTDDI() macro to construct NTDDI values from major, minor, service pack, and subversion values.
- Introduce NTDDI_MIN and NTDDI_MAX macros for tests that use NTDDI ranges.
Note: the build number in KUSER_SHARED_DATA is only populated on Windows 10 RTM and later. On older Windows versions, this field is reserved and is usually 0. For this reason, I chose not to expose it as a macro.
Historical note:
Investigation shows that this functionality, introduced between builds
1902 and 1906 of Windows NT 5.0 (future 2000) Beta 3, has always been
"nop-ed" and has remained this way till Windows Server 2003.
The value read from the "SafeMode" registry value is unconditionally
overridden afterwards, causing the notification DLLs to always be loaded.
In ReactOS, this functionality is restored, and only the notifications
tagged as such are loaded in SafeMode.
Furthermore:
Analysis of strings in Win2000 and WinXP/2003 winlogon.exe, show that
the "Safe" registry value doesn't exist for notifications; instead,
it is named "SafeMode".
The "Safe" value appears only for the SensLogn (SENS Winlogon Event) handler
registry entry. My hypothesis is that the value name is a typo for the
"SafeMode" value. It has been introduced in the `\Winlogon\Notify\SensLogn`
registry entry for SensLogn around Windows NT 5.0 build 1946.
Analysis of strings in Win2000 and WinXP/2003 winlogon.exe, show that
the "Enabled" registry value doesn't exist for notifications.
This value is actually only used by the ScCertProp (Smart Card Certificate
Propagation) notifications, in wlnotify.dll, to enable or disable certificate
progagation.[^1]
Note that whatever the "Enabled" registry value is, the notification DLL
is still loaded within winlogon.exe.
We however keep the `bEnabled` internal flag, so as to be able to disable
at runtime notifications that could not be delay-loaded, or that behave
erratically, etc.
[WLNOTIFY] Add a comment about the "Enabled" value in scard.c!SCardStartCertProp()
[^1]: For more information, see:
https://www.microfocus.com/documentation/securelogin/9.0/administration_guide/blm54qb.html?view=printhttps://www.betaarchive.com/wiki/index.php?title=Microsoft_KB_Archive/925884
Fix the shell32 apitests so that they pass on Windows Server 2003 - Windows 10. Many of these fixes are for Vista+, but the most important fixes are for ShellExecCmdLine and FindExecutable which had issues closing windows after tests and deleting test files. Failing to delete these files breaks the other test (i.e. running ShellExecCmdLine would break FindExecutable and vis-versa.)
- Move sdk\include\reactos\wine to sdk\include\wine
- Reorder the directories in include_directories() to be closer to alphabetical. This should make it easier to determine what global include directories can be removed in the future.
Introduce and use a minimal testing framework (minitest.h) based on
an updated version of `wine/test.h`.
Each notification handler becomes its own test. Useful macro and
function helpers have been introduced to simplify the code that is
duplicated for each handler.
See commit 38d07d3a24 (PR #8234) for the details of how to install
and use the notification dll.
These tests can exercise the notifications in the four cases, where
asynchronous events and user impersonation can be independently
enabled or disabled. To do this, the `Asynchronous` and `Impersonate`
registry `REG_DWORD` values, inside the `WLNotifyTests` subkey of:
`HKEY_LOCAL_MACHINE\Software\Microsoft\Windows NT\CurrentVersion\Winlogon\Notify`
must be set to their intended values[^1].
----
[^1]: https://learn.microsoft.com/en-us/windows/win32/secauthn/registry-entries
----
Current test results:
- When each notification is emitted, the DLL is loaded then unloaded:
```
err:(modules\rostests\win32\winlogon\wlntfytests\wlntfytests.c:1307)
WLNOTIFY(ac.b0): Entering `DllMain`(hInst: 0x10000000, dwReason: 0x1, pReserved: 0x00000000)
...
err:(modules\rostests\win32\winlogon\wlntfytests\wlntfytests.c:1307)
WLNOTIFY(ac.b0): Entering `DllMain`(hInst: 0x10000000, dwReason: 0x0, pReserved: 0x00000000)
```
This doesn't happen on Windows. And indeed, it should _not_ happen,
because otherwise the DLL would loose any of its internal global state
between consecutive notification calls. This currently happens in
ReactOS, where we can observe the following:
```
modules\rostests\win32\winlogon\wlntfytests\wlntfytests.c:788: Test failed: **** WLEventLogon: ERROR: Wrong state NON-INITIALIZED, expected Startup or Logoff
...
err:(modules\rostests\win32\winlogon\wlntfytests\wlntfytests.c:1036) **** WLEventLogon: Changing state NON-INITIALIZED to Logon
```
and similar for every other notification.
- Test results for each notification:
```
WLEventStartup: 30 tests executed (0 marked as todo, 1 failure), 0 skipped.
WLEventLogon: 30 tests executed (0 marked as todo, 4 failures), 2 skipped.
WLEventStartShell: 30 tests executed (0 marked as todo, 4 failures), 2 skipped.
-- Note: missing WLEventPostShell --
WLEventLock: 30 tests executed (0 marked as todo, 4 failures), 2 skipped.
WLEventUnlock: 30 tests executed (0 marked as todo, 4 failures), 2 skipped.
WLEventStartScreenSaver: 30 tests executed (0 marked as todo, 10 failures), 0 skipped.
WLEventStopScreenSaver: 30 tests executed (0 marked as todo, 9 failures), 0 skipped.
WLEventLogoff: 30 tests executed (0 marked as todo, 5 failures), 2 skipped.
WLEventShutdown: 31 tests executed (0 marked as todo, 5 failures), 0 skipped.
```
One more step to support ghost windows.
JIRA issue: CORE-19671
- Add IntForceMinimizeWindow helper
function.
- Use IntForceMinimizeWindow in
co_WinPosShowWindow function.
- Enhance ShowWindow testcase in
user32_apitest.
- Add mask to the return value of
GetWindowLong.
- Sync winerror.h to wine-10.0
- Fix a couple hacks that existed because we didn't do this sooner
- Fix an instance where Wine compares a pointer to S_OK to see if it's null
Previously, when leave was used in the except or finally block of a nested try block, it would jump back to the start of the except/finally block, resulting in an endless loop. This is fixed by jumping back to a label at the beginning of the try block, which is only visible from within the try block itself and from there to the end of the SEH block.
Fixes seh0055 testcase of MS SEH tests.
Add a `wlntfytests.dll` in `ReactOS\bin\suppl`, containing tests for Winlogon notifications.
To use it:
- Register the dll with `regsvr32 <path_to>\wlntfytests.dll` and reboot
the machine.
- With a debugger, observe the traces emitted by the dll, whenever any
Winlogon notification is emitted. If Winlogon debugging is enabled,
a debug-break is made everytime a notification is invoked.
- Once testing is done, unregister the dll with `regsvr32 /u <path_to>\wlntfytests.dll`
and reboot the machine.
For each received notification, the handlers dump the contents of the
`WLX_NOTIFICATION_INFO` structure, and verify that the notifications
are received in the correct expected order. This is done by keeping
what the previous received notification was, before comparing it with
the expected one(s) that must precede the new notification being received.
References:
- https://learn.microsoft.com/en-us/windows/win32/secauthn/creating-a-winlogon-notification-package
- https://learn.microsoft.com/en-us/windows/win32/secauthn/winlogon-notification-events
- https://rsdn.org/article/baseserv/winlogon.xml
