We need to save the missing registers in the trap frame before calling KiInitiateUserApc, because they will be used to create the CONTEXT that is saved on the user mode stack, where KiUserApcDispatcher returns back to using NtContinue. This fixes remaining failures in kernel32_apitest QueueUserAPC.
Both due-times and interrupt time are unsigned, but were treated as signed in KiInsertTimerTable, which led to very long (e.g. INFINITE) waits being interpreted as having a negative due-time and being completed instantly.
Mostly fixes kernel32_apitest QueueUserAPC
This is used in KiUserModeCallout instead of KiServiceExit2. The latter is broken, leaks non-volatile registers and will need to be modified to handle an exception frame, which we don't need/have here. It will also use sysret instead of iret and is generally simpler/faster.
Eventually it would be desirable to skip the entire trap frame setup and do everything in KiCallUserMode. This requires some cleanup and special handling for user APC delivery.
Fixes CORE-19505 'game "Hover!" from rapps does not start anymore'
It failed showing a msgbox
"There was an error allocating memory in Hover!"
the regression was introduced by
0.4.15-dev-5602-g c47506a5f4
Add missing heap initialization in DllMain entrypoints of msvcrt20.dll and msvcrt40.dll,
similarly to as it is done in msvcrt.dll (CRT).
The msvcrt.dll (CRT) initialization code was updated properly
during the last winesync, accordingly to the new changes,
but msvcrt20.dll/msvcrt40.dll one was not. So update it too.
This fixes the crash of HOVER 1.0 game from Rapps when
it tries to allocate a dynamical memory from the
unitialized heap via malloc() function exported from our msvcrt20.dll.
This is intentionally not part of ntdll_apitest, because that links to advapi32, which links to rpcrt4, which (wrongly!) links to ws2_32, which (wrongly!) links to user32, which breaks the test.
This fixes CORE-19497 'Regression, part of the statusbar with the sizegrip is not redrawn properly'
The regression was introduced / unhidden by 0.4.15-dev-5617-g 7866eb2886
Author of the patch: JIRA-user: I_Kill_Bugs
turned into PR by @Doug-Lyons
Thank you both!
KVM: https://reactos.org/testman/compare.php?ids=94433,94435
VBox: Builder Broken atm unfortunately
This can fail on optimized builds, where functions can end after a noreturn function call without an epilog / ret, but with an int 3. We simply fail gracefully, which is the right thing to do.
This also fixes delivering APCs from the system call handler, which previously would have clobbered rax. Also don't use the thread's TrapFrame member, which is not always set, when returning.
CORE-17256
- Implement support for complex input buffers.
This will enable the IDE driver with some hack
to execute the _STM (channel timing settings) control method.
- Fix a memory leak.
- Fix returned status code.
- Correctly calculate the output buffer size.
- Improve buffer validation.
- Implement support for async control requests.
- Implement local reference conversion.
- Implement support for subpackage argument conversion.
- Place some code into INIT section.
This is needed, because KiSystemCallHandler can be called multiple times for the same syscall entry, which would mess up the linkage. This replaces a previous hack and makes things cleaner.
In case of failure, ObInsertObject itself dereferences the object, so calling ObDereferenceObject one more time in the failure handler is redundant.
- Remove ObDereferenceObject call from CmpCreateRegistryRoot in case ObInsertObject fails.
- Add missing RootKey dereference if CmpCreateRegistryRoot failed to get the key cell (KeyCell).
CORE-17904
- Implement `GetFirmwareType` and improve existing API test for it
- Move (Get/Set)FirmwareEnvironmentVariableEx(A/W) to kernel32_vista_static and export them when NT version >= 6.2
Addendum to 4c8a2a8815. CORE-11954
Follow-up to #6635. The directory search without this flag didn't work.
JIRA issue: CORE-19493
- Add PRF_FIRSTDIRDEF flag for PathResolveW call.
- Delete needless macro definitions in PathQualifyExW.
- Remove an assertion in IsLFNDriveW.
- Fix shell32_apitest:FindExecutable testcase.
- Use an index for the table instead of ugly pointer offset
- Invert the check for GUI calls
- Fail with proper status code on invalid win32k syscall ids
This is not intended to fix any specific bug. It is just meant for cleanup.
No noticable change in behavior at runtime.
Some rcs had 7 of those, some only 6. So this will also slightly improve sync between the rcs.
This is required since while interrupts are enabled, another user APC could get queued and we want to guarantee that those are all delivered before returning to user mode.
- This removes an ancient polish-only workaround/extrawurst of 0.4.13-dev-1171-g c81af08 which made the dialog bigger than allowed in y, and which reduced its usability on small monitor resolution.
- Sync the dialogs dimensions back to en-US and all other localizations. Notice also that both dialogs have the same height again afterwards.
- Maximizes the available space for the dynamic parts of the dialog on the right side. Historically even 160 turned out to be too short for that, e.g. EDITTEXT 14009, the dynamic part of "Lokalizacja:". and also EDITTEXT 14007, the dynamic part of "Program:". Therefore we used 170 everywhere for those widths in all other languages. But due to the polish workaround it was crippled to 155 here. Not good!
- Notice also how now the leftmost checkboxes position of the Attributes is perfectly aligned with the other dynamic content of that column above after the fix. This will be important in the future when we will restore the vertical layout of Attribute-checkboxes in both dlgs (Plik vs. checkbox, and Katalog vs. checkbox) to match 2k3sp2
---------
Co-authored-by: Adam Słaboń <asaillen@protonmail.com>
The MSVC toolset will increment its minor version number from 19.39 (VS 2022 v17.9) to 19.40 (VS 2022 v17.10). The MSVC toolset will be version 19.40.33521.0 in the preview 1 release. For more details, see the MSVC Toolset Minor Version Number 14.40 in VS 2022 v17.10 blog post: https://aka.ms/msvc-ver14.40
Follow-up to #6656. An approach to
implement ShellExecuteEx correctly.
JIRA issue: CORE-19493
- Rewrite code by using AssocQueryStringW
and PathResolveW functions.
same as it is already the case in IDD_MIXED_CONTENT1, just 10 lines below it.
The controls name is not referenced from code. Will save a few bytes in the binary and shortens the rc-line.
* en-US.rc: create the TODO-list for the required menu-accel-changes
* some accelerator fixes, and add many FIXMEs for stuff that still needs to be improved
* beg the translators to shorten the "On Disk:" string in IDD_FILE_PROPERTIES and IDD_FOLDER_PROPERTIES
* strip superfluous spaces at some strings end and ':' within groupboxes, MS never does that
* bg-BG.rc & fr-FR.rc: Fix text-truncation, It got worse by 0.4.15-dev-5993-g a5bff2457a (#5239) A suggestion of Stasm, personally I dislike moving/resizing controls to dimenstions != en-US
* zh-CN.rc: China is the new land of capitalism, so CAPITALIZE the S. Grab another &W used in en-US/zh-TW
* he-IL.rc: strip a useless comment which just does repeat the string 1:1
* zh-TW: This taiwanese question-mark here is wrong, because the question spans over multiple controls, and the last one already ends with such a question mark.
* pt-PT.rc: Name the attributes-column 'Atributos', as that is the same name that is used in the file-properties-dlg
* sk-SK.rc: strip a comment that does not exist in any other lang
* Fix some lines which had 2 accelerators each, e.g. de-DE.
IDD_FOLDER_OPTIONS_GENERAL has 100% correct german accelerators now (checked against german XPSP3).
* es-ES.rc: remove a stone-age-workaround, and use the height from en-US again here. The text fits the single line fine now.
* no-NO.rc: Fix a regression of 0.4.9-dev-794-g bf88116252 which added the 70 to the 3 PUSHBUTTONS for all languages, but for no-NO.rc the third button was forgotten
* pt-PT.rc: Fix 2 wrong accelerators. They have been added by a pt-PT.rc translator, but Windows doesn't have them. Windows uses ENTER and ESCAPE just here. And other translators are aware of that. The thing that actually SHOULD have an accelerator in that dialog is the LTEXT "Abrir", but ironically that one was forgotten / intentionally removed by the pt-PT.rc translator upon translation. A pity!
* pt-PT.rc: pt-PT.rc was crippled regarding accelerators, mostly during 0.4.14-dev'ing. Wisdom of the elders was trampled down here.
* pt-PT.rc: fix unintended "::"
* tr-TR.rc: During 0.4.15-dev'ing a lot of accelerators and capitalization were changed in this file, and during that the IDS_* accelerators at the bottom in that file in the
/* context menus */
section were changed, but the reminder-menu-layouts IDM_* and FCIDM_SHVIEW_* were forgotten to be kept in sync.
That is such a common pitfall. Happened to all of us more than one time already.
Sync the capitalization for that file at least. While knowing that there is still a lot of async left for other files.
* tr-TR.rc: Point with 2 FIXMEs to some accelerator-issues that only somebody with a real turkish Windows will be able to fix properly. Actually most likely at least 6 lines in this file are involved into that collision. Try to mention them within the comment at least.
IDM_COPY
IDM_CREATELINK
IDM_DELETE
and
IDS_COPY
IDS_CREATELINK
IDS_DELETE
This bug does only affect master, and no older releases in this form.
* en-US.rc: And finally add a big fat TODO comment for my future self, regarding dlg redesign where attrubutes will need to be laid out vertically in IDD_FILE_PROPERTIES and IDD_FOLDER_PROPERTIES
Implementing correct FindExecutable...
JIRA issue: CORE-19493
- If there were filename extension, then skip to the extension
by using PathFindExtensionW.
- Use "open" verb if there is no default action if possible.
- Set outlen at CQueryAssociations::ReturnString.
Fixes the reg class key order for FS items. The existing code was close,
but for some reason used `//` as the path separator for SystemFileAssociations!
- Fixed SystemFileAssociations.
- Swapped the order of `*` and `AllFilesystemObjects`. This is the documented
order and can also be observed in Process Monitor.
https://learn.microsoft.com/en-us/windows/win32/shell/fa-associationarray#about-association-arrays
- Removed `(..., L"%s//%s", extension, wszClass)`, this does not seem to be
a valid thing (`.TestAAExtWeird` in my tests).
- Adds the `Unknown` class when appropriate. Not adding the `openas` verb
to `Unknown` rgs registration now to mimic Windows, because ROS
`CDefaultContextMenu` lacks verb de-duplication and the menu would end up
with two "Open With" entries. This just uses `(cidl == 1)` to simulate
Windows, while Windows on NT6 uses `MultiSelectModel=Single`, a NT6 feature
not implemented in ROS.
- The class order for folders was wrong and is still "wrong" in this PR,
but I chose to use the Windows menu display order until the exact mechanics
required in `CDefaultContextMenu` can be understood.
- Extracts the extension from ANSI PIDLs.
Added multiple new actions, most of them will use the system default implementation but can be forced to use a specific CLSID or IShellFolder implementation.
- /shgfi (SHGetFileInfo)
- /assocq (IQueryAssociations)
- /shellexec (ShellExecuteEx on pidl)
- /dumpmenu (Dumps the HMENU of a IContextMenu with its menu ids and verbs)
Added /explorerinstance as a new wait mode and made it the default. All the other wait modes are hacks that just works around bugs in ROS.
A hive whose KCBs have a reference count of 0, meaning nobody is using these keys anymore, will not get removed from the cache table.
As a result during a normal hive unloading operation you will get orphaned KCBs which results in an unload failure.
This is wrong, because this is what a normal hive unloading is supposed to do. What it cannot do of course is that it cannot
scramble the references of opened keys by the users who use the Registry, as it is the job of force unloading mechanism to do that.
Also remove a misleading debug print. Force unloading works as intended by scrambling the references of keys and marking the KCB for deletion,
which is what how a force unload works. Namely Windows does exactly that.
CORE-10705
- Annotate the CmpEnumerateOpenSubKeys function with SAL2
- When removing an orphaned cached KCB, ensure that it is locked before clearing it from cache table entries
- HKCR delete handling; delete the value in HKCU if it exists there, otherwise delete it in HKLM.
- Changed RegDeleteKeyValueW to just use the regular Reg functions to get HKCR handling for free.
Andrei Miloiu