- BasepGetModuleHandleExW returns BOOLEAN, not NTSTATUS. Treat it as such.
- Always set last error code on failure in BasepGetModuleHandleExW.
CORE-9241 #resolve
svn path=/trunk/; revision=66382
- Mispelling + whitespace fixes.
- Move the finalization of the boot logo display *before* we enable strings to be displayed (as on Windows), but *after* having increased the progress bar to 100% (contrary to Windows), so that we can display the progress bar at 100% without any problem and then do the correct transition:
* either GUI splash-screen --> text-mode emulation console with product name in console header + footer (aka. SOS mode)
* or SOS mode --> text-mode emulation without any graphics.
Since it is done before starting SMSS, we can correctly display the strings it may output (or start other native apps) without fearing that after 5 seconds, the screen clears up in front of you.
CORE-9112 #resolve #comment Fixed in r66378, thanks for reporting!
svn path=/trunk/; revision=66378
CORE-3655 #comment Hackfix the fact that when being compiled with GCC on Linux, compiled messages have their endlines encoded with LF instead of with CR-LF as required by Windows compatibility. This should be normally fixed in windmc (the GCC/binutils message compiler); I've enough complained about that, but not directly to GCC/binutils team. Until they fix their tool, I ask SVN to always store the .mc files with CR-LF endlines. This should "fix" the problem. Suggestion by Thomas Faber.
svn path=/trunk/; revision=66376
Add back support for allocator stack backtraces for GDI objects and dump a list of BTs, when the GDI object table is exhausted. Disabled by default.
svn path=/trunk/; revision=66374
- In PspExitThread, correctly set the ZeroInit field of the message before calling LpcRequestPort (which would otherwise complain with STATUS_INVALID_PARAMETER and fail to send the message)
- In NtRegisterThreadTerminatePort, dereference the port object in case of failure, not a null pointer
CORE-9234 #resolve
svn path=/trunk/; revision=66369
- Add a test for process termination behavior. This shows that CreateProcess with CREATE_SUSPENDED followed by TerminateProcess causes a handle leak.
CORE-9234
svn path=/trunk/; revision=66367
- Make BaseSetLastNTError return the converted Win32 error code. This will determine the upper 24 bits of EAX in functions that return BOOLEAN FALSE right after calling BaseSetLastNTError, e.g. Wow64EnableWow64FsRedirection. Fixes installers using WiX Toolset (e.g. VS2012 redist) on MSVC builds.
See http://wixtoolset.org/issues/4681/ for the WiX bug that causes this.
CORE-8010
svn path=/trunk/; revision=66365
OK, Thomas had to convince me that MAX_USHORT + MAX_USHORT + 0x400 + 0x6 cannot overflow an ULONG.
So remove useless check.
By Thomas :-)
svn path=/trunk/; revision=66358
Don't pretend that ExceptionRecord.ExceptionInformation[2] is a pointer to a string in case of Delphi exception. This can lead to invalid pointer dereference in DbgPrint then.
For instance:
(../../dll/win32/kernel32/client/except.c:371) Delphi Exception at address: 50075885
(../../dll/win32/kernel32/client/except.c:372) Exception-Object: 00E885A8
(../../dll/win32/kernel32/client/except.c:373) Exception text: 2
svn path=/trunk/; revision=66354
- Fix packing for system restore point API types (STATEMGRSTATUS in particular). Fixes VC++ 2012 redist installer.
CORE-8010 #resolve
svn path=/trunk/; revision=66352
- Add some simple tests for RtlRaiseStatus to the RtlException test
- Add missing volatile in (disabled) stack overflow test that caused an infinite loop on GCC builds
CORE-8773
svn path=/trunk/; revision=66351
- Refactor EnumFontFamilies test to be more generic, and check for some common expected system fonts. This shows that EnumFontFamilies behaves correctly, but that we're missing the System font in the enumeration.
CORE-9222
svn path=/trunk/; revision=66347
- Reject bits data early in SetBitmapBits when it's clear that they can't fit in the destination bitmap.
Fixes ugly Office 2007 rulers showing garbage.
svn path=/trunk/; revision=66342
