Our On-Screen Keyboard has a manifest and the buttons (the ones with BS_ICON styles) aren't rendered with the specific theme as it should be but instead it takes the classic theme.
The code relies on NM_CUSTOMDRAW notification, which is more intuitive and efficient than doing owner-drawn operations as NM_CUSTOMDRAW allows the controls to use styles whereas you cannot do it on owner-drawn controls.
CORE-15965
This finally fixes our duplicated getopt functions from different sources and gives us an up to date and the most compatible implementation.
isohybrid actually relies on a glibc-specific getopt behavior that we previously hacked into the reactos_support_code.c implementation derived from BSD/mingw-w64.
widl also needs getopt and previously used an even older BSD-derived code.
- The 2nd parameter is already passed in user-mode by GetThreadDesktop().
It is then used by NtUserGetThreadDesktop() as a fallback for console
threads.
- Lookup and validate the thread by using the IntTID2PTI() helper.
- Don't reference the desktop with too many access rights.
- Get rid of the old-school DECLARE_RETURN() & co. macros.
Co-authored-by: Hermès Bélusca-Maïto <hermes.belusca-maito@reactos.org>
It retrieves the handle to the desktop assigned to the specified
thread belonging to a console application, which is going to be
where the input thread of this console is.
Up to CMake 3.3 LINK_DEPENDS was ignored in non-Makefile generators. See also http://www.cmake.org/pipermail/cmake/2010-May/037206.html. Since CMake 3.4 it is no longer ignored, but requires CMP0005 policy to be set to NEW and it requires a full path which was already correctly used on GCC, but not on MSVC builds.
* [CMAKE] Fix for CMP0018
The new behaviour lets us explicitly disabling position independent code, avoiding the need to change undocumented variables
* LINK_DEPENDS is semicolon separated
* [CMAKE] Fix for CMP0005
it seems impossible to have cmake correct escape PROXY_CLSID_IS so move it to the header
unattend.inf only allows to add RunOnce entries, which are in theory only executed on the first boot after installation.
For almost a decade, this worked anyway to start regtest.cmd again after a ReactOS crash, because the registry and thus the deletion of the RunOnce entry wasn't stored on disk in case of a crash.
Thanks to Sylvain for the hint!
As Explorer processes the Run key right after RunOnce and therefore picks up regtest.cmd a second time during the first boot, exit the script just after adding the Run key.
Otherwise, rosautotest.exe would be started twice, with catastrophic results.
This is a rather graceful approach compared to sysreg2's 3 minute timeout before killing and restarting the entire VM.
Since we added autochk for FAT filesystems, the filesystem is often "fixed" after a reset with the consequence that ReactOS doesn't boot up anymore.
The sysreg2 restart code still remains for handling tests causing BSODs.
