Creating a view with an unaligned offset is legal in d3d and
apparently works, both on Windows and Linux, but may violate the
Vulkan specification:

VUID-VkBufferViewCreateInfo-offset-00926(ERROR / SPEC): msgNum: -833749292 - Validation Error: [ VUID-VkBufferViewCreateInfo-offset-00926 ] Object 0: handle = 0x3c000000003c, type = VK_OBJECT_TYPE_BUFFER; | MessageID = 0xce4dfed4 | vkCreateBufferView(): VkBufferViewCreateInfo offset (36) must be a multiple of VkPhysicalDeviceLimits::minTexelBufferOffsetAlignment (16). The Vulkan spec states: offset must be a multiple of VkPhysicalDeviceLimits::minTexelBufferOffsetAlignment (https://www.khronos.org/registry/vulkan/specs/1.2-khr-extensions/html/vkspec.html#VUID-VkBufferViewCreateInfo-offset-00926)

Without any reason to test unaligned buffer views here, just resolve this by changing the test.

Both the Msys2 and Cygwin runtimes make use of
FILE_DISPOSITON_POSIX_SEMANTICS in their implementations of the unlink()
system call. This enables these routines to behave similarly to POSIX where are
unlisted from the directory, if handles are still open.

https://github.com/msys2/msys2-runtime/blob/msys2-3.4.3/winsup/cygwin/syscalls.cc#L722
https://www.cygwin.com/git/?p=newlib-cygwin.git;a=blob;f=winsup/cygwin/syscalls.cc#l724Signed-off-by: Joel Holdsworth <joel@airwebreathe.org.uk>

In order to implement FILE_DISPOSITION_POSIX_SEMANTICS, it will be necessary
to add additional flags to closed_fd. In preparation for this, the unlink
member variable has been replaced with disp_flags which directly reflects the
flags defined in the FILE_DISPOSITION_INFORMATION_EX structure.
Signed-off-by: Joel Holdsworth <joel@airwebreathe.org.uk>

Signed-off-by: Joel Holdsworth <joel@airwebreathe.org.uk>

The FILE_DELETE_ON_CLOSE can be used with the FILE_DISPOSITION_ON_CLOSE flag
of FileDispositionInformationEx.
Signed-off-by: Joel Holdsworth <joel@airwebreathe.org.uk>

Signed-off-by: Joel Holdsworth <joel@airwebreathe.org.uk>

Both the Msys2 and Cygwin runtimes make use of
FILE_DISPOSITION_IGNORE_READONLY_ATTRIBUTE in their implementations of the
unlink() system call. This enables these routines to delete a read-only file
without first modifying the attributes.

https://github.com/msys2/msys2-runtime/blob/msys2-3.4.3/winsup/cygwin/syscalls.cc#L724
https://www.cygwin.com/git/?p=newlib-cygwin.git;a=blob;f=winsup/cygwin/syscalls.cc#l726

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=50771Signed-off-by: Joel Holdsworth <joel@airwebreathe.org.uk>

Signed-off-by: Joel Holdsworth <joel@airwebreathe.org.uk>

This is required by Msys2 when running gpg-agent.

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=54996Signed-off-by: Joel Holdsworth <joel@airwebreathe.org.uk>

Signed-off-by: Joel Holdsworth <joel@airwebreathe.org.uk>

This is far simpler than drawing with it.

The immediate motivation here is that the DXGI_FORMAT_R8G8B8A8_UINT test causes
a validation error when drawing, since the shader declares its resource as a
float4 texture, which does not match the UINT view format. This could be fixed
by using a different shader, but it seems more prudent to just simplify the test
to not draw at all.

This is far simpler than drawing with it.

The immediate motivation here is that the DXGI_FORMAT_R8G8B8A8_UINT test causes
a validation error when drawing, since the shader declares its resource as a
float4 texture, which does not match the UINT view format. This could be fixed
by using a different shader, but it seems more prudent to just simplify the test
to not draw at all.

This reverts commit a19c8712.

Fixes both a memory leak and an alloc/free mismatch, and the module heap
is the most appropriate place to hold the variable long-term.