The 0xe nibble value is reserved and makes hid-decode from hid-tools
crash while tryign to parse the descriptor.

0x0 is the correct way of expressing no special units are used.

whole_rect is in win32 virtual screen coordinates. It needs to be converted to x11 root
coordinates for XSetWMNormalHints().

This reverts commit 7a941fd5.
It doesn't work properly since explorer fails to launch, and it breaks
the tests.

Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

With IOleCommandTarget stub, since it's the interface used.
Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

And only if that fails try again by enumerating types.

Otherwise the next topology_branch_connect call will fail when it will
try to lookup the upstream element media type.

It only succeeds with converters, or if a decoder can directly output
the downstream media type.

If it fails we may have to add a converter after the decoder, and we
will call topology_branch_connect on the downstream branch for that.

Instead of modifying the method_mask parameter, in case we need to retry
with another transform.

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=53934

With a global fallback setting under HKCU\Software\Wine\Explorer.

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=56243

Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>

Shifting the rest of the variables by one towards indices of lesser value
overwrites the variable being cleared and effectively leaks it.
Signed-off-by: David Kahurani <k.kahurani@gmail.com>

The data is the path is invalid and therefore caution has to be
taken when adding data to this path to avoid lengthening it
unnecessarily. Also, don't assume there's a head node on the list
while counting number of nodes.
Signed-off-by: David Kahurani <k.kahurani@gmail.com>

This avoids a situation where the lengthening code, assuming the
data in the path is valid proceeds to lengthen the path further
while transforming data from a path_list into a path.
Signed-off-by: David Kahurani <k.kahurani@gmail.com>

This seems like a better interface and avoids duplicating code.
Signed-off-by: David Kahurani <k.kahurani@gmail.com>

This seems like a more effective interface and avoids
code duplication.
Signed-off-by: David Kahurani <k.kahurani@gmail.com>