Release 0.37.0 was not actually released.  From the git log, there are
two functional changes about optimization by using float points directly

	Bump version to 0.38.0
	pixman: Use maximum precision for pixman-bits-image, v2.
	Implement floating point gradient computation, v2.
	build: Add meson files to EXTRA_DIST
	editorconfig: use tabs for Makefiles
	Merge remote-tracking branch 'origin/master'
	Post release version bump to 37.1
	gitlab-ci: Add meson build to pipeline test
	meson: Add a meson build system
	Add .editorconfig file
	Bump version to 0.36.0

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>

Patch 0005 is now included in upstream commit 7c6066b ("pixman-private:
include <float.h> only in C code")

This will be required by qemu 2.11.  The two patches are copied from
packaging of current qemu 2.6

Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>