Upgrade to libxml 2.9.2 and libxslt 1.1.28

third_party/libxslt/win32/config.h

Index: third_party/libxslt/win32/config.h
diff --git a/third_party/libxslt/libxslt/win32config.h b/third_party/libxslt/win32/config.h
similarity index 99%
copy from third_party/libxslt/libxslt/win32config.h
copy to third_party/libxslt/win32/config.h
index 00ba2a0ed428894d922525a3ce09c05eb9035a77..9f60e55ca42b12a7e5c4a64884e6d6badd4300a8 100644
--- a/third_party/libxslt/libxslt/win32config.h
+++ b/third_party/libxslt/win32/config.h
@@ -28,12 +28,12 @@
 #include <math.h>
 #if defined _MSC_VER || defined __MINGW32__
 /* MS C-runtime has functions which can be used in order to determine if
-   a given floating-point variable contains NaN, (+-)INF. These are 
+   a given floating-point variable contains NaN, (+-)INF. These are
    preferred, because floating-point technology is considered propriatary
-   by MS and we can assume that their functions know more about their 
+   by MS and we can assume that their functions know more about their
    oddities than we do. */
 #include <float.h>
-/* Bjorn Reese figured a quite nice construct for isinf() using the 
+/* Bjorn Reese figured a quite nice construct for isinf() using the
    _fpclass() function. */
 #ifndef isinf
 #define isinf(d) ((_fpclass(d) == _FPCLASS_PINF) ? 1 \