This presumably always wanted to check for SAL_MAX_UINT16, not USHRT_MAX

The check against USHRT_MAX was introduced with
9c034462 "INTEGRATION: CWS cmcfixes30: #i72614#
crash in loading corrupt .doc", apparently to ensure the static_cast<sal_uInt16>
in the call to ShortToSVBT16 below (which already was that same
static_cast<sal_uInt16> back tehn) is OK.

Change-Id: Ie385f0f490cab1c59bce07fa25bd9b92e3ad286a
Reviewed-on: https://gerrit.libreoffice.org/48439Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

sw/source/filter/ww8/ww8scan.cxx

@@ -2305,8 +2305,9 @@ void WW8PLCF::GeneratePLCF(SvStream& rSt, sal_Int32 nPN, sal_Int32 ncpN)
 
     if (!failure)
     {
+        // Check arguments to ShortToSVBT16 in loop below will all be valid:
         sal_Int32 nResult;
-        failure = o3tl::checked_add(nPN, ncpN, nResult) || nResult > USHRT_MAX;
+        failure = o3tl::checked_add(nPN, ncpN, nResult) || nResult > SAL_MAX_UINT16;
     }
 
     if (!failure)