From 5c2894222beed4b30c6be0379cade228c42c5610 Mon Sep 17 00:00:00 2001
From: Stephan Bergmann <sbergman@redhat.com>
Date: Fri, 7 Oct 2016 18:12:26 +0200
Subject: New rtl::splitSurrogates, remove code duplication

Change-Id: Ic96b64244f817196ccdfe06b97f7f31291adf372
---
 include/rtl/character.hxx | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

(limited to 'include/rtl/character.hxx')

diff --git a/include/rtl/character.hxx b/include/rtl/character.hxx
index ba3088efdeda..4546c9f2d5e0 100644
--- a/include/rtl/character.hxx
+++ b/include/rtl/character.hxx
@@ -23,6 +23,7 @@
 #include <sal/config.h>
 
 #include <cassert>
+#include <cstddef>
 
 #include <sal/types.h>
 
@@ -308,6 +309,31 @@ inline sal_uInt32 combineSurrogates(sal_uInt32 high, sal_uInt32 low) {
         + (low - detail::surrogatesLowFirst) + 0x10000;
 }
 
+/** Split a Unicode code point into UTF-16 code units.
+
+    @param code  A Unicode code point.
+
+    @param output  A non-null pointer to an array with space for at least two
+    sal_Unicode UTF-16 code units.
+
+    @return  The number of UTF-16 code units placed into the output (either one
+    or two).
+
+    @since LibreOffice 5.3
+*/
+inline std::size_t splitSurrogates(sal_uInt32 code, sal_Unicode * output) {
+    assert(isUnicodeCodePoint(code));
+    assert(output != NULL);
+    if (code < 0x10000) {
+        output[0] = code;
+        return 1;
+    } else {
+        output[0] = getHighSurrogate(code);
+        output[1] = getLowSurrogate(code);
+        return 2;
+    }
+}
+
 }
 
 #endif
-- 
cgit