tdf#125497 allow backspace to remove CJK IVS.

Japanese users prefer to remove a CJK IVS character
when pressing the backspace instead of removing
the selector part of IVS.

Change-Id: I4313d69ed52d82c5a7e4e4823b1da06f1d90bdad
Reviewed-on: https://gerrit.libreoffice.org/72971
Tested-by: Jenkins
Reviewed-by: Mark Hung <marklh9@gmail.com>

editeng/source/editeng/impedit2.cxx

@@ -60,6 +60,7 @@
 #include <sot/exchange.hxx>
 #include <sot/formats.hxx>
 #include <svl/asiancfg.hxx>
+#include <rtl/character.hxx>
 #include <comphelper/lok.hxx>
 #include <unotools/configmgr.hxx>
 
@@ -2300,7 +2301,20 @@ EditPaM ImpEditEngine::DeleteLeftOrRight( const EditSelection& rSel, sal_uInt8 n
     {
         if ( nDelMode == DeleteMode::Simple )
         {
-            aDelStart = CursorLeft( aCurPos, i18n::CharacterIteratorMode::SKIPCHARACTER );
+            sal_uInt16 nCharMode = i18n::CharacterIteratorMode::SKIPCHARACTER;
+            // Check if we are deleting a CJK ideograph variance sequence (IVS).
+            sal_Int32 nIndex = aCurPos.GetIndex();
+            if (nIndex > 0)
+            {
+                const OUString& rString = aCurPos.GetNode()->GetString();
+                sal_Int32 nCode = rString.iterateCodePoints(&nIndex, -1);
+                if (rtl::isIVSSelector(nCode) && nIndex > 0 &&
+                        rtl::isCJKIVSCharacter(rString.iterateCodePoints(&nIndex, -1)))
+                {
+                    nCharMode = i18n::CharacterIteratorMode::SKIPCELL;
+                }
+            }
+            aDelStart = CursorLeft(aCurPos, nCharMode);
         }
         else if ( nDelMode == DeleteMode::RestOfWord )
         {

include/rtl/character.hxx

@@ -57,6 +57,35 @@ inline bool isAscii(sal_uInt32 code)
     return code <= 0x7F;
 }
 
+/** Check for Unicode variation sequence selectors
+
+    @param code  A Unicode code point.
+
+    @return  True if code is an Unicode variation sequence selector.
+
+    @since LibreOffice 6.3
+ */
+inline bool isIVSSelector(sal_uInt32 nCode)
+{
+    return (nCode >= 0xFE00 && nCode <= 0xFE0F)   // Variation Selectors block
+        || (nCode >= 0xE0100 && nCode <= 0xE01EF);// Variation Selectors Supplement block
+}
+
+/** Check for base characters of a CJK ideographic variation sequence (IVS)
+
+    @param code  A Unicode code point.
+
+    @return  True if code is an Unicode base character part of CJK IVS
+
+    @since LibreOffice 6.3
+ */
+inline bool isCJKIVSCharacter(sal_uInt32 nCode)
+{
+    return (nCode >= 0x4E00 && nCode <= 0x9FFF)       // CJK Unified Ideographs
+        || (nCode >= 0x3400 && nCode <= 0x4DBF)       // CJK Unified Ideographs Extension A
+        || (nCode >= 0x20000 && nCode <= 0x2A6DF);    // CJK Unified Ideographs Extension B
+}
+
 #if defined LIBO_INTERNAL_ONLY
 bool isAscii(char) = delete;
 bool isAscii(signed char) = delete;

sw/source/uibase/wrtsh/delete.cxx

@@ -44,23 +44,6 @@ inline void SwWrtShell::CloseMark( bool bOkFlag )
     EndAllAction();
 }
 
-namespace {
-
-bool isUnicodeVariationSequenceSelector( sal_uInt32 nCode )
-{
-    return ( nCode >= 0xFE00 && nCode <= 0xFE0F )   // Variation Selectors block
-        || ( nCode >= 0xE0100 && nCode <= 0xE01EF );// Variation Selectors Supplement block
-}
-
-// Return if the character might be a base character of a CJK ideographic variation sequence
-bool isCJKIVSCharacters( sal_uInt32 nCode )
-{
-    return ( nCode >= 0x4E00 && nCode <= 0x9FFF )       // CJK Unified Ideographs
-        || ( nCode >= 0x3400 && nCode <= 0x4DBF )       // CJK Unified Ideographs Extension A
-        || ( nCode >= 0x20000 && nCode <= 0x2A6DF );    // CJK Unified Ideographs Extension B
-}
-
-}
 
 
 // #i23725#
@@ -268,14 +251,14 @@ bool SwWrtShell::DelLeft()
                 nCode = sStr.iterateCodePoints( &nIndex );
             }
 
-            if ( isUnicodeVariationSequenceSelector( nCode ) )
+            if ( rtl::isIVSSelector( nCode ) )
             {
                 SwCursorShell::Push();
                 SwCursorShell::Left(1, CRSR_SKIP_CHARS);
                 OUString sStr = GetSelText();
                 sal_Int32 nIndex = 0;
                 nCode = sStr.iterateCodePoints( &nIndex );
-                if ( isCJKIVSCharacters( nCode ) )
+                if ( rtl::isCJKIVSCharacter( nCode ) )
                     SwCursorShell::Pop( SwCursorShell::PopMode::DeleteStack );
                 else
                     SwCursorShell::Pop( SwCursorShell::PopMode::DeleteCurrent ); // For the weak script.