tdf#97198 Make Calc function LEFT work with Unicode non-BMP characters

Change-Id: I66d24dd962f824079fd0aea2827bd5b975ffe698
Reviewed-on: https://gerrit.libreoffice.org/44684
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Eike Rathke <erack@redhat.com>

2 files changed, 92 insertions, 3 deletions

diff --git a/sc/qa/unit/data/functions/text/fods/left.fods b/sc/qa/unit/data/functions/text/fods/left.fods
index 535eff455432..45d659aab04e 100644
--- a/sc/qa/unit/data/functions/text/fods/left.fods
+++ b/sc/qa/unit/data/functions/text/fods/left.fods
@@ -1152,6 +1152,92 @@
      <table:table-cell table:style-name="ce17"/>
      <table:table-cell table:number-columns-repeated="4"/>
     </table:table-row>
+    <table:table-row table:style-name="ro5">
+     <table:table-cell table:style-name="ce14" table:formula="of:=LEFT([.I11];4)" office:value-type="string" office:string-value="ab😂d" calcext:value-type="string">
+      <text:p>ab😂d</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce17" office:value-type="string" calcext:value-type="string">
+      <text:p>ab😂d</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce38" table:formula="of:=[.A11]=[.B11]" office:value-type="boolean" office:boolean-value="true" calcext:value-type="boolean">
+      <text:p>WAAR</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce11" table:formula="of:=FORMULA([.A11])" office:value-type="string" office:string-value="=LEFT(I11;4)" calcext:value-type="string">
+      <text:p>=LEFT(I11;4)</text:p>
+     </table:table-cell>
+     <table:table-cell office:value-type="string" calcext:value-type="string">
+      <text:p>Excel 2016 returns the same result</text:p>
+     </table:table-cell>
+     <table:table-cell table:number-columns-repeated="3"/>
+     <table:table-cell office:value-type="string" calcext:value-type="string"><text:p>ab<text:span text:style-name="T1">😂de𝔖gh𝕬𝖈𝖍𝖙𝖚𝖓𝖌!</text:span></text:p>
+     </table:table-cell>
+     <table:table-cell table:number-columns-repeated="4"/>
+    </table:table-row>
+    <table:table-row table:style-name="ro6">
+     <table:table-cell table:style-name="ce14" table:formula="of:=LEFT([.I12];2)" office:value-type="string" office:string-value="&quot;𧌒" calcext:value-type="string">
+      <text:p>&quot;𧌒</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce17" office:value-type="string" calcext:value-type="string">
+      <text:p>&quot;𧌒</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce39" table:formula="of:=[.A12]=[.B12]" office:value-type="boolean" office:boolean-value="true" calcext:value-type="boolean">
+      <text:p>WAAR</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce11" table:formula="of:=FORMULA([.A12])" office:value-type="string" office:string-value="=LEFT(I12;2)" calcext:value-type="string">
+      <text:p>=LEFT(I12;2)</text:p>
+     </table:table-cell>
+     <table:table-cell office:value-type="string" calcext:value-type="string">
+      <text:p>Excel 2016 returns the same result</text:p>
+     </table:table-cell>
+     <table:table-cell table:number-columns-repeated="3"/>
+     <table:table-cell table:style-name="ce44" office:value-type="string" calcext:value-type="string"><text:p><text:span text:style-name="T2">&quot;</text:span><text:span text:style-name="T3">𧌒𧀬</text:span><text:span text:style-name="T4">&quot;</text:span></text:p>
+     </table:table-cell>
+     <table:table-cell table:number-columns-repeated="4"/>
+    </table:table-row>
+    <table:table-row table:style-name="ro3">
+     <table:table-cell table:style-name="ce14" table:formula="of:=LEFT([.I13];2)" office:value-type="string" office:string-value="&quot;u" calcext:value-type="string">
+      <text:p>&quot;u</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce17" office:value-type="string" calcext:value-type="string">
+      <text:p>&quot;u</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce40" table:formula="of:=[.A13]=[.B13]" office:value-type="boolean" office:boolean-value="true" calcext:value-type="boolean">
+      <text:p>WAAR</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce11" table:formula="of:=FORMULA([.A13])" office:value-type="string" office:string-value="=LEFT(I13;2)" calcext:value-type="string">
+      <text:p>=LEFT(I13;2)</text:p>
+     </table:table-cell>
+     <table:table-cell office:value-type="string" calcext:value-type="string">
+      <text:p>Excel 2016 returns the same result</text:p>
+     </table:table-cell>
+     <table:table-cell table:number-columns-repeated="3"/>
+     <table:table-cell table:style-name="ce45" office:value-type="string" calcext:value-type="string">
+      <text:p>&quot;üë&quot;</text:p>
+     </table:table-cell>
+     <table:table-cell table:number-columns-repeated="4"/>
+    </table:table-row>
+    <table:table-row table:style-name="ro5">
+     <table:table-cell table:style-name="ce14" table:formula="of:=LEFT([.I14];4)" office:value-type="string" office:string-value="𝕬𝖈𝖍𝖙" calcext:value-type="string">
+      <text:p>𝕬𝖈𝖍𝖙</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce17" office:value-type="string" calcext:value-type="string">
+      <text:p>𝕬𝖈𝖍𝖙</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce41" table:formula="of:=[.A14]=[.B14]" office:value-type="boolean" office:boolean-value="true" calcext:value-type="boolean">
+      <text:p>WAAR</text:p>
+     </table:table-cell>
+     <table:table-cell table:style-name="ce11" table:formula="of:=FORMULA([.A14])" office:value-type="string" office:string-value="=LEFT(I14;4)" calcext:value-type="string">
+      <text:p>=LEFT(I14;4)</text:p>
+     </table:table-cell>
+     <table:table-cell office:value-type="string" calcext:value-type="string">
+      <text:p>Excel 2016 returns the same result</text:p>
+     </table:table-cell>
+     <table:table-cell table:number-columns-repeated="3"/>
+     <table:table-cell table:style-name="ce46" office:value-type="string" calcext:value-type="string">
+      <text:p>𝕬𝖈𝖍𝖙𝖚𝖓𝖌!</text:p>
+     </table:table-cell>
+     <table:table-cell table:number-columns-repeated="4"/>
+    </table:table-row>
     <table:table-row table:style-name="ro5" table:number-rows-repeated="27">
      <table:table-cell table:style-name="ce14"/>
      <table:table-cell table:style-name="ce17"/>
@@ -1195,4 +1281,4 @@
    </table:named-expressions>
   </office:spreadsheet>
  </office:body>
-</office:document>
\ No newline at end of file
+</office:document>
diff --git a/sc/source/core/tool/interpr1.cxx b/sc/source/core/tool/interpr1.cxx
index 88b1fd43759d..2895c93f2a2e 100644
--- a/sc/source/core/tool/interpr1.cxx
+++ b/sc/source/core/tool/interpr1.cxx
@@ -8579,8 +8579,11 @@ void ScInterpreter::ScLeft()
         else
             n = 1;
         OUString aStr = GetString().getString();
-        n = std::min(n, aStr.getLength());
-        aStr = aStr.copy( 0, n );
+        sal_Int32 nIdx = 0;
+        sal_Int32 nCnt = 0;
+        while ( nIdx < aStr.getLength() && n > nCnt++ )
+            aStr.iterateCodePoints( &nIdx );
+        aStr = aStr.copy( 0, nIdx );
         PushString( aStr );
     }
 }