[Libreoffice-commits] core.git: sc/source

[Eike Rathke (via logerrit)]

 sc/source/core/tool/interpr3.cxx |    6 ++++++
 1 file changed, 6 insertions(+)

New commits:
commit 08b4ff03606540f8366217be8096e147f47dee49
Author:     Eike Rathke <er...@redhat.com>
AuthorDate: Sat Mar 4 19:36:30 2023 +0100
Commit:     Eike Rathke <er...@redhat.com>
CommitDate: Sat Mar 4 21:52:14 2023 +0000

    Comment that LARGE()/SMALL() rank svDoubleRef can not be hit in array mode
    
    Change-Id: I8df18b3164941cbf7f58317c2effec05ac0cb492
    Reviewed-on: https://gerrit.libreoffice.org/c/core/+/148255
    Reviewed-by: Eike Rathke <er...@redhat.com>
    Tested-by: Jenkins

diff --git a/sc/source/core/tool/interpr3.cxx b/sc/source/core/tool/interpr3.cxx
index c20d7e0327e1..25fdf94e7f78 100644
--- a/sc/source/core/tool/interpr3.cxx
+++ b/sc/source/core/tool/interpr3.cxx
@@ -3906,6 +3906,12 @@ std::vector<double> ScInterpreter::GetTopNumberArray( 
SCSIZE& rCol, SCSIZE& rRow
                     aArray.push_back(fCellVal);
                 while (aValIter.GetNext(fCellVal, nErr) && nErr == 
FormulaError::NONE);
             }
+            // Note that SMALL() and LARGE() rank parameters (2nd) have
+            // ParamClass::Value, so in array mode this is never hit and
+            // argument was converted to matrix instead, but for normal
+            // evaluation any non-numeric value including empty cell will
+            // result in error anyway, so just clear and propagate an existing
+            // error here already.
             if (aArray.size() != rCol * rRow)
             {
                 aArray.clear();