[libreoffice-commits] Hopefully this is the last of the cleanups for class ScDocument for now.

[Stephan Bergmann]

The changes of

> commit 872daba66c4be77c9df9d6fe6a1c501abe087d04
> Author: Joseph Powers <jpowers27 at cox.net>
> Date:   Thu Sep 8 21:05:58 2011 -0700
>
>     Hopefully this is the last of the cleanups for class ScDocument for now.

to

> diff --git a/binfilter/bf_sc/source/filter/xml/sc_XMLExportDDELinks.cxx b/binfilter/bf_sc/source/filter/xml/sc_XMLExportDDELinks.cxx
> index 343668d..3100f23 100644
> --- a/binfilter/bf_sc/source/filter/xml/sc_XMLExportDDELinks.cxx
> +++ b/binfilter/bf_sc/source/filter/xml/sc_XMLExportDDELinks.cxx
> @@ -132,15 +132,13 @@ void ScXMLExportDDELinks::WriteTable(const sal_Int32 nPos)
>              for(sal_Int32 nColumn = 0; nColumn < nColCount; nColumn++)
>              {
>                  if (nColumn == 0)
> -                    bPrevEmpty = rExport.GetDocument()->GetDdeLinkResult(pMatrix, static_cast<USHORT>(nColumn), static_cast<USHORT>(nRow),
> -                                                                        sPrevValue, fPrevValue, bPrevString);
> +                    bPrevEmpty = TRUE;
>                  else
>                  {
>                      double fValue;
>                      String sValue;
>                      sal_Bool bString(sal_True);
> -                    sal_Bool bEmpty = rExport.GetDocument()->GetDdeLinkResult(pMatrix, static_cast<USHORT>(nColumn), static_cast<USHORT>(nRow),
> -                                                                        sValue, fValue, bString);
> +                    sal_Bool bEmpty = TRUE;
>                      if (CellsEqual(bPrevEmpty, bPrevString, sPrevValue, fPrevValue,
>                                  bEmpty, bString, sValue, fValue))
>                          nRepeatColsCount++;

causes a warning now that fValue is uninitialized in the assignment

   fPrevValue = fValue;

(in code below the code shown above).  From bPrevString, bPrevEmpy, 
bString, and bEmpty always being true, it looks like fPrevValue and 
fValue have become unused anyway, but it looks a little suspicious to me 
that this code has degraded now to a glorious way of repeatedly calling 
WriteCell to only ever write empty cells.  I know too little about this 
area to tell whether this is as intended or got broken.  (And the 
respective code in sc/source/filter/xml/XMLExportDDELinks.cxx has 
diverted sufficiently by now to prevent direct comparison.)

Stephan