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

[Eike Rathke]

 sc/source/ui/docshell/impex.cxx |   13 -------------
 1 file changed, 13 deletions(-)

New commits:
commit a883783c2a1b0e7a8ee5b46387d1e1e3bc5e19bd
Author: Eike Rathke <erack at redhat.com>
Date:   Tue May 14 13:39:27 2013 +0200

    resolved fdo#60468 no special tab case workaround for CSV import
    
    This reverts 322cbc3818b0553254aab2dfb3c5b196fe814097 "Import mal-formed
    csv files gracefully." (code now moved to a different location) which
    was a workaround for a specific situation that breaks things with other
    documents.
    
    With the rework of CSV import for 3.6 the import of the one failing
    document https://bugzilla.novell.com/attachment.cgi?id=294589 from
    https://bugzilla.novell.com/show_bug.cgi?id=507322 works almost fine
    without this workaround, with one difference in row 1805 where data is
    completely broken and parts of two rows end up in one cell P1805. That
    row wasn't correct with the workaround either, see also cell P1805 in
    earlier versions.
    
    As a result now without that workaround there is one data row less for
    the broken document but correct data for correct documents.
    
    Change-Id: Iacf56993619cd92c1df81f0c49de0eeb04c2e4e4

diff --git a/sc/source/ui/docshell/impex.cxx b/sc/source/ui/docshell/impex.cxx
index 9b466bf..e09a139 100644
--- a/sc/source/ui/docshell/impex.cxx
+++ b/sc/source/ui/docshell/impex.cxx
@@ -2302,9 +2302,6 @@ OUString ReadCsvLine( SvStream &rStream, bool bEmbeddedLineBreak,
     {
         const sal_Unicode* pSeps = rFieldSeparators.GetBuffer();
 
-        // See if the separator(s) include tab.
-        bool bTabSep = lcl_UnicodeStrChr(pSeps, '\t') != NULL;
-
         QuoteType eQuoteState = FIELDEND_QUOTE;
         bool bFieldStart = true;
 
@@ -2319,16 +2316,6 @@ OUString ReadCsvLine( SvStream &rStream, bool bEmbeddedLineBreak,
             {
                 if (nQuotes)
                 {
-                    if (bTabSep && *p == '\t' && (nQuotes % 2) != 0)
-                    {
-                        // When tab-delimited, tab char ends quoted sequence
-                        // even if we haven't reached the end quote.  Doing
-                        // this helps keep mal-formed rows from damaging
-                        // other, well-formed rows.
-                        nQuotes = 0;
-                        break;
-                    }
-
                     if (*p == cFieldQuote)
                     {
                         if (bFieldStart)