Hi Eike,
But there seems to be ( possibly ) be a bug in the
lcl_ScRange_Parse_XL_Header which it seems will always sets some
SCA_VALID...2 bits e.g.
nFlags |= SCA_VALID_TAB2 | SCA_TAB2_ABSOLUTE - ( if only only single
cell address with sheet ref e.g. Sheet1!R1C1 )
or
nFlags |= SCA_VALID_TAB | SCA_VALID_TAB2; - ( if just something like
R1C1 is passed )
otoh the comments ( certainly for the 'Sheet!R1C1` case seem to indicate
this is intentional )
so I changed the routine ( once again ;-) ) while this uncertainty
exists
- USHORT nRes = aRange.Parse( aOne, pDoc, eConv );
+ USHORT nRes = aRange.ParseAny( aOne, pDoc, eConv );
+ USHORT nEndRangeBits = SCA_VALID_COL2 | SCA_VALID_ROW2 |
SCA_VALID_TAB2;
+ // nRes & (SCA_VALID_TAB | SCA_VALID_COL |
SCA_VALID_ROW |
SCA_TAB_3D | SCA_TAB_ABSOLUTE | SCA_ROW_ABSOLUTE | SCA_COL_ABSOLUTE);
+ USHORT nTmp1 = ( nRes & 0x070f );
+ USHORT nTmp2 = ( nRes & nEndRangeBits );
+ // If we have a valid single range with
+ // any of the address bits we are interested in
+ // set - set the equiv end range bits
+ if ( (nRes & SCA_VALID ) && nTmp1 && ( nTmp2 !=
nEndRangeBits ) )
+ nRes |= ( nTmp1 << 4 );
+
if ( (nRes & nMask) == nMask )
Append( aRange );
the above expects that a valid ( non-single ) case will have at least
nEndRangeBits set - that I think is not an unreasonable assumption given
ScRange::ParseAny tests the same bits in order to decide to try parsing
an address.
(p.s. I didn't forget about the suggestion for a define for 0x070f, it's
just I don't want to trigger a ripple re-compile right now )
Noel
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
