Hi Markus, On Wednesday, 2011-08-17 07:23:42 +0200, Markus Mohrhard wrote:
That was my intention. My idea was that we should create a global range name as default and not a local range name that points to the same cell. But after a talk to Kohei and your impression it seems that this is counter intuitive, so I'll change it.
Yes, if one has a sheet local name and copies the sheet or parts thereof s/he expects the same structure.
In a dbgutil build the shell displays Error: FormulaToken::SetIndex: virtual dummy called From File /lo/core/formula/source/core/api/token.cxx at Line 212Can it be that a debug build does not include the dgbutil messages?
Yes, that is independent from each other.
Apparently the FormulaToken used is not a FormulaIndexToken.Oh. I should not only override SetByte but also SetIndex. In calc all formula tokens are derived from ScToken which is derived from FormulaToken and not from FormulaIndexToken. I should no longer finish patchs after learning.c
Maybe that's due to some confusion when ScToken is actually used instead of the types provided in formula/*. The general formula/* implementation handles everything that doesn't need to access the document model or know about application details, ScToken derivates implement spreadsheet specific parts that need to know about addressing and such or use Calc data types. Actually the new ScNameToken implemented would belong to formula/* instead, but there is already FormulaIndexToken, which also now has changes for local/global names, so ScNameToken is superfluous and can be removed. ScRawToken::CreateToken() needs to be changed back to return a FormulaIndexToken.
I think best is that I add a unit test before I push any further modifications in this area.
Always good. Eike -- PGP/OpenPGP/GnuPG encrypted mail preferred in all private communication. Key ID: 0x293C05FD - 997A 4C60 CE41 0149 0DB3 9E96 2F1A D073 293C 05FD
Attachment:
pgpkXMEOOTJzf.pgp
Description: PGP signature