Currently, LibreOffice only uses the "ANSI" variants of the ODBC calls, never the Unicode variants (with a W suffix). However, much (but not all) of the code is actually prepared to use the Unicode variants.
Using the Unicode variants can be useful because it allows us not to mess with encodings, we get more-or-less directly the UTF-16 that our OUStrings need. Also, it allows to work around bugs in ODBC drivers, see e.g. the discussion in comment 13 of bug 68426. But it may trigger *other* bugs in ODBC drivers...
Some decision needs to be done when to use which variant. Possibly it should just be an "Advanced setting" (see menu edit / database / advanced settings), or even better something set in the same dialog tab as the connection charset.
See in directory connectivity/source/drivers/odbcbase/:
- OTools.cxx: argument _bUseWChar to various functions
- OPreparedStatement.cxx: constant useWChar
- OResultsetMetaData.cxx: this one is not ready;
need to use N3SQLColAttribute OR N3SQLColAttributeW
depending on conditions / configuration / ...
these to add the W variants to these two functions.
There is a good explanation on ODBC and Unicode in general at
the authoritative reference is at
Adding self to CC if not already on
Migrating Whiteboard tags to Keywords: (easyHack, difficultyInteresting, skillCpp, skillSql)
JanI is default CC for Easy Hacks (Add Jan; remove LibreOffice Dev List from CC)