Jan Dubois wrote: > VT_DATE values used to be returned as strings, but that doesn't give > you any control over how the date is formatted.
I would say the user has - although limited - control with the global LCID, right? > A long time ago I > changed Win32::OLE to return them as VARIANTs. If you include the > Win32::OLE::Variant module, they will still be stringified as before, > but you can then also treat them as objects and use the > Win32::OLE::Variant methods to format them the way you want to. > > It was discovered too late that this was a change in behavior for > applications that do *not* use the Win32::OLE::Variant module. Since > reverting the behavior would have broken programs written after the > change, I left the new behavior as is. Thank you very much for clarification! All works that nice behind the scenes that DBD::ADO could return Win32::OLE::Variant's. But I have a big yellow sticker on my desk with the letters 'DBD::Proxy' ... > In Win32::OLE 0.18 and later VT_CY and VT_DECIMAL values may also be > returned as Win32::OLE::Variant objects, but only if you ask for that > with: > > Win32::OLE->Option(Variant => 1); And Variant => 0 means that all types (even VT_DATE) are stringified? Will Win32::OLE 0.18 be available on CPAN? Thanks, Steffen
