Re: Numeric datatype doesn't match those from older driver version ?

On Thu, Dec 18, 2008 at 5:33 AM, Adam M <gnuman1(at)gmail(dot)com> wrote:
>
> http://en.wikipedia.org/wiki/ANSI_art
>
> I do blame Microsoft if they are calling UTF-8 ANSI? To me ANSI was
> back in the 1980s with BBSs and extended ASCII codes. ANSI terminals
> allowed scrolling around, etc..

It comes from the Microsoft conventions that sprung up when they added
the Unicode APIs. For function CreateFoo, they created CreateFooA and
CreateFooW - the A version being the ANSI, the W being the Wide
character (or Unicode UTF-16). If you program without the UNICODE
macro defined, CreateFoo maps to CreateFooA, otherwise to CreateFooW.

In the ODBC driver world, the ANSI drivers can be made to handle
multibyte characters through the ANSI interface, however, if you want
'proper' UTF-16 support used natively by modern versions of Windows,
you need the Unicode API extensions. psqlODBC kinda stretches this to
the limits with the multibyte support in the ANSI driver, which
doesn't play well with the UTF-16 support with some applications -
hence the use of two drivers.

--
Dave Page
EnterpriseDB UK: http://www.enterprisedb.com