Thread: Unicode v ANSI Calls

Hi there,

I am writing an application for which I am using unicode throughout.

However, I would still like my code to run on a Win9x machine without Microsofts Unicode Layer support.

What I have in mind is something like the following [psuedo code]:

Code:

  if RUNTIME OS is windows 2000 or above
    WinCallW(); // call wide function
  else
    WinCallA(); // call ANSI function

In other words, if the application detects it is running on Windows 2000 or above, it calls the wide function, or on 9.x it calls the ANSI equivalent.

While wide functions will never get called on Win 9.x, my question is will it run? Or will my app die with an error message because I have linked against a wide call? (I don't have any Win9.x platforms I could test it on).

PS. Please don't suggest I just define UNICODE -- it's not what I'm trying to achieve.

Thanks

Think I found the answer. Wikipedia tells me:

"Normally, the Win32 API provides both A (ANSI) and W (Wide-character) versions of most functions. On Windows 95/98/ME, only the A versions are implemented and attempting to call a W version will fail with an error code that indicates that function is unimplemented..."

From this my code [I think] would be fine as the W versions don't get called on 9.x platforms.

Any comments appreciated though...

>but it would probably be a good idea to test.

Yes. I think an old friend of mine may still have 98 laptop somewhere. May give him a call.


>Could you create a unicode and non-unicode version?

I'm kinda writing my own library for use in my own apps. My first project is going to be a DLL. I don't really want to have to tell other developers to use different DLLs for different platforms.


>you could also consider delay loading.

Thanks for the link. Looks interesting.

Cheers