What is the difference between the `A` and `W` functions in the Win32 API?

Question

What is the difference in calling the Win32 API function that have an A character appended to the end as opposed to the W character.

I know it means ASCII and WIDE CHARACTER or Unicode, but what is the difference in the output or the input?

For example, If I call GetDefaultCommConfigA, will it fill my COMMCONFIG structure with ASCII strings instead of WCHAR strings? (Or vice-versa for GetDefaultCommConfigW)

In other words, how do I know what Encoding the string is in, ASCII or UNICODE, it must be by the version of the function I call A or W? Correct?

I have found this question, but I don't think it answers my question.

Solution

The A functions use Ansi (not ASCII) strings as input and output, and the W functions use Unicode string instead (UCS-2 on NT4 and earlier, UTF-16 on W2K and later). Refer to MSDN for more details.

Comments

[BrendanMcK]

+1 - also another thing to keep in mind is that all the structures that are passed as parameters to these also come in A and W versions; so for example RegisterClassExA takes a WNDCLASSEXA which uses ANSI LPCSTRs, and while the W versions use LPCWSTRs - this prevents you from passing WCHARs to the A version, and vice versa - if you try, you'll get a compiler error about the types not matching - a fairly common beginner error is to use a cast to make the error "go away" instead of changing the code to use the right type without a cast.

[yu quan]

How does the function without "A" or "W" postfix work? Say "GetVolumeInformation" function, does it call the GetVolumeInformationA or GetVolumeInformationW?

[Remy Lebeau]

@yuquan kind of. The non-A/W functions are implemented as preprocessor macros that map to the appropriate A/W functions. They do not call the A/W functions, but rather are replaced with them during the preprocessor stage, depending on whether UNICODE is defined or not. So, for example, if UNICODE is defined, a call to GetVolumeInformation() is replaced with a call to GetVolumeInformationW(), otherwise it is replaced with a call to GetVolumeInformationA().

[Mark Ransom]

That macro substitution "feature" is a real pain. If you have your own class function called GetVolumeInformation it will be silently renamed to GetVolumeInformationA or GetVolumeInformationW. Using the IDE to look up the definition will take you to the macro definition, not the function.

[Remy Lebeau]

@MarkRansom Yup, and all because Microsoft won't get with the times and support C++ properly, such as by using function overloads and/or template functions rather than the preprocessor.

[Mark Ransom]

To be fair the Win32 API still has to work from C, especially back when wide-char support was first introduced, so a C++ only solution wasn't possible. But at this point there's no excuse for not creating something better and giving you a way of opting out of the macros.

[Remy Lebeau]

@MarkRansom "the Win32 API still has to work from C" - yes, but that doesn't mean it can't use proper wrappers for C++. It already does that for COM interfaces, so there is no reason why it can't do that for functions, too.

[Mark Ransom]

I think we're in agreement.

[Shayan]

What are ExA and ExW functions? What's the difference between these all?

[Remy Lebeau]

@Shayan The link in my answer, and earlier comments above, explain the A/W semantics. Those apply to ExA/ExW functions as well, which are usually just "Ex"-tended versions of earlier A/W functions.