Risks of running on single byte character app on a MBCS OS

Question

I have an MFC application where the Character Set is "Not Set". What are the risks associated with running this application on a OS that has a Multi Byte Character Set code page?

Answer 1

The "Character Set: Not Set" option defines neither _MBCS nor _UNICODE. That means that you're using the *A series of functions. Those will return MBCS strings, even when _MBCS is not defined.

If _MBCS doesn't affect the strings returned from the *A functions, then what does it do? It maps some <tchar.h> tcs* functions to their mbs* versions, instead of the str* or wcs* variants. E.g. without _MBCS, _tcsrev maps to strrev, not _mbsrev.

Therefore, you may not be able to reverse the multi-byte strings you receive from the OS, or edit them otherwise.

Answer 2

The risks are:

1) If you use non-English texts, and the user enters, say, Russian, some Russian characters that are not mapped to MBCS or require a Charset layer, will be retrieved as '?'

2) You need to inform the customer that the App. accepts English-only and no guarantee the App, when entering non-English characters, would accept or retrieve non-English texts correctly.

3) Any texts with foreign language characters may lead to unwanted or undesired conversion of accents and diacritics to something else.