Implication of using -fshort-wchar
c, c++, clang, gcc, wchar
Solution
`-fshort-wchar` is not usable if you want to interact with any part of the standard library or third-party library code using the correct (32-bit) definition of `wchar_t`. The situations where it's viable are very limited and probably pertain mostly to freestanding-implementation/embedded stuff.
In the case of porting the Windows application you're dealing with, I think you should just do a seatch-and-replace to change these uses of `wchar_t` to `WCHAR` (which you're free to define) or `char16_t` (defined by C11, and you can supply your own definition pre-C11).
Problem
While going through the file wchar.h on Mac OS X system, I found that wchar_t equivalent of str functions such as wcscpy, wcscat are poisoned when __cplusplust is not defined and max size of wchar_t is of 2 bytes (by using compiler option -fshort-wchar). It seems that for C program, it does not allow such functions to use if -fshort-wchar is defined. I would like to know that what is the implication of using wchar_t functions when -fshort-wchar is used? You may wonder that why I need to use -fshort-wchar. Because, I am porting an application initially written for Windows where size of wchar_t is two bytes. And that data held in wchar_t string is written on the file and also exchanged between two applications. What is a good way to handle the variability of wchar_t on different platforms? Assumption on Windows is that wchar_t is of 16 bits.