Mismatch in the definitions of fputwc (), putwc () and putwchar () in glibc

Question

Why fputwc(), putwc()and putwchar()take a type argument wchar_tinstead wint_t? This contradicts the corresponding narrow functions of the symbols fputc(), putc()and putchar(), which accept int, and not char.

Answer 1

This is because it wchar_tis required to store the value WEOF, and charnot required to store the value EOF.

For the charfunction fputc(), putc()and putchar()must take values ​​that can contain both values ​​in the range unsigned charand EOF, where EOFcan be a negative number therefore a intis required for both of them. ¹

While it wchar_titself should contain a symbol WEOF, as well as the largest locale. ²WEOF represents a value that is inside wchar_tbut does not overlap with any locale. ³

This becomes more confusing due to the names charand wchar_t, you should not see wchar_thow char, but more like int, whose size does not depend on but by the size of the largest locale (and by value WEOF). ⁴

---

¹ putchar, toupper, tolower .. int char?

² ISO/IEC 9899: 201x 7.19.2:

WEOF wint_t, . ( ) , , . , .

³ :

WEOF wint_t , .

⁴ ++ Reference:

wchar_t, , .

++, wchar_t - type (, , <cwchar> ).

c, typedef .