OpenBSD manual page server

Manual Page Search Parameters
MBRTOWC(3) Library Functions Manual MBRTOWC(3)

mbrtowc, mbrtoc32 — convert a multibyte character to a wide character (restartable)

#include <wchar.h>

size_t
mbrtowc(wchar_t * restrict wc, const char * restrict s, size_t n, mbstate_t * restrict mbs);

#include <uchar.h>

size_t
mbrtoc32(char32_t * restrict wc, const char * restrict s, size_t n, mbstate_t * restrict mbs);

The () and () functions examine at most n bytes of the multibyte character byte string pointed to by s, convert those bytes to a wide character, and store the wide character into *wc if wc is not NULL and s points to a valid character.

Conversion happens in accordance with the conversion state *mbs, which must be initialized to zero before the application's first call to () or (). If the previous call did not return (size_t)-1, mbs can safely be reused without reinitialization.

The input encoding that () and () use for s is determined by the LC_CTYPE category of the current locale. If the locale is changed without reinitialization of *mbs, the behaviour is undefined.

Unlike mbtowc(3), () and () accept an incomplete byte sequence pointed to by s which does not form a complete character but is potentially part of a valid character. In this case, both functions consume all such bytes. The conversion state saved in *mbs will be used to restart the suspended conversion during the next call.

On systems other than OpenBSD that support state-dependent encodings, s may point to a special sequence of bytes called a “shift sequence”. Shift sequences switch between character code sets available within an encoding scheme. One encoding scheme using shift sequences is ISO/IEC 2022-JP, which can switch e.g. from ASCII (which uses one byte per character) to JIS X 0208 (which uses two bytes per character). Shift sequence bytes correspond to no individual wide character, so () and () treat them as if they were part of the subsequent multibyte character. Therefore they do contribute to the number of bytes in the multibyte character.

The following arguments cause special processing:

wc == NULL
The conversion from a multibyte character to a wide character is performed and the conversion state may be affected, but the resulting wide character is discarded. This can be used to find out how many bytes are contained in the multibyte character pointed to by s.
s == NULL
The arguments wc and n are ignored and starting or continuing the conversion with an empty string is attempted, discarding the conversion result. If conversion succeeds, this call always returns zero. Unlike mbtowc(3), the value returned does not indicate whether the current encoding of the locale is state-dependent, i.e. uses shift sequences.
mbs == NULL
() and () each use their own internal state object instead of the mbs argument. Both internal state objects are initialized at startup time of the program, and no other libc function ever changes either of them.

If () or () is called with a NULL mbs argument and that call returns (size_t)-1, the internal conversion state of the respective function becomes permanently undefined and there is no way to reset it to any defined state. Consequently, after such a mishap, it is not safe to call the same function with a NULL mbs argument ever again until the program is terminated.

0
The bytes pointed to by s form a terminating NUL character. If wc is not NULL, a NUL wide character has been stored in the wchar_t object pointed to by wc.
positive
s points to a valid character, and the value returned is the number of bytes completing the character. If wc is not NULL, the corresponding wide character has been stored in the wchar_t object pointed to by wc.
(size_t)-1
s points to an illegal byte sequence which does not form a valid multibyte character in the current locale, or mbs points to an invalid or uninitialized object. errno is set to EILSEQ or EINVAL, respectively. The conversion state object pointed to by mbs is left in an undefined state and must be reinitialized before being used again.

Because applications using mbrtowc() or mbrtoc32() are shielded from the specifics of the multibyte character encoding scheme, it is impossible to repair byte sequences containing encoding errors. Such byte sequences must be treated as invalid and potentially malicious input. Applications must stop processing the byte string pointed to by s and either discard any wide characters already converted, or cope with truncated input.

(size_t)-2
s points to an incomplete byte sequence of length n which has been consumed and contains part of a valid multibyte character. The character may be completed by calling the same function again with s pointing to one or more subsequent bytes of the multibyte character and mbs pointing to the conversion state object used during conversion of the incomplete byte sequence.
(size_t)-3
The next character resulting from a previous call has been stored into wc, without consuming any additional bytes from s. This never happens for mbrtowc(), and on OpenBSD, it never happens for mbrtoc32() either.

mbrtowc() and mbrtoc32() cause an error in the following cases:

[]
s points to an invalid multibyte character.
[]
mbs points to an invalid or uninitialized mbstate_t object.

mbrlen(3), mbtowc(3), setlocale(3), wcrtomb(3)

mbrtowc() conforms to ISO/IEC 9899/AMD1:1995 (“ISO C90, Amendment 1”). The restrict qualifier was added at ISO/IEC 9899:1999 (“ISO C99”).

mbrtoc32() conforms to ISO/IEC 9899:2011 (“ISO C11”).

mbrtowc() has been available since OpenBSD 3.8 and has provided support for UTF-8 since OpenBSD 4.8.

mbrtoc32() has been available since OpenBSD 7.4.

mbrtowc() and mbrtoc32() are not suitable for programs that care about internals of the character encoding scheme used by the byte string pointed to by s.

It is possible that these functions fail because of locale configuration errors. An “invalid” character sequence may simply be encoded in a different encoding than that of the current locale.

The special cases for s == NULL and mbs == NULL do not make any sense. Instead of passing NULL for mbs, mbtowc(3) can be used.

Earlier versions of this man page implied that calling mbrtowc() with a NULL s argument would always set mbs to the initial conversion state. But this is true only if the previous call to mbrtowc() using mbs did not return (size_t)-1 or (size_t)-2. It is recommended to zero the mbstate_t object instead.

September 12, 2023 OpenBSD-current