wcsspn / wcscspn / wcspbrk / wcstok (3) - glibc

Split from the chapter glibc: strings, ctype and stdlib arithmetic into the per-function layout; see the method page and the glibc ledger for coverage status.

Declaration and gates

#include <wchar.h>

size_t  wcsspn(const wchar_t *wcs, const wchar_t *accept);
size_t  wcscspn(const wchar_t *wcs, const wchar_t *reject);
wchar_t *wcspbrk(const wchar_t *wcs, const wchar_t *accept);
wchar_t *wcstok(wchar_t *restrict wcs, const wchar_t *restrict delim,
                wchar_t **restrict ptr);

All four are declared unconditionally in /usr/include/wchar.h: wcscspn lines 212-213, wcsspn lines 216-217, wcspbrk lines 220-226, wcstok lines 242-244. No feature-test gates - unlike the narrow strtok, wcstok has had its context pointer since C99.

Semantics

wcsspn returns the length of the longest initial segment of wcs made entirely of characters from accept (wcsspn(3)); wcscspn the complement - the initial segment containing none of reject - i.e. the offset of the first rejected character, or wcslen(wcs) if none (wcscspn(3)). wcspbrk returns a pointer to the first occurrence of any accept character, NULL if none (wcspbrk(3)). wcstok splits wcs into tokens: it skips leading delimiters, terminates the token by overwriting the next delimiter with L’\0’, and keeps its place in the caller’s *ptr; NULL wcs continues from there, and NULL is returned when no tokens remain (wcstok(3)).

Gotchas and portability

• wcstok is not strtok-shaped: the third argument is mandatory and there is no context-free variant - code translated from strtok will not compile until the state pointer is threaded through, which is precisely what makes it MT-Safe where strtok is MT-Unsafe race:strtok (see strtok-strtok_r-strsep.md).
• Like strtok, wcstok destructively modifies the input - delimiters become L’\0’ - so it cannot tokenize constant strings (wcstok(3) NOTES).
• The span functions return counts, wcspbrk returns a pointer: wcschr(wcs, ...) style NULL checks on the span results are a type error waiting to happen.
• musl 1.2.5 implements all four (musl src/string/wcsspn.c, wcscspn.c, wcspbrk.c, wcstok.c).

References

• glibc manual: string.texi, nodes Search Functions, Finding Tokens in a String
• man3: wcsspn(3) (covers wcscspn context), wcspbrk(3), wcstok(3)
• header: /usr/include/wchar.h
• symbols: nm -D --with-symbol-versions libc.so.6