| Message ID | a64ad0a4-ea0f-4de3-aa2b-e377d3bef85d@inlv.org (mailing list archive) |
|---|---|
| State | Accepted |
| Delegated to: | Herbert Xu |
| Headers | show |
| Series | [v2] fix build on systems without memrchr(3) | expand |
Martijn Dekker <martijn@inlv.org> wrote: > > Op 22-06-2024 om 15:25 schreef Martijn Dekker: >> memrchr(3) is non-standard, and has been ported from glibc to FreeBSD, NetSBD >> and OpenBSD, but not to macOS, at least as of 12.7.5. So we need a test for >> it. As far as I can tell, *name is a zero-terminated C string, so it should >> work to use strrchr(3) as a fallback. > > Reading the code more closely, that's nonsense, because 'p' does not point to > the end of the string if metacharacters are found. > > Guess the best we can do is provide a simple local fallback implementation of > memrchr(3). Patch v2 attached. Patch applied. Thanks.
diff --git a/configure.ac b/configure.ac index 338d5bd..ba4856a 100644 --- a/configure.ac +++ b/configure.ac @@ -87,7 +87,7 @@ AC_CHECK_DECL([PRIdMAX],, dnl Checks for library functions. AC_CHECK_FUNCS(bsearch faccessat getpwnam getrlimit isalpha killpg \ - memfd_create mempcpy \ + memfd_create memrchr mempcpy \ sigsetmask stpcpy strchrnul strsignal strtod strtoimax \ strtoumax sysconf tee) diff --git a/src/expand.c b/src/expand.c index 6912e39..e3056a0 100644 --- a/src/expand.c +++ b/src/expand.c @@ -1633,6 +1633,17 @@ static char *expmeta_rmescapes(char *enddir, const char *name) return enddir - 1; } +#ifndef HAVE_MEMRCHR +static void *memrchr(const void *s, int c, size_t n) +{ + const unsigned char *str = (const unsigned char *)s; + for (const unsigned char *cp = str + n - 1; cp >= str; cp--) + if (*cp == c) + return (void *)cp; + return NULL; +} +#endif + /* * Do metacharacter (i.e. *, ?, [...]) expansion. */