https://sourceware.org/git/gitweb.cgi?p=newlib-cygwin.git;h=1df19b3caeaee95388a3c4d78559389bbcfc9ad5
commit 1df19b3caeaee95388a3c4d78559389bbcfc9ad5 Author: Corinna Vinschen <[email protected]> AuthorDate: Wed Feb 15 22:34:46 2023 +0100 Commit: Corinna Vinschen <[email protected]> CommitDate: Wed Feb 15 22:34:46 2023 +0100 Cygwin: glob: handle equivalence class expressions Handle [=x=] expressions in range brackets. Use the new is_unicode_equiv() function to perform the check. Signed-off-by: Corinna Vinschen <[email protected]> Diff: --- winsup/cygwin/glob.cc | 42 +++++++++++++++++++++++++++++++----------- 1 file changed, 31 insertions(+), 11 deletions(-) diff --git a/winsup/cygwin/glob.cc b/winsup/cygwin/glob.cc index 1e8eda903e70..c882cd59d2e2 100644 --- a/winsup/cygwin/glob.cc +++ b/winsup/cygwin/glob.cc @@ -158,6 +158,7 @@ typedef char Char; #define M_RNG META('-') #define M_SET META('[') #define M_NAMED META(':') +#define M_EQUIV META('=') #define ismeta(c) (((c)&M_QUOTE) != 0) static int compare(const void *, const void *); @@ -186,7 +187,7 @@ static void qprintf(const char *, Char *); /* Return value is either EOS, COLON, DOT, EQUALS, or LBRACKET if no class expression found. */ static inline Char -check_classes_expr(const Char *&cptr, char *classbuf = NULL, +check_classes_expr(const Char *&cptr, wint_t *classbuf = NULL, size_t classbufsize = 0) { const Char *ctype = NULL; @@ -206,7 +207,7 @@ check_classes_expr(const Char *&cptr, char *classbuf = NULL, if (clen < classbufsize) { for (idx = 0; idx < clen; ++idx) - classbuf[idx] = class_p[idx]; + classbuf[idx] = CHAR(class_p[idx]); classbuf[idx] = '\0'; } else ctype = NULL; @@ -480,6 +481,11 @@ globtilde(const Char *pattern, Char *patbuf, size_t patbuf_len, glob_t *pglob) return patbuf; } +static void +wcitoascii(char *dst, wint_t *src) +{ + while ((*dst++ = *src++)); +} /* * The main glob() routine: compiles the pattern (optionally processing @@ -523,19 +529,30 @@ glob0(const Char *pattern, glob_t *pglob, size_t *limit) *bufnext++ = M_NOT; c = *qpatnext; do { - char cclass[64]; - wctype_t type; + wint_t wclass[64]; Char ctype; - ctype = check_classes_expr(qpatnext, cclass, - sizeof cclass); + ctype = check_classes_expr(qpatnext, wclass, + 64); if (ctype) { - if (ctype == COLON && - (type = wctype (cclass))) { - *bufnext++ = M_NAMED; - *bufnext++ = CHAR (type); + wctype_t type; + + if (ctype == COLON) { + char cclass[64]; + + /* No worries, char classes are + ASCII-only anyway */ + wcitoascii (cclass, wclass); + if ((type = wctype (cclass))) { + *bufnext++ = M_NAMED; + *bufnext++ = CHAR (type); + } + } else if (ctype == EQUALS && + wclass[0] && !wclass[1]) { + *bufnext++ = M_EQUIV; + *bufnext++ = CHAR (wclass[0]); } - /* TODO: [. and [= are ignored yet */ + /* TODO: [. is ignored yet */ qpatnext++; continue; } @@ -858,6 +875,9 @@ match(Char *name, Char *pat, Char *patend) if ((c & M_MASK) == M_NAMED) { if (iswctype (k, *pat++)) ok = 1; + } else if ((c & M_MASK) == M_EQUIV) { + if (is_unicode_equiv (k, *pat++)) + ok = 1; } else if ((*pat & M_MASK) == M_RNG) { if (__collate_load_error ? CCHAR(c) <= CCHAR(k) && CCHAR(k) <= CCHAR(pat[1]) :
