1+ #include "glob.h"
2+ #include <stddef.h>
3+
4+ // This has been borrowed from the dual MIT/GPL licensed glob module from the Linux kernel under the MIT license.
5+ // See https://github.com/torvalds/linux/blob/d1bd5fa07667fcc3e38996ec42aef98761f23039/lib/glob.c
6+
7+ /**
8+ * glob_match - Shell-style pattern matching, like !fnmatch(pat, str, 0)
9+ * @pat: Shell-style pattern to match, e.g. "*.[ch]".
10+ * @str: String to match. The pattern must match the entire string.
11+ *
12+ * Perform shell-style glob matching, returning true (1) if the match
13+ * succeeds, or false (0) if it fails. Equivalent to !fnmatch(@pat, @str, 0).
14+ *
15+ * Pattern metacharacters are ?, *, [ and \.
16+ * (And, inside character classes, !, - and ].)
17+ *
18+ * This is small and simple implementation intended for device blacklists
19+ * where a string is matched against a number of patterns. Thus, it
20+ * does not preprocess the patterns. It is non-recursive, and run-time
21+ * is at most quadratic: strlen(@str)*strlen(@pat).
22+ *
23+ * An example of the worst case is glob_match("*aaaaa", "aaaaaaaaaa");
24+ * it takes 6 passes over the pattern before matching the string.
25+ *
26+ * Like !fnmatch(@pat, @str, 0) and unlike the shell, this does NOT
27+ * treat / or leading . specially; it isn't actually used for pathnames.
28+ *
29+ * Note that according to glob(7) (and unlike bash), character classes
30+ * are complemented by a leading !; this does not support the regex-style
31+ * [^a-z] syntax.
32+ *
33+ * An opening bracket without a matching close is matched literally.
34+ */
35+ #ifdef __cplusplus
36+ extern "C" {
37+ #endif
38+ bool glob_match (char const * pat , char const * str ) {
39+ /*
40+ * Backtrack to previous * on mismatch and retry starting one
41+ * character later in the string. Because * matches all characters
42+ * (no exception for /), it can be easily proved that there's
43+ * never a need to backtrack multiple levels.
44+ */
45+ char const * back_pat = NULL , * back_str = NULL ;
46+
47+ /*
48+ * Loop over each token (character or class) in pat, matching
49+ * it against the remaining unmatched tail of str. Return false
50+ * on mismatch, or true after matching the trailing nul bytes.
51+ */
52+ for (;;) {
53+ unsigned char c = * str ++ ;
54+ unsigned char d = * pat ++ ;
55+
56+ switch (d ) {
57+ case '?' : /* Wildcard: anything but nul */
58+ if (c == '\0' )
59+ return false;
60+ break ;
61+ case '*' : /* Any-length wildcard */
62+ if (* pat == '\0' ) /* Optimize trailing * case */
63+ return true;
64+ back_pat = pat ;
65+ back_str = -- str ; /* Allow zero-length match */
66+ break ;
67+ case '[' : { /* Character class */
68+ bool match = false, inverted = (* pat == '!' );
69+ char const * class = pat + inverted ;
70+ unsigned char a = * class ++ ;
71+
72+ /*
73+ * Iterate over each span in the character class.
74+ * A span is either a single character a, or a
75+ * range a-b. The first span may begin with ']'.
76+ */
77+ do {
78+ unsigned char b = a ;
79+
80+ if (a == '\0' ) /* Malformed */
81+ goto literal ;
82+
83+ if (class [0 ] == '-' && class [1 ] != ']' ) {
84+ b = class [1 ];
85+
86+ if (b == '\0' )
87+ goto literal ;
88+
89+ class += 2 ;
90+ /* Any special action if a > b? */
91+ }
92+ match |= (a <= c && c <= b );
93+ } while ((a = * class ++ ) != ']' );
94+
95+ if (match == inverted )
96+ goto backtrack ;
97+ pat = class ;
98+ } break ;
99+ case '\\' :
100+ d = * pat ++ ;
101+ //fallthrough;
102+ default : /* Literal character */
103+ literal :
104+ if (c == d ) {
105+ if (d == '\0' )
106+ return true;
107+ break ;
108+ }
109+ backtrack :
110+ if (c == '\0' || !back_pat )
111+ return false; /* No point continuing */
112+ /* Try again from last *, one character later in str. */
113+ pat = back_pat ;
114+ str = ++ back_str ;
115+ break ;
116+ }
117+ }
118+ }
119+
120+ #ifdef __cplusplus
121+ }
122+ ;
123+ #endif
0 commit comments