1 /* $Revision: 0.2.18.1 $
3 ** Do shell-style pattern matching for ?, \, [], and * characters.
4 ** Might not be robust in face of malformed patterns; e.g., "foo[a-"
5 ** could cause a segmentation violation. It is 8bit clean.
7 ** Written by Rich $alz, mirror!rs, Wed Nov 26 19:03:17 EST 1986.
8 ** Rich $alz is now <rsalz@osf.org>.
9 ** April, 1991: Replaced mutually-recursive calls with in-line code
10 ** for the star character.
12 ** Special thanks to Lars Mathiesen <thorinn@diku.dk> for the ABORT code.
13 ** This can greatly speed up failing wildcard patterns. For example:
14 ** pattern: -*-*-*-*-*-*-12-*-*-*-m-*-*-*
15 ** text 1: -adobe-courier-bold-o-normal--12-120-75-75-m-70-iso8859-1
16 ** text 2: -adobe-courier-bold-o-normal--12-120-75-75-X-70-iso8859-1
17 ** Text 1 matches with 51 calls, while text 2 fails with 54 calls. Without
18 ** the ABORT code, it takes 22310 calls to fail. Ugh. The following
19 ** explanation is from Lars:
20 ** The precondition that must be fulfilled is that DoMatch will consume
21 ** at least one character in text. This is true if *p is neither '*' nor
22 ** '\0'.) The last return has ABORT instead of FALSE to avoid quadratic
23 ** behaviour in cases like pattern "*a*b*c*d" with text "abcxxxxx". With
24 ** FALSE, each star-loop has to run to the end of the text; with ABORT
25 ** only the last one does.
27 ** Once the control of one instance of DoMatch enters the star-loop, that
28 ** instance will return either TRUE or ABORT, and any calling instance
29 ** will therefore return immediately after (without calling recursively
30 ** again). In effect, only one star-loop is ever active. It would be
31 ** possible to modify the code to maintain this context explicitly,
32 ** eliminating all recursive calls at the cost of some complication and
33 ** loss of clarity (and the ABORT stuff seems to be unclear enough by
34 ** itself). I think it would be unwise to try to get this into a
35 ** released version unless you have a good test data base to try it out
50 /* What character marks an inverted character class? */
51 #define NEGATE_CLASS '^'
52 /* Is "*" a common pattern? */
53 #define OPTIMIZE_JUST_STAR
54 /* Do tar(1) matching rules, which ignore a trailing slash? */
55 #undef MATCH_TAR_PATTERN
59 ** Match text and p, return TRUE, FALSE, or ABORT.
70 for ( ; *p; text++, p++) {
71 if (*text == '\0' && *p != '*')
75 /* Literal match with following character. */
79 if (toupper (*text) != toupper (*p))
87 /* Consecutive stars act just like one. */
90 /* Trailing star matches everything. */
93 if ((matched = DoMatch(text++, p)) != FALSE)
97 reverse = p[1] == NEGATE_CLASS ? TRUE : FALSE;
99 /* Inverted character class. */
102 if (p[1] == ']' || p[1] == '-')
103 if (toupper (*++p) == toupper(*text))
105 for (last = *p; *++p && *p != ']'; last = *p)
106 /* This next line requires a good C compiler. */
107 if (*p == '-' && p[1] != ']'
108 ? *text <= *++p && *text >= last
109 : toupper (*text) == toupper (*p))
111 if (matched == reverse)
117 #ifdef MATCH_TAR_PATTERN
120 #endif /* MATCH_TAR_ATTERN */
121 return *text == '\0';
126 ** User-level routine. Returns TRUE or FALSE.
133 #ifdef OPTIMIZE_JUST_STAR
134 if (p[0] == '*' && p[1] == '\0')
136 #endif /* OPTIMIZE_JUST_STAR */
137 return DoMatch(text, p) == TRUE;
145 /* Yes, we use gets not fgets. Sue me. */
155 printf("Wildmat tester. Enter pattern, then strings to test.\n");
156 printf("A blank line gets prompts for a new pattern; a blank pattern\n");
157 printf("exits the program.\n");
160 printf("\nEnter pattern: ");
161 (void)fflush(stdout);
162 if (gets(p) == NULL || p[0] == '\0')
165 printf("Enter text: ");
166 (void)fflush(stdout);
167 if (gets(text) == NULL)
170 /* Blank line; go back and get a new pattern. */
172 printf(" %s\n", wildmat(text, p) ? "YES" : "NO");
179 #endif /* defined(TEST) */