X-Git-Url: https://mattmccutchen.net/rsync/rsync.git/blobdiff_plain/fc96552d26e1d5960bcd85fcba4e1631ba15042d..3d0a159d96d153144fe3bd2a2c59a62ec2a62d89:/lib/wildmatch.c diff --git a/lib/wildmatch.c b/lib/wildmatch.c index 6bf74c7e..f3a1731e 100644 --- a/lib/wildmatch.c +++ b/lib/wildmatch.c @@ -22,7 +22,7 @@ #define CC_EQ(class, len, litmatch) ((len) == sizeof (litmatch)-1 \ && *(class) == *(litmatch) \ - && strncmp(class, litmatch, len) == 0) + && strncmp((char*)class, litmatch, len) == 0) #if defined STDC_HEADERS || !defined isascii # define ISASCII(c) 1 @@ -57,173 +57,312 @@ int wildmatch_iteration_count; #endif -static int domatch(const unsigned char *p, const unsigned char *text) +static int force_lower_case = 0; + +/* Match pattern "p" against the a virtually-joined string consisting + * of "text" and any strings in array "a". */ +static int dowild(const uchar *p, const uchar *text, const uchar*const *a) { - int matched, special; - unsigned char ch, prev; + uchar p_ch; #ifdef WILD_TEST_ITERATIONS wildmatch_iteration_count++; #endif - for ( ; (ch = *p) != '\0'; text++, p++) { - if (*text == '\0' && ch != '*') - return FALSE; - switch (ch) { + for ( ; (p_ch = *p) != '\0'; text++, p++) { + int matched, special; + uchar t_ch, prev_ch; + while ((t_ch = *text) == '\0') { + if (*a == NULL) { + if (p_ch != '*') + return ABORT_ALL; + break; + } + text = *a++; + } + if (force_lower_case && ISUPPER(t_ch)) + t_ch = tolower(t_ch); + switch (p_ch) { case '\\': /* Literal match with following character. Note that the test * in "default" handles the p[1] == '\0' failure case. */ - ch = *++p; + p_ch = *++p; /* FALLTHROUGH */ default: - if (*text != ch) + if (t_ch != p_ch) return FALSE; continue; case '?': /* Match anything but '/'. */ - if (*text == '/') + if (t_ch == '/') return FALSE; continue; case '*': if (*++p == '*') { while (*++p == '*') {} special = TRUE; - } - else + } else special = FALSE; if (*p == '\0') { /* Trailing "**" matches everything. Trailing "*" matches * only if there are no more slash characters. */ - return special? TRUE : strchr(text, '/') == NULL; + if (!special) { + do { + if (strchr((char*)text, '/') != NULL) + return FALSE; + } while ((text = *a++) != NULL); + } + return TRUE; } - for ( ; *text; text++) { - if ((matched = domatch(p, text)) != FALSE) { + while (1) { + if (t_ch == '\0') { + if ((text = *a++) == NULL) + break; + t_ch = *text; + continue; + } + if ((matched = dowild(p, text, a)) != FALSE) { if (!special || matched != ABORT_TO_STARSTAR) return matched; - } - else if (!special && *text == '/') + } else if (!special && t_ch == '/') return ABORT_TO_STARSTAR; + t_ch = *++text; } return ABORT_ALL; case '[': - ch = *++p; + p_ch = *++p; #ifdef NEGATE_CLASS2 - if (ch == NEGATE_CLASS2) - ch = NEGATE_CLASS; + if (p_ch == NEGATE_CLASS2) + p_ch = NEGATE_CLASS; #endif /* Assign literal TRUE/FALSE because of "matched" comparison. */ - special = ch == NEGATE_CLASS? TRUE : FALSE; + special = p_ch == NEGATE_CLASS? TRUE : FALSE; if (special) { /* Inverted character class. */ - ch = *++p; + p_ch = *++p; } - prev = 0; + prev_ch = 0; matched = FALSE; do { - if (!ch) + if (!p_ch) return ABORT_ALL; - if (ch == '\\') { - ch = *++p; - if (!ch) + if (p_ch == '\\') { + p_ch = *++p; + if (!p_ch) return ABORT_ALL; - if (*text == ch) + if (t_ch == p_ch) matched = TRUE; - } - else if (ch == '-' && prev && p[1] && p[1] != ']') { - ch = *++p; - if (ch == '\\') { - ch = *++p; - if (!ch) + } else if (p_ch == '-' && prev_ch && p[1] && p[1] != ']') { + p_ch = *++p; + if (p_ch == '\\') { + p_ch = *++p; + if (!p_ch) return ABORT_ALL; } - if (*text <= ch && *text >= prev) + if (t_ch <= p_ch && t_ch >= prev_ch) matched = TRUE; - ch = 0; /* This makes "prev" get set to 0. */ - } - else if (ch == '[' && p[1] == ':') { - unsigned const char *s = p += 2; + p_ch = 0; /* This makes "prev_ch" get set to 0. */ + } else if (p_ch == '[' && p[1] == ':') { + const uchar *s; int i; - while ((ch = *p) && ch != ']') p++; - if (!ch) + for (s = p += 2; (p_ch = *p) && p_ch != ']'; p++) {} /*SHARED ITERATOR*/ + if (!p_ch) return ABORT_ALL; i = p - s - 1; if (i < 0 || p[-1] != ':') { /* Didn't find ":]", so treat like a normal set. */ p = s - 2; - ch = '['; - if (*text == ch) + p_ch = '['; + if (t_ch == p_ch) matched = TRUE; continue; } if (CC_EQ(s,i, "alnum")) { - if (ISALNUM(*text)) + if (ISALNUM(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "alpha")) { - if (ISALPHA(*text)) + } else if (CC_EQ(s,i, "alpha")) { + if (ISALPHA(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "blank")) { - if (ISBLANK(*text)) + } else if (CC_EQ(s,i, "blank")) { + if (ISBLANK(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "cntrl")) { - if (ISCNTRL(*text)) + } else if (CC_EQ(s,i, "cntrl")) { + if (ISCNTRL(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "digit")) { - if (ISDIGIT(*text)) + } else if (CC_EQ(s,i, "digit")) { + if (ISDIGIT(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "graph")) { - if (ISGRAPH(*text)) + } else if (CC_EQ(s,i, "graph")) { + if (ISGRAPH(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "lower")) { - if (ISLOWER(*text)) + } else if (CC_EQ(s,i, "lower")) { + if (ISLOWER(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "print")) { - if (ISPRINT(*text)) + } else if (CC_EQ(s,i, "print")) { + if (ISPRINT(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "punct")) { - if (ISPUNCT(*text)) + } else if (CC_EQ(s,i, "punct")) { + if (ISPUNCT(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "space")) { - if (ISSPACE(*text)) + } else if (CC_EQ(s,i, "space")) { + if (ISSPACE(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "upper")) { - if (ISUPPER(*text)) + } else if (CC_EQ(s,i, "upper")) { + if (ISUPPER(t_ch)) matched = TRUE; - } - else if (CC_EQ(s,i, "xdigit")) { - if (ISXDIGIT(*text)) + } else if (CC_EQ(s,i, "xdigit")) { + if (ISXDIGIT(t_ch)) matched = TRUE; - } - else /* malformed [:class:] string */ + } else /* malformed [:class:] string */ return ABORT_ALL; - ch = 0; /* This makes "prev" get set to 0. */ - } - else if (*text == ch) + p_ch = 0; /* This makes "prev_ch" get set to 0. */ + } else if (t_ch == p_ch) matched = TRUE; - } while (prev = ch, (ch = *++p) != ']'); - if (matched == special) + } while (prev_ch = p_ch, (p_ch = *++p) != ']'); + if (matched == special || t_ch == '/') return FALSE; continue; } } - return *text == '\0'; + do { + if (*text) + return FALSE; + } while ((text = *a++) != NULL); + + return TRUE; +} + +/* Match literal string "s" against the a virtually-joined string consisting + * of "text" and any strings in array "a". */ +static int doliteral(const uchar *s, const uchar *text, const uchar*const *a) +{ + for ( ; *s != '\0'; text++, s++) { + while (*text == '\0') { + if ((text = *a++) == NULL) + return FALSE; + } + if (*text != *s) + return FALSE; + } + + do { + if (*text) + return FALSE; + } while ((text = *a++) != NULL); + + return TRUE; +} + +/* Return the last "count" path elements from the concatenated string. + * We return a string pointer to the start of the string, and update the + * array pointer-pointer to point to any remaining string elements. */ +static const uchar *trailing_N_elements(const uchar*const **a_ptr, int count) +{ + const uchar*const *a = *a_ptr; + const uchar*const *first_a = a; + + while (*a) + a++; + + while (a != first_a) { + const uchar *s = *--a; + s += strlen((char*)s); + while (--s >= *a) { + if (*s == '/' && !--count) { + *a_ptr = a+1; + return s+1; + } + } + } + + if (count == 1) { + *a_ptr = a+1; + return *a; + } + + return NULL; +} + +/* Match the "pattern" against the "text" string. */ +int wildmatch(const char *pattern, const char *text) +{ + static const uchar *nomore[1]; /* A NULL pointer. */ +#ifdef WILD_TEST_ITERATIONS + wildmatch_iteration_count = 0; +#endif + return dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE; } -/* Find the pattern (p) in the text string (t). */ -int wildmatch(const char *p, const char *t) +/* Match the "pattern" against the forced-to-lower-case "text" string. */ +int iwildmatch(const char *pattern, const char *text) { + static const uchar *nomore[1]; /* A NULL pointer. */ + int ret; #ifdef WILD_TEST_ITERATIONS wildmatch_iteration_count = 0; #endif - return domatch((const unsigned char*)p, (const unsigned char*)t) == TRUE; + force_lower_case = 1; + ret = dowild((const uchar*)pattern, (const uchar*)text, nomore) == TRUE; + force_lower_case = 0; + return ret; +} + +/* Match pattern "p" against the a virtually-joined string consisting + * of all the pointers in array "texts" (which has a NULL pointer at the + * end). The int "where" can be 0 (normal matching), > 0 (match only + * the trailing N slash-separated filename components of "texts"), or < 0 + * (match the "pattern" at the start or after any slash in "texts"). */ +int wildmatch_array(const char *pattern, const char*const *texts, int where) +{ + const uchar *p = (const uchar*)pattern; + const uchar*const *a = (const uchar*const*)texts; + const uchar *text; + int matched; + +#ifdef WILD_TEST_ITERATIONS + wildmatch_iteration_count = 0; +#endif + + if (where > 0) + text = trailing_N_elements(&a, where); + else + text = *a++; + if (!text) + return FALSE; + + if ((matched = dowild(p, text, a)) != TRUE && where < 0 + && matched != ABORT_ALL) { + while (1) { + if (*text == '\0') { + if ((text = (uchar*)*a++) == NULL) + return FALSE; + continue; + } + if (*text++ == '/' && (matched = dowild(p, text, a)) != FALSE + && matched != ABORT_TO_STARSTAR) + break; + } + } + return matched == TRUE; +} + +/* Match literal string "s" against the a virtually-joined string consisting + * of all the pointers in array "texts" (which has a NULL pointer at the + * end). The int "where" can be 0 (normal matching), or > 0 (match + * only the trailing N slash-separated filename components of "texts"). */ +int litmatch_array(const char *string, const char*const *texts, int where) +{ + const uchar *s = (const uchar*)string; + const uchar*const *a = (const uchar* const*)texts; + const uchar *text; + + if (where > 0) + text = trailing_N_elements(&a, where); + else + text = *a++; + if (!text) + return FALSE; + + return doliteral(s, text, a) == TRUE; }