exclude.c

Rsync 3.0.5 source code

/* Each time rsync changes to a new directory it call this function to
 * handle all the per-dir merge-files.  The "dir" value is the current path
 * relative to curr_dir (which might not be null-terminated).  We copy it
 * into dirbuf so that we can easily append a file name on the end. */
void *push_local_filters(const char *dir, unsigned int dirlen)
{
	struct filter_list_struct *ap, *push;
	int i;

	set_filter_dir(dir, dirlen);

	if (!mergelist_cnt)
		return NULL;

	push = new_array(struct filter_list_struct, mergelist_cnt);
	if (!push)
		out_of_memory("push_local_filters");

	for (i = 0, ap = push; i < mergelist_cnt; i++) {
		memcpy(ap++, mergelist_parents[i]->u.mergelist,
		       sizeof (struct filter_list_struct));
	}

	/* Note: parse_filter_file() might increase mergelist_cnt, so keep
	 * this loop separate from the above loop. */
	for (i = 0; i < mergelist_cnt; i++) {
		struct filter_struct *ex = mergelist_parents[i];
		struct filter_list_struct *lp = ex->u.mergelist;

		if (verbose > 2) {
			rprintf(FINFO, "[%s] pushing filter list%s\n",
				who_am_i(), lp->debug_type);
		}

		lp->tail = NULL; /* Switch any local rules to inherited. */
		if (ex->match_flags & MATCHFLG_NO_INHERIT)
			lp->head = NULL;

		if (ex->match_flags & MATCHFLG_FINISH_SETUP) {
			ex->match_flags &= ~MATCHFLG_FINISH_SETUP;
			if (setup_merge_file(ex, lp))
				set_filter_dir(dir, dirlen);
		}

		if (strlcpy(dirbuf + dirbuf_len, ex->pattern,
		    MAXPATHLEN - dirbuf_len) < MAXPATHLEN - dirbuf_len) {
			parse_filter_file(lp, dirbuf, ex->match_flags,
					  XFLG_ANCHORED2ABS);
		} else {
			io_error |= IOERR_GENERAL;
			rprintf(FERROR,
			    "cannot add local filter rules in long-named directory: %s\n",
			    full_fname(dirbuf));
		}
		dirbuf[dirbuf_len] = '\0';
	}

	return (void*)push;
}

void pop_local_filters(void *mem)
{
	struct filter_list_struct *ap, *pop = (struct filter_list_struct*)mem;
	int i;

	for (i = mergelist_cnt; i-- > 0; ) {
		struct filter_struct *ex = mergelist_parents[i];
		struct filter_list_struct *lp = ex->u.mergelist;

		if (verbose > 2) {
			rprintf(FINFO, "[%s] popping filter list%s\n",
				who_am_i(), lp->debug_type);
		}

		clear_filter_list(lp);
	}

	if (!pop)
		return;

	for (i = 0, ap = pop; i < mergelist_cnt; i++) {
		memcpy(mergelist_parents[i]->u.mergelist, ap++,
		       sizeof (struct filter_list_struct));
	}

	free(pop);
}

void change_local_filter_dir(const char *dname, int dlen, int dir_depth)
{
	static int cur_depth = -1;
	static void *filt_array[MAXPATHLEN/2+1];

	if (!dname) {
		for ( ; cur_depth >= 0; cur_depth--) {
			if (filt_array[cur_depth]) {
				pop_local_filters(filt_array[cur_depth]);
				filt_array[cur_depth] = NULL;
			}
		}
		return;
	}

	assert(dir_depth < MAXPATHLEN/2+1);

	for ( ; cur_depth >= dir_depth; cur_depth--) {
		if (filt_array[cur_depth]) {
			pop_local_filters(filt_array[cur_depth]);
			filt_array[cur_depth] = NULL;
		}
	}

	cur_depth = dir_depth;
	filt_array[cur_depth] = push_local_filters(dname, dlen);
}

static int rule_matches(const char *fname, struct filter_struct *ex, int name_is_dir)
{
	int slash_handling, str_cnt = 0, anchored_match = 0;
	int ret_match = ex->match_flags & MATCHFLG_NEGATE ? 0 : 1;
	char *p, *pattern = ex->pattern;
	const char *strings[16]; /* more than enough */
	const char *name = fname + (*fname == '/');

	if (!*name)
		return 0;

	if (!ex->u.slash_cnt && !(ex->match_flags & MATCHFLG_WILD2)) {
		/* If the pattern does not have any slashes AND it does
		 * not have a "**" (which could match a slash), then we
		 * just match the name portion of the path. */
		if ((p = strrchr(name,'/')) != NULL)
			name = p+1;
	} else if (ex->match_flags & MATCHFLG_ABS_PATH && *fname != '/'
	    && curr_dir_len > module_dirlen + 1) {
		/* If we're matching against an absolute-path pattern,
		 * we need to prepend our full path info. */
		strings[str_cnt++] = curr_dir + module_dirlen + 1;
		strings[str_cnt++] = "/";
	} else if (ex->match_flags & MATCHFLG_WILD2_PREFIX && *fname != '/') {
		/* Allow "**"+"/" to match at the start of the string. */
		strings[str_cnt++] = "/";
	}
	strings[str_cnt++] = name;
	if (name_is_dir) {
		/* Allow a trailing "/"+"***" to match the directory. */
		if (ex->match_flags & MATCHFLG_WILD3_SUFFIX)
			strings[str_cnt++] = "/";
	} else if (ex->match_flags & MATCHFLG_DIRECTORY)
		return !ret_match;
	strings[str_cnt] = NULL;

	if (*pattern == '/') {
		anchored_match = 1;
		pattern++;
	}

	if (!anchored_match && ex->u.slash_cnt
	    && !(ex->match_flags & MATCHFLG_WILD2)) {
		/* A non-anchored match with an infix slash and no "**"
		 * needs to match the last slash_cnt+1 name elements. */
		slash_handling = ex->u.slash_cnt + 1;
	} else if (!anchored_match && !(ex->match_flags & MATCHFLG_WILD2_PREFIX)
				   && ex->match_flags & MATCHFLG_WILD2) {
		/* A non-anchored match with an infix or trailing "**" (but not
		 * a prefixed "**") needs to try matching after every slash. */
		slash_handling = -1;
	} else {
		/* The pattern matches only at the start of the path or name. */
		slash_handling = 0;
	}

	if (ex->match_flags & MATCHFLG_WILD) {
		if (wildmatch_array(pattern, strings, slash_handling))
			return ret_match;
	} else if (str_cnt > 1) {
		if (litmatch_array(pattern, strings, slash_handling))
			return ret_match;
	} else if (anchored_match) {
		if (strcmp(name, pattern) == 0)
			return ret_match;
	} else {
		int l1 = strlen(name);
		int l2 = strlen(pattern);
		if (l2 <= l1 &&
		    strcmp(name+(l1-l2),pattern) == 0 &&
		    (l1==l2 || name[l1-(l2+1)] == '/')) {
			return ret_match;
		}
	}

	return !ret_match;
}


static void report_filter_result(enum logcode code, char const *name,
                                 struct filter_struct const *ent,
                                 int name_is_dir, const char *type)
{
	/* If a trailing slash is present to match only directories,
	 * then it is stripped out by add_rule().  So as a special
	 * case we add it back in here. */

	if (verbose >= 2) {
		static char *actions[2][2]
		    = { {"show", "hid"}, {"risk", "protect"} };
		const char *w = who_am_i();
		rprintf(code, "[%s] %sing %s %s because of pattern %s%s%s\n",
		    w, actions[*w!='s'][!(ent->match_flags&MATCHFLG_INCLUDE)],
		    name_is_dir ? "directory" : "file", name, ent->pattern,
		    ent->match_flags & MATCHFLG_DIRECTORY ? "/" : "", type);
	}
}


/*
 * Return -1 if file "name" is defined to be excluded by the specified
 * exclude list, 1 if it is included, and 0 if it was not matched.
 */
int check_filter(struct filter_list_struct *listp, enum logcode code,
		 const char *name, int name_is_dir)
{
	struct filter_struct *ent;

	for (ent = listp->head; ent; ent = ent->next) {
		if (ignore_perishable && ent->match_flags & MATCHFLG_PERISHABLE)
			continue;
		if (ent->match_flags & MATCHFLG_PERDIR_MERGE) {
			int rc = check_filter(ent->u.mergelist, code, name,
					      name_is_dir);
			if (rc)
				return rc;
			continue;
		}
		if (ent->match_flags & MATCHFLG_CVS_IGNORE) {
			int rc = check_filter(&cvs_filter_list, code, name,
					      name_is_dir);
			if (rc)
				return rc;
			continue;
		}
		if (rule_matches(name, ent, name_is_dir)) {
			report_filter_result(code, name, ent, name_is_dir,
					     listp->debug_type);
			return ent->match_flags & MATCHFLG_INCLUDE ? 1 : -1;
		}
	}

	return 0;
}

#define RULE_STRCMP(s,r) rule_strcmp((s), (r), sizeof (r) - 1)

static const uchar *rule_strcmp(const uchar *str, const char *rule, int rule_len)
{
	if (strncmp((char*)str, rule, rule_len) != 0)
		return NULL;
	if (isspace(str[rule_len]) || str[rule_len] == '_' || !str[rule_len])
		return str + rule_len - 1;
	if (str[rule_len] == ',')
		return str + rule_len;
	return NULL;
}

/* Get the next include/exclude arg from the string.  The token will not
 * be '\0' terminated, so use the returned length to limit the string.
 * Also, be sure to add this length to the returned pointer before passing
 * it back to ask for the next token.  This routine parses the "!" (list-
 * clearing) token and (depending on the mflags) the various prefixes.
 * The *mflags_ptr value will be set on exit to the new MATCHFLG_* bits
 * for the current token. */
static const char *parse_rule_tok(const char *p, uint32 mflags, int xflags,
				  unsigned int *len_ptr, uint32 *mflags_ptr)
{
	const uchar *s = (const uchar *)p;
	uint32 new_mflags;
	unsigned int len;

	if (mflags & MATCHFLG_WORD_SPLIT) {
		/* Skip over any initial whitespace. */
		while (isspace(*s))
			s++;
		/* Update to point to real start of rule. */
		p = (const char *)s;
	}
	if (!*s)
		return NULL;

	new_mflags = mflags & MATCHFLGS_FROM_CONTAINER;

	/* Figure out what kind of a filter rule "s" is pointing at.  Note
	 * that if MATCHFLG_NO_PREFIXES is set, the rule is either an include
	 * or an exclude based on the inheritance of the MATCHFLG_INCLUDE
	 * flag (above).  XFLG_OLD_PREFIXES indicates a compatibility mode
	 * for old include/exclude patterns where just "+ " and "- " are
	 * allowed as optional prefixes.  */
	if (mflags & MATCHFLG_NO_PREFIXES) {
		if (*s == '!' && mflags & MATCHFLG_CVS_IGNORE)
			new_mflags |= MATCHFLG_CLEAR_LIST; /* Tentative! */
	} else if (xflags & XFLG_OLD_PREFIXES) {
		if (*s == '-' && s[1] == ' ') {
			new_mflags &= ~MATCHFLG_INCLUDE;
			s += 2;
		} else if (*s == '+' && s[1] == ' ') {
			new_mflags |= MATCHFLG_INCLUDE;
			s += 2;
		} else if (*s == '!')
			new_mflags |= MATCHFLG_CLEAR_LIST; /* Tentative! */
	} else {
		char ch = 0, *mods = "";
		switch (*s) {
		case 'c':
			if ((s = RULE_STRCMP(s, "clear")) != NULL)
				ch = '!';
			break;
		case 'd':
			if ((s = RULE_STRCMP(s, "dir-merge")) != NULL)
				ch = ':';
			break;
		case 'e':
			if ((s = RULE_STRCMP(s, "exclude")) != NULL)
				ch = '-';
			break;
		case 'h':
			if ((s = RULE_STRCMP(s, "hide")) != NULL)
				ch = 'H';
			break;
		case 'i':
			if ((s = RULE_STRCMP(s, "include")) != NULL)
				ch = '+';
			break;
		case 'm':
			if ((s = RULE_STRCMP(s, "merge")) != NULL)
				ch = '.';
			break;
		case 'p':
			if ((s = RULE_STRCMP(s, "protect")) != NULL)
				ch = 'P';
			break;
		case 'r':
			if ((s = RULE_STRCMP(s, "risk")) != NULL)
				ch = 'R';
			break;
		case 's':
			if ((s = RULE_STRCMP(s, "show")) != NULL)
				ch = 'S';
			break;
		default:
			ch = *s;
			if (s[1] == ',')
				s++;
			break;
		}
		switch (ch) {
		case ':':
			new_mflags |= MATCHFLG_PERDIR_MERGE
				    | MATCHFLG_FINISH_SETUP;
			/* FALL THROUGH */
		case '.':
			new_mflags |= MATCHFLG_MERGE_FILE;
			mods = MODIFIERS_INCL_EXCL MODIFIERS_MERGE_FILE;
			break;
		case '+':
			new_mflags |= MATCHFLG_INCLUDE;
			/* FALL THROUGH */
		case '-':
			mods = MODIFIERS_INCL_EXCL;
			break;
		case 'S':
			new_mflags |= MATCHFLG_INCLUDE;
			/* FALL THROUGH */
		case 'H':
			new_mflags |= MATCHFLG_SENDER_SIDE;
			mods = MODIFIERS_HIDE_PROTECT;
			break;
		case 'R':
			new_mflags |= MATCHFLG_INCLUDE;
			/* FALL THROUGH */
		case 'P':
			new_mflags |= MATCHFLG_RECEIVER_SIDE;
			mods = MODIFIERS_HIDE_PROTECT;
			break;
		case '!':
			new_mflags |= MATCHFLG_CLEAR_LIST;
			mods = NULL;
			break;
		default:
			rprintf(FERROR, "Unknown filter rule: `%s'\n", p);
			exit_cleanup(RERR_SYNTAX);
		}
		while (mods && *++s && *s != ' ' && *s != '_') {
			if (strchr(mods, *s) == NULL) {
				if (mflags & MATCHFLG_WORD_SPLIT && isspace(*s)) {
					s--;
					break;
				}
			    invalid:
				rprintf(FERROR,
					"invalid modifier sequence at '%c' in filter rule: %s\n",
					*s, p);
				exit_cleanup(RERR_SYNTAX);
			}
			switch (*s) {
			case '-':
				if (new_mflags & MATCHFLG_NO_PREFIXES)
				    goto invalid;
				new_mflags |= MATCHFLG_NO_PREFIXES;
				break;
			case '+':
				if (new_mflags & MATCHFLG_NO_PREFIXES)
				    goto invalid;
				new_mflags |= MATCHFLG_NO_PREFIXES
					    | MATCHFLG_INCLUDE;
				break;
			case '/':
				new_mflags |= MATCHFLG_ABS_PATH;
				break;
			case '!':
				new_mflags |= MATCHFLG_NEGATE;
				break;
			case 'C':
				if (new_mflags & MATCHFLG_NO_PREFIXES)
				    goto invalid;
				new_mflags |= MATCHFLG_NO_PREFIXES
					    | MATCHFLG_WORD_SPLIT
					    | MATCHFLG_NO_INHERIT
					    | MATCHFLG_CVS_IGNORE;
				break;