-/* -*- c-file-style: "linux" -*-
+/*
+ * The filter include/exclude routines.
*
- * Copyright (C) 1996-2001 by Andrew Tridgell <tridge@samba.org>
- * Copyright (C) 1996 by Paul Mackerras
- * Copyright (C) 2002 by Martin Pool
+ * Copyright (C) 1996-2001 Andrew Tridgell <tridge@samba.org>
+ * Copyright (C) 1996 Paul Mackerras
+ * Copyright (C) 2002 Martin Pool
+ * Copyright (C) 2003, 2004, 2005, 2006 Wayne Davison
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street - Fifth Floor, Boston, MA 02110-1301, USA.
*/
-/* a lot of this stuff was originally derived from GNU tar, although
- it has now changed so much that it is hard to tell :) */
-
-/* include/exclude cluestick added by Martin Pool <mbp@samba.org> */
-
#include "rsync.h"
extern int verbose;
extern int recurse;
extern int io_error;
extern int local_server;
+extern int prune_empty_dirs;
extern int delete_mode;
extern int delete_excluded;
extern int cvs_exclude;
extern unsigned int module_dirlen;
struct filter_list_struct filter_list = { 0, 0, "" };
-struct filter_list_struct cvs_filter_list = { 0, 0, " [cvsignore]" };
-struct filter_list_struct server_filter_list = { 0, 0, " [server]" };
+struct filter_list_struct cvs_filter_list = { 0, 0, " [global CVS]" };
+struct filter_list_struct server_filter_list = { 0, 0, " [daemon]" };
/* Need room enough for ":MODS " prefix plus some room to grow. */
#define MAX_RULE_PREFIX (16)
#define MODIFIERS_MERGE_FILE "-+Cenw"
-#define MODIFIERS_INCL_EXCL "/!C"
+#define MODIFIERS_INCL_EXCL "/!Crs"
+#define MODIFIERS_HIDE_PROTECT "/!"
/* The dirbuf is set by push_local_filters() to the current subdirectory
* relative to curr_dir that is being processed. The path always has a
listp->debug_type);
}
+ /* These flags also indicate that we're reading a list that
+ * needs to be filtered now, not post-filtered later. */
+ if (xflags & (XFLG_ANCHORED2ABS|XFLG_ABS_IF_SLASH)) {
+ uint32 mf = mflags & (MATCHFLG_RECEIVER_SIDE|MATCHFLG_SENDER_SIDE);
+ if (am_sender) {
+ if (mf == MATCHFLG_RECEIVER_SIDE)
+ return;
+ } else {
+ if (mf == MATCHFLG_SENDER_SIDE)
+ return;
+ }
+ }
+
if (!(ret = new(struct filter_struct)))
out_of_memory("add_rule");
memset(ret, 0, sizeof ret[0]);
- if (xflags & XFLG_ANCHORED2ABS && *pat == '/'
- && !(mflags & (MATCHFLG_ABS_PATH | MATCHFLG_MERGE_FILE))) {
+ if (!(mflags & (MATCHFLG_ABS_PATH | MATCHFLG_MERGE_FILE))
+ && ((xflags & (XFLG_ANCHORED2ABS|XFLG_ABS_IF_SLASH) && *pat == '/')
+ || (xflags & XFLG_ABS_IF_SLASH && strchr(pat, '/') != NULL))) {
mflags |= MATCHFLG_ABS_PATH;
- ex_len = dirbuf_len - module_dirlen - 1;
+ if (*pat == '/')
+ ex_len = dirbuf_len - module_dirlen - 1;
+ else
+ ex_len = 0;
} else
ex_len = 0;
if (!(ret->pattern = new_array(char, ex_len + pat_len + 1)))
/* If the pattern starts with **, note that. */
if (cp == ret->pattern)
mflags |= MATCHFLG_WILD2_PREFIX;
+ /* If the pattern ends with ***, note that. */
+ if (pat_len >= 3
+ && ret->pattern[pat_len-3] == '*'
+ && ret->pattern[pat_len-2] == '*'
+ && ret->pattern[pat_len-1] == '*')
+ mflags |= MATCHFLG_WILD3_SUFFIX;
}
}
strlcpy(to, merge_file, *len_ptr + 1);
merge_file = to;
}
- if (!sanitize_path(fn, merge_file, r, dirbuf_depth)) {
+ if (!sanitize_path(fn, merge_file, r, dirbuf_depth, NULL)) {
rprintf(FERROR, "merge-file name overflows: %s\n",
- safe_fname(merge_file));
+ merge_file);
return NULL;
}
+ fn_len = strlen(fn);
} else {
strlcpy(fn, merge_file, len_ptr ? *len_ptr + 1 : MAXPATHLEN);
- clean_fname(fn, 1);
+ fn_len = clean_fname(fn, 1);
}
-
- fn_len = strlen(fn);
- if (fn == buf)
- goto done;
-
- if (dirbuf_len + fn_len >= MAXPATHLEN) {
- rprintf(FERROR, "merge-file name overflows: %s\n",
- safe_fname(fn));
- return NULL;
+
+ /* If the name isn't in buf yet, it's wasn't absolute. */
+ if (fn != buf) {
+ if (dirbuf_len + fn_len >= MAXPATHLEN) {
+ rprintf(FERROR, "merge-file name overflows: %s\n", fn);
+ return NULL;
+ }
+ memcpy(buf, dirbuf + prefix_skip, dirbuf_len - prefix_skip);
+ memcpy(buf + dirbuf_len - prefix_skip, fn, fn_len + 1);
+ fn_len = clean_fname(buf, 1);
}
- memcpy(buf, dirbuf + prefix_skip, dirbuf_len - prefix_skip);
- memcpy(buf + dirbuf_len - prefix_skip, fn, fn_len + 1);
- fn_len = clean_fname(buf, 1);
- done:
if (len_ptr)
*len_ptr = fn_len;
return buf;
static int rule_matches(char *name, struct filter_struct *ex, int name_is_dir)
{
- char *p, full_name[MAXPATHLEN];
- int match_start = 0;
+ int slash_handling, str_cnt = 0, anchored_match = 0;
int ret_match = ex->match_flags & MATCHFLG_NEGATE ? 0 : 1;
- char *pattern = ex->pattern;
+ char *p, *pattern = ex->pattern;
+ const char *strings[16]; /* more than enough */
+ if (*name == '/')
+ name++;
if (!*name)
return 0;
- /* 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 (!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 && *name != '/'
+ } else if (ex->match_flags & MATCHFLG_ABS_PATH && *name != '/'
&& curr_dir_len > module_dirlen + 1) {
- pathjoin(full_name, sizeof full_name,
- curr_dir + module_dirlen + 1, name);
- name = full_name;
+ /* 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 && *name != '/') {
+ /* Allow "**"+"/" to match at the start of the string. */
+ strings[str_cnt++] = "/";
}
-
- if (ex->match_flags & MATCHFLG_DIRECTORY && !name_is_dir)
+ 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 == '/') {
- match_start = 1;
+ anchored_match = 1;
pattern++;
- if (*name == '/')
- name++;
}
- if (ex->match_flags & MATCHFLG_WILD) {
+ 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. */
- if (!match_start && ex->u.slash_cnt
- && !(ex->match_flags & MATCHFLG_WILD2)) {
- int cnt = ex->u.slash_cnt + 1;
- for (p = name + strlen(name) - 1; p >= name; p--) {
- if (*p == '/' && !--cnt)
- break;
- }
- name = p+1;
- }
- if (wildmatch(pattern, name))
+ 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;
- if (ex->match_flags & MATCHFLG_WILD2_PREFIX) {
- /* If the **-prefixed pattern has a '/' as the next
- * character, then try to match the rest of the
- * pattern at the root. */
- if (pattern[2] == '/' && wildmatch(pattern+3, name))
- return ret_match;
- }
- else if (!match_start && 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. */
- while ((name = strchr(name, '/')) != NULL) {
- name++;
- if (wildmatch(pattern, name))
- return ret_match;
- }
- }
- } else if (match_start) {
- if (strcmp(name,pattern) == 0)
+ } 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);
* case we add it back in here. */
if (verbose >= 2) {
- rprintf(FINFO, "[%s] %scluding %s %s because of pattern %s%s%s\n",
- who_am_i(),
- ent->match_flags & MATCHFLG_INCLUDE ? "in" : "ex",
- name_is_dir ? "directory" : "file", name, ent->pattern,
- ent->match_flags & MATCHFLG_DIRECTORY ? "/" : "", type);
+ static char *actions[2][2]
+ = { {"show", "hid"}, {"risk", "protect"} };
+ const char *w = who_am_i();
+ rprintf(FINFO, "[%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 0;
}
-#define RULE_MATCH(s,r) rule_match((s), (r), sizeof (r) - 1)
+#define RULE_STRCMP(s,r) rule_strcmp((s), (r), sizeof (r) - 1)
-static const char *rule_match(const char *str, const char *rule, int rule_len)
+static const uchar *rule_strcmp(const uchar *str, const char *rule, int rule_len)
{
- if (strncmp(str, rule, rule_len) != 0)
+ 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;
} else if (*s == '+' && s[1] == ' ') {
new_mflags |= MATCHFLG_INCLUDE;
s += 2;
- }
- if (*s == '!')
+ } else if (*s == '!')
new_mflags |= MATCHFLG_CLEAR_LIST; /* Tentative! */
} else {
char ch = 0, *mods = "";
switch (*s) {
case 'c':
- if ((s = RULE_MATCH(s, "clear")) != NULL)
+ if ((s = RULE_STRCMP(s, "clear")) != NULL)
ch = '!';
break;
case 'd':
- if ((s = RULE_MATCH(s, "dir-merge")) != NULL)
+ if ((s = RULE_STRCMP(s, "dir-merge")) != NULL)
ch = ':';
break;
case 'e':
- if ((s = RULE_MATCH(s, "exclude")) != NULL)
+ 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_MATCH(s, "include")) != NULL)
+ if ((s = RULE_STRCMP(s, "include")) != NULL)
ch = '+';
break;
case 'm':
- if ((s = RULE_MATCH(s, "merge")) != NULL)
+ 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 '-':
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);
+ rprintf(FERROR, "Unknown filter rule: `%s'\n", p);
exit_cleanup(RERR_SYNTAX);
}
while (mods && *++s && *s != ' ' && *s != '_') {
case 'n':
new_mflags |= MATCHFLG_NO_INHERIT;
break;
+ case 'r':
+ new_mflags |= MATCHFLG_RECEIVER_SIDE;
+ break;
+ case 's':
+ new_mflags |= MATCHFLG_SENDER_SIDE;
+ break;
case 'w':
new_mflags |= MATCHFLG_WORD_SPLIT;
break;
len = strlen((char*)s);
if (new_mflags & MATCHFLG_CLEAR_LIST) {
- if (!(xflags & XFLG_OLD_PREFIXES) && len) {
+ if (!(mflags & MATCHFLG_NO_PREFIXES)
+ && !(xflags & XFLG_OLD_PREFIXES) && len) {
rprintf(FERROR,
"'!' rule has trailing characters: %s\n", p);
exit_cleanup(RERR_SYNTAX);
exit_cleanup(RERR_SYNTAX);
}
+ /* --delete-excluded turns an un-modified include/exclude into a
+ * sender-side rule. We also affect per-dir merge files that take
+ * no prefixes as a simple optimization. */
+ if (delete_excluded
+ && !(new_mflags & (MATCHFLG_RECEIVER_SIDE|MATCHFLG_SENDER_SIDE))
+ && (!(new_mflags & MATCHFLG_PERDIR_MERGE)
+ || new_mflags & MATCHFLG_NO_PREFIXES))
+ new_mflags |= MATCHFLG_SENDER_SIDE;
+
*len_ptr = len;
*mflags_ptr = new_mflags;
return (const char *)s;
}
-static char default_cvsignore[] =
+static char default_cvsignore[] =
/* These default ignored items come from the CVS manual. */
"RCS SCCS CVS CVS.adm RCSLOG cvslog.* tags TAGS"
" .make.state .nse_depinfo *~ #* .#* ,* _$* *$"
" *.a *.olb *.o *.obj *.so *.exe"
" *.Z *.elc *.ln core"
/* The rest we added to suit ourself. */
- " .svn/";
+ " .svn/ .bzr/";
static void get_cvs_excludes(uint32 mflags)
{
&pat_len, &new_mflags);
if (!cp)
break;
+
+ pattern = cp + pat_len;
+
if (pat_len >= MAXPATHLEN) {
- rprintf(FERROR, "discarding over-long filter: %s\n",
- cp);
+ rprintf(FERROR, "discarding over-long filter: %.*s\n",
+ (int)pat_len, cp);
continue;
}
- pattern = cp + pat_len;
if (new_mflags & MATCHFLG_CLEAR_LIST) {
if (verbose > 2) {
}
len = pat_len;
if (new_mflags & MATCHFLG_EXCLUDE_SELF) {
- const char *name = strrchr(cp, '/');
- if (name)
- len -= ++name - cp;
- else
- name = cp;
+ const char *name = cp + len;
+ while (name > cp && name[-1] != '/') name--;
+ len -= name - cp;
add_rule(listp, name, len, 0, 0);
new_mflags &= ~MATCHFLG_EXCLUDE_SELF;
len = pat_len;
uint32 mflags, int xflags)
{
FILE *fp;
- char line[MAXPATHLEN+MAX_RULE_PREFIX+1]; /* +1 for trailing slash. */
+ char line[BIGPATHBUFLEN];
char *eob = line + sizeof line - 1;
int word_split = mflags & MATCHFLG_WORD_SPLIT;
if (verbose > 2) {
rprintf(FINFO, "[%s] parse_filter_file(%s,%x,%x)%s\n",
- who_am_i(), safe_fname(fname), mflags, xflags,
+ who_am_i(), fname, mflags, xflags,
fp ? "" : " [not found]");
}
rsyserr(FERROR, errno,
"failed to open %sclude file %s",
mflags & MATCHFLG_INCLUDE ? "in" : "ex",
- safe_fname(fname));
+ fname);
exit_cleanup(RERR_FILEIO);
}
return;
int ch, overflow = 0;
while (1) {
if ((ch = getc(fp)) == EOF) {
- if (ferror(fp) && errno == EINTR)
+ if (ferror(fp) && errno == EINTR) {
+ clearerr(fp);
continue;
+ }
break;
}
if (word_split && isspace(ch))
{
static char buf[MAX_RULE_PREFIX+1];
char *op = buf;
- int legal_len = for_xfer && protocol_version < 29 ? 1 : MAX_RULE_PREFIX;
+ int legal_len = for_xfer && protocol_version < 29 ? 1 : MAX_RULE_PREFIX-1;
if (match_flags & MATCHFLG_PERDIR_MERGE) {
if (legal_len == 1)
}
if (match_flags & MATCHFLG_EXCLUDE_SELF)
*op++ = 'e';
- if (legal_len)
- *op++ = ' ';
+ if (match_flags & MATCHFLG_SENDER_SIDE
+ && (!for_xfer || protocol_version >= 29))
+ *op++ = 's';
+ if (match_flags & MATCHFLG_RECEIVER_SIDE
+ && (!for_xfer || protocol_version >= 29
+ || (delete_excluded && am_sender)))
+ *op++ = 'r';
if (op - buf > legal_len)
return NULL;
+ if (legal_len)
+ *op++ = ' ';
*op = '\0';
if (plen_ptr)
*plen_ptr = op - buf;
static void send_rules(int f_out, struct filter_list_struct *flp)
{
- struct filter_struct *ent;
+ struct filter_struct *ent, *prev = NULL;
for (ent = flp->head; ent; ent = ent->next) {
unsigned int len, plen, dlen;
+ int elide = 0;
char *p;
+ /* Note we need to check delete_excluded here in addition to
+ * the code in parse_rule_tok() because some rules may have
+ * been added before we found the --delete-excluded option.
+ * We must also elide any CVS merge-file rules to avoid a
+ * backward compatibility problem, and we elide any no-prefix
+ * merge files as an optimization (since they can only have
+ * include/exclude rules). */
+ if (ent->match_flags & MATCHFLG_SENDER_SIDE)
+ elide = am_sender ? 1 : -1;
+ if (ent->match_flags & MATCHFLG_RECEIVER_SIDE)
+ elide = elide ? 0 : am_sender ? -1 : 1;
+ else if (delete_excluded && !elide
+ && (!(ent->match_flags & MATCHFLG_PERDIR_MERGE)
+ || ent->match_flags & MATCHFLG_NO_PREFIXES))
+ elide = am_sender ? 1 : -1;
+ if (elide < 0) {
+ if (prev)
+ prev->next = ent->next;
+ else
+ flp->head = ent->next;
+ } else
+ prev = ent;
+ if (elide > 0)
+ continue;
if (ent->match_flags & MATCHFLG_CVS_IGNORE
&& !(ent->match_flags & MATCHFLG_MERGE_FILE)) {
- if (am_sender || protocol_version < 29) {
- send_rules(f_out, &cvs_filter_list);
+ int f = am_sender || protocol_version < 29 ? f_out : -2;
+ send_rules(f, &cvs_filter_list);
+ if (f == f_out)
continue;
- }
}
p = get_rule_prefix(ent->match_flags, ent->pattern, 1, &plen);
if (!p) {
"filter rules are too modern for remote rsync.\n");
exit_cleanup(RERR_SYNTAX);
}
+ if (f_out < 0)
+ continue;
len = strlen(ent->pattern);
dlen = ent->match_flags & MATCHFLG_DIRECTORY ? 1 : 0;
if (!(plen + len + dlen))
if (dlen)
write_byte(f_out, '/');
}
+ flp->tail = prev;
}
/* This is only called by the client. */
void send_filter_list(int f_out)
{
- int receiver_wants_list = delete_mode && !delete_excluded;
+ int receiver_wants_list = prune_empty_dirs
+ || (delete_mode && (!delete_excluded || protocol_version >= 29));
if (local_server || (am_sender && !receiver_wants_list))
f_out = -1;
if (list_only == 1 && !recurse)
parse_rule(&filter_list, "/*/*", MATCHFLG_NO_PREFIXES, 0);
- if (f_out >= 0) {
- send_rules(f_out, &filter_list);
+ send_rules(f_out, &filter_list);
+
+ if (f_out >= 0)
write_int(f_out, 0);
- }
if (cvs_exclude) {
if (!am_sender || protocol_version < 29)
/* This is only called by the server. */
void recv_filter_list(int f_in)
{
- char line[MAXPATHLEN+MAX_RULE_PREFIX+1]; /* +1 for trailing slash. */
+ char line[BIGPATHBUFLEN];
int xflags = protocol_version >= 29 ? 0 : XFLG_OLD_PREFIXES;
+ int receiver_wants_list = prune_empty_dirs
+ || (delete_mode
+ && (!delete_excluded || protocol_version >= 29));
unsigned int len;
- int receiver_wants_list = delete_mode && !delete_excluded;
if (!local_server && (am_sender || receiver_wants_list)) {
while ((len = read_int(f_in)) != 0) {
if (len >= sizeof line)
- overflow("recv_rules");
+ overflow_exit("recv_rules");
read_sbuf(f_in, line, len);
parse_rule(&filter_list, line, 0, xflags);
}
if (local_server || am_sender)
parse_rule(&filter_list, "-C", 0, 0);
}
+
+ if (local_server) /* filter out any rules that aren't for us. */
+ send_rules(-1, &filter_list);
}