X-Git-Url: https://mattmccutchen.net/rsync/rsync.git/blobdiff_plain/0eeb1cf83ae21e57cd4fc90fb21febdfcfab328a..d3d07a5e860f1cde0e234ec7a1aff7111a2c514f:/exclude.c diff --git a/exclude.c b/exclude.c index 88f2cbbb..027ea7a6 100644 --- a/exclude.c +++ b/exclude.c @@ -1,12 +1,14 @@ -/* -*- c-file-style: "linux" -*- +/* + * The filter include/exclude routines. * - * Copyright (C) 1996-2001 by Andrew Tridgell - * Copyright (C) 1996 by Paul Mackerras - * Copyright (C) 2002 by Martin Pool + * Copyright (C) 1996-2001 Andrew Tridgell + * Copyright (C) 1996 Paul Mackerras + * Copyright (C) 2002 Martin Pool + * Copyright (C) 2003-2008 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 - * the Free Software Foundation; either version 2 of the License, or + * the Free Software Foundation; either version 3 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, @@ -14,36 +16,42 @@ * 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, visit the http://fsf.org website. */ -/* 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 */ - #include "rsync.h" extern int verbose; +extern int am_server; +extern int am_sender; extern int eol_nulls; -extern int list_only; -extern int recurse; extern int io_error; +extern int local_server; +extern int prune_empty_dirs; +extern int ignore_perishable; +extern int delete_mode; +extern int delete_excluded; +extern int cvs_exclude; extern int sanitize_paths; extern int protocol_version; +extern int module_id; extern char curr_dir[]; extern unsigned int curr_dir_len; extern unsigned int module_dirlen; struct filter_list_struct filter_list = { 0, 0, "" }; -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 "/!Crsp" +#define MODIFIERS_HIDE_PROTECT "/!p" + /* 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 * trailing slash appended, and the variable dirbuf_len contains the length @@ -106,38 +114,49 @@ static void free_filter(struct filter_struct *ex) /* Build a filter structure given a filter pattern. The value in "pat" * is not null-terminated. */ -static void make_filter(struct filter_list_struct *listp, const char *pat, - unsigned int pat_len, unsigned int mflags) +static void add_rule(struct filter_list_struct *listp, const char *pat, + unsigned int pat_len, uint32 mflags, int xflags) { struct filter_struct *ret; const char *cp; unsigned int ex_len; if (verbose > 2) { - rprintf(FINFO, "[%s] make_filter(%.*s, %s%s)\n", - who_am_i(), (int)pat_len, pat, - mflags & MATCHFLG_PERDIR_MERGE ? "per-dir-merge" - : mflags & MATCHFLG_INCLUDE ? "include" : "exclude", + rprintf(FINFO, "[%s] add_rule(%s%.*s%s)%s\n", + who_am_i(), get_rule_prefix(mflags, pat, 0, NULL), + (int)pat_len, pat, + (mflags & MATCHFLG_DIRECTORY) ? "/" : "", listp->debug_type); } - ret = new(struct filter_struct); - if (!ret) - out_of_memory("make_filter"); + /* 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; + } + } - memset(ret, 0, sizeof ret[0]); + if (!(ret = new0(struct filter_struct))) + out_of_memory("add_rule"); - if (mflags & MATCHFLG_ABS_PATH) { - if (*pat != '/') { - mflags &= ~MATCHFLG_ABS_PATH; - ex_len = 0; - } else + 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; + if (*pat == '/') ex_len = dirbuf_len - module_dirlen - 1; + else + ex_len = 0; } else ex_len = 0; - ret->pattern = new_array(char, ex_len + pat_len + 1); - if (!ret->pattern) - out_of_memory("make_filter"); + if (!(ret->pattern = new_array(char, ex_len + pat_len + 1))) + out_of_memory("add_rule"); if (ex_len) memcpy(ret->pattern, dirbuf + module_dirlen, ex_len); strlcpy(ret->pattern + ex_len, pat, pat_len + 1); @@ -150,6 +169,12 @@ static void make_filter(struct filter_list_struct *listp, const char *pat, /* 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; } } @@ -174,9 +199,9 @@ static void make_filter(struct filter_list_struct *listp, const char *pat, struct filter_struct *ex = mergelist_parents[i]; const char *s = strrchr(ex->pattern, '/'); if (s) - s++; + s++; else - s = ex->pattern; + s = ex->pattern; len = strlen(s); if (len == pat_len - (cp - ret->pattern) && memcmp(s, cp, len) == 0) { @@ -186,10 +211,10 @@ static void make_filter(struct filter_list_struct *listp, const char *pat, } if (!(lp = new_array(struct filter_list_struct, 1))) - out_of_memory("make_filter"); + out_of_memory("add_rule"); lp->head = lp->tail = NULL; - if (asprintf(&lp->debug_type, " (per-dir %s)", cp) < 0) - out_of_memory("make_filter"); + if (asprintf(&lp->debug_type, " [per-dir %s]", cp) < 0) + out_of_memory("add_rule"); ret->u.mergelist = lp; if (mergelist_cnt == mergelist_size) { @@ -198,7 +223,7 @@ static void make_filter(struct filter_list_struct *listp, const char *pat, struct filter_struct *, mergelist_size); if (!mergelist_parents) - out_of_memory("make_filter"); + out_of_memory("add_rule"); } mergelist_parents[mergelist_cnt++] = ret; } else { @@ -275,24 +300,23 @@ static char *parse_merge_name(const char *merge_file, unsigned int *len_ptr, 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, CFN_COLLAPSE_DOT_DOT_DIRS); } - - fn_len = strlen(fn); - if (fn == buf) - goto done; - if (dirbuf_len + fn_len >= MAXPATHLEN) { - rprintf(FERROR, "merge-file name overflows: %s\n", 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, CFN_COLLAPSE_DOT_DOT_DIRS); } - 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; @@ -312,7 +336,7 @@ void set_filter_dir(const char *dir, unsigned int dirlen) len = 0; memcpy(dirbuf + len, dir, dirlen); dirbuf[dirlen + len] = '\0'; - dirbuf_len = clean_fname(dirbuf, 1); + dirbuf_len = clean_fname(dirbuf, CFN_COLLAPSE_DOT_DOT_DIRS); if (dirbuf_len > 1 && dirbuf[dirbuf_len-1] == '.' && dirbuf[dirbuf_len-2] == '/') dirbuf_len -= 2; @@ -329,7 +353,7 @@ void set_filter_dir(const char *dir, unsigned int dirlen) * dirs from that point through the parent dir of the transfer dir looking * for the per-dir merge-file in each one. */ static BOOL setup_merge_file(struct filter_struct *ex, - struct filter_list_struct *lp, int flags) + struct filter_list_struct *lp) { char buf[MAXPATHLEN]; char *x, *y, *pat = ex->pattern; @@ -348,7 +372,7 @@ static BOOL setup_merge_file(struct filter_struct *ex, else pathjoin(buf, MAXPATHLEN, dirbuf, x); - len = clean_fname(buf, 1); + len = clean_fname(buf, CFN_COLLAPSE_DOT_DOT_DIRS); if (len != 1 && len < MAXPATHLEN-1) { buf[len++] = '/'; buf[len] = '\0'; @@ -365,7 +389,7 @@ static BOOL setup_merge_file(struct filter_struct *ex, *y = '\0'; dirbuf_len = y - dirbuf; strlcpy(x, ex->pattern, MAXPATHLEN - (x - buf)); - add_filter_file(lp, buf, flags | XFLG_ABS_PATH); + parse_filter_file(lp, buf, ex->match_flags, XFLG_ANCHORED2ABS); if (ex->match_flags & MATCHFLG_NO_INHERIT) lp->head = NULL; lp->tail = NULL; @@ -400,12 +424,11 @@ void *push_local_filters(const char *dir, unsigned int dirlen) sizeof (struct filter_list_struct)); } - /* Note: add_filter_file() might increase mergelist_cnt, so keep + /* 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; - int flags = 0; if (verbose > 2) { rprintf(FINFO, "[%s] pushing filter list%s\n", @@ -415,25 +438,18 @@ void *push_local_filters(const char *dir, unsigned int dirlen) lp->tail = NULL; /* Switch any local rules to inherited. */ if (ex->match_flags & MATCHFLG_NO_INHERIT) lp->head = NULL; - if (ex->match_flags & MATCHFLG_WORD_SPLIT) - flags |= XFLG_WORD_SPLIT; - if (ex->match_flags & MATCHFLG_NO_PREFIXES) - flags |= XFLG_NO_PREFIXES; - if (ex->match_flags & MATCHFLG_INCLUDE) - flags |= XFLG_DEF_INCLUDE; - else if (ex->match_flags & MATCHFLG_NO_PREFIXES) - flags |= XFLG_DEF_EXCLUDE; if (ex->match_flags & MATCHFLG_FINISH_SETUP) { ex->match_flags &= ~MATCHFLG_FINISH_SETUP; - if (setup_merge_file(ex, lp, flags)) + if (setup_merge_file(ex, lp)) set_filter_dir(dir, dirlen); } if (strlcpy(dirbuf + dirbuf_len, ex->pattern, - MAXPATHLEN - dirbuf_len) < MAXPATHLEN - dirbuf_len) - add_filter_file(lp, dirbuf, flags | XFLG_ABS_PATH); - else { + MAXPATHLEN - dirbuf_len) < MAXPATHLEN - dirbuf_len) { + parse_filter_file(lp, dirbuf, ex->match_flags, + XFLG_ANCHORED2ABS); + } else { io_error |= IOERR_GENERAL; rprintf(FINFO, "cannot add local filter rules in long-named directory: %s\n", @@ -473,84 +489,110 @@ void pop_local_filters(void *mem) free(pop); } -static int rule_matches(char *name, struct filter_struct *ex, int name_is_dir) +void change_local_filter_dir(const char *dname, int dlen, int dir_depth) { - char *p, full_name[MAXPATHLEN]; - int match_start = 0; - char *pattern = ex->pattern; + 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(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 */ + char *name = fname + (*fname == '/'); 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 && *fname != '/' && 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 && *fname != '/') { + /* Allow "**"+"/" to match at the start of the string. */ + strings[str_cnt++] = "/"; } - - if (ex->match_flags & MATCHFLG_DIRECTORY && !name_is_dir) - return 0; + 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)) - return 1; - 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 1; - } - 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 1; - } - } - } else if (match_start) { - if (strcmp(name,pattern) == 0) - return 1; + 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 1; + return ret_match; } } - return 0; + return !ret_match; } @@ -559,15 +601,17 @@ static void report_filter_result(char const *name, int name_is_dir, const char *type) { /* If a trailing slash is present to match only directories, - * then it is stripped out by make_filter. So as a special + * then it is stripped out by add_rule(). So as a special * 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); } } @@ -581,6 +625,8 @@ int check_filter(struct filter_list_struct *listp, 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, name, name_is_dir); @@ -588,6 +634,13 @@ int check_filter(struct filter_list_struct *listp, char *name, int name_is_dir) return rc; continue; } + if (ent->match_flags & MATCHFLG_CVS_IGNORE) { + int rc = check_filter(&cvs_filter_list, name, + name_is_dir); + if (rc) + return rc; + continue; + } if (rule_matches(name, ent, name_is_dir)) { report_filter_result(name, ent, name_is_dir, listp->debug_type); @@ -598,23 +651,34 @@ int check_filter(struct filter_list_struct *listp, char *name, int name_is_dir) 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 (if xflags does NOT contain XFLG_NO_PREFIXES) the - * +/- prefixes for overriding the include/exclude mode. The *flag_ptr - * value will also be set to the MATCHFLG_* bits for the current token. - */ -static const char *get_filter_tok(const char *p, int xflags, - unsigned int *len_ptr, unsigned int *flag_ptr) + * 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 unsigned char *s = (const unsigned char *)p; - unsigned int len, mflags = 0; - int empty_pat_is_OK = 0; + const uchar *s = (const uchar *)p; + uint32 new_mflags; + unsigned int len; - if (xflags & XFLG_WORD_SPLIT) { + if (mflags & MATCHFLG_WORD_SPLIT) { /* Skip over any initial whitespace. */ while (isspace(*s)) s++; @@ -624,86 +688,172 @@ static const char *get_filter_tok(const char *p, int xflags, if (!*s) return NULL; - /* Figure out what kind of a filter rule "s" is pointing at. */ - if (!(xflags & (XFLG_DEF_INCLUDE | XFLG_DEF_EXCLUDE))) { - char *mods = ""; + 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 ':': - mflags |= MATCHFLG_PERDIR_MERGE - | MATCHFLG_FINISH_SETUP; + new_mflags |= MATCHFLG_PERDIR_MERGE + | MATCHFLG_FINISH_SETUP; /* FALL THROUGH */ case '.': - mflags |= MATCHFLG_MERGE_FILE; - mods = "-+Cens"; + new_mflags |= MATCHFLG_MERGE_FILE; + mods = MODIFIERS_INCL_EXCL MODIFIERS_MERGE_FILE; break; case '+': - mflags |= MATCHFLG_INCLUDE; - break; + 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 '!': - mflags |= MATCHFLG_CLEAR_LIST; + 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 != '=' && *s != '_') { + while (mods && *++s && *s != ' ' && *s != '_') { if (strchr(mods, *s) == NULL) { - if (xflags & XFLG_WORD_SPLIT && isspace(*s)) { + if (mflags & MATCHFLG_WORD_SPLIT && isspace(*s)) { s--; break; } + invalid: rprintf(FERROR, - "unknown option '%c' in filter rule: %s\n", + "invalid modifier sequence at '%c' in filter rule: %s\n", *s, p); exit_cleanup(RERR_SYNTAX); } switch (*s) { case '-': - mflags |= MATCHFLG_NO_PREFIXES; + if (new_mflags & MATCHFLG_NO_PREFIXES) + goto invalid; + new_mflags |= MATCHFLG_NO_PREFIXES; break; case '+': - mflags |= MATCHFLG_NO_PREFIXES - | MATCHFLG_INCLUDE; + 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': - empty_pat_is_OK = 1; - mflags |= MATCHFLG_NO_PREFIXES - | MATCHFLG_WORD_SPLIT - | MATCHFLG_NO_INHERIT; + if (new_mflags & MATCHFLG_NO_PREFIXES) + goto invalid; + new_mflags |= MATCHFLG_NO_PREFIXES + | MATCHFLG_WORD_SPLIT + | MATCHFLG_NO_INHERIT + | MATCHFLG_CVS_IGNORE; break; case 'e': - mflags |= MATCHFLG_EXCLUDE_SELF; + new_mflags |= MATCHFLG_EXCLUDE_SELF; break; case 'n': - mflags |= MATCHFLG_NO_INHERIT; + new_mflags |= MATCHFLG_NO_INHERIT; + break; + case 'p': + new_mflags |= MATCHFLG_PERISHABLE; + break; + case 'r': + new_mflags |= MATCHFLG_RECEIVER_SIDE; break; case 's': - mflags |= MATCHFLG_WORD_SPLIT; + new_mflags |= MATCHFLG_SENDER_SIDE; + break; + case 'w': + new_mflags |= MATCHFLG_WORD_SPLIT; break; } } if (*s) s++; - } else if (!(xflags & XFLG_NO_PREFIXES) - && (*s == '-' || *s == '+') && s[1] == ' ') { - if (*s == '+') - mflags |= MATCHFLG_INCLUDE; - s += 2; - } else { - if (xflags & XFLG_DEF_INCLUDE) - mflags |= MATCHFLG_INCLUDE; - if (*s == '!') - mflags |= MATCHFLG_CLEAR_LIST; /* Tentative! */ } - if (xflags & XFLG_DIRECTORY) - mflags |= MATCHFLG_DIRECTORY; - - if (xflags & XFLG_WORD_SPLIT) { - const unsigned char *cp = s; + if (mflags & MATCHFLG_WORD_SPLIT) { + const uchar *cp = s; /* Token ends at whitespace or the end of the string. */ while (!isspace(*cp) && *cp != '\0') cp++; @@ -711,32 +861,71 @@ static const char *get_filter_tok(const char *p, int xflags, } else len = strlen((char*)s); - if (mflags & MATCHFLG_CLEAR_LIST) { - if (!(xflags & (XFLG_DEF_INCLUDE | XFLG_DEF_EXCLUDE)) && len) { + if (new_mflags & MATCHFLG_CLEAR_LIST) { + if (!(mflags & MATCHFLG_NO_PREFIXES) + && !(xflags & XFLG_OLD_PREFIXES) && len) { rprintf(FERROR, "'!' rule has trailing characters: %s\n", p); exit_cleanup(RERR_SYNTAX); } if (len > 1) - mflags &= ~MATCHFLG_CLEAR_LIST; - } else if (!len && !empty_pat_is_OK) { + new_mflags &= ~MATCHFLG_CLEAR_LIST; + } else if (!len && !(new_mflags & MATCHFLG_CVS_IGNORE)) { rprintf(FERROR, "unexpected end of filter rule: %s\n", p); exit_cleanup(RERR_SYNTAX); } - if (xflags & XFLG_ABS_PATH) - mflags |= MATCHFLG_ABS_PATH; + /* --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; - *flag_ptr = mflags; + *mflags_ptr = new_mflags; return (const char *)s; } -void add_filter(struct filter_list_struct *listp, const char *pattern, - int xflags) +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 *~ #* .#* ,* _$* *$" + " *.old *.bak *.BAK *.orig *.rej .del-*" + " *.a *.olb *.o *.obj *.so *.exe" + " *.Z *.elc *.ln core" + /* The rest we added to suit ourself. */ + " .svn/ .git/ .bzr/"; + +static void get_cvs_excludes(uint32 mflags) { - unsigned int pat_len, mflags; + static int initialized = 0; + char *p, fname[MAXPATHLEN]; + + if (initialized) + return; + initialized = 1; + + parse_rule(&cvs_filter_list, default_cvsignore, + mflags | (protocol_version >= 30 ? MATCHFLG_PERISHABLE : 0), + 0); + + p = module_id >= 0 && lp_use_chroot(module_id) ? "/" : getenv("HOME"); + if (p && pathjoin(fname, MAXPATHLEN, p, ".cvsignore") < MAXPATHLEN) + parse_filter_file(&cvs_filter_list, fname, mflags, 0); + + parse_rule(&cvs_filter_list, getenv("CVSIGNORE"), mflags, 0); +} + + +void parse_rule(struct filter_list_struct *listp, const char *pattern, + uint32 mflags, int xflags) +{ + unsigned int pat_len; + uint32 new_mflags; const char *cp, *p; if (!pattern) @@ -744,17 +933,20 @@ void add_filter(struct filter_list_struct *listp, const char *pattern, while (1) { /* Remember that the returned string is NOT '\0' terminated! */ - cp = get_filter_tok(pattern, xflags, &pat_len, &mflags); + cp = parse_rule_tok(pattern, mflags, xflags, + &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 (mflags & MATCHFLG_CLEAR_LIST) { + if (new_mflags & MATCHFLG_CLEAR_LIST) { if (verbose > 2) { rprintf(FINFO, "[%s] clearing filter list%s\n", @@ -764,87 +956,98 @@ void add_filter(struct filter_list_struct *listp, const char *pattern, continue; } - if (!pat_len) { - cp = ".cvsignore"; - pat_len = 10; - } - - if (mflags & MATCHFLG_MERGE_FILE) { - unsigned int len = pat_len; - if (mflags & MATCHFLG_EXCLUDE_SELF) { - const char *name = strrchr(cp, '/'); - if (name) - len -= ++name - cp; - else - name = cp; - make_filter(listp, name, len, 0); - mflags &= ~MATCHFLG_EXCLUDE_SELF; + if (new_mflags & MATCHFLG_MERGE_FILE) { + unsigned int len; + if (!pat_len) { + cp = ".cvsignore"; + pat_len = 10; + } + len = pat_len; + if (new_mflags & MATCHFLG_EXCLUDE_SELF) { + 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; } - if (mflags & MATCHFLG_PERDIR_MERGE) { + if (new_mflags & MATCHFLG_PERDIR_MERGE) { if (parent_dirscan) { - if (!(p = parse_merge_name(cp, &len, module_dirlen))) + if (!(p = parse_merge_name(cp, &len, + module_dirlen))) continue; - make_filter(listp, p, len, mflags); + add_rule(listp, p, len, new_mflags, 0); continue; } } else { - int flgs = XFLG_FATAL_ERRORS; if (!(p = parse_merge_name(cp, &len, 0))) continue; - if (mflags & MATCHFLG_INCLUDE) - flgs |= XFLG_DEF_INCLUDE; - else if (mflags & MATCHFLG_NO_PREFIXES) - flgs |= XFLG_DEF_EXCLUDE; - add_filter_file(listp, p, flgs); + parse_filter_file(listp, p, new_mflags, + XFLG_FATAL_ERRORS); continue; } } - make_filter(listp, cp, pat_len, mflags); + add_rule(listp, cp, pat_len, new_mflags, xflags); + + if (new_mflags & MATCHFLG_CVS_IGNORE + && !(new_mflags & MATCHFLG_MERGE_FILE)) + get_cvs_excludes(new_mflags); } } -void add_filter_file(struct filter_list_struct *listp, const char *fname, - int xflags) +void parse_filter_file(struct filter_list_struct *listp, const char *fname, + 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 = xflags & XFLG_WORD_SPLIT; + int word_split = mflags & MATCHFLG_WORD_SPLIT; if (!fname || !*fname) return; - if (*fname != '-' || fname[1]) - fp = fopen(fname, "rb"); - else + if (*fname != '-' || fname[1] || am_server) { + if (server_filter_list.head) { + strlcpy(line, fname, sizeof line); + clean_fname(line, CFN_COLLAPSE_DOT_DOT_DIRS); + if (check_filter(&server_filter_list, line, 0) < 0) + fp = NULL; + else + fp = fopen(line, "rb"); + } else + fp = fopen(fname, "rb"); + } else fp = stdin; + + if (verbose > 2) { + rprintf(FINFO, "[%s] parse_filter_file(%s,%x,%x)%s\n", + who_am_i(), fname, mflags, xflags, + fp ? "" : " [not found]"); + } + if (!fp) { if (xflags & XFLG_FATAL_ERRORS) { rsyserr(FERROR, errno, "failed to open %sclude file %s", - xflags & XFLG_DEF_INCLUDE ? "in" : "ex", - safe_fname(fname)); + mflags & MATCHFLG_INCLUDE ? "in" : "ex", + fname); exit_cleanup(RERR_FILEIO); } return; } dirbuf[dirbuf_len] = '\0'; - if (verbose > 2) { - rprintf(FINFO, "[%s] add_filter_file(%s,%d)\n", - who_am_i(), safe_fname(fname), xflags); - } - while (1) { char *s = line; 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)) @@ -863,126 +1066,191 @@ void add_filter_file(struct filter_list_struct *listp, const char *fname, *s = '\0'; /* Skip an empty token and (when line parsing) comments. */ if (*line && (word_split || (*line != ';' && *line != '#'))) - add_filter(listp, line, xflags); + parse_rule(listp, line, mflags, xflags); if (ch == EOF) break; } fclose(fp); } -char *get_rule_prefix(int match_flags, const char *pat, unsigned int *plen_ptr) +/* If the "for_xfer" flag is set, the prefix is made compatible with the + * current protocol_version (if possible) or a NULL is returned (if not + * possible). */ +char *get_rule_prefix(int match_flags, const char *pat, int for_xfer, + unsigned int *plen_ptr) { static char buf[MAX_RULE_PREFIX+1]; char *op = buf; + int legal_len = for_xfer && protocol_version < 29 ? 1 : MAX_RULE_PREFIX-1; if (match_flags & MATCHFLG_PERDIR_MERGE) { + if (legal_len == 1) + return NULL; *op++ = ':'; - if (match_flags & MATCHFLG_WORD_SPLIT) - *op++ = 's'; + } else if (match_flags & MATCHFLG_INCLUDE) + *op++ = '+'; + else if (legal_len != 1 + || ((*pat == '-' || *pat == '+') && pat[1] == ' ')) + *op++ = '-'; + else + legal_len = 0; + + if (match_flags & MATCHFLG_NEGATE) + *op++ = '!'; + if (match_flags & MATCHFLG_CVS_IGNORE) + *op++ = 'C'; + else { if (match_flags & MATCHFLG_NO_INHERIT) *op++ = 'n'; - if (match_flags & MATCHFLG_EXCLUDE_SELF) - *op++ = 'e'; + if (match_flags & MATCHFLG_WORD_SPLIT) + *op++ = 'w'; if (match_flags & MATCHFLG_NO_PREFIXES) { if (match_flags & MATCHFLG_INCLUDE) *op++ = '+'; else *op++ = '-'; } - *op++ = ' '; - } else if (match_flags & MATCHFLG_INCLUDE) { - *op++ = '+'; - *op++ = ' '; - } else if (protocol_version >= 29 - || ((*pat == '-' || *pat == '+') && pat[1] == ' ')) { - *op++ = '-'; - *op++ = ' '; } + if (match_flags & MATCHFLG_EXCLUDE_SELF) + *op++ = 'e'; + 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 (match_flags & MATCHFLG_PERISHABLE) { + if (!for_xfer || protocol_version >= 30) + *op++ = 'p'; + else if (am_sender) + return NULL; + } + if (op - buf > legal_len) + return NULL; + if (legal_len) + *op++ = ' '; *op = '\0'; if (plen_ptr) *plen_ptr = op - buf; - if (op - buf > MAX_RULE_PREFIX) - overflow("get_rule_prefix"); return buf; } -void send_filter_list(int f) +static void send_rules(int f_out, struct filter_list_struct *flp) { - struct filter_struct *ent; - - /* This is a complete hack - blame Rusty. FIXME! - * Remove this hack when older rsyncs (below 2.6.4) are gone. */ - if (list_only == 1 && !recurse) - add_filter(&filter_list, "/*/*", XFLG_DEF_EXCLUDE); + struct filter_struct *ent, *prev = NULL; - for (ent = filter_list.head; ent; ent = ent->next) { + for (ent = flp->head; ent; ent = ent->next) { unsigned int len, plen, dlen; + int elide = 0; char *p; - len = strlen(ent->pattern); - if (len == 0 || len >= MAXPATHLEN) + /* 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; - p = get_rule_prefix(ent->match_flags, ent->pattern, &plen); - if (protocol_version < 29 && *p == ':') { - if (strcmp(p, ":sn- ") == 0 - && strcmp(ent->pattern, ".cvsignore") == 0) + if (ent->match_flags & MATCHFLG_CVS_IGNORE + && !(ent->match_flags & MATCHFLG_MERGE_FILE)) { + 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) { rprintf(FERROR, - "remote rsync is too old to understand per-directory merge files.\n"); - exit_cleanup(RERR_SYNTAX); + "filter rules are too modern for remote rsync.\n"); + exit_cleanup(RERR_PROTOCOL); } + if (f_out < 0) + continue; + len = strlen(ent->pattern); dlen = ent->match_flags & MATCHFLG_DIRECTORY ? 1 : 0; - write_int(f, plen + len + dlen); + if (!(plen + len + dlen)) + continue; + write_int(f_out, plen + len + dlen); if (plen) - write_buf(f, p, plen); - write_buf(f, ent->pattern, len); + write_buf(f_out, p, plen); + write_buf(f_out, ent->pattern, len); if (dlen) - write_byte(f, '/'); + write_byte(f_out, '/'); } - - write_int(f, 0); + flp->tail = prev; } - -void recv_filter_list(int f) +/* This is only called by the client. */ +void send_filter_list(int f_out) { - char line[MAXPATHLEN+MAX_RULE_PREFIX+1]; /* +1 for trailing slash. */ - unsigned int xflags = protocol_version >= 29 ? 0 : XFLG_DEF_EXCLUDE; - unsigned int l; - - while ((l = read_int(f)) != 0) { - if (l >= sizeof line) - overflow("recv_filter_list"); - read_sbuf(f, line, l); - add_filter(&filter_list, line, xflags); + 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 (cvs_exclude && am_sender) { + if (protocol_version >= 29) + parse_rule(&filter_list, ":C", 0, 0); + parse_rule(&filter_list, "-C", 0, 0); } -} + send_rules(f_out, &filter_list); -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 *~ #* .#* ,* _$* *$" - " *.old *.bak *.BAK *.orig *.rej .del-*" - " *.a *.olb *.o *.obj *.so *.exe" - " *.Z *.elc *.ln core" - /* The rest we added to suit ourself. */ - " .svn/"; + if (f_out >= 0) + write_int(f_out, 0); + + if (cvs_exclude) { + if (!am_sender || protocol_version < 29) + parse_rule(&filter_list, ":C", 0, 0); + if (!am_sender) + parse_rule(&filter_list, "-C", 0, 0); + } +} -void add_cvs_excludes(void) +/* This is only called by the server. */ +void recv_filter_list(int f_in) { - static unsigned int cvs_flags = XFLG_WORD_SPLIT | XFLG_NO_PREFIXES - | XFLG_DEF_EXCLUDE; - char fname[MAXPATHLEN]; - char *p; + 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; - add_filter(&filter_list, ":C", 0); - add_filter(&filter_list, default_cvsignore, cvs_flags); + if (!local_server && (am_sender || receiver_wants_list)) { + while ((len = read_int(f_in)) != 0) { + if (len >= sizeof line) + overflow_exit("recv_rules"); + read_sbuf(f_in, line, len); + parse_rule(&filter_list, line, 0, xflags); + } + } - if ((p = getenv("HOME")) - && pathjoin(fname, sizeof fname, p, ".cvsignore") < sizeof fname) { - add_filter_file(&filter_list, fname, cvs_flags); + if (cvs_exclude) { + if (local_server || am_sender || protocol_version < 29) + parse_rule(&filter_list, ":C", 0, 0); + if (local_server || am_sender) + parse_rule(&filter_list, "-C", 0, 0); } - add_filter(&filter_list, getenv("CVSIGNORE"), cvs_flags); + if (local_server) /* filter out any rules that aren't for us. */ + send_rules(-1, &filter_list); }