X-Git-Url: https://mattmccutchen.net/rsync/rsync-patches.git/blobdiff_plain/fe6407b5a9f04c88f74ea95e959aaa38334e41a6..5397d015f4a69a0762e71a7446671df623b4fe00:/link-by-hash.diff diff --git a/link-by-hash.diff b/link-by-hash.diff index daffd6f..94cae34 100644 --- a/link-by-hash.diff +++ b/link-by-hash.diff @@ -11,8 +11,8 @@ will only store one copy of the unique contents of each file, regardless of the file's name. ---- Makefile.in 15 May 2004 00:48:11 -0000 1.101 -+++ Makefile.in 15 May 2004 20:29:17 -0000 +--- orig/Makefile.in 2004-11-03 11:56:03 ++++ Makefile.in 2004-07-03 20:20:15 @@ -35,7 +35,7 @@ OBJS1=rsync.o generator.o receiver.o cle main.o checksum.o match.o syscall.o log.o backup.o OBJS2=options.o flist.o io.o compat.o hlink.o token.o uidlist.o socket.o \ @@ -22,9 +22,9 @@ the file's name. DAEMON_OBJ = params.o loadparm.o clientserver.o access.o connection.o authenticate.o popt_OBJS=popt/findme.o popt/popt.o popt/poptconfig.o \ popt/popthelp.o popt/poptparse.o ---- /dev/null 1 Jan 1970 00:00:00 -0000 -+++ hashlink.c 15 May 2004 20:29:17 -0000 -@@ -0,0 +1,342 @@ +--- orig/hashlink.c 2004-09-24 16:44:25 ++++ hashlink.c 2004-09-24 16:44:25 +@@ -0,0 +1,340 @@ +/* + Copyright (C) Cronosys, LLC 2004 + @@ -49,7 +49,7 @@ the file's name. + +extern char *link_by_hash_dir; + -+#ifdef HAVE_LINK ++#if HAVE_LINK + +char* make_hash_name(struct file_struct *file) +{ @@ -111,11 +111,11 @@ the file's name. + long this_fnbr; + + *fnbr = 0; -+ ++ + /* Build a list of potential candidates and open + * them. */ + if ((d = opendir(hashname)) == NULL) { -+ rsyserr(FERROR, errno, "opendir \"%s\"", hashname); ++ rsyserr(FERROR, errno, "opendir failed: \"%s\"", hashname); + free(hashname); + return NULL; + } @@ -134,7 +134,7 @@ the file's name. + asprintf(&hashfile->name,"%s/%s",hashname, + di->d_name); + if (do_stat(hashfile->name,&st) == -1) { -+ rsyserr(FERROR, errno, "%s: %s", hashfile->name); ++ rsyserr(FERROR, errno, "stat failed: %s", hashfile->name); + kill_hashfile(hashfile); + continue; + } @@ -145,7 +145,7 @@ the file's name. + hashfile->nlink = st.st_nlink; + hashfile->fd = open(hashfile->name,O_RDONLY|O_BINARY); + if (hashfile->fd == -1) { -+ rsyserr(FERROR, errno, "%s", hashfile->name); ++ rsyserr(FERROR, errno, "open failed: %s", hashfile->name); + kill_hashfile(hashfile); + continue; + } @@ -210,11 +210,10 @@ the file's name. + /* There are no matches. */ + return NULL; + } -+ + } + + if (amt == -1) { -+ rsyserr(FERROR, errno, "%s"); ++ rsyserr(FERROR, errno, "read failed in compare_hashfiles()"); + kill_hashfiles(files); + return NULL; + } @@ -250,7 +249,7 @@ the file's name. +int link_by_hash(char *fnametmp,char *fname,struct file_struct *file) +{ + STRUCT_STAT st; -+ char *hashname = make_hash_name(file); ++ char *hashname = make_hash_name(file); + int first = 0, rc; + char *linkname; + long last_fnbr; @@ -266,7 +265,7 @@ the file's name. + dirname = strdup(hashname); + *strrchr(dirname,'/') = 0; + if (do_mkdir(dirname, 0755) == -1 && errno != EEXIST) { -+ rsyserr(FERROR, errno, "mkdir %s", dirname); ++ rsyserr(FERROR, errno, "mkdir failed: %s", dirname); + free(hashname); + free(dirname); + return robust_rename(fnametmp,fname,0644); @@ -274,7 +273,7 @@ the file's name. + free(dirname); + + if (do_mkdir(hashname, 0755) == -1 && errno != EEXIST) { -+ rsyserr(FERROR, errno, "mkdir %s", hashname); ++ rsyserr(FERROR, errno, "mkdir failed: %s", hashname); + free(hashname); + return robust_rename(fnametmp,fname,0644); + } @@ -282,13 +281,11 @@ the file's name. + first = 1; + asprintf(&linkname,"%s/0",hashname); + rprintf(FINFO, "(1) linkname = %s\n", linkname); -+ + } else { + struct hashfile_struct *hashfiles, *hashfile; -+ int fd; + + if (do_stat(fnametmp,&st) == -1) { -+ rsyserr(FERROR, errno, "%s", fname); ++ rsyserr(FERROR, errno, "stat failed: %s", fname); + return -1; + } + hashfiles = find_hashfiles(hashname, st.st_size, &last_fnbr); @@ -298,15 +295,16 @@ the file's name. + asprintf(&linkname,"%s/0",hashname); + rprintf(FINFO, "(2) linkname = %s\n", linkname); + } else { -+ ++ int fd; + /* Search for one identical to us. */ + if ((fd = open(fnametmp,O_RDONLY|O_BINARY)) == -1) { -+ rsyserr(FERROR, errno, "%s", fnametmp); ++ rsyserr(FERROR, errno, "open failed: %s", fnametmp); + kill_hashfiles(hashfiles); + return -1; + } + hashfile = compare_hashfiles(fd, hashfiles); + hashfiles = NULL; ++ close(fd); + + if (hashfile) { + first = 0; @@ -325,6 +323,7 @@ the file's name. + if (!first) { + rprintf(FINFO, "link-by-hash (existing): \"%s\" -> %s\n", + linkname, full_fname(fname)); ++ robust_unlink(fname); + rc = do_link(linkname, fname); + if (rc == -1) { + if (errno == EMLINK) { @@ -337,7 +336,6 @@ the file's name. + } else { + rsyserr(FERROR, errno, "link \"%s\" -> \"%s\"", + linkname, full_fname(fname)); -+ robust_unlink(fname); + rc = robust_rename(fnametmp,fname,0644); + } + } else { @@ -367,49 +365,51 @@ the file's name. +} + +#endif ---- options.c 6 May 2004 21:08:01 -0000 1.148 -+++ options.c 15 May 2004 20:29:17 -0000 -@@ -121,6 +121,7 @@ char *log_format = NULL; +--- orig/options.c 2005-02-11 18:21:45 ++++ options.c 2005-01-28 19:32:26 +@@ -131,6 +131,7 @@ char *log_format = NULL; char *password_file = NULL; char *rsync_path = RSYNC_PATH; char *backup_dir = NULL; +char *link_by_hash_dir = NULL; char backup_dir_buf[MAXPATHLEN]; - int rsync_port = RSYNC_PORT; - int link_dest = 0; -@@ -266,6 +267,7 @@ void usage(enum logcode F) - rprintf(F," -T --temp-dir=DIR create temporary files in directory DIR\n"); + int rsync_port = 0; + int compare_dest = 0; +@@ -305,6 +306,7 @@ void usage(enum logcode F) rprintf(F," --compare-dest=DIR also compare destination files relative to DIR\n"); - rprintf(F," --link-dest=DIR create hardlinks to DIR for unchanged files\n"); + rprintf(F," --copy-dest=DIR ... and include copies of unchanged files\n"); + rprintf(F," --link-dest=DIR hardlink to files in DIR when unchanged\n"); + rprintf(F," --link-by-hash=DIR create hardlinks by hash to DIR for regular files\n"); - rprintf(F," -P equivalent to --partial --progress\n"); rprintf(F," -z, --compress compress file data\n"); - rprintf(F," -C, --cvs-exclude auto ignore files in the same way CVS does\n"); -@@ -305,7 +307,7 @@ void usage(enum logcode F) - enum {OPT_VERSION = 1000, OPT_SENDER, OPT_EXCLUDE, OPT_EXCLUDE_FROM, - OPT_DELETE_AFTER, OPT_DELETE_EXCLUDED, OPT_LINK_DEST, + rprintf(F," -C, --cvs-exclude auto-ignore files the same way CVS does\n"); + rprintf(F," -f, --filter=RULE add a file-filtering RULE\n"); +@@ -343,7 +345,7 @@ void usage(enum logcode F) + enum {OPT_VERSION = 1000, OPT_DAEMON, OPT_SENDER, OPT_EXCLUDE, OPT_EXCLUDE_FROM, + OPT_FILTER, OPT_COMPARE_DEST, OPT_COPY_DEST, OPT_LINK_DEST, OPT_INCLUDE, OPT_INCLUDE_FROM, OPT_MODIFY_WINDOW, -- OPT_READ_BATCH, OPT_WRITE_BATCH, -+ OPT_READ_BATCH, OPT_WRITE_BATCH, OPT_LINK_BY_HASH, +- OPT_READ_BATCH, OPT_WRITE_BATCH, OPT_TIMEOUT, OPT_MAX_SIZE, ++ OPT_READ_BATCH, OPT_WRITE_BATCH, OPT_TIMEOUT, OPT_MAX_SIZE, OPT_LINK_BY_HASH, OPT_REFUSED_BASE = 9000}; static struct poptOption long_options[] = { -@@ -362,6 +364,7 @@ static struct poptOption long_options[] - {"temp-dir", 'T', POPT_ARG_STRING, &tmpdir, 0, 0, 0 }, - {"compare-dest", 0, POPT_ARG_STRING, &compare_dest, 0, 0, 0 }, - {"link-dest", 0, POPT_ARG_STRING, &compare_dest, OPT_LINK_DEST, 0, 0 }, -+ {"link-by-hash", 0, POPT_ARG_STRING, 0, OPT_LINK_BY_HASH, 0, 0}, +@@ -411,6 +413,7 @@ static struct poptOption long_options[] + {"compare-dest", 0, POPT_ARG_STRING, 0, OPT_COMPARE_DEST, 0, 0 }, + {"copy-dest", 0, POPT_ARG_STRING, 0, OPT_COPY_DEST, 0, 0 }, + {"link-dest", 0, POPT_ARG_STRING, 0, OPT_LINK_DEST, 0, 0 }, ++ {"link-by-hash", 0, POPT_ARG_STRING, 0, OPT_LINK_BY_HASH, 0, 0}, /* TODO: Should this take an optional int giving the compression level? */ {"compress", 'z', POPT_ARG_NONE, &do_compression, 0, 0, 0 }, - {"daemon", 0, POPT_ARG_NONE, &daemon_opt, 0, 0, 0 }, -@@ -584,6 +587,19 @@ int parse_arguments(int *argc, const cha - return 0; - #endif + {"stats", 0, POPT_ARG_NONE, &do_stats, 0, 0, 0 }, +@@ -831,6 +834,21 @@ int parse_arguments(int *argc, const cha + basis_dir[basis_dir_cnt++] = (char *)arg; + break; + case OPT_LINK_BY_HASH: +#if HAVE_LINK -+ link_by_hash_dir = (char *)poptGetOptArg(pc); -+ checksum_seed = FIXED_CHECKSUM_SEED; ++ arg = poptGetOptArg(pc); ++ if (sanitize_paths) ++ arg = sanitize_path(NULL, arg, NULL, 0); ++ link_by_hash_dir = (char *)arg; + break; +#else + snprintf(err_buf, sizeof err_buf, @@ -421,149 +421,131 @@ the file's name. + default: /* A large opt value means that set_refuse_options() - * turned this option off (opt-BASE is its index). */ -@@ -951,6 +967,11 @@ void server_options(char **args,int *arg - */ - args[ac++] = link_dest ? "--link-dest" : "--compare-dest"; - args[ac++] = compare_dest; -+ } -+ + * turned this option off. */ +@@ -1366,6 +1384,11 @@ void server_options(char **args,int *arg + } + } + + if (link_by_hash_dir && am_sender) { + args[ac++] = "--link-by-hash"; + args[ac++] = link_by_hash_dir; - } - ++ } ++ if (files_from && (!am_sender || remote_filesfrom_file)) { ---- receiver.c 15 May 2004 19:31:10 -0000 1.78 -+++ receiver.c 15 May 2004 20:29:17 -0000 -@@ -46,6 +46,7 @@ extern int module_id; - extern int ignore_errors; - extern int orig_umask; - extern int keep_partial; + if (remote_filesfrom_file) { + args[ac++] = "--files-from"; +--- orig/receiver.c 2005-02-11 10:53:14 ++++ receiver.c 2005-01-15 21:29:13 +@@ -35,6 +35,7 @@ extern int preserve_hard_links; + extern int preserve_perms; + extern int io_error; + extern char *tmpdir; +extern char *link_by_hash_dir; - - static void delete_one(char *fn, int is_dir) - { -@@ -192,10 +193,11 @@ static int get_tmpname(char *fnametmp, c + extern char *partial_dir; + extern char *basis_dir[]; + extern int basis_dir_cnt; +@@ -137,12 +138,13 @@ static int get_tmpname(char *fnametmp, c - static int receive_data(int f_in,struct map_struct *mapbuf,int fd,char *fname, -- OFF_T total_size) -+ OFF_T total_size,char *md4) + static int receive_data(int f_in, char *fname_r, int fd_r, OFF_T size_r, +- char *fname, int fd, OFF_T total_size) ++ char *fname, int fd, OFF_T total_size, char *md4) { - int i; + static char file_sum1[MD4_SUM_LENGTH]; + static char file_sum2[MD4_SUM_LENGTH]; + struct map_struct *mapbuf; struct sum_struct sum; + struct mdfour mdfour_data; - unsigned int len; + int32 len; OFF_T offset = 0; OFF_T offset2; -@@ -205,7 +207,9 @@ static int receive_data(int f_in,struct - char *map=NULL; +@@ -162,6 +164,9 @@ static int receive_data(int f_in, char * + } else + mapbuf = NULL; - read_sum_head(f_in, &sum); -- + if (md4) + mdfour_begin(&mdfour_data); -+ - sum_init(); ++ + sum_init(checksum_seed); while ((i = recv_token(f_in, &data)) != 0) { -@@ -222,6 +226,8 @@ static int receive_data(int f_in,struct +@@ -178,6 +183,8 @@ static int receive_data(int f_in, char * cleanup_got_literal = 1; - sum_update(data,i); + sum_update(data, i); + if (md4) + mdfour_update(&mdfour_data,data,i); - if (fd != -1 && write_file(fd,data,i) != i) { - rsyserr(FERROR, errno, "write failed on %s", -@@ -249,6 +255,8 @@ static int receive_data(int f_in,struct + if (fd != -1 && write_file(fd,data,i) != i) + goto report_write_error; +@@ -204,6 +211,8 @@ static int receive_data(int f_in, char * see_token(map, len); - sum_update(map,len); + sum_update(map, len); + if (md4) + mdfour_update(&mdfour_data,map,len); } - if (fd != -1 && write_file(fd,map,len) != (int) len) { -@@ -271,6 +279,8 @@ static int receive_data(int f_in,struct + if (inplace) { +@@ -244,6 +253,8 @@ static int receive_data(int f_in, char * } sum_end(file_sum1); + if (md4) + mdfour_result(&mdfour_data, (unsigned char*)md4); - read_buf(f_in,file_sum2,MD4_SUM_LENGTH); - if (verbose > 2) { -@@ -374,7 +384,7 @@ int recv_files(int f_in,struct file_list - if (fd1 != -1 && do_fstat(fd1,&st) != 0) { - rsyserr(FERROR, errno, "fstat %s failed", - full_fname(fnamecmp)); -- receive_data(f_in,NULL,-1,NULL,file->length); -+ receive_data(f_in,NULL,-1,NULL,file->length,NULL); - close(fd1); - continue; - } -@@ -387,7 +397,7 @@ int recv_files(int f_in,struct file_list - */ - rprintf(FERROR,"recv_files: %s is a directory\n", - full_fname(fnamecmp)); -- receive_data(f_in, NULL, -1, NULL, file->length); -+ receive_data(f_in,NULL,-1,NULL,file->length,NULL); - close(fd1); - continue; - } -@@ -439,7 +449,7 @@ int recv_files(int f_in,struct file_list - if (fd2 == -1) { - rsyserr(FERROR, errno, "mkstemp %s failed", - full_fname(fnametmp)); -- receive_data(f_in,mapbuf,-1,NULL,file->length); -+ receive_data(f_in,mapbuf,-1,NULL,file->length,NULL); - if (mapbuf) unmap_file(mapbuf); - if (fd1 != -1) close(fd1); - continue; -@@ -452,7 +462,12 @@ int recv_files(int f_in,struct file_list - } + if (mapbuf) + unmap_file(mapbuf); +@@ -259,7 +270,7 @@ static int receive_data(int f_in, char * + + static void discard_receive_data(int f_in, OFF_T length) + { +- receive_data(f_in, NULL, -1, 0, NULL, -1, length); ++ receive_data(f_in, NULL, -1, 0, NULL, -1, length, NULL); + } + + +@@ -501,8 +512,12 @@ int recv_files(int f_in, struct file_lis + rprintf(FINFO, "%s\n", safe_fname(fname)); /* recv file data */ -- recv_ok = receive_data(f_in,mapbuf,fd2,fname,file->length); -+#ifdef HAVE_LINK -+ if (link_by_hash_dir) { -+ file->u.sum = (char*)malloc (MD4_SUM_LENGTH); -+ } ++#if HAVE_LINK ++ if (link_by_hash_dir) ++ file->u.sum = (char*)malloc(MD4_SUM_LENGTH); +#endif -+ recv_ok = receive_data(f_in,mapbuf,fd2,fname,file->length,file->u.sum); + recv_ok = receive_data(f_in, fnamecmp, fd1, st.st_size, +- fname, fd2, file->length); ++ fname, fd2, file->length, file->u.sum); log_recv(file, &initial_stats); ---- rsync.c 15 May 2004 19:31:10 -0000 1.139 -+++ rsync.c 15 May 2004 20:29:17 -0000 -@@ -31,6 +31,7 @@ extern int am_generator; - extern int preserve_uid; - extern int preserve_gid; +--- orig/rsync.c 2005-02-07 20:41:57 ++++ rsync.c 2005-02-07 21:11:30 +@@ -36,6 +36,7 @@ extern int force_delete; + extern int recurse; + extern int keep_dirlinks; extern int make_backups; +extern char *link_by_hash_dir; + extern char *backup_dir; + extern int inplace; - - /* -@@ -236,8 +237,12 @@ void finish_transfer(char *fname, char * - if (make_backups && !make_backup(fname)) - return; - -- /* move tmp file over real file */ +@@ -287,7 +288,12 @@ void finish_transfer(char *fname, char * + rprintf(FINFO, "renaming %s to %s\n", + safe_fname(fnametmp), safe_fname(fname)); + } - ret = robust_rename(fnametmp, fname, file->mode & INITACCESSPERMS); -+#ifdef HAVE_LINK ++#if HAVE_LINK + if (link_by_hash_dir) -+ ret = link_by_hash(fnametmp,fname,file); ++ ret = link_by_hash(fnametmp, fname, file); + else +#endif + ret = robust_rename(fnametmp, fname, file->mode & INITACCESSPERMS); if (ret < 0) { rsyserr(FERROR, errno, "%s %s -> \"%s\"", ret == -2 ? "copy" : "rename", ---- rsync.h 13 May 2004 18:51:22 -0000 1.203 -+++ rsync.h 15 May 2004 20:29:18 -0000 -@@ -521,6 +521,14 @@ struct stats { +--- orig/rsync.h 2005-02-07 20:41:57 ++++ rsync.h 2004-07-03 20:20:15 +@@ -600,6 +600,14 @@ struct stats { int current_file_index; }; @@ -576,5 +558,15 @@ the file's name. +}; + - /* we need this function because of the silly way in which duplicate - entries are handled in the file lists - we can't change this + #include "byteorder.h" + #include "lib/mdfour.h" +--- orig/rsync.yo 2005-02-11 23:14:49 ++++ rsync.yo 2005-01-28 19:32:45 +@@ -354,6 +354,7 @@ to the detailed description below for a + --compare-dest=DIR also compare received files relative to DIR + --copy-dest=DIR ... and include copies of unchanged files + --link-dest=DIR hardlink to files in DIR when unchanged ++ --link-by-hash=DIR create hardlinks by hash to DIR for regular files + -z, --compress compress file data + -C, --cvs-exclude auto-ignore files in the same way CVS does + -f, --filter=RULE add a file-filtering RULE