| 1 | Wayne Davison wrote: |
| 2 | |
| 3 | I greatly simplified the changes to generator.c, making the patch |
| 4 | easier to maintain and fixing the failing test in the testsuite. |
| 5 | Very lightly tested (by me). |
| 6 | |
| 7 | --- Makefile.in 15 May 2004 00:48:11 -0000 1.101 |
| 8 | +++ Makefile.in 29 Jun 2004 17:46:12 -0000 |
| 9 | @@ -32,7 +32,7 @@ ZLIBOBJ=zlib/deflate.o zlib/infblock.o z |
| 10 | zlib/inflate.o zlib/inftrees.o zlib/infutil.o zlib/trees.o \ |
| 11 | zlib/zutil.o zlib/adler32.o |
| 12 | OBJS1=rsync.o generator.o receiver.o cleanup.o sender.o exclude.o util.o \ |
| 13 | - main.o checksum.o match.o syscall.o log.o backup.o |
| 14 | + main.o checksum.o match.o syscall.o log.o backup.o alternate.o |
| 15 | OBJS2=options.o flist.o io.o compat.o hlink.o token.o uidlist.o socket.o \ |
| 16 | fileio.o batch.o clientname.o |
| 17 | OBJS3=progress.o pipe.o |
| 18 | --- /dev/null 1 Jan 1970 00:00:00 -0000 |
| 19 | +++ alternate.c 29 Jun 2004 17:46:12 -0000 |
| 20 | @@ -0,0 +1,105 @@ |
| 21 | +#include "rsync.h" |
| 22 | + |
| 23 | +extern char *compare_dest; |
| 24 | +extern int verbose; |
| 25 | + |
| 26 | +/* Alternate methods for opening files, if local doesn't exist */ |
| 27 | +/* Sanity check that we are about to open regular file */ |
| 28 | +static int do_open_regular(char *fname) |
| 29 | +{ |
| 30 | + STRUCT_STAT st; |
| 31 | + |
| 32 | + if (do_stat(fname, &st) == 0 && S_ISREG(st.st_mode)) |
| 33 | + return do_open(fname, O_RDONLY, 0); |
| 34 | + |
| 35 | + return -1; |
| 36 | +} |
| 37 | + |
| 38 | +static void split_names(char *fname, char **dirname, char **basename) |
| 39 | +{ |
| 40 | + char *slash = strrchr(fname, '/'); |
| 41 | + if (slash) { |
| 42 | + *dirname = fname; |
| 43 | + *slash = '\0'; |
| 44 | + *basename = slash+1; |
| 45 | + } else { |
| 46 | + *basename = fname; |
| 47 | + *dirname = "."; |
| 48 | + } |
| 49 | +} |
| 50 | + |
| 51 | +static unsigned int measure_name(const char *name, const char *basename, |
| 52 | + const char *ext) |
| 53 | +{ |
| 54 | + int namelen = strlen(name); |
| 55 | + int extlen = strlen(ext); |
| 56 | + unsigned int score = 0; |
| 57 | + |
| 58 | + /* Extensions must match */ |
| 59 | + if (namelen <= extlen || strcmp(name + namelen - extlen, ext) != 0) |
| 60 | + return 0; |
| 61 | + |
| 62 | + /* Now score depends on similarity of prefix */ |
| 63 | + for (; *name == *basename && *name; name++, basename++) |
| 64 | + score++; |
| 65 | + return score; |
| 66 | +} |
| 67 | + |
| 68 | +int check_alternate_base_fuzzy(char **fname_ptr, char *buf, |
| 69 | + STRUCT_STAT *st_ptr) |
| 70 | +{ |
| 71 | + DIR *d; |
| 72 | + struct dirent *di; |
| 73 | + char *basename, *dirname; |
| 74 | + char mangled_name[MAXPATHLEN]; |
| 75 | + char bestname[MAXPATHLEN]; |
| 76 | + unsigned int bestscore = 0; |
| 77 | + const char *ext; |
| 78 | + |
| 79 | + strlcpy(mangled_name, *fname_ptr, sizeof mangled_name); |
| 80 | + |
| 81 | + split_names(mangled_name, &dirname, &basename); |
| 82 | + if (!(d = opendir(dirname))) { |
| 83 | + rsyserr(FERROR, errno, "recv_generator opendir(%s)", dirname); |
| 84 | + return -1; |
| 85 | + } |
| 86 | + |
| 87 | + /* Get final extension, eg. .gz; never full basename though. */ |
| 88 | + ext = strrchr(basename + 1, '.'); |
| 89 | + if (!ext) |
| 90 | + ext = basename + strlen(basename); /* ext = "" */ |
| 91 | + |
| 92 | + while ((di = readdir(d)) != NULL) { |
| 93 | + const char *dname = d_name(di); |
| 94 | + unsigned int score; |
| 95 | + |
| 96 | + if (dname[0] == '.' && (dname[1] == '\0' |
| 97 | + || (dname[1] == '.' && dname[2] == '\0'))) |
| 98 | + continue; |
| 99 | + |
| 100 | + score = measure_name(dname, basename, ext); |
| 101 | + if (verbose > 4) { |
| 102 | + rprintf(FINFO, "[%s] fuzzy score for %s = %u\n", |
| 103 | + who_am_i(), dname, score); |
| 104 | + } |
| 105 | + if (score > bestscore) { |
| 106 | + strlcpy(bestname, dname, sizeof bestname); |
| 107 | + bestscore = score; |
| 108 | + } |
| 109 | + } |
| 110 | + closedir(d); |
| 111 | + |
| 112 | + /* Found a candidate. */ |
| 113 | + if (bestscore != 0) { |
| 114 | + pathjoin(buf, MAXPATHLEN, dirname, bestname); |
| 115 | + if (verbose > 2) { |
| 116 | + rprintf(FINFO, "[%s] fuzzy match %s->%s\n", |
| 117 | + who_am_i(), *fname_ptr, buf); |
| 118 | + } |
| 119 | + *fname_ptr = buf; |
| 120 | + if (st_ptr) |
| 121 | + return link_stat(buf, st_ptr, 0); |
| 122 | + return do_open_regular(buf); |
| 123 | + } |
| 124 | + return -1; |
| 125 | +} |
| 126 | --- generator.c 29 Jun 2004 16:22:54 -0000 1.91 |
| 127 | +++ generator.c 29 Jun 2004 17:46:12 -0000 |
| 128 | @@ -41,6 +41,7 @@ extern int ignore_times; |
| 129 | extern int size_only; |
| 130 | extern int io_timeout; |
| 131 | extern int protocol_version; |
| 132 | +extern int fuzzy; |
| 133 | extern int always_checksum; |
| 134 | extern char *compare_dest; |
| 135 | extern int link_dest; |
| 136 | @@ -271,7 +272,7 @@ void recv_generator(char *fname, struct |
| 137 | int fd; |
| 138 | STRUCT_STAT st; |
| 139 | struct map_struct *mapbuf; |
| 140 | - int statret; |
| 141 | + int statret, fuzzy_file = 0; |
| 142 | char *fnamecmp; |
| 143 | char fnamecmpbuf[MAXPATHLEN]; |
| 144 | |
| 145 | @@ -447,6 +448,15 @@ void recv_generator(char *fname, struct |
| 146 | fnamecmp = fnamecmpbuf; |
| 147 | } |
| 148 | |
| 149 | + if (statret == -1 && fuzzy) { |
| 150 | + statret = check_alternate_base_fuzzy(&fnamecmp, fnamecmpbuf, |
| 151 | + &st); |
| 152 | + if (!S_ISREG(st.st_mode)) |
| 153 | + statret = -1; |
| 154 | + else |
| 155 | + fuzzy_file = 1; |
| 156 | + } |
| 157 | + |
| 158 | if (statret == -1) { |
| 159 | if (preserve_hard_links && hard_link_check(file, HL_SKIP)) |
| 160 | return; |
| 161 | @@ -489,7 +499,7 @@ void recv_generator(char *fname, struct |
| 162 | return; |
| 163 | } |
| 164 | |
| 165 | - if (skip_file(fname, file, &st)) { |
| 166 | + if (!fuzzy_file && skip_file(fname, file, &st)) { |
| 167 | if (fnamecmp == fname) |
| 168 | set_perms(fname, file, &st, PERMS_REPORT); |
| 169 | return; |
| 170 | --- options.c 20 Jun 2004 19:47:05 -0000 1.157 |
| 171 | +++ options.c 29 Jun 2004 17:46:13 -0000 |
| 172 | @@ -94,6 +94,7 @@ int ignore_errors = 0; |
| 173 | int modify_window = 0; |
| 174 | int blocking_io = -1; |
| 175 | int checksum_seed = 0; |
| 176 | +int fuzzy = 0; |
| 177 | unsigned int block_size = 0; |
| 178 | |
| 179 | |
| 180 | @@ -270,6 +271,7 @@ void usage(enum logcode F) |
| 181 | rprintf(F," -T --temp-dir=DIR create temporary files in directory DIR\n"); |
| 182 | rprintf(F," --compare-dest=DIR also compare destination files relative to DIR\n"); |
| 183 | rprintf(F," --link-dest=DIR create hardlinks to DIR for unchanged files\n"); |
| 184 | + rprintf(F," --fuzzy use similar file as basis if basis doesn't exist\n"); |
| 185 | rprintf(F," -P equivalent to --partial --progress\n"); |
| 186 | rprintf(F," -z, --compress compress file data\n"); |
| 187 | rprintf(F," -C, --cvs-exclude auto ignore files in the same way CVS does\n"); |
| 188 | @@ -368,6 +370,7 @@ static struct poptOption long_options[] |
| 189 | {"temp-dir", 'T', POPT_ARG_STRING, &tmpdir, 0, 0, 0 }, |
| 190 | {"compare-dest", 0, POPT_ARG_STRING, &compare_dest, 0, 0, 0 }, |
| 191 | {"link-dest", 0, POPT_ARG_STRING, &compare_dest, OPT_LINK_DEST, 0, 0 }, |
| 192 | + {"fuzzy", 0, POPT_ARG_NONE, &fuzzy, 0, 0, 0 }, |
| 193 | /* TODO: Should this take an optional int giving the compression level? */ |
| 194 | {"compress", 'z', POPT_ARG_NONE, &do_compression, 0, 0, 0 }, |
| 195 | {"daemon", 0, POPT_ARG_NONE, &daemon_opt, 0, 0, 0 }, |
| 196 | @@ -989,6 +992,9 @@ void server_options(char **args,int *arg |
| 197 | } |
| 198 | } |
| 199 | |
| 200 | + if (fuzzy && am_sender) |
| 201 | + args[ac++] = "--fuzzy"; |
| 202 | + |
| 203 | *argc = ac; |
| 204 | return; |
| 205 | |
| 206 | --- receiver.c 29 Jun 2004 15:12:01 -0000 1.83 |
| 207 | +++ receiver.c 29 Jun 2004 17:46:13 -0000 |
| 208 | @@ -48,6 +48,7 @@ extern int ignore_errors; |
| 209 | extern int orig_umask; |
| 210 | extern int keep_partial; |
| 211 | extern int checksum_seed; |
| 212 | +extern int fuzzy; |
| 213 | |
| 214 | static void delete_one(char *fn, int is_dir) |
| 215 | { |
| 216 | @@ -377,6 +378,11 @@ int recv_files(int f_in,struct file_list |
| 217 | fd1 = do_open(fnamecmp, O_RDONLY, 0); |
| 218 | } |
| 219 | |
| 220 | + if (fd1 == -1 && fuzzy) { |
| 221 | + fd1 = check_alternate_base_fuzzy(&fnamecmp, fnamecmpbuf, |
| 222 | + NULL); |
| 223 | + } |
| 224 | + |
| 225 | if (fd1 != -1 && do_fstat(fd1,&st) != 0) { |
| 226 | rsyserr(FERROR, errno, "fstat %s failed", |
| 227 | full_fname(fnamecmp)); |
| 228 | --- rsync.yo 5 Jun 2004 16:16:30 -0000 1.171 |
| 229 | +++ rsync.yo 29 Jun 2004 17:46:14 -0000 |
| 230 | @@ -325,6 +325,7 @@ verb( |
| 231 | -T --temp-dir=DIR create temporary files in directory DIR |
| 232 | --compare-dest=DIR also compare received files relative to DIR |
| 233 | --link-dest=DIR create hardlinks to DIR for unchanged files |
| 234 | + --fuzzy use similar file as basis if basis is gone |
| 235 | -P equivalent to --partial --progress |
| 236 | -z, --compress compress file data |
| 237 | -C, --cvs-exclude auto ignore files in the same way CVS does |