1 Depends-On-Patch: g2r-basis-filename.diff
3 The changes to generator.c were greatly simplified, making the patch
4 easier to maintain and fixing the failing test in the testsuite.
7 Be sure to run "make proto" before "make".
9 --- orig/generator.c 2004-09-07 21:55:02
10 +++ generator.c 2004-07-30 18:11:15
11 @@ -43,6 +43,7 @@ extern int ignore_times;
13 extern int io_timeout;
14 extern int protocol_version;
15 +extern int fuzzy_basis;
16 extern int always_checksum;
17 extern char *partial_dir;
18 extern char *compare_dest;
19 @@ -244,6 +245,93 @@ static void generate_and_send_sums(int f
23 +static void split_names(char *fname, char **dirname, char **basename)
25 + char *slash = strrchr(fname, '/');
29 + *basename = slash+1;
37 +static unsigned int measure_name(const char *name, const char *basename,
40 + int namelen = strlen(name);
41 + int extlen = strlen(ext);
42 + unsigned int score = 0;
44 + /* Extensions must match */
45 + if (namelen <= extlen || strcmp(name + namelen - extlen, ext) != 0)
48 + /* Now score depends on similarity of prefix */
49 + for (; *name == *basename && *name; name++, basename++)
55 +static int find_fuzzy(const char *fname, char *buf, STRUCT_STAT *st_ptr)
59 + char *basename, *dirname;
60 + char mangled_name[MAXPATHLEN];
61 + char bestname[MAXPATHLEN];
62 + unsigned int bestscore = 0;
65 + strlcpy(mangled_name, fname, sizeof mangled_name);
67 + split_names(mangled_name, &dirname, &basename);
68 + if (!(d = opendir(dirname))) {
69 + rsyserr(FERROR, errno, "recv_generator opendir(%s)", dirname);
73 + /* Get final extension, eg. .gz; never full basename though. */
74 + ext = strrchr(basename + 1, '.');
76 + ext = basename + strlen(basename); /* ext = "" */
78 + while ((di = readdir(d)) != NULL) {
79 + const char *dname = d_name(di);
82 + if (dname[0] == '.' && (dname[1] == '\0'
83 + || (dname[1] == '.' && dname[2] == '\0')))
86 + score = measure_name(dname, basename, ext);
88 + rprintf(FINFO, "[%s] fuzzy score for %s = %u\n",
89 + who_am_i(), dname, score);
91 + if (score > bestscore) {
92 + strlcpy(bestname, dname, sizeof bestname);
98 + /* Found a candidate. */
99 + if (bestscore != 0) {
100 + pathjoin(buf, MAXPATHLEN, dirname, bestname);
102 + rprintf(FINFO, "[%s] fuzzy match %s->%s\n",
103 + who_am_i(), fname, buf);
105 + return link_stat(buf, st_ptr, 0);
112 * Acts on file number @p i from @p flist, whose name is @p fname.
113 @@ -468,6 +556,15 @@ static void recv_generator(char *fname,
117 + if (statret == -1 && fuzzy_basis) {
118 + if (find_fuzzy(fname, fnamecmpbuf, &st) == 0
119 + && S_ISREG(st.st_mode)) {
121 + fnamecmp = fnamecmpbuf;
122 + fnamecmp_type = G2R_FUZZY;
127 if (preserve_hard_links && hard_link_check(file, HL_SKIP))
129 @@ -494,7 +591,7 @@ static void recv_generator(char *fname,
133 - if (skip_file(fnamecmp, file, &st)) {
134 + if (fnamecmp_type != G2R_FUZZY && skip_file(fnamecmp, file, &st)) {
135 if (fnamecmp_type == G2R_FNAME)
136 set_perms(fname, file, &st, PERMS_REPORT);
138 @@ -570,6 +667,21 @@ prepare_to_open:
140 if (f_out_name >= 0) {
141 write_byte(f_out_name, fnamecmp_type);
142 + if (fnamecmp_type == G2R_FUZZY) {
143 + uchar lenbuf[3], *lb = lenbuf;
144 + int len = strlen(fnamecmpbuf);
146 +#if MAXPATHLEN > 0x7FFF
147 + *lb++ = len / 0x10000 + 0x80;
148 + *lb++ = len / 0x100;
150 + *lb++ = len / 0x100 + 0x80;
154 + write_buf(f_out_name, lenbuf, lb - lenbuf + 1);
155 + write_buf(f_out_name, fnamecmpbuf, len);
157 io_flush(NORMAL_FLUSH); /* XXX make this more efficient! */
160 --- orig/main.c 2004-07-22 00:10:43
161 +++ main.c 2004-07-22 00:32:31
162 @@ -48,6 +48,7 @@ extern int keep_dirlinks;
163 extern int preserve_hard_links;
164 extern int protocol_version;
166 +extern int fuzzy_basis;
167 extern int relative_paths;
168 extern int rsync_port;
169 extern int whole_file;
170 @@ -463,7 +464,7 @@ static int do_recv(int f_in,int f_out,st
173 int error_pipe[2], name_pipe[2];
174 - BOOL need_name_pipe = compare_dest && !dry_run;
175 + BOOL need_name_pipe = (compare_dest || fuzzy_basis) && !dry_run;
177 /* The receiving side mustn't obey this, or an existing symlink that
178 * points to an identical file won't be replaced by the referent. */
179 --- orig/options.c 2004-09-07 21:45:30
180 +++ options.c 2004-07-16 20:14:12
181 @@ -85,6 +85,7 @@ int safe_symlinks = 0;
182 int copy_unsafe_links = 0;
185 +int fuzzy_basis = 0;
186 size_t bwlimit_writemax = 0;
187 int delete_after = 0;
188 int only_existing = 0;
189 @@ -279,6 +280,7 @@ void usage(enum logcode F)
190 rprintf(F," -T, --temp-dir=DIR create temporary files in directory DIR\n");
191 rprintf(F," --compare-dest=DIR also compare destination files relative to DIR\n");
192 rprintf(F," --link-dest=DIR create hardlinks to DIR for unchanged files\n");
193 + rprintf(F," --fuzzy use similar file as basis if basis doesn't exist\n");
194 rprintf(F," -P equivalent to --partial --progress\n");
195 rprintf(F," -z, --compress compress file data\n");
196 rprintf(F," -C, --cvs-exclude auto ignore files in the same way CVS does\n");
197 @@ -378,6 +380,7 @@ static struct poptOption long_options[]
198 {"temp-dir", 'T', POPT_ARG_STRING, &tmpdir, 0, 0, 0 },
199 {"compare-dest", 0, POPT_ARG_STRING, &compare_dest, 0, 0, 0 },
200 {"link-dest", 0, POPT_ARG_STRING, &compare_dest, OPT_LINK_DEST, 0, 0 },
201 + {"fuzzy", 0, POPT_ARG_NONE, &fuzzy_basis, 0, 0, 0 },
202 /* TODO: Should this take an optional int giving the compression level? */
203 {"compress", 'z', POPT_ARG_NONE, &do_compression, 0, 0, 0 },
204 {"daemon", 0, POPT_ARG_NONE, &daemon_opt, 0, 0, 0 },
205 @@ -1092,6 +1095,9 @@ void server_options(char **args,int *arg
209 + if (fuzzy_basis && am_sender)
210 + args[ac++] = "--fuzzy";
215 --- orig/receiver.c 2004-09-07 21:57:20
216 +++ receiver.c 2004-07-30 18:21:38
217 @@ -323,6 +323,27 @@ static int receive_data(int f_in, char *
221 +static void read_gen_name(int fd, char *buf)
223 + int len = read_byte(fd);
225 +#if MAXPATHLEN > 32767
227 + read_buf(fd, (char *)lenbuf, 2);
228 + len = (len & ~0x80) * 0x10000 + lenbuf[0] * 0x100 + lenbuf[1];
230 + len = (len & ~0x80) * 0x100 + read_byte(fd);
233 + if (len >= MAXPATHLEN) {
234 + rprintf(FERROR, "bogus data on generator name pipe\n");
235 + exit_cleanup(RERR_PROTOCOL);
238 + read_sbuf(fd, buf, len);
242 static void discard_receive_data(int f_in, OFF_T length)
244 receive_data(f_in, NULL, -1, 0, NULL, -1, length);
245 @@ -452,6 +473,10 @@ int recv_files(int f_in, struct file_lis
247 fnamecmp = get_backup_name(fname);
250 + read_gen_name(f_in_name, fnamecmpbuf);
251 + fnamecmp = fnamecmpbuf;
254 pathjoin(fnamecmpbuf, sizeof fnamecmpbuf,
255 compare_dest, fname);
256 --- orig/rsync.h 2004-09-07 21:52:22
257 +++ rsync.h 2004-09-07 22:02:37
259 #define G2R_FNAME 0x80
260 #define G2R_PARTIAL_DIR 0x81
261 #define G2R_BACKUP 0x82
262 +#define G2R_FUZZY 0x83
265 /* Log-message categories. FLOG is only used on the daemon side to
266 --- orig/rsync.yo 2004-09-07 19:54:51
267 +++ rsync.yo 2004-07-03 19:27:25
268 @@ -356,6 +356,7 @@ verb(
269 -T --temp-dir=DIR create temporary files in directory DIR
270 --compare-dest=DIR also compare received files relative to DIR
271 --link-dest=DIR create hardlinks to DIR for unchanged files
272 + --fuzzy use similar file as basis if basis is gone
273 -P equivalent to --partial --progress
274 -z, --compress compress file data
275 -C, --cvs-exclude auto ignore files in the same way CVS does