Commit | Line | Data |
---|---|---|
824abc86 | 1 | Depends-On-Patch: g2r-basis-filename.diff |
8c5b8235 | 2 | |
824abc86 | 3 | The changes to generator.c were greatly simplified, making the patch |
8c5b8235 | 4 | easier to maintain and fixing the failing test in the testsuite. |
58118c25 | 5 | Very lightly tested. |
241013b4 | 6 | |
824abc86 WD |
7 | Be sure to run "make proto" before "make". |
8 | ||
a6587818 | 9 | --- orig/generator.c 2004-09-20 19:57:58 |
09fb2223 | 10 | +++ generator.c 2004-07-30 18:11:15 |
f48a237e | 11 | @@ -43,6 +43,7 @@ extern int ignore_times; |
58118c25 WD |
12 | extern int size_only; |
13 | extern int io_timeout; | |
14 | extern int protocol_version; | |
09fb2223 | 15 | +extern int fuzzy_basis; |
58118c25 | 16 | extern int always_checksum; |
b952a177 | 17 | extern char *partial_dir; |
58118c25 | 18 | extern char *compare_dest; |
f48a237e | 19 | @@ -244,6 +245,93 @@ static void generate_and_send_sums(int f |
58118c25 WD |
20 | } |
21 | ||
22 | ||
47dd7a31 WD |
23 | +static void split_names(char *fname, char **dirname, char **basename) |
24 | +{ | |
25 | + char *slash = strrchr(fname, '/'); | |
26 | + if (slash) { | |
27 | + *dirname = fname; | |
28 | + *slash = '\0'; | |
29 | + *basename = slash+1; | |
30 | + } else { | |
31 | + *basename = fname; | |
32 | + *dirname = "."; | |
33 | + } | |
34 | +} | |
35 | + | |
58118c25 | 36 | + |
47dd7a31 WD |
37 | +static unsigned int measure_name(const char *name, const char *basename, |
38 | + const char *ext) | |
39 | +{ | |
40 | + int namelen = strlen(name); | |
41 | + int extlen = strlen(ext); | |
42 | + unsigned int score = 0; | |
43 | + | |
44 | + /* Extensions must match */ | |
45 | + if (namelen <= extlen || strcmp(name + namelen - extlen, ext) != 0) | |
46 | + return 0; | |
47 | + | |
48 | + /* Now score depends on similarity of prefix */ | |
49 | + for (; *name == *basename && *name; name++, basename++) | |
50 | + score++; | |
51 | + return score; | |
52 | +} | |
53 | + | |
58118c25 | 54 | + |
09fb2223 | 55 | +static int find_fuzzy(const char *fname, char *buf, STRUCT_STAT *st_ptr) |
47dd7a31 WD |
56 | +{ |
57 | + DIR *d; | |
58 | + struct dirent *di; | |
59 | + char *basename, *dirname; | |
60 | + char mangled_name[MAXPATHLEN]; | |
61 | + char bestname[MAXPATHLEN]; | |
62 | + unsigned int bestscore = 0; | |
63 | + const char *ext; | |
64 | + | |
09fb2223 | 65 | + strlcpy(mangled_name, fname, sizeof mangled_name); |
47dd7a31 WD |
66 | + |
67 | + split_names(mangled_name, &dirname, &basename); | |
68 | + if (!(d = opendir(dirname))) { | |
69 | + rsyserr(FERROR, errno, "recv_generator opendir(%s)", dirname); | |
70 | + return -1; | |
71 | + } | |
72 | + | |
73 | + /* Get final extension, eg. .gz; never full basename though. */ | |
74 | + ext = strrchr(basename + 1, '.'); | |
75 | + if (!ext) | |
76 | + ext = basename + strlen(basename); /* ext = "" */ | |
77 | + | |
78 | + while ((di = readdir(d)) != NULL) { | |
79 | + const char *dname = d_name(di); | |
80 | + unsigned int score; | |
81 | + | |
82 | + if (dname[0] == '.' && (dname[1] == '\0' | |
83 | + || (dname[1] == '.' && dname[2] == '\0'))) | |
84 | + continue; | |
85 | + | |
86 | + score = measure_name(dname, basename, ext); | |
87 | + if (verbose > 4) { | |
8c5b8235 WD |
88 | + rprintf(FINFO, "[%s] fuzzy score for %s = %u\n", |
89 | + who_am_i(), dname, score); | |
47dd7a31 WD |
90 | + } |
91 | + if (score > bestscore) { | |
8c5b8235 | 92 | + strlcpy(bestname, dname, sizeof bestname); |
47dd7a31 WD |
93 | + bestscore = score; |
94 | + } | |
95 | + } | |
96 | + closedir(d); | |
97 | + | |
98 | + /* Found a candidate. */ | |
99 | + if (bestscore != 0) { | |
8c5b8235 | 100 | + pathjoin(buf, MAXPATHLEN, dirname, bestname); |
47dd7a31 | 101 | + if (verbose > 2) { |
8c5b8235 | 102 | + rprintf(FINFO, "[%s] fuzzy match %s->%s\n", |
09fb2223 | 103 | + who_am_i(), fname, buf); |
47dd7a31 | 104 | + } |
58118c25 | 105 | + return link_stat(buf, st_ptr, 0); |
47dd7a31 WD |
106 | + } |
107 | + return -1; | |
108 | +} | |
58118c25 WD |
109 | + |
110 | ||
111 | /* | |
112 | * Acts on file number @p i from @p flist, whose name is @p fname. | |
f48a237e WD |
113 | @@ -468,6 +556,15 @@ static void recv_generator(char *fname, |
114 | } else | |
115 | partialptr = NULL; | |
824abc86 | 116 | |
09fb2223 WD |
117 | + if (statret == -1 && fuzzy_basis) { |
118 | + if (find_fuzzy(fname, fnamecmpbuf, &st) == 0 | |
119 | + && S_ISREG(st.st_mode)) { | |
120 | + statret = 0; | |
121 | + fnamecmp = fnamecmpbuf; | |
122 | + fnamecmp_type = G2R_FUZZY; | |
123 | + } | |
824abc86 WD |
124 | + } |
125 | + | |
09fb2223 WD |
126 | if (statret == -1) { |
127 | if (preserve_hard_links && hard_link_check(file, HL_SKIP)) | |
7628f156 | 128 | return; |
f48a237e | 129 | @@ -494,7 +591,7 @@ static void recv_generator(char *fname, |
241013b4 MP |
130 | return; |
131 | } | |
132 | ||
ec8473c6 | 133 | - if (skip_file(fnamecmp, file, &st)) { |
09fb2223 WD |
134 | + if (fnamecmp_type != G2R_FUZZY && skip_file(fnamecmp, file, &st)) { |
135 | if (fnamecmp_type == G2R_FNAME) | |
8c5b8235 WD |
136 | set_perms(fname, file, &st, PERMS_REPORT); |
137 | return; | |
f48a237e | 138 | @@ -570,6 +667,21 @@ prepare_to_open: |
09fb2223 WD |
139 | notify_others: |
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); | |
145 | + if (len > 0x7F) { | |
146 | +#if MAXPATHLEN > 0x7FFF | |
147 | + *lb++ = len / 0x10000 + 0x80; | |
148 | + *lb++ = len / 0x100; | |
149 | +#else | |
150 | + *lb++ = len / 0x100 + 0x80; | |
151 | +#endif | |
152 | + } | |
153 | + *lb = len; | |
154 | + write_buf(f_out_name, lenbuf, lb - lenbuf + 1); | |
155 | + write_buf(f_out_name, fnamecmpbuf, len); | |
156 | + } | |
157 | io_flush(NORMAL_FLUSH); /* XXX make this more efficient! */ | |
158 | } | |
159 | ||
2f5fa77e WD |
160 | --- orig/main.c 2004-07-22 00:10:43 |
161 | +++ main.c 2004-07-22 00:32:31 | |
d5753a22 | 162 | @@ -48,6 +48,7 @@ extern int keep_dirlinks; |
495f1899 WD |
163 | extern int preserve_hard_links; |
164 | extern int protocol_version; | |
165 | extern int recurse; | |
09fb2223 | 166 | +extern int fuzzy_basis; |
495f1899 WD |
167 | extern int relative_paths; |
168 | extern int rsync_port; | |
169 | extern int whole_file; | |
f635ed27 | 170 | @@ -463,7 +464,7 @@ static int do_recv(int f_in,int f_out,st |
495f1899 WD |
171 | int pid; |
172 | int status = 0; | |
173 | int error_pipe[2], name_pipe[2]; | |
c2530f70 | 174 | - BOOL need_name_pipe = compare_dest && !dry_run; |
09fb2223 | 175 | + BOOL need_name_pipe = (compare_dest || fuzzy_basis) && !dry_run; |
495f1899 | 176 | |
d5753a22 WD |
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. */ | |
a6587818 | 179 | --- orig/options.c 2004-09-20 05:10:48 |
f6c3b300 | 180 | +++ options.c 2004-07-16 20:14:12 |
f635ed27 | 181 | @@ -85,6 +85,7 @@ int safe_symlinks = 0; |
f6c3b300 WD |
182 | int copy_unsafe_links = 0; |
183 | int size_only = 0; | |
184 | int bwlimit = 0; | |
09fb2223 | 185 | +int fuzzy_basis = 0; |
f6c3b300 WD |
186 | size_t bwlimit_writemax = 0; |
187 | int delete_after = 0; | |
188 | int only_existing = 0; | |
f635ed27 | 189 | @@ -279,6 +280,7 @@ void usage(enum logcode F) |
c2530f70 | 190 | rprintf(F," -T, --temp-dir=DIR create temporary files in directory DIR\n"); |
f0533c4c WD |
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"); | |
f635ed27 | 197 | @@ -378,6 +380,7 @@ static struct poptOption long_options[] |
f0533c4c WD |
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 }, | |
09fb2223 | 201 | + {"fuzzy", 0, POPT_ARG_NONE, &fuzzy_basis, 0, 0, 0 }, |
f0533c4c WD |
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 }, | |
a6587818 | 205 | @@ -1095,6 +1098,9 @@ void server_options(char **args,int *arg |
f74d2272 | 206 | } |
241013b4 | 207 | } |
7b675ff5 | 208 | |
09fb2223 | 209 | + if (fuzzy_basis && am_sender) |
241013b4 | 210 | + args[ac++] = "--fuzzy"; |
7b675ff5 | 211 | + |
241013b4 | 212 | *argc = ac; |
f74d2272 | 213 | return; |
7b675ff5 | 214 | |
f48a237e | 215 | --- orig/receiver.c 2004-09-07 21:57:20 |
09fb2223 | 216 | +++ receiver.c 2004-07-30 18:21:38 |
a6587818 | 217 | @@ -319,6 +319,27 @@ static int receive_data(int f_in, char * |
09fb2223 WD |
218 | } |
219 | ||
220 | ||
221 | +static void read_gen_name(int fd, char *buf) | |
222 | +{ | |
223 | + int len = read_byte(fd); | |
224 | + if (len & 0x80) { | |
225 | +#if MAXPATHLEN > 32767 | |
226 | + uchar lenbuf[2]; | |
227 | + read_buf(fd, (char *)lenbuf, 2); | |
228 | + len = (len & ~0x80) * 0x10000 + lenbuf[0] * 0x100 + lenbuf[1]; | |
229 | +#else | |
230 | + len = (len & ~0x80) * 0x100 + read_byte(fd); | |
231 | +#endif | |
232 | + } | |
233 | + if (len >= MAXPATHLEN) { | |
234 | + rprintf(FERROR, "bogus data on generator name pipe\n"); | |
235 | + exit_cleanup(RERR_PROTOCOL); | |
236 | + } | |
237 | + | |
238 | + read_sbuf(fd, buf, len); | |
239 | +} | |
240 | + | |
241 | + | |
242 | static void discard_receive_data(int f_in, OFF_T length) | |
243 | { | |
244 | receive_data(f_in, NULL, -1, 0, NULL, -1, length); | |
a6587818 | 245 | @@ -448,6 +469,10 @@ int recv_files(int f_in, struct file_lis |
f48a237e WD |
246 | case G2R_BACKUP: |
247 | fnamecmp = get_backup_name(fname); | |
09fb2223 WD |
248 | break; |
249 | + case G2R_FUZZY: | |
250 | + read_gen_name(f_in_name, fnamecmpbuf); | |
251 | + fnamecmp = fnamecmpbuf; | |
252 | + break; | |
253 | default: | |
254 | pathjoin(fnamecmpbuf, sizeof fnamecmpbuf, | |
255 | compare_dest, fname); | |
f48a237e WD |
256 | --- orig/rsync.h 2004-09-07 21:52:22 |
257 | +++ rsync.h 2004-09-07 22:02:37 | |
258 | @@ -121,6 +121,7 @@ | |
09fb2223 WD |
259 | #define G2R_FNAME 0x80 |
260 | #define G2R_PARTIAL_DIR 0x81 | |
f48a237e WD |
261 | #define G2R_BACKUP 0x82 |
262 | +#define G2R_FUZZY 0x83 | |
09fb2223 WD |
263 | |
264 | ||
265 | /* Log-message categories. FLOG is only used on the daemon side to | |
a6587818 | 266 | --- orig/rsync.yo 2004-09-20 05:10:48 |
824abc86 | 267 | +++ rsync.yo 2004-07-03 19:27:25 |
67c08134 | 268 | @@ -356,6 +356,7 @@ verb( |
f0533c4c WD |
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 |