1: This patch adds the --detect-renamed option which makes rsync notice files
2: that either (1) match in size & modify-time (plus the basename, if possible)
3: or (2) match in size & checksum (when --checksum was also specified) and use
4: each match as an alternate basis file to speed up the transfer.
5:
6: The algorithm attempts to scan the receiving-side's files in an efficient
7: manner. If --delete[-before] is enabled, we'll take advantage of the
8: pre-transfer delete pass to prepare any alternate-basis-file matches we
9: might find. If --delete-before is not enabled, rsync does the rename scan
10: during the regular file-sending scan (scanning each directory right before
11: the generator starts updating files from that dir). In this latter mode,
12: rsync might delay the updating of a file (if no alternate-basis match was
13: yet found) until the full scan of the receiving side is complete, at which
14: point any delayed files are processed.
15:
16: I chose to hard-link the alternate-basis files into a ".~tmp~" subdir that
17: takes advantage of rsync's pre-existing partial-dir logic. This uses less
18: memory than trying to keep track of the matches internally, and also allows
19: any deletions or file-updates to occur normally without interfering with
20: these alternate-basis discoveries.
21:
22: To use this patch, run these commands for a successful build:
23:
24: patch -p1 <patches/detect-renamed.diff
25: ./configure (optional if already run)
26: make
27:
28: TODO:
29:
30: The routine that makes missing directories for files that get renamed
31: down into a new sub-hierarchy doesn't properly handle the case where some
32: path elements might exist but not be a dir yet. We need to either change
33: our stash-ahead algorithm (to not require unknown path elements) or we
34: need to create a better path-making routine.
35:
36: We need to never return a match from fattr_find() that has a basis
37: file. This will ensure that we don't try to give a renamed file to
38: a file that can't use it, while missing out on giving it to a file
39: that could use it.
40:
41: based-on: e94bad1c156fc3910f24e2b3b71a81b0b0bdeb70
42: diff --git a/backup.c b/backup.c
43: --- a/backup.c
44: +++ b/backup.c
45: @@ -162,7 +162,7 @@ char *get_backup_name(const char *fname)
46: int ret;
47: if (backup_dir_len > 1)
48: backup_dir_buf[backup_dir_len-1] = '\0';
49: - ret = make_path(backup_dir_buf, 0);
50: + ret = make_path(backup_dir_buf, ACCESSPERMS, 0);
51: if (backup_dir_len > 1)
52: backup_dir_buf[backup_dir_len-1] = '/';
53: if (ret < 0)
54: diff --git a/compat.c b/compat.c
55: --- a/compat.c
56: +++ b/compat.c
57: @@ -39,6 +39,7 @@ extern int checksum_seed;
58: extern int basis_dir_cnt;
59: extern int prune_empty_dirs;
60: extern int protocol_version;
61: +extern int detect_renamed;
62: extern int protect_args;
63: extern int preserve_uid;
64: extern int preserve_gid;
65: @@ -159,6 +160,7 @@ void set_allow_inc_recurse(void)
66: allow_inc_recurse = 0;
67: else if (!am_sender
68: && (delete_before || delete_after
69: + || detect_renamed
70: || delay_updates || prune_empty_dirs))
71: allow_inc_recurse = 0;
72: else if (am_server && !local_server
73: diff --git a/delete.c b/delete.c
74: --- a/delete.c
75: +++ b/delete.c
76: @@ -25,6 +25,7 @@
77: extern int am_root;
78: extern int make_backups;
79: extern int max_delete;
80: +extern int detect_renamed;
81: extern char *backup_dir;
82: extern char *backup_suffix;
83: extern int backup_suffix_len;
84: @@ -44,6 +45,8 @@ static inline int is_backup_file(char *fn)
85: * its contents, otherwise just checks for content. Returns DR_SUCCESS or
86: * DR_NOT_EMPTY. Note that fname must point to a MAXPATHLEN buffer! (The
87: * buffer is used for recursion, but returned unchanged.)
88: + *
89: + * Note: --detect-rename may use this routine with DEL_NO_DELETIONS set!
90: */
91: static enum delret delete_dir_contents(char *fname, uint16 flags)
92: {
93: @@ -63,7 +66,9 @@ static enum delret delete_dir_contents(char *fname, uint16 flags)
94: save_filters = push_local_filters(fname, dlen);
95:
96: non_perishable_cnt = 0;
97: + file_extra_cnt += SUM_EXTRA_CNT;
98: dirlist = get_dirlist(fname, dlen, 0);
99: + file_extra_cnt -= SUM_EXTRA_CNT;
100: ret = non_perishable_cnt ? DR_NOT_EMPTY : DR_SUCCESS;
101:
102: if (!dirlist->used)
103: @@ -103,7 +108,8 @@ static enum delret delete_dir_contents(char *fname, uint16 flags)
104: if (S_ISDIR(fp->mode)) {
105: if (delete_dir_contents(fname, flags | DEL_RECURSE) != DR_SUCCESS)
106: ret = DR_NOT_EMPTY;
107: - }
108: + } else if (detect_renamed && S_ISREG(fp->mode))
109: + look_for_rename(fp, fname);
110: if (delete_item(fname, fp->mode, flags) != DR_SUCCESS)
111: ret = DR_NOT_EMPTY;
112: }
113: @@ -126,6 +132,8 @@ static enum delret delete_dir_contents(char *fname, uint16 flags)
114: *
115: * Note that fbuf must point to a MAXPATHLEN buffer if the mode indicates it's
116: * a directory! (The buffer is used for recursion, but returned unchanged.)
117: + *
118: + * Also note: --detect-rename may use this routine with DEL_NO_DELETIONS set!
119: */
120: enum delret delete_item(char *fbuf, uint16 mode, uint16 flags)
121: {
122: @@ -153,6 +161,9 @@ enum delret delete_item(char *fbuf, uint16 mode, uint16 flags)
123: /* OK: try to delete the directory. */
124: }
125:
126: + if (flags & DEL_NO_DELETIONS)
127: + return DR_SUCCESS;
128: +
129: if (!(flags & DEL_MAKE_ROOM) && max_delete >= 0 && stats.deleted_files >= max_delete) {
130: skipped_deletes++;
131: return DR_AT_LIMIT;
132: diff --git a/flist.c b/flist.c
133: --- a/flist.c
134: +++ b/flist.c
135: @@ -64,6 +64,7 @@ extern int non_perishable_cnt;
136: extern int prune_empty_dirs;
137: extern int copy_links;
138: extern int copy_unsafe_links;
139: +extern int detect_renamed;
140: extern int protocol_version;
141: extern int sanitize_paths;
142: extern int munge_symlinks;
143: @@ -130,6 +131,8 @@ static int64 tmp_dev = -1, tmp_ino;
144: #endif
145: static char tmp_sum[MAX_DIGEST_LEN];
146:
147: +struct file_list the_fattr_list;
148: +
149: static char empty_sum[MAX_DIGEST_LEN];
150: static int flist_count_offset; /* for --delete --progress */
151: static int show_filelist_progress;
152: @@ -277,6 +280,45 @@ static inline int is_excluded(const char *fname, int is_dir, int filter_level)
153: return name_is_excluded(fname, is_dir ? NAME_IS_DIR : NAME_IS_FILE, filter_level);
154: }
155:
156: +static int fattr_compare(struct file_struct **file1, struct file_struct **file2)
157: +{
158: + struct file_struct *f1 = *file1;
159: + struct file_struct *f2 = *file2;
160: + int64 len1 = F_LENGTH(f1), len2 = F_LENGTH(f2);
161: + int diff;
162: +
163: + if (!f1->basename || !S_ISREG(f1->mode) || !len1) {
164: + if (!f2->basename || !S_ISREG(f2->mode) || !len2)
165: + return 0;
166: + return 1;
167: + }
168: + if (!f2->basename || !S_ISREG(f2->mode) || !len2)
169: + return -1;
170: +
171: + /* Don't use diff for values that are longer than an int. */
172: + if (len1 != len2)
173: + return len1 < len2 ? -1 : 1;
174: +
175: + if (always_checksum) {
176: + diff = u_memcmp(F_SUM(f1), F_SUM(f2), flist_csum_len);
177: + if (diff)
178: + return diff;
179: + } else if (f1->modtime != f2->modtime)
180: + return f1->modtime < f2->modtime ? -1 : 1;
181: +
182: + diff = u_strcmp(f1->basename, f2->basename);
183: + if (diff)
184: + return diff;
185: +
186: + if (f1->dirname == f2->dirname)
187: + return 0;
188: + if (!f1->dirname)
189: + return -1;
190: + if (!f2->dirname)
191: + return 1;
192: + return u_strcmp(f1->dirname, f2->dirname);
193: +}
194: +
195: static void send_directory(int f, struct file_list *flist,
196: char *fbuf, int len, int flags);
197:
198: @@ -2675,6 +2717,23 @@ struct file_list *recv_file_list(int f, int dir_ndx)
199: * for a non-relative transfer in recv_file_entry(). */
200: flist_sort_and_clean(flist, relative_paths);
201:
202: + if (detect_renamed) {
203: + int j = flist->used;
204: + the_fattr_list.used = j;
205: + the_fattr_list.files = new_array(struct file_struct *, j);
206: + memcpy(the_fattr_list.files, flist->files,
207: + j * sizeof (struct file_struct *));
208: + qsort(the_fattr_list.files, j,
209: + sizeof the_fattr_list.files[0], (int (*)())fattr_compare);
210: + the_fattr_list.low = 0;
211: + while (j-- > 0) {
212: + struct file_struct *fp = the_fattr_list.files[j];
213: + if (fp->basename && S_ISREG(fp->mode) && F_LENGTH(fp))
214: + break;
215: + }
216: + the_fattr_list.high = j;
217: + }
218: +
219: if (protocol_version < 30) {
220: /* Recv the io_error flag */
221: int err = read_int(f);
222: diff --git a/generator.c b/generator.c
223: --- a/generator.c
224: +++ b/generator.c
225: @@ -79,6 +79,7 @@ extern int always_checksum;
226: extern int flist_csum_len;
227: extern char *partial_dir;
228: extern int alt_dest_type;
229: +extern int detect_renamed;
230: extern int whole_file;
231: extern int list_only;
232: extern int read_batch;
233: @@ -97,11 +98,13 @@ extern char *tmpdir;
234: extern char *basis_dir[MAX_BASIS_DIRS+1];
235: extern struct file_list *cur_flist, *first_flist, *dir_flist;
236: extern filter_rule_list filter_list, daemon_filter_list;
237: +extern struct file_list the_fattr_list;
238:
239: int maybe_ATTRS_REPORT = 0;
240: int maybe_ATTRS_ACCURATE_TIME = 0;
241:
242: static dev_t dev_zero;
243: +static int unexplored_dirs = 1;
244: static int deldelay_size = 0, deldelay_cnt = 0;
245: static char *deldelay_buf = NULL;
246: static int deldelay_fd = -1;
247: @@ -269,14 +272,19 @@ static void do_delayed_deletions(char *delbuf)
248: * all the --delete-WHEN options. Note that the fbuf pointer must point to a
249: * MAXPATHLEN buffer with the name of the directory in it (the functions we
250: * call will append names onto the end, but the old dir value will be restored
251: - * on exit). */
252: -static void delete_in_dir(char *fbuf, struct file_struct *file, dev_t *fs_dev)
253: + * on exit).
254: + *
255: + * Note: --detect-rename may use this routine with DEL_NO_DELETIONS set!
256: + */
257: +static void delete_in_dir(char *fbuf, struct file_struct *file, dev_t *fs_dev,
258: + int del_flags)
259: {
260: static int already_warned = 0;
261: static struct hashtable *dev_tbl;
262: struct file_list *dirlist;
263: - char delbuf[MAXPATHLEN];
264: - int dlen, i;
265: + char *p, delbuf[MAXPATHLEN];
266: + unsigned remainder;
267: + int dlen, i, restore_dot = 0;
268:
269: if (!fbuf) {
270: change_local_filter_dir(NULL, 0, 0);
271: @@ -290,17 +298,22 @@ static void delete_in_dir(char *fbuf, struct file_struct *file, dev_t *fs_dev)
272: maybe_send_keepalive(time(NULL), MSK_ALLOW_FLUSH);
273:
274: if (io_error & IOERR_GENERAL && !ignore_errors) {
275: - if (already_warned)
276: + if (!already_warned) {
277: + rprintf(FINFO,
278: + "IO error encountered -- skipping file deletion\n");
279: + already_warned = 1;
280: + }
281: + if (!detect_renamed)
282: return;
283: - rprintf(FINFO,
284: - "IO error encountered -- skipping file deletion\n");
285: - already_warned = 1;
286: - return;
287: + del_flags |= DEL_NO_DELETIONS;
288: }
289:
290: dlen = strlen(fbuf);
291: change_local_filter_dir(fbuf, dlen, F_DEPTH(file));
292:
293: + if (detect_renamed)
294: + unexplored_dirs--;
295: +
296: if (one_file_system) {
297: if (!dev_tbl)
298: dev_tbl = hashtable_create(16, HT_KEY64);
299: @@ -316,6 +329,14 @@ static void delete_in_dir(char *fbuf, struct file_struct *file, dev_t *fs_dev)
300:
301: dirlist = get_dirlist(fbuf, dlen, 0);
302:
303: + p = fbuf + dlen;
304: + if (dlen == 1 && *fbuf == '.') {
305: + restore_dot = 1;
306: + p = fbuf;
307: + } else if (dlen != 1 || *fbuf != '/')
308: + *p++ = '/';
309: + remainder = MAXPATHLEN - (p - fbuf);
310: +
311: /* If an item in dirlist is not found in flist, delete it
312: * from the filesystem. */
313: for (i = dirlist->used; i--; ) {
314: @@ -328,6 +349,10 @@ static void delete_in_dir(char *fbuf, struct file_struct *file, dev_t *fs_dev)
315: f_name(fp, NULL));
316: continue;
317: }
318: + if (detect_renamed && S_ISREG(fp->mode)) {
319: + strlcpy(p, fp->basename, remainder);
320: + look_for_rename(fp, fbuf);
321: + }
322: /* Here we want to match regardless of file type. Replacement
323: * of a file with one of another type is handled separately by
324: * a delete_item call with a DEL_MAKE_ROOM flag. */
325: @@ -336,14 +361,19 @@ static void delete_in_dir(char *fbuf, struct file_struct *file, dev_t *fs_dev)
326: if (!(fp->mode & S_IWUSR) && !am_root && fp->flags & FLAG_OWNED_BY_US)
327: flags |= DEL_NO_UID_WRITE;
328: f_name(fp, delbuf);
329: - if (delete_during == 2) {
330: - if (!remember_delete(fp, delbuf, flags))
331: + if (delete_during == 2 && !(del_flags & DEL_NO_DELETIONS)) {
332: + if (!remember_delete(fp, delbuf, del_flags | flags))
333: break;
334: } else
335: - delete_item(delbuf, fp->mode, flags);
336: - }
337: + delete_item(delbuf, fp->mode, del_flags | flags);
338: + } else if (detect_renamed && S_ISDIR(fp->mode))
339: + unexplored_dirs++;
340: }
341:
342: + if (restore_dot)
343: + fbuf[0] = '.';
344: + fbuf[dlen] = '\0';
345: +
346: flist_free(dirlist);
347: }
348:
349: @@ -379,14 +409,125 @@ static void do_delete_pass(void)
350: || !S_ISDIR(st.st_mode))
351: continue;
352:
353: - delete_in_dir(fbuf, file, &st.st_dev);
354: + delete_in_dir(fbuf, file, &st.st_dev, 0);
355: }
356: - delete_in_dir(NULL, NULL, &dev_zero);
357: + delete_in_dir(NULL, NULL, &dev_zero, 0);
358:
359: if (INFO_GTE(FLIST, 2) && !am_server)
360: rprintf(FINFO, " \r");
361: }
362:
363: +/* Search for a regular file that matches either (1) the size & modified
364: + * time (plus the basename, if possible) or (2) the size & checksum. If
365: + * we find an exact match down to the dirname, return -1 because we found
366: + * an up-to-date file in the transfer, not a renamed file. */
367: +static int fattr_find(struct file_struct *f, char *fname)
368: +{
369: + int low = the_fattr_list.low, high = the_fattr_list.high;
370: + int mid, ok_match = -1, good_match = -1;
371: + struct file_struct *fmid;
372: + int diff;
373: +
374: + while (low <= high) {
375: + mid = (low + high) / 2;
376: + fmid = the_fattr_list.files[mid];
377: + if (F_LENGTH(fmid) != F_LENGTH(f)) {
378: + if (F_LENGTH(fmid) < F_LENGTH(f))
379: + low = mid + 1;
380: + else
381: + high = mid - 1;
382: + continue;
383: + }
384: + if (always_checksum) {
385: + /* We use the FLAG_FILE_SENT flag to indicate when we
386: + * have computed the checksum for an entry. */
387: + if (!(f->flags & FLAG_FILE_SENT)) {
388: + STRUCT_STAT st;
389: + if (fmid->modtime == f->modtime
390: + && f_name_cmp(fmid, f) == 0)
391: + return -1; /* assume we can't help */
392: + st.st_size = F_LENGTH(f);
393: + st.st_mtime = f->modtime;
394: + file_checksum(fname, &st, F_SUM(f));
395: + f->flags |= FLAG_FILE_SENT;
396: + }
397: + diff = u_memcmp(F_SUM(fmid), F_SUM(f), flist_csum_len);
398: + if (diff) {
399: + if (diff < 0)
400: + low = mid + 1;
401: + else
402: + high = mid - 1;
403: + continue;
404: + }
405: + } else {
406: + if (fmid->modtime != f->modtime) {
407: + if (fmid->modtime < f->modtime)
408: + low = mid + 1;
409: + else
410: + high = mid - 1;
411: + continue;
412: + }
413: + }
414: + ok_match = mid;
415: + diff = u_strcmp(fmid->basename, f->basename);
416: + if (diff == 0) {
417: + good_match = mid;
418: + if (fmid->dirname == f->dirname)
419: + return -1; /* file is up-to-date */
420: + if (!fmid->dirname) {
421: + low = mid + 1;
422: + continue;
423: + }
424: + if (!f->dirname) {
425: + high = mid - 1;
426: + continue;
427: + }
428: + diff = u_strcmp(fmid->dirname, f->dirname);
429: + if (diff == 0)
430: + return -1; /* file is up-to-date */
431: + }
432: + if (diff < 0)
433: + low = mid + 1;
434: + else
435: + high = mid - 1;
436: + }
437: +
438: + return good_match >= 0 ? good_match : ok_match;
439: +}
440: +
441: +void look_for_rename(struct file_struct *file, char *fname)
442: +{
443: + struct file_struct *fp;
444: + char *partialptr, *fn;
445: + STRUCT_STAT st;
446: + int ndx;
447: +
448: + if (!partial_dir || (ndx = fattr_find(file, fname)) < 0)
449: + return;
450: +
451: + fp = the_fattr_list.files[ndx];
452: + fn = f_name(fp, NULL);
453: + /* We don't provide an alternate-basis file if there is a basis file. */
454: + if (link_stat(fn, &st, 0) == 0)
455: + return;
456: +
457: + if (!dry_run) {
458: + if ((partialptr = partial_dir_fname(fn)) == NULL
459: + || !handle_partial_dir(partialptr, PDIR_CREATE))
460: + return;
461: + /* We only use the file if we can hard-link it into our tmp dir. */
462: + if (link(fname, partialptr) != 0) {
463: + if (errno != EEXIST)
464: + handle_partial_dir(partialptr, PDIR_DELETE);
465: + return;
466: + }
467: + }
468: +
469: + /* I think this falls into the -vv category with "%s is uptodate", etc. */
470: + if (INFO_GTE(MISC, 2))
471: + rprintf(FINFO, "found renamed: %s => %s\n", fname, fn);
472: +}
473: +
474: static inline int mtime_differs(STRUCT_STAT *stp, struct file_struct *file)
475: {
476: #ifdef ST_MTIME_NSEC
477: @@ -1187,6 +1328,7 @@ static void list_file_entry(struct file_struct *f)
478: }
479: }
480:
481: +static struct bitbag *delayed_bits = NULL;
482: static int phase = 0;
483: static int dflt_perms;
484:
485: @@ -1323,7 +1465,7 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
486: && do_stat(dn, &sx.st) < 0) {
487: if (dry_run)
488: goto parent_is_dry_missing;
489: - if (make_path(fname, MKP_DROP_NAME | MKP_SKIP_SLASH) < 0) {
490: + if (make_path(fname, ACCESSPERMS, MKP_DROP_NAME | MKP_SKIP_SLASH) < 0) {
491: rsyserr(FERROR_XFER, errno,
492: "recv_generator: mkdir %s failed",
493: full_fname(dn));
494: @@ -1459,7 +1601,7 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
495: }
496: if (real_ret != 0 && do_mkdir(fname,file->mode|added_perms) < 0 && errno != EEXIST) {
497: if (!relative_paths || errno != ENOENT
498: - || make_path(fname, MKP_DROP_NAME | MKP_SKIP_SLASH) < 0
499: + || make_path(fname, ACCESSPERMS, MKP_DROP_NAME | MKP_SKIP_SLASH) < 0
500: || (do_mkdir(fname, file->mode|added_perms) < 0 && errno != EEXIST)) {
501: rsyserr(FERROR_XFER, errno,
502: "recv_generator: mkdir %s failed",
503: @@ -1507,9 +1649,12 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
504: }
505: else if (delete_during && f_out != -1 && !phase
506: && !(file->flags & FLAG_MISSING_DIR)) {
507: - if (file->flags & FLAG_CONTENT_DIR)
508: - delete_in_dir(fname, file, &real_sx.st.st_dev);
509: - else
510: + if (file->flags & FLAG_CONTENT_DIR) {
511: + if (detect_renamed && real_ret != 0)
512: + unexplored_dirs++;
513: + delete_in_dir(fname, file, &real_sx.st.st_dev,
514: + delete_during < 0 ? DEL_NO_DELETIONS : 0);
515: + } else
516: change_local_filter_dir(fname, strlen(fname), F_DEPTH(file));
517: }
518: prior_dir_file = file;
519: @@ -1786,8 +1931,14 @@ static void recv_generator(char *fname, struct file_struct *file, int ndx,
520: goto cleanup;
521: }
522: #endif
523: - if (stat_errno == ENOENT)
524: + if (stat_errno == ENOENT) {
525: + if (detect_renamed && unexplored_dirs > 0
526: + && F_LENGTH(file)) {
527: + bitbag_set_bit(delayed_bits, ndx);
528: + return;
529: + }
530: goto notify_others;
531: + }
532: rsyserr(FERROR_XFER, stat_errno, "recv_generator: failed to stat %s",
533: full_fname(fname));
534: goto cleanup;
535: @@ -2251,6 +2402,12 @@ void generate_files(int f_out, const char *local_name)
536: if (DEBUG_GTE(GENR, 1))
537: rprintf(FINFO, "generator starting pid=%d\n", (int)getpid());
538:
539: + if (detect_renamed) {
540: + delayed_bits = bitbag_create(cur_flist->used);
541: + if (!delete_before && !delete_during)
542: + delete_during = -1;
543: + }
544: +
545: if (delete_before && !solo_file && cur_flist->used > 0)
546: do_delete_pass();
547: if (delete_during == 2) {
548: @@ -2259,7 +2416,7 @@ void generate_files(int f_out, const char *local_name)
549: }
550: info_levels[INFO_FLIST] = info_levels[INFO_PROGRESS] = 0;
551:
552: - if (append_mode > 0 || whole_file < 0)
553: + if (append_mode > 0 || detect_renamed || whole_file < 0)
554: whole_file = 0;
555: if (DEBUG_GTE(FLIST, 1)) {
556: rprintf(FINFO, "delta-transmission %s\n",
557: @@ -2295,7 +2452,7 @@ void generate_files(int f_out, const char *local_name)
558: dirdev = MAKEDEV(DEV_MAJOR(devp), DEV_MINOR(devp));
559: } else
560: dirdev = MAKEDEV(0, 0);
561: - delete_in_dir(fbuf, fp, &dirdev);
562: + delete_in_dir(fbuf, fp, &dirdev, 0);
563: } else
564: change_local_filter_dir(fbuf, strlen(fbuf), F_DEPTH(fp));
565: }
566: @@ -2342,7 +2499,21 @@ void generate_files(int f_out, const char *local_name)
567: } while ((cur_flist = cur_flist->next) != NULL);
568:
569: if (delete_during)
570: - delete_in_dir(NULL, NULL, &dev_zero);
571: + delete_in_dir(NULL, NULL, &dev_zero, 0);
572: + if (detect_renamed) {
573: + if (delete_during < 0)
574: + delete_during = 0;
575: + detect_renamed = 0;
576: +
577: + for (i = -1; (i = bitbag_next_bit(delayed_bits, i)) >= 0; ) {
578: + struct file_struct *file = cur_flist->files[i];
579: + if (local_name)
580: + strlcpy(fbuf, local_name, sizeof fbuf);
581: + else
582: + f_name(file, fbuf);
583: + recv_generator(fbuf, file, i, itemizing, code, f_out);
584: + }
585: + }
586: phase++;
587: if (DEBUG_GTE(GENR, 1))
588: rprintf(FINFO, "generate_files phase=%d\n", phase);
589: diff --git a/main.c b/main.c
590: --- a/main.c
591: +++ b/main.c
592: @@ -721,7 +721,7 @@ static char *get_local_name(struct file_list *flist, char *dest_path)
593: trailing_slash = cp && !cp[1];
594:
595: if (mkpath_dest_arg && statret < 0 && (cp || file_total > 1)) {
596: - int ret = make_path(dest_path, file_total > 1 && !trailing_slash ? 0 : MKP_DROP_NAME);
597: + int ret = make_path(dest_path, ACCESSPERMS, file_total > 1 && !trailing_slash ? 0 : MKP_DROP_NAME);
598: if (ret < 0)
599: goto mkdir_error;
600: if (INFO_GTE(NAME, 1)) {
601: diff --git a/options.c b/options.c
602: --- a/options.c
603: +++ b/options.c
604: @@ -84,6 +84,7 @@ int am_server = 0;
605: int am_sender = 0;
606: int am_starting_up = 1;
607: int relative_paths = -1;
608: +int detect_renamed = 0;
609: int implied_dirs = 1;
610: int missing_args = 0; /* 0 = FERROR_XFER, 1 = ignore, 2 = delete */
611: int numeric_ids = 0;
612: @@ -733,6 +734,7 @@ static struct poptOption long_options[] = {
613: {"compare-dest", 0, POPT_ARG_STRING, 0, OPT_COMPARE_DEST, 0, 0 },
614: {"copy-dest", 0, POPT_ARG_STRING, 0, OPT_COPY_DEST, 0, 0 },
615: {"link-dest", 0, POPT_ARG_STRING, 0, OPT_LINK_DEST, 0, 0 },
616: + {"detect-renamed", 0, POPT_ARG_NONE, &detect_renamed, 0, 0, 0 },
617: {"fuzzy", 'y', POPT_ARG_NONE, 0, 'y', 0, 0 },
618: {"no-fuzzy", 0, POPT_ARG_VAL, &fuzzy_basis, 0, 0, 0 },
619: {"no-y", 0, POPT_ARG_VAL, &fuzzy_basis, 0, 0, 0 },
620: @@ -2346,7 +2348,7 @@ int parse_arguments(int *argc_p, const char ***argv_p)
621: inplace = 1;
622: }
623:
624: - if (delay_updates && !partial_dir)
625: + if ((delay_updates || detect_renamed) && !partial_dir)
626: partial_dir = tmp_partialdir;
627:
628: if (inplace) {
629: @@ -2355,6 +2357,7 @@ int parse_arguments(int *argc_p, const char ***argv_p)
630: snprintf(err_buf, sizeof err_buf,
631: "--%s cannot be used with --%s\n",
632: append_mode ? "append" : "inplace",
633: + detect_renamed ? "detect-renamed" :
634: delay_updates ? "delay-updates" : "partial-dir");
635: return 0;
636: }
637: @@ -2760,6 +2763,8 @@ void server_options(char **args, int *argc_p)
638: args[ac++] = "--super";
639: if (size_only)
640: args[ac++] = "--size-only";
641: + if (detect_renamed)
642: + args[ac++] = "--detect-renamed";
643: if (do_stats)
644: args[ac++] = "--stats";
645: } else {
646: diff --git a/receiver.c b/receiver.c
647: --- a/receiver.c
648: +++ b/receiver.c
649: @@ -217,7 +217,7 @@ int open_tmpfile(char *fnametmp, const char *fname, struct file_struct *file)
650: * information should have been previously transferred, but that may
651: * not be the case with -R */
652: if (fd == -1 && relative_paths && errno == ENOENT
653: - && make_path(fnametmp, MKP_SKIP_SLASH | MKP_DROP_NAME) == 0) {
654: + && make_path(fnametmp, ACCESSPERMS, MKP_SKIP_SLASH | MKP_DROP_NAME) == 0) {
655: /* Get back to name with XXXXXX in it. */
656: get_tmpname(fnametmp, fname, False);
657: fd = do_mkstemp(fnametmp, (file->mode|added_perms) & INITACCESSPERMS);
658: diff --git a/rsync.1.md b/rsync.1.md
659: --- a/rsync.1.md
660: +++ b/rsync.1.md
661: @@ -421,6 +421,7 @@ detailed description below for a complete description.
662: --modify-window=NUM, -@ set the accuracy for mod-time comparisons
663: --temp-dir=DIR, -T create temporary files in directory DIR
664: --fuzzy, -y find similar file for basis if no dest file
665: +--detect-renamed try to find renamed files to speed the xfer
666: --compare-dest=DIR also compare destination files relative to DIR
667: --copy-dest=DIR ... and include copies of unchanged files
668: --link-dest=DIR hardlink to files in DIR when unchanged
669: @@ -2247,6 +2248,22 @@ your home directory (remove the '=' for that).
670: fuzzy-match files, so either use `--delete-after` or specify some filename
671: exclusions if you need to prevent this.
672:
673: +0. `--detect-renamed`
674: +
675: + With this option, for each new source file (call it `src/S`), rsync looks
676: + for a file `dest/D` anywhere in the destination that passes the quick check
677: + with `src/S`. If such a `dest/D` is found, rsync uses it as an alternate
678: + basis for transferring `S`. The idea is that if `src/S` was renamed from
679: + `src/D` (as opposed to `src/S` passing the quick check with `dest/D` by
680: + coincidence), the delta-transfer algorithm will find that all the data
681: + matches between `src/S` and `dest/D`, and the transfer will be really fast.
682: +
683: + By default, alternate-basis files are hard-linked into a directory named
684: + ".~tmp~" in each file's destination directory, but if you've specified the
685: + `--partial-dir` option, that directory will be used instead. These
686: + otential alternate-basis files will be removed as the transfer progresses.
687: + This option conflicts with `--inplace` and `--append`.
688: +
689: 0. `--compare-dest=DIR`
690:
691: This option instructs rsync to use _DIR_ on the destination machine as an
692: diff --git a/rsync.h b/rsync.h
693: --- a/rsync.h
694: +++ b/rsync.h
695: @@ -272,7 +272,7 @@ enum msgcode {
696: #define NDX_DEL_STATS -3
697: #define NDX_FLIST_OFFSET -101
698:
699: -/* For calling delete_item() and delete_dir_contents(). */
700: +/* For calling delete_item(), delete_dir_contents(), and delete_in_dir(). */
701: #define DEL_NO_UID_WRITE (1<<0) /* file/dir has our uid w/o write perm */
702: #define DEL_RECURSE (1<<1) /* if dir, delete all contents */
703: #define DEL_DIR_IS_EMPTY (1<<2) /* internal delete_FUNCTIONS use only */
704: @@ -282,6 +282,7 @@ enum msgcode {
705: #define DEL_FOR_DEVICE (1<<6) /* making room for a replacement device */
706: #define DEL_FOR_SPECIAL (1<<7) /* making room for a replacement special */
707: #define DEL_FOR_BACKUP (1<<8) /* the delete is for a backup operation */
708: +#define DEL_NO_DELETIONS (1<<9) /* just check for renames w/o deleting */
709:
710: #define DEL_MAKE_ROOM (DEL_FOR_FILE|DEL_FOR_DIR|DEL_FOR_SYMLINK|DEL_FOR_DEVICE|DEL_FOR_SPECIAL)
711:
712: diff --git a/util.c b/util.c
713: --- a/util.c
714: +++ b/util.c
715: @@ -182,7 +182,7 @@ int set_times(const char *fname, STRUCT_STAT *stp)
716: /* Create any necessary directories in fname. Any missing directories are
717: * created with default permissions. Returns < 0 on error, or the number
718: * of directories created. */
719: -int make_path(char *fname, int flags)
720: +int make_path(char *fname, mode_t mode, int flags)
721: {
722: char *end, *p;
723: int ret = 0;
724: @@ -213,7 +213,7 @@ int make_path(char *fname, int flags)
725: else
726: errno = ENOTDIR;
727: }
728: - } else if (do_mkdir(fname, ACCESSPERMS) == 0) {
729: + } else if (do_mkdir(fname, mode) == 0) {
730: ret++;
731: break;
732: }
733: @@ -252,7 +252,7 @@ int make_path(char *fname, int flags)
734: p += strlen(p);
735: if (ret < 0) /* Skip mkdir on error, but keep restoring the path. */
736: continue;
737: - if (do_mkdir(fname, ACCESSPERMS) < 0)
738: + if (do_mkdir(fname, mode) < 0)
739: ret = -ret - 1;
740: else
741: ret++;
742: @@ -1162,6 +1162,32 @@ char *normalize_path(char *path, BOOL force_newbuf, unsigned int *len_ptr)
743: return path;
744: }
745:
746: +/* We need to supply our own strcmp function for file list comparisons
747: + * to ensure that signed/unsigned usage is consistent between machines. */
748: +int u_strcmp(const char *p1, const char *p2)
749: +{
750: + for ( ; *p1; p1++, p2++) {
751: + if (*p1 != *p2)
752: + break;
753: + }
754: +
755: + return (int)*(uchar*)p1 - (int)*(uchar*)p2;
756: +}
757: +
758: +/* We need a memcmp function compares unsigned-byte values. */
759: +int u_memcmp(const void *p1, const void *p2, size_t len)
760: +{
761: + const uchar *u1 = p1;
762: + const uchar *u2 = p2;
763: +
764: + while (len--) {
765: + if (*u1 != *u2)
766: + return (int)*u1 - (int)*u2;
767: + }
768: +
769: + return 0;
770: +}
771: +
772: /**
773: * Return a quoted string with the full pathname of the indicated filename.
774: * The string " (in MODNAME)" may also be appended. The returned pointer
775: @@ -1255,7 +1281,7 @@ int handle_partial_dir(const char *fname, int create)
776: }
777: statret = -1;
778: }
779: - if (statret < 0 && do_mkdir(dir, 0700) < 0) {
780: + if (statret < 0 && make_path(dir, 0700, 0) < 0) {
781: *fn = '/';
782: return 0;
783: }
784: diff -Nurp a/rsync.1 b/rsync.1
785: --- a/rsync.1
786: +++ b/rsync.1
787: @@ -497,6 +497,7 @@ detailed description below for a complet
788: --modify-window=NUM, -@ set the accuracy for mod-time comparisons
789: --temp-dir=DIR, -T create temporary files in directory DIR
790: --fuzzy, -y find similar file for basis if no dest file
791: +--detect-renamed try to find renamed files to speed the xfer
792: --compare-dest=DIR also compare destination files relative to DIR
793: --copy-dest=DIR ... and include copies of unchanged files
794: --link-dest=DIR hardlink to files in DIR when unchanged
795: @@ -2290,6 +2291,20 @@ alternate destination directories that a
796: Note that the use of the \fB\-\-delete\fP option might get rid of any potential
797: fuzzy-match files, so either use \fB\-\-delete-after\fP or specify some filename
798: exclusions if you need to prevent this.
799: +.IP "\fB\-\-detect-renamed\fP"
800: +With this option, for each new source file (call it \fBsrc/S\fP), rsync looks
801: +for a file \fBdest/D\fP anywhere in the destination that passes the quick check
802: +with \fBsrc/S\fP. If such a \fBdest/D\fP is found, rsync uses it as an alternate
803: +basis for transferring \fBS\fP. The idea is that if \fBsrc/S\fP was renamed from
804: +\fBsrc/D\fP (as opposed to \fBsrc/S\fP passing the quick check with \fBdest/D\fP by
805: +coincidence), the delta-transfer algorithm will find that all the data
806: +matches between \fBsrc/S\fP and \fBdest/D\fP, and the transfer will be really fast.
807: +.IP
808: +By default, alternate-basis files are hard-linked into a directory named
809: +".~tmp~" in each file's destination directory, but if you've specified the
810: +\fB\-\-partial-dir\fP option, that directory will be used instead. These
811: +otential alternate-basis files will be removed as the transfer progresses.
812: +This option conflicts with \fB\-\-inplace\fP and \fB\-\-append\fP.
813: .IP "\fB\-\-compare-dest=DIR\fP"
814: This option instructs rsync to use \fIDIR\fP on the destination machine as an
815: additional hierarchy to compare destination files against doing transfers
816: diff -Nurp a/rsync.1.html b/rsync.1.html
817: --- a/rsync.1.html
818: +++ b/rsync.1.html
819: @@ -412,6 +412,7 @@ detailed description below for a complet
820: --modify-window=NUM, -@ set the accuracy for mod-time comparisons
821: --temp-dir=DIR, -T create temporary files in directory DIR
822: --fuzzy, -y find similar file for basis if no dest file
823: +--detect-renamed try to find renamed files to speed the xfer
824: --compare-dest=DIR also compare destination files relative to DIR
825: --copy-dest=DIR ... and include copies of unchanged files
826: --link-dest=DIR hardlink to files in DIR when unchanged
827: @@ -2135,6 +2136,21 @@ fuzzy-match files, so either use <code>-
828: exclusions if you need to prevent this.</p>
829: </dd>
830:
831: +<dt><code>--detect-renamed</code></dt><dd>
832: +<p>With this option, for each new source file (call it <code>src/S</code>), rsync looks
833: +for a file <code>dest/D</code> anywhere in the destination that passes the quick check
834: +with <code>src/S</code>. If such a <code>dest/D</code> is found, rsync uses it as an alternate
835: +basis for transferring <code>S</code>. The idea is that if <code>src/S</code> was renamed from
836: +<code>src/D</code> (as opposed to <code>src/S</code> passing the quick check with <code>dest/D</code> by
837: +coincidence), the delta-transfer algorithm will find that all the data
838: +matches between <code>src/S</code> and <code>dest/D</code>, and the transfer will be really fast.</p>
839: +<p>By default, alternate-basis files are hard-linked into a directory named
840: +".~tmp~" in each file's destination directory, but if you've specified the
841: +<code>--partial-dir</code> option, that directory will be used instead. These
842: +otential alternate-basis files will be removed as the transfer progresses.
843: +This option conflicts with <code>--inplace</code> and <code>--append</code>.</p>
844: +</dd>
845: +
846: <dt><code>--compare-dest=DIR</code></dt><dd>
847: <p>This option instructs rsync to use <u>DIR</u> on the destination machine as an
848: additional hierarchy to compare destination files against doing transfers
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/icons/cvsweb/back.gif){kind=icon}
Return to detect-renamed.diff CVS log ![[TXT]](/icons/cvsweb/text.gif){kind=icon}
![[DIR]](/icons/cvsweb/dir.gif){kind=icon}
Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / rsync / patches