Commit | Line | Data |
---|---|---|
46ef7d1d | 1 | -*- indented-text -*- |
a0365806 | 2 | |
259c3e72 MP |
3 | BUGS --------------------------------------------------------------- |
4 | ||
5 | There seems to be a bug with hardlinks | |
6 | ||
7 | mbp/2 build$ ls -l /tmp/a /tmp/b -i | |
8 | /tmp/a: | |
9 | total 32 | |
10 | 2568307 -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a1 | |
11 | 2568307 -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a2 | |
12 | 2568307 -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a3 | |
13 | 2568310 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 a4 | |
14 | 2568310 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 a5 | |
15 | 2568310 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b1 | |
16 | 2568310 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b2 | |
17 | 2568310 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b3 | |
18 | ||
19 | /tmp/b: | |
20 | total 32 | |
21 | 2568309 -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a1 | |
22 | 2568309 -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a2 | |
23 | 2568309 -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a3 | |
24 | 2568311 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 a4 | |
25 | 2568311 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 a5 | |
26 | 2568311 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b1 | |
27 | 2568311 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b2 | |
28 | 2568311 -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b3 | |
29 | mbp/2 build$ rm -r /tmp/b && ./rsync -avH /tmp/a/ /tmp/b | |
30 | building file list ... done | |
31 | created directory /tmp/b | |
32 | ./ | |
33 | a1 | |
34 | a4 | |
35 | a2 => a1 | |
36 | a3 => a2 | |
37 | wrote 350 bytes read 52 bytes 804.00 bytes/sec | |
38 | total size is 232 speedup is 0.58 | |
39 | mbp/2 build$ rm -r /tmp/b | |
40 | mbp/2 build$ ls -l /tmp/b | |
41 | ls: /tmp/b: No such file or directory | |
42 | mbp/2 build$ rm -r /tmp/b && ./rsync -avH /tmp/a/ /tmp/b | |
43 | rm: cannot remove `/tmp/b': No such file or directory | |
44 | mbp/2 build$ rm -f -r /tmp/b && ./rsync -avH /tmp/a/ /tmp/b | |
45 | building file list ... done | |
46 | created directory /tmp/b | |
47 | ./ | |
48 | a1 | |
49 | a4 | |
50 | a2 => a1 | |
51 | a3 => a2 | |
52 | wrote 350 bytes read 52 bytes 804.00 bytes/sec | |
53 | total size is 232 speedup is 0.58 | |
54 | mbp/2 build$ ls -l /tmp/b | |
55 | total 32 | |
56 | -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a1 | |
57 | -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a2 | |
58 | -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a3 | |
59 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 a4 | |
60 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 a5 | |
61 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b1 | |
62 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b2 | |
63 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b3 | |
64 | mbp/2 build$ ls -l /tmp/a | |
65 | total 32 | |
66 | -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a1 | |
67 | -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a2 | |
68 | -rw-rw-r-- 3 mbp mbp 29 Mar 25 17:30 a3 | |
69 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 a4 | |
70 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 a5 | |
71 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b1 | |
72 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b2 | |
73 | -rw-rw-r-- 5 mbp mbp 29 Mar 25 17:30 b3 | |
46ef7d1d | 74 | |
33d213bb | 75 | |
e4724e5c MP |
76 | Progress indicator can produce corrupt output when transferring directories: |
77 | ||
78 | main/binary-arm/ | |
79 | main/binary-arm/admin/ | |
80 | main/binary-arm/base/ | |
81 | main/binary-arm/comm/8.56kB/s 0:00:52 | |
82 | main/binary-arm/devel/ | |
83 | main/binary-arm/doc/ | |
84 | main/binary-arm/editors/ | |
85 | main/binary-arm/electronics/s 0:00:53 | |
86 | main/binary-arm/games/ | |
87 | main/binary-arm/graphics/ | |
88 | main/binary-arm/hamradio/ | |
89 | main/binary-arm/interpreters/ | |
90 | main/binary-arm/libs/6.61kB/s 0:00:54 | |
91 | main/binary-arm/mail/ | |
92 | main/binary-arm/math/ | |
93 | main/binary-arm/misc/ | |
94 | ||
7e28fca1 | 95 | |
e4724e5c | 96 | lchmod |
e4724e5c | 97 | I don't think we handle this properly on systems that don't have the |
7e28fca1 MP |
98 | call. Are there any such? |
99 | ||
e4724e5c | 100 | |
5ba268ef MP |
101 | Cross-test versions |
102 | Part of the regression suite should be making sure that we don't | |
103 | break backwards compatibility: old clients vs new servers and so | |
104 | on. Ideally we would test the cross product of versions. | |
105 | ||
106 | It might be sufficient to test downloads from well-known public | |
107 | rsync servers running different versions of rsync. This will give | |
108 | some testing and also be the most common case for having different | |
109 | versions and not being able to upgrade. | |
110 | ||
8bd1a73e MP |
111 | Do not rely on having a group called "nobody" |
112 | ||
113 | http://www.linuxbase.org/spec/refspecs/LSB_1.1.0/gLSB/usernames.html | |
114 | ||
115 | On Debian it's "nogroup" | |
e4724e5c | 116 | |
3d90ec14 MP |
117 | DAEMON -------------------------------------------------------------- |
118 | ||
119 | server-imposed bandwidth limits | |
120 | ||
121 | rsyncd over ssh | |
122 | ||
123 | There are already some patches to do this. | |
124 | ||
125 | BitKeeper uses a server whose login shell is set to bkd. That's | |
126 | probably a reasonable approach. | |
127 | ||
128 | ||
129 | FEATURES ------------------------------------------------------------ | |
33d213bb | 130 | |
33d213bb | 131 | |
32e83406 | 132 | --dry-run is too dry |
642a979a MP |
133 | |
134 | Mark Santcroos points out that -n fails to list files which have | |
a577af90 | 135 | only metadata changes, though it probably should. |
642a979a MP |
136 | |
137 | There may be a Debian bug about this as well. | |
138 | ||
139 | ||
a2d2e5c0 MP |
140 | use chroot |
141 | ||
142 | If the platform doesn't support it, then don't even try. | |
143 | ||
144 | If running as non-root, then don't fail, just give a warning. | |
145 | (There was a thread about this a while ago?) | |
146 | ||
147 | http://lists.samba.org/pipermail/rsync/2001-August/thread.html | |
148 | http://lists.samba.org/pipermail/rsync/2001-September/thread.html | |
149 | ||
642a979a | 150 | |
595f2d4d MP |
151 | supplementary groups |
152 | ||
153 | Perhaps allow supplementary groups to be specified in rsyncd.conf; | |
154 | then make the first one the primary gid and all the rest be | |
155 | supplementary gids. | |
156 | ||
157 | ||
8f4455f2 MP |
158 | File list structure in memory |
159 | ||
160 | Rather than one big array, perhaps have a tree in memory mirroring | |
a577af90 | 161 | the directory tree. |
8f4455f2 MP |
162 | |
163 | This might make sorting much faster! (I'm not sure it's a big CPU | |
a577af90 | 164 | problem, mind you.) |
8f4455f2 MP |
165 | |
166 | It might also reduce memory use in storing repeated directory names | |
a577af90 | 167 | -- again I'm not sure this is a problem. |
0e5a1f83 | 168 | |
a2d2e5c0 MP |
169 | Performance |
170 | ||
171 | Traverse just one directory at a time. Tridge says it's possible. | |
a6a3c3df MP |
172 | |
173 | At the moment rsync reads the whole file list into memory at the | |
174 | start, which makes us use a lot of memory and also not pipeline | |
175 | network access as much as we could. | |
176 | ||
0e5a1f83 MP |
177 | |
178 | Handling duplicate names | |
179 | ||
b3e6c815 | 180 | We need to be careful of duplicate names getting into the file list. |
d2e9d069 MP |
181 | See clean_flist(). This could happen if multiple arguments include |
182 | the same file. Bad. | |
b3e6c815 MP |
183 | |
184 | I think duplicates are only a problem if they're both flowing | |
185 | through the pipeline at the same time. For example we might have | |
186 | updated the first occurrence after reading the checksums for the | |
187 | second. So possibly we just need to make sure that we don't have | |
a577af90 | 188 | both in the pipeline at the same time. |
b3e6c815 MP |
189 | |
190 | Possibly if we did one directory at a time that would be sufficient. | |
191 | ||
192 | Alternatively we could pre-process the arguments to make sure no | |
58379559 MP |
193 | duplicates will ever be inserted. There could be some bad cases |
194 | when we're collapsing symlinks. | |
b3e6c815 MP |
195 | |
196 | We could have a hash table. | |
197 | ||
d2e9d069 MP |
198 | The root of the problem is that we do not want more than one file |
199 | list entry referring to the same file. At first glance there are | |
200 | several ways this could happen: symlinks, hardlinks, and repeated | |
201 | names on the command line. | |
202 | ||
203 | If names are repeated on the command line, they may be present in | |
204 | different forms, perhaps by traversing directory paths in different | |
205 | ways, traversing paths including symlinks. Also we need to allow | |
206 | for expansion of globs by rsync. | |
207 | ||
208 | At the moment, clean_flist() requires having the entire file list in | |
209 | memory. Duplicate names are detected just by a string comparison. | |
210 | ||
211 | We don't need to worry about hard links causing duplicates because | |
212 | files are never updated in place. Similarly for symlinks. | |
213 | ||
214 | I think even if we're using a different symlink mode we don't need | |
215 | to worry. | |
216 | ||
0e5a1f83 MP |
217 | Unless we're really clever this will introduce a protocol |
218 | incompatibility, so we need to be able to accept the old format as | |
219 | well. | |
220 | ||
221 | ||
a6a3c3df MP |
222 | Memory accounting |
223 | ||
224 | At exit, show how much memory was used for the file list, etc. | |
225 | ||
b3e6c815 MP |
226 | Also we do a wierd exponential-growth allocation in flist.c. I'm |
227 | not sure this makes sense with modern mallocs. At any rate it will | |
228 | make us allocate a huge amount of memory for large file lists. | |
229 | ||
0e5a1f83 | 230 | |
a6a3c3df MP |
231 | Hard-link handling |
232 | ||
233 | At the moment hardlink handling is very expensive, so it's off by | |
a577af90 | 234 | default. It does not need to be so. |
a6a3c3df | 235 | |
0e5a1f83 MP |
236 | Since most of the solutions are rather intertwined with the file |
237 | list it is probably better to fix that first, although fixing | |
238 | hardlinks is possibly simpler. | |
239 | ||
a6a3c3df MP |
240 | We can rule out hardlinked directories since they will probably |
241 | screw us up in all kinds of ways. They simply should not be used. | |
242 | ||
243 | At the moment rsync only cares about hardlinks to regular files. I | |
244 | guess you could also use them for sockets, devices and other beasts, | |
245 | but I have not seen them. | |
246 | ||
247 | When trying to reproduce hard links, we only need to worry about | |
a577af90 | 248 | files that have more than one name (nlinks>1 && !S_ISDIR). |
a6a3c3df MP |
249 | |
250 | The basic point of this is to discover alternate names that refer to | |
251 | the same file. All operations, including creating the file and | |
252 | writing modifications to it need only to be done for the first name. | |
253 | For all later names, we just create the link and then leave it | |
254 | alone. | |
255 | ||
256 | If hard links are to be preserved: | |
257 | ||
258 | Before the generator/receiver fork, the list of files is received | |
259 | from the sender (recv_file_list), and a table for detecting hard | |
260 | links is built. | |
261 | ||
262 | The generator looks for hard links within the file list and does | |
263 | not send checksums for them, though it does send other metadata. | |
264 | ||
265 | The sender sends the device number and inode with file entries, so | |
266 | that files are uniquely identified. | |
267 | ||
268 | The receiver goes through and creates hard links (do_hard_links) | |
269 | after all data has been written, but before directory permissions | |
270 | are set. | |
271 | ||
272 | At the moment device and inum are sent as 4-byte integers, which | |
273 | will probably cause problems on large filesystems. On Linux the | |
274 | kernel uses 64-bit ino_t's internally, and people will soon have | |
275 | filesystems big enough to use them. We ought to follow NFS4 in | |
276 | using 64-bit device and inode identification, perhaps with a | |
277 | protocol version bump. | |
278 | ||
279 | Once we've seen all the names for a particular file, we no longer | |
280 | need to think about it and we can deallocate the memory. | |
281 | ||
282 | We can also have the case where there are links to a file that are | |
283 | not in the tree being transferred. There's nothing we can do about | |
284 | that. Because we rename the destination into place after writing, | |
285 | any hardlinks to the old file are always going to be orphaned. In | |
286 | fact that is almost necessary because otherwise we'd get really | |
287 | confused if we were generating checksums for one name of a file and | |
288 | modifying another. | |
289 | ||
290 | At the moment the code seems to make a whole second copy of the file | |
291 | list, which seems unnecessary. | |
292 | ||
293 | We should have a test case that exercises hard links. Since it | |
294 | might be hard to compare ./tls output where the inodes change we | |
295 | might need a little program to check whether several names refer to | |
296 | the same file. | |
a2d2e5c0 | 297 | |
a2d2e5c0 | 298 | |
bde47ca7 | 299 | |
411acbbc | 300 | Handling IPv6 on old machines |
bde47ca7 | 301 | |
411acbbc MP |
302 | The KAME IPv6 patch is nice in theory but has proved a bit of a |
303 | nightmare in practice. The basic idea of their patch is that rsync | |
304 | is rewritten to use the new getaddrinfo()/getnameinfo() interface, | |
305 | rather than gethostbyname()/gethostbyaddr() as in rsync 2.4.6. | |
306 | Systems that don't have the new interface are handled by providing | |
307 | our own implementation in lib/, which is selectively linked in. | |
c7d692c3 | 308 | |
411acbbc MP |
309 | The problem with this is that it is really hard to get right on |
310 | platforms that have a half-working implementation, so redefining | |
311 | these functions clashes with system headers, and leaving them out | |
312 | breaks. This affects at least OSF/1, RedHat 5, and Cobalt, which | |
313 | are moderately improtant. | |
314 | ||
315 | Perhaps the simplest solution would be to have two different files | |
316 | implementing the same interface, and choose either the new or the | |
317 | old API. This is probably necessary for systems that e.g. have | |
318 | IPv6, but gethostbyaddr() can't handle it. The Linux manpage claims | |
319 | this is currently the case. | |
320 | ||
321 | In fact, our internal sockets interface (things like | |
322 | open_socket_out(), etc) is much narrower than the getaddrinfo() | |
323 | interface, and so probably simpler to get right. In addition, the | |
324 | old code is known to work well on old machines. | |
325 | ||
326 | We could drop the rather large lib/getaddrinfo files. | |
327 | ||
328 | ||
329 | Other IPv6 stuff: | |
330 | ||
c33e3e39 MP |
331 | Implement suggestions from http://www.kame.net/newsletter/19980604/ |
332 | and ftp://ftp.iij.ad.jp/pub/RFC/rfc2553.txt | |
333 | ||
334 | If a host has multiple addresses, then listen try to connect to all | |
335 | in order until we get through. (getaddrinfo may return multiple | |
c10b0bdd | 336 | addresses.) This is kind of implemented already. |
c33e3e39 MP |
337 | |
338 | Possibly also when starting as a server we may need to listen on | |
339 | multiple passive addresses. This might be a bit harder, because we | |
340 | may need to select on all of them. Hm. | |
341 | ||
a2d2e5c0 MP |
342 | Define a syntax for IPv6 literal addresses. Since they include |
343 | colons, they tend to break most naming systems, including ours. | |
344 | Based on the HTTP IPv6 syntax, I think we should use | |
345 | ||
a577af90 | 346 | rsync://[::1]/foo/bar [::1]::bar |
a2d2e5c0 MP |
347 | |
348 | which should just take a small change to the parser code. | |
349 | ||
b17dd0c4 | 350 | |
5aafd07b MP |
351 | Errors |
352 | ||
353 | If we hang or get SIGINT, then explain where we were up to. Perhaps | |
354 | have a static buffer that contains the current function name, or | |
355 | some kind of description of what we were trying to do. This is a | |
356 | little easier on people than needing to run strace/truss. | |
357 | ||
358 | "The dungeon collapses! You are killed." Rather than "unexpected | |
359 | eof" give a message that is more detailed if possible and also more | |
a577af90 | 360 | helpful. |
5aafd07b | 361 | |
89b0a3d9 MP |
362 | If we get an error writing to a socket, then we should perhaps |
363 | continue trying to read to see if an error message comes across | |
364 | explaining why the socket is closed. I'm not sure if this would | |
365 | work, but it would certainly make our messages more helpful. | |
366 | ||
b17dd0c4 | 367 | What happens if a directory is missing -x attributes. Do we lose |
a577af90 PG |
368 | our load? (Debian #28416) Probably fixed now, but a test case would |
369 | be good. | |
b17dd0c4 MP |
370 | |
371 | ||
5575de14 MP |
372 | File attributes |
373 | ||
5575de14 MP |
374 | Transfer ACLs. Need to think of a standard representation. |
375 | Probably better not to even try to convert between NT and POSIX. | |
376 | Possibly can share some code with Samba. | |
5aafd07b | 377 | |
28a69e25 MP |
378 | Empty directories |
379 | ||
380 | With the current common --include '*/' --exclude '*' pattern, people | |
381 | can end up with many empty directories. We might avoid this by | |
382 | lazily creating such directories. | |
383 | ||
c6e27b60 | 384 | |
28a69e25 MP |
385 | zlib |
386 | ||
a577af90 | 387 | Perhaps don't use our own zlib. |
c6e27b60 MP |
388 | |
389 | Advantages: | |
390 | ||
391 | - will automatically be up to date with bugfixes in zlib | |
392 | ||
393 | - can leave it out for small rsync on e.g. recovery disks | |
394 | ||
395 | - can use a shared library | |
396 | ||
397 | - avoids people breaking rsync by trying to do this themselves and | |
398 | messing up | |
399 | ||
400 | Should we ship zlib for systems that don't have it, or require | |
401 | people to install it separately? | |
402 | ||
403 | Apparently this will make us incompatible with versions of rsync | |
404 | that use the patched version of rsync. Probably the simplest way to | |
405 | do this is to just disable gzip (with a warning) when talking to old | |
406 | versions. | |
407 | ||
5ba268ef MP |
408 | |
409 | logging | |
410 | ||
411 | Perhaps flush stdout after each filename, so that people trying to | |
412 | monitor progress in a log file can do so more easily. See | |
413 | http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=48108 | |
414 | ||
415 | At the connections that just get a list of modules are not logged, | |
416 | but they should be. | |
417 | ||
418 | If a child of the rsync daemon dies with a signal, we should notice | |
419 | that when we reap it and log a message. | |
420 | ||
421 | Keep stderr and stdout properly separated (Debian #23626) | |
422 | ||
595f2d4d MP |
423 | After we get the @RSYNCD greeting from the server, we know it's |
424 | version but we have not yet sent the command line, so we could just | |
a577af90 | 425 | remove the -z option if the server is too old. |
595f2d4d MP |
426 | |
427 | For ssh invocation it's not so simple, because we actually use the | |
428 | command line to start the remote process. However, we only actually | |
429 | do compression in token.c, and we could therefore once we discover | |
430 | the remote version emit an error if it's too old. I'm not sure if | |
431 | that's a good tradeoff or not. | |
432 | ||
28a69e25 | 433 | |
92325ada MP |
434 | proxy authentication |
435 | ||
436 | Allow RSYNC_PROXY to be http://user:pass@proxy.foo:3128/, and do | |
a577af90 | 437 | HTTP Basic Proxy-Authentication. |
92325ada MP |
438 | |
439 | Multiple schemes are possible, up to and including the insanity that | |
440 | is NTLM, but Basic probably covers most cases. | |
441 | ||
442 | SOCKS | |
443 | ||
444 | Add --with-socks, and then perhaps a command-line option to put them | |
445 | on or off. This might be more reliable than LD_PRELOAD hacks. | |
446 | ||
5ba268ef MP |
447 | FAT support |
448 | ||
a577af90 PG |
449 | rsync to a FAT partition on a Unix machine doesn't work very well at |
450 | the moment. I think we get errors about invalid filenames and | |
5ba268ef MP |
451 | perhaps also trying to do atomic renames. |
452 | ||
a577af90 PG |
453 | I guess the code to do this is currently #ifdef'd on Windows; |
454 | perhaps we ought to intelligently fall back to it on Unix too. | |
5ba268ef MP |
455 | |
456 | ||
27741d9f MP |
457 | Better statistics: |
458 | ||
459 | <Rasmus> mbp: hey, how about an rsync option that just gives you the | |
460 | summary without the list of files? And perhaps gives more | |
461 | information like the number of new files, number of changed, | |
a577af90 PG |
462 | deleted, etc. ? <mbp> Rasmus: nice idea <mbp> there is --stats |
463 | <mbp> but at the moment it's very tridge-oriented <mbp> rather than | |
464 | user-friendly <mbp> it would be nice to improve it <mbp> that would | |
465 | also work well with --dryrun | |
27741d9f | 466 | |
e53fe9a2 MP |
467 | TDB: |
468 | ||
469 | Rather than storing the file list in memory, store it in a TDB. | |
470 | ||
471 | This *might* make memory usage lower while building the file list. | |
472 | ||
473 | Hashtable lookup will mean files are not transmitted in order, | |
474 | though... hm. | |
475 | ||
476 | This would neatly eliminate one of the major post-fork shared data | |
477 | structures. | |
478 | ||
479 | ||
97e1254a MP |
480 | chmod: |
481 | ||
a577af90 PG |
482 | On 12 Mar 2002, Dave Dykstra <dwd@bell-labs.com> wrote: > If we |
483 | would add an option to do that functionality, I would vote for one > | |
484 | that was more general which could mask off any set of permission bits | |
485 | and > possibly add any set of bits. Perhaps a chmod-like syntax if it | |
486 | could be > implemented simply. | |
97e1254a | 487 | |
a577af90 | 488 | I think that would be good too. For example, people uploading files |
97e1254a MP |
489 | to a web server might like to say |
490 | ||
491 | rsync -avzP --chmod a+rX ./ sourcefrog.net:/home/www/sourcefrog/ | |
492 | ||
493 | Ideally the patch would implement as many of the gnu chmod semantics | |
494 | as possible. I think the mode parser should be a separate function | |
a577af90 PG |
495 | that passes back something like (mask,set) description to the rest |
496 | of the program. For bonus points there would be a test case for the | |
97e1254a MP |
497 | parser. |
498 | ||
8bd1a73e MP |
499 | Possibly also --chown |
500 | ||
36692011 MP |
501 | (Debian #23628) |
502 | ||
97e1254a | 503 | |
3c1edccb MP |
504 | --diff |
505 | ||
506 | Allow people to specify the diff command. (Might want to use wdiff, | |
507 | gnudiff, etc.) | |
508 | ||
509 | Just diff the temporary file with the destination file, and delete | |
510 | the tmp file rather than moving it into place. | |
511 | ||
512 | Interaction with --partial. | |
513 | ||
514 | Security interactions with daemon mode? | |
515 | ||
516 | (Suggestion from david.e.sewell) | |
517 | ||
518 | ||
a628b069 MP |
519 | Incorrect timestamps (Debian #100295) |
520 | ||
521 | A bit hard to believe, but apparently it happens. | |
6d19c674 MP |
522 | |
523 | ||
524 | Check "refuse options works" | |
525 | ||
526 | We need a test case for this... | |
527 | ||
528 | Was this broken when we changed to popt? | |
6479c2ed MP |
529 | |
530 | ||
bd0ad74f MP |
531 | PERFORMANCE ---------------------------------------------------------- |
532 | ||
533 | MD4 file_sum | |
534 | ||
535 | If we're doing a local transfer, or using -W, then perhaps don't | |
536 | send the file checksum. If we're doing a local transfer, then | |
537 | calculating MD4 checksums uses 90% of CPU and is unlikely to be | |
538 | useful. | |
539 | ||
540 | Indeed for transfers over zlib or ssh we can also rely on the | |
541 | transport to have quite strong protection against corruption. | |
542 | ||
543 | Perhaps we should have an option to disable this, analogous to | |
a577af90 PG |
544 | --whole-file, although it would default to disabled. The file |
545 | checksum takes up a definite space in the protocol -- we can either | |
546 | set it to 0, or perhaps just leave it out. | |
bd0ad74f MP |
547 | |
548 | MD4 | |
549 | ||
550 | Perhaps borrow an assembler MD4 from someone? | |
551 | ||
552 | Make sure we call MD4 with properly-sized blocks whenever possible | |
553 | to avoid copying into the residue region? | |
554 | ||
6479c2ed MP |
555 | String area code |
556 | ||
557 | Test whether this is actually faster than just using malloc(). If | |
558 | it's not (anymore), throw it out. | |
a628b069 MP |
559 | |
560 | ||
a2d2e5c0 MP |
561 | PLATFORMS ------------------------------------------------------------ |
562 | ||
563 | Win32 | |
564 | ||
565 | Don't detach, because this messes up --srvany. | |
566 | ||
567 | http://sources.redhat.com/ml/cygwin/2001-08/msg00234.html | |
568 | ||
5ba268ef | 569 | |
0e23e41d MP |
570 | DEVELOPMENT ---------------------------------------------------------- |
571 | ||
572 | Splint | |
573 | ||
574 | Build rsync with SPLINT to try to find security holes. Add | |
575 | annotations as necessary. Keep track of the number of warnings | |
576 | found initially, and see how many of them are real bugs, or real | |
577 | security bugs. Knowing the percentage of likely hits would be | |
578 | really interesting for other projects. | |
579 | ||
f5a95bb5 MP |
580 | Torture test |
581 | ||
582 | Something that just keeps running rsync continuously over a data set | |
583 | likely to generate problems. | |
584 | ||
585 | Cross-testing | |
586 | ||
587 | Run current rsync versions against significant past releases. | |
588 | ||
43a4dc10 MP |
589 | Memory debugger |
590 | ||
3a79260d | 591 | jra recommends Valgrind: |
43a4dc10 MP |
592 | |
593 | http://devel-home.kde.org/~sewardj/ | |
594 | ||
25ff30e8 MP |
595 | Release script |
596 | ||
597 | Update spec files | |
598 | ||
599 | Build tar file; upload | |
600 | ||
601 | Send announcement to mailing list and c.o.l.a. | |
602 | ||
603 | Make freshmeat announcement | |
604 | ||
605 | Update web site | |
606 | ||
607 | ||
608 | ||
e9c4c301 MP |
609 | TESTING -------------------------------------------------------------- |
610 | ||
611 | Cross-test versions | |
612 | ||
613 | Part of the regression suite should be making sure that we don't | |
a577af90 PG |
614 | break backwards compatibility: old clients vs new servers and so on. |
615 | Ideally we would test both up and down from the current release to | |
616 | all old versions. | |
25ff30e8 MP |
617 | |
618 | We might need to omit broken old versions, or versions in which | |
619 | particular functionality is broken | |
e9c4c301 MP |
620 | |
621 | It might be sufficient to test downloads from well-known public | |
622 | rsync servers running different versions of rsync. This will give | |
623 | some testing and also be the most common case for having different | |
624 | versions and not being able to upgrade. | |
625 | ||
25ff30e8 MP |
626 | |
627 | Test on kernel source | |
628 | ||
629 | Download all versions of kernel; unpack, sync between them. Also | |
630 | sync between uncompressed tarballs. Compare directories after | |
631 | transfer. | |
632 | ||
633 | Use local mode; ssh; daemon; --whole-file and --no-whole-file. | |
634 | ||
635 | Use awk to pull out the 'speedup' number for each transfer. Make | |
636 | sure it is >= x. | |
637 | ||
638 | ||
e9c4c301 MP |
639 | Test large files |
640 | ||
641 | Sparse and non-sparse | |
642 | ||
643 | Mutator program | |
644 | ||
645 | Insert bytes, delete bytes, swap blocks, ... | |
646 | ||
647 | configure option to enable dangerous tests | |
648 | ||
649 | If tests are skipped, say why. | |
650 | ||
b73b51a9 MP |
651 | Test daemon feature to disallow particular options. |
652 | ||
599dc93c MP |
653 | Pipe program that makes slow/jerky connections. |
654 | ||
a577af90 PG |
655 | Versions of read() and write() that corrupt the stream, or abruptly |
656 | fail | |
76533c52 | 657 | |
6479c2ed MP |
658 | Separate makefile target to run rough tests -- or perhaps just run |
659 | them every time? | |
660 | ||
717eb9b8 MP |
661 | Test "refuse options" works |
662 | ||
663 | What about for --recursive? | |
664 | ||
665 | If you specify an unrecognized option here, you should get an error. | |
666 | ||
e9c4c301 | 667 | |
7c583c73 MP |
668 | DOCUMENTATION -------------------------------------------------------- |
669 | ||
670 | Update README | |
671 | ||
b73b51a9 MP |
672 | Keep list of open issues and todos on the web site |
673 | ||
674 | Update web site from CVS | |
675 | ||
5af50297 MP |
676 | |
677 | Perhaps redo manual as SGML | |
678 | ||
679 | The man page is getting rather large, and there is more information | |
680 | that ought to be added. | |
681 | ||
682 | TexInfo source is probably a dying format. | |
683 | ||
684 | Linuxdoc looks like the most likely contender. I know DocBook is | |
685 | favoured by some people, but it's so bloody verbose, even with emacs | |
686 | support. | |
687 | ||
688 | ||
a2d2e5c0 MP |
689 | BUILD FARM ----------------------------------------------------------- |
690 | ||
691 | Add machines | |
692 | ||
a2d2e5c0 MP |
693 | Cygwin (on different versions of Win32?) |
694 | ||
695 | HP-UX variants (via HP?) | |
33d213bb | 696 | |
5aafd07b MP |
697 | SCO |
698 | ||
46ef7d1d | 699 | |
62b68c80 | 700 | LOGGING -------------------------------------------------------------- |
a2d2e5c0 | 701 | |
62b68c80 MP |
702 | Perhaps flush stdout after each filename, so that people trying to |
703 | monitor progress in a log file can do so more easily. See | |
704 | http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=48108 | |
a2d2e5c0 | 705 | |
62b68c80 MP |
706 | At the connections that just get a list of modules are not logged, |
707 | but they should be. | |
708 | ||
709 | If a child of the rsync daemon dies with a signal, we should notice | |
710 | that when we reap it and log a message. | |
711 | ||
712 | Keep stderr and stdout properly separated (Debian #23626) | |
713 | ||
714 | Use a separate function for reporting errors; prefix it with | |
715 | "rsync:" or "rsync(remote)", or perhaps even "rsync(local | |
716 | generator): ". | |
a2d2e5c0 | 717 | |
50f2f002 MP |
718 | verbose output |
719 | ||
720 | Indicate whether files are new, updated, or deleted | |
721 | ||
d834adc1 MP |
722 | At end of transfer, show how many files were or were not transferred |
723 | correctly. | |
724 | ||
62b68c80 MP |
725 | -vv |
726 | ||
727 | Explain *why* every file is transferred or not (e.g. "local mtime | |
728 | 123123 newer than 1283198") | |
729 | ||
730 | ||
8ff9d697 MP |
731 | debugging of daemon |
732 | ||
733 | Add an rsyncd.conf parameter to turn on debugging on the server. | |
734 | ||
735 | ||
62b68c80 MP |
736 | |
737 | NICE ----------------------------------------------------------------- | |
738 | ||
739 | --no-detach and --no-fork options | |
740 | ||
741 | Very useful for debugging. Also good when running under a | |
742 | daemon-monitoring process that tries to restart the service when the | |
743 | parent exits. | |
744 | ||
745 | hang/timeout friendliness | |
3d90ec14 | 746 | |
a2d2e5c0 MP |
747 | internationalization |
748 | ||
749 | Change to using gettext(). Probably need to ship this for platforms | |
a577af90 | 750 | that don't have it. |
a2d2e5c0 MP |
751 | |
752 | Solicit translations. | |
753 | ||
3d90ec14 MP |
754 | Does anyone care? Before we bother modifying the code, we ought to |
755 | get the manual translated first, because that's possibly more useful | |
756 | and at any rate demonstrates desire. | |
757 | ||
a577af90 | 758 | rsyncsh |
46ef7d1d MP |
759 | |
760 | Write a small emulation of interactive ftp as a Pythonn program | |
761 | that calls rsync. Commands such as "cd", "ls", "ls *.c" etc map | |
762 | fairly directly into rsync commands: it just needs to remember the | |
763 | current host, directory and so on. We can probably even do | |
764 | completion of remote filenames. | |
25ff30e8 MP |
765 | |
766 | ||
767 | RELATED PROJECTS ----------------------------------------------------- | |
768 | ||
769 | http://rsync.samba.org/rsync-and-debian/ | |
770 | ||
771 | rsyncable gzip patch | |
772 | ||
773 | Exhaustive, tortuous testing | |
774 | ||
775 | Cleanups? | |
776 | ||
777 | rsyncsplit as alternative to real integration with gzip? | |
778 | ||
779 | reverse rsync over HTTP Range | |
780 | ||
781 | Goswin Brederlow suggested this on Debian; I think tridge and I | |
782 | talked about it previous in relation to rproxy. | |
a577af90 PG |
783 | |
784 |