2 from __future__ import absolute_import, print_function
3 from binascii import hexlify
4 from errno import EACCES
6 import math, os, stat, sys, time
8 from bup import compat, hashsplit, git, options, index, client, metadata
9 from bup import hlinkdb
10 from bup.compat import argv_bytes, environ
11 from bup.hashsplit import GIT_MODE_TREE, GIT_MODE_FILE, GIT_MODE_SYMLINK
12 from bup.helpers import (add_error, grafted_path_components, handle_ctrl_c,
13 hostname, istty2, log, parse_date_or_fatal, parse_num,
14 path_components, progress, qprogress, resolve_parent,
15 saved_errors, stripped_path_components,
17 from bup.io import byte_stream, path_msg
18 from bup.pwdgrp import userfullname, username
19 from bup.tree import StackDir
23 bup save [-tc] [-n name] <filenames...>
25 r,remote= hostname:/path/to/repo of remote repository
26 t,tree output a tree id
27 c,commit output a commit id
28 n,name= name of backup set to update (if any)
29 d,date= date for the commit (seconds since the epoch)
30 v,verbose increase log output (can be used more than once)
31 q,quiet don't show progress meter
32 smaller= only back up files smaller than n bytes
33 bwlimit= maximum bytes/sec to transmit to server
34 f,indexfile= the name of the index file (normally BUP_DIR/bupindex)
35 strip strips the path to every filename given
36 strip-path= path-prefix to be stripped when saving
37 graft= a graft point *old_path*=*new_path* (can be used more than once)
38 #,compress= set compression level to # (0-9, 9 is highest) [1]
44 after_nondir_metadata_stat = None
46 def before_saving_regular_file(name):
52 # Hack around lack of nonlocal vars in python 2
55 o = options.Options(optspec)
56 opt, flags, extra = o.parse_bytes(argv[1:])
59 opt.indexfile = argv_bytes(opt.indexfile)
61 opt.name = argv_bytes(opt.name)
63 opt.remote = argv_bytes(opt.remote)
65 opt.strip_path = argv_bytes(opt.strip_path)
67 git.check_repo_or_die()
68 if not (opt.tree or opt.commit or opt.name):
69 o.fatal("use one or more of -t, -c, -n")
71 o.fatal("no filenames given")
73 extra = [argv_bytes(x) for x in extra]
75 opt.progress = (istty2 and not opt.quiet)
76 opt.smaller = parse_num(opt.smaller or 0)
78 client.bwlimit = parse_num(opt.bwlimit)
81 date = parse_date_or_fatal(opt.date, o.fatal)
85 if opt.strip and opt.strip_path:
86 o.fatal("--strip is incompatible with --strip-path")
91 o.fatal("--strip is incompatible with --graft")
94 o.fatal("--strip-path is incompatible with --graft")
96 for (option, parameter) in flags:
97 if option == "--graft":
98 parameter = argv_bytes(parameter)
99 splitted_parameter = parameter.split(b'=')
100 if len(splitted_parameter) != 2:
101 o.fatal("a graft point must be of the form old_path=new_path")
102 old_path, new_path = splitted_parameter
103 if not (old_path and new_path):
104 o.fatal("a graft point cannot be empty")
105 graft_points.append((resolve_parent(old_path),
106 resolve_parent(new_path)))
108 is_reverse = environ.get(b'BUP_SERVER_REVERSE')
109 if is_reverse and opt.remote:
110 o.fatal("don't use -r in reverse mode; it's automatic")
113 if name and not valid_save_name(name):
114 o.fatal("'%s' is not a valid branch name" % path_msg(name))
115 refname = name and b'refs/heads/%s' % name or None
116 if opt.remote or is_reverse:
118 cli = client.Client(opt.remote)
119 except client.ClientError as e:
122 oldref = refname and cli.read_ref(refname) or None
123 w = cli.new_packwriter(compression_level=opt.compress)
126 oldref = refname and git.read_ref(refname) or None
127 w = git.PackWriter(compression_level=opt.compress)
132 # Metadata is stored in a file named .bupm in each directory. The
133 # first metadata entry will be the metadata for the current directory.
134 # The remaining entries will be for each of the other directory
135 # elements, in the order they're listed in the index.
137 # Since the git tree elements are sorted according to
138 # git.shalist_item_sort_key, the metalist items are accumulated as
139 # (sort_key, metadata) tuples, and then sorted when the .bupm file is
140 # created. The sort_key should have been computed using the element's
141 # mangled name and git mode (after hashsplitting), but the code isn't
142 # actually doing that but rather uses the element's real name and mode.
143 # This makes things a bit more difficult when reading it back, see
144 # vfs.ordered_tree_entries().
146 # Maintain a stack of information representing the current location in
147 # the archive being constructed. The current path is recorded in
148 # parts, which will be something like
149 # [StackDir(name=''), StackDir(name='home'), StackDir(name='someuser')],
150 # and the accumulated content and metadata for files in the dirs is stored
151 # in the .items member of the StackDir.
156 def _push(part, metadata):
157 # Enter a new archive directory -- make it the current directory.
158 item = StackDir(part, metadata)
162 def _pop(force_tree=None, dir_metadata=None):
163 # Leave the current archive directory and add its tree to its parent.
165 # FIXME: only test if collision is possible (i.e. given --strip, etc.)?
173 if name in names_seen:
174 parent_path = b'/'.join(x.name for x in stack) + b'/'
175 add_error('error: ignoring duplicate path %s in %s'
176 % (path_msg(name), path_msg(parent_path)))
181 # if set, overrides the original metadata pushed for this dir.
182 if dir_metadata is None:
183 dir_metadata = item.meta
184 metalist = [(b'', dir_metadata)]
185 metalist += [(git.shalist_item_sort_key((entry.mode, entry.name, None)),
187 for entry in clean_list if entry.mode != GIT_MODE_TREE]
188 metalist.sort(key = lambda x: x[0])
189 metadata = BytesIO(b''.join(m[1].encode() for m in metalist))
190 mode, id = hashsplit.split_to_blob_or_tree(w.new_blob, w.new_tree,
192 keep_boundaries=False)
193 shalist = [(mode, b'.bupm', id)]
194 shalist += [(entry.gitmode,
195 git.mangle_name(entry.name, entry.mode, entry.gitmode),
197 for entry in clean_list]
199 tree = w.new_tree(shalist)
201 stack[-1].append(item.name, GIT_MODE_TREE, GIT_MODE_TREE, tree, None)
205 _nonlocal['count'] = 0
206 _nonlocal['subcount'] = 0
207 _nonlocal['lastremain'] = None
209 def progress_report(n):
210 _nonlocal['subcount'] += n
211 cc = _nonlocal['count'] + _nonlocal['subcount']
212 pct = total and (cc*100.0/total) or 0
214 elapsed = now - tstart
215 kps = elapsed and int(cc/1024./elapsed)
216 kps_frac = 10 ** int(math.log(kps+1, 10) - 1)
217 kps = int(kps/kps_frac)*kps_frac
219 remain = elapsed*1.0/cc * (total-cc)
222 if (_nonlocal['lastremain'] and (remain > _nonlocal['lastremain'])
223 and ((remain - _nonlocal['lastremain'])/_nonlocal['lastremain'] < 0.05)):
224 remain = _nonlocal['lastremain']
226 _nonlocal['lastremain'] = remain
227 hours = int(remain/60/60)
228 mins = int(remain/60 - hours*60)
229 secs = int(remain - hours*60*60 - mins*60)
234 kpsstr = '%dk/s' % kps
236 remainstr = '%dh%dm' % (hours, mins)
238 remainstr = '%dm%d' % (mins, secs)
240 remainstr = '%ds' % secs
241 qprogress('Saving: %.2f%% (%d/%dk, %d/%d files) %s %s\r'
242 % (pct, cc/1024, total/1024, fcount, ftotal,
246 indexfile = opt.indexfile or git.repo(b'bupindex')
247 r = index.Reader(indexfile)
249 msr = index.MetaStoreReader(indexfile + b'.meta')
250 except IOError as ex:
251 if ex.errno != EACCES:
253 log('error: cannot access %r; have you run bup index?'
254 % path_msg(indexfile))
256 hlink_db = hlinkdb.HLinkDB(indexfile + b'.hlink')
258 def already_saved(ent):
259 return ent.is_valid() and w.exists(ent.sha) and ent.sha
261 def wantrecurse_pre(ent):
262 return not already_saved(ent)
264 def wantrecurse_during(ent):
265 return not already_saved(ent) or ent.sha_missing()
267 def find_hardlink_target(hlink_db, ent):
268 if hlink_db and not stat.S_ISDIR(ent.mode) and ent.nlink > 1:
269 link_paths = hlink_db.node_paths(ent.dev, ent.ino)
275 for (transname,ent) in r.filter(extra, wantrecurse=wantrecurse_pre):
276 if not (ftotal % 10024):
277 qprogress('Reading index: %d\r' % ftotal)
278 exists = ent.exists()
279 hashvalid = already_saved(ent)
280 ent.set_sha_missing(not hashvalid)
281 if not opt.smaller or ent.size < opt.smaller:
282 if exists and not hashvalid:
285 progress('Reading index: %d, done.\n' % ftotal)
286 hashsplit.progress_callback = progress_report
288 # Root collisions occur when strip or graft options map more than one
289 # path to the same directory (paths which originally had separate
290 # parents). When that situation is detected, use empty metadata for
291 # the parent. Otherwise, use the metadata for the common parent.
292 # Collision example: "bup save ... --strip /foo /foo/bar /bar".
294 # FIXME: Add collision tests, or handle collisions some other way.
296 # FIXME: Detect/handle strip/graft name collisions (other than root),
297 # i.e. if '/foo/bar' and '/bar' both map to '/'.
300 root_collision = None
305 for (transname,ent) in r.filter(extra, wantrecurse=wantrecurse_during):
306 (dir, file) = os.path.split(ent.name)
307 exists = (ent.flags & index.IX_EXISTS)
308 hashvalid = already_saved(ent)
309 wasmissing = ent.sha_missing()
315 if ent.sha == index.EMPTY_SHA:
322 log('%s %-70s\n' % (status, path_msg(ent.name)))
323 elif not stat.S_ISDIR(ent.mode) and lastdir != dir:
324 if not lastdir.startswith(dir):
325 log('%s %-70s\n' % (status, path_msg(os.path.join(dir, b''))))
334 if opt.smaller and ent.size >= opt.smaller:
335 if exists and not hashvalid:
337 log('skipping large file "%s"\n' % path_msg(ent.name))
338 lastskip_name = ent.name
341 assert(dir.startswith(b'/'))
343 dirp = stripped_path_components(dir, extra)
345 dirp = stripped_path_components(dir, [opt.strip_path])
347 dirp = grafted_path_components(graft_points, dir)
349 dirp = path_components(dir)
351 # At this point, dirp contains a representation of the archive
352 # path that looks like [(archive_dir_name, real_fs_path), ...].
353 # So given "bup save ... --strip /foo/bar /foo/bar/baz", dirp
354 # might look like this at some point:
355 # [('', '/foo/bar'), ('baz', '/foo/bar/baz'), ...].
357 # This dual representation supports stripping/grafting, where the
358 # archive path may not have a direct correspondence with the
359 # filesystem. The root directory is represented by an initial
360 # component named '', and any component that doesn't have a
361 # corresponding filesystem directory (due to grafting, for
362 # example) will have a real_fs_path of None, i.e. [('', None),
365 if first_root == None:
367 elif first_root != dirp[0]:
368 root_collision = True
370 # If switching to a new sub-tree, finish the current sub-tree.
371 while [x.name for x in stack] > [x[0] for x in dirp]:
374 # If switching to a new sub-tree, start a new sub-tree.
375 for path_component in dirp[len(stack):]:
376 dir_name, fs_path = path_component
377 # Not indexed, so just grab the FS metadata or use empty metadata.
379 meta = metadata.from_path(fs_path, normalized=True) \
380 if fs_path else metadata.Metadata()
381 except (OSError, IOError) as e:
383 lastskip_name = dir_name
384 meta = metadata.Metadata()
385 _push(dir_name, meta)
389 continue # We're at the top level -- keep the current root dir
390 # Since there's no filename, this is a subdir -- finish it.
391 oldtree = already_saved(ent) # may be None
392 newtree = _pop(force_tree = oldtree)
394 if lastskip_name and lastskip_name.startswith(ent.name):
397 ent.validate(GIT_MODE_TREE, newtree)
399 if exists and wasmissing:
400 _nonlocal['count'] += oldsize
403 # it's not a directory
405 meta = msr.metadata_at(ent.meta_ofs)
406 meta.hardlink_target = find_hardlink_target(hlink_db, ent)
407 # Restore the times that were cleared to 0 in the metastore.
408 (meta.atime, meta.mtime, meta.ctime) = (ent.atime, ent.mtime, ent.ctime)
409 stack[-1].append(file, ent.mode, ent.gitmode, ent.sha, meta)
412 hlink = find_hardlink_target(hlink_db, ent)
414 meta = metadata.from_path(ent.name, hardlink_target=hlink,
416 after_stat=after_nondir_metadata_stat)
417 except (OSError, IOError) as e:
419 lastskip_name = ent.name
421 if stat.S_IFMT(ent.mode) != stat.S_IFMT(meta.mode):
422 # The mode changed since we indexed the file, this is bad.
423 # This can cause two issues:
424 # 1) We e.g. think the file is a regular file, but now it's
425 # something else (a device, socket, FIFO or symlink, etc.)
426 # and _read_ from it when we shouldn't.
427 # 2) We then record it as valid, but don't update the index
428 # metadata, and on a subsequent save it has 'hashvalid'
429 # but is recorded as the file type from the index, when
430 # the content is something else ...
431 # Avoid all of these consistency issues by just skipping such
432 # things - it really ought to not happen anyway.
433 add_error("%s: mode changed since indexing, skipping." % path_msg(ent.name))
434 lastskip_name = ent.name
436 if stat.S_ISREG(ent.mode):
438 # If the file changes while we're reading it, then our reading
439 # may stop at some point, but the stat() above may have gotten
440 # a different size already. Recalculate the meta size so that
441 # the repository records the accurate size in the metadata, even
442 # if the other stat() data might be slightly older than the file
443 # content (which we can't fix, this is inherently racy, but we
444 # can prevent the size mismatch.)
447 meta.size += len(data)
448 return w.new_blob(data)
449 before_saving_regular_file(ent.name)
450 with hashsplit.open_noatime(ent.name) as f:
451 (mode, id) = hashsplit.split_to_blob_or_tree(
452 new_blob, w.new_tree, [f],
453 keep_boundaries=False)
454 except (IOError, OSError) as e:
455 add_error('%s: %s' % (ent.name, e))
456 lastskip_name = ent.name
457 elif stat.S_ISDIR(ent.mode):
458 assert(0) # handled above
459 elif stat.S_ISLNK(ent.mode):
460 mode, id = (GIT_MODE_SYMLINK, w.new_blob(meta.symlink_target))
462 # Everything else should be fully described by its
463 # metadata, so just record an empty blob, so the paths
464 # in the tree and .bupm will match up.
465 (mode, id) = (GIT_MODE_FILE, w.new_blob(b''))
468 ent.validate(mode, id)
470 stack[-1].append(file, ent.mode, ent.gitmode, id, meta)
472 if exists and wasmissing:
473 _nonlocal['count'] += oldsize
474 _nonlocal['subcount'] = 0
478 pct = total and _nonlocal['count']*100.0/total or 100
479 progress('Saving: %.2f%% (%d/%dk, %d/%d files), done. \n'
480 % (pct, _nonlocal['count']/1024, total/1024, fcount, ftotal))
482 while len(stack) > 1: # _pop() all the parts above the root
485 # Finish the root directory.
486 # When there's a collision, use empty metadata for the root.
487 tree = _pop(dir_metadata = metadata.Metadata() if root_collision else None)
490 out = byte_stream(sys.stdout)
493 out.write(hexlify(tree))
495 if opt.commit or name:
496 if compat.py_maj > 2:
497 # Strip b prefix from python 3 bytes reprs to preserve previous format
498 msgcmd = b'[%s]' % b', '.join([repr(argv_bytes(x))[1:].encode('ascii')
502 msg = b'bup save\n\nGenerated by command:\n%s\n' % msgcmd
503 userline = (b'%s <%s@%s>' % (userfullname(), username(), hostname()))
504 commit = w.new_commit(tree, oldref, userline, date, None,
505 userline, date, None, msg)
507 out.write(hexlify(commit))
511 w.close() # must close before we can update the ref
515 cli.update_ref(refname, commit, oldref)
517 git.update_ref(refname, commit, oldref)
523 log('WARNING: %d errors encountered while saving.\n' % len(saved_errors))