3 import sys, stat, time, os, errno, re
4 from bup import metadata, options, git, index, drecurse, hlinkdb
5 from bup.helpers import *
6 from bup.hashsplit import GIT_MODE_TREE, GIT_MODE_FILE
16 self.cur = self.i.next()
22 def check_index(reader):
24 log('check: checking forward iteration...\n')
27 for e in reader.forward_iter():
30 log('%08x+%-4d %r\n' % (e.children_ofs, e.children_n,
32 assert(e.children_ofs)
33 assert(e.name.endswith('/'))
34 assert(not d.get(e.children_ofs))
36 if e.flags & index.IX_HASHVALID:
37 assert(e.sha != index.EMPTY_SHA)
39 assert(not e or e.name == '/') # last entry is *always* /
40 log('check: checking normal iteration...\n')
47 log('index error! at %r\n' % e)
49 log('check: passed.\n')
52 def clear_index(indexfile):
53 indexfiles = [indexfile, indexfile + '.meta', indexfile + '.hlink']
54 for indexfile in indexfiles:
55 path = git.repo(indexfile)
59 log('clear: removed %s\n' % path)
61 if e.errno != errno.ENOENT:
65 def update_index(top, excluded_paths, exclude_rxs):
66 # tmax and start must be epoch nanoseconds.
67 tmax = (time.time() - 1) * 10**9
68 ri = index.Reader(indexfile)
69 msw = index.MetaStoreWriter(indexfile + '.meta')
70 wi = index.Writer(indexfile, msw, tmax)
71 rig = IterHelper(ri.iter(name=top))
72 tstart = int(time.time()) * 10**9
74 hlinks = hlinkdb.HLinkDB(indexfile + '.hlink')
79 return (GIT_MODE_FILE, index.FAKE_SHA)
82 bup_dir = os.path.abspath(git.repo())
83 for (path,pst) in drecurse.recursive_dirlist([top], xdev=opt.xdev,
85 excluded_paths=excluded_paths,
86 exclude_rxs=exclude_rxs):
87 if opt.verbose>=2 or (opt.verbose==1 and stat.S_ISDIR(pst.st_mode)):
88 sys.stdout.write('%s\n' % path)
90 qprogress('Indexing: %d\r' % total)
91 elif not (total % 128):
92 qprogress('Indexing: %d\r' % total)
94 while rig.cur and rig.cur.name > path: # deleted paths
98 if rig.cur.nlink > 1 and not stat.S_ISDIR(rig.cur.mode):
99 hlinks.del_path(rig.cur.name)
101 if rig.cur and rig.cur.name == path: # paths that already existed
103 meta = metadata.from_path(path, statinfo=pst)
104 except (OSError, IOError), e:
108 if not stat.S_ISDIR(rig.cur.mode) and rig.cur.nlink > 1:
109 hlinks.del_path(rig.cur.name)
110 if not stat.S_ISDIR(pst.st_mode) and pst.st_nlink > 1:
111 hlinks.add_path(path, pst.st_dev, pst.st_ino)
112 # Clear these so they don't bloat the store -- they're
113 # already in the index (since they vary a lot and they're
114 # fixed length). If you've noticed "tmax", you might
115 # wonder why it's OK to do this, since that code may
116 # adjust (mangle) the index mtime and ctime -- producing
117 # fake values which must not end up in a .bupm. However,
118 # it looks like that shouldn't be possible: (1) When
119 # "save" validates the index entry, it always reads the
120 # metadata from the filesytem. (2) Metadata is only
121 # read/used from the index if hashvalid is true. (3) index
122 # always invalidates "faked" entries, because "old != new"
124 meta.ctime = meta.mtime = meta.atime = 0
125 meta_ofs = msw.store(meta)
126 rig.cur.from_stat(pst, meta_ofs, tstart,
127 check_device=opt.check_device)
128 if not (rig.cur.flags & index.IX_HASHVALID):
130 (rig.cur.gitmode, rig.cur.sha) = hashgen(path)
131 rig.cur.flags |= index.IX_HASHVALID
138 meta = metadata.from_path(path, statinfo=pst)
139 except (OSError, IOError), e:
142 # See same assignment to 0, above, for rationale.
143 meta.atime = meta.mtime = meta.ctime = 0
144 meta_ofs = msw.store(meta)
145 wi.add(path, pst, meta_ofs, hashgen = hashgen)
146 if not stat.S_ISDIR(pst.st_mode) and pst.st_nlink > 1:
147 hlinks.add_path(path, pst.st_dev, pst.st_ino)
149 progress('Indexing: %d, done.\n' % total)
151 hlinks.prepare_save()
159 log('check: before merging: oldfile\n')
161 log('check: before merging: newfile\n')
163 mi = index.Writer(indexfile, msw, tmax)
165 for e in index.merge(ri, wr):
166 # FIXME: shouldn't we remove deleted entries eventually? When?
181 bup index <-p|m|s|u> [options...] <filenames...>
184 p,print print the index entries for the given names (also works with -u)
185 m,modified print only added/deleted/modified files (implies -p)
186 s,status print each filename with a status char (A/M/D) (implies -p)
187 u,update recursively update the index entries for the given file/dir names (default if no mode is specified)
188 check carefully check index file integrity
189 clear clear the default index
191 H,hash print the hash for each object next to its name
192 l,long print more information about each file
193 no-check-device don't invalidate an entry if the containing device changes
194 fake-valid mark all index entries as up-to-date even if they aren't
195 fake-invalid mark all index entries as invalid
196 f,indexfile= the name of the index file (normally BUP_DIR/bupindex)
197 exclude= a path to exclude from the backup (can be used more than once)
198 exclude-from= a file that contains exclude paths (can be used more than once)
199 exclude-rx= skip paths that match the unanchored regular expression
200 v,verbose increase log output (can be used more than once)
201 x,xdev,one-file-system don't cross filesystem boundaries
203 o = options.Options(optspec)
204 (opt, flags, extra) = o.parse(sys.argv[1:])
206 if not (opt.modified or \
213 if (opt.fake_valid or opt.fake_invalid) and not opt.update:
214 o.fatal('--fake-{in,}valid are meaningless without -u')
215 if opt.fake_valid and opt.fake_invalid:
216 o.fatal('--fake-valid is incompatible with --fake-invalid')
217 if opt.clear and opt.indexfile:
218 o.fatal('cannot clear an external index (via -f)')
220 # FIXME: remove this once we account for timestamp races, i.e. index;
221 # touch new-file; index. It's possible for this to happen quickly
222 # enough that new-file ends up with the same timestamp as the first
223 # index, and then bup will ignore it.
224 tick_start = time.time()
225 time.sleep(1 - (tick_start - int(tick_start)))
227 git.check_repo_or_die()
228 indexfile = opt.indexfile or git.repo('bupindex')
233 log('check: starting initial check.\n')
234 check_index(index.Reader(indexfile))
237 log('clear: clearing index.\n')
238 clear_index(indexfile)
240 excluded_paths = parse_excludes(flags, o.fatal)
241 exclude_rxs = parse_rx_excludes(flags, o.fatal)
242 paths = index.reduce_paths(extra)
246 o.fatal('update mode (-u) requested but no paths given')
247 for (rp,path) in paths:
248 update_index(rp, excluded_paths, exclude_rxs)
250 if opt['print'] or opt.status or opt.modified:
251 for (name, ent) in index.Reader(indexfile).filter(extra or ['']):
253 and (ent.is_valid() or ent.is_deleted() or not ent.mode)):
259 elif not ent.is_valid():
260 if ent.sha == index.EMPTY_SHA:
267 line += ent.sha.encode('hex') + ' '
269 line += "%7s %7s " % (oct(ent.mode), oct(ent.gitmode))
270 print line + (name or './')
272 if opt.check and (opt['print'] or opt.status or opt.modified or opt.update):
273 log('check: starting final check.\n')
274 check_index(index.Reader(indexfile))
277 log('WARNING: %d errors encountered.\n' % len(saved_errors))