3 import sys, stat, time, os, errno
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']
55 for indexfile in indexfiles:
56 path = git.repo(indexfile)
60 log('clear: removed %s\n' % path)
63 if e.errno != errno.ENOENT:
67 def update_index(top, excluded_paths):
68 # tmax and start must be epoch nanoseconds.
69 tmax = (time.time() - 1) * 10**9
70 ri = index.Reader(indexfile)
71 msw = index.MetaStoreWriter(indexfile + '.meta')
72 wi = index.Writer(indexfile, msw, tmax)
73 rig = IterHelper(ri.iter(name=top))
74 tstart = int(time.time()) * 10**9
76 hlinks = hlinkdb.HLinkDB(indexfile + '.hlink')
81 return (GIT_MODE_FILE, index.FAKE_SHA)
84 bup_dir = os.path.abspath(git.repo())
85 for (path,pst) in drecurse.recursive_dirlist([top], xdev=opt.xdev,
87 excluded_paths=excluded_paths):
88 if opt.verbose>=2 or (opt.verbose==1 and stat.S_ISDIR(pst.st_mode)):
89 sys.stdout.write('%s\n' % path)
91 qprogress('Indexing: %d\r' % total)
92 elif not (total % 128):
93 qprogress('Indexing: %d\r' % total)
95 while rig.cur and rig.cur.name > path: # deleted paths
99 if rig.cur.nlink > 1 and not stat.S_ISDIR(rig.cur.mode):
100 hlinks.del_path(rig.cur.name)
102 if rig.cur and rig.cur.name == path: # paths that already existed
103 if not stat.S_ISDIR(rig.cur.mode) and rig.cur.nlink > 1:
104 hlinks.del_path(rig.cur.name)
105 if not stat.S_ISDIR(pst.st_mode) and pst.st_nlink > 1:
106 hlinks.add_path(path, pst.st_dev, pst.st_ino)
107 meta = metadata.from_path(path, statinfo=pst)
108 # Clear these so they don't bloat the store -- they're
109 # already in the index (since they vary a lot and they're
110 # fixed length). If you've noticed "tmax", you might
111 # wonder why it's OK to do this, since that code may
112 # adjust (mangle) the index mtime and ctime -- producing
113 # fake values which must not end up in a .bupm. However,
114 # it looks like that shouldn't be possible: (1) When
115 # "save" validates the index entry, it always reads the
116 # metadata from the filesytem. (2) Metadata is only
117 # read/used from the index if hashvalid is true. (3) index
118 # always invalidates "faked" entries, because "old != new"
120 meta.ctime = meta.mtime = meta.atime = 0
121 meta_ofs = msw.store(meta)
122 rig.cur.from_stat(pst, meta_ofs, tstart)
123 if not (rig.cur.flags & index.IX_HASHVALID):
125 (rig.cur.gitmode, rig.cur.sha) = hashgen(path)
126 rig.cur.flags |= index.IX_HASHVALID
132 meta = metadata.from_path(path, statinfo=pst)
133 # See same assignment to 0, above, for rationale.
134 meta.atime = meta.mtime = meta.ctime = 0
135 meta_ofs = msw.store(meta)
136 wi.add(path, pst, meta_ofs, hashgen = hashgen)
137 if not stat.S_ISDIR(pst.st_mode) and pst.st_nlink > 1:
138 hlinks.add_path(path, pst.st_dev, pst.st_ino)
140 progress('Indexing: %d, done.\n' % total)
142 hlinks.prepare_save()
150 log('check: before merging: oldfile\n')
152 log('check: before merging: newfile\n')
154 mi = index.Writer(indexfile, msw, tmax)
156 for e in index.merge(ri, wr):
157 # FIXME: shouldn't we remove deleted entries eventually? When?
172 bup index <-p|m|s|u> [options...] <filenames...>
175 p,print print the index entries for the given names (also works with -u)
176 m,modified print only added/deleted/modified files (implies -p)
177 s,status print each filename with a status char (A/M/D) (implies -p)
178 u,update recursively update the index entries for the given file/dir names (default if no mode is specified)
179 check carefully check index file integrity
180 clear clear the index
182 H,hash print the hash for each object next to its name
183 l,long print more information about each file
184 fake-valid mark all index entries as up-to-date even if they aren't
185 fake-invalid mark all index entries as invalid
186 f,indexfile= the name of the index file (normally BUP_DIR/bupindex)
187 exclude= a path to exclude from the backup (can be used more than once)
188 exclude-from= a file that contains exclude paths (can be used more than once)
189 v,verbose increase log output (can be used more than once)
190 x,xdev,one-file-system don't cross filesystem boundaries
192 o = options.Options(optspec)
193 (opt, flags, extra) = o.parse(sys.argv[1:])
195 if not (opt.modified or \
202 if (opt.fake_valid or opt.fake_invalid) and not opt.update:
203 o.fatal('--fake-{in,}valid are meaningless without -u')
204 if opt.fake_valid and opt.fake_invalid:
205 o.fatal('--fake-valid is incompatible with --fake-invalid')
207 # FIXME: remove this once we account for timestamp races, i.e. index;
208 # touch new-file; index. It's possible for this to happen quickly
209 # enough that new-file ends up with the same timestamp as the first
210 # index, and then bup will ignore it.
211 tick_start = time.time()
212 time.sleep(1 - (tick_start - int(tick_start)))
214 git.check_repo_or_die()
215 indexfile = opt.indexfile or git.repo('bupindex')
220 log('check: starting initial check.\n')
221 check_index(index.Reader(indexfile))
224 log('clear: clearing index.\n')
225 clear_index(indexfile)
227 excluded_paths = drecurse.parse_excludes(flags)
229 paths = index.reduce_paths(extra)
233 o.fatal('update mode (-u) requested but no paths given')
234 for (rp,path) in paths:
235 update_index(rp, excluded_paths)
237 if opt['print'] or opt.status or opt.modified:
238 for (name, ent) in index.Reader(indexfile).filter(extra or ['']):
240 and (ent.is_valid() or ent.is_deleted() or not ent.mode)):
246 elif not ent.is_valid():
247 if ent.sha == index.EMPTY_SHA:
254 line += ent.sha.encode('hex') + ' '
256 line += "%7s %7s " % (oct(ent.mode), oct(ent.gitmode))
257 print line + (name or './')
259 if opt.check and (opt['print'] or opt.status or opt.modified or opt.update):
260 log('check: starting final check.\n')
261 check_index(index.Reader(indexfile))
264 log('WARNING: %d errors encountered.\n' % len(saved_errors))