3 from bup import hashsplit, git, options, client
4 from bup.helpers import *
8 bup split [-tcb] [-n name] [--bench] [filenames...]
10 r,remote= remote repository path
11 b,blobs output a series of blob ids
12 t,tree output a tree id
13 c,commit output a commit id
14 n,name= name of backup set to update (if any)
15 d,date= date for the commit (seconds since the epoch)
16 q,quiet don't print progress messages
17 v,verbose increase log output (can be used more than once)
18 git-ids read a list of git object ids from stdin and split their contents
19 keep-boundaries don't let one chunk span two input files
20 noop don't actually save the data anywhere
21 copy just copy input to output, hashsplitting along the way
22 bench print benchmark timings to stderr
23 max-pack-size= maximum bytes in a single pack
24 max-pack-objects= maximum number of objects in a single pack
25 fanout= maximum number of blobs in a single tree
26 bwlimit= maximum bytes/sec to transmit to server
28 o = options.Options('bup split', optspec)
29 (opt, flags, extra) = o.parse(sys.argv[1:])
32 git.check_repo_or_die()
33 if not (opt.blobs or opt.tree or opt.commit or opt.name or
34 opt.noop or opt.copy):
35 o.fatal("use one or more of -b, -t, -c, -n, -N, --copy")
36 if (opt.noop or opt.copy) and (opt.blobs or opt.tree or
37 opt.commit or opt.name):
38 o.fatal('-N and --copy are incompatible with -b, -t, -c, -n')
39 if extra and opt.git_ids:
40 o.fatal("don't provide filenames when using --git-ids")
43 git.verbose = opt.verbose - 1
46 hashsplit.max_pack_size = parse_num(opt.max_pack_size)
47 if opt.max_pack_objects:
48 hashsplit.max_pack_objects = parse_num(opt.max_pack_objects)
50 hashsplit.fanout = parse_num(opt.fanout)
54 client.bwlimit = parse_num(opt.bwlimit)
56 date = parse_date_or_fatal(opt.date, o.fatal)
61 last_prog = total_bytes = 0
62 def prog(filenum, nbytes):
63 global last_prog, total_bytes
66 if now - last_prog < 0.2:
69 progress('Splitting: file #%d, %d kbytes\r'
70 % (filenum+1, total_bytes/1024))
72 progress('Splitting: %d kbytes\r' % (total_bytes/1024))
76 is_reverse = os.environ.get('BUP_SERVER_REVERSE')
77 if is_reverse and opt.remote:
78 o.fatal("don't use -r in reverse mode; it's automatic")
79 start_time = time.time()
81 refname = opt.name and 'refs/heads/%s' % opt.name or None
82 if opt.noop or opt.copy:
83 cli = pack_writer = oldref = None
84 elif opt.remote or is_reverse:
85 cli = client.Client(opt.remote)
86 oldref = refname and cli.read_ref(refname) or None
87 pack_writer = cli.new_packwriter()
90 oldref = refname and git.read_ref(refname) or None
91 pack_writer = git.PackWriter()
94 # the input is actually a series of git object ids that we should retrieve
97 # This is a bit messy, but basically it converts from a series of
98 # CatPipe.get() iterators into a series of file-type objects.
99 # It would be less ugly if either CatPipe.get() returned a file-like object
100 # (not very efficient), or split_to_shalist() expected an iterator instead
104 def __init__(self, it):
106 def read(self, size):
111 line = sys.stdin.readline()
117 it = cp.get(line.strip())
118 next(it) # skip the file type
120 add_error('error: %s' % e)
125 # the input either comes from a series of files or from stdin.
126 files = extra and (open(fn) for fn in extra) or [sys.stdin]
129 shalist = hashsplit.split_to_shalist(pack_writer, files,
130 keep_boundaries=opt.keep_boundaries,
132 tree = pack_writer.new_tree(shalist)
135 for (blob, bits) in hashsplit.hashsplit_iter(files,
136 keep_boundaries=opt.keep_boundaries,
138 hashsplit.total_split += len(blob)
140 sys.stdout.write(str(blob))
141 megs = hashsplit.total_split/1024/1024
142 if not opt.quiet and last != megs:
143 progress('%d Mbytes read\r' % megs)
145 progress('%d Mbytes read, done.\n' % megs)
150 for (mode,name,bin) in shalist:
151 print bin.encode('hex')
153 print tree.encode('hex')
154 if opt.commit or opt.name:
155 msg = 'bup split\n\nGenerated by command:\n%r' % sys.argv
156 ref = opt.name and ('refs/heads/%s' % opt.name) or None
157 commit = pack_writer.new_commit(oldref, tree, date, msg)
159 print commit.encode('hex')
162 pack_writer.close() # must close before we can update the ref
166 cli.update_ref(refname, commit, oldref)
168 git.update_ref(refname, commit, oldref)
173 secs = time.time() - start_time
174 size = hashsplit.total_split
176 log('\nbup: %.2fkbytes in %.2f secs = %.2f kbytes/sec\n'
177 % (size/1024., secs, size/1024./secs))
180 log('WARNING: %d errors encountered while saving.\n' % len(saved_errors))