2 import sys, math, struct, glob, sha
7 SHA_PER_PAGE=PAGE_SIZE/200.
17 def merge(idxlist, total, bits, table):
18 iters = [iter(i) for i in idxlist]
19 iters = [[next(it), it] for it in iters]
23 if (count % 10000) == 0:
24 log('\rMerging: %.2f%% (%d/%d)'
25 % (count*100.0/total, count, total))
29 prefix = git.extract_bits(e, bits)
31 e = iters[0][0] = next(iters[0][1])
40 iters = iters[1:i] + [iters[0]] + iters[i:]
41 log('\rMerging: done. \n')
44 def do_midx(outdir, outfilename, infilenames):
47 sum = sha.sha('\0'.join(infilenames)).hexdigest()
48 outfilename = '%s/midx-%s.midx' % (outdir, sum)
52 for name in infilenames:
53 ix = git.PackIndex(name)
58 log('%s: no new .idx files: nothing to do.\n' % outdir)
61 log('Merging %d indexes (%d objects).\n' % (len(infilenames), total))
62 pages = total/SHA_PER_PAGE
63 bits = int(math.ceil(math.log(pages, 2)))
65 log('table size: %d (%d bits)\n' % (entries*8, bits))
70 os.unlink(outfilename)
73 f = open(outfilename + '.tmp', 'w+')
74 f.write('MIDX\0\0\0\1')
75 f.write(struct.pack('!I', bits))
76 assert(f.tell() == 12)
77 f.write('\0'*8*entries)
79 for e in merge(inp, total, bits, table):
82 f.write('\0'.join([os.path.basename(p) for p in infilenames]))
85 f.write(struct.pack('!%dQ' % entries, *table))
87 os.rename(outfilename + '.tmp', outfilename)
89 # this is just for testing
91 p = git.PackMidx(outfilename)
92 assert(len(p.idxnames) == len(infilenames))
94 assert(len(p) == total)
96 for i in merge(inp, total, bits, table):
97 assert(i == pi.next())
103 bup midx [options...] <idxnames...>
105 o,output= output midx filename (default: auto-generated)
106 a,auto automatically create .midx from any unindexed .idx files
107 f,force automatically create .midx from *all* .idx files
109 o = options.Options('bup midx', optspec)
110 (opt, flags, extra) = o.parse(sys.argv[1:])
112 if extra and (opt.auto or opt.force):
113 log("bup midx: you can't use -f/-a and also provide filenames\n")
116 git.check_repo_or_die()
119 do_midx(git.repo('objects/pack'), opt.output, extra)
120 elif opt.auto or opt.force:
121 paths = [git.repo('objects/pack')]
122 paths += glob.glob(git.repo('index-cache/*/.'))
125 do_midx(path, opt.output, glob.glob('%s/*.idx' % path))
128 m = git.MultiPackIndex(path)
130 for pack in m.packs: # only .idx files without a .midx are open
131 if pack.name.endswith('.idx'):
132 needed[pack.name] = 1
134 do_midx(path, opt.output, needed.keys())
136 log("bup midx: you must use -f or -a or provide input filenames\n")