2 import sys, math, struct, glob, resource
3 from bup import options, git
4 from bup.helpers import *
7 SHA_PER_PAGE=PAGE_SIZE/20.
10 bup midx [options...] <idxnames...>
12 o,output= output midx filename (default: auto-generated)
13 a,auto automatically create .midx from any unindexed .idx files
14 f,force automatically create .midx from *all* .idx files
15 max-files= maximum number of idx files to open at once [-1]
19 for i in xrange(0, len(l), count):
24 mf = min(resource.getrlimit(resource.RLIMIT_NOFILE))
26 mf -= 20 # just a safety margin
28 mf -= 6 # minimum safety margin
32 def merge(idxlist, bits, table):
34 for e in git.idxmerge(idxlist):
36 prefix = git.extract_bits(e, bits)
41 def _do_midx(outdir, outfilename, infilenames):
44 sum = Sha1('\0'.join(infilenames)).hexdigest()
45 outfilename = '%s/midx-%s.midx' % (outdir, sum)
50 for name in infilenames:
51 ix = git.open_idx(name)
57 log('Merging %d indexes (%d objects).\n' % (len(infilenames), total))
58 if (not opt.force and (total < 1024 and len(infilenames) < 3)) \
59 or len(infilenames) < 2 \
60 or (opt.force and not total):
61 log('midx: nothing to do.\n')
64 pages = int(total/SHA_PER_PAGE) or 1
65 bits = int(math.ceil(math.log(pages, 2)))
67 log('Table size: %d (%d bits)\n' % (entries*4, bits))
72 os.unlink(outfilename)
75 f = open(outfilename + '.tmp', 'w+')
76 f.write('MIDX\0\0\0\2')
77 f.write(struct.pack('!I', bits))
78 assert(f.tell() == 12)
79 f.write('\0'*4*entries)
81 for e in merge(inp, bits, table):
84 f.write('\0'.join(os.path.basename(p) for p in allfilenames.keys()))
87 f.write(struct.pack('!%dI' % entries, *table))
89 os.rename(outfilename + '.tmp', outfilename)
91 # this is just for testing
93 p = git.PackMidx(outfilename)
94 assert(len(p.idxnames) == len(infilenames))
96 assert(len(p) == total)
98 for i in merge(inp, total, bits, table):
99 assert(i == pi.next())
102 return total,outfilename
105 def do_midx(outdir, outfilename, infilenames):
106 rv = _do_midx(outdir, outfilename, infilenames)
111 def do_midx_dir(path):
114 if opt.force and not opt.auto:
115 midxs = [] # don't use existing midx files
117 midxs = glob.glob('%s/*.midx' % path)
120 m = git.open_idx(mname)
121 contents[mname] = [('%s/%s' % (path,i)) for i in m.idxnames]
122 sizes[mname] = len(m)
124 # sort the biggest midxes first, so that we can eliminate smaller
125 # redundant ones that come later in the list
126 midxs.sort(lambda x,y: -cmp(sizes[x], sizes[y]))
130 for iname in contents[mname]:
131 if not already.get(iname):
135 log('%r is redundant\n' % mname)
139 midxs = [k for k in midxs if not already.get(k)]
140 idxs = [k for k in glob.glob('%s/*.idx' % path) if not already.get(k)]
143 i = git.open_idx(iname)
144 sizes[iname] = len(i)
146 all = [(sizes[n],n) for n in (midxs + idxs)]
148 # FIXME: what are the optimal values? Does this make sense?
149 DESIRED_HWM = opt.force and 1 or 5
150 DESIRED_LWM = opt.force and 1 or 2
151 existed = dict((name,1) for sz,name in all)
152 log('%d indexes; want no more than %d.\n' % (len(all), DESIRED_HWM))
153 if len(all) <= DESIRED_HWM:
154 log('Nothing to do.\n')
155 while len(all) > DESIRED_HWM:
157 part1 = [name for sz,name in all[:len(all)-DESIRED_LWM+1]]
158 part2 = all[len(all)-DESIRED_LWM+1:]
159 all = list(do_midx_group(path, part1)) + part2
160 if len(all) > DESIRED_HWM:
161 log('\nStill too many indexes (%d > %d). Merging again.\n'
162 % (len(all), DESIRED_HWM))
165 if not existed.get(name):
169 def do_midx_group(outdir, infiles):
170 for sublist in _group(infiles, opt.max_files):
171 rv = _do_midx(path, None, sublist)
177 o = options.Options('bup midx', optspec)
178 (opt, flags, extra) = o.parse(sys.argv[1:])
180 if extra and (opt.auto or opt.force):
181 o.fatal("you can't use -f/-a and also provide filenames")
183 git.check_repo_or_die()
185 if opt.max_files < 0:
186 opt.max_files = max_files()
187 assert(opt.max_files >= 5)
190 do_midx(git.repo('objects/pack'), opt.output, extra)
191 elif opt.auto or opt.force:
192 paths = [git.repo('objects/pack')]
193 paths += glob.glob(git.repo('index-cache/*/.'))
195 log('midx: scanning %s\n' % path)
199 o.fatal("you must use -f or -a or provide input filenames")