3 bup_python="$(dirname "$0")/bup-python" || exit $?
4 exec "$bup_python" "$0" ${1+"$@"}
8 from __future__ import print_function
9 from collections import defaultdict
10 from itertools import groupby
11 from sys import stderr
12 from time import localtime, strftime, time
15 from bup import git, options
16 from bup.gc import bup_gc
17 from bup.helpers import die_if_errors, log, partition, period_as_secs
18 from bup.rm import bup_rm
21 def branches(refnames=()):
22 return ((name[11:], sha.encode('hex')) for (name,sha)
23 in git.list_refs(patterns=('refs/heads/' + n for n in refnames),
26 def save_name(branch, utc):
27 return branch + '/' + strftime('%Y-%m-%d-%H%M%S', localtime(utc))
29 def classify_saves(saves, period_start):
30 """For each (utc, id) in saves, yield (True, (utc, id)) if the save
31 should be kept and (False, (utc, id)) if the save should be removed.
32 The ids are binary hashes.
35 def retain_newest_in_region(region):
36 for save in region[0:1]:
38 for save in region[1:]:
41 matches, rest = partition(lambda s: s[0] >= period_start['all'], saves)
45 tm_ranges = ((period_start['dailies'], lambda s: localtime(s[0]).tm_yday),
46 (period_start['monthlies'], lambda s: localtime(s[0]).tm_mon),
47 (period_start['yearlies'], lambda s: localtime(s[0]).tm_year))
49 # Break the decreasing utc sorted saves up into the respective
50 # period ranges (dailies, monthlies, ...). Within each range,
51 # group the saves by the period scale (days, months, ...), and
52 # then yield a "keep" action (True, utc) for the newest save in
53 # each group, and a "drop" action (False, utc) for the rest.
54 for pstart, time_region_id in tm_ranges:
55 matches, rest = partition(lambda s: s[0] >= pstart, rest)
56 for region_id, region_saves in groupby(matches, time_region_id):
57 for action in retain_newest_in_region(list(region_saves)):
60 # Finally, drop any saves older than the specified periods
66 bup prune-older [options...] [BRANCH...]
68 keep-all-for= retain all saves within the PERIOD
69 keep-dailies-for= retain the newest save per day within the PERIOD
70 keep-monthlies-for= retain the newest save per month within the PERIOD
71 keep-yearlies-for= retain the newest save per year within the PERIOD
72 wrt= end all periods at this number of seconds since the epoch
73 pretend don't prune, just report intended actions to standard output
74 gc collect garbage after removals [1]
75 gc-threshold= only rewrite a packfile if it's over this percent garbage [10]
76 #,compress= set compression level to # (0-9, 9 is highest) [1]
77 v,verbose increase log output (can be used more than once)
78 unsafe use the command even though it may be DANGEROUS
81 o = options.Options(optspec)
82 opt, flags, roots = o.parse(sys.argv[1:])
85 o.fatal('refusing to run dangerous, experimental command without --unsafe')
87 now = int(time()) if not opt.wrt else opt.wrt
88 if not isinstance(now, (int, long)):
89 o.fatal('--wrt value ' + str(now) + ' is not an integer')
92 for period, extent in (('all', opt.keep_all_for),
93 ('dailies', opt.keep_dailies_for),
94 ('monthlies', opt.keep_monthlies_for),
95 ('yearlies', opt.keep_yearlies_for)):
97 secs = period_as_secs(extent)
99 o.fatal('%r is not a valid period' % extent)
100 period_start[period] = now - secs
103 o.fatal('at least one keep argument is required')
105 period_start = defaultdict(lambda: float('inf'), period_start)
108 epoch_ymd = strftime('%Y-%m-%d-%H%M%S', localtime(0))
109 for kind in ['all', 'dailies', 'monthlies', 'yearlies']:
110 period_utc = period_start[kind]
111 if period_utc != float('inf'):
112 if not (period_utc > float('-inf')):
113 log('keeping all ' + kind)
116 when = strftime('%Y-%m-%d-%H%M%S', localtime(period_utc))
117 log('keeping ' + kind + ' since ' + when + '\n')
118 except ValueError as ex:
120 log('keeping %s since %d seconds before %s\n'
121 %(kind, abs(period_utc), epoch_ymd))
123 log('keeping %s since %d seconds after %s\n'
124 %(kind, period_utc, epoch_ymd))
126 log('keeping %s since %s\n' % (kind, epoch_ymd))
128 git.check_repo_or_die()
130 # This could be more efficient, but for now just build the whole list
131 # in memory and let bup_rm() do some redundant work.
134 author_secs = f.readline().strip()
135 return int(author_secs)
138 for branch, branch_id in branches(roots):
140 saves = ((utc, oidx.decode('hex')) for (oidx, utc) in
141 git.rev_list(branch_id, format='%at', parse=parse_info))
142 for keep_save, (utc, id) in classify_saves(saves, period_start):
143 assert(keep_save in (False, True))
144 # FIXME: base removals on hashes
146 print('+' if keep_save else '-', save_name(branch, utc))
148 removals.append(save_name(branch, utc))
152 bup_rm(removals, compression=opt.compress, verbosity=opt.verbose)
155 bup_gc(threshold=opt.gc_threshold,
156 compression=opt.compress,
157 verbosity=opt.verbose)