annotate contrib/hg-relink @ 8849:80cc4b1a62d0

compare grep result between target and its parent I found that typical case is that grep target is added at (*) revision in the tree shown below. +--- 1(*) --- 3 0 +--- 2 ------ 4 Now, I expect 'hg grep --all' to show only rev:1 which is first appearance of target line. But 'hg grep --all' will tell: target line dis-appeared at 3 => 4 target line appeared at 2 => 3 target line dis-appeared at 1 => 2 target line appeared at 0 => 1 because current 'hg grep' implementation compares not between target revision and its parent, but between neighbor revisions in walkthrough order. I checked performance of this patch by "hg grep --follow --all walkchangerevs" on whole Mercurial repo, and patched version could complete as fast as un-patched one.
author FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
date Tue, 19 May 2009 16:49:54 +0900
parents 46293a0c7e9f
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
4249
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
1 #!/usr/bin/env python
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
2 #
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
3 # Copyright (C) 2007 Brendan Cully <brendan@kublai.com>
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
4 #
8225
46293a0c7e9f updated license to be explicit about GPL version 2
Martin Geisler <mg@lazybytes.net>
parents: 4270
diff changeset
5 # This software may be used and distributed according to the terms of the
46293a0c7e9f updated license to be explicit about GPL version 2
Martin Geisler <mg@lazybytes.net>
parents: 4270
diff changeset
6 # GNU General Public License version 2, incorporated herein by reference.
4249
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
7
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
8 import os, sys
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
9
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
10 class ConfigError(Exception): pass
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
11
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
12 def usage():
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
13 print """relink <source> <destination>
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
14 Recreate hard links between source and destination repositories"""
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
15
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
16 class Config:
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
17 def __init__(self, args):
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
18 if len(args) != 3:
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
19 raise ConfigError("wrong number of arguments")
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
20 self.src = os.path.abspath(args[1])
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
21 self.dst = os.path.abspath(args[2])
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
22 for d in (self.src, self.dst):
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
23 if not os.path.exists(os.path.join(d, '.hg')):
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
24 raise ConfigError("%s: not a mercurial repository" % d)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
25
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
26 def collect(src):
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
27 seplen = len(os.path.sep)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
28 candidates = []
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
29 for dirpath, dirnames, filenames in os.walk(src):
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
30 relpath = dirpath[len(src) + seplen:]
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
31 for filename in filenames:
4270
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
32 if not filename.endswith('.i'):
4249
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
33 continue
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
34 st = os.stat(os.path.join(dirpath, filename))
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
35 candidates.append((os.path.join(relpath, filename), st))
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
36
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
37 return candidates
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
38
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
39 def prune(candidates, dst):
4270
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
40 def getdatafile(path):
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
41 if not path.endswith('.i'):
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
42 return None, None
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
43 df = path[:-1] + 'd'
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
44 try:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
45 st = os.stat(df)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
46 except OSError:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
47 return None, None
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
48 return df, st
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
49
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
50 def linkfilter(dst, st):
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
51 try:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
52 ts = os.stat(dst)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
53 except OSError:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
54 # Destination doesn't have this file?
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
55 return False
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
56 if st.st_ino == ts.st_ino:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
57 return False
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
58 if st.st_dev != ts.st_dev:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
59 # No point in continuing
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
60 raise Exception('Source and destination are on different devices')
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
61 if st.st_size != ts.st_size:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
62 # TODO: compare revlog heads
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
63 return False
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
64 return st
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
65
4249
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
66 targets = []
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
67 for fn, st in candidates:
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
68 tgt = os.path.join(dst, fn)
4270
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
69 ts = linkfilter(tgt, st)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
70 if not ts:
4249
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
71 continue
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
72 targets.append((fn, ts.st_size))
4270
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
73 df, ts = getdatafile(tgt)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
74 if df:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
75 targets.append((fn[:-1] + 'd', ts.st_size))
4249
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
76
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
77 return targets
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
78
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
79 def relink(src, dst, files):
4270
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
80 def relinkfile(src, dst):
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
81 bak = dst + '.bak'
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
82 os.rename(dst, bak)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
83 try:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
84 os.link(src, dst)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
85 except OSError:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
86 os.rename(bak, dst)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
87 raise
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
88 os.remove(bak)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
89
4249
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
90 CHUNKLEN = 65536
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
91 relinked = 0
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
92 savedbytes = 0
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
93
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
94 for f, sz in files:
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
95 source = os.path.join(src, f)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
96 tgt = os.path.join(dst, f)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
97 sfp = file(source)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
98 dfp = file(tgt)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
99 sin = sfp.read(CHUNKLEN)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
100 while sin:
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
101 din = dfp.read(CHUNKLEN)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
102 if sin != din:
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
103 break
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
104 sin = sfp.read(CHUNKLEN)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
105 if sin:
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
106 continue
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
107 try:
4270
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
108 relinkfile(source, tgt)
4249
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
109 print 'Relinked %s' % f
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
110 relinked += 1
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
111 savedbytes += sz
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
112 except OSError, inst:
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
113 print '%s: %s' % (tgt, str(inst))
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
114
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
115 print 'Relinked %d files (%d bytes reclaimed)' % (relinked, savedbytes)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
116
4270
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
117 try:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
118 cfg = Config(sys.argv)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
119 except ConfigError, inst:
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
120 print str(inst)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
121 usage()
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
122 sys.exit(1)
29eb88bd5c8d hg-relink: do not compare .d files
Brendan Cully <brendan@kublai.com>
parents: 4249
diff changeset
123
4249
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
124 src = os.path.join(cfg.src, '.hg')
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
125 dst = os.path.join(cfg.dst, '.hg')
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
126 candidates = collect(src)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
127 targets = prune(candidates, dst)
7663780b55a7 Add hg-relink script to contrib
Brendan Cully <brendan@kublai.com>
parents:
diff changeset
128 relink(src, dst, targets)