Show More
@@ -1,160 +1,159 | |||||
1 | # Mercurial extension to provide 'hg relink' command |
|
1 | # Mercurial extension to provide 'hg relink' command | |
2 | # |
|
2 | # | |
3 | # Copyright (C) 2007 Brendan Cully <brendan@kublai.com> |
|
3 | # Copyright (C) 2007 Brendan Cully <brendan@kublai.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | """recreates hardlinks between repository clones""" |
|
8 | """recreates hardlinks between repository clones""" | |
9 |
|
9 | |||
10 | from mercurial import cmdutil, hg, util |
|
10 | from mercurial import cmdutil, hg, util | |
11 | from mercurial.i18n import _ |
|
11 | from mercurial.i18n import _ | |
12 | import os, stat |
|
12 | import os, stat | |
13 |
|
13 | |||
14 | def relink(ui, repo, origin=None, **opts): |
|
14 | def relink(ui, repo, origin=None, **opts): | |
15 | """recreate hardlinks between two repositories |
|
15 | """recreate hardlinks between two repositories | |
16 |
|
16 | |||
17 | When repositories are cloned locally, their data files will be |
|
17 | When repositories are cloned locally, their data files will be | |
18 | hardlinked so that they only use the space of a single repository. |
|
18 | hardlinked so that they only use the space of a single repository. | |
19 |
|
19 | |||
20 | Unfortunately, subsequent pulls into either repository will break |
|
20 | Unfortunately, subsequent pulls into either repository will break | |
21 | hardlinks for any files touched by the new changesets, even if |
|
21 | hardlinks for any files touched by the new changesets, even if | |
22 | both repositories end up pulling the same changes. |
|
22 | both repositories end up pulling the same changes. | |
23 |
|
23 | |||
24 | Similarly, passing --rev to "hg clone" will fail to use any |
|
24 | Similarly, passing --rev to "hg clone" will fail to use any | |
25 | hardlinks, falling back to a complete copy of the source |
|
25 | hardlinks, falling back to a complete copy of the source | |
26 | repository. |
|
26 | repository. | |
27 |
|
27 | |||
28 | This command lets you recreate those hardlinks and reclaim that |
|
28 | This command lets you recreate those hardlinks and reclaim that | |
29 | wasted space. |
|
29 | wasted space. | |
30 |
|
30 | |||
31 | This repository will be relinked to share space with ORIGIN, which |
|
31 | This repository will be relinked to share space with ORIGIN, which | |
32 | must be on the same local disk. If ORIGIN is omitted, looks for |
|
32 | must be on the same local disk. If ORIGIN is omitted, looks for | |
33 | "default-relink", then "default", in [paths]. |
|
33 | "default-relink", then "default", in [paths]. | |
34 |
|
34 | |||
35 | Do not attempt any read operations on this repository while the |
|
35 | Do not attempt any read operations on this repository while the | |
36 | command is running. (Both repositories will be locked against |
|
36 | command is running. (Both repositories will be locked against | |
37 | writes.) |
|
37 | writes.) | |
38 | """ |
|
38 | """ | |
39 | if not hasattr(util, 'samefile') or not hasattr(util, 'samedevice'): |
|
39 | if not hasattr(util, 'samefile') or not hasattr(util, 'samedevice'): | |
40 | raise util.Abort(_('hardlinks are not supported on this system')) |
|
40 | raise util.Abort(_('hardlinks are not supported on this system')) | |
41 | src = hg.repository( |
|
41 | src = hg.repository( | |
42 | cmdutil.remoteui(repo, opts), |
|
42 | cmdutil.remoteui(repo, opts), | |
43 | ui.expandpath(origin or 'default-relink', origin or 'default')) |
|
43 | ui.expandpath(origin or 'default-relink', origin or 'default')) | |
44 | if not src.local(): |
|
44 | if not src.local(): | |
45 | raise util.Abort('must specify local origin repository') |
|
45 | raise util.Abort('must specify local origin repository') | |
46 | ui.status(_('relinking %s to %s\n') % (src.store.path, repo.store.path)) |
|
46 | ui.status(_('relinking %s to %s\n') % (src.store.path, repo.store.path)) | |
47 | locallock = repo.lock() |
|
47 | locallock = repo.lock() | |
48 | try: |
|
48 | try: | |
49 | remotelock = src.lock() |
|
49 | remotelock = src.lock() | |
50 | try: |
|
50 | try: | |
51 | candidates = sorted(collect(src.store.path, ui)) |
|
51 | candidates = sorted(collect(src.store.path, ui)) | |
52 | targets = prune(candidates, src.store.path, repo.store.path, ui) |
|
52 | targets = prune(candidates, src.store.path, repo.store.path, ui) | |
53 | do_relink(src.store.path, repo.store.path, targets, ui) |
|
53 | do_relink(src.store.path, repo.store.path, targets, ui) | |
54 | finally: |
|
54 | finally: | |
55 | remotelock.release() |
|
55 | remotelock.release() | |
56 | finally: |
|
56 | finally: | |
57 | locallock.release() |
|
57 | locallock.release() | |
58 |
|
58 | |||
59 | def collect(src, ui): |
|
59 | def collect(src, ui): | |
60 | seplen = len(os.path.sep) |
|
60 | seplen = len(os.path.sep) | |
61 | candidates = [] |
|
61 | candidates = [] | |
62 | for dirpath, dirnames, filenames in os.walk(src): |
|
62 | for dirpath, dirnames, filenames in os.walk(src): | |
63 | relpath = dirpath[len(src) + seplen:] |
|
63 | relpath = dirpath[len(src) + seplen:] | |
64 | for filename in filenames: |
|
64 | for filename in filenames: | |
65 | if not filename[-2:] in ('.d', '.i'): |
|
65 | if not filename[-2:] in ('.d', '.i'): | |
66 | continue |
|
66 | continue | |
67 | st = os.stat(os.path.join(dirpath, filename)) |
|
67 | st = os.stat(os.path.join(dirpath, filename)) | |
68 | if not stat.S_ISREG(st.st_mode): |
|
68 | if not stat.S_ISREG(st.st_mode): | |
69 | continue |
|
69 | continue | |
70 | candidates.append((os.path.join(relpath, filename), st)) |
|
70 | candidates.append((os.path.join(relpath, filename), st)) | |
71 |
|
71 | |||
72 | ui.status(_('collected %d candidate storage files\n') % len(candidates)) |
|
72 | ui.status(_('collected %d candidate storage files\n') % len(candidates)) | |
73 | return candidates |
|
73 | return candidates | |
74 |
|
74 | |||
75 | def prune(candidates, src, dst, ui): |
|
75 | def prune(candidates, src, dst, ui): | |
76 | def linkfilter(src, dst, st): |
|
76 | def linkfilter(src, dst, st): | |
77 | try: |
|
77 | try: | |
78 | ts = os.stat(dst) |
|
78 | ts = os.stat(dst) | |
79 | except OSError: |
|
79 | except OSError: | |
80 | # Destination doesn't have this file? |
|
80 | # Destination doesn't have this file? | |
81 | return False |
|
81 | return False | |
82 | if util.samefile(src, dst): |
|
82 | if util.samefile(src, dst): | |
83 | return False |
|
83 | return False | |
84 | if not util.samedevice(src, dst): |
|
84 | if not util.samedevice(src, dst): | |
85 | # No point in continuing |
|
85 | # No point in continuing | |
86 | raise util.Abort( |
|
86 | raise util.Abort( | |
87 | _('source and destination are on different devices')) |
|
87 | _('source and destination are on different devices')) | |
88 | if st.st_size != ts.st_size: |
|
88 | if st.st_size != ts.st_size: | |
89 | return False |
|
89 | return False | |
90 | return st |
|
90 | return st | |
91 |
|
91 | |||
92 | targets = [] |
|
92 | targets = [] | |
93 | for fn, st in candidates: |
|
93 | for fn, st in candidates: | |
94 | srcpath = os.path.join(src, fn) |
|
94 | srcpath = os.path.join(src, fn) | |
95 | tgt = os.path.join(dst, fn) |
|
95 | tgt = os.path.join(dst, fn) | |
96 | ts = linkfilter(srcpath, tgt, st) |
|
96 | ts = linkfilter(srcpath, tgt, st) | |
97 | if not ts: |
|
97 | if not ts: | |
98 | ui.debug(_('not linkable: %s\n') % fn) |
|
98 | ui.debug(_('not linkable: %s\n') % fn) | |
99 | continue |
|
99 | continue | |
100 | targets.append((fn, ts.st_size)) |
|
100 | targets.append((fn, ts.st_size)) | |
101 |
|
101 | |||
102 | ui.status(_('pruned down to %d probably relinkable files\n') % len(targets)) |
|
102 | ui.status(_('pruned down to %d probably relinkable files\n') % len(targets)) | |
103 | return targets |
|
103 | return targets | |
104 |
|
104 | |||
105 | def do_relink(src, dst, files, ui): |
|
105 | def do_relink(src, dst, files, ui): | |
106 | def relinkfile(src, dst): |
|
106 | def relinkfile(src, dst): | |
107 | bak = dst + '.bak' |
|
107 | bak = dst + '.bak' | |
108 | os.rename(dst, bak) |
|
108 | os.rename(dst, bak) | |
109 | try: |
|
109 | try: | |
110 | util.os_link(src, dst) |
|
110 | util.os_link(src, dst) | |
111 | except OSError: |
|
111 | except OSError: | |
112 | os.rename(bak, dst) |
|
112 | os.rename(bak, dst) | |
113 | raise |
|
113 | raise | |
114 | os.remove(bak) |
|
114 | os.remove(bak) | |
115 |
|
115 | |||
116 | CHUNKLEN = 65536 |
|
116 | CHUNKLEN = 65536 | |
117 | relinked = 0 |
|
117 | relinked = 0 | |
118 | savedbytes = 0 |
|
118 | savedbytes = 0 | |
119 | f = '' |
|
|||
120 |
|
119 | |||
121 | pos = 0 |
|
120 | pos = 0 | |
122 | total = len(files) |
|
121 | total = len(files) | |
123 | for f, sz in files: |
|
122 | for f, sz in files: | |
124 | pos += 1 |
|
123 | pos += 1 | |
125 | source = os.path.join(src, f) |
|
124 | source = os.path.join(src, f) | |
126 | tgt = os.path.join(dst, f) |
|
125 | tgt = os.path.join(dst, f) | |
127 | # Binary mode, so that read() works correctly, especially on Windows |
|
126 | # Binary mode, so that read() works correctly, especially on Windows | |
128 | sfp = file(source, 'rb') |
|
127 | sfp = file(source, 'rb') | |
129 | dfp = file(tgt, 'rb') |
|
128 | dfp = file(tgt, 'rb') | |
130 | sin = sfp.read(CHUNKLEN) |
|
129 | sin = sfp.read(CHUNKLEN) | |
131 | while sin: |
|
130 | while sin: | |
132 | din = dfp.read(CHUNKLEN) |
|
131 | din = dfp.read(CHUNKLEN) | |
133 | if sin != din: |
|
132 | if sin != din: | |
134 | break |
|
133 | break | |
135 | sin = sfp.read(CHUNKLEN) |
|
134 | sin = sfp.read(CHUNKLEN) | |
136 | sfp.close() |
|
135 | sfp.close() | |
137 | dfp.close() |
|
136 | dfp.close() | |
138 | if sin: |
|
137 | if sin: | |
139 | ui.debug(_('not linkable: %s\n') % f) |
|
138 | ui.debug(_('not linkable: %s\n') % f) | |
140 | continue |
|
139 | continue | |
141 | try: |
|
140 | try: | |
142 | relinkfile(source, tgt) |
|
141 | relinkfile(source, tgt) | |
143 | ui.progress(_('relinking'), pos, f, _(' files'), total) |
|
142 | ui.progress(_('relinking'), pos, f, _(' files'), total) | |
144 | relinked += 1 |
|
143 | relinked += 1 | |
145 | savedbytes += sz |
|
144 | savedbytes += sz | |
146 | except OSError, inst: |
|
145 | except OSError, inst: | |
147 | ui.warn('%s: %s\n' % (tgt, str(inst))) |
|
146 | ui.warn('%s: %s\n' % (tgt, str(inst))) | |
148 |
|
147 | |||
149 | ui.progress(_('relinking'), None) |
|
148 | ui.progress(_('relinking'), None) | |
150 |
|
149 | |||
151 | ui.status(_('relinked %d files (%d bytes reclaimed)\n') % |
|
150 | ui.status(_('relinked %d files (%d bytes reclaimed)\n') % | |
152 | (relinked, savedbytes)) |
|
151 | (relinked, savedbytes)) | |
153 |
|
152 | |||
154 | cmdtable = { |
|
153 | cmdtable = { | |
155 | 'relink': ( |
|
154 | 'relink': ( | |
156 | relink, |
|
155 | relink, | |
157 | [], |
|
156 | [], | |
158 | _('[ORIGIN]') |
|
157 | _('[ORIGIN]') | |
159 | ) |
|
158 | ) | |
160 | } |
|
159 | } |
General Comments 0
You need to be logged in to leave comments.
Login now