Show More
@@ -1,211 +1,215 b'' | |||
|
1 | 1 | # Mercurial extension to provide 'hg relink' command |
|
2 | 2 | # |
|
3 | 3 | # Copyright (C) 2007 Brendan Cully <brendan@kublai.com> |
|
4 | 4 | # |
|
5 | 5 | # This software may be used and distributed according to the terms of the |
|
6 | 6 | # GNU General Public License version 2 or any later version. |
|
7 | 7 | |
|
8 | 8 | """recreates hardlinks between repository clones""" |
|
9 | 9 | from __future__ import absolute_import |
|
10 | 10 | |
|
11 | 11 | import os |
|
12 | 12 | import stat |
|
13 | 13 | |
|
14 | 14 | from mercurial.i18n import _ |
|
15 | 15 | from mercurial.pycompat import open |
|
16 | 16 | from mercurial import ( |
|
17 | 17 | error, |
|
18 | 18 | hg, |
|
19 | 19 | registrar, |
|
20 | 20 | util, |
|
21 | 21 | ) |
|
22 |
from mercurial.utils import |
|
|
22 | from mercurial.utils import ( | |
|
23 | stringutil, | |
|
24 | urlutil, | |
|
25 | ) | |
|
23 | 26 | |
|
24 | 27 | cmdtable = {} |
|
25 | 28 | command = registrar.command(cmdtable) |
|
26 | 29 | # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for |
|
27 | 30 | # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should |
|
28 | 31 | # be specifying the version(s) of Mercurial they are tested with, or |
|
29 | 32 | # leave the attribute unspecified. |
|
30 | 33 | testedwith = b'ships-with-hg-core' |
|
31 | 34 | |
|
32 | 35 | |
|
33 | 36 | @command( |
|
34 | 37 | b'relink', [], _(b'[ORIGIN]'), helpcategory=command.CATEGORY_MAINTENANCE |
|
35 | 38 | ) |
|
36 | 39 | def relink(ui, repo, origin=None, **opts): |
|
37 | 40 | """recreate hardlinks between two repositories |
|
38 | 41 | |
|
39 | 42 | When repositories are cloned locally, their data files will be |
|
40 | 43 | hardlinked so that they only use the space of a single repository. |
|
41 | 44 | |
|
42 | 45 | Unfortunately, subsequent pulls into either repository will break |
|
43 | 46 | hardlinks for any files touched by the new changesets, even if |
|
44 | 47 | both repositories end up pulling the same changes. |
|
45 | 48 | |
|
46 | 49 | Similarly, passing --rev to "hg clone" will fail to use any |
|
47 | 50 | hardlinks, falling back to a complete copy of the source |
|
48 | 51 | repository. |
|
49 | 52 | |
|
50 | 53 | This command lets you recreate those hardlinks and reclaim that |
|
51 | 54 | wasted space. |
|
52 | 55 | |
|
53 | 56 | This repository will be relinked to share space with ORIGIN, which |
|
54 | 57 | must be on the same local disk. If ORIGIN is omitted, looks for |
|
55 | 58 | "default-relink", then "default", in [paths]. |
|
56 | 59 | |
|
57 | 60 | Do not attempt any read operations on this repository while the |
|
58 | 61 | command is running. (Both repositories will be locked against |
|
59 | 62 | writes.) |
|
60 | 63 | """ |
|
61 | 64 | if not util.safehasattr(util, b'samefile') or not util.safehasattr( |
|
62 | 65 | util, b'samedevice' |
|
63 | 66 | ): |
|
64 | 67 | raise error.Abort(_(b'hardlinks are not supported on this system')) |
|
65 | src = hg.repository( | |
|
66 | repo.baseui, | |
|
67 |
|
|
|
68 | ) | |
|
68 | ||
|
69 | if origin is None and b'default-relink' in ui.paths: | |
|
70 | origin = b'default-relink' | |
|
71 | path, __ = urlutil.get_unique_pull_path(b'relink', repo, ui, origin) | |
|
72 | src = hg.repository(repo.baseui, path) | |
|
69 | 73 | ui.status(_(b'relinking %s to %s\n') % (src.store.path, repo.store.path)) |
|
70 | 74 | if repo.root == src.root: |
|
71 | 75 | ui.status(_(b'there is nothing to relink\n')) |
|
72 | 76 | return |
|
73 | 77 | |
|
74 | 78 | if not util.samedevice(src.store.path, repo.store.path): |
|
75 | 79 | # No point in continuing |
|
76 | 80 | raise error.Abort(_(b'source and destination are on different devices')) |
|
77 | 81 | |
|
78 | 82 | with repo.lock(), src.lock(): |
|
79 | 83 | candidates = sorted(collect(src, ui)) |
|
80 | 84 | targets = prune(candidates, src.store.path, repo.store.path, ui) |
|
81 | 85 | do_relink(src.store.path, repo.store.path, targets, ui) |
|
82 | 86 | |
|
83 | 87 | |
|
84 | 88 | def collect(src, ui): |
|
85 | 89 | seplen = len(os.path.sep) |
|
86 | 90 | candidates = [] |
|
87 | 91 | live = len(src[b'tip'].manifest()) |
|
88 | 92 | # Your average repository has some files which were deleted before |
|
89 | 93 | # the tip revision. We account for that by assuming that there are |
|
90 | 94 | # 3 tracked files for every 2 live files as of the tip version of |
|
91 | 95 | # the repository. |
|
92 | 96 | # |
|
93 | 97 | # mozilla-central as of 2010-06-10 had a ratio of just over 7:5. |
|
94 | 98 | total = live * 3 // 2 |
|
95 | 99 | src = src.store.path |
|
96 | 100 | progress = ui.makeprogress(_(b'collecting'), unit=_(b'files'), total=total) |
|
97 | 101 | pos = 0 |
|
98 | 102 | ui.status( |
|
99 | 103 | _(b"tip has %d files, estimated total number of files: %d\n") |
|
100 | 104 | % (live, total) |
|
101 | 105 | ) |
|
102 | 106 | for dirpath, dirnames, filenames in os.walk(src): |
|
103 | 107 | dirnames.sort() |
|
104 | 108 | relpath = dirpath[len(src) + seplen :] |
|
105 | 109 | for filename in sorted(filenames): |
|
106 | 110 | if filename[-2:] not in (b'.d', b'.i'): |
|
107 | 111 | continue |
|
108 | 112 | st = os.stat(os.path.join(dirpath, filename)) |
|
109 | 113 | if not stat.S_ISREG(st.st_mode): |
|
110 | 114 | continue |
|
111 | 115 | pos += 1 |
|
112 | 116 | candidates.append((os.path.join(relpath, filename), st)) |
|
113 | 117 | progress.update(pos, item=filename) |
|
114 | 118 | |
|
115 | 119 | progress.complete() |
|
116 | 120 | ui.status(_(b'collected %d candidate storage files\n') % len(candidates)) |
|
117 | 121 | return candidates |
|
118 | 122 | |
|
119 | 123 | |
|
120 | 124 | def prune(candidates, src, dst, ui): |
|
121 | 125 | def linkfilter(src, dst, st): |
|
122 | 126 | try: |
|
123 | 127 | ts = os.stat(dst) |
|
124 | 128 | except OSError: |
|
125 | 129 | # Destination doesn't have this file? |
|
126 | 130 | return False |
|
127 | 131 | if util.samefile(src, dst): |
|
128 | 132 | return False |
|
129 | 133 | if not util.samedevice(src, dst): |
|
130 | 134 | # No point in continuing |
|
131 | 135 | raise error.Abort( |
|
132 | 136 | _(b'source and destination are on different devices') |
|
133 | 137 | ) |
|
134 | 138 | if st.st_size != ts.st_size: |
|
135 | 139 | return False |
|
136 | 140 | return st |
|
137 | 141 | |
|
138 | 142 | targets = [] |
|
139 | 143 | progress = ui.makeprogress( |
|
140 | 144 | _(b'pruning'), unit=_(b'files'), total=len(candidates) |
|
141 | 145 | ) |
|
142 | 146 | pos = 0 |
|
143 | 147 | for fn, st in candidates: |
|
144 | 148 | pos += 1 |
|
145 | 149 | srcpath = os.path.join(src, fn) |
|
146 | 150 | tgt = os.path.join(dst, fn) |
|
147 | 151 | ts = linkfilter(srcpath, tgt, st) |
|
148 | 152 | if not ts: |
|
149 | 153 | ui.debug(b'not linkable: %s\n' % fn) |
|
150 | 154 | continue |
|
151 | 155 | targets.append((fn, ts.st_size)) |
|
152 | 156 | progress.update(pos, item=fn) |
|
153 | 157 | |
|
154 | 158 | progress.complete() |
|
155 | 159 | ui.status( |
|
156 | 160 | _(b'pruned down to %d probably relinkable files\n') % len(targets) |
|
157 | 161 | ) |
|
158 | 162 | return targets |
|
159 | 163 | |
|
160 | 164 | |
|
161 | 165 | def do_relink(src, dst, files, ui): |
|
162 | 166 | def relinkfile(src, dst): |
|
163 | 167 | bak = dst + b'.bak' |
|
164 | 168 | os.rename(dst, bak) |
|
165 | 169 | try: |
|
166 | 170 | util.oslink(src, dst) |
|
167 | 171 | except OSError: |
|
168 | 172 | os.rename(bak, dst) |
|
169 | 173 | raise |
|
170 | 174 | os.remove(bak) |
|
171 | 175 | |
|
172 | 176 | CHUNKLEN = 65536 |
|
173 | 177 | relinked = 0 |
|
174 | 178 | savedbytes = 0 |
|
175 | 179 | |
|
176 | 180 | progress = ui.makeprogress( |
|
177 | 181 | _(b'relinking'), unit=_(b'files'), total=len(files) |
|
178 | 182 | ) |
|
179 | 183 | pos = 0 |
|
180 | 184 | for f, sz in files: |
|
181 | 185 | pos += 1 |
|
182 | 186 | source = os.path.join(src, f) |
|
183 | 187 | tgt = os.path.join(dst, f) |
|
184 | 188 | # Binary mode, so that read() works correctly, especially on Windows |
|
185 | 189 | sfp = open(source, b'rb') |
|
186 | 190 | dfp = open(tgt, b'rb') |
|
187 | 191 | sin = sfp.read(CHUNKLEN) |
|
188 | 192 | while sin: |
|
189 | 193 | din = dfp.read(CHUNKLEN) |
|
190 | 194 | if sin != din: |
|
191 | 195 | break |
|
192 | 196 | sin = sfp.read(CHUNKLEN) |
|
193 | 197 | sfp.close() |
|
194 | 198 | dfp.close() |
|
195 | 199 | if sin: |
|
196 | 200 | ui.debug(b'not linkable: %s\n' % f) |
|
197 | 201 | continue |
|
198 | 202 | try: |
|
199 | 203 | relinkfile(source, tgt) |
|
200 | 204 | progress.update(pos, item=f) |
|
201 | 205 | relinked += 1 |
|
202 | 206 | savedbytes += sz |
|
203 | 207 | except OSError as inst: |
|
204 | 208 | ui.warn(b'%s: %s\n' % (tgt, stringutil.forcebytestr(inst))) |
|
205 | 209 | |
|
206 | 210 | progress.complete() |
|
207 | 211 | |
|
208 | 212 | ui.status( |
|
209 | 213 | _(b'relinked %d files (%s reclaimed)\n') |
|
210 | 214 | % (relinked, util.bytecount(savedbytes)) |
|
211 | 215 | ) |
General Comments 0
You need to be logged in to leave comments.
Login now