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