Show More
@@ -1,694 +1,694 b'' | |||||
1 | # narrowcommands.py - command modifications for narrowhg extension |
|
1 | # narrowcommands.py - command modifications for narrowhg extension | |
2 | # |
|
2 | # | |
3 | # Copyright 2017 Google, Inc. |
|
3 | # Copyright 2017 Google, Inc. | |
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 | from __future__ import absolute_import |
|
7 | from __future__ import absolute_import | |
8 |
|
8 | |||
9 | import itertools |
|
9 | import itertools | |
10 | import os |
|
10 | import os | |
11 |
|
11 | |||
12 | from mercurial.i18n import _ |
|
12 | from mercurial.i18n import _ | |
13 | from mercurial.node import ( |
|
13 | from mercurial.node import ( | |
14 | hex, |
|
14 | hex, | |
15 | short, |
|
15 | short, | |
16 | ) |
|
16 | ) | |
17 | from mercurial import ( |
|
17 | from mercurial import ( | |
18 | bundle2, |
|
18 | bundle2, | |
19 | cmdutil, |
|
19 | cmdutil, | |
20 | commands, |
|
20 | commands, | |
21 | discovery, |
|
21 | discovery, | |
22 | encoding, |
|
22 | encoding, | |
23 | error, |
|
23 | error, | |
24 | exchange, |
|
24 | exchange, | |
25 | extensions, |
|
25 | extensions, | |
26 | hg, |
|
26 | hg, | |
27 | narrowspec, |
|
27 | narrowspec, | |
28 | pathutil, |
|
28 | pathutil, | |
29 | pycompat, |
|
29 | pycompat, | |
30 | registrar, |
|
30 | registrar, | |
31 | repair, |
|
31 | repair, | |
32 | repoview, |
|
32 | repoview, | |
33 | requirements, |
|
33 | requirements, | |
34 | sparse, |
|
34 | sparse, | |
35 | util, |
|
35 | util, | |
36 | wireprototypes, |
|
36 | wireprototypes, | |
37 | ) |
|
37 | ) | |
38 | from mercurial.utils import ( |
|
38 | from mercurial.utils import ( | |
39 | urlutil, |
|
39 | urlutil, | |
40 | ) |
|
40 | ) | |
41 |
|
41 | |||
42 | table = {} |
|
42 | table = {} | |
43 | command = registrar.command(table) |
|
43 | command = registrar.command(table) | |
44 |
|
44 | |||
45 |
|
45 | |||
46 | def setup(): |
|
46 | def setup(): | |
47 | """Wraps user-facing mercurial commands with narrow-aware versions.""" |
|
47 | """Wraps user-facing mercurial commands with narrow-aware versions.""" | |
48 |
|
48 | |||
49 | entry = extensions.wrapcommand(commands.table, b'clone', clonenarrowcmd) |
|
49 | entry = extensions.wrapcommand(commands.table, b'clone', clonenarrowcmd) | |
50 | entry[1].append( |
|
50 | entry[1].append( | |
51 | (b'', b'narrow', None, _(b"create a narrow clone of select files")) |
|
51 | (b'', b'narrow', None, _(b"create a narrow clone of select files")) | |
52 | ) |
|
52 | ) | |
53 | entry[1].append( |
|
53 | entry[1].append( | |
54 | ( |
|
54 | ( | |
55 | b'', |
|
55 | b'', | |
56 | b'depth', |
|
56 | b'depth', | |
57 | b'', |
|
57 | b'', | |
58 | _(b"limit the history fetched by distance from heads"), |
|
58 | _(b"limit the history fetched by distance from heads"), | |
59 | ) |
|
59 | ) | |
60 | ) |
|
60 | ) | |
61 | entry[1].append((b'', b'narrowspec', b'', _(b"read narrowspecs from file"))) |
|
61 | entry[1].append((b'', b'narrowspec', b'', _(b"read narrowspecs from file"))) | |
62 | # TODO(durin42): unify sparse/narrow --include/--exclude logic a bit |
|
62 | # TODO(durin42): unify sparse/narrow --include/--exclude logic a bit | |
63 | if b'sparse' not in extensions.enabled(): |
|
63 | if b'sparse' not in extensions.enabled(): | |
64 | entry[1].append( |
|
64 | entry[1].append( | |
65 | (b'', b'include', [], _(b"specifically fetch this file/directory")) |
|
65 | (b'', b'include', [], _(b"specifically fetch this file/directory")) | |
66 | ) |
|
66 | ) | |
67 | entry[1].append( |
|
67 | entry[1].append( | |
68 | ( |
|
68 | ( | |
69 | b'', |
|
69 | b'', | |
70 | b'exclude', |
|
70 | b'exclude', | |
71 | [], |
|
71 | [], | |
72 | _(b"do not fetch this file/directory, even if included"), |
|
72 | _(b"do not fetch this file/directory, even if included"), | |
73 | ) |
|
73 | ) | |
74 | ) |
|
74 | ) | |
75 |
|
75 | |||
76 | entry = extensions.wrapcommand(commands.table, b'pull', pullnarrowcmd) |
|
76 | entry = extensions.wrapcommand(commands.table, b'pull', pullnarrowcmd) | |
77 | entry[1].append( |
|
77 | entry[1].append( | |
78 | ( |
|
78 | ( | |
79 | b'', |
|
79 | b'', | |
80 | b'depth', |
|
80 | b'depth', | |
81 | b'', |
|
81 | b'', | |
82 | _(b"limit the history fetched by distance from heads"), |
|
82 | _(b"limit the history fetched by distance from heads"), | |
83 | ) |
|
83 | ) | |
84 | ) |
|
84 | ) | |
85 |
|
85 | |||
86 | extensions.wrapcommand(commands.table, b'archive', archivenarrowcmd) |
|
86 | extensions.wrapcommand(commands.table, b'archive', archivenarrowcmd) | |
87 |
|
87 | |||
88 |
|
88 | |||
89 | def clonenarrowcmd(orig, ui, repo, *args, **opts): |
|
89 | def clonenarrowcmd(orig, ui, repo, *args, **opts): | |
90 | """Wraps clone command, so 'hg clone' first wraps localrepo.clone().""" |
|
90 | """Wraps clone command, so 'hg clone' first wraps localrepo.clone().""" | |
91 | opts = pycompat.byteskwargs(opts) |
|
91 | opts = pycompat.byteskwargs(opts) | |
92 | wrappedextraprepare = util.nullcontextmanager() |
|
92 | wrappedextraprepare = util.nullcontextmanager() | |
93 | narrowspecfile = opts[b'narrowspec'] |
|
93 | narrowspecfile = opts[b'narrowspec'] | |
94 |
|
94 | |||
95 | if narrowspecfile: |
|
95 | if narrowspecfile: | |
96 | filepath = os.path.join(encoding.getcwd(), narrowspecfile) |
|
96 | filepath = os.path.join(encoding.getcwd(), narrowspecfile) | |
97 | ui.status(_(b"reading narrowspec from '%s'\n") % filepath) |
|
97 | ui.status(_(b"reading narrowspec from '%s'\n") % filepath) | |
98 | try: |
|
98 | try: | |
99 | fdata = util.readfile(filepath) |
|
99 | fdata = util.readfile(filepath) | |
100 | except IOError as inst: |
|
100 | except IOError as inst: | |
101 | raise error.Abort( |
|
101 | raise error.Abort( | |
102 | _(b"cannot read narrowspecs from '%s': %s") |
|
102 | _(b"cannot read narrowspecs from '%s': %s") | |
103 | % (filepath, encoding.strtolocal(inst.strerror)) |
|
103 | % (filepath, encoding.strtolocal(inst.strerror)) | |
104 | ) |
|
104 | ) | |
105 |
|
105 | |||
106 | includes, excludes, profiles = sparse.parseconfig(ui, fdata, b'narrow') |
|
106 | includes, excludes, profiles = sparse.parseconfig(ui, fdata, b'narrow') | |
107 | if profiles: |
|
107 | if profiles: | |
108 | raise error.ConfigError( |
|
108 | raise error.ConfigError( | |
109 | _( |
|
109 | _( | |
110 | b"cannot specify other files using '%include' in" |
|
110 | b"cannot specify other files using '%include' in" | |
111 | b" narrowspec" |
|
111 | b" narrowspec" | |
112 | ) |
|
112 | ) | |
113 | ) |
|
113 | ) | |
114 |
|
114 | |||
115 | narrowspec.validatepatterns(includes) |
|
115 | narrowspec.validatepatterns(includes) | |
116 | narrowspec.validatepatterns(excludes) |
|
116 | narrowspec.validatepatterns(excludes) | |
117 |
|
117 | |||
118 | # narrowspec is passed so we should assume that user wants narrow clone |
|
118 | # narrowspec is passed so we should assume that user wants narrow clone | |
119 | opts[b'narrow'] = True |
|
119 | opts[b'narrow'] = True | |
120 | opts[b'include'].extend(includes) |
|
120 | opts[b'include'].extend(includes) | |
121 | opts[b'exclude'].extend(excludes) |
|
121 | opts[b'exclude'].extend(excludes) | |
122 |
|
122 | |||
123 | if opts[b'narrow']: |
|
123 | if opts[b'narrow']: | |
124 |
|
124 | |||
125 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): |
|
125 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): | |
126 | orig(pullop, kwargs) |
|
126 | orig(pullop, kwargs) | |
127 |
|
127 | |||
128 | if opts.get(b'depth'): |
|
128 | if opts.get(b'depth'): | |
129 | kwargs[b'depth'] = opts[b'depth'] |
|
129 | kwargs[b'depth'] = opts[b'depth'] | |
130 |
|
130 | |||
131 | wrappedextraprepare = extensions.wrappedfunction( |
|
131 | wrappedextraprepare = extensions.wrappedfunction( | |
132 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen |
|
132 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen | |
133 | ) |
|
133 | ) | |
134 |
|
134 | |||
135 | with wrappedextraprepare: |
|
135 | with wrappedextraprepare: | |
136 | return orig(ui, repo, *args, **pycompat.strkwargs(opts)) |
|
136 | return orig(ui, repo, *args, **pycompat.strkwargs(opts)) | |
137 |
|
137 | |||
138 |
|
138 | |||
139 | def pullnarrowcmd(orig, ui, repo, *args, **opts): |
|
139 | def pullnarrowcmd(orig, ui, repo, *args, **opts): | |
140 | """Wraps pull command to allow modifying narrow spec.""" |
|
140 | """Wraps pull command to allow modifying narrow spec.""" | |
141 | wrappedextraprepare = util.nullcontextmanager() |
|
141 | wrappedextraprepare = util.nullcontextmanager() | |
142 | if requirements.NARROW_REQUIREMENT in repo.requirements: |
|
142 | if requirements.NARROW_REQUIREMENT in repo.requirements: | |
143 |
|
143 | |||
144 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): |
|
144 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): | |
145 | orig(pullop, kwargs) |
|
145 | orig(pullop, kwargs) | |
146 | if opts.get('depth'): |
|
146 | if opts.get('depth'): | |
147 | kwargs[b'depth'] = opts['depth'] |
|
147 | kwargs[b'depth'] = opts['depth'] | |
148 |
|
148 | |||
149 | wrappedextraprepare = extensions.wrappedfunction( |
|
149 | wrappedextraprepare = extensions.wrappedfunction( | |
150 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen |
|
150 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen | |
151 | ) |
|
151 | ) | |
152 |
|
152 | |||
153 | with wrappedextraprepare: |
|
153 | with wrappedextraprepare: | |
154 | return orig(ui, repo, *args, **opts) |
|
154 | return orig(ui, repo, *args, **opts) | |
155 |
|
155 | |||
156 |
|
156 | |||
157 | def archivenarrowcmd(orig, ui, repo, *args, **opts): |
|
157 | def archivenarrowcmd(orig, ui, repo, *args, **opts): | |
158 | """Wraps archive command to narrow the default includes.""" |
|
158 | """Wraps archive command to narrow the default includes.""" | |
159 | if requirements.NARROW_REQUIREMENT in repo.requirements: |
|
159 | if requirements.NARROW_REQUIREMENT in repo.requirements: | |
160 | repo_includes, repo_excludes = repo.narrowpats |
|
160 | repo_includes, repo_excludes = repo.narrowpats | |
161 | includes = set(opts.get('include', [])) |
|
161 | includes = set(opts.get('include', [])) | |
162 | excludes = set(opts.get('exclude', [])) |
|
162 | excludes = set(opts.get('exclude', [])) | |
163 | includes, excludes, unused_invalid = narrowspec.restrictpatterns( |
|
163 | includes, excludes, unused_invalid = narrowspec.restrictpatterns( | |
164 | includes, excludes, repo_includes, repo_excludes |
|
164 | includes, excludes, repo_includes, repo_excludes | |
165 | ) |
|
165 | ) | |
166 | if includes: |
|
166 | if includes: | |
167 | opts['include'] = includes |
|
167 | opts['include'] = includes | |
168 | if excludes: |
|
168 | if excludes: | |
169 | opts['exclude'] = excludes |
|
169 | opts['exclude'] = excludes | |
170 | return orig(ui, repo, *args, **opts) |
|
170 | return orig(ui, repo, *args, **opts) | |
171 |
|
171 | |||
172 |
|
172 | |||
173 | def pullbundle2extraprepare(orig, pullop, kwargs): |
|
173 | def pullbundle2extraprepare(orig, pullop, kwargs): | |
174 | repo = pullop.repo |
|
174 | repo = pullop.repo | |
175 | if requirements.NARROW_REQUIREMENT not in repo.requirements: |
|
175 | if requirements.NARROW_REQUIREMENT not in repo.requirements: | |
176 | return orig(pullop, kwargs) |
|
176 | return orig(pullop, kwargs) | |
177 |
|
177 | |||
178 | if wireprototypes.NARROWCAP not in pullop.remote.capabilities(): |
|
178 | if wireprototypes.NARROWCAP not in pullop.remote.capabilities(): | |
179 | raise error.Abort(_(b"server does not support narrow clones")) |
|
179 | raise error.Abort(_(b"server does not support narrow clones")) | |
180 | orig(pullop, kwargs) |
|
180 | orig(pullop, kwargs) | |
181 | kwargs[b'narrow'] = True |
|
181 | kwargs[b'narrow'] = True | |
182 | include, exclude = repo.narrowpats |
|
182 | include, exclude = repo.narrowpats | |
183 | kwargs[b'oldincludepats'] = include |
|
183 | kwargs[b'oldincludepats'] = include | |
184 | kwargs[b'oldexcludepats'] = exclude |
|
184 | kwargs[b'oldexcludepats'] = exclude | |
185 | if include: |
|
185 | if include: | |
186 | kwargs[b'includepats'] = include |
|
186 | kwargs[b'includepats'] = include | |
187 | if exclude: |
|
187 | if exclude: | |
188 | kwargs[b'excludepats'] = exclude |
|
188 | kwargs[b'excludepats'] = exclude | |
189 | # calculate known nodes only in ellipses cases because in non-ellipses cases |
|
189 | # calculate known nodes only in ellipses cases because in non-ellipses cases | |
190 | # we have all the nodes |
|
190 | # we have all the nodes | |
191 | if wireprototypes.ELLIPSESCAP1 in pullop.remote.capabilities(): |
|
191 | if wireprototypes.ELLIPSESCAP1 in pullop.remote.capabilities(): | |
192 | kwargs[b'known'] = [ |
|
192 | kwargs[b'known'] = [ | |
193 | hex(ctx.node()) |
|
193 | hex(ctx.node()) | |
194 | for ctx in repo.set(b'::%ln', pullop.common) |
|
194 | for ctx in repo.set(b'::%ln', pullop.common) | |
195 | if ctx.node() != repo.nullid |
|
195 | if ctx.node() != repo.nullid | |
196 | ] |
|
196 | ] | |
197 | if not kwargs[b'known']: |
|
197 | if not kwargs[b'known']: | |
198 | # Mercurial serializes an empty list as '' and deserializes it as |
|
198 | # Mercurial serializes an empty list as '' and deserializes it as | |
199 | # [''], so delete it instead to avoid handling the empty string on |
|
199 | # [''], so delete it instead to avoid handling the empty string on | |
200 | # the server. |
|
200 | # the server. | |
201 | del kwargs[b'known'] |
|
201 | del kwargs[b'known'] | |
202 |
|
202 | |||
203 |
|
203 | |||
204 | extensions.wrapfunction( |
|
204 | extensions.wrapfunction( | |
205 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare |
|
205 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare | |
206 | ) |
|
206 | ) | |
207 |
|
207 | |||
208 |
|
208 | |||
209 | def _narrow( |
|
209 | def _narrow( | |
210 | ui, |
|
210 | ui, | |
211 | repo, |
|
211 | repo, | |
212 | remote, |
|
212 | remote, | |
213 | commoninc, |
|
213 | commoninc, | |
214 | oldincludes, |
|
214 | oldincludes, | |
215 | oldexcludes, |
|
215 | oldexcludes, | |
216 | newincludes, |
|
216 | newincludes, | |
217 | newexcludes, |
|
217 | newexcludes, | |
218 | force, |
|
218 | force, | |
219 | backup, |
|
219 | backup, | |
220 | ): |
|
220 | ): | |
221 | oldmatch = narrowspec.match(repo.root, oldincludes, oldexcludes) |
|
221 | oldmatch = narrowspec.match(repo.root, oldincludes, oldexcludes) | |
222 | newmatch = narrowspec.match(repo.root, newincludes, newexcludes) |
|
222 | newmatch = narrowspec.match(repo.root, newincludes, newexcludes) | |
223 |
|
223 | |||
224 | # This is essentially doing "hg outgoing" to find all local-only |
|
224 | # This is essentially doing "hg outgoing" to find all local-only | |
225 | # commits. We will then check that the local-only commits don't |
|
225 | # commits. We will then check that the local-only commits don't | |
226 | # have any changes to files that will be untracked. |
|
226 | # have any changes to files that will be untracked. | |
227 | unfi = repo.unfiltered() |
|
227 | unfi = repo.unfiltered() | |
228 | outgoing = discovery.findcommonoutgoing(unfi, remote, commoninc=commoninc) |
|
228 | outgoing = discovery.findcommonoutgoing(unfi, remote, commoninc=commoninc) | |
229 | ui.status(_(b'looking for local changes to affected paths\n')) |
|
229 | ui.status(_(b'looking for local changes to affected paths\n')) | |
230 | progress = ui.makeprogress( |
|
230 | progress = ui.makeprogress( | |
231 | topic=_(b'changesets'), |
|
231 | topic=_(b'changesets'), | |
232 | unit=_(b'changesets'), |
|
232 | unit=_(b'changesets'), | |
233 | total=len(outgoing.missing) + len(outgoing.excluded), |
|
233 | total=len(outgoing.missing) + len(outgoing.excluded), | |
234 | ) |
|
234 | ) | |
235 | localnodes = [] |
|
235 | localnodes = [] | |
236 | with progress: |
|
236 | with progress: | |
237 | for n in itertools.chain(outgoing.missing, outgoing.excluded): |
|
237 | for n in itertools.chain(outgoing.missing, outgoing.excluded): | |
238 | progress.increment() |
|
238 | progress.increment() | |
239 | if any(oldmatch(f) and not newmatch(f) for f in unfi[n].files()): |
|
239 | if any(oldmatch(f) and not newmatch(f) for f in unfi[n].files()): | |
240 | localnodes.append(n) |
|
240 | localnodes.append(n) | |
241 | revstostrip = unfi.revs(b'descendants(%ln)', localnodes) |
|
241 | revstostrip = unfi.revs(b'descendants(%ln)', localnodes) | |
242 | hiddenrevs = repoview.filterrevs(repo, b'visible') |
|
242 | hiddenrevs = repoview.filterrevs(repo, b'visible') | |
243 | visibletostrip = list( |
|
243 | visibletostrip = list( | |
244 | repo.changelog.node(r) for r in (revstostrip - hiddenrevs) |
|
244 | repo.changelog.node(r) for r in (revstostrip - hiddenrevs) | |
245 | ) |
|
245 | ) | |
246 | if visibletostrip: |
|
246 | if visibletostrip: | |
247 | ui.status( |
|
247 | ui.status( | |
248 | _( |
|
248 | _( | |
249 | b'The following changeset(s) or their ancestors have ' |
|
249 | b'The following changeset(s) or their ancestors have ' | |
250 | b'local changes not on the remote:\n' |
|
250 | b'local changes not on the remote:\n' | |
251 | ) |
|
251 | ) | |
252 | ) |
|
252 | ) | |
253 | maxnodes = 10 |
|
253 | maxnodes = 10 | |
254 | if ui.verbose or len(visibletostrip) <= maxnodes: |
|
254 | if ui.verbose or len(visibletostrip) <= maxnodes: | |
255 | for n in visibletostrip: |
|
255 | for n in visibletostrip: | |
256 | ui.status(b'%s\n' % short(n)) |
|
256 | ui.status(b'%s\n' % short(n)) | |
257 | else: |
|
257 | else: | |
258 | for n in visibletostrip[:maxnodes]: |
|
258 | for n in visibletostrip[:maxnodes]: | |
259 | ui.status(b'%s\n' % short(n)) |
|
259 | ui.status(b'%s\n' % short(n)) | |
260 | ui.status( |
|
260 | ui.status( | |
261 | _(b'...and %d more, use --verbose to list all\n') |
|
261 | _(b'...and %d more, use --verbose to list all\n') | |
262 | % (len(visibletostrip) - maxnodes) |
|
262 | % (len(visibletostrip) - maxnodes) | |
263 | ) |
|
263 | ) | |
264 | if not force: |
|
264 | if not force: | |
265 | raise error.StateError( |
|
265 | raise error.StateError( | |
266 | _(b'local changes found'), |
|
266 | _(b'local changes found'), | |
267 | hint=_(b'use --force-delete-local-changes to ignore'), |
|
267 | hint=_(b'use --force-delete-local-changes to ignore'), | |
268 | ) |
|
268 | ) | |
269 |
|
269 | |||
270 | with ui.uninterruptible(): |
|
270 | with ui.uninterruptible(): | |
271 | if revstostrip: |
|
271 | if revstostrip: | |
272 | tostrip = [unfi.changelog.node(r) for r in revstostrip] |
|
272 | tostrip = [unfi.changelog.node(r) for r in revstostrip] | |
273 | if repo[b'.'].node() in tostrip: |
|
273 | if repo[b'.'].node() in tostrip: | |
274 | # stripping working copy, so move to a different commit first |
|
274 | # stripping working copy, so move to a different commit first | |
275 | urev = max( |
|
275 | urev = max( | |
276 | repo.revs( |
|
276 | repo.revs( | |
277 | b'(::%n) - %ln + null', |
|
277 | b'(::%n) - %ln + null', | |
278 | repo[b'.'].node(), |
|
278 | repo[b'.'].node(), | |
279 | visibletostrip, |
|
279 | visibletostrip, | |
280 | ) |
|
280 | ) | |
281 | ) |
|
281 | ) | |
282 | hg.clean(repo, urev) |
|
282 | hg.clean(repo, urev) | |
283 | overrides = {(b'devel', b'strip-obsmarkers'): False} |
|
283 | overrides = {(b'devel', b'strip-obsmarkers'): False} | |
284 | if backup: |
|
284 | if backup: | |
285 | ui.status(_(b'moving unwanted changesets to backup\n')) |
|
285 | ui.status(_(b'moving unwanted changesets to backup\n')) | |
286 | else: |
|
286 | else: | |
287 | ui.status(_(b'deleting unwanted changesets\n')) |
|
287 | ui.status(_(b'deleting unwanted changesets\n')) | |
288 | with ui.configoverride(overrides, b'narrow'): |
|
288 | with ui.configoverride(overrides, b'narrow'): | |
289 | repair.strip(ui, unfi, tostrip, topic=b'narrow', backup=backup) |
|
289 | repair.strip(ui, unfi, tostrip, topic=b'narrow', backup=backup) | |
290 |
|
290 | |||
291 | todelete = [] |
|
291 | todelete = [] | |
292 |
for t, f |
|
292 | for t, f, size in repo.store.datafiles(): | |
293 | if f.startswith(b'data/'): |
|
293 | if f.startswith(b'data/'): | |
294 | file = f[5:-2] |
|
294 | file = f[5:-2] | |
295 | if not newmatch(file): |
|
295 | if not newmatch(file): | |
296 | todelete.append(f) |
|
296 | todelete.append(f) | |
297 | elif f.startswith(b'meta/'): |
|
297 | elif f.startswith(b'meta/'): | |
298 | dir = f[5:-13] |
|
298 | dir = f[5:-13] | |
299 | dirs = sorted(pathutil.dirs({dir})) + [dir] |
|
299 | dirs = sorted(pathutil.dirs({dir})) + [dir] | |
300 | include = True |
|
300 | include = True | |
301 | for d in dirs: |
|
301 | for d in dirs: | |
302 | visit = newmatch.visitdir(d) |
|
302 | visit = newmatch.visitdir(d) | |
303 | if not visit: |
|
303 | if not visit: | |
304 | include = False |
|
304 | include = False | |
305 | break |
|
305 | break | |
306 | if visit == b'all': |
|
306 | if visit == b'all': | |
307 | break |
|
307 | break | |
308 | if not include: |
|
308 | if not include: | |
309 | todelete.append(f) |
|
309 | todelete.append(f) | |
310 |
|
310 | |||
311 | repo.destroying() |
|
311 | repo.destroying() | |
312 |
|
312 | |||
313 | with repo.transaction(b'narrowing'): |
|
313 | with repo.transaction(b'narrowing'): | |
314 | # Update narrowspec before removing revlogs, so repo won't be |
|
314 | # Update narrowspec before removing revlogs, so repo won't be | |
315 | # corrupt in case of crash |
|
315 | # corrupt in case of crash | |
316 | repo.setnarrowpats(newincludes, newexcludes) |
|
316 | repo.setnarrowpats(newincludes, newexcludes) | |
317 |
|
317 | |||
318 | for f in todelete: |
|
318 | for f in todelete: | |
319 | ui.status(_(b'deleting %s\n') % f) |
|
319 | ui.status(_(b'deleting %s\n') % f) | |
320 | util.unlinkpath(repo.svfs.join(f)) |
|
320 | util.unlinkpath(repo.svfs.join(f)) | |
321 | repo.store.markremoved(f) |
|
321 | repo.store.markremoved(f) | |
322 |
|
322 | |||
323 | ui.status(_(b'deleting unwanted files from working copy\n')) |
|
323 | ui.status(_(b'deleting unwanted files from working copy\n')) | |
324 | with repo.dirstate.parentchange(): |
|
324 | with repo.dirstate.parentchange(): | |
325 | narrowspec.updateworkingcopy(repo, assumeclean=True) |
|
325 | narrowspec.updateworkingcopy(repo, assumeclean=True) | |
326 | narrowspec.copytoworkingcopy(repo) |
|
326 | narrowspec.copytoworkingcopy(repo) | |
327 |
|
327 | |||
328 | repo.destroyed() |
|
328 | repo.destroyed() | |
329 |
|
329 | |||
330 |
|
330 | |||
331 | def _widen( |
|
331 | def _widen( | |
332 | ui, |
|
332 | ui, | |
333 | repo, |
|
333 | repo, | |
334 | remote, |
|
334 | remote, | |
335 | commoninc, |
|
335 | commoninc, | |
336 | oldincludes, |
|
336 | oldincludes, | |
337 | oldexcludes, |
|
337 | oldexcludes, | |
338 | newincludes, |
|
338 | newincludes, | |
339 | newexcludes, |
|
339 | newexcludes, | |
340 | ): |
|
340 | ): | |
341 | # for now we assume that if a server has ellipses enabled, we will be |
|
341 | # for now we assume that if a server has ellipses enabled, we will be | |
342 | # exchanging ellipses nodes. In future we should add ellipses as a client |
|
342 | # exchanging ellipses nodes. In future we should add ellipses as a client | |
343 | # side requirement (maybe) to distinguish a client is shallow or not and |
|
343 | # side requirement (maybe) to distinguish a client is shallow or not and | |
344 | # then send that information to server whether we want ellipses or not. |
|
344 | # then send that information to server whether we want ellipses or not. | |
345 | # Theoretically a non-ellipses repo should be able to use narrow |
|
345 | # Theoretically a non-ellipses repo should be able to use narrow | |
346 | # functionality from an ellipses enabled server |
|
346 | # functionality from an ellipses enabled server | |
347 | remotecap = remote.capabilities() |
|
347 | remotecap = remote.capabilities() | |
348 | ellipsesremote = any( |
|
348 | ellipsesremote = any( | |
349 | cap in remotecap for cap in wireprototypes.SUPPORTED_ELLIPSESCAP |
|
349 | cap in remotecap for cap in wireprototypes.SUPPORTED_ELLIPSESCAP | |
350 | ) |
|
350 | ) | |
351 |
|
351 | |||
352 | # check whether we are talking to a server which supports old version of |
|
352 | # check whether we are talking to a server which supports old version of | |
353 | # ellipses capabilities |
|
353 | # ellipses capabilities | |
354 | isoldellipses = ( |
|
354 | isoldellipses = ( | |
355 | ellipsesremote |
|
355 | ellipsesremote | |
356 | and wireprototypes.ELLIPSESCAP1 in remotecap |
|
356 | and wireprototypes.ELLIPSESCAP1 in remotecap | |
357 | and wireprototypes.ELLIPSESCAP not in remotecap |
|
357 | and wireprototypes.ELLIPSESCAP not in remotecap | |
358 | ) |
|
358 | ) | |
359 |
|
359 | |||
360 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): |
|
360 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): | |
361 | orig(pullop, kwargs) |
|
361 | orig(pullop, kwargs) | |
362 | # The old{in,ex}cludepats have already been set by orig() |
|
362 | # The old{in,ex}cludepats have already been set by orig() | |
363 | kwargs[b'includepats'] = newincludes |
|
363 | kwargs[b'includepats'] = newincludes | |
364 | kwargs[b'excludepats'] = newexcludes |
|
364 | kwargs[b'excludepats'] = newexcludes | |
365 |
|
365 | |||
366 | wrappedextraprepare = extensions.wrappedfunction( |
|
366 | wrappedextraprepare = extensions.wrappedfunction( | |
367 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen |
|
367 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen | |
368 | ) |
|
368 | ) | |
369 |
|
369 | |||
370 | # define a function that narrowbundle2 can call after creating the |
|
370 | # define a function that narrowbundle2 can call after creating the | |
371 | # backup bundle, but before applying the bundle from the server |
|
371 | # backup bundle, but before applying the bundle from the server | |
372 | def setnewnarrowpats(): |
|
372 | def setnewnarrowpats(): | |
373 | repo.setnarrowpats(newincludes, newexcludes) |
|
373 | repo.setnarrowpats(newincludes, newexcludes) | |
374 |
|
374 | |||
375 | repo.setnewnarrowpats = setnewnarrowpats |
|
375 | repo.setnewnarrowpats = setnewnarrowpats | |
376 | # silence the devel-warning of applying an empty changegroup |
|
376 | # silence the devel-warning of applying an empty changegroup | |
377 | overrides = {(b'devel', b'all-warnings'): False} |
|
377 | overrides = {(b'devel', b'all-warnings'): False} | |
378 |
|
378 | |||
379 | common = commoninc[0] |
|
379 | common = commoninc[0] | |
380 | with ui.uninterruptible(): |
|
380 | with ui.uninterruptible(): | |
381 | if ellipsesremote: |
|
381 | if ellipsesremote: | |
382 | ds = repo.dirstate |
|
382 | ds = repo.dirstate | |
383 | p1, p2 = ds.p1(), ds.p2() |
|
383 | p1, p2 = ds.p1(), ds.p2() | |
384 | with ds.parentchange(): |
|
384 | with ds.parentchange(): | |
385 | ds.setparents(repo.nullid, repo.nullid) |
|
385 | ds.setparents(repo.nullid, repo.nullid) | |
386 | if isoldellipses: |
|
386 | if isoldellipses: | |
387 | with wrappedextraprepare: |
|
387 | with wrappedextraprepare: | |
388 | exchange.pull(repo, remote, heads=common) |
|
388 | exchange.pull(repo, remote, heads=common) | |
389 | else: |
|
389 | else: | |
390 | known = [] |
|
390 | known = [] | |
391 | if ellipsesremote: |
|
391 | if ellipsesremote: | |
392 | known = [ |
|
392 | known = [ | |
393 | ctx.node() |
|
393 | ctx.node() | |
394 | for ctx in repo.set(b'::%ln', common) |
|
394 | for ctx in repo.set(b'::%ln', common) | |
395 | if ctx.node() != repo.nullid |
|
395 | if ctx.node() != repo.nullid | |
396 | ] |
|
396 | ] | |
397 | with remote.commandexecutor() as e: |
|
397 | with remote.commandexecutor() as e: | |
398 | bundle = e.callcommand( |
|
398 | bundle = e.callcommand( | |
399 | b'narrow_widen', |
|
399 | b'narrow_widen', | |
400 | { |
|
400 | { | |
401 | b'oldincludes': oldincludes, |
|
401 | b'oldincludes': oldincludes, | |
402 | b'oldexcludes': oldexcludes, |
|
402 | b'oldexcludes': oldexcludes, | |
403 | b'newincludes': newincludes, |
|
403 | b'newincludes': newincludes, | |
404 | b'newexcludes': newexcludes, |
|
404 | b'newexcludes': newexcludes, | |
405 | b'cgversion': b'03', |
|
405 | b'cgversion': b'03', | |
406 | b'commonheads': common, |
|
406 | b'commonheads': common, | |
407 | b'known': known, |
|
407 | b'known': known, | |
408 | b'ellipses': ellipsesremote, |
|
408 | b'ellipses': ellipsesremote, | |
409 | }, |
|
409 | }, | |
410 | ).result() |
|
410 | ).result() | |
411 |
|
411 | |||
412 | trmanager = exchange.transactionmanager( |
|
412 | trmanager = exchange.transactionmanager( | |
413 | repo, b'widen', remote.url() |
|
413 | repo, b'widen', remote.url() | |
414 | ) |
|
414 | ) | |
415 | with trmanager, repo.ui.configoverride(overrides, b'widen'): |
|
415 | with trmanager, repo.ui.configoverride(overrides, b'widen'): | |
416 | op = bundle2.bundleoperation( |
|
416 | op = bundle2.bundleoperation( | |
417 | repo, trmanager.transaction, source=b'widen' |
|
417 | repo, trmanager.transaction, source=b'widen' | |
418 | ) |
|
418 | ) | |
419 | # TODO: we should catch error.Abort here |
|
419 | # TODO: we should catch error.Abort here | |
420 | bundle2.processbundle(repo, bundle, op=op) |
|
420 | bundle2.processbundle(repo, bundle, op=op) | |
421 |
|
421 | |||
422 | if ellipsesremote: |
|
422 | if ellipsesremote: | |
423 | with ds.parentchange(): |
|
423 | with ds.parentchange(): | |
424 | ds.setparents(p1, p2) |
|
424 | ds.setparents(p1, p2) | |
425 |
|
425 | |||
426 | with repo.transaction(b'widening'), repo.dirstate.parentchange(): |
|
426 | with repo.transaction(b'widening'), repo.dirstate.parentchange(): | |
427 | repo.setnewnarrowpats() |
|
427 | repo.setnewnarrowpats() | |
428 | narrowspec.updateworkingcopy(repo) |
|
428 | narrowspec.updateworkingcopy(repo) | |
429 | narrowspec.copytoworkingcopy(repo) |
|
429 | narrowspec.copytoworkingcopy(repo) | |
430 |
|
430 | |||
431 |
|
431 | |||
432 | # TODO(rdamazio): Make new matcher format and update description |
|
432 | # TODO(rdamazio): Make new matcher format and update description | |
433 | @command( |
|
433 | @command( | |
434 | b'tracked', |
|
434 | b'tracked', | |
435 | [ |
|
435 | [ | |
436 | (b'', b'addinclude', [], _(b'new paths to include')), |
|
436 | (b'', b'addinclude', [], _(b'new paths to include')), | |
437 | (b'', b'removeinclude', [], _(b'old paths to no longer include')), |
|
437 | (b'', b'removeinclude', [], _(b'old paths to no longer include')), | |
438 | ( |
|
438 | ( | |
439 | b'', |
|
439 | b'', | |
440 | b'auto-remove-includes', |
|
440 | b'auto-remove-includes', | |
441 | False, |
|
441 | False, | |
442 | _(b'automatically choose unused includes to remove'), |
|
442 | _(b'automatically choose unused includes to remove'), | |
443 | ), |
|
443 | ), | |
444 | (b'', b'addexclude', [], _(b'new paths to exclude')), |
|
444 | (b'', b'addexclude', [], _(b'new paths to exclude')), | |
445 | (b'', b'import-rules', b'', _(b'import narrowspecs from a file')), |
|
445 | (b'', b'import-rules', b'', _(b'import narrowspecs from a file')), | |
446 | (b'', b'removeexclude', [], _(b'old paths to no longer exclude')), |
|
446 | (b'', b'removeexclude', [], _(b'old paths to no longer exclude')), | |
447 | ( |
|
447 | ( | |
448 | b'', |
|
448 | b'', | |
449 | b'clear', |
|
449 | b'clear', | |
450 | False, |
|
450 | False, | |
451 | _(b'whether to replace the existing narrowspec'), |
|
451 | _(b'whether to replace the existing narrowspec'), | |
452 | ), |
|
452 | ), | |
453 | ( |
|
453 | ( | |
454 | b'', |
|
454 | b'', | |
455 | b'force-delete-local-changes', |
|
455 | b'force-delete-local-changes', | |
456 | False, |
|
456 | False, | |
457 | _(b'forces deletion of local changes when narrowing'), |
|
457 | _(b'forces deletion of local changes when narrowing'), | |
458 | ), |
|
458 | ), | |
459 | ( |
|
459 | ( | |
460 | b'', |
|
460 | b'', | |
461 | b'backup', |
|
461 | b'backup', | |
462 | True, |
|
462 | True, | |
463 | _(b'back up local changes when narrowing'), |
|
463 | _(b'back up local changes when narrowing'), | |
464 | ), |
|
464 | ), | |
465 | ( |
|
465 | ( | |
466 | b'', |
|
466 | b'', | |
467 | b'update-working-copy', |
|
467 | b'update-working-copy', | |
468 | False, |
|
468 | False, | |
469 | _(b'update working copy when the store has changed'), |
|
469 | _(b'update working copy when the store has changed'), | |
470 | ), |
|
470 | ), | |
471 | ] |
|
471 | ] | |
472 | + commands.remoteopts, |
|
472 | + commands.remoteopts, | |
473 | _(b'[OPTIONS]... [REMOTE]'), |
|
473 | _(b'[OPTIONS]... [REMOTE]'), | |
474 | inferrepo=True, |
|
474 | inferrepo=True, | |
475 | helpcategory=command.CATEGORY_MAINTENANCE, |
|
475 | helpcategory=command.CATEGORY_MAINTENANCE, | |
476 | ) |
|
476 | ) | |
477 | def trackedcmd(ui, repo, remotepath=None, *pats, **opts): |
|
477 | def trackedcmd(ui, repo, remotepath=None, *pats, **opts): | |
478 | """show or change the current narrowspec |
|
478 | """show or change the current narrowspec | |
479 |
|
479 | |||
480 | With no argument, shows the current narrowspec entries, one per line. Each |
|
480 | With no argument, shows the current narrowspec entries, one per line. Each | |
481 | line will be prefixed with 'I' or 'X' for included or excluded patterns, |
|
481 | line will be prefixed with 'I' or 'X' for included or excluded patterns, | |
482 | respectively. |
|
482 | respectively. | |
483 |
|
483 | |||
484 | The narrowspec is comprised of expressions to match remote files and/or |
|
484 | The narrowspec is comprised of expressions to match remote files and/or | |
485 | directories that should be pulled into your client. |
|
485 | directories that should be pulled into your client. | |
486 | The narrowspec has *include* and *exclude* expressions, with excludes always |
|
486 | The narrowspec has *include* and *exclude* expressions, with excludes always | |
487 | trumping includes: that is, if a file matches an exclude expression, it will |
|
487 | trumping includes: that is, if a file matches an exclude expression, it will | |
488 | be excluded even if it also matches an include expression. |
|
488 | be excluded even if it also matches an include expression. | |
489 | Excluding files that were never included has no effect. |
|
489 | Excluding files that were never included has no effect. | |
490 |
|
490 | |||
491 | Each included or excluded entry is in the format described by |
|
491 | Each included or excluded entry is in the format described by | |
492 | 'hg help patterns'. |
|
492 | 'hg help patterns'. | |
493 |
|
493 | |||
494 | The options allow you to add or remove included and excluded expressions. |
|
494 | The options allow you to add or remove included and excluded expressions. | |
495 |
|
495 | |||
496 | If --clear is specified, then all previous includes and excludes are DROPPED |
|
496 | If --clear is specified, then all previous includes and excludes are DROPPED | |
497 | and replaced by the new ones specified to --addinclude and --addexclude. |
|
497 | and replaced by the new ones specified to --addinclude and --addexclude. | |
498 | If --clear is specified without any further options, the narrowspec will be |
|
498 | If --clear is specified without any further options, the narrowspec will be | |
499 | empty and will not match any files. |
|
499 | empty and will not match any files. | |
500 |
|
500 | |||
501 | If --auto-remove-includes is specified, then those includes that don't match |
|
501 | If --auto-remove-includes is specified, then those includes that don't match | |
502 | any files modified by currently visible local commits (those not shared by |
|
502 | any files modified by currently visible local commits (those not shared by | |
503 | the remote) will be added to the set of explicitly specified includes to |
|
503 | the remote) will be added to the set of explicitly specified includes to | |
504 | remove. |
|
504 | remove. | |
505 |
|
505 | |||
506 | --import-rules accepts a path to a file containing rules, allowing you to |
|
506 | --import-rules accepts a path to a file containing rules, allowing you to | |
507 | add --addinclude, --addexclude rules in bulk. Like the other include and |
|
507 | add --addinclude, --addexclude rules in bulk. Like the other include and | |
508 | exclude switches, the changes are applied immediately. |
|
508 | exclude switches, the changes are applied immediately. | |
509 | """ |
|
509 | """ | |
510 | opts = pycompat.byteskwargs(opts) |
|
510 | opts = pycompat.byteskwargs(opts) | |
511 | if requirements.NARROW_REQUIREMENT not in repo.requirements: |
|
511 | if requirements.NARROW_REQUIREMENT not in repo.requirements: | |
512 | raise error.InputError( |
|
512 | raise error.InputError( | |
513 | _( |
|
513 | _( | |
514 | b'the tracked command is only supported on ' |
|
514 | b'the tracked command is only supported on ' | |
515 | b'repositories cloned with --narrow' |
|
515 | b'repositories cloned with --narrow' | |
516 | ) |
|
516 | ) | |
517 | ) |
|
517 | ) | |
518 |
|
518 | |||
519 | # Before supporting, decide whether it "hg tracked --clear" should mean |
|
519 | # Before supporting, decide whether it "hg tracked --clear" should mean | |
520 | # tracking no paths or all paths. |
|
520 | # tracking no paths or all paths. | |
521 | if opts[b'clear']: |
|
521 | if opts[b'clear']: | |
522 | raise error.InputError(_(b'the --clear option is not yet supported')) |
|
522 | raise error.InputError(_(b'the --clear option is not yet supported')) | |
523 |
|
523 | |||
524 | # import rules from a file |
|
524 | # import rules from a file | |
525 | newrules = opts.get(b'import_rules') |
|
525 | newrules = opts.get(b'import_rules') | |
526 | if newrules: |
|
526 | if newrules: | |
527 | try: |
|
527 | try: | |
528 | filepath = os.path.join(encoding.getcwd(), newrules) |
|
528 | filepath = os.path.join(encoding.getcwd(), newrules) | |
529 | fdata = util.readfile(filepath) |
|
529 | fdata = util.readfile(filepath) | |
530 | except IOError as inst: |
|
530 | except IOError as inst: | |
531 | raise error.StorageError( |
|
531 | raise error.StorageError( | |
532 | _(b"cannot read narrowspecs from '%s': %s") |
|
532 | _(b"cannot read narrowspecs from '%s': %s") | |
533 | % (filepath, encoding.strtolocal(inst.strerror)) |
|
533 | % (filepath, encoding.strtolocal(inst.strerror)) | |
534 | ) |
|
534 | ) | |
535 | includepats, excludepats, profiles = sparse.parseconfig( |
|
535 | includepats, excludepats, profiles = sparse.parseconfig( | |
536 | ui, fdata, b'narrow' |
|
536 | ui, fdata, b'narrow' | |
537 | ) |
|
537 | ) | |
538 | if profiles: |
|
538 | if profiles: | |
539 | raise error.InputError( |
|
539 | raise error.InputError( | |
540 | _( |
|
540 | _( | |
541 | b"including other spec files using '%include' " |
|
541 | b"including other spec files using '%include' " | |
542 | b"is not supported in narrowspec" |
|
542 | b"is not supported in narrowspec" | |
543 | ) |
|
543 | ) | |
544 | ) |
|
544 | ) | |
545 | opts[b'addinclude'].extend(includepats) |
|
545 | opts[b'addinclude'].extend(includepats) | |
546 | opts[b'addexclude'].extend(excludepats) |
|
546 | opts[b'addexclude'].extend(excludepats) | |
547 |
|
547 | |||
548 | addedincludes = narrowspec.parsepatterns(opts[b'addinclude']) |
|
548 | addedincludes = narrowspec.parsepatterns(opts[b'addinclude']) | |
549 | removedincludes = narrowspec.parsepatterns(opts[b'removeinclude']) |
|
549 | removedincludes = narrowspec.parsepatterns(opts[b'removeinclude']) | |
550 | addedexcludes = narrowspec.parsepatterns(opts[b'addexclude']) |
|
550 | addedexcludes = narrowspec.parsepatterns(opts[b'addexclude']) | |
551 | removedexcludes = narrowspec.parsepatterns(opts[b'removeexclude']) |
|
551 | removedexcludes = narrowspec.parsepatterns(opts[b'removeexclude']) | |
552 | autoremoveincludes = opts[b'auto_remove_includes'] |
|
552 | autoremoveincludes = opts[b'auto_remove_includes'] | |
553 |
|
553 | |||
554 | update_working_copy = opts[b'update_working_copy'] |
|
554 | update_working_copy = opts[b'update_working_copy'] | |
555 | only_show = not ( |
|
555 | only_show = not ( | |
556 | addedincludes |
|
556 | addedincludes | |
557 | or removedincludes |
|
557 | or removedincludes | |
558 | or addedexcludes |
|
558 | or addedexcludes | |
559 | or removedexcludes |
|
559 | or removedexcludes | |
560 | or newrules |
|
560 | or newrules | |
561 | or autoremoveincludes |
|
561 | or autoremoveincludes | |
562 | or update_working_copy |
|
562 | or update_working_copy | |
563 | ) |
|
563 | ) | |
564 |
|
564 | |||
565 | oldincludes, oldexcludes = repo.narrowpats |
|
565 | oldincludes, oldexcludes = repo.narrowpats | |
566 |
|
566 | |||
567 | # filter the user passed additions and deletions into actual additions and |
|
567 | # filter the user passed additions and deletions into actual additions and | |
568 | # deletions of excludes and includes |
|
568 | # deletions of excludes and includes | |
569 | addedincludes -= oldincludes |
|
569 | addedincludes -= oldincludes | |
570 | removedincludes &= oldincludes |
|
570 | removedincludes &= oldincludes | |
571 | addedexcludes -= oldexcludes |
|
571 | addedexcludes -= oldexcludes | |
572 | removedexcludes &= oldexcludes |
|
572 | removedexcludes &= oldexcludes | |
573 |
|
573 | |||
574 | widening = addedincludes or removedexcludes |
|
574 | widening = addedincludes or removedexcludes | |
575 | narrowing = removedincludes or addedexcludes |
|
575 | narrowing = removedincludes or addedexcludes | |
576 |
|
576 | |||
577 | # Only print the current narrowspec. |
|
577 | # Only print the current narrowspec. | |
578 | if only_show: |
|
578 | if only_show: | |
579 | ui.pager(b'tracked') |
|
579 | ui.pager(b'tracked') | |
580 | fm = ui.formatter(b'narrow', opts) |
|
580 | fm = ui.formatter(b'narrow', opts) | |
581 | for i in sorted(oldincludes): |
|
581 | for i in sorted(oldincludes): | |
582 | fm.startitem() |
|
582 | fm.startitem() | |
583 | fm.write(b'status', b'%s ', b'I', label=b'narrow.included') |
|
583 | fm.write(b'status', b'%s ', b'I', label=b'narrow.included') | |
584 | fm.write(b'pat', b'%s\n', i, label=b'narrow.included') |
|
584 | fm.write(b'pat', b'%s\n', i, label=b'narrow.included') | |
585 | for i in sorted(oldexcludes): |
|
585 | for i in sorted(oldexcludes): | |
586 | fm.startitem() |
|
586 | fm.startitem() | |
587 | fm.write(b'status', b'%s ', b'X', label=b'narrow.excluded') |
|
587 | fm.write(b'status', b'%s ', b'X', label=b'narrow.excluded') | |
588 | fm.write(b'pat', b'%s\n', i, label=b'narrow.excluded') |
|
588 | fm.write(b'pat', b'%s\n', i, label=b'narrow.excluded') | |
589 | fm.end() |
|
589 | fm.end() | |
590 | return 0 |
|
590 | return 0 | |
591 |
|
591 | |||
592 | if update_working_copy: |
|
592 | if update_working_copy: | |
593 | with repo.wlock(), repo.lock(), repo.transaction( |
|
593 | with repo.wlock(), repo.lock(), repo.transaction( | |
594 | b'narrow-wc' |
|
594 | b'narrow-wc' | |
595 | ), repo.dirstate.parentchange(): |
|
595 | ), repo.dirstate.parentchange(): | |
596 | narrowspec.updateworkingcopy(repo) |
|
596 | narrowspec.updateworkingcopy(repo) | |
597 | narrowspec.copytoworkingcopy(repo) |
|
597 | narrowspec.copytoworkingcopy(repo) | |
598 | return 0 |
|
598 | return 0 | |
599 |
|
599 | |||
600 | if not (widening or narrowing or autoremoveincludes): |
|
600 | if not (widening or narrowing or autoremoveincludes): | |
601 | ui.status(_(b"nothing to widen or narrow\n")) |
|
601 | ui.status(_(b"nothing to widen or narrow\n")) | |
602 | return 0 |
|
602 | return 0 | |
603 |
|
603 | |||
604 | with repo.wlock(), repo.lock(): |
|
604 | with repo.wlock(), repo.lock(): | |
605 | cmdutil.bailifchanged(repo) |
|
605 | cmdutil.bailifchanged(repo) | |
606 |
|
606 | |||
607 | # Find the revisions we have in common with the remote. These will |
|
607 | # Find the revisions we have in common with the remote. These will | |
608 | # be used for finding local-only changes for narrowing. They will |
|
608 | # be used for finding local-only changes for narrowing. They will | |
609 | # also define the set of revisions to update for widening. |
|
609 | # also define the set of revisions to update for widening. | |
610 | r = urlutil.get_unique_pull_path(b'tracked', repo, ui, remotepath) |
|
610 | r = urlutil.get_unique_pull_path(b'tracked', repo, ui, remotepath) | |
611 | url, branches = r |
|
611 | url, branches = r | |
612 | ui.status(_(b'comparing with %s\n') % urlutil.hidepassword(url)) |
|
612 | ui.status(_(b'comparing with %s\n') % urlutil.hidepassword(url)) | |
613 | remote = hg.peer(repo, opts, url) |
|
613 | remote = hg.peer(repo, opts, url) | |
614 |
|
614 | |||
615 | try: |
|
615 | try: | |
616 | # check narrow support before doing anything if widening needs to be |
|
616 | # check narrow support before doing anything if widening needs to be | |
617 | # performed. In future we should also abort if client is ellipses and |
|
617 | # performed. In future we should also abort if client is ellipses and | |
618 | # server does not support ellipses |
|
618 | # server does not support ellipses | |
619 | if ( |
|
619 | if ( | |
620 | widening |
|
620 | widening | |
621 | and wireprototypes.NARROWCAP not in remote.capabilities() |
|
621 | and wireprototypes.NARROWCAP not in remote.capabilities() | |
622 | ): |
|
622 | ): | |
623 | raise error.Abort(_(b"server does not support narrow clones")) |
|
623 | raise error.Abort(_(b"server does not support narrow clones")) | |
624 |
|
624 | |||
625 | commoninc = discovery.findcommonincoming(repo, remote) |
|
625 | commoninc = discovery.findcommonincoming(repo, remote) | |
626 |
|
626 | |||
627 | if autoremoveincludes: |
|
627 | if autoremoveincludes: | |
628 | outgoing = discovery.findcommonoutgoing( |
|
628 | outgoing = discovery.findcommonoutgoing( | |
629 | repo, remote, commoninc=commoninc |
|
629 | repo, remote, commoninc=commoninc | |
630 | ) |
|
630 | ) | |
631 | ui.status(_(b'looking for unused includes to remove\n')) |
|
631 | ui.status(_(b'looking for unused includes to remove\n')) | |
632 | localfiles = set() |
|
632 | localfiles = set() | |
633 | for n in itertools.chain(outgoing.missing, outgoing.excluded): |
|
633 | for n in itertools.chain(outgoing.missing, outgoing.excluded): | |
634 | localfiles.update(repo[n].files()) |
|
634 | localfiles.update(repo[n].files()) | |
635 | suggestedremovals = [] |
|
635 | suggestedremovals = [] | |
636 | for include in sorted(oldincludes): |
|
636 | for include in sorted(oldincludes): | |
637 | match = narrowspec.match(repo.root, [include], oldexcludes) |
|
637 | match = narrowspec.match(repo.root, [include], oldexcludes) | |
638 | if not any(match(f) for f in localfiles): |
|
638 | if not any(match(f) for f in localfiles): | |
639 | suggestedremovals.append(include) |
|
639 | suggestedremovals.append(include) | |
640 | if suggestedremovals: |
|
640 | if suggestedremovals: | |
641 | for s in suggestedremovals: |
|
641 | for s in suggestedremovals: | |
642 | ui.status(b'%s\n' % s) |
|
642 | ui.status(b'%s\n' % s) | |
643 | if ( |
|
643 | if ( | |
644 | ui.promptchoice( |
|
644 | ui.promptchoice( | |
645 | _( |
|
645 | _( | |
646 | b'remove these unused includes (yn)?' |
|
646 | b'remove these unused includes (yn)?' | |
647 | b'$$ &Yes $$ &No' |
|
647 | b'$$ &Yes $$ &No' | |
648 | ) |
|
648 | ) | |
649 | ) |
|
649 | ) | |
650 | == 0 |
|
650 | == 0 | |
651 | ): |
|
651 | ): | |
652 | removedincludes.update(suggestedremovals) |
|
652 | removedincludes.update(suggestedremovals) | |
653 | narrowing = True |
|
653 | narrowing = True | |
654 | else: |
|
654 | else: | |
655 | ui.status(_(b'found no unused includes\n')) |
|
655 | ui.status(_(b'found no unused includes\n')) | |
656 |
|
656 | |||
657 | if narrowing: |
|
657 | if narrowing: | |
658 | newincludes = oldincludes - removedincludes |
|
658 | newincludes = oldincludes - removedincludes | |
659 | newexcludes = oldexcludes | addedexcludes |
|
659 | newexcludes = oldexcludes | addedexcludes | |
660 | _narrow( |
|
660 | _narrow( | |
661 | ui, |
|
661 | ui, | |
662 | repo, |
|
662 | repo, | |
663 | remote, |
|
663 | remote, | |
664 | commoninc, |
|
664 | commoninc, | |
665 | oldincludes, |
|
665 | oldincludes, | |
666 | oldexcludes, |
|
666 | oldexcludes, | |
667 | newincludes, |
|
667 | newincludes, | |
668 | newexcludes, |
|
668 | newexcludes, | |
669 | opts[b'force_delete_local_changes'], |
|
669 | opts[b'force_delete_local_changes'], | |
670 | opts[b'backup'], |
|
670 | opts[b'backup'], | |
671 | ) |
|
671 | ) | |
672 | # _narrow() updated the narrowspec and _widen() below needs to |
|
672 | # _narrow() updated the narrowspec and _widen() below needs to | |
673 | # use the updated values as its base (otherwise removed includes |
|
673 | # use the updated values as its base (otherwise removed includes | |
674 | # and addedexcludes will be lost in the resulting narrowspec) |
|
674 | # and addedexcludes will be lost in the resulting narrowspec) | |
675 | oldincludes = newincludes |
|
675 | oldincludes = newincludes | |
676 | oldexcludes = newexcludes |
|
676 | oldexcludes = newexcludes | |
677 |
|
677 | |||
678 | if widening: |
|
678 | if widening: | |
679 | newincludes = oldincludes | addedincludes |
|
679 | newincludes = oldincludes | addedincludes | |
680 | newexcludes = oldexcludes - removedexcludes |
|
680 | newexcludes = oldexcludes - removedexcludes | |
681 | _widen( |
|
681 | _widen( | |
682 | ui, |
|
682 | ui, | |
683 | repo, |
|
683 | repo, | |
684 | remote, |
|
684 | remote, | |
685 | commoninc, |
|
685 | commoninc, | |
686 | oldincludes, |
|
686 | oldincludes, | |
687 | oldexcludes, |
|
687 | oldexcludes, | |
688 | newincludes, |
|
688 | newincludes, | |
689 | newexcludes, |
|
689 | newexcludes, | |
690 | ) |
|
690 | ) | |
691 | finally: |
|
691 | finally: | |
692 | remote.close() |
|
692 | remote.close() | |
693 |
|
693 | |||
694 | return 0 |
|
694 | return 0 |
@@ -1,399 +1,399 b'' | |||||
1 | from __future__ import absolute_import |
|
1 | from __future__ import absolute_import | |
2 |
|
2 | |||
3 | import threading |
|
3 | import threading | |
4 |
|
4 | |||
5 | from mercurial.node import ( |
|
5 | from mercurial.node import ( | |
6 | hex, |
|
6 | hex, | |
7 | sha1nodeconstants, |
|
7 | sha1nodeconstants, | |
8 | ) |
|
8 | ) | |
9 | from mercurial.pycompat import getattr |
|
9 | from mercurial.pycompat import getattr | |
10 | from mercurial import ( |
|
10 | from mercurial import ( | |
11 | mdiff, |
|
11 | mdiff, | |
12 | pycompat, |
|
12 | pycompat, | |
13 | revlog, |
|
13 | revlog, | |
14 | ) |
|
14 | ) | |
15 | from . import ( |
|
15 | from . import ( | |
16 | basestore, |
|
16 | basestore, | |
17 | constants, |
|
17 | constants, | |
18 | shallowutil, |
|
18 | shallowutil, | |
19 | ) |
|
19 | ) | |
20 |
|
20 | |||
21 |
|
21 | |||
22 | class ChainIndicies(object): |
|
22 | class ChainIndicies(object): | |
23 | """A static class for easy reference to the delta chain indicies.""" |
|
23 | """A static class for easy reference to the delta chain indicies.""" | |
24 |
|
24 | |||
25 | # The filename of this revision delta |
|
25 | # The filename of this revision delta | |
26 | NAME = 0 |
|
26 | NAME = 0 | |
27 | # The mercurial file node for this revision delta |
|
27 | # The mercurial file node for this revision delta | |
28 | NODE = 1 |
|
28 | NODE = 1 | |
29 | # The filename of the delta base's revision. This is useful when delta |
|
29 | # The filename of the delta base's revision. This is useful when delta | |
30 | # between different files (like in the case of a move or copy, we can delta |
|
30 | # between different files (like in the case of a move or copy, we can delta | |
31 | # against the original file content). |
|
31 | # against the original file content). | |
32 | BASENAME = 2 |
|
32 | BASENAME = 2 | |
33 | # The mercurial file node for the delta base revision. This is the nullid if |
|
33 | # The mercurial file node for the delta base revision. This is the nullid if | |
34 | # this delta is a full text. |
|
34 | # this delta is a full text. | |
35 | BASENODE = 3 |
|
35 | BASENODE = 3 | |
36 | # The actual delta or full text data. |
|
36 | # The actual delta or full text data. | |
37 | DATA = 4 |
|
37 | DATA = 4 | |
38 |
|
38 | |||
39 |
|
39 | |||
40 | class unioncontentstore(basestore.baseunionstore): |
|
40 | class unioncontentstore(basestore.baseunionstore): | |
41 | def __init__(self, *args, **kwargs): |
|
41 | def __init__(self, *args, **kwargs): | |
42 | super(unioncontentstore, self).__init__(*args, **kwargs) |
|
42 | super(unioncontentstore, self).__init__(*args, **kwargs) | |
43 |
|
43 | |||
44 | self.stores = args |
|
44 | self.stores = args | |
45 | self.writestore = kwargs.get('writestore') |
|
45 | self.writestore = kwargs.get('writestore') | |
46 |
|
46 | |||
47 | # If allowincomplete==True then the union store can return partial |
|
47 | # If allowincomplete==True then the union store can return partial | |
48 | # delta chains, otherwise it will throw a KeyError if a full |
|
48 | # delta chains, otherwise it will throw a KeyError if a full | |
49 | # deltachain can't be found. |
|
49 | # deltachain can't be found. | |
50 | self.allowincomplete = kwargs.get('allowincomplete', False) |
|
50 | self.allowincomplete = kwargs.get('allowincomplete', False) | |
51 |
|
51 | |||
52 | def get(self, name, node): |
|
52 | def get(self, name, node): | |
53 | """Fetches the full text revision contents of the given name+node pair. |
|
53 | """Fetches the full text revision contents of the given name+node pair. | |
54 | If the full text doesn't exist, throws a KeyError. |
|
54 | If the full text doesn't exist, throws a KeyError. | |
55 |
|
55 | |||
56 | Under the hood, this uses getdeltachain() across all the stores to build |
|
56 | Under the hood, this uses getdeltachain() across all the stores to build | |
57 | up a full chain to produce the full text. |
|
57 | up a full chain to produce the full text. | |
58 | """ |
|
58 | """ | |
59 | chain = self.getdeltachain(name, node) |
|
59 | chain = self.getdeltachain(name, node) | |
60 |
|
60 | |||
61 | if chain[-1][ChainIndicies.BASENODE] != sha1nodeconstants.nullid: |
|
61 | if chain[-1][ChainIndicies.BASENODE] != sha1nodeconstants.nullid: | |
62 | # If we didn't receive a full chain, throw |
|
62 | # If we didn't receive a full chain, throw | |
63 | raise KeyError((name, hex(node))) |
|
63 | raise KeyError((name, hex(node))) | |
64 |
|
64 | |||
65 | # The last entry in the chain is a full text, so we start our delta |
|
65 | # The last entry in the chain is a full text, so we start our delta | |
66 | # applies with that. |
|
66 | # applies with that. | |
67 | fulltext = chain.pop()[ChainIndicies.DATA] |
|
67 | fulltext = chain.pop()[ChainIndicies.DATA] | |
68 |
|
68 | |||
69 | text = fulltext |
|
69 | text = fulltext | |
70 | while chain: |
|
70 | while chain: | |
71 | delta = chain.pop()[ChainIndicies.DATA] |
|
71 | delta = chain.pop()[ChainIndicies.DATA] | |
72 | text = mdiff.patches(text, [delta]) |
|
72 | text = mdiff.patches(text, [delta]) | |
73 |
|
73 | |||
74 | return text |
|
74 | return text | |
75 |
|
75 | |||
76 | @basestore.baseunionstore.retriable |
|
76 | @basestore.baseunionstore.retriable | |
77 | def getdelta(self, name, node): |
|
77 | def getdelta(self, name, node): | |
78 | """Return the single delta entry for the given name/node pair.""" |
|
78 | """Return the single delta entry for the given name/node pair.""" | |
79 | for store in self.stores: |
|
79 | for store in self.stores: | |
80 | try: |
|
80 | try: | |
81 | return store.getdelta(name, node) |
|
81 | return store.getdelta(name, node) | |
82 | except KeyError: |
|
82 | except KeyError: | |
83 | pass |
|
83 | pass | |
84 |
|
84 | |||
85 | raise KeyError((name, hex(node))) |
|
85 | raise KeyError((name, hex(node))) | |
86 |
|
86 | |||
87 | def getdeltachain(self, name, node): |
|
87 | def getdeltachain(self, name, node): | |
88 | """Returns the deltachain for the given name/node pair. |
|
88 | """Returns the deltachain for the given name/node pair. | |
89 |
|
89 | |||
90 | Returns an ordered list of: |
|
90 | Returns an ordered list of: | |
91 |
|
91 | |||
92 | [(name, node, deltabasename, deltabasenode, deltacontent),...] |
|
92 | [(name, node, deltabasename, deltabasenode, deltacontent),...] | |
93 |
|
93 | |||
94 | where the chain is terminated by a full text entry with a nullid |
|
94 | where the chain is terminated by a full text entry with a nullid | |
95 | deltabasenode. |
|
95 | deltabasenode. | |
96 | """ |
|
96 | """ | |
97 | chain = self._getpartialchain(name, node) |
|
97 | chain = self._getpartialchain(name, node) | |
98 | while chain[-1][ChainIndicies.BASENODE] != sha1nodeconstants.nullid: |
|
98 | while chain[-1][ChainIndicies.BASENODE] != sha1nodeconstants.nullid: | |
99 | x, x, deltabasename, deltabasenode, x = chain[-1] |
|
99 | x, x, deltabasename, deltabasenode, x = chain[-1] | |
100 | try: |
|
100 | try: | |
101 | morechain = self._getpartialchain(deltabasename, deltabasenode) |
|
101 | morechain = self._getpartialchain(deltabasename, deltabasenode) | |
102 | chain.extend(morechain) |
|
102 | chain.extend(morechain) | |
103 | except KeyError: |
|
103 | except KeyError: | |
104 | # If we allow incomplete chains, don't throw. |
|
104 | # If we allow incomplete chains, don't throw. | |
105 | if not self.allowincomplete: |
|
105 | if not self.allowincomplete: | |
106 | raise |
|
106 | raise | |
107 | break |
|
107 | break | |
108 |
|
108 | |||
109 | return chain |
|
109 | return chain | |
110 |
|
110 | |||
111 | @basestore.baseunionstore.retriable |
|
111 | @basestore.baseunionstore.retriable | |
112 | def getmeta(self, name, node): |
|
112 | def getmeta(self, name, node): | |
113 | """Returns the metadata dict for given node.""" |
|
113 | """Returns the metadata dict for given node.""" | |
114 | for store in self.stores: |
|
114 | for store in self.stores: | |
115 | try: |
|
115 | try: | |
116 | return store.getmeta(name, node) |
|
116 | return store.getmeta(name, node) | |
117 | except KeyError: |
|
117 | except KeyError: | |
118 | pass |
|
118 | pass | |
119 | raise KeyError((name, hex(node))) |
|
119 | raise KeyError((name, hex(node))) | |
120 |
|
120 | |||
121 | def getmetrics(self): |
|
121 | def getmetrics(self): | |
122 | metrics = [s.getmetrics() for s in self.stores] |
|
122 | metrics = [s.getmetrics() for s in self.stores] | |
123 | return shallowutil.sumdicts(*metrics) |
|
123 | return shallowutil.sumdicts(*metrics) | |
124 |
|
124 | |||
125 | @basestore.baseunionstore.retriable |
|
125 | @basestore.baseunionstore.retriable | |
126 | def _getpartialchain(self, name, node): |
|
126 | def _getpartialchain(self, name, node): | |
127 | """Returns a partial delta chain for the given name/node pair. |
|
127 | """Returns a partial delta chain for the given name/node pair. | |
128 |
|
128 | |||
129 | A partial chain is a chain that may not be terminated in a full-text. |
|
129 | A partial chain is a chain that may not be terminated in a full-text. | |
130 | """ |
|
130 | """ | |
131 | for store in self.stores: |
|
131 | for store in self.stores: | |
132 | try: |
|
132 | try: | |
133 | return store.getdeltachain(name, node) |
|
133 | return store.getdeltachain(name, node) | |
134 | except KeyError: |
|
134 | except KeyError: | |
135 | pass |
|
135 | pass | |
136 |
|
136 | |||
137 | raise KeyError((name, hex(node))) |
|
137 | raise KeyError((name, hex(node))) | |
138 |
|
138 | |||
139 | def add(self, name, node, data): |
|
139 | def add(self, name, node, data): | |
140 | raise RuntimeError( |
|
140 | raise RuntimeError( | |
141 | b"cannot add content only to remotefilelog contentstore" |
|
141 | b"cannot add content only to remotefilelog contentstore" | |
142 | ) |
|
142 | ) | |
143 |
|
143 | |||
144 | def getmissing(self, keys): |
|
144 | def getmissing(self, keys): | |
145 | missing = keys |
|
145 | missing = keys | |
146 | for store in self.stores: |
|
146 | for store in self.stores: | |
147 | if missing: |
|
147 | if missing: | |
148 | missing = store.getmissing(missing) |
|
148 | missing = store.getmissing(missing) | |
149 | return missing |
|
149 | return missing | |
150 |
|
150 | |||
151 | def addremotefilelognode(self, name, node, data): |
|
151 | def addremotefilelognode(self, name, node, data): | |
152 | if self.writestore: |
|
152 | if self.writestore: | |
153 | self.writestore.addremotefilelognode(name, node, data) |
|
153 | self.writestore.addremotefilelognode(name, node, data) | |
154 | else: |
|
154 | else: | |
155 | raise RuntimeError(b"no writable store configured") |
|
155 | raise RuntimeError(b"no writable store configured") | |
156 |
|
156 | |||
157 | def markledger(self, ledger, options=None): |
|
157 | def markledger(self, ledger, options=None): | |
158 | for store in self.stores: |
|
158 | for store in self.stores: | |
159 | store.markledger(ledger, options) |
|
159 | store.markledger(ledger, options) | |
160 |
|
160 | |||
161 |
|
161 | |||
162 | class remotefilelogcontentstore(basestore.basestore): |
|
162 | class remotefilelogcontentstore(basestore.basestore): | |
163 | def __init__(self, *args, **kwargs): |
|
163 | def __init__(self, *args, **kwargs): | |
164 | super(remotefilelogcontentstore, self).__init__(*args, **kwargs) |
|
164 | super(remotefilelogcontentstore, self).__init__(*args, **kwargs) | |
165 | self._threaddata = threading.local() |
|
165 | self._threaddata = threading.local() | |
166 |
|
166 | |||
167 | def get(self, name, node): |
|
167 | def get(self, name, node): | |
168 | # return raw revision text |
|
168 | # return raw revision text | |
169 | data = self._getdata(name, node) |
|
169 | data = self._getdata(name, node) | |
170 |
|
170 | |||
171 | offset, size, flags = shallowutil.parsesizeflags(data) |
|
171 | offset, size, flags = shallowutil.parsesizeflags(data) | |
172 | content = data[offset : offset + size] |
|
172 | content = data[offset : offset + size] | |
173 |
|
173 | |||
174 | ancestormap = shallowutil.ancestormap(data) |
|
174 | ancestormap = shallowutil.ancestormap(data) | |
175 | p1, p2, linknode, copyfrom = ancestormap[node] |
|
175 | p1, p2, linknode, copyfrom = ancestormap[node] | |
176 | copyrev = None |
|
176 | copyrev = None | |
177 | if copyfrom: |
|
177 | if copyfrom: | |
178 | copyrev = hex(p1) |
|
178 | copyrev = hex(p1) | |
179 |
|
179 | |||
180 | self._updatemetacache(node, size, flags) |
|
180 | self._updatemetacache(node, size, flags) | |
181 |
|
181 | |||
182 | # lfs tracks renames in its own metadata, remove hg copy metadata, |
|
182 | # lfs tracks renames in its own metadata, remove hg copy metadata, | |
183 | # because copy metadata will be re-added by lfs flag processor. |
|
183 | # because copy metadata will be re-added by lfs flag processor. | |
184 | if flags & revlog.REVIDX_EXTSTORED: |
|
184 | if flags & revlog.REVIDX_EXTSTORED: | |
185 | copyrev = copyfrom = None |
|
185 | copyrev = copyfrom = None | |
186 | revision = shallowutil.createrevlogtext(content, copyfrom, copyrev) |
|
186 | revision = shallowutil.createrevlogtext(content, copyfrom, copyrev) | |
187 | return revision |
|
187 | return revision | |
188 |
|
188 | |||
189 | def getdelta(self, name, node): |
|
189 | def getdelta(self, name, node): | |
190 | # Since remotefilelog content stores only contain full texts, just |
|
190 | # Since remotefilelog content stores only contain full texts, just | |
191 | # return that. |
|
191 | # return that. | |
192 | revision = self.get(name, node) |
|
192 | revision = self.get(name, node) | |
193 | return ( |
|
193 | return ( | |
194 | revision, |
|
194 | revision, | |
195 | name, |
|
195 | name, | |
196 | sha1nodeconstants.nullid, |
|
196 | sha1nodeconstants.nullid, | |
197 | self.getmeta(name, node), |
|
197 | self.getmeta(name, node), | |
198 | ) |
|
198 | ) | |
199 |
|
199 | |||
200 | def getdeltachain(self, name, node): |
|
200 | def getdeltachain(self, name, node): | |
201 | # Since remotefilelog content stores just contain full texts, we return |
|
201 | # Since remotefilelog content stores just contain full texts, we return | |
202 | # a fake delta chain that just consists of a single full text revision. |
|
202 | # a fake delta chain that just consists of a single full text revision. | |
203 | # The nullid in the deltabasenode slot indicates that the revision is a |
|
203 | # The nullid in the deltabasenode slot indicates that the revision is a | |
204 | # fulltext. |
|
204 | # fulltext. | |
205 | revision = self.get(name, node) |
|
205 | revision = self.get(name, node) | |
206 | return [(name, node, None, sha1nodeconstants.nullid, revision)] |
|
206 | return [(name, node, None, sha1nodeconstants.nullid, revision)] | |
207 |
|
207 | |||
208 | def getmeta(self, name, node): |
|
208 | def getmeta(self, name, node): | |
209 | self._sanitizemetacache() |
|
209 | self._sanitizemetacache() | |
210 | if node != self._threaddata.metacache[0]: |
|
210 | if node != self._threaddata.metacache[0]: | |
211 | data = self._getdata(name, node) |
|
211 | data = self._getdata(name, node) | |
212 | offset, size, flags = shallowutil.parsesizeflags(data) |
|
212 | offset, size, flags = shallowutil.parsesizeflags(data) | |
213 | self._updatemetacache(node, size, flags) |
|
213 | self._updatemetacache(node, size, flags) | |
214 | return self._threaddata.metacache[1] |
|
214 | return self._threaddata.metacache[1] | |
215 |
|
215 | |||
216 | def add(self, name, node, data): |
|
216 | def add(self, name, node, data): | |
217 | raise RuntimeError( |
|
217 | raise RuntimeError( | |
218 | b"cannot add content only to remotefilelog contentstore" |
|
218 | b"cannot add content only to remotefilelog contentstore" | |
219 | ) |
|
219 | ) | |
220 |
|
220 | |||
221 | def _sanitizemetacache(self): |
|
221 | def _sanitizemetacache(self): | |
222 | metacache = getattr(self._threaddata, 'metacache', None) |
|
222 | metacache = getattr(self._threaddata, 'metacache', None) | |
223 | if metacache is None: |
|
223 | if metacache is None: | |
224 | self._threaddata.metacache = (None, None) # (node, meta) |
|
224 | self._threaddata.metacache = (None, None) # (node, meta) | |
225 |
|
225 | |||
226 | def _updatemetacache(self, node, size, flags): |
|
226 | def _updatemetacache(self, node, size, flags): | |
227 | self._sanitizemetacache() |
|
227 | self._sanitizemetacache() | |
228 | if node == self._threaddata.metacache[0]: |
|
228 | if node == self._threaddata.metacache[0]: | |
229 | return |
|
229 | return | |
230 | meta = {constants.METAKEYFLAG: flags, constants.METAKEYSIZE: size} |
|
230 | meta = {constants.METAKEYFLAG: flags, constants.METAKEYSIZE: size} | |
231 | self._threaddata.metacache = (node, meta) |
|
231 | self._threaddata.metacache = (node, meta) | |
232 |
|
232 | |||
233 |
|
233 | |||
234 | class remotecontentstore(object): |
|
234 | class remotecontentstore(object): | |
235 | def __init__(self, ui, fileservice, shared): |
|
235 | def __init__(self, ui, fileservice, shared): | |
236 | self._fileservice = fileservice |
|
236 | self._fileservice = fileservice | |
237 | # type(shared) is usually remotefilelogcontentstore |
|
237 | # type(shared) is usually remotefilelogcontentstore | |
238 | self._shared = shared |
|
238 | self._shared = shared | |
239 |
|
239 | |||
240 | def get(self, name, node): |
|
240 | def get(self, name, node): | |
241 | self._fileservice.prefetch( |
|
241 | self._fileservice.prefetch( | |
242 | [(name, hex(node))], force=True, fetchdata=True |
|
242 | [(name, hex(node))], force=True, fetchdata=True | |
243 | ) |
|
243 | ) | |
244 | return self._shared.get(name, node) |
|
244 | return self._shared.get(name, node) | |
245 |
|
245 | |||
246 | def getdelta(self, name, node): |
|
246 | def getdelta(self, name, node): | |
247 | revision = self.get(name, node) |
|
247 | revision = self.get(name, node) | |
248 | return ( |
|
248 | return ( | |
249 | revision, |
|
249 | revision, | |
250 | name, |
|
250 | name, | |
251 | sha1nodeconstants.nullid, |
|
251 | sha1nodeconstants.nullid, | |
252 | self._shared.getmeta(name, node), |
|
252 | self._shared.getmeta(name, node), | |
253 | ) |
|
253 | ) | |
254 |
|
254 | |||
255 | def getdeltachain(self, name, node): |
|
255 | def getdeltachain(self, name, node): | |
256 | # Since our remote content stores just contain full texts, we return a |
|
256 | # Since our remote content stores just contain full texts, we return a | |
257 | # fake delta chain that just consists of a single full text revision. |
|
257 | # fake delta chain that just consists of a single full text revision. | |
258 | # The nullid in the deltabasenode slot indicates that the revision is a |
|
258 | # The nullid in the deltabasenode slot indicates that the revision is a | |
259 | # fulltext. |
|
259 | # fulltext. | |
260 | revision = self.get(name, node) |
|
260 | revision = self.get(name, node) | |
261 | return [(name, node, None, sha1nodeconstants.nullid, revision)] |
|
261 | return [(name, node, None, sha1nodeconstants.nullid, revision)] | |
262 |
|
262 | |||
263 | def getmeta(self, name, node): |
|
263 | def getmeta(self, name, node): | |
264 | self._fileservice.prefetch( |
|
264 | self._fileservice.prefetch( | |
265 | [(name, hex(node))], force=True, fetchdata=True |
|
265 | [(name, hex(node))], force=True, fetchdata=True | |
266 | ) |
|
266 | ) | |
267 | return self._shared.getmeta(name, node) |
|
267 | return self._shared.getmeta(name, node) | |
268 |
|
268 | |||
269 | def add(self, name, node, data): |
|
269 | def add(self, name, node, data): | |
270 | raise RuntimeError(b"cannot add to a remote store") |
|
270 | raise RuntimeError(b"cannot add to a remote store") | |
271 |
|
271 | |||
272 | def getmissing(self, keys): |
|
272 | def getmissing(self, keys): | |
273 | return keys |
|
273 | return keys | |
274 |
|
274 | |||
275 | def markledger(self, ledger, options=None): |
|
275 | def markledger(self, ledger, options=None): | |
276 | pass |
|
276 | pass | |
277 |
|
277 | |||
278 |
|
278 | |||
279 | class manifestrevlogstore(object): |
|
279 | class manifestrevlogstore(object): | |
280 | def __init__(self, repo): |
|
280 | def __init__(self, repo): | |
281 | self._store = repo.store |
|
281 | self._store = repo.store | |
282 | self._svfs = repo.svfs |
|
282 | self._svfs = repo.svfs | |
283 | self._revlogs = dict() |
|
283 | self._revlogs = dict() | |
284 | self._cl = revlog.revlog(self._svfs, radix=b'00changelog.i') |
|
284 | self._cl = revlog.revlog(self._svfs, radix=b'00changelog.i') | |
285 | self._repackstartlinkrev = 0 |
|
285 | self._repackstartlinkrev = 0 | |
286 |
|
286 | |||
287 | def get(self, name, node): |
|
287 | def get(self, name, node): | |
288 | return self._revlog(name).rawdata(node) |
|
288 | return self._revlog(name).rawdata(node) | |
289 |
|
289 | |||
290 | def getdelta(self, name, node): |
|
290 | def getdelta(self, name, node): | |
291 | revision = self.get(name, node) |
|
291 | revision = self.get(name, node) | |
292 | return revision, name, self._cl.nullid, self.getmeta(name, node) |
|
292 | return revision, name, self._cl.nullid, self.getmeta(name, node) | |
293 |
|
293 | |||
294 | def getdeltachain(self, name, node): |
|
294 | def getdeltachain(self, name, node): | |
295 | revision = self.get(name, node) |
|
295 | revision = self.get(name, node) | |
296 | return [(name, node, None, self._cl.nullid, revision)] |
|
296 | return [(name, node, None, self._cl.nullid, revision)] | |
297 |
|
297 | |||
298 | def getmeta(self, name, node): |
|
298 | def getmeta(self, name, node): | |
299 | rl = self._revlog(name) |
|
299 | rl = self._revlog(name) | |
300 | rev = rl.rev(node) |
|
300 | rev = rl.rev(node) | |
301 | return { |
|
301 | return { | |
302 | constants.METAKEYFLAG: rl.flags(rev), |
|
302 | constants.METAKEYFLAG: rl.flags(rev), | |
303 | constants.METAKEYSIZE: rl.rawsize(rev), |
|
303 | constants.METAKEYSIZE: rl.rawsize(rev), | |
304 | } |
|
304 | } | |
305 |
|
305 | |||
306 | def getancestors(self, name, node, known=None): |
|
306 | def getancestors(self, name, node, known=None): | |
307 | if known is None: |
|
307 | if known is None: | |
308 | known = set() |
|
308 | known = set() | |
309 | if node in known: |
|
309 | if node in known: | |
310 | return [] |
|
310 | return [] | |
311 |
|
311 | |||
312 | rl = self._revlog(name) |
|
312 | rl = self._revlog(name) | |
313 | ancestors = {} |
|
313 | ancestors = {} | |
314 | missing = {node} |
|
314 | missing = {node} | |
315 | for ancrev in rl.ancestors([rl.rev(node)], inclusive=True): |
|
315 | for ancrev in rl.ancestors([rl.rev(node)], inclusive=True): | |
316 | ancnode = rl.node(ancrev) |
|
316 | ancnode = rl.node(ancrev) | |
317 | missing.discard(ancnode) |
|
317 | missing.discard(ancnode) | |
318 |
|
318 | |||
319 | p1, p2 = rl.parents(ancnode) |
|
319 | p1, p2 = rl.parents(ancnode) | |
320 | if p1 != self._cl.nullid and p1 not in known: |
|
320 | if p1 != self._cl.nullid and p1 not in known: | |
321 | missing.add(p1) |
|
321 | missing.add(p1) | |
322 | if p2 != self._cl.nullid and p2 not in known: |
|
322 | if p2 != self._cl.nullid and p2 not in known: | |
323 | missing.add(p2) |
|
323 | missing.add(p2) | |
324 |
|
324 | |||
325 | linknode = self._cl.node(rl.linkrev(ancrev)) |
|
325 | linknode = self._cl.node(rl.linkrev(ancrev)) | |
326 | ancestors[rl.node(ancrev)] = (p1, p2, linknode, b'') |
|
326 | ancestors[rl.node(ancrev)] = (p1, p2, linknode, b'') | |
327 | if not missing: |
|
327 | if not missing: | |
328 | break |
|
328 | break | |
329 | return ancestors |
|
329 | return ancestors | |
330 |
|
330 | |||
331 | def getnodeinfo(self, name, node): |
|
331 | def getnodeinfo(self, name, node): | |
332 | cl = self._cl |
|
332 | cl = self._cl | |
333 | rl = self._revlog(name) |
|
333 | rl = self._revlog(name) | |
334 | parents = rl.parents(node) |
|
334 | parents = rl.parents(node) | |
335 | linkrev = rl.linkrev(rl.rev(node)) |
|
335 | linkrev = rl.linkrev(rl.rev(node)) | |
336 | return (parents[0], parents[1], cl.node(linkrev), None) |
|
336 | return (parents[0], parents[1], cl.node(linkrev), None) | |
337 |
|
337 | |||
338 | def add(self, *args): |
|
338 | def add(self, *args): | |
339 | raise RuntimeError(b"cannot add to a revlog store") |
|
339 | raise RuntimeError(b"cannot add to a revlog store") | |
340 |
|
340 | |||
341 | def _revlog(self, name): |
|
341 | def _revlog(self, name): | |
342 | rl = self._revlogs.get(name) |
|
342 | rl = self._revlogs.get(name) | |
343 | if rl is None: |
|
343 | if rl is None: | |
344 | revlogname = b'00manifesttree' |
|
344 | revlogname = b'00manifesttree' | |
345 | if name != b'': |
|
345 | if name != b'': | |
346 | revlogname = b'meta/%s/00manifest' % name |
|
346 | revlogname = b'meta/%s/00manifest' % name | |
347 | rl = revlog.revlog(self._svfs, radix=revlogname) |
|
347 | rl = revlog.revlog(self._svfs, radix=revlogname) | |
348 | self._revlogs[name] = rl |
|
348 | self._revlogs[name] = rl | |
349 | return rl |
|
349 | return rl | |
350 |
|
350 | |||
351 | def getmissing(self, keys): |
|
351 | def getmissing(self, keys): | |
352 | missing = [] |
|
352 | missing = [] | |
353 | for name, node in keys: |
|
353 | for name, node in keys: | |
354 | mfrevlog = self._revlog(name) |
|
354 | mfrevlog = self._revlog(name) | |
355 | if node not in mfrevlog.nodemap: |
|
355 | if node not in mfrevlog.nodemap: | |
356 | missing.append((name, node)) |
|
356 | missing.append((name, node)) | |
357 |
|
357 | |||
358 | return missing |
|
358 | return missing | |
359 |
|
359 | |||
360 | def setrepacklinkrevrange(self, startrev, endrev): |
|
360 | def setrepacklinkrevrange(self, startrev, endrev): | |
361 | self._repackstartlinkrev = startrev |
|
361 | self._repackstartlinkrev = startrev | |
362 | self._repackendlinkrev = endrev |
|
362 | self._repackendlinkrev = endrev | |
363 |
|
363 | |||
364 | def markledger(self, ledger, options=None): |
|
364 | def markledger(self, ledger, options=None): | |
365 | if options and options.get(constants.OPTION_PACKSONLY): |
|
365 | if options and options.get(constants.OPTION_PACKSONLY): | |
366 | return |
|
366 | return | |
367 | treename = b'' |
|
367 | treename = b'' | |
368 | rl = revlog.revlog(self._svfs, radix=b'00manifesttree') |
|
368 | rl = revlog.revlog(self._svfs, radix=b'00manifesttree') | |
369 | startlinkrev = self._repackstartlinkrev |
|
369 | startlinkrev = self._repackstartlinkrev | |
370 | endlinkrev = self._repackendlinkrev |
|
370 | endlinkrev = self._repackendlinkrev | |
371 | for rev in pycompat.xrange(len(rl) - 1, -1, -1): |
|
371 | for rev in pycompat.xrange(len(rl) - 1, -1, -1): | |
372 | linkrev = rl.linkrev(rev) |
|
372 | linkrev = rl.linkrev(rev) | |
373 | if linkrev < startlinkrev: |
|
373 | if linkrev < startlinkrev: | |
374 | break |
|
374 | break | |
375 | if linkrev > endlinkrev: |
|
375 | if linkrev > endlinkrev: | |
376 | continue |
|
376 | continue | |
377 | node = rl.node(rev) |
|
377 | node = rl.node(rev) | |
378 | ledger.markdataentry(self, treename, node) |
|
378 | ledger.markdataentry(self, treename, node) | |
379 | ledger.markhistoryentry(self, treename, node) |
|
379 | ledger.markhistoryentry(self, treename, node) | |
380 |
|
380 | |||
381 |
for t, path |
|
381 | for t, path, size in self._store.datafiles(): | |
382 | if path[:5] != b'meta/' or path[-2:] != b'.i': |
|
382 | if path[:5] != b'meta/' or path[-2:] != b'.i': | |
383 | continue |
|
383 | continue | |
384 |
|
384 | |||
385 | treename = path[5 : -len(b'/00manifest')] |
|
385 | treename = path[5 : -len(b'/00manifest')] | |
386 |
|
386 | |||
387 | rl = revlog.revlog(self._svfs, indexfile=path[:-2]) |
|
387 | rl = revlog.revlog(self._svfs, indexfile=path[:-2]) | |
388 | for rev in pycompat.xrange(len(rl) - 1, -1, -1): |
|
388 | for rev in pycompat.xrange(len(rl) - 1, -1, -1): | |
389 | linkrev = rl.linkrev(rev) |
|
389 | linkrev = rl.linkrev(rev) | |
390 | if linkrev < startlinkrev: |
|
390 | if linkrev < startlinkrev: | |
391 | break |
|
391 | break | |
392 | if linkrev > endlinkrev: |
|
392 | if linkrev > endlinkrev: | |
393 | continue |
|
393 | continue | |
394 | node = rl.node(rev) |
|
394 | node = rl.node(rev) | |
395 | ledger.markdataentry(self, treename, node) |
|
395 | ledger.markdataentry(self, treename, node) | |
396 | ledger.markhistoryentry(self, treename, node) |
|
396 | ledger.markhistoryentry(self, treename, node) | |
397 |
|
397 | |||
398 | def cleanup(self, ledger): |
|
398 | def cleanup(self, ledger): | |
399 | pass |
|
399 | pass |
@@ -1,441 +1,441 b'' | |||||
1 | # remotefilelogserver.py - server logic for a remotefilelog server |
|
1 | # remotefilelogserver.py - server logic for a remotefilelog server | |
2 | # |
|
2 | # | |
3 | # Copyright 2013 Facebook, Inc. |
|
3 | # Copyright 2013 Facebook, Inc. | |
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 | from __future__ import absolute_import |
|
7 | from __future__ import absolute_import | |
8 |
|
8 | |||
9 | import errno |
|
9 | import errno | |
10 | import os |
|
10 | import os | |
11 | import stat |
|
11 | import stat | |
12 | import time |
|
12 | import time | |
13 | import zlib |
|
13 | import zlib | |
14 |
|
14 | |||
15 | from mercurial.i18n import _ |
|
15 | from mercurial.i18n import _ | |
16 | from mercurial.node import bin, hex |
|
16 | from mercurial.node import bin, hex | |
17 | from mercurial.pycompat import open |
|
17 | from mercurial.pycompat import open | |
18 | from mercurial import ( |
|
18 | from mercurial import ( | |
19 | changegroup, |
|
19 | changegroup, | |
20 | changelog, |
|
20 | changelog, | |
21 | context, |
|
21 | context, | |
22 | error, |
|
22 | error, | |
23 | extensions, |
|
23 | extensions, | |
24 | match, |
|
24 | match, | |
25 | pycompat, |
|
25 | pycompat, | |
26 | scmutil, |
|
26 | scmutil, | |
27 | store, |
|
27 | store, | |
28 | streamclone, |
|
28 | streamclone, | |
29 | util, |
|
29 | util, | |
30 | wireprotoserver, |
|
30 | wireprotoserver, | |
31 | wireprototypes, |
|
31 | wireprototypes, | |
32 | wireprotov1server, |
|
32 | wireprotov1server, | |
33 | ) |
|
33 | ) | |
34 | from . import ( |
|
34 | from . import ( | |
35 | constants, |
|
35 | constants, | |
36 | shallowutil, |
|
36 | shallowutil, | |
37 | ) |
|
37 | ) | |
38 |
|
38 | |||
39 | _sshv1server = wireprotoserver.sshv1protocolhandler |
|
39 | _sshv1server = wireprotoserver.sshv1protocolhandler | |
40 |
|
40 | |||
41 |
|
41 | |||
42 | def setupserver(ui, repo): |
|
42 | def setupserver(ui, repo): | |
43 | """Sets up a normal Mercurial repo so it can serve files to shallow repos.""" |
|
43 | """Sets up a normal Mercurial repo so it can serve files to shallow repos.""" | |
44 | onetimesetup(ui) |
|
44 | onetimesetup(ui) | |
45 |
|
45 | |||
46 | # don't send files to shallow clients during pulls |
|
46 | # don't send files to shallow clients during pulls | |
47 | def generatefiles( |
|
47 | def generatefiles( | |
48 | orig, self, changedfiles, linknodes, commonrevs, source, *args, **kwargs |
|
48 | orig, self, changedfiles, linknodes, commonrevs, source, *args, **kwargs | |
49 | ): |
|
49 | ): | |
50 | caps = self._bundlecaps or [] |
|
50 | caps = self._bundlecaps or [] | |
51 | if constants.BUNDLE2_CAPABLITY in caps: |
|
51 | if constants.BUNDLE2_CAPABLITY in caps: | |
52 | # only send files that don't match the specified patterns |
|
52 | # only send files that don't match the specified patterns | |
53 | includepattern = None |
|
53 | includepattern = None | |
54 | excludepattern = None |
|
54 | excludepattern = None | |
55 | for cap in self._bundlecaps or []: |
|
55 | for cap in self._bundlecaps or []: | |
56 | if cap.startswith(b"includepattern="): |
|
56 | if cap.startswith(b"includepattern="): | |
57 | includepattern = cap[len(b"includepattern=") :].split(b'\0') |
|
57 | includepattern = cap[len(b"includepattern=") :].split(b'\0') | |
58 | elif cap.startswith(b"excludepattern="): |
|
58 | elif cap.startswith(b"excludepattern="): | |
59 | excludepattern = cap[len(b"excludepattern=") :].split(b'\0') |
|
59 | excludepattern = cap[len(b"excludepattern=") :].split(b'\0') | |
60 |
|
60 | |||
61 | m = match.always() |
|
61 | m = match.always() | |
62 | if includepattern or excludepattern: |
|
62 | if includepattern or excludepattern: | |
63 | m = match.match( |
|
63 | m = match.match( | |
64 | repo.root, b'', None, includepattern, excludepattern |
|
64 | repo.root, b'', None, includepattern, excludepattern | |
65 | ) |
|
65 | ) | |
66 |
|
66 | |||
67 | changedfiles = list([f for f in changedfiles if not m(f)]) |
|
67 | changedfiles = list([f for f in changedfiles if not m(f)]) | |
68 | return orig( |
|
68 | return orig( | |
69 | self, changedfiles, linknodes, commonrevs, source, *args, **kwargs |
|
69 | self, changedfiles, linknodes, commonrevs, source, *args, **kwargs | |
70 | ) |
|
70 | ) | |
71 |
|
71 | |||
72 | extensions.wrapfunction( |
|
72 | extensions.wrapfunction( | |
73 | changegroup.cgpacker, b'generatefiles', generatefiles |
|
73 | changegroup.cgpacker, b'generatefiles', generatefiles | |
74 | ) |
|
74 | ) | |
75 |
|
75 | |||
76 |
|
76 | |||
77 | onetime = False |
|
77 | onetime = False | |
78 |
|
78 | |||
79 |
|
79 | |||
80 | def onetimesetup(ui): |
|
80 | def onetimesetup(ui): | |
81 | """Configures the wireprotocol for both clients and servers.""" |
|
81 | """Configures the wireprotocol for both clients and servers.""" | |
82 | global onetime |
|
82 | global onetime | |
83 | if onetime: |
|
83 | if onetime: | |
84 | return |
|
84 | return | |
85 | onetime = True |
|
85 | onetime = True | |
86 |
|
86 | |||
87 | # support file content requests |
|
87 | # support file content requests | |
88 | wireprotov1server.wireprotocommand( |
|
88 | wireprotov1server.wireprotocommand( | |
89 | b'x_rfl_getflogheads', b'path', permission=b'pull' |
|
89 | b'x_rfl_getflogheads', b'path', permission=b'pull' | |
90 | )(getflogheads) |
|
90 | )(getflogheads) | |
91 | wireprotov1server.wireprotocommand( |
|
91 | wireprotov1server.wireprotocommand( | |
92 | b'x_rfl_getfiles', b'', permission=b'pull' |
|
92 | b'x_rfl_getfiles', b'', permission=b'pull' | |
93 | )(getfiles) |
|
93 | )(getfiles) | |
94 | wireprotov1server.wireprotocommand( |
|
94 | wireprotov1server.wireprotocommand( | |
95 | b'x_rfl_getfile', b'file node', permission=b'pull' |
|
95 | b'x_rfl_getfile', b'file node', permission=b'pull' | |
96 | )(getfile) |
|
96 | )(getfile) | |
97 |
|
97 | |||
98 | class streamstate(object): |
|
98 | class streamstate(object): | |
99 | match = None |
|
99 | match = None | |
100 | shallowremote = False |
|
100 | shallowremote = False | |
101 | noflatmf = False |
|
101 | noflatmf = False | |
102 |
|
102 | |||
103 | state = streamstate() |
|
103 | state = streamstate() | |
104 |
|
104 | |||
105 | def stream_out_shallow(repo, proto, other): |
|
105 | def stream_out_shallow(repo, proto, other): | |
106 | includepattern = None |
|
106 | includepattern = None | |
107 | excludepattern = None |
|
107 | excludepattern = None | |
108 | raw = other.get(b'includepattern') |
|
108 | raw = other.get(b'includepattern') | |
109 | if raw: |
|
109 | if raw: | |
110 | includepattern = raw.split(b'\0') |
|
110 | includepattern = raw.split(b'\0') | |
111 | raw = other.get(b'excludepattern') |
|
111 | raw = other.get(b'excludepattern') | |
112 | if raw: |
|
112 | if raw: | |
113 | excludepattern = raw.split(b'\0') |
|
113 | excludepattern = raw.split(b'\0') | |
114 |
|
114 | |||
115 | oldshallow = state.shallowremote |
|
115 | oldshallow = state.shallowremote | |
116 | oldmatch = state.match |
|
116 | oldmatch = state.match | |
117 | oldnoflatmf = state.noflatmf |
|
117 | oldnoflatmf = state.noflatmf | |
118 | try: |
|
118 | try: | |
119 | state.shallowremote = True |
|
119 | state.shallowremote = True | |
120 | state.match = match.always() |
|
120 | state.match = match.always() | |
121 | state.noflatmf = other.get(b'noflatmanifest') == b'True' |
|
121 | state.noflatmf = other.get(b'noflatmanifest') == b'True' | |
122 | if includepattern or excludepattern: |
|
122 | if includepattern or excludepattern: | |
123 | state.match = match.match( |
|
123 | state.match = match.match( | |
124 | repo.root, b'', None, includepattern, excludepattern |
|
124 | repo.root, b'', None, includepattern, excludepattern | |
125 | ) |
|
125 | ) | |
126 | streamres = wireprotov1server.stream(repo, proto) |
|
126 | streamres = wireprotov1server.stream(repo, proto) | |
127 |
|
127 | |||
128 | # Force the first value to execute, so the file list is computed |
|
128 | # Force the first value to execute, so the file list is computed | |
129 | # within the try/finally scope |
|
129 | # within the try/finally scope | |
130 | first = next(streamres.gen) |
|
130 | first = next(streamres.gen) | |
131 | second = next(streamres.gen) |
|
131 | second = next(streamres.gen) | |
132 |
|
132 | |||
133 | def gen(): |
|
133 | def gen(): | |
134 | yield first |
|
134 | yield first | |
135 | yield second |
|
135 | yield second | |
136 | for value in streamres.gen: |
|
136 | for value in streamres.gen: | |
137 | yield value |
|
137 | yield value | |
138 |
|
138 | |||
139 | return wireprototypes.streamres(gen()) |
|
139 | return wireprototypes.streamres(gen()) | |
140 | finally: |
|
140 | finally: | |
141 | state.shallowremote = oldshallow |
|
141 | state.shallowremote = oldshallow | |
142 | state.match = oldmatch |
|
142 | state.match = oldmatch | |
143 | state.noflatmf = oldnoflatmf |
|
143 | state.noflatmf = oldnoflatmf | |
144 |
|
144 | |||
145 | wireprotov1server.commands[b'stream_out_shallow'] = ( |
|
145 | wireprotov1server.commands[b'stream_out_shallow'] = ( | |
146 | stream_out_shallow, |
|
146 | stream_out_shallow, | |
147 | b'*', |
|
147 | b'*', | |
148 | ) |
|
148 | ) | |
149 |
|
149 | |||
150 | # don't clone filelogs to shallow clients |
|
150 | # don't clone filelogs to shallow clients | |
151 | def _walkstreamfiles(orig, repo, matcher=None): |
|
151 | def _walkstreamfiles(orig, repo, matcher=None): | |
152 | if state.shallowremote: |
|
152 | if state.shallowremote: | |
153 | # if we are shallow ourselves, stream our local commits |
|
153 | # if we are shallow ourselves, stream our local commits | |
154 | if shallowutil.isenabled(repo): |
|
154 | if shallowutil.isenabled(repo): | |
155 | striplen = len(repo.store.path) + 1 |
|
155 | striplen = len(repo.store.path) + 1 | |
156 | readdir = repo.store.rawvfs.readdir |
|
156 | readdir = repo.store.rawvfs.readdir | |
157 | visit = [os.path.join(repo.store.path, b'data')] |
|
157 | visit = [os.path.join(repo.store.path, b'data')] | |
158 | while visit: |
|
158 | while visit: | |
159 | p = visit.pop() |
|
159 | p = visit.pop() | |
160 | for f, kind, st in readdir(p, stat=True): |
|
160 | for f, kind, st in readdir(p, stat=True): | |
161 | fp = p + b'/' + f |
|
161 | fp = p + b'/' + f | |
162 | if kind == stat.S_IFREG: |
|
162 | if kind == stat.S_IFREG: | |
163 | if not fp.endswith(b'.i') and not fp.endswith( |
|
163 | if not fp.endswith(b'.i') and not fp.endswith( | |
164 | b'.d' |
|
164 | b'.d' | |
165 | ): |
|
165 | ): | |
166 | n = util.pconvert(fp[striplen:]) |
|
166 | n = util.pconvert(fp[striplen:]) | |
167 | d = store.decodedir(n) |
|
167 | d = store.decodedir(n) | |
168 | t = store.FILETYPE_OTHER |
|
168 | t = store.FILETYPE_OTHER | |
169 |
yield (t, d, |
|
169 | yield (t, d, st.st_size) | |
170 | if kind == stat.S_IFDIR: |
|
170 | if kind == stat.S_IFDIR: | |
171 | visit.append(fp) |
|
171 | visit.append(fp) | |
172 |
|
172 | |||
173 | if scmutil.istreemanifest(repo): |
|
173 | if scmutil.istreemanifest(repo): | |
174 |
for (t, u |
|
174 | for (t, u, s) in repo.store.datafiles(): | |
175 | if u.startswith(b'meta/') and ( |
|
175 | if u.startswith(b'meta/') and ( | |
176 | u.endswith(b'.i') or u.endswith(b'.d') |
|
176 | u.endswith(b'.i') or u.endswith(b'.d') | |
177 | ): |
|
177 | ): | |
178 |
yield (t, u, |
|
178 | yield (t, u, s) | |
179 |
|
179 | |||
180 | # Return .d and .i files that do not match the shallow pattern |
|
180 | # Return .d and .i files that do not match the shallow pattern | |
181 | match = state.match |
|
181 | match = state.match | |
182 | if match and not match.always(): |
|
182 | if match and not match.always(): | |
183 |
for (t, u |
|
183 | for (t, u, s) in repo.store.datafiles(): | |
184 | f = u[5:-2] # trim data/... and .i/.d |
|
184 | f = u[5:-2] # trim data/... and .i/.d | |
185 | if not state.match(f): |
|
185 | if not state.match(f): | |
186 |
yield (t, u, |
|
186 | yield (t, u, s) | |
187 |
|
187 | |||
188 | for x in repo.store.topfiles(): |
|
188 | for x in repo.store.topfiles(): | |
189 | if state.noflatmf and x[1][:11] == b'00manifest.': |
|
189 | if state.noflatmf and x[1][:11] == b'00manifest.': | |
190 | continue |
|
190 | continue | |
191 | yield x |
|
191 | yield x | |
192 |
|
192 | |||
193 | elif shallowutil.isenabled(repo): |
|
193 | elif shallowutil.isenabled(repo): | |
194 | # don't allow cloning from a shallow repo to a full repo |
|
194 | # don't allow cloning from a shallow repo to a full repo | |
195 | # since it would require fetching every version of every |
|
195 | # since it would require fetching every version of every | |
196 | # file in order to create the revlogs. |
|
196 | # file in order to create the revlogs. | |
197 | raise error.Abort( |
|
197 | raise error.Abort( | |
198 | _(b"Cannot clone from a shallow repo to a full repo.") |
|
198 | _(b"Cannot clone from a shallow repo to a full repo.") | |
199 | ) |
|
199 | ) | |
200 | else: |
|
200 | else: | |
201 | for x in orig(repo, matcher): |
|
201 | for x in orig(repo, matcher): | |
202 | yield x |
|
202 | yield x | |
203 |
|
203 | |||
204 | extensions.wrapfunction(streamclone, b'_walkstreamfiles', _walkstreamfiles) |
|
204 | extensions.wrapfunction(streamclone, b'_walkstreamfiles', _walkstreamfiles) | |
205 |
|
205 | |||
206 | # expose remotefilelog capabilities |
|
206 | # expose remotefilelog capabilities | |
207 | def _capabilities(orig, repo, proto): |
|
207 | def _capabilities(orig, repo, proto): | |
208 | caps = orig(repo, proto) |
|
208 | caps = orig(repo, proto) | |
209 | if shallowutil.isenabled(repo) or ui.configbool( |
|
209 | if shallowutil.isenabled(repo) or ui.configbool( | |
210 | b'remotefilelog', b'server' |
|
210 | b'remotefilelog', b'server' | |
211 | ): |
|
211 | ): | |
212 | if isinstance(proto, _sshv1server): |
|
212 | if isinstance(proto, _sshv1server): | |
213 | # legacy getfiles method which only works over ssh |
|
213 | # legacy getfiles method which only works over ssh | |
214 | caps.append(constants.NETWORK_CAP_LEGACY_SSH_GETFILES) |
|
214 | caps.append(constants.NETWORK_CAP_LEGACY_SSH_GETFILES) | |
215 | caps.append(b'x_rfl_getflogheads') |
|
215 | caps.append(b'x_rfl_getflogheads') | |
216 | caps.append(b'x_rfl_getfile') |
|
216 | caps.append(b'x_rfl_getfile') | |
217 | return caps |
|
217 | return caps | |
218 |
|
218 | |||
219 | extensions.wrapfunction(wireprotov1server, b'_capabilities', _capabilities) |
|
219 | extensions.wrapfunction(wireprotov1server, b'_capabilities', _capabilities) | |
220 |
|
220 | |||
221 | def _adjustlinkrev(orig, self, *args, **kwargs): |
|
221 | def _adjustlinkrev(orig, self, *args, **kwargs): | |
222 | # When generating file blobs, taking the real path is too slow on large |
|
222 | # When generating file blobs, taking the real path is too slow on large | |
223 | # repos, so force it to just return the linkrev directly. |
|
223 | # repos, so force it to just return the linkrev directly. | |
224 | repo = self._repo |
|
224 | repo = self._repo | |
225 | if util.safehasattr(repo, b'forcelinkrev') and repo.forcelinkrev: |
|
225 | if util.safehasattr(repo, b'forcelinkrev') and repo.forcelinkrev: | |
226 | return self._filelog.linkrev(self._filelog.rev(self._filenode)) |
|
226 | return self._filelog.linkrev(self._filelog.rev(self._filenode)) | |
227 | return orig(self, *args, **kwargs) |
|
227 | return orig(self, *args, **kwargs) | |
228 |
|
228 | |||
229 | extensions.wrapfunction( |
|
229 | extensions.wrapfunction( | |
230 | context.basefilectx, b'_adjustlinkrev', _adjustlinkrev |
|
230 | context.basefilectx, b'_adjustlinkrev', _adjustlinkrev | |
231 | ) |
|
231 | ) | |
232 |
|
232 | |||
233 | def _iscmd(orig, cmd): |
|
233 | def _iscmd(orig, cmd): | |
234 | if cmd == b'x_rfl_getfiles': |
|
234 | if cmd == b'x_rfl_getfiles': | |
235 | return False |
|
235 | return False | |
236 | return orig(cmd) |
|
236 | return orig(cmd) | |
237 |
|
237 | |||
238 | extensions.wrapfunction(wireprotoserver, b'iscmd', _iscmd) |
|
238 | extensions.wrapfunction(wireprotoserver, b'iscmd', _iscmd) | |
239 |
|
239 | |||
240 |
|
240 | |||
241 | def _loadfileblob(repo, cachepath, path, node): |
|
241 | def _loadfileblob(repo, cachepath, path, node): | |
242 | filecachepath = os.path.join(cachepath, path, hex(node)) |
|
242 | filecachepath = os.path.join(cachepath, path, hex(node)) | |
243 | if not os.path.exists(filecachepath) or os.path.getsize(filecachepath) == 0: |
|
243 | if not os.path.exists(filecachepath) or os.path.getsize(filecachepath) == 0: | |
244 | filectx = repo.filectx(path, fileid=node) |
|
244 | filectx = repo.filectx(path, fileid=node) | |
245 | if filectx.node() == repo.nullid: |
|
245 | if filectx.node() == repo.nullid: | |
246 | repo.changelog = changelog.changelog(repo.svfs) |
|
246 | repo.changelog = changelog.changelog(repo.svfs) | |
247 | filectx = repo.filectx(path, fileid=node) |
|
247 | filectx = repo.filectx(path, fileid=node) | |
248 |
|
248 | |||
249 | text = createfileblob(filectx) |
|
249 | text = createfileblob(filectx) | |
250 | # TODO configurable compression engines |
|
250 | # TODO configurable compression engines | |
251 | text = zlib.compress(text) |
|
251 | text = zlib.compress(text) | |
252 |
|
252 | |||
253 | # everything should be user & group read/writable |
|
253 | # everything should be user & group read/writable | |
254 | oldumask = os.umask(0o002) |
|
254 | oldumask = os.umask(0o002) | |
255 | try: |
|
255 | try: | |
256 | dirname = os.path.dirname(filecachepath) |
|
256 | dirname = os.path.dirname(filecachepath) | |
257 | if not os.path.exists(dirname): |
|
257 | if not os.path.exists(dirname): | |
258 | try: |
|
258 | try: | |
259 | os.makedirs(dirname) |
|
259 | os.makedirs(dirname) | |
260 | except OSError as ex: |
|
260 | except OSError as ex: | |
261 | if ex.errno != errno.EEXIST: |
|
261 | if ex.errno != errno.EEXIST: | |
262 | raise |
|
262 | raise | |
263 |
|
263 | |||
264 | f = None |
|
264 | f = None | |
265 | try: |
|
265 | try: | |
266 | f = util.atomictempfile(filecachepath, b"wb") |
|
266 | f = util.atomictempfile(filecachepath, b"wb") | |
267 | f.write(text) |
|
267 | f.write(text) | |
268 | except (IOError, OSError): |
|
268 | except (IOError, OSError): | |
269 | # Don't abort if the user only has permission to read, |
|
269 | # Don't abort if the user only has permission to read, | |
270 | # and not write. |
|
270 | # and not write. | |
271 | pass |
|
271 | pass | |
272 | finally: |
|
272 | finally: | |
273 | if f: |
|
273 | if f: | |
274 | f.close() |
|
274 | f.close() | |
275 | finally: |
|
275 | finally: | |
276 | os.umask(oldumask) |
|
276 | os.umask(oldumask) | |
277 | else: |
|
277 | else: | |
278 | with open(filecachepath, b"rb") as f: |
|
278 | with open(filecachepath, b"rb") as f: | |
279 | text = f.read() |
|
279 | text = f.read() | |
280 | return text |
|
280 | return text | |
281 |
|
281 | |||
282 |
|
282 | |||
283 | def getflogheads(repo, proto, path): |
|
283 | def getflogheads(repo, proto, path): | |
284 | """A server api for requesting a filelog's heads""" |
|
284 | """A server api for requesting a filelog's heads""" | |
285 | flog = repo.file(path) |
|
285 | flog = repo.file(path) | |
286 | heads = flog.heads() |
|
286 | heads = flog.heads() | |
287 | return b'\n'.join((hex(head) for head in heads if head != repo.nullid)) |
|
287 | return b'\n'.join((hex(head) for head in heads if head != repo.nullid)) | |
288 |
|
288 | |||
289 |
|
289 | |||
290 | def getfile(repo, proto, file, node): |
|
290 | def getfile(repo, proto, file, node): | |
291 | """A server api for requesting a particular version of a file. Can be used |
|
291 | """A server api for requesting a particular version of a file. Can be used | |
292 | in batches to request many files at once. The return protocol is: |
|
292 | in batches to request many files at once. The return protocol is: | |
293 | <errorcode>\0<data/errormsg> where <errorcode> is 0 for success or |
|
293 | <errorcode>\0<data/errormsg> where <errorcode> is 0 for success or | |
294 | non-zero for an error. |
|
294 | non-zero for an error. | |
295 |
|
295 | |||
296 | data is a compressed blob with revlog flag and ancestors information. See |
|
296 | data is a compressed blob with revlog flag and ancestors information. See | |
297 | createfileblob for its content. |
|
297 | createfileblob for its content. | |
298 | """ |
|
298 | """ | |
299 | if shallowutil.isenabled(repo): |
|
299 | if shallowutil.isenabled(repo): | |
300 | return b'1\0' + _(b'cannot fetch remote files from shallow repo') |
|
300 | return b'1\0' + _(b'cannot fetch remote files from shallow repo') | |
301 | cachepath = repo.ui.config(b"remotefilelog", b"servercachepath") |
|
301 | cachepath = repo.ui.config(b"remotefilelog", b"servercachepath") | |
302 | if not cachepath: |
|
302 | if not cachepath: | |
303 | cachepath = os.path.join(repo.path, b"remotefilelogcache") |
|
303 | cachepath = os.path.join(repo.path, b"remotefilelogcache") | |
304 | node = bin(node.strip()) |
|
304 | node = bin(node.strip()) | |
305 | if node == repo.nullid: |
|
305 | if node == repo.nullid: | |
306 | return b'0\0' |
|
306 | return b'0\0' | |
307 | return b'0\0' + _loadfileblob(repo, cachepath, file, node) |
|
307 | return b'0\0' + _loadfileblob(repo, cachepath, file, node) | |
308 |
|
308 | |||
309 |
|
309 | |||
310 | def getfiles(repo, proto): |
|
310 | def getfiles(repo, proto): | |
311 | """A server api for requesting particular versions of particular files.""" |
|
311 | """A server api for requesting particular versions of particular files.""" | |
312 | if shallowutil.isenabled(repo): |
|
312 | if shallowutil.isenabled(repo): | |
313 | raise error.Abort(_(b'cannot fetch remote files from shallow repo')) |
|
313 | raise error.Abort(_(b'cannot fetch remote files from shallow repo')) | |
314 | if not isinstance(proto, _sshv1server): |
|
314 | if not isinstance(proto, _sshv1server): | |
315 | raise error.Abort(_(b'cannot fetch remote files over non-ssh protocol')) |
|
315 | raise error.Abort(_(b'cannot fetch remote files over non-ssh protocol')) | |
316 |
|
316 | |||
317 | def streamer(): |
|
317 | def streamer(): | |
318 | fin = proto._fin |
|
318 | fin = proto._fin | |
319 |
|
319 | |||
320 | cachepath = repo.ui.config(b"remotefilelog", b"servercachepath") |
|
320 | cachepath = repo.ui.config(b"remotefilelog", b"servercachepath") | |
321 | if not cachepath: |
|
321 | if not cachepath: | |
322 | cachepath = os.path.join(repo.path, b"remotefilelogcache") |
|
322 | cachepath = os.path.join(repo.path, b"remotefilelogcache") | |
323 |
|
323 | |||
324 | while True: |
|
324 | while True: | |
325 | request = fin.readline()[:-1] |
|
325 | request = fin.readline()[:-1] | |
326 | if not request: |
|
326 | if not request: | |
327 | break |
|
327 | break | |
328 |
|
328 | |||
329 | node = bin(request[:40]) |
|
329 | node = bin(request[:40]) | |
330 | if node == repo.nullid: |
|
330 | if node == repo.nullid: | |
331 | yield b'0\n' |
|
331 | yield b'0\n' | |
332 | continue |
|
332 | continue | |
333 |
|
333 | |||
334 | path = request[40:] |
|
334 | path = request[40:] | |
335 |
|
335 | |||
336 | text = _loadfileblob(repo, cachepath, path, node) |
|
336 | text = _loadfileblob(repo, cachepath, path, node) | |
337 |
|
337 | |||
338 | yield b'%d\n%s' % (len(text), text) |
|
338 | yield b'%d\n%s' % (len(text), text) | |
339 |
|
339 | |||
340 | # it would be better to only flush after processing a whole batch |
|
340 | # it would be better to only flush after processing a whole batch | |
341 | # but currently we don't know if there are more requests coming |
|
341 | # but currently we don't know if there are more requests coming | |
342 | proto._fout.flush() |
|
342 | proto._fout.flush() | |
343 |
|
343 | |||
344 | return wireprototypes.streamres(streamer()) |
|
344 | return wireprototypes.streamres(streamer()) | |
345 |
|
345 | |||
346 |
|
346 | |||
347 | def createfileblob(filectx): |
|
347 | def createfileblob(filectx): | |
348 | """ |
|
348 | """ | |
349 | format: |
|
349 | format: | |
350 | v0: |
|
350 | v0: | |
351 | str(len(rawtext)) + '\0' + rawtext + ancestortext |
|
351 | str(len(rawtext)) + '\0' + rawtext + ancestortext | |
352 | v1: |
|
352 | v1: | |
353 | 'v1' + '\n' + metalist + '\0' + rawtext + ancestortext |
|
353 | 'v1' + '\n' + metalist + '\0' + rawtext + ancestortext | |
354 | metalist := metalist + '\n' + meta | meta |
|
354 | metalist := metalist + '\n' + meta | meta | |
355 | meta := sizemeta | flagmeta |
|
355 | meta := sizemeta | flagmeta | |
356 | sizemeta := METAKEYSIZE + str(len(rawtext)) |
|
356 | sizemeta := METAKEYSIZE + str(len(rawtext)) | |
357 | flagmeta := METAKEYFLAG + str(flag) |
|
357 | flagmeta := METAKEYFLAG + str(flag) | |
358 |
|
358 | |||
359 | note: sizemeta must exist. METAKEYFLAG and METAKEYSIZE must have a |
|
359 | note: sizemeta must exist. METAKEYFLAG and METAKEYSIZE must have a | |
360 | length of 1. |
|
360 | length of 1. | |
361 | """ |
|
361 | """ | |
362 | flog = filectx.filelog() |
|
362 | flog = filectx.filelog() | |
363 | frev = filectx.filerev() |
|
363 | frev = filectx.filerev() | |
364 | revlogflags = flog._revlog.flags(frev) |
|
364 | revlogflags = flog._revlog.flags(frev) | |
365 | if revlogflags == 0: |
|
365 | if revlogflags == 0: | |
366 | # normal files |
|
366 | # normal files | |
367 | text = filectx.data() |
|
367 | text = filectx.data() | |
368 | else: |
|
368 | else: | |
369 | # lfs, read raw revision data |
|
369 | # lfs, read raw revision data | |
370 | text = flog.rawdata(frev) |
|
370 | text = flog.rawdata(frev) | |
371 |
|
371 | |||
372 | repo = filectx._repo |
|
372 | repo = filectx._repo | |
373 |
|
373 | |||
374 | ancestors = [filectx] |
|
374 | ancestors = [filectx] | |
375 |
|
375 | |||
376 | try: |
|
376 | try: | |
377 | repo.forcelinkrev = True |
|
377 | repo.forcelinkrev = True | |
378 | ancestors.extend([f for f in filectx.ancestors()]) |
|
378 | ancestors.extend([f for f in filectx.ancestors()]) | |
379 |
|
379 | |||
380 | ancestortext = b"" |
|
380 | ancestortext = b"" | |
381 | for ancestorctx in ancestors: |
|
381 | for ancestorctx in ancestors: | |
382 | parents = ancestorctx.parents() |
|
382 | parents = ancestorctx.parents() | |
383 | p1 = repo.nullid |
|
383 | p1 = repo.nullid | |
384 | p2 = repo.nullid |
|
384 | p2 = repo.nullid | |
385 | if len(parents) > 0: |
|
385 | if len(parents) > 0: | |
386 | p1 = parents[0].filenode() |
|
386 | p1 = parents[0].filenode() | |
387 | if len(parents) > 1: |
|
387 | if len(parents) > 1: | |
388 | p2 = parents[1].filenode() |
|
388 | p2 = parents[1].filenode() | |
389 |
|
389 | |||
390 | copyname = b"" |
|
390 | copyname = b"" | |
391 | rename = ancestorctx.renamed() |
|
391 | rename = ancestorctx.renamed() | |
392 | if rename: |
|
392 | if rename: | |
393 | copyname = rename[0] |
|
393 | copyname = rename[0] | |
394 | linknode = ancestorctx.node() |
|
394 | linknode = ancestorctx.node() | |
395 | ancestortext += b"%s%s%s%s%s\0" % ( |
|
395 | ancestortext += b"%s%s%s%s%s\0" % ( | |
396 | ancestorctx.filenode(), |
|
396 | ancestorctx.filenode(), | |
397 | p1, |
|
397 | p1, | |
398 | p2, |
|
398 | p2, | |
399 | linknode, |
|
399 | linknode, | |
400 | copyname, |
|
400 | copyname, | |
401 | ) |
|
401 | ) | |
402 | finally: |
|
402 | finally: | |
403 | repo.forcelinkrev = False |
|
403 | repo.forcelinkrev = False | |
404 |
|
404 | |||
405 | header = shallowutil.buildfileblobheader(len(text), revlogflags) |
|
405 | header = shallowutil.buildfileblobheader(len(text), revlogflags) | |
406 |
|
406 | |||
407 | return b"%s\0%s%s" % (header, text, ancestortext) |
|
407 | return b"%s\0%s%s" % (header, text, ancestortext) | |
408 |
|
408 | |||
409 |
|
409 | |||
410 | def gcserver(ui, repo): |
|
410 | def gcserver(ui, repo): | |
411 | if not repo.ui.configbool(b"remotefilelog", b"server"): |
|
411 | if not repo.ui.configbool(b"remotefilelog", b"server"): | |
412 | return |
|
412 | return | |
413 |
|
413 | |||
414 | neededfiles = set() |
|
414 | neededfiles = set() | |
415 | heads = repo.revs(b"heads(tip~25000:) - null") |
|
415 | heads = repo.revs(b"heads(tip~25000:) - null") | |
416 |
|
416 | |||
417 | cachepath = repo.vfs.join(b"remotefilelogcache") |
|
417 | cachepath = repo.vfs.join(b"remotefilelogcache") | |
418 | for head in heads: |
|
418 | for head in heads: | |
419 | mf = repo[head].manifest() |
|
419 | mf = repo[head].manifest() | |
420 | for filename, filenode in pycompat.iteritems(mf): |
|
420 | for filename, filenode in pycompat.iteritems(mf): | |
421 | filecachepath = os.path.join(cachepath, filename, hex(filenode)) |
|
421 | filecachepath = os.path.join(cachepath, filename, hex(filenode)) | |
422 | neededfiles.add(filecachepath) |
|
422 | neededfiles.add(filecachepath) | |
423 |
|
423 | |||
424 | # delete unneeded older files |
|
424 | # delete unneeded older files | |
425 | days = repo.ui.configint(b"remotefilelog", b"serverexpiration") |
|
425 | days = repo.ui.configint(b"remotefilelog", b"serverexpiration") | |
426 | expiration = time.time() - (days * 24 * 60 * 60) |
|
426 | expiration = time.time() - (days * 24 * 60 * 60) | |
427 |
|
427 | |||
428 | progress = ui.makeprogress(_(b"removing old server cache"), unit=b"files") |
|
428 | progress = ui.makeprogress(_(b"removing old server cache"), unit=b"files") | |
429 | progress.update(0) |
|
429 | progress.update(0) | |
430 | for root, dirs, files in os.walk(cachepath): |
|
430 | for root, dirs, files in os.walk(cachepath): | |
431 | for file in files: |
|
431 | for file in files: | |
432 | filepath = os.path.join(root, file) |
|
432 | filepath = os.path.join(root, file) | |
433 | progress.increment() |
|
433 | progress.increment() | |
434 | if filepath in neededfiles: |
|
434 | if filepath in neededfiles: | |
435 | continue |
|
435 | continue | |
436 |
|
436 | |||
437 | stat = os.stat(filepath) |
|
437 | stat = os.stat(filepath) | |
438 | if stat.st_mtime < expiration: |
|
438 | if stat.st_mtime < expiration: | |
439 | os.remove(filepath) |
|
439 | os.remove(filepath) | |
440 |
|
440 | |||
441 | progress.complete() |
|
441 | progress.complete() |
@@ -1,563 +1,563 b'' | |||||
1 | # repair.py - functions for repository repair for mercurial |
|
1 | # repair.py - functions for repository repair for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005, 2006 Chris Mason <mason@suse.com> |
|
3 | # Copyright 2005, 2006 Chris Mason <mason@suse.com> | |
4 | # Copyright 2007 Olivia Mackall |
|
4 | # Copyright 2007 Olivia Mackall | |
5 | # |
|
5 | # | |
6 | # This software may be used and distributed according to the terms of the |
|
6 | # This software may be used and distributed according to the terms of the | |
7 | # GNU General Public License version 2 or any later version. |
|
7 | # GNU General Public License version 2 or any later version. | |
8 |
|
8 | |||
9 | from __future__ import absolute_import |
|
9 | from __future__ import absolute_import | |
10 |
|
10 | |||
11 | import errno |
|
11 | import errno | |
12 |
|
12 | |||
13 | from .i18n import _ |
|
13 | from .i18n import _ | |
14 | from .node import ( |
|
14 | from .node import ( | |
15 | hex, |
|
15 | hex, | |
16 | short, |
|
16 | short, | |
17 | ) |
|
17 | ) | |
18 | from . import ( |
|
18 | from . import ( | |
19 | bundle2, |
|
19 | bundle2, | |
20 | changegroup, |
|
20 | changegroup, | |
21 | discovery, |
|
21 | discovery, | |
22 | error, |
|
22 | error, | |
23 | exchange, |
|
23 | exchange, | |
24 | obsolete, |
|
24 | obsolete, | |
25 | obsutil, |
|
25 | obsutil, | |
26 | pathutil, |
|
26 | pathutil, | |
27 | phases, |
|
27 | phases, | |
28 | pycompat, |
|
28 | pycompat, | |
29 | requirements, |
|
29 | requirements, | |
30 | scmutil, |
|
30 | scmutil, | |
31 | util, |
|
31 | util, | |
32 | ) |
|
32 | ) | |
33 | from .utils import ( |
|
33 | from .utils import ( | |
34 | hashutil, |
|
34 | hashutil, | |
35 | stringutil, |
|
35 | stringutil, | |
36 | urlutil, |
|
36 | urlutil, | |
37 | ) |
|
37 | ) | |
38 |
|
38 | |||
39 |
|
39 | |||
40 | def backupbundle( |
|
40 | def backupbundle( | |
41 | repo, bases, heads, node, suffix, compress=True, obsolescence=True |
|
41 | repo, bases, heads, node, suffix, compress=True, obsolescence=True | |
42 | ): |
|
42 | ): | |
43 | """create a bundle with the specified revisions as a backup""" |
|
43 | """create a bundle with the specified revisions as a backup""" | |
44 |
|
44 | |||
45 | backupdir = b"strip-backup" |
|
45 | backupdir = b"strip-backup" | |
46 | vfs = repo.vfs |
|
46 | vfs = repo.vfs | |
47 | if not vfs.isdir(backupdir): |
|
47 | if not vfs.isdir(backupdir): | |
48 | vfs.mkdir(backupdir) |
|
48 | vfs.mkdir(backupdir) | |
49 |
|
49 | |||
50 | # Include a hash of all the nodes in the filename for uniqueness |
|
50 | # Include a hash of all the nodes in the filename for uniqueness | |
51 | allcommits = repo.set(b'%ln::%ln', bases, heads) |
|
51 | allcommits = repo.set(b'%ln::%ln', bases, heads) | |
52 | allhashes = sorted(c.hex() for c in allcommits) |
|
52 | allhashes = sorted(c.hex() for c in allcommits) | |
53 | totalhash = hashutil.sha1(b''.join(allhashes)).digest() |
|
53 | totalhash = hashutil.sha1(b''.join(allhashes)).digest() | |
54 | name = b"%s/%s-%s-%s.hg" % ( |
|
54 | name = b"%s/%s-%s-%s.hg" % ( | |
55 | backupdir, |
|
55 | backupdir, | |
56 | short(node), |
|
56 | short(node), | |
57 | hex(totalhash[:4]), |
|
57 | hex(totalhash[:4]), | |
58 | suffix, |
|
58 | suffix, | |
59 | ) |
|
59 | ) | |
60 |
|
60 | |||
61 | cgversion = changegroup.localversion(repo) |
|
61 | cgversion = changegroup.localversion(repo) | |
62 | comp = None |
|
62 | comp = None | |
63 | if cgversion != b'01': |
|
63 | if cgversion != b'01': | |
64 | bundletype = b"HG20" |
|
64 | bundletype = b"HG20" | |
65 | if compress: |
|
65 | if compress: | |
66 | comp = b'BZ' |
|
66 | comp = b'BZ' | |
67 | elif compress: |
|
67 | elif compress: | |
68 | bundletype = b"HG10BZ" |
|
68 | bundletype = b"HG10BZ" | |
69 | else: |
|
69 | else: | |
70 | bundletype = b"HG10UN" |
|
70 | bundletype = b"HG10UN" | |
71 |
|
71 | |||
72 | outgoing = discovery.outgoing(repo, missingroots=bases, ancestorsof=heads) |
|
72 | outgoing = discovery.outgoing(repo, missingroots=bases, ancestorsof=heads) | |
73 | contentopts = { |
|
73 | contentopts = { | |
74 | b'cg.version': cgversion, |
|
74 | b'cg.version': cgversion, | |
75 | b'obsolescence': obsolescence, |
|
75 | b'obsolescence': obsolescence, | |
76 | b'phases': True, |
|
76 | b'phases': True, | |
77 | } |
|
77 | } | |
78 | return bundle2.writenewbundle( |
|
78 | return bundle2.writenewbundle( | |
79 | repo.ui, |
|
79 | repo.ui, | |
80 | repo, |
|
80 | repo, | |
81 | b'strip', |
|
81 | b'strip', | |
82 | name, |
|
82 | name, | |
83 | bundletype, |
|
83 | bundletype, | |
84 | outgoing, |
|
84 | outgoing, | |
85 | contentopts, |
|
85 | contentopts, | |
86 | vfs, |
|
86 | vfs, | |
87 | compression=comp, |
|
87 | compression=comp, | |
88 | ) |
|
88 | ) | |
89 |
|
89 | |||
90 |
|
90 | |||
91 | def _collectfiles(repo, striprev): |
|
91 | def _collectfiles(repo, striprev): | |
92 | """find out the filelogs affected by the strip""" |
|
92 | """find out the filelogs affected by the strip""" | |
93 | files = set() |
|
93 | files = set() | |
94 |
|
94 | |||
95 | for x in pycompat.xrange(striprev, len(repo)): |
|
95 | for x in pycompat.xrange(striprev, len(repo)): | |
96 | files.update(repo[x].files()) |
|
96 | files.update(repo[x].files()) | |
97 |
|
97 | |||
98 | return sorted(files) |
|
98 | return sorted(files) | |
99 |
|
99 | |||
100 |
|
100 | |||
101 | def _collectrevlog(revlog, striprev): |
|
101 | def _collectrevlog(revlog, striprev): | |
102 | _, brokenset = revlog.getstrippoint(striprev) |
|
102 | _, brokenset = revlog.getstrippoint(striprev) | |
103 | return [revlog.linkrev(r) for r in brokenset] |
|
103 | return [revlog.linkrev(r) for r in brokenset] | |
104 |
|
104 | |||
105 |
|
105 | |||
106 | def _collectbrokencsets(repo, files, striprev): |
|
106 | def _collectbrokencsets(repo, files, striprev): | |
107 | """return the changesets which will be broken by the truncation""" |
|
107 | """return the changesets which will be broken by the truncation""" | |
108 | s = set() |
|
108 | s = set() | |
109 |
|
109 | |||
110 | for revlog in manifestrevlogs(repo): |
|
110 | for revlog in manifestrevlogs(repo): | |
111 | s.update(_collectrevlog(revlog, striprev)) |
|
111 | s.update(_collectrevlog(revlog, striprev)) | |
112 | for fname in files: |
|
112 | for fname in files: | |
113 | s.update(_collectrevlog(repo.file(fname), striprev)) |
|
113 | s.update(_collectrevlog(repo.file(fname), striprev)) | |
114 |
|
114 | |||
115 | return s |
|
115 | return s | |
116 |
|
116 | |||
117 |
|
117 | |||
118 | def strip(ui, repo, nodelist, backup=True, topic=b'backup'): |
|
118 | def strip(ui, repo, nodelist, backup=True, topic=b'backup'): | |
119 | # This function requires the caller to lock the repo, but it operates |
|
119 | # This function requires the caller to lock the repo, but it operates | |
120 | # within a transaction of its own, and thus requires there to be no current |
|
120 | # within a transaction of its own, and thus requires there to be no current | |
121 | # transaction when it is called. |
|
121 | # transaction when it is called. | |
122 | if repo.currenttransaction() is not None: |
|
122 | if repo.currenttransaction() is not None: | |
123 | raise error.ProgrammingError(b'cannot strip from inside a transaction') |
|
123 | raise error.ProgrammingError(b'cannot strip from inside a transaction') | |
124 |
|
124 | |||
125 | # Simple way to maintain backwards compatibility for this |
|
125 | # Simple way to maintain backwards compatibility for this | |
126 | # argument. |
|
126 | # argument. | |
127 | if backup in [b'none', b'strip']: |
|
127 | if backup in [b'none', b'strip']: | |
128 | backup = False |
|
128 | backup = False | |
129 |
|
129 | |||
130 | repo = repo.unfiltered() |
|
130 | repo = repo.unfiltered() | |
131 | repo.destroying() |
|
131 | repo.destroying() | |
132 | vfs = repo.vfs |
|
132 | vfs = repo.vfs | |
133 | # load bookmark before changelog to avoid side effect from outdated |
|
133 | # load bookmark before changelog to avoid side effect from outdated | |
134 | # changelog (see repo._refreshchangelog) |
|
134 | # changelog (see repo._refreshchangelog) | |
135 | repo._bookmarks |
|
135 | repo._bookmarks | |
136 | cl = repo.changelog |
|
136 | cl = repo.changelog | |
137 |
|
137 | |||
138 | # TODO handle undo of merge sets |
|
138 | # TODO handle undo of merge sets | |
139 | if isinstance(nodelist, bytes): |
|
139 | if isinstance(nodelist, bytes): | |
140 | nodelist = [nodelist] |
|
140 | nodelist = [nodelist] | |
141 | striplist = [cl.rev(node) for node in nodelist] |
|
141 | striplist = [cl.rev(node) for node in nodelist] | |
142 | striprev = min(striplist) |
|
142 | striprev = min(striplist) | |
143 |
|
143 | |||
144 | files = _collectfiles(repo, striprev) |
|
144 | files = _collectfiles(repo, striprev) | |
145 | saverevs = _collectbrokencsets(repo, files, striprev) |
|
145 | saverevs = _collectbrokencsets(repo, files, striprev) | |
146 |
|
146 | |||
147 | # Some revisions with rev > striprev may not be descendants of striprev. |
|
147 | # Some revisions with rev > striprev may not be descendants of striprev. | |
148 | # We have to find these revisions and put them in a bundle, so that |
|
148 | # We have to find these revisions and put them in a bundle, so that | |
149 | # we can restore them after the truncations. |
|
149 | # we can restore them after the truncations. | |
150 | # To create the bundle we use repo.changegroupsubset which requires |
|
150 | # To create the bundle we use repo.changegroupsubset which requires | |
151 | # the list of heads and bases of the set of interesting revisions. |
|
151 | # the list of heads and bases of the set of interesting revisions. | |
152 | # (head = revision in the set that has no descendant in the set; |
|
152 | # (head = revision in the set that has no descendant in the set; | |
153 | # base = revision in the set that has no ancestor in the set) |
|
153 | # base = revision in the set that has no ancestor in the set) | |
154 | tostrip = set(striplist) |
|
154 | tostrip = set(striplist) | |
155 | saveheads = set(saverevs) |
|
155 | saveheads = set(saverevs) | |
156 | for r in cl.revs(start=striprev + 1): |
|
156 | for r in cl.revs(start=striprev + 1): | |
157 | if any(p in tostrip for p in cl.parentrevs(r)): |
|
157 | if any(p in tostrip for p in cl.parentrevs(r)): | |
158 | tostrip.add(r) |
|
158 | tostrip.add(r) | |
159 |
|
159 | |||
160 | if r not in tostrip: |
|
160 | if r not in tostrip: | |
161 | saverevs.add(r) |
|
161 | saverevs.add(r) | |
162 | saveheads.difference_update(cl.parentrevs(r)) |
|
162 | saveheads.difference_update(cl.parentrevs(r)) | |
163 | saveheads.add(r) |
|
163 | saveheads.add(r) | |
164 | saveheads = [cl.node(r) for r in saveheads] |
|
164 | saveheads = [cl.node(r) for r in saveheads] | |
165 |
|
165 | |||
166 | # compute base nodes |
|
166 | # compute base nodes | |
167 | if saverevs: |
|
167 | if saverevs: | |
168 | descendants = set(cl.descendants(saverevs)) |
|
168 | descendants = set(cl.descendants(saverevs)) | |
169 | saverevs.difference_update(descendants) |
|
169 | saverevs.difference_update(descendants) | |
170 | savebases = [cl.node(r) for r in saverevs] |
|
170 | savebases = [cl.node(r) for r in saverevs] | |
171 | stripbases = [cl.node(r) for r in tostrip] |
|
171 | stripbases = [cl.node(r) for r in tostrip] | |
172 |
|
172 | |||
173 | stripobsidx = obsmarkers = () |
|
173 | stripobsidx = obsmarkers = () | |
174 | if repo.ui.configbool(b'devel', b'strip-obsmarkers'): |
|
174 | if repo.ui.configbool(b'devel', b'strip-obsmarkers'): | |
175 | obsmarkers = obsutil.exclusivemarkers(repo, stripbases) |
|
175 | obsmarkers = obsutil.exclusivemarkers(repo, stripbases) | |
176 | if obsmarkers: |
|
176 | if obsmarkers: | |
177 | stripobsidx = [ |
|
177 | stripobsidx = [ | |
178 | i for i, m in enumerate(repo.obsstore) if m in obsmarkers |
|
178 | i for i, m in enumerate(repo.obsstore) if m in obsmarkers | |
179 | ] |
|
179 | ] | |
180 |
|
180 | |||
181 | newbmtarget, updatebm = _bookmarkmovements(repo, tostrip) |
|
181 | newbmtarget, updatebm = _bookmarkmovements(repo, tostrip) | |
182 |
|
182 | |||
183 | backupfile = None |
|
183 | backupfile = None | |
184 | node = nodelist[-1] |
|
184 | node = nodelist[-1] | |
185 | if backup: |
|
185 | if backup: | |
186 | backupfile = _createstripbackup(repo, stripbases, node, topic) |
|
186 | backupfile = _createstripbackup(repo, stripbases, node, topic) | |
187 | # create a changegroup for all the branches we need to keep |
|
187 | # create a changegroup for all the branches we need to keep | |
188 | tmpbundlefile = None |
|
188 | tmpbundlefile = None | |
189 | if saveheads: |
|
189 | if saveheads: | |
190 | # do not compress temporary bundle if we remove it from disk later |
|
190 | # do not compress temporary bundle if we remove it from disk later | |
191 | # |
|
191 | # | |
192 | # We do not include obsolescence, it might re-introduce prune markers |
|
192 | # We do not include obsolescence, it might re-introduce prune markers | |
193 | # we are trying to strip. This is harmless since the stripped markers |
|
193 | # we are trying to strip. This is harmless since the stripped markers | |
194 | # are already backed up and we did not touched the markers for the |
|
194 | # are already backed up and we did not touched the markers for the | |
195 | # saved changesets. |
|
195 | # saved changesets. | |
196 | tmpbundlefile = backupbundle( |
|
196 | tmpbundlefile = backupbundle( | |
197 | repo, |
|
197 | repo, | |
198 | savebases, |
|
198 | savebases, | |
199 | saveheads, |
|
199 | saveheads, | |
200 | node, |
|
200 | node, | |
201 | b'temp', |
|
201 | b'temp', | |
202 | compress=False, |
|
202 | compress=False, | |
203 | obsolescence=False, |
|
203 | obsolescence=False, | |
204 | ) |
|
204 | ) | |
205 |
|
205 | |||
206 | with ui.uninterruptible(): |
|
206 | with ui.uninterruptible(): | |
207 | try: |
|
207 | try: | |
208 | with repo.transaction(b"strip") as tr: |
|
208 | with repo.transaction(b"strip") as tr: | |
209 | # TODO this code violates the interface abstraction of the |
|
209 | # TODO this code violates the interface abstraction of the | |
210 | # transaction and makes assumptions that file storage is |
|
210 | # transaction and makes assumptions that file storage is | |
211 | # using append-only files. We'll need some kind of storage |
|
211 | # using append-only files. We'll need some kind of storage | |
212 | # API to handle stripping for us. |
|
212 | # API to handle stripping for us. | |
213 | oldfiles = set(tr._offsetmap.keys()) |
|
213 | oldfiles = set(tr._offsetmap.keys()) | |
214 | oldfiles.update(tr._newfiles) |
|
214 | oldfiles.update(tr._newfiles) | |
215 |
|
215 | |||
216 | tr.startgroup() |
|
216 | tr.startgroup() | |
217 | cl.strip(striprev, tr) |
|
217 | cl.strip(striprev, tr) | |
218 | stripmanifest(repo, striprev, tr, files) |
|
218 | stripmanifest(repo, striprev, tr, files) | |
219 |
|
219 | |||
220 | for fn in files: |
|
220 | for fn in files: | |
221 | repo.file(fn).strip(striprev, tr) |
|
221 | repo.file(fn).strip(striprev, tr) | |
222 | tr.endgroup() |
|
222 | tr.endgroup() | |
223 |
|
223 | |||
224 | entries = tr.readjournal() |
|
224 | entries = tr.readjournal() | |
225 |
|
225 | |||
226 | for file, troffset in entries: |
|
226 | for file, troffset in entries: | |
227 | if file in oldfiles: |
|
227 | if file in oldfiles: | |
228 | continue |
|
228 | continue | |
229 | with repo.svfs(file, b'a', checkambig=True) as fp: |
|
229 | with repo.svfs(file, b'a', checkambig=True) as fp: | |
230 | fp.truncate(troffset) |
|
230 | fp.truncate(troffset) | |
231 | if troffset == 0: |
|
231 | if troffset == 0: | |
232 | repo.store.markremoved(file) |
|
232 | repo.store.markremoved(file) | |
233 |
|
233 | |||
234 | deleteobsmarkers(repo.obsstore, stripobsidx) |
|
234 | deleteobsmarkers(repo.obsstore, stripobsidx) | |
235 | del repo.obsstore |
|
235 | del repo.obsstore | |
236 | repo.invalidatevolatilesets() |
|
236 | repo.invalidatevolatilesets() | |
237 | repo._phasecache.filterunknown(repo) |
|
237 | repo._phasecache.filterunknown(repo) | |
238 |
|
238 | |||
239 | if tmpbundlefile: |
|
239 | if tmpbundlefile: | |
240 | ui.note(_(b"adding branch\n")) |
|
240 | ui.note(_(b"adding branch\n")) | |
241 | f = vfs.open(tmpbundlefile, b"rb") |
|
241 | f = vfs.open(tmpbundlefile, b"rb") | |
242 | gen = exchange.readbundle(ui, f, tmpbundlefile, vfs) |
|
242 | gen = exchange.readbundle(ui, f, tmpbundlefile, vfs) | |
243 | # silence internal shuffling chatter |
|
243 | # silence internal shuffling chatter | |
244 | maybe_silent = ( |
|
244 | maybe_silent = ( | |
245 | repo.ui.silent() |
|
245 | repo.ui.silent() | |
246 | if not repo.ui.verbose |
|
246 | if not repo.ui.verbose | |
247 | else util.nullcontextmanager() |
|
247 | else util.nullcontextmanager() | |
248 | ) |
|
248 | ) | |
249 | with maybe_silent: |
|
249 | with maybe_silent: | |
250 | tmpbundleurl = b'bundle:' + vfs.join(tmpbundlefile) |
|
250 | tmpbundleurl = b'bundle:' + vfs.join(tmpbundlefile) | |
251 | txnname = b'strip' |
|
251 | txnname = b'strip' | |
252 | if not isinstance(gen, bundle2.unbundle20): |
|
252 | if not isinstance(gen, bundle2.unbundle20): | |
253 | txnname = b"strip\n%s" % urlutil.hidepassword( |
|
253 | txnname = b"strip\n%s" % urlutil.hidepassword( | |
254 | tmpbundleurl |
|
254 | tmpbundleurl | |
255 | ) |
|
255 | ) | |
256 | with repo.transaction(txnname) as tr: |
|
256 | with repo.transaction(txnname) as tr: | |
257 | bundle2.applybundle( |
|
257 | bundle2.applybundle( | |
258 | repo, gen, tr, source=b'strip', url=tmpbundleurl |
|
258 | repo, gen, tr, source=b'strip', url=tmpbundleurl | |
259 | ) |
|
259 | ) | |
260 | f.close() |
|
260 | f.close() | |
261 |
|
261 | |||
262 | with repo.transaction(b'repair') as tr: |
|
262 | with repo.transaction(b'repair') as tr: | |
263 | bmchanges = [(m, repo[newbmtarget].node()) for m in updatebm] |
|
263 | bmchanges = [(m, repo[newbmtarget].node()) for m in updatebm] | |
264 | repo._bookmarks.applychanges(repo, tr, bmchanges) |
|
264 | repo._bookmarks.applychanges(repo, tr, bmchanges) | |
265 |
|
265 | |||
266 | # remove undo files |
|
266 | # remove undo files | |
267 | for undovfs, undofile in repo.undofiles(): |
|
267 | for undovfs, undofile in repo.undofiles(): | |
268 | try: |
|
268 | try: | |
269 | undovfs.unlink(undofile) |
|
269 | undovfs.unlink(undofile) | |
270 | except OSError as e: |
|
270 | except OSError as e: | |
271 | if e.errno != errno.ENOENT: |
|
271 | if e.errno != errno.ENOENT: | |
272 | ui.warn( |
|
272 | ui.warn( | |
273 | _(b'error removing %s: %s\n') |
|
273 | _(b'error removing %s: %s\n') | |
274 | % ( |
|
274 | % ( | |
275 | undovfs.join(undofile), |
|
275 | undovfs.join(undofile), | |
276 | stringutil.forcebytestr(e), |
|
276 | stringutil.forcebytestr(e), | |
277 | ) |
|
277 | ) | |
278 | ) |
|
278 | ) | |
279 |
|
279 | |||
280 | except: # re-raises |
|
280 | except: # re-raises | |
281 | if backupfile: |
|
281 | if backupfile: | |
282 | ui.warn( |
|
282 | ui.warn( | |
283 | _(b"strip failed, backup bundle stored in '%s'\n") |
|
283 | _(b"strip failed, backup bundle stored in '%s'\n") | |
284 | % vfs.join(backupfile) |
|
284 | % vfs.join(backupfile) | |
285 | ) |
|
285 | ) | |
286 | if tmpbundlefile: |
|
286 | if tmpbundlefile: | |
287 | ui.warn( |
|
287 | ui.warn( | |
288 | _(b"strip failed, unrecovered changes stored in '%s'\n") |
|
288 | _(b"strip failed, unrecovered changes stored in '%s'\n") | |
289 | % vfs.join(tmpbundlefile) |
|
289 | % vfs.join(tmpbundlefile) | |
290 | ) |
|
290 | ) | |
291 | ui.warn( |
|
291 | ui.warn( | |
292 | _( |
|
292 | _( | |
293 | b"(fix the problem, then recover the changesets with " |
|
293 | b"(fix the problem, then recover the changesets with " | |
294 | b"\"hg unbundle '%s'\")\n" |
|
294 | b"\"hg unbundle '%s'\")\n" | |
295 | ) |
|
295 | ) | |
296 | % vfs.join(tmpbundlefile) |
|
296 | % vfs.join(tmpbundlefile) | |
297 | ) |
|
297 | ) | |
298 | raise |
|
298 | raise | |
299 | else: |
|
299 | else: | |
300 | if tmpbundlefile: |
|
300 | if tmpbundlefile: | |
301 | # Remove temporary bundle only if there were no exceptions |
|
301 | # Remove temporary bundle only if there were no exceptions | |
302 | vfs.unlink(tmpbundlefile) |
|
302 | vfs.unlink(tmpbundlefile) | |
303 |
|
303 | |||
304 | repo.destroyed() |
|
304 | repo.destroyed() | |
305 | # return the backup file path (or None if 'backup' was False) so |
|
305 | # return the backup file path (or None if 'backup' was False) so | |
306 | # extensions can use it |
|
306 | # extensions can use it | |
307 | return backupfile |
|
307 | return backupfile | |
308 |
|
308 | |||
309 |
|
309 | |||
310 | def softstrip(ui, repo, nodelist, backup=True, topic=b'backup'): |
|
310 | def softstrip(ui, repo, nodelist, backup=True, topic=b'backup'): | |
311 | """perform a "soft" strip using the archived phase""" |
|
311 | """perform a "soft" strip using the archived phase""" | |
312 | tostrip = [c.node() for c in repo.set(b'sort(%ln::)', nodelist)] |
|
312 | tostrip = [c.node() for c in repo.set(b'sort(%ln::)', nodelist)] | |
313 | if not tostrip: |
|
313 | if not tostrip: | |
314 | return None |
|
314 | return None | |
315 |
|
315 | |||
316 | backupfile = None |
|
316 | backupfile = None | |
317 | if backup: |
|
317 | if backup: | |
318 | node = tostrip[0] |
|
318 | node = tostrip[0] | |
319 | backupfile = _createstripbackup(repo, tostrip, node, topic) |
|
319 | backupfile = _createstripbackup(repo, tostrip, node, topic) | |
320 |
|
320 | |||
321 | newbmtarget, updatebm = _bookmarkmovements(repo, tostrip) |
|
321 | newbmtarget, updatebm = _bookmarkmovements(repo, tostrip) | |
322 | with repo.transaction(b'strip') as tr: |
|
322 | with repo.transaction(b'strip') as tr: | |
323 | phases.retractboundary(repo, tr, phases.archived, tostrip) |
|
323 | phases.retractboundary(repo, tr, phases.archived, tostrip) | |
324 | bmchanges = [(m, repo[newbmtarget].node()) for m in updatebm] |
|
324 | bmchanges = [(m, repo[newbmtarget].node()) for m in updatebm] | |
325 | repo._bookmarks.applychanges(repo, tr, bmchanges) |
|
325 | repo._bookmarks.applychanges(repo, tr, bmchanges) | |
326 | return backupfile |
|
326 | return backupfile | |
327 |
|
327 | |||
328 |
|
328 | |||
329 | def _bookmarkmovements(repo, tostrip): |
|
329 | def _bookmarkmovements(repo, tostrip): | |
330 | # compute necessary bookmark movement |
|
330 | # compute necessary bookmark movement | |
331 | bm = repo._bookmarks |
|
331 | bm = repo._bookmarks | |
332 | updatebm = [] |
|
332 | updatebm = [] | |
333 | for m in bm: |
|
333 | for m in bm: | |
334 | rev = repo[bm[m]].rev() |
|
334 | rev = repo[bm[m]].rev() | |
335 | if rev in tostrip: |
|
335 | if rev in tostrip: | |
336 | updatebm.append(m) |
|
336 | updatebm.append(m) | |
337 | newbmtarget = None |
|
337 | newbmtarget = None | |
338 | # If we need to move bookmarks, compute bookmark |
|
338 | # If we need to move bookmarks, compute bookmark | |
339 | # targets. Otherwise we can skip doing this logic. |
|
339 | # targets. Otherwise we can skip doing this logic. | |
340 | if updatebm: |
|
340 | if updatebm: | |
341 | # For a set s, max(parents(s) - s) is the same as max(heads(::s - s)), |
|
341 | # For a set s, max(parents(s) - s) is the same as max(heads(::s - s)), | |
342 | # but is much faster |
|
342 | # but is much faster | |
343 | newbmtarget = repo.revs(b'max(parents(%ld) - (%ld))', tostrip, tostrip) |
|
343 | newbmtarget = repo.revs(b'max(parents(%ld) - (%ld))', tostrip, tostrip) | |
344 | if newbmtarget: |
|
344 | if newbmtarget: | |
345 | newbmtarget = repo[newbmtarget.first()].node() |
|
345 | newbmtarget = repo[newbmtarget.first()].node() | |
346 | else: |
|
346 | else: | |
347 | newbmtarget = b'.' |
|
347 | newbmtarget = b'.' | |
348 | return newbmtarget, updatebm |
|
348 | return newbmtarget, updatebm | |
349 |
|
349 | |||
350 |
|
350 | |||
351 | def _createstripbackup(repo, stripbases, node, topic): |
|
351 | def _createstripbackup(repo, stripbases, node, topic): | |
352 | # backup the changeset we are about to strip |
|
352 | # backup the changeset we are about to strip | |
353 | vfs = repo.vfs |
|
353 | vfs = repo.vfs | |
354 | cl = repo.changelog |
|
354 | cl = repo.changelog | |
355 | backupfile = backupbundle(repo, stripbases, cl.heads(), node, topic) |
|
355 | backupfile = backupbundle(repo, stripbases, cl.heads(), node, topic) | |
356 | repo.ui.status(_(b"saved backup bundle to %s\n") % vfs.join(backupfile)) |
|
356 | repo.ui.status(_(b"saved backup bundle to %s\n") % vfs.join(backupfile)) | |
357 | repo.ui.log( |
|
357 | repo.ui.log( | |
358 | b"backupbundle", b"saved backup bundle to %s\n", vfs.join(backupfile) |
|
358 | b"backupbundle", b"saved backup bundle to %s\n", vfs.join(backupfile) | |
359 | ) |
|
359 | ) | |
360 | return backupfile |
|
360 | return backupfile | |
361 |
|
361 | |||
362 |
|
362 | |||
363 | def safestriproots(ui, repo, nodes): |
|
363 | def safestriproots(ui, repo, nodes): | |
364 | """return list of roots of nodes where descendants are covered by nodes""" |
|
364 | """return list of roots of nodes where descendants are covered by nodes""" | |
365 | torev = repo.unfiltered().changelog.rev |
|
365 | torev = repo.unfiltered().changelog.rev | |
366 | revs = {torev(n) for n in nodes} |
|
366 | revs = {torev(n) for n in nodes} | |
367 | # tostrip = wanted - unsafe = wanted - ancestors(orphaned) |
|
367 | # tostrip = wanted - unsafe = wanted - ancestors(orphaned) | |
368 | # orphaned = affected - wanted |
|
368 | # orphaned = affected - wanted | |
369 | # affected = descendants(roots(wanted)) |
|
369 | # affected = descendants(roots(wanted)) | |
370 | # wanted = revs |
|
370 | # wanted = revs | |
371 | revset = b'%ld - ( ::( (roots(%ld):: and not _phase(%s)) -%ld) )' |
|
371 | revset = b'%ld - ( ::( (roots(%ld):: and not _phase(%s)) -%ld) )' | |
372 | tostrip = set(repo.revs(revset, revs, revs, phases.internal, revs)) |
|
372 | tostrip = set(repo.revs(revset, revs, revs, phases.internal, revs)) | |
373 | notstrip = revs - tostrip |
|
373 | notstrip = revs - tostrip | |
374 | if notstrip: |
|
374 | if notstrip: | |
375 | nodestr = b', '.join(sorted(short(repo[n].node()) for n in notstrip)) |
|
375 | nodestr = b', '.join(sorted(short(repo[n].node()) for n in notstrip)) | |
376 | ui.warn( |
|
376 | ui.warn( | |
377 | _(b'warning: orphaned descendants detected, not stripping %s\n') |
|
377 | _(b'warning: orphaned descendants detected, not stripping %s\n') | |
378 | % nodestr |
|
378 | % nodestr | |
379 | ) |
|
379 | ) | |
380 | return [c.node() for c in repo.set(b'roots(%ld)', tostrip)] |
|
380 | return [c.node() for c in repo.set(b'roots(%ld)', tostrip)] | |
381 |
|
381 | |||
382 |
|
382 | |||
383 | class stripcallback(object): |
|
383 | class stripcallback(object): | |
384 | """used as a transaction postclose callback""" |
|
384 | """used as a transaction postclose callback""" | |
385 |
|
385 | |||
386 | def __init__(self, ui, repo, backup, topic): |
|
386 | def __init__(self, ui, repo, backup, topic): | |
387 | self.ui = ui |
|
387 | self.ui = ui | |
388 | self.repo = repo |
|
388 | self.repo = repo | |
389 | self.backup = backup |
|
389 | self.backup = backup | |
390 | self.topic = topic or b'backup' |
|
390 | self.topic = topic or b'backup' | |
391 | self.nodelist = [] |
|
391 | self.nodelist = [] | |
392 |
|
392 | |||
393 | def addnodes(self, nodes): |
|
393 | def addnodes(self, nodes): | |
394 | self.nodelist.extend(nodes) |
|
394 | self.nodelist.extend(nodes) | |
395 |
|
395 | |||
396 | def __call__(self, tr): |
|
396 | def __call__(self, tr): | |
397 | roots = safestriproots(self.ui, self.repo, self.nodelist) |
|
397 | roots = safestriproots(self.ui, self.repo, self.nodelist) | |
398 | if roots: |
|
398 | if roots: | |
399 | strip(self.ui, self.repo, roots, self.backup, self.topic) |
|
399 | strip(self.ui, self.repo, roots, self.backup, self.topic) | |
400 |
|
400 | |||
401 |
|
401 | |||
402 | def delayedstrip(ui, repo, nodelist, topic=None, backup=True): |
|
402 | def delayedstrip(ui, repo, nodelist, topic=None, backup=True): | |
403 | """like strip, but works inside transaction and won't strip irreverent revs |
|
403 | """like strip, but works inside transaction and won't strip irreverent revs | |
404 |
|
404 | |||
405 | nodelist must explicitly contain all descendants. Otherwise a warning will |
|
405 | nodelist must explicitly contain all descendants. Otherwise a warning will | |
406 | be printed that some nodes are not stripped. |
|
406 | be printed that some nodes are not stripped. | |
407 |
|
407 | |||
408 | Will do a backup if `backup` is True. The last non-None "topic" will be |
|
408 | Will do a backup if `backup` is True. The last non-None "topic" will be | |
409 | used as the backup topic name. The default backup topic name is "backup". |
|
409 | used as the backup topic name. The default backup topic name is "backup". | |
410 | """ |
|
410 | """ | |
411 | tr = repo.currenttransaction() |
|
411 | tr = repo.currenttransaction() | |
412 | if not tr: |
|
412 | if not tr: | |
413 | nodes = safestriproots(ui, repo, nodelist) |
|
413 | nodes = safestriproots(ui, repo, nodelist) | |
414 | return strip(ui, repo, nodes, backup=backup, topic=topic) |
|
414 | return strip(ui, repo, nodes, backup=backup, topic=topic) | |
415 | # transaction postclose callbacks are called in alphabet order. |
|
415 | # transaction postclose callbacks are called in alphabet order. | |
416 | # use '\xff' as prefix so we are likely to be called last. |
|
416 | # use '\xff' as prefix so we are likely to be called last. | |
417 | callback = tr.getpostclose(b'\xffstrip') |
|
417 | callback = tr.getpostclose(b'\xffstrip') | |
418 | if callback is None: |
|
418 | if callback is None: | |
419 | callback = stripcallback(ui, repo, backup=backup, topic=topic) |
|
419 | callback = stripcallback(ui, repo, backup=backup, topic=topic) | |
420 | tr.addpostclose(b'\xffstrip', callback) |
|
420 | tr.addpostclose(b'\xffstrip', callback) | |
421 | if topic: |
|
421 | if topic: | |
422 | callback.topic = topic |
|
422 | callback.topic = topic | |
423 | callback.addnodes(nodelist) |
|
423 | callback.addnodes(nodelist) | |
424 |
|
424 | |||
425 |
|
425 | |||
426 | def stripmanifest(repo, striprev, tr, files): |
|
426 | def stripmanifest(repo, striprev, tr, files): | |
427 | for revlog in manifestrevlogs(repo): |
|
427 | for revlog in manifestrevlogs(repo): | |
428 | revlog.strip(striprev, tr) |
|
428 | revlog.strip(striprev, tr) | |
429 |
|
429 | |||
430 |
|
430 | |||
431 | def manifestrevlogs(repo): |
|
431 | def manifestrevlogs(repo): | |
432 | yield repo.manifestlog.getstorage(b'') |
|
432 | yield repo.manifestlog.getstorage(b'') | |
433 | if scmutil.istreemanifest(repo): |
|
433 | if scmutil.istreemanifest(repo): | |
434 | # This logic is safe if treemanifest isn't enabled, but also |
|
434 | # This logic is safe if treemanifest isn't enabled, but also | |
435 | # pointless, so we skip it if treemanifest isn't enabled. |
|
435 | # pointless, so we skip it if treemanifest isn't enabled. | |
436 |
for t, unencoded |
|
436 | for t, unencoded, size in repo.store.datafiles(): | |
437 | if unencoded.startswith(b'meta/') and unencoded.endswith( |
|
437 | if unencoded.startswith(b'meta/') and unencoded.endswith( | |
438 | b'00manifest.i' |
|
438 | b'00manifest.i' | |
439 | ): |
|
439 | ): | |
440 | dir = unencoded[5:-12] |
|
440 | dir = unencoded[5:-12] | |
441 | yield repo.manifestlog.getstorage(dir) |
|
441 | yield repo.manifestlog.getstorage(dir) | |
442 |
|
442 | |||
443 |
|
443 | |||
444 | def rebuildfncache(ui, repo, only_data=False): |
|
444 | def rebuildfncache(ui, repo, only_data=False): | |
445 | """Rebuilds the fncache file from repo history. |
|
445 | """Rebuilds the fncache file from repo history. | |
446 |
|
446 | |||
447 | Missing entries will be added. Extra entries will be removed. |
|
447 | Missing entries will be added. Extra entries will be removed. | |
448 | """ |
|
448 | """ | |
449 | repo = repo.unfiltered() |
|
449 | repo = repo.unfiltered() | |
450 |
|
450 | |||
451 | if requirements.FNCACHE_REQUIREMENT not in repo.requirements: |
|
451 | if requirements.FNCACHE_REQUIREMENT not in repo.requirements: | |
452 | ui.warn( |
|
452 | ui.warn( | |
453 | _( |
|
453 | _( | |
454 | b'(not rebuilding fncache because repository does not ' |
|
454 | b'(not rebuilding fncache because repository does not ' | |
455 | b'support fncache)\n' |
|
455 | b'support fncache)\n' | |
456 | ) |
|
456 | ) | |
457 | ) |
|
457 | ) | |
458 | return |
|
458 | return | |
459 |
|
459 | |||
460 | with repo.lock(): |
|
460 | with repo.lock(): | |
461 | fnc = repo.store.fncache |
|
461 | fnc = repo.store.fncache | |
462 | fnc.ensureloaded(warn=ui.warn) |
|
462 | fnc.ensureloaded(warn=ui.warn) | |
463 |
|
463 | |||
464 | oldentries = set(fnc.entries) |
|
464 | oldentries = set(fnc.entries) | |
465 | newentries = set() |
|
465 | newentries = set() | |
466 | seenfiles = set() |
|
466 | seenfiles = set() | |
467 |
|
467 | |||
468 | if only_data: |
|
468 | if only_data: | |
469 | # Trust the listing of .i from the fncache, but not the .d. This is |
|
469 | # Trust the listing of .i from the fncache, but not the .d. This is | |
470 | # much faster, because we only need to stat every possible .d files, |
|
470 | # much faster, because we only need to stat every possible .d files, | |
471 | # instead of reading the full changelog |
|
471 | # instead of reading the full changelog | |
472 | for f in fnc: |
|
472 | for f in fnc: | |
473 | if f[:5] == b'data/' and f[-2:] == b'.i': |
|
473 | if f[:5] == b'data/' and f[-2:] == b'.i': | |
474 | seenfiles.add(f[5:-2]) |
|
474 | seenfiles.add(f[5:-2]) | |
475 | newentries.add(f) |
|
475 | newentries.add(f) | |
476 | dataf = f[:-2] + b'.d' |
|
476 | dataf = f[:-2] + b'.d' | |
477 | if repo.store._exists(dataf): |
|
477 | if repo.store._exists(dataf): | |
478 | newentries.add(dataf) |
|
478 | newentries.add(dataf) | |
479 | else: |
|
479 | else: | |
480 | progress = ui.makeprogress( |
|
480 | progress = ui.makeprogress( | |
481 | _(b'rebuilding'), unit=_(b'changesets'), total=len(repo) |
|
481 | _(b'rebuilding'), unit=_(b'changesets'), total=len(repo) | |
482 | ) |
|
482 | ) | |
483 | for rev in repo: |
|
483 | for rev in repo: | |
484 | progress.update(rev) |
|
484 | progress.update(rev) | |
485 |
|
485 | |||
486 | ctx = repo[rev] |
|
486 | ctx = repo[rev] | |
487 | for f in ctx.files(): |
|
487 | for f in ctx.files(): | |
488 | # This is to minimize I/O. |
|
488 | # This is to minimize I/O. | |
489 | if f in seenfiles: |
|
489 | if f in seenfiles: | |
490 | continue |
|
490 | continue | |
491 | seenfiles.add(f) |
|
491 | seenfiles.add(f) | |
492 |
|
492 | |||
493 | i = b'data/%s.i' % f |
|
493 | i = b'data/%s.i' % f | |
494 | d = b'data/%s.d' % f |
|
494 | d = b'data/%s.d' % f | |
495 |
|
495 | |||
496 | if repo.store._exists(i): |
|
496 | if repo.store._exists(i): | |
497 | newentries.add(i) |
|
497 | newentries.add(i) | |
498 | if repo.store._exists(d): |
|
498 | if repo.store._exists(d): | |
499 | newentries.add(d) |
|
499 | newentries.add(d) | |
500 |
|
500 | |||
501 | progress.complete() |
|
501 | progress.complete() | |
502 |
|
502 | |||
503 | if requirements.TREEMANIFEST_REQUIREMENT in repo.requirements: |
|
503 | if requirements.TREEMANIFEST_REQUIREMENT in repo.requirements: | |
504 | # This logic is safe if treemanifest isn't enabled, but also |
|
504 | # This logic is safe if treemanifest isn't enabled, but also | |
505 | # pointless, so we skip it if treemanifest isn't enabled. |
|
505 | # pointless, so we skip it if treemanifest isn't enabled. | |
506 | for dir in pathutil.dirs(seenfiles): |
|
506 | for dir in pathutil.dirs(seenfiles): | |
507 | i = b'meta/%s/00manifest.i' % dir |
|
507 | i = b'meta/%s/00manifest.i' % dir | |
508 | d = b'meta/%s/00manifest.d' % dir |
|
508 | d = b'meta/%s/00manifest.d' % dir | |
509 |
|
509 | |||
510 | if repo.store._exists(i): |
|
510 | if repo.store._exists(i): | |
511 | newentries.add(i) |
|
511 | newentries.add(i) | |
512 | if repo.store._exists(d): |
|
512 | if repo.store._exists(d): | |
513 | newentries.add(d) |
|
513 | newentries.add(d) | |
514 |
|
514 | |||
515 | addcount = len(newentries - oldentries) |
|
515 | addcount = len(newentries - oldentries) | |
516 | removecount = len(oldentries - newentries) |
|
516 | removecount = len(oldentries - newentries) | |
517 | for p in sorted(oldentries - newentries): |
|
517 | for p in sorted(oldentries - newentries): | |
518 | ui.write(_(b'removing %s\n') % p) |
|
518 | ui.write(_(b'removing %s\n') % p) | |
519 | for p in sorted(newentries - oldentries): |
|
519 | for p in sorted(newentries - oldentries): | |
520 | ui.write(_(b'adding %s\n') % p) |
|
520 | ui.write(_(b'adding %s\n') % p) | |
521 |
|
521 | |||
522 | if addcount or removecount: |
|
522 | if addcount or removecount: | |
523 | ui.write( |
|
523 | ui.write( | |
524 | _(b'%d items added, %d removed from fncache\n') |
|
524 | _(b'%d items added, %d removed from fncache\n') | |
525 | % (addcount, removecount) |
|
525 | % (addcount, removecount) | |
526 | ) |
|
526 | ) | |
527 | fnc.entries = newentries |
|
527 | fnc.entries = newentries | |
528 | fnc._dirty = True |
|
528 | fnc._dirty = True | |
529 |
|
529 | |||
530 | with repo.transaction(b'fncache') as tr: |
|
530 | with repo.transaction(b'fncache') as tr: | |
531 | fnc.write(tr) |
|
531 | fnc.write(tr) | |
532 | else: |
|
532 | else: | |
533 | ui.write(_(b'fncache already up to date\n')) |
|
533 | ui.write(_(b'fncache already up to date\n')) | |
534 |
|
534 | |||
535 |
|
535 | |||
536 | def deleteobsmarkers(obsstore, indices): |
|
536 | def deleteobsmarkers(obsstore, indices): | |
537 | """Delete some obsmarkers from obsstore and return how many were deleted |
|
537 | """Delete some obsmarkers from obsstore and return how many were deleted | |
538 |
|
538 | |||
539 | 'indices' is a list of ints which are the indices |
|
539 | 'indices' is a list of ints which are the indices | |
540 | of the markers to be deleted. |
|
540 | of the markers to be deleted. | |
541 |
|
541 | |||
542 | Every invocation of this function completely rewrites the obsstore file, |
|
542 | Every invocation of this function completely rewrites the obsstore file, | |
543 | skipping the markers we want to be removed. The new temporary file is |
|
543 | skipping the markers we want to be removed. The new temporary file is | |
544 | created, remaining markers are written there and on .close() this file |
|
544 | created, remaining markers are written there and on .close() this file | |
545 | gets atomically renamed to obsstore, thus guaranteeing consistency.""" |
|
545 | gets atomically renamed to obsstore, thus guaranteeing consistency.""" | |
546 | if not indices: |
|
546 | if not indices: | |
547 | # we don't want to rewrite the obsstore with the same content |
|
547 | # we don't want to rewrite the obsstore with the same content | |
548 | return |
|
548 | return | |
549 |
|
549 | |||
550 | left = [] |
|
550 | left = [] | |
551 | current = obsstore._all |
|
551 | current = obsstore._all | |
552 | n = 0 |
|
552 | n = 0 | |
553 | for i, m in enumerate(current): |
|
553 | for i, m in enumerate(current): | |
554 | if i in indices: |
|
554 | if i in indices: | |
555 | n += 1 |
|
555 | n += 1 | |
556 | continue |
|
556 | continue | |
557 | left.append(m) |
|
557 | left.append(m) | |
558 |
|
558 | |||
559 | newobsstorefile = obsstore.svfs(b'obsstore', b'w', atomictemp=True) |
|
559 | newobsstorefile = obsstore.svfs(b'obsstore', b'w', atomictemp=True) | |
560 | for bytes in obsolete.encodemarkers(left, True, obsstore._version): |
|
560 | for bytes in obsolete.encodemarkers(left, True, obsstore._version): | |
561 | newobsstorefile.write(bytes) |
|
561 | newobsstorefile.write(bytes) | |
562 | newobsstorefile.close() |
|
562 | newobsstorefile.close() | |
563 | return n |
|
563 | return n |
@@ -1,886 +1,886 b'' | |||||
1 | # censor code related to censoring revision |
|
1 | # censor code related to censoring revision | |
2 | # coding: utf8 |
|
2 | # coding: utf8 | |
3 | # |
|
3 | # | |
4 | # Copyright 2021 Pierre-Yves David <pierre-yves.david@octobus.net> |
|
4 | # Copyright 2021 Pierre-Yves David <pierre-yves.david@octobus.net> | |
5 | # Copyright 2015 Google, Inc <martinvonz@google.com> |
|
5 | # Copyright 2015 Google, Inc <martinvonz@google.com> | |
6 | # |
|
6 | # | |
7 | # This software may be used and distributed according to the terms of the |
|
7 | # This software may be used and distributed according to the terms of the | |
8 | # GNU General Public License version 2 or any later version. |
|
8 | # GNU General Public License version 2 or any later version. | |
9 |
|
9 | |||
10 | import binascii |
|
10 | import binascii | |
11 | import contextlib |
|
11 | import contextlib | |
12 | import os |
|
12 | import os | |
13 | import struct |
|
13 | import struct | |
14 |
|
14 | |||
15 | from ..node import ( |
|
15 | from ..node import ( | |
16 | nullrev, |
|
16 | nullrev, | |
17 | ) |
|
17 | ) | |
18 | from .constants import ( |
|
18 | from .constants import ( | |
19 | COMP_MODE_PLAIN, |
|
19 | COMP_MODE_PLAIN, | |
20 | ENTRY_DATA_COMPRESSED_LENGTH, |
|
20 | ENTRY_DATA_COMPRESSED_LENGTH, | |
21 | ENTRY_DATA_COMPRESSION_MODE, |
|
21 | ENTRY_DATA_COMPRESSION_MODE, | |
22 | ENTRY_DATA_OFFSET, |
|
22 | ENTRY_DATA_OFFSET, | |
23 | ENTRY_DATA_UNCOMPRESSED_LENGTH, |
|
23 | ENTRY_DATA_UNCOMPRESSED_LENGTH, | |
24 | ENTRY_DELTA_BASE, |
|
24 | ENTRY_DELTA_BASE, | |
25 | ENTRY_LINK_REV, |
|
25 | ENTRY_LINK_REV, | |
26 | ENTRY_NODE_ID, |
|
26 | ENTRY_NODE_ID, | |
27 | ENTRY_PARENT_1, |
|
27 | ENTRY_PARENT_1, | |
28 | ENTRY_PARENT_2, |
|
28 | ENTRY_PARENT_2, | |
29 | ENTRY_SIDEDATA_COMPRESSED_LENGTH, |
|
29 | ENTRY_SIDEDATA_COMPRESSED_LENGTH, | |
30 | ENTRY_SIDEDATA_COMPRESSION_MODE, |
|
30 | ENTRY_SIDEDATA_COMPRESSION_MODE, | |
31 | ENTRY_SIDEDATA_OFFSET, |
|
31 | ENTRY_SIDEDATA_OFFSET, | |
32 | REVIDX_ISCENSORED, |
|
32 | REVIDX_ISCENSORED, | |
33 | REVLOGV0, |
|
33 | REVLOGV0, | |
34 | REVLOGV1, |
|
34 | REVLOGV1, | |
35 | ) |
|
35 | ) | |
36 | from ..i18n import _ |
|
36 | from ..i18n import _ | |
37 |
|
37 | |||
38 | from .. import ( |
|
38 | from .. import ( | |
39 | error, |
|
39 | error, | |
40 | mdiff, |
|
40 | mdiff, | |
41 | pycompat, |
|
41 | pycompat, | |
42 | revlogutils, |
|
42 | revlogutils, | |
43 | util, |
|
43 | util, | |
44 | ) |
|
44 | ) | |
45 | from ..utils import ( |
|
45 | from ..utils import ( | |
46 | storageutil, |
|
46 | storageutil, | |
47 | ) |
|
47 | ) | |
48 | from . import ( |
|
48 | from . import ( | |
49 | constants, |
|
49 | constants, | |
50 | deltas, |
|
50 | deltas, | |
51 | ) |
|
51 | ) | |
52 |
|
52 | |||
53 |
|
53 | |||
54 | def v1_censor(rl, tr, censornode, tombstone=b''): |
|
54 | def v1_censor(rl, tr, censornode, tombstone=b''): | |
55 | """censors a revision in a "version 1" revlog""" |
|
55 | """censors a revision in a "version 1" revlog""" | |
56 | assert rl._format_version == constants.REVLOGV1, rl._format_version |
|
56 | assert rl._format_version == constants.REVLOGV1, rl._format_version | |
57 |
|
57 | |||
58 | # avoid cycle |
|
58 | # avoid cycle | |
59 | from .. import revlog |
|
59 | from .. import revlog | |
60 |
|
60 | |||
61 | censorrev = rl.rev(censornode) |
|
61 | censorrev = rl.rev(censornode) | |
62 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') |
|
62 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') | |
63 |
|
63 | |||
64 | # Rewriting the revlog in place is hard. Our strategy for censoring is |
|
64 | # Rewriting the revlog in place is hard. Our strategy for censoring is | |
65 | # to create a new revlog, copy all revisions to it, then replace the |
|
65 | # to create a new revlog, copy all revisions to it, then replace the | |
66 | # revlogs on transaction close. |
|
66 | # revlogs on transaction close. | |
67 | # |
|
67 | # | |
68 | # This is a bit dangerous. We could easily have a mismatch of state. |
|
68 | # This is a bit dangerous. We could easily have a mismatch of state. | |
69 | newrl = revlog.revlog( |
|
69 | newrl = revlog.revlog( | |
70 | rl.opener, |
|
70 | rl.opener, | |
71 | target=rl.target, |
|
71 | target=rl.target, | |
72 | radix=rl.radix, |
|
72 | radix=rl.radix, | |
73 | postfix=b'tmpcensored', |
|
73 | postfix=b'tmpcensored', | |
74 | censorable=True, |
|
74 | censorable=True, | |
75 | ) |
|
75 | ) | |
76 | newrl._format_version = rl._format_version |
|
76 | newrl._format_version = rl._format_version | |
77 | newrl._format_flags = rl._format_flags |
|
77 | newrl._format_flags = rl._format_flags | |
78 | newrl._generaldelta = rl._generaldelta |
|
78 | newrl._generaldelta = rl._generaldelta | |
79 | newrl._parse_index = rl._parse_index |
|
79 | newrl._parse_index = rl._parse_index | |
80 |
|
80 | |||
81 | for rev in rl.revs(): |
|
81 | for rev in rl.revs(): | |
82 | node = rl.node(rev) |
|
82 | node = rl.node(rev) | |
83 | p1, p2 = rl.parents(node) |
|
83 | p1, p2 = rl.parents(node) | |
84 |
|
84 | |||
85 | if rev == censorrev: |
|
85 | if rev == censorrev: | |
86 | newrl.addrawrevision( |
|
86 | newrl.addrawrevision( | |
87 | tombstone, |
|
87 | tombstone, | |
88 | tr, |
|
88 | tr, | |
89 | rl.linkrev(censorrev), |
|
89 | rl.linkrev(censorrev), | |
90 | p1, |
|
90 | p1, | |
91 | p2, |
|
91 | p2, | |
92 | censornode, |
|
92 | censornode, | |
93 | constants.REVIDX_ISCENSORED, |
|
93 | constants.REVIDX_ISCENSORED, | |
94 | ) |
|
94 | ) | |
95 |
|
95 | |||
96 | if newrl.deltaparent(rev) != nullrev: |
|
96 | if newrl.deltaparent(rev) != nullrev: | |
97 | m = _(b'censored revision stored as delta; cannot censor') |
|
97 | m = _(b'censored revision stored as delta; cannot censor') | |
98 | h = _( |
|
98 | h = _( | |
99 | b'censoring of revlogs is not fully implemented;' |
|
99 | b'censoring of revlogs is not fully implemented;' | |
100 | b' please report this bug' |
|
100 | b' please report this bug' | |
101 | ) |
|
101 | ) | |
102 | raise error.Abort(m, hint=h) |
|
102 | raise error.Abort(m, hint=h) | |
103 | continue |
|
103 | continue | |
104 |
|
104 | |||
105 | if rl.iscensored(rev): |
|
105 | if rl.iscensored(rev): | |
106 | if rl.deltaparent(rev) != nullrev: |
|
106 | if rl.deltaparent(rev) != nullrev: | |
107 | m = _( |
|
107 | m = _( | |
108 | b'cannot censor due to censored ' |
|
108 | b'cannot censor due to censored ' | |
109 | b'revision having delta stored' |
|
109 | b'revision having delta stored' | |
110 | ) |
|
110 | ) | |
111 | raise error.Abort(m) |
|
111 | raise error.Abort(m) | |
112 | rawtext = rl._chunk(rev) |
|
112 | rawtext = rl._chunk(rev) | |
113 | else: |
|
113 | else: | |
114 | rawtext = rl.rawdata(rev) |
|
114 | rawtext = rl.rawdata(rev) | |
115 |
|
115 | |||
116 | newrl.addrawrevision( |
|
116 | newrl.addrawrevision( | |
117 | rawtext, tr, rl.linkrev(rev), p1, p2, node, rl.flags(rev) |
|
117 | rawtext, tr, rl.linkrev(rev), p1, p2, node, rl.flags(rev) | |
118 | ) |
|
118 | ) | |
119 |
|
119 | |||
120 | tr.addbackup(rl._indexfile, location=b'store') |
|
120 | tr.addbackup(rl._indexfile, location=b'store') | |
121 | if not rl._inline: |
|
121 | if not rl._inline: | |
122 | tr.addbackup(rl._datafile, location=b'store') |
|
122 | tr.addbackup(rl._datafile, location=b'store') | |
123 |
|
123 | |||
124 | rl.opener.rename(newrl._indexfile, rl._indexfile) |
|
124 | rl.opener.rename(newrl._indexfile, rl._indexfile) | |
125 | if not rl._inline: |
|
125 | if not rl._inline: | |
126 | rl.opener.rename(newrl._datafile, rl._datafile) |
|
126 | rl.opener.rename(newrl._datafile, rl._datafile) | |
127 |
|
127 | |||
128 | rl.clearcaches() |
|
128 | rl.clearcaches() | |
129 | rl._loadindex() |
|
129 | rl._loadindex() | |
130 |
|
130 | |||
131 |
|
131 | |||
132 | def v2_censor(revlog, tr, censornode, tombstone=b''): |
|
132 | def v2_censor(revlog, tr, censornode, tombstone=b''): | |
133 | """censors a revision in a "version 2" revlog""" |
|
133 | """censors a revision in a "version 2" revlog""" | |
134 | assert revlog._format_version != REVLOGV0, revlog._format_version |
|
134 | assert revlog._format_version != REVLOGV0, revlog._format_version | |
135 | assert revlog._format_version != REVLOGV1, revlog._format_version |
|
135 | assert revlog._format_version != REVLOGV1, revlog._format_version | |
136 |
|
136 | |||
137 | censor_revs = {revlog.rev(censornode)} |
|
137 | censor_revs = {revlog.rev(censornode)} | |
138 | _rewrite_v2(revlog, tr, censor_revs, tombstone) |
|
138 | _rewrite_v2(revlog, tr, censor_revs, tombstone) | |
139 |
|
139 | |||
140 |
|
140 | |||
141 | def _rewrite_v2(revlog, tr, censor_revs, tombstone=b''): |
|
141 | def _rewrite_v2(revlog, tr, censor_revs, tombstone=b''): | |
142 | """rewrite a revlog to censor some of its content |
|
142 | """rewrite a revlog to censor some of its content | |
143 |
|
143 | |||
144 | General principle |
|
144 | General principle | |
145 |
|
145 | |||
146 | We create new revlog files (index/data/sidedata) to copy the content of |
|
146 | We create new revlog files (index/data/sidedata) to copy the content of | |
147 | the existing data without the censored data. |
|
147 | the existing data without the censored data. | |
148 |
|
148 | |||
149 | We need to recompute new delta for any revision that used the censored |
|
149 | We need to recompute new delta for any revision that used the censored | |
150 | revision as delta base. As the cumulative size of the new delta may be |
|
150 | revision as delta base. As the cumulative size of the new delta may be | |
151 | large, we store them in a temporary file until they are stored in their |
|
151 | large, we store them in a temporary file until they are stored in their | |
152 | final destination. |
|
152 | final destination. | |
153 |
|
153 | |||
154 | All data before the censored data can be blindly copied. The rest needs |
|
154 | All data before the censored data can be blindly copied. The rest needs | |
155 | to be copied as we go and the associated index entry needs adjustement. |
|
155 | to be copied as we go and the associated index entry needs adjustement. | |
156 | """ |
|
156 | """ | |
157 | assert revlog._format_version != REVLOGV0, revlog._format_version |
|
157 | assert revlog._format_version != REVLOGV0, revlog._format_version | |
158 | assert revlog._format_version != REVLOGV1, revlog._format_version |
|
158 | assert revlog._format_version != REVLOGV1, revlog._format_version | |
159 |
|
159 | |||
160 | old_index = revlog.index |
|
160 | old_index = revlog.index | |
161 | docket = revlog._docket |
|
161 | docket = revlog._docket | |
162 |
|
162 | |||
163 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') |
|
163 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') | |
164 |
|
164 | |||
165 | first_excl_rev = min(censor_revs) |
|
165 | first_excl_rev = min(censor_revs) | |
166 |
|
166 | |||
167 | first_excl_entry = revlog.index[first_excl_rev] |
|
167 | first_excl_entry = revlog.index[first_excl_rev] | |
168 | index_cutoff = revlog.index.entry_size * first_excl_rev |
|
168 | index_cutoff = revlog.index.entry_size * first_excl_rev | |
169 | data_cutoff = first_excl_entry[ENTRY_DATA_OFFSET] >> 16 |
|
169 | data_cutoff = first_excl_entry[ENTRY_DATA_OFFSET] >> 16 | |
170 | sidedata_cutoff = revlog.sidedata_cut_off(first_excl_rev) |
|
170 | sidedata_cutoff = revlog.sidedata_cut_off(first_excl_rev) | |
171 |
|
171 | |||
172 | with pycompat.unnamedtempfile(mode=b"w+b") as tmp_storage: |
|
172 | with pycompat.unnamedtempfile(mode=b"w+b") as tmp_storage: | |
173 | # rev β (new_base, data_start, data_end, compression_mode) |
|
173 | # rev β (new_base, data_start, data_end, compression_mode) | |
174 | rewritten_entries = _precompute_rewritten_delta( |
|
174 | rewritten_entries = _precompute_rewritten_delta( | |
175 | revlog, |
|
175 | revlog, | |
176 | old_index, |
|
176 | old_index, | |
177 | censor_revs, |
|
177 | censor_revs, | |
178 | tmp_storage, |
|
178 | tmp_storage, | |
179 | ) |
|
179 | ) | |
180 |
|
180 | |||
181 | all_files = _setup_new_files( |
|
181 | all_files = _setup_new_files( | |
182 | revlog, |
|
182 | revlog, | |
183 | index_cutoff, |
|
183 | index_cutoff, | |
184 | data_cutoff, |
|
184 | data_cutoff, | |
185 | sidedata_cutoff, |
|
185 | sidedata_cutoff, | |
186 | ) |
|
186 | ) | |
187 |
|
187 | |||
188 | # we dont need to open the old index file since its content already |
|
188 | # we dont need to open the old index file since its content already | |
189 | # exist in a usable form in `old_index`. |
|
189 | # exist in a usable form in `old_index`. | |
190 | with all_files() as open_files: |
|
190 | with all_files() as open_files: | |
191 | ( |
|
191 | ( | |
192 | old_data_file, |
|
192 | old_data_file, | |
193 | old_sidedata_file, |
|
193 | old_sidedata_file, | |
194 | new_index_file, |
|
194 | new_index_file, | |
195 | new_data_file, |
|
195 | new_data_file, | |
196 | new_sidedata_file, |
|
196 | new_sidedata_file, | |
197 | ) = open_files |
|
197 | ) = open_files | |
198 |
|
198 | |||
199 | # writing the censored revision |
|
199 | # writing the censored revision | |
200 |
|
200 | |||
201 | # Writing all subsequent revisions |
|
201 | # Writing all subsequent revisions | |
202 | for rev in range(first_excl_rev, len(old_index)): |
|
202 | for rev in range(first_excl_rev, len(old_index)): | |
203 | if rev in censor_revs: |
|
203 | if rev in censor_revs: | |
204 | _rewrite_censor( |
|
204 | _rewrite_censor( | |
205 | revlog, |
|
205 | revlog, | |
206 | old_index, |
|
206 | old_index, | |
207 | open_files, |
|
207 | open_files, | |
208 | rev, |
|
208 | rev, | |
209 | tombstone, |
|
209 | tombstone, | |
210 | ) |
|
210 | ) | |
211 | else: |
|
211 | else: | |
212 | _rewrite_simple( |
|
212 | _rewrite_simple( | |
213 | revlog, |
|
213 | revlog, | |
214 | old_index, |
|
214 | old_index, | |
215 | open_files, |
|
215 | open_files, | |
216 | rev, |
|
216 | rev, | |
217 | rewritten_entries, |
|
217 | rewritten_entries, | |
218 | tmp_storage, |
|
218 | tmp_storage, | |
219 | ) |
|
219 | ) | |
220 | docket.write(transaction=None, stripping=True) |
|
220 | docket.write(transaction=None, stripping=True) | |
221 |
|
221 | |||
222 |
|
222 | |||
223 | def _precompute_rewritten_delta( |
|
223 | def _precompute_rewritten_delta( | |
224 | revlog, |
|
224 | revlog, | |
225 | old_index, |
|
225 | old_index, | |
226 | excluded_revs, |
|
226 | excluded_revs, | |
227 | tmp_storage, |
|
227 | tmp_storage, | |
228 | ): |
|
228 | ): | |
229 | """Compute new delta for revisions whose delta is based on revision that |
|
229 | """Compute new delta for revisions whose delta is based on revision that | |
230 | will not survive as is. |
|
230 | will not survive as is. | |
231 |
|
231 | |||
232 | Return a mapping: {rev β (new_base, data_start, data_end, compression_mode)} |
|
232 | Return a mapping: {rev β (new_base, data_start, data_end, compression_mode)} | |
233 | """ |
|
233 | """ | |
234 | dc = deltas.deltacomputer(revlog) |
|
234 | dc = deltas.deltacomputer(revlog) | |
235 | rewritten_entries = {} |
|
235 | rewritten_entries = {} | |
236 | first_excl_rev = min(excluded_revs) |
|
236 | first_excl_rev = min(excluded_revs) | |
237 | with revlog._segmentfile._open_read() as dfh: |
|
237 | with revlog._segmentfile._open_read() as dfh: | |
238 | for rev in range(first_excl_rev, len(old_index)): |
|
238 | for rev in range(first_excl_rev, len(old_index)): | |
239 | if rev in excluded_revs: |
|
239 | if rev in excluded_revs: | |
240 | # this revision will be preserved as is, so we don't need to |
|
240 | # this revision will be preserved as is, so we don't need to | |
241 | # consider recomputing a delta. |
|
241 | # consider recomputing a delta. | |
242 | continue |
|
242 | continue | |
243 | entry = old_index[rev] |
|
243 | entry = old_index[rev] | |
244 | if entry[ENTRY_DELTA_BASE] not in excluded_revs: |
|
244 | if entry[ENTRY_DELTA_BASE] not in excluded_revs: | |
245 | continue |
|
245 | continue | |
246 | # This is a revision that use the censored revision as the base |
|
246 | # This is a revision that use the censored revision as the base | |
247 | # for its delta. We need a need new deltas |
|
247 | # for its delta. We need a need new deltas | |
248 | if entry[ENTRY_DATA_UNCOMPRESSED_LENGTH] == 0: |
|
248 | if entry[ENTRY_DATA_UNCOMPRESSED_LENGTH] == 0: | |
249 | # this revision is empty, we can delta against nullrev |
|
249 | # this revision is empty, we can delta against nullrev | |
250 | rewritten_entries[rev] = (nullrev, 0, 0, COMP_MODE_PLAIN) |
|
250 | rewritten_entries[rev] = (nullrev, 0, 0, COMP_MODE_PLAIN) | |
251 | else: |
|
251 | else: | |
252 |
|
252 | |||
253 | text = revlog.rawdata(rev, _df=dfh) |
|
253 | text = revlog.rawdata(rev, _df=dfh) | |
254 | info = revlogutils.revisioninfo( |
|
254 | info = revlogutils.revisioninfo( | |
255 | node=entry[ENTRY_NODE_ID], |
|
255 | node=entry[ENTRY_NODE_ID], | |
256 | p1=revlog.node(entry[ENTRY_PARENT_1]), |
|
256 | p1=revlog.node(entry[ENTRY_PARENT_1]), | |
257 | p2=revlog.node(entry[ENTRY_PARENT_2]), |
|
257 | p2=revlog.node(entry[ENTRY_PARENT_2]), | |
258 | btext=[text], |
|
258 | btext=[text], | |
259 | textlen=len(text), |
|
259 | textlen=len(text), | |
260 | cachedelta=None, |
|
260 | cachedelta=None, | |
261 | flags=entry[ENTRY_DATA_OFFSET] & 0xFFFF, |
|
261 | flags=entry[ENTRY_DATA_OFFSET] & 0xFFFF, | |
262 | ) |
|
262 | ) | |
263 | d = dc.finddeltainfo( |
|
263 | d = dc.finddeltainfo( | |
264 | info, dfh, excluded_bases=excluded_revs, target_rev=rev |
|
264 | info, dfh, excluded_bases=excluded_revs, target_rev=rev | |
265 | ) |
|
265 | ) | |
266 | default_comp = revlog._docket.default_compression_header |
|
266 | default_comp = revlog._docket.default_compression_header | |
267 | comp_mode, d = deltas.delta_compression(default_comp, d) |
|
267 | comp_mode, d = deltas.delta_compression(default_comp, d) | |
268 | # using `tell` is a bit lazy, but we are not here for speed |
|
268 | # using `tell` is a bit lazy, but we are not here for speed | |
269 | start = tmp_storage.tell() |
|
269 | start = tmp_storage.tell() | |
270 | tmp_storage.write(d.data[1]) |
|
270 | tmp_storage.write(d.data[1]) | |
271 | end = tmp_storage.tell() |
|
271 | end = tmp_storage.tell() | |
272 | rewritten_entries[rev] = (d.base, start, end, comp_mode) |
|
272 | rewritten_entries[rev] = (d.base, start, end, comp_mode) | |
273 | return rewritten_entries |
|
273 | return rewritten_entries | |
274 |
|
274 | |||
275 |
|
275 | |||
276 | def _setup_new_files( |
|
276 | def _setup_new_files( | |
277 | revlog, |
|
277 | revlog, | |
278 | index_cutoff, |
|
278 | index_cutoff, | |
279 | data_cutoff, |
|
279 | data_cutoff, | |
280 | sidedata_cutoff, |
|
280 | sidedata_cutoff, | |
281 | ): |
|
281 | ): | |
282 | """ |
|
282 | """ | |
283 |
|
283 | |||
284 | return a context manager to open all the relevant files: |
|
284 | return a context manager to open all the relevant files: | |
285 | - old_data_file, |
|
285 | - old_data_file, | |
286 | - old_sidedata_file, |
|
286 | - old_sidedata_file, | |
287 | - new_index_file, |
|
287 | - new_index_file, | |
288 | - new_data_file, |
|
288 | - new_data_file, | |
289 | - new_sidedata_file, |
|
289 | - new_sidedata_file, | |
290 |
|
290 | |||
291 | The old_index_file is not here because it is accessed through the |
|
291 | The old_index_file is not here because it is accessed through the | |
292 | `old_index` object if the caller function. |
|
292 | `old_index` object if the caller function. | |
293 | """ |
|
293 | """ | |
294 | docket = revlog._docket |
|
294 | docket = revlog._docket | |
295 | old_index_filepath = revlog.opener.join(docket.index_filepath()) |
|
295 | old_index_filepath = revlog.opener.join(docket.index_filepath()) | |
296 | old_data_filepath = revlog.opener.join(docket.data_filepath()) |
|
296 | old_data_filepath = revlog.opener.join(docket.data_filepath()) | |
297 | old_sidedata_filepath = revlog.opener.join(docket.sidedata_filepath()) |
|
297 | old_sidedata_filepath = revlog.opener.join(docket.sidedata_filepath()) | |
298 |
|
298 | |||
299 | new_index_filepath = revlog.opener.join(docket.new_index_file()) |
|
299 | new_index_filepath = revlog.opener.join(docket.new_index_file()) | |
300 | new_data_filepath = revlog.opener.join(docket.new_data_file()) |
|
300 | new_data_filepath = revlog.opener.join(docket.new_data_file()) | |
301 | new_sidedata_filepath = revlog.opener.join(docket.new_sidedata_file()) |
|
301 | new_sidedata_filepath = revlog.opener.join(docket.new_sidedata_file()) | |
302 |
|
302 | |||
303 | util.copyfile(old_index_filepath, new_index_filepath, nb_bytes=index_cutoff) |
|
303 | util.copyfile(old_index_filepath, new_index_filepath, nb_bytes=index_cutoff) | |
304 | util.copyfile(old_data_filepath, new_data_filepath, nb_bytes=data_cutoff) |
|
304 | util.copyfile(old_data_filepath, new_data_filepath, nb_bytes=data_cutoff) | |
305 | util.copyfile( |
|
305 | util.copyfile( | |
306 | old_sidedata_filepath, |
|
306 | old_sidedata_filepath, | |
307 | new_sidedata_filepath, |
|
307 | new_sidedata_filepath, | |
308 | nb_bytes=sidedata_cutoff, |
|
308 | nb_bytes=sidedata_cutoff, | |
309 | ) |
|
309 | ) | |
310 | revlog.opener.register_file(docket.index_filepath()) |
|
310 | revlog.opener.register_file(docket.index_filepath()) | |
311 | revlog.opener.register_file(docket.data_filepath()) |
|
311 | revlog.opener.register_file(docket.data_filepath()) | |
312 | revlog.opener.register_file(docket.sidedata_filepath()) |
|
312 | revlog.opener.register_file(docket.sidedata_filepath()) | |
313 |
|
313 | |||
314 | docket.index_end = index_cutoff |
|
314 | docket.index_end = index_cutoff | |
315 | docket.data_end = data_cutoff |
|
315 | docket.data_end = data_cutoff | |
316 | docket.sidedata_end = sidedata_cutoff |
|
316 | docket.sidedata_end = sidedata_cutoff | |
317 |
|
317 | |||
318 | # reload the revlog internal information |
|
318 | # reload the revlog internal information | |
319 | revlog.clearcaches() |
|
319 | revlog.clearcaches() | |
320 | revlog._loadindex(docket=docket) |
|
320 | revlog._loadindex(docket=docket) | |
321 |
|
321 | |||
322 | @contextlib.contextmanager |
|
322 | @contextlib.contextmanager | |
323 | def all_files_opener(): |
|
323 | def all_files_opener(): | |
324 | # hide opening in an helper function to please check-code, black |
|
324 | # hide opening in an helper function to please check-code, black | |
325 | # and various python version at the same time |
|
325 | # and various python version at the same time | |
326 | with open(old_data_filepath, 'rb') as old_data_file: |
|
326 | with open(old_data_filepath, 'rb') as old_data_file: | |
327 | with open(old_sidedata_filepath, 'rb') as old_sidedata_file: |
|
327 | with open(old_sidedata_filepath, 'rb') as old_sidedata_file: | |
328 | with open(new_index_filepath, 'r+b') as new_index_file: |
|
328 | with open(new_index_filepath, 'r+b') as new_index_file: | |
329 | with open(new_data_filepath, 'r+b') as new_data_file: |
|
329 | with open(new_data_filepath, 'r+b') as new_data_file: | |
330 | with open( |
|
330 | with open( | |
331 | new_sidedata_filepath, 'r+b' |
|
331 | new_sidedata_filepath, 'r+b' | |
332 | ) as new_sidedata_file: |
|
332 | ) as new_sidedata_file: | |
333 | new_index_file.seek(0, os.SEEK_END) |
|
333 | new_index_file.seek(0, os.SEEK_END) | |
334 | assert new_index_file.tell() == index_cutoff |
|
334 | assert new_index_file.tell() == index_cutoff | |
335 | new_data_file.seek(0, os.SEEK_END) |
|
335 | new_data_file.seek(0, os.SEEK_END) | |
336 | assert new_data_file.tell() == data_cutoff |
|
336 | assert new_data_file.tell() == data_cutoff | |
337 | new_sidedata_file.seek(0, os.SEEK_END) |
|
337 | new_sidedata_file.seek(0, os.SEEK_END) | |
338 | assert new_sidedata_file.tell() == sidedata_cutoff |
|
338 | assert new_sidedata_file.tell() == sidedata_cutoff | |
339 | yield ( |
|
339 | yield ( | |
340 | old_data_file, |
|
340 | old_data_file, | |
341 | old_sidedata_file, |
|
341 | old_sidedata_file, | |
342 | new_index_file, |
|
342 | new_index_file, | |
343 | new_data_file, |
|
343 | new_data_file, | |
344 | new_sidedata_file, |
|
344 | new_sidedata_file, | |
345 | ) |
|
345 | ) | |
346 |
|
346 | |||
347 | return all_files_opener |
|
347 | return all_files_opener | |
348 |
|
348 | |||
349 |
|
349 | |||
350 | def _rewrite_simple( |
|
350 | def _rewrite_simple( | |
351 | revlog, |
|
351 | revlog, | |
352 | old_index, |
|
352 | old_index, | |
353 | all_files, |
|
353 | all_files, | |
354 | rev, |
|
354 | rev, | |
355 | rewritten_entries, |
|
355 | rewritten_entries, | |
356 | tmp_storage, |
|
356 | tmp_storage, | |
357 | ): |
|
357 | ): | |
358 | """append a normal revision to the index after the rewritten one(s)""" |
|
358 | """append a normal revision to the index after the rewritten one(s)""" | |
359 | ( |
|
359 | ( | |
360 | old_data_file, |
|
360 | old_data_file, | |
361 | old_sidedata_file, |
|
361 | old_sidedata_file, | |
362 | new_index_file, |
|
362 | new_index_file, | |
363 | new_data_file, |
|
363 | new_data_file, | |
364 | new_sidedata_file, |
|
364 | new_sidedata_file, | |
365 | ) = all_files |
|
365 | ) = all_files | |
366 | entry = old_index[rev] |
|
366 | entry = old_index[rev] | |
367 | flags = entry[ENTRY_DATA_OFFSET] & 0xFFFF |
|
367 | flags = entry[ENTRY_DATA_OFFSET] & 0xFFFF | |
368 | old_data_offset = entry[ENTRY_DATA_OFFSET] >> 16 |
|
368 | old_data_offset = entry[ENTRY_DATA_OFFSET] >> 16 | |
369 |
|
369 | |||
370 | if rev not in rewritten_entries: |
|
370 | if rev not in rewritten_entries: | |
371 | old_data_file.seek(old_data_offset) |
|
371 | old_data_file.seek(old_data_offset) | |
372 | new_data_size = entry[ENTRY_DATA_COMPRESSED_LENGTH] |
|
372 | new_data_size = entry[ENTRY_DATA_COMPRESSED_LENGTH] | |
373 | new_data = old_data_file.read(new_data_size) |
|
373 | new_data = old_data_file.read(new_data_size) | |
374 | data_delta_base = entry[ENTRY_DELTA_BASE] |
|
374 | data_delta_base = entry[ENTRY_DELTA_BASE] | |
375 | d_comp_mode = entry[ENTRY_DATA_COMPRESSION_MODE] |
|
375 | d_comp_mode = entry[ENTRY_DATA_COMPRESSION_MODE] | |
376 | else: |
|
376 | else: | |
377 | ( |
|
377 | ( | |
378 | data_delta_base, |
|
378 | data_delta_base, | |
379 | start, |
|
379 | start, | |
380 | end, |
|
380 | end, | |
381 | d_comp_mode, |
|
381 | d_comp_mode, | |
382 | ) = rewritten_entries[rev] |
|
382 | ) = rewritten_entries[rev] | |
383 | new_data_size = end - start |
|
383 | new_data_size = end - start | |
384 | tmp_storage.seek(start) |
|
384 | tmp_storage.seek(start) | |
385 | new_data = tmp_storage.read(new_data_size) |
|
385 | new_data = tmp_storage.read(new_data_size) | |
386 |
|
386 | |||
387 | # It might be faster to group continuous read/write operation, |
|
387 | # It might be faster to group continuous read/write operation, | |
388 | # however, this is censor, an operation that is not focussed |
|
388 | # however, this is censor, an operation that is not focussed | |
389 | # around stellar performance. So I have not written this |
|
389 | # around stellar performance. So I have not written this | |
390 | # optimisation yet. |
|
390 | # optimisation yet. | |
391 | new_data_offset = new_data_file.tell() |
|
391 | new_data_offset = new_data_file.tell() | |
392 | new_data_file.write(new_data) |
|
392 | new_data_file.write(new_data) | |
393 |
|
393 | |||
394 | sidedata_size = entry[ENTRY_SIDEDATA_COMPRESSED_LENGTH] |
|
394 | sidedata_size = entry[ENTRY_SIDEDATA_COMPRESSED_LENGTH] | |
395 | new_sidedata_offset = new_sidedata_file.tell() |
|
395 | new_sidedata_offset = new_sidedata_file.tell() | |
396 | if 0 < sidedata_size: |
|
396 | if 0 < sidedata_size: | |
397 | old_sidedata_offset = entry[ENTRY_SIDEDATA_OFFSET] |
|
397 | old_sidedata_offset = entry[ENTRY_SIDEDATA_OFFSET] | |
398 | old_sidedata_file.seek(old_sidedata_offset) |
|
398 | old_sidedata_file.seek(old_sidedata_offset) | |
399 | new_sidedata = old_sidedata_file.read(sidedata_size) |
|
399 | new_sidedata = old_sidedata_file.read(sidedata_size) | |
400 | new_sidedata_file.write(new_sidedata) |
|
400 | new_sidedata_file.write(new_sidedata) | |
401 |
|
401 | |||
402 | data_uncompressed_length = entry[ENTRY_DATA_UNCOMPRESSED_LENGTH] |
|
402 | data_uncompressed_length = entry[ENTRY_DATA_UNCOMPRESSED_LENGTH] | |
403 | sd_com_mode = entry[ENTRY_SIDEDATA_COMPRESSION_MODE] |
|
403 | sd_com_mode = entry[ENTRY_SIDEDATA_COMPRESSION_MODE] | |
404 | assert data_delta_base <= rev, (data_delta_base, rev) |
|
404 | assert data_delta_base <= rev, (data_delta_base, rev) | |
405 |
|
405 | |||
406 | new_entry = revlogutils.entry( |
|
406 | new_entry = revlogutils.entry( | |
407 | flags=flags, |
|
407 | flags=flags, | |
408 | data_offset=new_data_offset, |
|
408 | data_offset=new_data_offset, | |
409 | data_compressed_length=new_data_size, |
|
409 | data_compressed_length=new_data_size, | |
410 | data_uncompressed_length=data_uncompressed_length, |
|
410 | data_uncompressed_length=data_uncompressed_length, | |
411 | data_delta_base=data_delta_base, |
|
411 | data_delta_base=data_delta_base, | |
412 | link_rev=entry[ENTRY_LINK_REV], |
|
412 | link_rev=entry[ENTRY_LINK_REV], | |
413 | parent_rev_1=entry[ENTRY_PARENT_1], |
|
413 | parent_rev_1=entry[ENTRY_PARENT_1], | |
414 | parent_rev_2=entry[ENTRY_PARENT_2], |
|
414 | parent_rev_2=entry[ENTRY_PARENT_2], | |
415 | node_id=entry[ENTRY_NODE_ID], |
|
415 | node_id=entry[ENTRY_NODE_ID], | |
416 | sidedata_offset=new_sidedata_offset, |
|
416 | sidedata_offset=new_sidedata_offset, | |
417 | sidedata_compressed_length=sidedata_size, |
|
417 | sidedata_compressed_length=sidedata_size, | |
418 | data_compression_mode=d_comp_mode, |
|
418 | data_compression_mode=d_comp_mode, | |
419 | sidedata_compression_mode=sd_com_mode, |
|
419 | sidedata_compression_mode=sd_com_mode, | |
420 | ) |
|
420 | ) | |
421 | revlog.index.append(new_entry) |
|
421 | revlog.index.append(new_entry) | |
422 | entry_bin = revlog.index.entry_binary(rev) |
|
422 | entry_bin = revlog.index.entry_binary(rev) | |
423 | new_index_file.write(entry_bin) |
|
423 | new_index_file.write(entry_bin) | |
424 |
|
424 | |||
425 | revlog._docket.index_end = new_index_file.tell() |
|
425 | revlog._docket.index_end = new_index_file.tell() | |
426 | revlog._docket.data_end = new_data_file.tell() |
|
426 | revlog._docket.data_end = new_data_file.tell() | |
427 | revlog._docket.sidedata_end = new_sidedata_file.tell() |
|
427 | revlog._docket.sidedata_end = new_sidedata_file.tell() | |
428 |
|
428 | |||
429 |
|
429 | |||
430 | def _rewrite_censor( |
|
430 | def _rewrite_censor( | |
431 | revlog, |
|
431 | revlog, | |
432 | old_index, |
|
432 | old_index, | |
433 | all_files, |
|
433 | all_files, | |
434 | rev, |
|
434 | rev, | |
435 | tombstone, |
|
435 | tombstone, | |
436 | ): |
|
436 | ): | |
437 | """rewrite and append a censored revision""" |
|
437 | """rewrite and append a censored revision""" | |
438 | ( |
|
438 | ( | |
439 | old_data_file, |
|
439 | old_data_file, | |
440 | old_sidedata_file, |
|
440 | old_sidedata_file, | |
441 | new_index_file, |
|
441 | new_index_file, | |
442 | new_data_file, |
|
442 | new_data_file, | |
443 | new_sidedata_file, |
|
443 | new_sidedata_file, | |
444 | ) = all_files |
|
444 | ) = all_files | |
445 | entry = old_index[rev] |
|
445 | entry = old_index[rev] | |
446 |
|
446 | |||
447 | # XXX consider trying the default compression too |
|
447 | # XXX consider trying the default compression too | |
448 | new_data_size = len(tombstone) |
|
448 | new_data_size = len(tombstone) | |
449 | new_data_offset = new_data_file.tell() |
|
449 | new_data_offset = new_data_file.tell() | |
450 | new_data_file.write(tombstone) |
|
450 | new_data_file.write(tombstone) | |
451 |
|
451 | |||
452 | # we are not adding any sidedata as they might leak info about the censored version |
|
452 | # we are not adding any sidedata as they might leak info about the censored version | |
453 |
|
453 | |||
454 | link_rev = entry[ENTRY_LINK_REV] |
|
454 | link_rev = entry[ENTRY_LINK_REV] | |
455 |
|
455 | |||
456 | p1 = entry[ENTRY_PARENT_1] |
|
456 | p1 = entry[ENTRY_PARENT_1] | |
457 | p2 = entry[ENTRY_PARENT_2] |
|
457 | p2 = entry[ENTRY_PARENT_2] | |
458 |
|
458 | |||
459 | new_entry = revlogutils.entry( |
|
459 | new_entry = revlogutils.entry( | |
460 | flags=constants.REVIDX_ISCENSORED, |
|
460 | flags=constants.REVIDX_ISCENSORED, | |
461 | data_offset=new_data_offset, |
|
461 | data_offset=new_data_offset, | |
462 | data_compressed_length=new_data_size, |
|
462 | data_compressed_length=new_data_size, | |
463 | data_uncompressed_length=new_data_size, |
|
463 | data_uncompressed_length=new_data_size, | |
464 | data_delta_base=rev, |
|
464 | data_delta_base=rev, | |
465 | link_rev=link_rev, |
|
465 | link_rev=link_rev, | |
466 | parent_rev_1=p1, |
|
466 | parent_rev_1=p1, | |
467 | parent_rev_2=p2, |
|
467 | parent_rev_2=p2, | |
468 | node_id=entry[ENTRY_NODE_ID], |
|
468 | node_id=entry[ENTRY_NODE_ID], | |
469 | sidedata_offset=0, |
|
469 | sidedata_offset=0, | |
470 | sidedata_compressed_length=0, |
|
470 | sidedata_compressed_length=0, | |
471 | data_compression_mode=COMP_MODE_PLAIN, |
|
471 | data_compression_mode=COMP_MODE_PLAIN, | |
472 | sidedata_compression_mode=COMP_MODE_PLAIN, |
|
472 | sidedata_compression_mode=COMP_MODE_PLAIN, | |
473 | ) |
|
473 | ) | |
474 | revlog.index.append(new_entry) |
|
474 | revlog.index.append(new_entry) | |
475 | entry_bin = revlog.index.entry_binary(rev) |
|
475 | entry_bin = revlog.index.entry_binary(rev) | |
476 | new_index_file.write(entry_bin) |
|
476 | new_index_file.write(entry_bin) | |
477 | revlog._docket.index_end = new_index_file.tell() |
|
477 | revlog._docket.index_end = new_index_file.tell() | |
478 | revlog._docket.data_end = new_data_file.tell() |
|
478 | revlog._docket.data_end = new_data_file.tell() | |
479 |
|
479 | |||
480 |
|
480 | |||
481 | def _get_filename_from_filelog_index(path): |
|
481 | def _get_filename_from_filelog_index(path): | |
482 | # Drop the extension and the `data/` prefix |
|
482 | # Drop the extension and the `data/` prefix | |
483 | path_part = path.rsplit(b'.', 1)[0].split(b'/', 1) |
|
483 | path_part = path.rsplit(b'.', 1)[0].split(b'/', 1) | |
484 | if len(path_part) < 2: |
|
484 | if len(path_part) < 2: | |
485 | msg = _(b"cannot recognize filelog from filename: '%s'") |
|
485 | msg = _(b"cannot recognize filelog from filename: '%s'") | |
486 | msg %= path |
|
486 | msg %= path | |
487 | raise error.Abort(msg) |
|
487 | raise error.Abort(msg) | |
488 |
|
488 | |||
489 | return path_part[1] |
|
489 | return path_part[1] | |
490 |
|
490 | |||
491 |
|
491 | |||
492 | def _filelog_from_filename(repo, path): |
|
492 | def _filelog_from_filename(repo, path): | |
493 | """Returns the filelog for the given `path`. Stolen from `engine.py`""" |
|
493 | """Returns the filelog for the given `path`. Stolen from `engine.py`""" | |
494 |
|
494 | |||
495 | from .. import filelog # avoid cycle |
|
495 | from .. import filelog # avoid cycle | |
496 |
|
496 | |||
497 | fl = filelog.filelog(repo.svfs, path) |
|
497 | fl = filelog.filelog(repo.svfs, path) | |
498 | return fl |
|
498 | return fl | |
499 |
|
499 | |||
500 |
|
500 | |||
501 | def _write_swapped_parents(repo, rl, rev, offset, fp): |
|
501 | def _write_swapped_parents(repo, rl, rev, offset, fp): | |
502 | """Swaps p1 and p2 and overwrites the revlog entry for `rev` in `fp`""" |
|
502 | """Swaps p1 and p2 and overwrites the revlog entry for `rev` in `fp`""" | |
503 | from ..pure import parsers # avoid cycle |
|
503 | from ..pure import parsers # avoid cycle | |
504 |
|
504 | |||
505 | if repo._currentlock(repo._lockref) is None: |
|
505 | if repo._currentlock(repo._lockref) is None: | |
506 | # Let's be paranoid about it |
|
506 | # Let's be paranoid about it | |
507 | msg = "repo needs to be locked to rewrite parents" |
|
507 | msg = "repo needs to be locked to rewrite parents" | |
508 | raise error.ProgrammingError(msg) |
|
508 | raise error.ProgrammingError(msg) | |
509 |
|
509 | |||
510 | index_format = parsers.IndexObject.index_format |
|
510 | index_format = parsers.IndexObject.index_format | |
511 | entry = rl.index[rev] |
|
511 | entry = rl.index[rev] | |
512 | new_entry = list(entry) |
|
512 | new_entry = list(entry) | |
513 | new_entry[5], new_entry[6] = entry[6], entry[5] |
|
513 | new_entry[5], new_entry[6] = entry[6], entry[5] | |
514 | packed = index_format.pack(*new_entry[:8]) |
|
514 | packed = index_format.pack(*new_entry[:8]) | |
515 | fp.seek(offset) |
|
515 | fp.seek(offset) | |
516 | fp.write(packed) |
|
516 | fp.write(packed) | |
517 |
|
517 | |||
518 |
|
518 | |||
519 | def _reorder_filelog_parents(repo, fl, to_fix): |
|
519 | def _reorder_filelog_parents(repo, fl, to_fix): | |
520 | """ |
|
520 | """ | |
521 | Swaps p1 and p2 for all `to_fix` revisions of filelog `fl` and writes the |
|
521 | Swaps p1 and p2 for all `to_fix` revisions of filelog `fl` and writes the | |
522 | new version to disk, overwriting the old one with a rename. |
|
522 | new version to disk, overwriting the old one with a rename. | |
523 | """ |
|
523 | """ | |
524 | from ..pure import parsers # avoid cycle |
|
524 | from ..pure import parsers # avoid cycle | |
525 |
|
525 | |||
526 | ui = repo.ui |
|
526 | ui = repo.ui | |
527 | assert len(to_fix) > 0 |
|
527 | assert len(to_fix) > 0 | |
528 | rl = fl._revlog |
|
528 | rl = fl._revlog | |
529 | if rl._format_version != constants.REVLOGV1: |
|
529 | if rl._format_version != constants.REVLOGV1: | |
530 | msg = "expected version 1 revlog, got version '%d'" % rl._format_version |
|
530 | msg = "expected version 1 revlog, got version '%d'" % rl._format_version | |
531 | raise error.ProgrammingError(msg) |
|
531 | raise error.ProgrammingError(msg) | |
532 |
|
532 | |||
533 | index_file = rl._indexfile |
|
533 | index_file = rl._indexfile | |
534 | new_file_path = index_file + b'.tmp-parents-fix' |
|
534 | new_file_path = index_file + b'.tmp-parents-fix' | |
535 | repaired_msg = _(b"repaired revision %d of 'filelog %s'\n") |
|
535 | repaired_msg = _(b"repaired revision %d of 'filelog %s'\n") | |
536 |
|
536 | |||
537 | with ui.uninterruptible(): |
|
537 | with ui.uninterruptible(): | |
538 | try: |
|
538 | try: | |
539 | util.copyfile( |
|
539 | util.copyfile( | |
540 | rl.opener.join(index_file), |
|
540 | rl.opener.join(index_file), | |
541 | rl.opener.join(new_file_path), |
|
541 | rl.opener.join(new_file_path), | |
542 | checkambig=rl._checkambig, |
|
542 | checkambig=rl._checkambig, | |
543 | ) |
|
543 | ) | |
544 |
|
544 | |||
545 | with rl.opener(new_file_path, mode=b"r+") as fp: |
|
545 | with rl.opener(new_file_path, mode=b"r+") as fp: | |
546 | if rl._inline: |
|
546 | if rl._inline: | |
547 | index = parsers.InlinedIndexObject(fp.read()) |
|
547 | index = parsers.InlinedIndexObject(fp.read()) | |
548 | for rev in fl.revs(): |
|
548 | for rev in fl.revs(): | |
549 | if rev in to_fix: |
|
549 | if rev in to_fix: | |
550 | offset = index._calculate_index(rev) |
|
550 | offset = index._calculate_index(rev) | |
551 | _write_swapped_parents(repo, rl, rev, offset, fp) |
|
551 | _write_swapped_parents(repo, rl, rev, offset, fp) | |
552 | ui.write(repaired_msg % (rev, index_file)) |
|
552 | ui.write(repaired_msg % (rev, index_file)) | |
553 | else: |
|
553 | else: | |
554 | index_format = parsers.IndexObject.index_format |
|
554 | index_format = parsers.IndexObject.index_format | |
555 | for rev in to_fix: |
|
555 | for rev in to_fix: | |
556 | offset = rev * index_format.size |
|
556 | offset = rev * index_format.size | |
557 | _write_swapped_parents(repo, rl, rev, offset, fp) |
|
557 | _write_swapped_parents(repo, rl, rev, offset, fp) | |
558 | ui.write(repaired_msg % (rev, index_file)) |
|
558 | ui.write(repaired_msg % (rev, index_file)) | |
559 |
|
559 | |||
560 | rl.opener.rename(new_file_path, index_file) |
|
560 | rl.opener.rename(new_file_path, index_file) | |
561 | rl.clearcaches() |
|
561 | rl.clearcaches() | |
562 | rl._loadindex() |
|
562 | rl._loadindex() | |
563 | finally: |
|
563 | finally: | |
564 | util.tryunlink(new_file_path) |
|
564 | util.tryunlink(new_file_path) | |
565 |
|
565 | |||
566 |
|
566 | |||
567 | def _is_revision_affected(fl, filerev, metadata_cache=None): |
|
567 | def _is_revision_affected(fl, filerev, metadata_cache=None): | |
568 | full_text = lambda: fl._revlog.rawdata(filerev) |
|
568 | full_text = lambda: fl._revlog.rawdata(filerev) | |
569 | parent_revs = lambda: fl._revlog.parentrevs(filerev) |
|
569 | parent_revs = lambda: fl._revlog.parentrevs(filerev) | |
570 | return _is_revision_affected_inner( |
|
570 | return _is_revision_affected_inner( | |
571 | full_text, parent_revs, filerev, metadata_cache |
|
571 | full_text, parent_revs, filerev, metadata_cache | |
572 | ) |
|
572 | ) | |
573 |
|
573 | |||
574 |
|
574 | |||
575 | def _is_revision_affected_inner( |
|
575 | def _is_revision_affected_inner( | |
576 | full_text, |
|
576 | full_text, | |
577 | parents_revs, |
|
577 | parents_revs, | |
578 | filerev, |
|
578 | filerev, | |
579 | metadata_cache=None, |
|
579 | metadata_cache=None, | |
580 | ): |
|
580 | ): | |
581 | """Mercurial currently (5.9rc0) uses `p1 == nullrev and p2 != nullrev` as a |
|
581 | """Mercurial currently (5.9rc0) uses `p1 == nullrev and p2 != nullrev` as a | |
582 | special meaning compared to the reverse in the context of filelog-based |
|
582 | special meaning compared to the reverse in the context of filelog-based | |
583 | copytracing. issue6528 exists because new code assumed that parent ordering |
|
583 | copytracing. issue6528 exists because new code assumed that parent ordering | |
584 | didn't matter, so this detects if the revision contains metadata (since |
|
584 | didn't matter, so this detects if the revision contains metadata (since | |
585 | it's only used for filelog-based copytracing) and its parents are in the |
|
585 | it's only used for filelog-based copytracing) and its parents are in the | |
586 | "wrong" order.""" |
|
586 | "wrong" order.""" | |
587 | try: |
|
587 | try: | |
588 | raw_text = full_text() |
|
588 | raw_text = full_text() | |
589 | except error.CensoredNodeError: |
|
589 | except error.CensoredNodeError: | |
590 | # We don't care about censored nodes as they never carry metadata |
|
590 | # We don't care about censored nodes as they never carry metadata | |
591 | return False |
|
591 | return False | |
592 | has_meta = raw_text.startswith(b'\x01\n') |
|
592 | has_meta = raw_text.startswith(b'\x01\n') | |
593 | if metadata_cache is not None: |
|
593 | if metadata_cache is not None: | |
594 | metadata_cache[filerev] = has_meta |
|
594 | metadata_cache[filerev] = has_meta | |
595 | if has_meta: |
|
595 | if has_meta: | |
596 | (p1, p2) = parents_revs() |
|
596 | (p1, p2) = parents_revs() | |
597 | if p1 != nullrev and p2 == nullrev: |
|
597 | if p1 != nullrev and p2 == nullrev: | |
598 | return True |
|
598 | return True | |
599 | return False |
|
599 | return False | |
600 |
|
600 | |||
601 |
|
601 | |||
602 | def _is_revision_affected_fast(repo, fl, filerev, metadata_cache): |
|
602 | def _is_revision_affected_fast(repo, fl, filerev, metadata_cache): | |
603 | rl = fl._revlog |
|
603 | rl = fl._revlog | |
604 | is_censored = lambda: rl.iscensored(filerev) |
|
604 | is_censored = lambda: rl.iscensored(filerev) | |
605 | delta_base = lambda: rl.deltaparent(filerev) |
|
605 | delta_base = lambda: rl.deltaparent(filerev) | |
606 | delta = lambda: rl._chunk(filerev) |
|
606 | delta = lambda: rl._chunk(filerev) | |
607 | full_text = lambda: rl.rawdata(filerev) |
|
607 | full_text = lambda: rl.rawdata(filerev) | |
608 | parent_revs = lambda: rl.parentrevs(filerev) |
|
608 | parent_revs = lambda: rl.parentrevs(filerev) | |
609 | return _is_revision_affected_fast_inner( |
|
609 | return _is_revision_affected_fast_inner( | |
610 | is_censored, |
|
610 | is_censored, | |
611 | delta_base, |
|
611 | delta_base, | |
612 | delta, |
|
612 | delta, | |
613 | full_text, |
|
613 | full_text, | |
614 | parent_revs, |
|
614 | parent_revs, | |
615 | filerev, |
|
615 | filerev, | |
616 | metadata_cache, |
|
616 | metadata_cache, | |
617 | ) |
|
617 | ) | |
618 |
|
618 | |||
619 |
|
619 | |||
620 | def _is_revision_affected_fast_inner( |
|
620 | def _is_revision_affected_fast_inner( | |
621 | is_censored, |
|
621 | is_censored, | |
622 | delta_base, |
|
622 | delta_base, | |
623 | delta, |
|
623 | delta, | |
624 | full_text, |
|
624 | full_text, | |
625 | parent_revs, |
|
625 | parent_revs, | |
626 | filerev, |
|
626 | filerev, | |
627 | metadata_cache, |
|
627 | metadata_cache, | |
628 | ): |
|
628 | ): | |
629 | """Optimization fast-path for `_is_revision_affected`. |
|
629 | """Optimization fast-path for `_is_revision_affected`. | |
630 |
|
630 | |||
631 | `metadata_cache` is a dict of `{rev: has_metadata}` which allows any |
|
631 | `metadata_cache` is a dict of `{rev: has_metadata}` which allows any | |
632 | revision to check if its base has metadata, saving computation of the full |
|
632 | revision to check if its base has metadata, saving computation of the full | |
633 | text, instead looking at the current delta. |
|
633 | text, instead looking at the current delta. | |
634 |
|
634 | |||
635 | This optimization only works if the revisions are looked at in order.""" |
|
635 | This optimization only works if the revisions are looked at in order.""" | |
636 |
|
636 | |||
637 | if is_censored(): |
|
637 | if is_censored(): | |
638 | # Censored revisions don't contain metadata, so they cannot be affected |
|
638 | # Censored revisions don't contain metadata, so they cannot be affected | |
639 | metadata_cache[filerev] = False |
|
639 | metadata_cache[filerev] = False | |
640 | return False |
|
640 | return False | |
641 |
|
641 | |||
642 | p1, p2 = parent_revs() |
|
642 | p1, p2 = parent_revs() | |
643 | if p1 == nullrev or p2 != nullrev: |
|
643 | if p1 == nullrev or p2 != nullrev: | |
644 | return False |
|
644 | return False | |
645 |
|
645 | |||
646 | delta_parent = delta_base() |
|
646 | delta_parent = delta_base() | |
647 | parent_has_metadata = metadata_cache.get(delta_parent) |
|
647 | parent_has_metadata = metadata_cache.get(delta_parent) | |
648 | if parent_has_metadata is None: |
|
648 | if parent_has_metadata is None: | |
649 | return _is_revision_affected_inner( |
|
649 | return _is_revision_affected_inner( | |
650 | full_text, |
|
650 | full_text, | |
651 | parent_revs, |
|
651 | parent_revs, | |
652 | filerev, |
|
652 | filerev, | |
653 | metadata_cache, |
|
653 | metadata_cache, | |
654 | ) |
|
654 | ) | |
655 |
|
655 | |||
656 | chunk = delta() |
|
656 | chunk = delta() | |
657 | if not len(chunk): |
|
657 | if not len(chunk): | |
658 | # No diff for this revision |
|
658 | # No diff for this revision | |
659 | return parent_has_metadata |
|
659 | return parent_has_metadata | |
660 |
|
660 | |||
661 | header_length = 12 |
|
661 | header_length = 12 | |
662 | if len(chunk) < header_length: |
|
662 | if len(chunk) < header_length: | |
663 | raise error.Abort(_(b"patch cannot be decoded")) |
|
663 | raise error.Abort(_(b"patch cannot be decoded")) | |
664 |
|
664 | |||
665 | start, _end, _length = struct.unpack(b">lll", chunk[:header_length]) |
|
665 | start, _end, _length = struct.unpack(b">lll", chunk[:header_length]) | |
666 |
|
666 | |||
667 | if start < 2: # len(b'\x01\n') == 2 |
|
667 | if start < 2: # len(b'\x01\n') == 2 | |
668 | # This delta does *something* to the metadata marker (if any). |
|
668 | # This delta does *something* to the metadata marker (if any). | |
669 | # Check it the slow way |
|
669 | # Check it the slow way | |
670 | is_affected = _is_revision_affected_inner( |
|
670 | is_affected = _is_revision_affected_inner( | |
671 | full_text, |
|
671 | full_text, | |
672 | parent_revs, |
|
672 | parent_revs, | |
673 | filerev, |
|
673 | filerev, | |
674 | metadata_cache, |
|
674 | metadata_cache, | |
675 | ) |
|
675 | ) | |
676 | return is_affected |
|
676 | return is_affected | |
677 |
|
677 | |||
678 | # The diff did not remove or add the metadata header, it's then in the same |
|
678 | # The diff did not remove or add the metadata header, it's then in the same | |
679 | # situation as its parent |
|
679 | # situation as its parent | |
680 | metadata_cache[filerev] = parent_has_metadata |
|
680 | metadata_cache[filerev] = parent_has_metadata | |
681 | return parent_has_metadata |
|
681 | return parent_has_metadata | |
682 |
|
682 | |||
683 |
|
683 | |||
684 | def _from_report(ui, repo, context, from_report, dry_run): |
|
684 | def _from_report(ui, repo, context, from_report, dry_run): | |
685 | """ |
|
685 | """ | |
686 | Fix the revisions given in the `from_report` file, but still checks if the |
|
686 | Fix the revisions given in the `from_report` file, but still checks if the | |
687 | revisions are indeed affected to prevent an unfortunate cyclic situation |
|
687 | revisions are indeed affected to prevent an unfortunate cyclic situation | |
688 | where we'd swap well-ordered parents again. |
|
688 | where we'd swap well-ordered parents again. | |
689 |
|
689 | |||
690 | See the doc for `debug_fix_issue6528` for the format documentation. |
|
690 | See the doc for `debug_fix_issue6528` for the format documentation. | |
691 | """ |
|
691 | """ | |
692 | ui.write(_(b"loading report file '%s'\n") % from_report) |
|
692 | ui.write(_(b"loading report file '%s'\n") % from_report) | |
693 |
|
693 | |||
694 | with context(), open(from_report, mode='rb') as f: |
|
694 | with context(), open(from_report, mode='rb') as f: | |
695 | for line in f.read().split(b'\n'): |
|
695 | for line in f.read().split(b'\n'): | |
696 | if not line: |
|
696 | if not line: | |
697 | continue |
|
697 | continue | |
698 | filenodes, filename = line.split(b' ', 1) |
|
698 | filenodes, filename = line.split(b' ', 1) | |
699 | fl = _filelog_from_filename(repo, filename) |
|
699 | fl = _filelog_from_filename(repo, filename) | |
700 | to_fix = set( |
|
700 | to_fix = set( | |
701 | fl.rev(binascii.unhexlify(n)) for n in filenodes.split(b',') |
|
701 | fl.rev(binascii.unhexlify(n)) for n in filenodes.split(b',') | |
702 | ) |
|
702 | ) | |
703 | excluded = set() |
|
703 | excluded = set() | |
704 |
|
704 | |||
705 | for filerev in to_fix: |
|
705 | for filerev in to_fix: | |
706 | if _is_revision_affected(fl, filerev): |
|
706 | if _is_revision_affected(fl, filerev): | |
707 | msg = b"found affected revision %d for filelog '%s'\n" |
|
707 | msg = b"found affected revision %d for filelog '%s'\n" | |
708 | ui.warn(msg % (filerev, filename)) |
|
708 | ui.warn(msg % (filerev, filename)) | |
709 | else: |
|
709 | else: | |
710 | msg = _(b"revision %s of file '%s' is not affected\n") |
|
710 | msg = _(b"revision %s of file '%s' is not affected\n") | |
711 | msg %= (binascii.hexlify(fl.node(filerev)), filename) |
|
711 | msg %= (binascii.hexlify(fl.node(filerev)), filename) | |
712 | ui.warn(msg) |
|
712 | ui.warn(msg) | |
713 | excluded.add(filerev) |
|
713 | excluded.add(filerev) | |
714 |
|
714 | |||
715 | to_fix = to_fix - excluded |
|
715 | to_fix = to_fix - excluded | |
716 | if not to_fix: |
|
716 | if not to_fix: | |
717 | msg = _(b"no affected revisions were found for '%s'\n") |
|
717 | msg = _(b"no affected revisions were found for '%s'\n") | |
718 | ui.write(msg % filename) |
|
718 | ui.write(msg % filename) | |
719 | continue |
|
719 | continue | |
720 | if not dry_run: |
|
720 | if not dry_run: | |
721 | _reorder_filelog_parents(repo, fl, sorted(to_fix)) |
|
721 | _reorder_filelog_parents(repo, fl, sorted(to_fix)) | |
722 |
|
722 | |||
723 |
|
723 | |||
724 | def filter_delta_issue6528(revlog, deltas_iter): |
|
724 | def filter_delta_issue6528(revlog, deltas_iter): | |
725 | """filter incomind deltas to repaire issue 6528 on the fly""" |
|
725 | """filter incomind deltas to repaire issue 6528 on the fly""" | |
726 | metadata_cache = {} |
|
726 | metadata_cache = {} | |
727 |
|
727 | |||
728 | deltacomputer = deltas.deltacomputer(revlog) |
|
728 | deltacomputer = deltas.deltacomputer(revlog) | |
729 |
|
729 | |||
730 | for rev, d in enumerate(deltas_iter, len(revlog)): |
|
730 | for rev, d in enumerate(deltas_iter, len(revlog)): | |
731 | ( |
|
731 | ( | |
732 | node, |
|
732 | node, | |
733 | p1_node, |
|
733 | p1_node, | |
734 | p2_node, |
|
734 | p2_node, | |
735 | linknode, |
|
735 | linknode, | |
736 | deltabase, |
|
736 | deltabase, | |
737 | delta, |
|
737 | delta, | |
738 | flags, |
|
738 | flags, | |
739 | sidedata, |
|
739 | sidedata, | |
740 | ) = d |
|
740 | ) = d | |
741 |
|
741 | |||
742 | if not revlog.index.has_node(deltabase): |
|
742 | if not revlog.index.has_node(deltabase): | |
743 | raise error.LookupError( |
|
743 | raise error.LookupError( | |
744 | deltabase, revlog.radix, _(b'unknown parent') |
|
744 | deltabase, revlog.radix, _(b'unknown parent') | |
745 | ) |
|
745 | ) | |
746 | base_rev = revlog.rev(deltabase) |
|
746 | base_rev = revlog.rev(deltabase) | |
747 | if not revlog.index.has_node(p1_node): |
|
747 | if not revlog.index.has_node(p1_node): | |
748 | raise error.LookupError(p1_node, revlog.radix, _(b'unknown parent')) |
|
748 | raise error.LookupError(p1_node, revlog.radix, _(b'unknown parent')) | |
749 | p1_rev = revlog.rev(p1_node) |
|
749 | p1_rev = revlog.rev(p1_node) | |
750 | if not revlog.index.has_node(p2_node): |
|
750 | if not revlog.index.has_node(p2_node): | |
751 | raise error.LookupError(p2_node, revlog.radix, _(b'unknown parent')) |
|
751 | raise error.LookupError(p2_node, revlog.radix, _(b'unknown parent')) | |
752 | p2_rev = revlog.rev(p2_node) |
|
752 | p2_rev = revlog.rev(p2_node) | |
753 |
|
753 | |||
754 | is_censored = lambda: bool(flags & REVIDX_ISCENSORED) |
|
754 | is_censored = lambda: bool(flags & REVIDX_ISCENSORED) | |
755 | delta_base = lambda: revlog.rev(delta_base) |
|
755 | delta_base = lambda: revlog.rev(delta_base) | |
756 | delta_base = lambda: base_rev |
|
756 | delta_base = lambda: base_rev | |
757 | parent_revs = lambda: (p1_rev, p2_rev) |
|
757 | parent_revs = lambda: (p1_rev, p2_rev) | |
758 |
|
758 | |||
759 | def full_text(): |
|
759 | def full_text(): | |
760 | # note: being able to reuse the full text computation in the |
|
760 | # note: being able to reuse the full text computation in the | |
761 | # underlying addrevision would be useful however this is a bit too |
|
761 | # underlying addrevision would be useful however this is a bit too | |
762 | # intrusive the for the "quick" issue6528 we are writing before the |
|
762 | # intrusive the for the "quick" issue6528 we are writing before the | |
763 | # 5.8 release |
|
763 | # 5.8 release | |
764 | textlen = mdiff.patchedsize(revlog.size(base_rev), delta) |
|
764 | textlen = mdiff.patchedsize(revlog.size(base_rev), delta) | |
765 |
|
765 | |||
766 | revinfo = revlogutils.revisioninfo( |
|
766 | revinfo = revlogutils.revisioninfo( | |
767 | node, |
|
767 | node, | |
768 | p1_node, |
|
768 | p1_node, | |
769 | p2_node, |
|
769 | p2_node, | |
770 | [None], |
|
770 | [None], | |
771 | textlen, |
|
771 | textlen, | |
772 | (base_rev, delta), |
|
772 | (base_rev, delta), | |
773 | flags, |
|
773 | flags, | |
774 | ) |
|
774 | ) | |
775 | # cached by the global "writing" context |
|
775 | # cached by the global "writing" context | |
776 | assert revlog._writinghandles is not None |
|
776 | assert revlog._writinghandles is not None | |
777 | if revlog._inline: |
|
777 | if revlog._inline: | |
778 | fh = revlog._writinghandles[0] |
|
778 | fh = revlog._writinghandles[0] | |
779 | else: |
|
779 | else: | |
780 | fh = revlog._writinghandles[1] |
|
780 | fh = revlog._writinghandles[1] | |
781 | return deltacomputer.buildtext(revinfo, fh) |
|
781 | return deltacomputer.buildtext(revinfo, fh) | |
782 |
|
782 | |||
783 | is_affected = _is_revision_affected_fast_inner( |
|
783 | is_affected = _is_revision_affected_fast_inner( | |
784 | is_censored, |
|
784 | is_censored, | |
785 | delta_base, |
|
785 | delta_base, | |
786 | lambda: delta, |
|
786 | lambda: delta, | |
787 | full_text, |
|
787 | full_text, | |
788 | parent_revs, |
|
788 | parent_revs, | |
789 | rev, |
|
789 | rev, | |
790 | metadata_cache, |
|
790 | metadata_cache, | |
791 | ) |
|
791 | ) | |
792 | if is_affected: |
|
792 | if is_affected: | |
793 | d = ( |
|
793 | d = ( | |
794 | node, |
|
794 | node, | |
795 | p2_node, |
|
795 | p2_node, | |
796 | p1_node, |
|
796 | p1_node, | |
797 | linknode, |
|
797 | linknode, | |
798 | deltabase, |
|
798 | deltabase, | |
799 | delta, |
|
799 | delta, | |
800 | flags, |
|
800 | flags, | |
801 | sidedata, |
|
801 | sidedata, | |
802 | ) |
|
802 | ) | |
803 | yield d |
|
803 | yield d | |
804 |
|
804 | |||
805 |
|
805 | |||
806 | def repair_issue6528( |
|
806 | def repair_issue6528( | |
807 | ui, repo, dry_run=False, to_report=None, from_report=None, paranoid=False |
|
807 | ui, repo, dry_run=False, to_report=None, from_report=None, paranoid=False | |
808 | ): |
|
808 | ): | |
809 | from .. import store # avoid cycle |
|
809 | from .. import store # avoid cycle | |
810 |
|
810 | |||
811 | @contextlib.contextmanager |
|
811 | @contextlib.contextmanager | |
812 | def context(): |
|
812 | def context(): | |
813 | if dry_run or to_report: # No need for locking |
|
813 | if dry_run or to_report: # No need for locking | |
814 | yield |
|
814 | yield | |
815 | else: |
|
815 | else: | |
816 | with repo.wlock(), repo.lock(): |
|
816 | with repo.wlock(), repo.lock(): | |
817 | yield |
|
817 | yield | |
818 |
|
818 | |||
819 | if from_report: |
|
819 | if from_report: | |
820 | return _from_report(ui, repo, context, from_report, dry_run) |
|
820 | return _from_report(ui, repo, context, from_report, dry_run) | |
821 |
|
821 | |||
822 | report_entries = [] |
|
822 | report_entries = [] | |
823 |
|
823 | |||
824 | with context(): |
|
824 | with context(): | |
825 | files = list( |
|
825 | files = list( | |
826 | (file_type, path) |
|
826 | (file_type, path) | |
827 |
for (file_type, path |
|
827 | for (file_type, path, _s) in repo.store.datafiles() | |
828 | if path.endswith(b'.i') and file_type & store.FILEFLAGS_FILELOG |
|
828 | if path.endswith(b'.i') and file_type & store.FILEFLAGS_FILELOG | |
829 | ) |
|
829 | ) | |
830 |
|
830 | |||
831 | progress = ui.makeprogress( |
|
831 | progress = ui.makeprogress( | |
832 | _(b"looking for affected revisions"), |
|
832 | _(b"looking for affected revisions"), | |
833 | unit=_(b"filelogs"), |
|
833 | unit=_(b"filelogs"), | |
834 | total=len(files), |
|
834 | total=len(files), | |
835 | ) |
|
835 | ) | |
836 | found_nothing = True |
|
836 | found_nothing = True | |
837 |
|
837 | |||
838 | for file_type, path in files: |
|
838 | for file_type, path in files: | |
839 | if ( |
|
839 | if ( | |
840 | not path.endswith(b'.i') |
|
840 | not path.endswith(b'.i') | |
841 | or not file_type & store.FILEFLAGS_FILELOG |
|
841 | or not file_type & store.FILEFLAGS_FILELOG | |
842 | ): |
|
842 | ): | |
843 | continue |
|
843 | continue | |
844 | progress.increment() |
|
844 | progress.increment() | |
845 | filename = _get_filename_from_filelog_index(path) |
|
845 | filename = _get_filename_from_filelog_index(path) | |
846 | fl = _filelog_from_filename(repo, filename) |
|
846 | fl = _filelog_from_filename(repo, filename) | |
847 |
|
847 | |||
848 | # Set of filerevs (or hex filenodes if `to_report`) that need fixing |
|
848 | # Set of filerevs (or hex filenodes if `to_report`) that need fixing | |
849 | to_fix = set() |
|
849 | to_fix = set() | |
850 | metadata_cache = {} |
|
850 | metadata_cache = {} | |
851 | for filerev in fl.revs(): |
|
851 | for filerev in fl.revs(): | |
852 | affected = _is_revision_affected_fast( |
|
852 | affected = _is_revision_affected_fast( | |
853 | repo, fl, filerev, metadata_cache |
|
853 | repo, fl, filerev, metadata_cache | |
854 | ) |
|
854 | ) | |
855 | if paranoid: |
|
855 | if paranoid: | |
856 | slow = _is_revision_affected(fl, filerev) |
|
856 | slow = _is_revision_affected(fl, filerev) | |
857 | if slow != affected: |
|
857 | if slow != affected: | |
858 | msg = _(b"paranoid check failed for '%s' at node %s") |
|
858 | msg = _(b"paranoid check failed for '%s' at node %s") | |
859 | node = binascii.hexlify(fl.node(filerev)) |
|
859 | node = binascii.hexlify(fl.node(filerev)) | |
860 | raise error.Abort(msg % (filename, node)) |
|
860 | raise error.Abort(msg % (filename, node)) | |
861 | if affected: |
|
861 | if affected: | |
862 | msg = b"found affected revision %d for filelog '%s'\n" |
|
862 | msg = b"found affected revision %d for filelog '%s'\n" | |
863 | ui.warn(msg % (filerev, path)) |
|
863 | ui.warn(msg % (filerev, path)) | |
864 | found_nothing = False |
|
864 | found_nothing = False | |
865 | if not dry_run: |
|
865 | if not dry_run: | |
866 | if to_report: |
|
866 | if to_report: | |
867 | to_fix.add(binascii.hexlify(fl.node(filerev))) |
|
867 | to_fix.add(binascii.hexlify(fl.node(filerev))) | |
868 | else: |
|
868 | else: | |
869 | to_fix.add(filerev) |
|
869 | to_fix.add(filerev) | |
870 |
|
870 | |||
871 | if to_fix: |
|
871 | if to_fix: | |
872 | to_fix = sorted(to_fix) |
|
872 | to_fix = sorted(to_fix) | |
873 | if to_report: |
|
873 | if to_report: | |
874 | report_entries.append((filename, to_fix)) |
|
874 | report_entries.append((filename, to_fix)) | |
875 | else: |
|
875 | else: | |
876 | _reorder_filelog_parents(repo, fl, to_fix) |
|
876 | _reorder_filelog_parents(repo, fl, to_fix) | |
877 |
|
877 | |||
878 | if found_nothing: |
|
878 | if found_nothing: | |
879 | ui.write(_(b"no affected revisions were found\n")) |
|
879 | ui.write(_(b"no affected revisions were found\n")) | |
880 |
|
880 | |||
881 | if to_report and report_entries: |
|
881 | if to_report and report_entries: | |
882 | with open(to_report, mode="wb") as f: |
|
882 | with open(to_report, mode="wb") as f: | |
883 | for path, to_fix in report_entries: |
|
883 | for path, to_fix in report_entries: | |
884 | f.write(b"%s %s\n" % (b",".join(to_fix), path)) |
|
884 | f.write(b"%s %s\n" % (b",".join(to_fix), path)) | |
885 |
|
885 | |||
886 | progress.complete() |
|
886 | progress.complete() |
@@ -1,839 +1,850 b'' | |||||
1 | # store.py - repository store handling for Mercurial |
|
1 | # store.py - repository store handling for Mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2008 Olivia Mackall <olivia@selenic.com> |
|
3 | # Copyright 2008 Olivia Mackall <olivia@selenic.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 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import errno |
|
10 | import errno | |
11 | import functools |
|
11 | import functools | |
12 | import os |
|
12 | import os | |
13 | import re |
|
13 | import re | |
14 | import stat |
|
14 | import stat | |
15 |
|
15 | |||
16 | from .i18n import _ |
|
16 | from .i18n import _ | |
17 | from .pycompat import getattr |
|
17 | from .pycompat import getattr | |
18 | from .node import hex |
|
18 | from .node import hex | |
19 | from . import ( |
|
19 | from . import ( | |
20 | changelog, |
|
20 | changelog, | |
21 | error, |
|
21 | error, | |
22 | manifest, |
|
22 | manifest, | |
23 | policy, |
|
23 | policy, | |
24 | pycompat, |
|
24 | pycompat, | |
25 | util, |
|
25 | util, | |
26 | vfs as vfsmod, |
|
26 | vfs as vfsmod, | |
27 | ) |
|
27 | ) | |
28 | from .utils import hashutil |
|
28 | from .utils import hashutil | |
29 |
|
29 | |||
30 | parsers = policy.importmod('parsers') |
|
30 | parsers = policy.importmod('parsers') | |
31 | # how much bytes should be read from fncache in one read |
|
31 | # how much bytes should be read from fncache in one read | |
32 | # It is done to prevent loading large fncache files into memory |
|
32 | # It is done to prevent loading large fncache files into memory | |
33 | fncache_chunksize = 10 ** 6 |
|
33 | fncache_chunksize = 10 ** 6 | |
34 |
|
34 | |||
35 |
|
35 | |||
36 | def _matchtrackedpath(path, matcher): |
|
36 | def _matchtrackedpath(path, matcher): | |
37 | """parses a fncache entry and returns whether the entry is tracking a path |
|
37 | """parses a fncache entry and returns whether the entry is tracking a path | |
38 | matched by matcher or not. |
|
38 | matched by matcher or not. | |
39 |
|
39 | |||
40 | If matcher is None, returns True""" |
|
40 | If matcher is None, returns True""" | |
41 |
|
41 | |||
42 | if matcher is None: |
|
42 | if matcher is None: | |
43 | return True |
|
43 | return True | |
44 | path = decodedir(path) |
|
44 | path = decodedir(path) | |
45 | if path.startswith(b'data/'): |
|
45 | if path.startswith(b'data/'): | |
46 | return matcher(path[len(b'data/') : -len(b'.i')]) |
|
46 | return matcher(path[len(b'data/') : -len(b'.i')]) | |
47 | elif path.startswith(b'meta/'): |
|
47 | elif path.startswith(b'meta/'): | |
48 | return matcher.visitdir(path[len(b'meta/') : -len(b'/00manifest.i')]) |
|
48 | return matcher.visitdir(path[len(b'meta/') : -len(b'/00manifest.i')]) | |
49 |
|
49 | |||
50 | raise error.ProgrammingError(b"cannot decode path %s" % path) |
|
50 | raise error.ProgrammingError(b"cannot decode path %s" % path) | |
51 |
|
51 | |||
52 |
|
52 | |||
53 | # This avoids a collision between a file named foo and a dir named |
|
53 | # This avoids a collision between a file named foo and a dir named | |
54 | # foo.i or foo.d |
|
54 | # foo.i or foo.d | |
55 | def _encodedir(path): |
|
55 | def _encodedir(path): | |
56 | """ |
|
56 | """ | |
57 | >>> _encodedir(b'data/foo.i') |
|
57 | >>> _encodedir(b'data/foo.i') | |
58 | 'data/foo.i' |
|
58 | 'data/foo.i' | |
59 | >>> _encodedir(b'data/foo.i/bla.i') |
|
59 | >>> _encodedir(b'data/foo.i/bla.i') | |
60 | 'data/foo.i.hg/bla.i' |
|
60 | 'data/foo.i.hg/bla.i' | |
61 | >>> _encodedir(b'data/foo.i.hg/bla.i') |
|
61 | >>> _encodedir(b'data/foo.i.hg/bla.i') | |
62 | 'data/foo.i.hg.hg/bla.i' |
|
62 | 'data/foo.i.hg.hg/bla.i' | |
63 | >>> _encodedir(b'data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n') |
|
63 | >>> _encodedir(b'data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n') | |
64 | 'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n' |
|
64 | 'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n' | |
65 | """ |
|
65 | """ | |
66 | return ( |
|
66 | return ( | |
67 | path.replace(b".hg/", b".hg.hg/") |
|
67 | path.replace(b".hg/", b".hg.hg/") | |
68 | .replace(b".i/", b".i.hg/") |
|
68 | .replace(b".i/", b".i.hg/") | |
69 | .replace(b".d/", b".d.hg/") |
|
69 | .replace(b".d/", b".d.hg/") | |
70 | ) |
|
70 | ) | |
71 |
|
71 | |||
72 |
|
72 | |||
73 | encodedir = getattr(parsers, 'encodedir', _encodedir) |
|
73 | encodedir = getattr(parsers, 'encodedir', _encodedir) | |
74 |
|
74 | |||
75 |
|
75 | |||
76 | def decodedir(path): |
|
76 | def decodedir(path): | |
77 | """ |
|
77 | """ | |
78 | >>> decodedir(b'data/foo.i') |
|
78 | >>> decodedir(b'data/foo.i') | |
79 | 'data/foo.i' |
|
79 | 'data/foo.i' | |
80 | >>> decodedir(b'data/foo.i.hg/bla.i') |
|
80 | >>> decodedir(b'data/foo.i.hg/bla.i') | |
81 | 'data/foo.i/bla.i' |
|
81 | 'data/foo.i/bla.i' | |
82 | >>> decodedir(b'data/foo.i.hg.hg/bla.i') |
|
82 | >>> decodedir(b'data/foo.i.hg.hg/bla.i') | |
83 | 'data/foo.i.hg/bla.i' |
|
83 | 'data/foo.i.hg/bla.i' | |
84 | """ |
|
84 | """ | |
85 | if b".hg/" not in path: |
|
85 | if b".hg/" not in path: | |
86 | return path |
|
86 | return path | |
87 | return ( |
|
87 | return ( | |
88 | path.replace(b".d.hg/", b".d/") |
|
88 | path.replace(b".d.hg/", b".d/") | |
89 | .replace(b".i.hg/", b".i/") |
|
89 | .replace(b".i.hg/", b".i/") | |
90 | .replace(b".hg.hg/", b".hg/") |
|
90 | .replace(b".hg.hg/", b".hg/") | |
91 | ) |
|
91 | ) | |
92 |
|
92 | |||
93 |
|
93 | |||
94 | def _reserved(): |
|
94 | def _reserved(): | |
95 | """characters that are problematic for filesystems |
|
95 | """characters that are problematic for filesystems | |
96 |
|
96 | |||
97 | * ascii escapes (0..31) |
|
97 | * ascii escapes (0..31) | |
98 | * ascii hi (126..255) |
|
98 | * ascii hi (126..255) | |
99 | * windows specials |
|
99 | * windows specials | |
100 |
|
100 | |||
101 | these characters will be escaped by encodefunctions |
|
101 | these characters will be escaped by encodefunctions | |
102 | """ |
|
102 | """ | |
103 | winreserved = [ord(x) for x in u'\\:*?"<>|'] |
|
103 | winreserved = [ord(x) for x in u'\\:*?"<>|'] | |
104 | for x in range(32): |
|
104 | for x in range(32): | |
105 | yield x |
|
105 | yield x | |
106 | for x in range(126, 256): |
|
106 | for x in range(126, 256): | |
107 | yield x |
|
107 | yield x | |
108 | for x in winreserved: |
|
108 | for x in winreserved: | |
109 | yield x |
|
109 | yield x | |
110 |
|
110 | |||
111 |
|
111 | |||
112 | def _buildencodefun(): |
|
112 | def _buildencodefun(): | |
113 | """ |
|
113 | """ | |
114 | >>> enc, dec = _buildencodefun() |
|
114 | >>> enc, dec = _buildencodefun() | |
115 |
|
115 | |||
116 | >>> enc(b'nothing/special.txt') |
|
116 | >>> enc(b'nothing/special.txt') | |
117 | 'nothing/special.txt' |
|
117 | 'nothing/special.txt' | |
118 | >>> dec(b'nothing/special.txt') |
|
118 | >>> dec(b'nothing/special.txt') | |
119 | 'nothing/special.txt' |
|
119 | 'nothing/special.txt' | |
120 |
|
120 | |||
121 | >>> enc(b'HELLO') |
|
121 | >>> enc(b'HELLO') | |
122 | '_h_e_l_l_o' |
|
122 | '_h_e_l_l_o' | |
123 | >>> dec(b'_h_e_l_l_o') |
|
123 | >>> dec(b'_h_e_l_l_o') | |
124 | 'HELLO' |
|
124 | 'HELLO' | |
125 |
|
125 | |||
126 | >>> enc(b'hello:world?') |
|
126 | >>> enc(b'hello:world?') | |
127 | 'hello~3aworld~3f' |
|
127 | 'hello~3aworld~3f' | |
128 | >>> dec(b'hello~3aworld~3f') |
|
128 | >>> dec(b'hello~3aworld~3f') | |
129 | 'hello:world?' |
|
129 | 'hello:world?' | |
130 |
|
130 | |||
131 | >>> enc(b'the\\x07quick\\xADshot') |
|
131 | >>> enc(b'the\\x07quick\\xADshot') | |
132 | 'the~07quick~adshot' |
|
132 | 'the~07quick~adshot' | |
133 | >>> dec(b'the~07quick~adshot') |
|
133 | >>> dec(b'the~07quick~adshot') | |
134 | 'the\\x07quick\\xadshot' |
|
134 | 'the\\x07quick\\xadshot' | |
135 | """ |
|
135 | """ | |
136 | e = b'_' |
|
136 | e = b'_' | |
137 | xchr = pycompat.bytechr |
|
137 | xchr = pycompat.bytechr | |
138 | asciistr = list(map(xchr, range(127))) |
|
138 | asciistr = list(map(xchr, range(127))) | |
139 | capitals = list(range(ord(b"A"), ord(b"Z") + 1)) |
|
139 | capitals = list(range(ord(b"A"), ord(b"Z") + 1)) | |
140 |
|
140 | |||
141 | cmap = {x: x for x in asciistr} |
|
141 | cmap = {x: x for x in asciistr} | |
142 | for x in _reserved(): |
|
142 | for x in _reserved(): | |
143 | cmap[xchr(x)] = b"~%02x" % x |
|
143 | cmap[xchr(x)] = b"~%02x" % x | |
144 | for x in capitals + [ord(e)]: |
|
144 | for x in capitals + [ord(e)]: | |
145 | cmap[xchr(x)] = e + xchr(x).lower() |
|
145 | cmap[xchr(x)] = e + xchr(x).lower() | |
146 |
|
146 | |||
147 | dmap = {} |
|
147 | dmap = {} | |
148 | for k, v in pycompat.iteritems(cmap): |
|
148 | for k, v in pycompat.iteritems(cmap): | |
149 | dmap[v] = k |
|
149 | dmap[v] = k | |
150 |
|
150 | |||
151 | def decode(s): |
|
151 | def decode(s): | |
152 | i = 0 |
|
152 | i = 0 | |
153 | while i < len(s): |
|
153 | while i < len(s): | |
154 | for l in pycompat.xrange(1, 4): |
|
154 | for l in pycompat.xrange(1, 4): | |
155 | try: |
|
155 | try: | |
156 | yield dmap[s[i : i + l]] |
|
156 | yield dmap[s[i : i + l]] | |
157 | i += l |
|
157 | i += l | |
158 | break |
|
158 | break | |
159 | except KeyError: |
|
159 | except KeyError: | |
160 | pass |
|
160 | pass | |
161 | else: |
|
161 | else: | |
162 | raise KeyError |
|
162 | raise KeyError | |
163 |
|
163 | |||
164 | return ( |
|
164 | return ( | |
165 | lambda s: b''.join( |
|
165 | lambda s: b''.join( | |
166 | [cmap[s[c : c + 1]] for c in pycompat.xrange(len(s))] |
|
166 | [cmap[s[c : c + 1]] for c in pycompat.xrange(len(s))] | |
167 | ), |
|
167 | ), | |
168 | lambda s: b''.join(list(decode(s))), |
|
168 | lambda s: b''.join(list(decode(s))), | |
169 | ) |
|
169 | ) | |
170 |
|
170 | |||
171 |
|
171 | |||
172 | _encodefname, _decodefname = _buildencodefun() |
|
172 | _encodefname, _decodefname = _buildencodefun() | |
173 |
|
173 | |||
174 |
|
174 | |||
175 | def encodefilename(s): |
|
175 | def encodefilename(s): | |
176 | """ |
|
176 | """ | |
177 | >>> encodefilename(b'foo.i/bar.d/bla.hg/hi:world?/HELLO') |
|
177 | >>> encodefilename(b'foo.i/bar.d/bla.hg/hi:world?/HELLO') | |
178 | 'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o' |
|
178 | 'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o' | |
179 | """ |
|
179 | """ | |
180 | return _encodefname(encodedir(s)) |
|
180 | return _encodefname(encodedir(s)) | |
181 |
|
181 | |||
182 |
|
182 | |||
183 | def decodefilename(s): |
|
183 | def decodefilename(s): | |
184 | """ |
|
184 | """ | |
185 | >>> decodefilename(b'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o') |
|
185 | >>> decodefilename(b'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o') | |
186 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' |
|
186 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' | |
187 | """ |
|
187 | """ | |
188 | return decodedir(_decodefname(s)) |
|
188 | return decodedir(_decodefname(s)) | |
189 |
|
189 | |||
190 |
|
190 | |||
191 | def _buildlowerencodefun(): |
|
191 | def _buildlowerencodefun(): | |
192 | """ |
|
192 | """ | |
193 | >>> f = _buildlowerencodefun() |
|
193 | >>> f = _buildlowerencodefun() | |
194 | >>> f(b'nothing/special.txt') |
|
194 | >>> f(b'nothing/special.txt') | |
195 | 'nothing/special.txt' |
|
195 | 'nothing/special.txt' | |
196 | >>> f(b'HELLO') |
|
196 | >>> f(b'HELLO') | |
197 | 'hello' |
|
197 | 'hello' | |
198 | >>> f(b'hello:world?') |
|
198 | >>> f(b'hello:world?') | |
199 | 'hello~3aworld~3f' |
|
199 | 'hello~3aworld~3f' | |
200 | >>> f(b'the\\x07quick\\xADshot') |
|
200 | >>> f(b'the\\x07quick\\xADshot') | |
201 | 'the~07quick~adshot' |
|
201 | 'the~07quick~adshot' | |
202 | """ |
|
202 | """ | |
203 | xchr = pycompat.bytechr |
|
203 | xchr = pycompat.bytechr | |
204 | cmap = {xchr(x): xchr(x) for x in pycompat.xrange(127)} |
|
204 | cmap = {xchr(x): xchr(x) for x in pycompat.xrange(127)} | |
205 | for x in _reserved(): |
|
205 | for x in _reserved(): | |
206 | cmap[xchr(x)] = b"~%02x" % x |
|
206 | cmap[xchr(x)] = b"~%02x" % x | |
207 | for x in range(ord(b"A"), ord(b"Z") + 1): |
|
207 | for x in range(ord(b"A"), ord(b"Z") + 1): | |
208 | cmap[xchr(x)] = xchr(x).lower() |
|
208 | cmap[xchr(x)] = xchr(x).lower() | |
209 |
|
209 | |||
210 | def lowerencode(s): |
|
210 | def lowerencode(s): | |
211 | return b"".join([cmap[c] for c in pycompat.iterbytestr(s)]) |
|
211 | return b"".join([cmap[c] for c in pycompat.iterbytestr(s)]) | |
212 |
|
212 | |||
213 | return lowerencode |
|
213 | return lowerencode | |
214 |
|
214 | |||
215 |
|
215 | |||
216 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() |
|
216 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() | |
217 |
|
217 | |||
218 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 |
|
218 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 | |
219 | _winres3 = (b'aux', b'con', b'prn', b'nul') # length 3 |
|
219 | _winres3 = (b'aux', b'con', b'prn', b'nul') # length 3 | |
220 | _winres4 = (b'com', b'lpt') # length 4 (with trailing 1..9) |
|
220 | _winres4 = (b'com', b'lpt') # length 4 (with trailing 1..9) | |
221 |
|
221 | |||
222 |
|
222 | |||
223 | def _auxencode(path, dotencode): |
|
223 | def _auxencode(path, dotencode): | |
224 | """ |
|
224 | """ | |
225 | Encodes filenames containing names reserved by Windows or which end in |
|
225 | Encodes filenames containing names reserved by Windows or which end in | |
226 | period or space. Does not touch other single reserved characters c. |
|
226 | period or space. Does not touch other single reserved characters c. | |
227 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. |
|
227 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. | |
228 | Additionally encodes space or period at the beginning, if dotencode is |
|
228 | Additionally encodes space or period at the beginning, if dotencode is | |
229 | True. Parameter path is assumed to be all lowercase. |
|
229 | True. Parameter path is assumed to be all lowercase. | |
230 | A segment only needs encoding if a reserved name appears as a |
|
230 | A segment only needs encoding if a reserved name appears as a | |
231 | basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux" |
|
231 | basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux" | |
232 | doesn't need encoding. |
|
232 | doesn't need encoding. | |
233 |
|
233 | |||
234 | >>> s = b'.foo/aux.txt/txt.aux/con/prn/nul/foo.' |
|
234 | >>> s = b'.foo/aux.txt/txt.aux/con/prn/nul/foo.' | |
235 | >>> _auxencode(s.split(b'/'), True) |
|
235 | >>> _auxencode(s.split(b'/'), True) | |
236 | ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e'] |
|
236 | ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e'] | |
237 | >>> s = b'.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' |
|
237 | >>> s = b'.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' | |
238 | >>> _auxencode(s.split(b'/'), False) |
|
238 | >>> _auxencode(s.split(b'/'), False) | |
239 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] |
|
239 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] | |
240 | >>> _auxencode([b'foo. '], True) |
|
240 | >>> _auxencode([b'foo. '], True) | |
241 | ['foo.~20'] |
|
241 | ['foo.~20'] | |
242 | >>> _auxencode([b' .foo'], True) |
|
242 | >>> _auxencode([b' .foo'], True) | |
243 | ['~20.foo'] |
|
243 | ['~20.foo'] | |
244 | """ |
|
244 | """ | |
245 | for i, n in enumerate(path): |
|
245 | for i, n in enumerate(path): | |
246 | if not n: |
|
246 | if not n: | |
247 | continue |
|
247 | continue | |
248 | if dotencode and n[0] in b'. ': |
|
248 | if dotencode and n[0] in b'. ': | |
249 | n = b"~%02x" % ord(n[0:1]) + n[1:] |
|
249 | n = b"~%02x" % ord(n[0:1]) + n[1:] | |
250 | path[i] = n |
|
250 | path[i] = n | |
251 | else: |
|
251 | else: | |
252 | l = n.find(b'.') |
|
252 | l = n.find(b'.') | |
253 | if l == -1: |
|
253 | if l == -1: | |
254 | l = len(n) |
|
254 | l = len(n) | |
255 | if (l == 3 and n[:3] in _winres3) or ( |
|
255 | if (l == 3 and n[:3] in _winres3) or ( | |
256 | l == 4 |
|
256 | l == 4 | |
257 | and n[3:4] <= b'9' |
|
257 | and n[3:4] <= b'9' | |
258 | and n[3:4] >= b'1' |
|
258 | and n[3:4] >= b'1' | |
259 | and n[:3] in _winres4 |
|
259 | and n[:3] in _winres4 | |
260 | ): |
|
260 | ): | |
261 | # encode third letter ('aux' -> 'au~78') |
|
261 | # encode third letter ('aux' -> 'au~78') | |
262 | ec = b"~%02x" % ord(n[2:3]) |
|
262 | ec = b"~%02x" % ord(n[2:3]) | |
263 | n = n[0:2] + ec + n[3:] |
|
263 | n = n[0:2] + ec + n[3:] | |
264 | path[i] = n |
|
264 | path[i] = n | |
265 | if n[-1] in b'. ': |
|
265 | if n[-1] in b'. ': | |
266 | # encode last period or space ('foo...' -> 'foo..~2e') |
|
266 | # encode last period or space ('foo...' -> 'foo..~2e') | |
267 | path[i] = n[:-1] + b"~%02x" % ord(n[-1:]) |
|
267 | path[i] = n[:-1] + b"~%02x" % ord(n[-1:]) | |
268 | return path |
|
268 | return path | |
269 |
|
269 | |||
270 |
|
270 | |||
271 | _maxstorepathlen = 120 |
|
271 | _maxstorepathlen = 120 | |
272 | _dirprefixlen = 8 |
|
272 | _dirprefixlen = 8 | |
273 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 |
|
273 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 | |
274 |
|
274 | |||
275 |
|
275 | |||
276 | def _hashencode(path, dotencode): |
|
276 | def _hashencode(path, dotencode): | |
277 | digest = hex(hashutil.sha1(path).digest()) |
|
277 | digest = hex(hashutil.sha1(path).digest()) | |
278 | le = lowerencode(path[5:]).split(b'/') # skips prefix 'data/' or 'meta/' |
|
278 | le = lowerencode(path[5:]).split(b'/') # skips prefix 'data/' or 'meta/' | |
279 | parts = _auxencode(le, dotencode) |
|
279 | parts = _auxencode(le, dotencode) | |
280 | basename = parts[-1] |
|
280 | basename = parts[-1] | |
281 | _root, ext = os.path.splitext(basename) |
|
281 | _root, ext = os.path.splitext(basename) | |
282 | sdirs = [] |
|
282 | sdirs = [] | |
283 | sdirslen = 0 |
|
283 | sdirslen = 0 | |
284 | for p in parts[:-1]: |
|
284 | for p in parts[:-1]: | |
285 | d = p[:_dirprefixlen] |
|
285 | d = p[:_dirprefixlen] | |
286 | if d[-1] in b'. ': |
|
286 | if d[-1] in b'. ': | |
287 | # Windows can't access dirs ending in period or space |
|
287 | # Windows can't access dirs ending in period or space | |
288 | d = d[:-1] + b'_' |
|
288 | d = d[:-1] + b'_' | |
289 | if sdirslen == 0: |
|
289 | if sdirslen == 0: | |
290 | t = len(d) |
|
290 | t = len(d) | |
291 | else: |
|
291 | else: | |
292 | t = sdirslen + 1 + len(d) |
|
292 | t = sdirslen + 1 + len(d) | |
293 | if t > _maxshortdirslen: |
|
293 | if t > _maxshortdirslen: | |
294 | break |
|
294 | break | |
295 | sdirs.append(d) |
|
295 | sdirs.append(d) | |
296 | sdirslen = t |
|
296 | sdirslen = t | |
297 | dirs = b'/'.join(sdirs) |
|
297 | dirs = b'/'.join(sdirs) | |
298 | if len(dirs) > 0: |
|
298 | if len(dirs) > 0: | |
299 | dirs += b'/' |
|
299 | dirs += b'/' | |
300 | res = b'dh/' + dirs + digest + ext |
|
300 | res = b'dh/' + dirs + digest + ext | |
301 | spaceleft = _maxstorepathlen - len(res) |
|
301 | spaceleft = _maxstorepathlen - len(res) | |
302 | if spaceleft > 0: |
|
302 | if spaceleft > 0: | |
303 | filler = basename[:spaceleft] |
|
303 | filler = basename[:spaceleft] | |
304 | res = b'dh/' + dirs + filler + digest + ext |
|
304 | res = b'dh/' + dirs + filler + digest + ext | |
305 | return res |
|
305 | return res | |
306 |
|
306 | |||
307 |
|
307 | |||
308 | def _hybridencode(path, dotencode): |
|
308 | def _hybridencode(path, dotencode): | |
309 | """encodes path with a length limit |
|
309 | """encodes path with a length limit | |
310 |
|
310 | |||
311 | Encodes all paths that begin with 'data/', according to the following. |
|
311 | Encodes all paths that begin with 'data/', according to the following. | |
312 |
|
312 | |||
313 | Default encoding (reversible): |
|
313 | Default encoding (reversible): | |
314 |
|
314 | |||
315 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal |
|
315 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal | |
316 | characters are encoded as '~xx', where xx is the two digit hex code |
|
316 | characters are encoded as '~xx', where xx is the two digit hex code | |
317 | of the character (see encodefilename). |
|
317 | of the character (see encodefilename). | |
318 | Relevant path components consisting of Windows reserved filenames are |
|
318 | Relevant path components consisting of Windows reserved filenames are | |
319 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). |
|
319 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). | |
320 |
|
320 | |||
321 | Hashed encoding (not reversible): |
|
321 | Hashed encoding (not reversible): | |
322 |
|
322 | |||
323 | If the default-encoded path is longer than _maxstorepathlen, a |
|
323 | If the default-encoded path is longer than _maxstorepathlen, a | |
324 | non-reversible hybrid hashing of the path is done instead. |
|
324 | non-reversible hybrid hashing of the path is done instead. | |
325 | This encoding uses up to _dirprefixlen characters of all directory |
|
325 | This encoding uses up to _dirprefixlen characters of all directory | |
326 | levels of the lowerencoded path, but not more levels than can fit into |
|
326 | levels of the lowerencoded path, but not more levels than can fit into | |
327 | _maxshortdirslen. |
|
327 | _maxshortdirslen. | |
328 | Then follows the filler followed by the sha digest of the full path. |
|
328 | Then follows the filler followed by the sha digest of the full path. | |
329 | The filler is the beginning of the basename of the lowerencoded path |
|
329 | The filler is the beginning of the basename of the lowerencoded path | |
330 | (the basename is everything after the last path separator). The filler |
|
330 | (the basename is everything after the last path separator). The filler | |
331 | is as long as possible, filling in characters from the basename until |
|
331 | is as long as possible, filling in characters from the basename until | |
332 | the encoded path has _maxstorepathlen characters (or all chars of the |
|
332 | the encoded path has _maxstorepathlen characters (or all chars of the | |
333 | basename have been taken). |
|
333 | basename have been taken). | |
334 | The extension (e.g. '.i' or '.d') is preserved. |
|
334 | The extension (e.g. '.i' or '.d') is preserved. | |
335 |
|
335 | |||
336 | The string 'data/' at the beginning is replaced with 'dh/', if the hashed |
|
336 | The string 'data/' at the beginning is replaced with 'dh/', if the hashed | |
337 | encoding was used. |
|
337 | encoding was used. | |
338 | """ |
|
338 | """ | |
339 | path = encodedir(path) |
|
339 | path = encodedir(path) | |
340 | ef = _encodefname(path).split(b'/') |
|
340 | ef = _encodefname(path).split(b'/') | |
341 | res = b'/'.join(_auxencode(ef, dotencode)) |
|
341 | res = b'/'.join(_auxencode(ef, dotencode)) | |
342 | if len(res) > _maxstorepathlen: |
|
342 | if len(res) > _maxstorepathlen: | |
343 | res = _hashencode(path, dotencode) |
|
343 | res = _hashencode(path, dotencode) | |
344 | return res |
|
344 | return res | |
345 |
|
345 | |||
346 |
|
346 | |||
347 | def _pathencode(path): |
|
347 | def _pathencode(path): | |
348 | de = encodedir(path) |
|
348 | de = encodedir(path) | |
349 | if len(path) > _maxstorepathlen: |
|
349 | if len(path) > _maxstorepathlen: | |
350 | return _hashencode(de, True) |
|
350 | return _hashencode(de, True) | |
351 | ef = _encodefname(de).split(b'/') |
|
351 | ef = _encodefname(de).split(b'/') | |
352 | res = b'/'.join(_auxencode(ef, True)) |
|
352 | res = b'/'.join(_auxencode(ef, True)) | |
353 | if len(res) > _maxstorepathlen: |
|
353 | if len(res) > _maxstorepathlen: | |
354 | return _hashencode(de, True) |
|
354 | return _hashencode(de, True) | |
355 | return res |
|
355 | return res | |
356 |
|
356 | |||
357 |
|
357 | |||
358 | _pathencode = getattr(parsers, 'pathencode', _pathencode) |
|
358 | _pathencode = getattr(parsers, 'pathencode', _pathencode) | |
359 |
|
359 | |||
360 |
|
360 | |||
361 | def _plainhybridencode(f): |
|
361 | def _plainhybridencode(f): | |
362 | return _hybridencode(f, False) |
|
362 | return _hybridencode(f, False) | |
363 |
|
363 | |||
364 |
|
364 | |||
365 | def _calcmode(vfs): |
|
365 | def _calcmode(vfs): | |
366 | try: |
|
366 | try: | |
367 | # files in .hg/ will be created using this mode |
|
367 | # files in .hg/ will be created using this mode | |
368 | mode = vfs.stat().st_mode |
|
368 | mode = vfs.stat().st_mode | |
369 | # avoid some useless chmods |
|
369 | # avoid some useless chmods | |
370 | if (0o777 & ~util.umask) == (0o777 & mode): |
|
370 | if (0o777 & ~util.umask) == (0o777 & mode): | |
371 | mode = None |
|
371 | mode = None | |
372 | except OSError: |
|
372 | except OSError: | |
373 | mode = None |
|
373 | mode = None | |
374 | return mode |
|
374 | return mode | |
375 |
|
375 | |||
376 |
|
376 | |||
377 | _data = [ |
|
377 | _data = [ | |
378 | b'bookmarks', |
|
378 | b'bookmarks', | |
379 | b'narrowspec', |
|
379 | b'narrowspec', | |
380 | b'data', |
|
380 | b'data', | |
381 | b'meta', |
|
381 | b'meta', | |
382 | b'00manifest.d', |
|
382 | b'00manifest.d', | |
383 | b'00manifest.i', |
|
383 | b'00manifest.i', | |
384 | b'00changelog.d', |
|
384 | b'00changelog.d', | |
385 | b'00changelog.i', |
|
385 | b'00changelog.i', | |
386 | b'phaseroots', |
|
386 | b'phaseroots', | |
387 | b'obsstore', |
|
387 | b'obsstore', | |
388 | b'requires', |
|
388 | b'requires', | |
389 | ] |
|
389 | ] | |
390 |
|
390 | |||
391 | REVLOG_FILES_MAIN_EXT = (b'.i', b'i.tmpcensored') |
|
391 | REVLOG_FILES_MAIN_EXT = (b'.i', b'i.tmpcensored') | |
392 | REVLOG_FILES_OTHER_EXT = ( |
|
392 | REVLOG_FILES_OTHER_EXT = ( | |
393 | b'.idx', |
|
393 | b'.idx', | |
394 | b'.d', |
|
394 | b'.d', | |
395 | b'.dat', |
|
395 | b'.dat', | |
396 | b'.n', |
|
396 | b'.n', | |
397 | b'.nd', |
|
397 | b'.nd', | |
398 | b'.sda', |
|
398 | b'.sda', | |
399 | b'd.tmpcensored', |
|
399 | b'd.tmpcensored', | |
400 | ) |
|
400 | ) | |
401 | # files that are "volatile" and might change between listing and streaming |
|
401 | # files that are "volatile" and might change between listing and streaming | |
402 | # |
|
402 | # | |
403 | # note: the ".nd" file are nodemap data and won't "change" but they might be |
|
403 | # note: the ".nd" file are nodemap data and won't "change" but they might be | |
404 | # deleted. |
|
404 | # deleted. | |
405 | REVLOG_FILES_VOLATILE_EXT = (b'.n', b'.nd') |
|
405 | REVLOG_FILES_VOLATILE_EXT = (b'.n', b'.nd') | |
406 |
|
406 | |||
407 | # some exception to the above matching |
|
407 | # some exception to the above matching | |
408 | # |
|
408 | # | |
409 | # XXX This is currently not in use because of issue6542 |
|
409 | # XXX This is currently not in use because of issue6542 | |
410 | EXCLUDED = re.compile(b'.*undo\.[^/]+\.(nd?|i)$') |
|
410 | EXCLUDED = re.compile(b'.*undo\.[^/]+\.(nd?|i)$') | |
411 |
|
411 | |||
412 |
|
412 | |||
413 | def is_revlog(f, kind, st): |
|
413 | def is_revlog(f, kind, st): | |
414 | if kind != stat.S_IFREG: |
|
414 | if kind != stat.S_IFREG: | |
415 | return None |
|
415 | return None | |
416 | return revlog_type(f) |
|
416 | return revlog_type(f) | |
417 |
|
417 | |||
418 |
|
418 | |||
419 | def revlog_type(f): |
|
419 | def revlog_type(f): | |
420 | # XXX we need to filter `undo.` created by the transaction here, however |
|
420 | # XXX we need to filter `undo.` created by the transaction here, however | |
421 | # being naive about it also filter revlog for `undo.*` files, leading to |
|
421 | # being naive about it also filter revlog for `undo.*` files, leading to | |
422 | # issue6542. So we no longer use EXCLUDED. |
|
422 | # issue6542. So we no longer use EXCLUDED. | |
423 | if f.endswith(REVLOG_FILES_MAIN_EXT): |
|
423 | if f.endswith(REVLOG_FILES_MAIN_EXT): | |
424 | return FILEFLAGS_REVLOG_MAIN |
|
424 | return FILEFLAGS_REVLOG_MAIN | |
425 | elif f.endswith(REVLOG_FILES_OTHER_EXT): |
|
425 | elif f.endswith(REVLOG_FILES_OTHER_EXT): | |
426 | t = FILETYPE_FILELOG_OTHER |
|
426 | t = FILETYPE_FILELOG_OTHER | |
427 | if f.endswith(REVLOG_FILES_VOLATILE_EXT): |
|
427 | if f.endswith(REVLOG_FILES_VOLATILE_EXT): | |
428 | t |= FILEFLAGS_VOLATILE |
|
428 | t |= FILEFLAGS_VOLATILE | |
429 | return t |
|
429 | return t | |
430 | return None |
|
430 | return None | |
431 |
|
431 | |||
432 |
|
432 | |||
433 | # the file is part of changelog data |
|
433 | # the file is part of changelog data | |
434 | FILEFLAGS_CHANGELOG = 1 << 13 |
|
434 | FILEFLAGS_CHANGELOG = 1 << 13 | |
435 | # the file is part of manifest data |
|
435 | # the file is part of manifest data | |
436 | FILEFLAGS_MANIFESTLOG = 1 << 12 |
|
436 | FILEFLAGS_MANIFESTLOG = 1 << 12 | |
437 | # the file is part of filelog data |
|
437 | # the file is part of filelog data | |
438 | FILEFLAGS_FILELOG = 1 << 11 |
|
438 | FILEFLAGS_FILELOG = 1 << 11 | |
439 | # file that are not directly part of a revlog |
|
439 | # file that are not directly part of a revlog | |
440 | FILEFLAGS_OTHER = 1 << 10 |
|
440 | FILEFLAGS_OTHER = 1 << 10 | |
441 |
|
441 | |||
442 | # the main entry point for a revlog |
|
442 | # the main entry point for a revlog | |
443 | FILEFLAGS_REVLOG_MAIN = 1 << 1 |
|
443 | FILEFLAGS_REVLOG_MAIN = 1 << 1 | |
444 | # a secondary file for a revlog |
|
444 | # a secondary file for a revlog | |
445 | FILEFLAGS_REVLOG_OTHER = 1 << 0 |
|
445 | FILEFLAGS_REVLOG_OTHER = 1 << 0 | |
446 |
|
446 | |||
447 | # files that are "volatile" and might change between listing and streaming |
|
447 | # files that are "volatile" and might change between listing and streaming | |
448 | FILEFLAGS_VOLATILE = 1 << 20 |
|
448 | FILEFLAGS_VOLATILE = 1 << 20 | |
449 |
|
449 | |||
450 | FILETYPE_CHANGELOG_MAIN = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_MAIN |
|
450 | FILETYPE_CHANGELOG_MAIN = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_MAIN | |
451 | FILETYPE_CHANGELOG_OTHER = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_OTHER |
|
451 | FILETYPE_CHANGELOG_OTHER = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_OTHER | |
452 | FILETYPE_MANIFESTLOG_MAIN = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_MAIN |
|
452 | FILETYPE_MANIFESTLOG_MAIN = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_MAIN | |
453 | FILETYPE_MANIFESTLOG_OTHER = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_OTHER |
|
453 | FILETYPE_MANIFESTLOG_OTHER = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_OTHER | |
454 | FILETYPE_FILELOG_MAIN = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_MAIN |
|
454 | FILETYPE_FILELOG_MAIN = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_MAIN | |
455 | FILETYPE_FILELOG_OTHER = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_OTHER |
|
455 | FILETYPE_FILELOG_OTHER = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_OTHER | |
456 | FILETYPE_OTHER = FILEFLAGS_OTHER |
|
456 | FILETYPE_OTHER = FILEFLAGS_OTHER | |
457 |
|
457 | |||
458 |
|
458 | |||
459 | class basicstore(object): |
|
459 | class basicstore(object): | |
460 | '''base class for local repository stores''' |
|
460 | '''base class for local repository stores''' | |
461 |
|
461 | |||
462 | def __init__(self, path, vfstype): |
|
462 | def __init__(self, path, vfstype): | |
463 | vfs = vfstype(path) |
|
463 | vfs = vfstype(path) | |
464 | self.path = vfs.base |
|
464 | self.path = vfs.base | |
465 | self.createmode = _calcmode(vfs) |
|
465 | self.createmode = _calcmode(vfs) | |
466 | vfs.createmode = self.createmode |
|
466 | vfs.createmode = self.createmode | |
467 | self.rawvfs = vfs |
|
467 | self.rawvfs = vfs | |
468 | self.vfs = vfsmod.filtervfs(vfs, encodedir) |
|
468 | self.vfs = vfsmod.filtervfs(vfs, encodedir) | |
469 | self.opener = self.vfs |
|
469 | self.opener = self.vfs | |
470 |
|
470 | |||
471 | def join(self, f): |
|
471 | def join(self, f): | |
472 | return self.path + b'/' + encodedir(f) |
|
472 | return self.path + b'/' + encodedir(f) | |
473 |
|
473 | |||
474 | def _walk(self, relpath, recurse): |
|
474 | def _walk(self, relpath, recurse): | |
475 |
'''yields ( |
|
475 | '''yields (revlog_type, unencoded, size)''' | |
476 | path = self.path |
|
476 | path = self.path | |
477 | if relpath: |
|
477 | if relpath: | |
478 | path += b'/' + relpath |
|
478 | path += b'/' + relpath | |
479 | striplen = len(self.path) + 1 |
|
479 | striplen = len(self.path) + 1 | |
480 | l = [] |
|
480 | l = [] | |
481 | if self.rawvfs.isdir(path): |
|
481 | if self.rawvfs.isdir(path): | |
482 | visit = [path] |
|
482 | visit = [path] | |
483 | readdir = self.rawvfs.readdir |
|
483 | readdir = self.rawvfs.readdir | |
484 | while visit: |
|
484 | while visit: | |
485 | p = visit.pop() |
|
485 | p = visit.pop() | |
486 | for f, kind, st in readdir(p, stat=True): |
|
486 | for f, kind, st in readdir(p, stat=True): | |
487 | fp = p + b'/' + f |
|
487 | fp = p + b'/' + f | |
488 | rl_type = is_revlog(f, kind, st) |
|
488 | rl_type = is_revlog(f, kind, st) | |
489 | if rl_type is not None: |
|
489 | if rl_type is not None: | |
490 | n = util.pconvert(fp[striplen:]) |
|
490 | n = util.pconvert(fp[striplen:]) | |
491 |
l.append((rl_type, decodedir(n), |
|
491 | l.append((rl_type, decodedir(n), st.st_size)) | |
492 | elif kind == stat.S_IFDIR and recurse: |
|
492 | elif kind == stat.S_IFDIR and recurse: | |
493 | visit.append(fp) |
|
493 | visit.append(fp) | |
494 | l.sort() |
|
494 | l.sort() | |
495 | return l |
|
495 | return l | |
496 |
|
496 | |||
497 | def changelog(self, trypending, concurrencychecker=None): |
|
497 | def changelog(self, trypending, concurrencychecker=None): | |
498 | return changelog.changelog( |
|
498 | return changelog.changelog( | |
499 | self.vfs, |
|
499 | self.vfs, | |
500 | trypending=trypending, |
|
500 | trypending=trypending, | |
501 | concurrencychecker=concurrencychecker, |
|
501 | concurrencychecker=concurrencychecker, | |
502 | ) |
|
502 | ) | |
503 |
|
503 | |||
504 | def manifestlog(self, repo, storenarrowmatch): |
|
504 | def manifestlog(self, repo, storenarrowmatch): | |
505 | rootstore = manifest.manifestrevlog(repo.nodeconstants, self.vfs) |
|
505 | rootstore = manifest.manifestrevlog(repo.nodeconstants, self.vfs) | |
506 | return manifest.manifestlog(self.vfs, repo, rootstore, storenarrowmatch) |
|
506 | return manifest.manifestlog(self.vfs, repo, rootstore, storenarrowmatch) | |
507 |
|
507 | |||
508 | def datafiles(self, matcher=None): |
|
508 | def datafiles(self, matcher=None, undecodable=None): | |
|
509 | """Like walk, but excluding the changelog and root manifest. | |||
|
510 | ||||
|
511 | When [undecodable] is None, revlogs names that can't be | |||
|
512 | decoded cause an exception. When it is provided, it should | |||
|
513 | be a list and the filenames that can't be decoded are added | |||
|
514 | to it instead. This is very rarely needed.""" | |||
509 | files = self._walk(b'data', True) + self._walk(b'meta', True) |
|
515 | files = self._walk(b'data', True) + self._walk(b'meta', True) | |
510 |
for (t, u, |
|
516 | for (t, u, s) in files: | |
511 |
yield (FILEFLAGS_FILELOG | t, u, |
|
517 | yield (FILEFLAGS_FILELOG | t, u, s) | |
512 |
|
518 | |||
513 | def topfiles(self): |
|
519 | def topfiles(self): | |
514 | # yield manifest before changelog |
|
520 | # yield manifest before changelog | |
515 | files = reversed(self._walk(b'', False)) |
|
521 | files = reversed(self._walk(b'', False)) | |
516 |
for (t, u, |
|
522 | for (t, u, s) in files: | |
517 | if u.startswith(b'00changelog'): |
|
523 | if u.startswith(b'00changelog'): | |
518 |
yield (FILEFLAGS_CHANGELOG | t, u, |
|
524 | yield (FILEFLAGS_CHANGELOG | t, u, s) | |
519 | elif u.startswith(b'00manifest'): |
|
525 | elif u.startswith(b'00manifest'): | |
520 |
yield (FILEFLAGS_MANIFESTLOG | t, u, |
|
526 | yield (FILEFLAGS_MANIFESTLOG | t, u, s) | |
521 | else: |
|
527 | else: | |
522 |
yield (FILETYPE_OTHER | t, u, |
|
528 | yield (FILETYPE_OTHER | t, u, s) | |
523 |
|
529 | |||
524 | def walk(self, matcher=None): |
|
530 | def walk(self, matcher=None): | |
525 | """return file related to data storage (ie: revlogs) |
|
531 | """return file related to data storage (ie: revlogs) | |
526 |
|
532 | |||
527 |
yields (file_type, unencoded, |
|
533 | yields (file_type, unencoded, size) | |
528 |
|
534 | |||
529 | if a matcher is passed, storage files of only those tracked paths |
|
535 | if a matcher is passed, storage files of only those tracked paths | |
530 | are passed with matches the matcher |
|
536 | are passed with matches the matcher | |
531 | """ |
|
537 | """ | |
532 | # yield data files first |
|
538 | # yield data files first | |
533 | for x in self.datafiles(matcher): |
|
539 | for x in self.datafiles(matcher): | |
534 | yield x |
|
540 | yield x | |
535 | for x in self.topfiles(): |
|
541 | for x in self.topfiles(): | |
536 | yield x |
|
542 | yield x | |
537 |
|
543 | |||
538 | def copylist(self): |
|
544 | def copylist(self): | |
539 | return _data |
|
545 | return _data | |
540 |
|
546 | |||
541 | def write(self, tr): |
|
547 | def write(self, tr): | |
542 | pass |
|
548 | pass | |
543 |
|
549 | |||
544 | def invalidatecaches(self): |
|
550 | def invalidatecaches(self): | |
545 | pass |
|
551 | pass | |
546 |
|
552 | |||
547 | def markremoved(self, fn): |
|
553 | def markremoved(self, fn): | |
548 | pass |
|
554 | pass | |
549 |
|
555 | |||
550 | def __contains__(self, path): |
|
556 | def __contains__(self, path): | |
551 | '''Checks if the store contains path''' |
|
557 | '''Checks if the store contains path''' | |
552 | path = b"/".join((b"data", path)) |
|
558 | path = b"/".join((b"data", path)) | |
553 | # file? |
|
559 | # file? | |
554 | if self.vfs.exists(path + b".i"): |
|
560 | if self.vfs.exists(path + b".i"): | |
555 | return True |
|
561 | return True | |
556 | # dir? |
|
562 | # dir? | |
557 | if not path.endswith(b"/"): |
|
563 | if not path.endswith(b"/"): | |
558 | path = path + b"/" |
|
564 | path = path + b"/" | |
559 | return self.vfs.exists(path) |
|
565 | return self.vfs.exists(path) | |
560 |
|
566 | |||
561 |
|
567 | |||
562 | class encodedstore(basicstore): |
|
568 | class encodedstore(basicstore): | |
563 | def __init__(self, path, vfstype): |
|
569 | def __init__(self, path, vfstype): | |
564 | vfs = vfstype(path + b'/store') |
|
570 | vfs = vfstype(path + b'/store') | |
565 | self.path = vfs.base |
|
571 | self.path = vfs.base | |
566 | self.createmode = _calcmode(vfs) |
|
572 | self.createmode = _calcmode(vfs) | |
567 | vfs.createmode = self.createmode |
|
573 | vfs.createmode = self.createmode | |
568 | self.rawvfs = vfs |
|
574 | self.rawvfs = vfs | |
569 | self.vfs = vfsmod.filtervfs(vfs, encodefilename) |
|
575 | self.vfs = vfsmod.filtervfs(vfs, encodefilename) | |
570 | self.opener = self.vfs |
|
576 | self.opener = self.vfs | |
571 |
|
577 | |||
572 | # note: topfiles would also need a decode phase. It is just that in |
|
578 | # note: topfiles would also need a decode phase. It is just that in | |
573 | # practice we do not have any file outside of `data/` that needs encoding. |
|
579 | # practice we do not have any file outside of `data/` that needs encoding. | |
574 | # However that might change so we should probably add a test and encoding |
|
580 | # However that might change so we should probably add a test and encoding | |
575 | # decoding for it too. see issue6548 |
|
581 | # decoding for it too. see issue6548 | |
576 |
|
582 | |||
577 | def datafiles(self, matcher=None): |
|
583 | def datafiles(self, matcher=None, undecodable=None): | |
578 |
for t, |
|
584 | for t, f1, size in super(encodedstore, self).datafiles(): | |
579 | try: |
|
585 | try: | |
580 |
|
|
586 | f2 = decodefilename(f1) | |
581 | except KeyError: |
|
587 | except KeyError: | |
582 |
|
|
588 | if undecodable is None: | |
583 | if a is not None and not _matchtrackedpath(a, matcher): |
|
589 | msg = _(b'undecodable revlog name %s') % f1 | |
|
590 | raise error.StorageError(msg) | |||
|
591 | else: | |||
|
592 | undecodable.append(f1) | |||
584 | continue |
|
593 | continue | |
585 | yield t, a, b, size |
|
594 | if not _matchtrackedpath(f2, matcher): | |
|
595 | continue | |||
|
596 | yield t, f2, size | |||
586 |
|
597 | |||
587 | def join(self, f): |
|
598 | def join(self, f): | |
588 | return self.path + b'/' + encodefilename(f) |
|
599 | return self.path + b'/' + encodefilename(f) | |
589 |
|
600 | |||
590 | def copylist(self): |
|
601 | def copylist(self): | |
591 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in _data] |
|
602 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in _data] | |
592 |
|
603 | |||
593 |
|
604 | |||
594 | class fncache(object): |
|
605 | class fncache(object): | |
595 | # the filename used to be partially encoded |
|
606 | # the filename used to be partially encoded | |
596 | # hence the encodedir/decodedir dance |
|
607 | # hence the encodedir/decodedir dance | |
597 | def __init__(self, vfs): |
|
608 | def __init__(self, vfs): | |
598 | self.vfs = vfs |
|
609 | self.vfs = vfs | |
599 | self.entries = None |
|
610 | self.entries = None | |
600 | self._dirty = False |
|
611 | self._dirty = False | |
601 | # set of new additions to fncache |
|
612 | # set of new additions to fncache | |
602 | self.addls = set() |
|
613 | self.addls = set() | |
603 |
|
614 | |||
604 | def ensureloaded(self, warn=None): |
|
615 | def ensureloaded(self, warn=None): | |
605 | """read the fncache file if not already read. |
|
616 | """read the fncache file if not already read. | |
606 |
|
617 | |||
607 | If the file on disk is corrupted, raise. If warn is provided, |
|
618 | If the file on disk is corrupted, raise. If warn is provided, | |
608 | warn and keep going instead.""" |
|
619 | warn and keep going instead.""" | |
609 | if self.entries is None: |
|
620 | if self.entries is None: | |
610 | self._load(warn) |
|
621 | self._load(warn) | |
611 |
|
622 | |||
612 | def _load(self, warn=None): |
|
623 | def _load(self, warn=None): | |
613 | '''fill the entries from the fncache file''' |
|
624 | '''fill the entries from the fncache file''' | |
614 | self._dirty = False |
|
625 | self._dirty = False | |
615 | try: |
|
626 | try: | |
616 | fp = self.vfs(b'fncache', mode=b'rb') |
|
627 | fp = self.vfs(b'fncache', mode=b'rb') | |
617 | except IOError: |
|
628 | except IOError: | |
618 | # skip nonexistent file |
|
629 | # skip nonexistent file | |
619 | self.entries = set() |
|
630 | self.entries = set() | |
620 | return |
|
631 | return | |
621 |
|
632 | |||
622 | self.entries = set() |
|
633 | self.entries = set() | |
623 | chunk = b'' |
|
634 | chunk = b'' | |
624 | for c in iter(functools.partial(fp.read, fncache_chunksize), b''): |
|
635 | for c in iter(functools.partial(fp.read, fncache_chunksize), b''): | |
625 | chunk += c |
|
636 | chunk += c | |
626 | try: |
|
637 | try: | |
627 | p = chunk.rindex(b'\n') |
|
638 | p = chunk.rindex(b'\n') | |
628 | self.entries.update(decodedir(chunk[: p + 1]).splitlines()) |
|
639 | self.entries.update(decodedir(chunk[: p + 1]).splitlines()) | |
629 | chunk = chunk[p + 1 :] |
|
640 | chunk = chunk[p + 1 :] | |
630 | except ValueError: |
|
641 | except ValueError: | |
631 | # substring '\n' not found, maybe the entry is bigger than the |
|
642 | # substring '\n' not found, maybe the entry is bigger than the | |
632 | # chunksize, so let's keep iterating |
|
643 | # chunksize, so let's keep iterating | |
633 | pass |
|
644 | pass | |
634 |
|
645 | |||
635 | if chunk: |
|
646 | if chunk: | |
636 | msg = _(b"fncache does not ends with a newline") |
|
647 | msg = _(b"fncache does not ends with a newline") | |
637 | if warn: |
|
648 | if warn: | |
638 | warn(msg + b'\n') |
|
649 | warn(msg + b'\n') | |
639 | else: |
|
650 | else: | |
640 | raise error.Abort( |
|
651 | raise error.Abort( | |
641 | msg, |
|
652 | msg, | |
642 | hint=_( |
|
653 | hint=_( | |
643 | b"use 'hg debugrebuildfncache' to " |
|
654 | b"use 'hg debugrebuildfncache' to " | |
644 | b"rebuild the fncache" |
|
655 | b"rebuild the fncache" | |
645 | ), |
|
656 | ), | |
646 | ) |
|
657 | ) | |
647 | self._checkentries(fp, warn) |
|
658 | self._checkentries(fp, warn) | |
648 | fp.close() |
|
659 | fp.close() | |
649 |
|
660 | |||
650 | def _checkentries(self, fp, warn): |
|
661 | def _checkentries(self, fp, warn): | |
651 | """make sure there is no empty string in entries""" |
|
662 | """make sure there is no empty string in entries""" | |
652 | if b'' in self.entries: |
|
663 | if b'' in self.entries: | |
653 | fp.seek(0) |
|
664 | fp.seek(0) | |
654 | for n, line in enumerate(util.iterfile(fp)): |
|
665 | for n, line in enumerate(util.iterfile(fp)): | |
655 | if not line.rstrip(b'\n'): |
|
666 | if not line.rstrip(b'\n'): | |
656 | t = _(b'invalid entry in fncache, line %d') % (n + 1) |
|
667 | t = _(b'invalid entry in fncache, line %d') % (n + 1) | |
657 | if warn: |
|
668 | if warn: | |
658 | warn(t + b'\n') |
|
669 | warn(t + b'\n') | |
659 | else: |
|
670 | else: | |
660 | raise error.Abort(t) |
|
671 | raise error.Abort(t) | |
661 |
|
672 | |||
662 | def write(self, tr): |
|
673 | def write(self, tr): | |
663 | if self._dirty: |
|
674 | if self._dirty: | |
664 | assert self.entries is not None |
|
675 | assert self.entries is not None | |
665 | self.entries = self.entries | self.addls |
|
676 | self.entries = self.entries | self.addls | |
666 | self.addls = set() |
|
677 | self.addls = set() | |
667 | tr.addbackup(b'fncache') |
|
678 | tr.addbackup(b'fncache') | |
668 | fp = self.vfs(b'fncache', mode=b'wb', atomictemp=True) |
|
679 | fp = self.vfs(b'fncache', mode=b'wb', atomictemp=True) | |
669 | if self.entries: |
|
680 | if self.entries: | |
670 | fp.write(encodedir(b'\n'.join(self.entries) + b'\n')) |
|
681 | fp.write(encodedir(b'\n'.join(self.entries) + b'\n')) | |
671 | fp.close() |
|
682 | fp.close() | |
672 | self._dirty = False |
|
683 | self._dirty = False | |
673 | if self.addls: |
|
684 | if self.addls: | |
674 | # if we have just new entries, let's append them to the fncache |
|
685 | # if we have just new entries, let's append them to the fncache | |
675 | tr.addbackup(b'fncache') |
|
686 | tr.addbackup(b'fncache') | |
676 | fp = self.vfs(b'fncache', mode=b'ab', atomictemp=True) |
|
687 | fp = self.vfs(b'fncache', mode=b'ab', atomictemp=True) | |
677 | if self.addls: |
|
688 | if self.addls: | |
678 | fp.write(encodedir(b'\n'.join(self.addls) + b'\n')) |
|
689 | fp.write(encodedir(b'\n'.join(self.addls) + b'\n')) | |
679 | fp.close() |
|
690 | fp.close() | |
680 | self.entries = None |
|
691 | self.entries = None | |
681 | self.addls = set() |
|
692 | self.addls = set() | |
682 |
|
693 | |||
683 | def add(self, fn): |
|
694 | def add(self, fn): | |
684 | if self.entries is None: |
|
695 | if self.entries is None: | |
685 | self._load() |
|
696 | self._load() | |
686 | if fn not in self.entries: |
|
697 | if fn not in self.entries: | |
687 | self.addls.add(fn) |
|
698 | self.addls.add(fn) | |
688 |
|
699 | |||
689 | def remove(self, fn): |
|
700 | def remove(self, fn): | |
690 | if self.entries is None: |
|
701 | if self.entries is None: | |
691 | self._load() |
|
702 | self._load() | |
692 | if fn in self.addls: |
|
703 | if fn in self.addls: | |
693 | self.addls.remove(fn) |
|
704 | self.addls.remove(fn) | |
694 | return |
|
705 | return | |
695 | try: |
|
706 | try: | |
696 | self.entries.remove(fn) |
|
707 | self.entries.remove(fn) | |
697 | self._dirty = True |
|
708 | self._dirty = True | |
698 | except KeyError: |
|
709 | except KeyError: | |
699 | pass |
|
710 | pass | |
700 |
|
711 | |||
701 | def __contains__(self, fn): |
|
712 | def __contains__(self, fn): | |
702 | if fn in self.addls: |
|
713 | if fn in self.addls: | |
703 | return True |
|
714 | return True | |
704 | if self.entries is None: |
|
715 | if self.entries is None: | |
705 | self._load() |
|
716 | self._load() | |
706 | return fn in self.entries |
|
717 | return fn in self.entries | |
707 |
|
718 | |||
708 | def __iter__(self): |
|
719 | def __iter__(self): | |
709 | if self.entries is None: |
|
720 | if self.entries is None: | |
710 | self._load() |
|
721 | self._load() | |
711 | return iter(self.entries | self.addls) |
|
722 | return iter(self.entries | self.addls) | |
712 |
|
723 | |||
713 |
|
724 | |||
714 | class _fncachevfs(vfsmod.proxyvfs): |
|
725 | class _fncachevfs(vfsmod.proxyvfs): | |
715 | def __init__(self, vfs, fnc, encode): |
|
726 | def __init__(self, vfs, fnc, encode): | |
716 | vfsmod.proxyvfs.__init__(self, vfs) |
|
727 | vfsmod.proxyvfs.__init__(self, vfs) | |
717 | self.fncache = fnc |
|
728 | self.fncache = fnc | |
718 | self.encode = encode |
|
729 | self.encode = encode | |
719 |
|
730 | |||
720 | def __call__(self, path, mode=b'r', *args, **kw): |
|
731 | def __call__(self, path, mode=b'r', *args, **kw): | |
721 | encoded = self.encode(path) |
|
732 | encoded = self.encode(path) | |
722 | if mode not in (b'r', b'rb') and ( |
|
733 | if mode not in (b'r', b'rb') and ( | |
723 | path.startswith(b'data/') or path.startswith(b'meta/') |
|
734 | path.startswith(b'data/') or path.startswith(b'meta/') | |
724 | ): |
|
735 | ): | |
725 | # do not trigger a fncache load when adding a file that already is |
|
736 | # do not trigger a fncache load when adding a file that already is | |
726 | # known to exist. |
|
737 | # known to exist. | |
727 | notload = self.fncache.entries is None and self.vfs.exists(encoded) |
|
738 | notload = self.fncache.entries is None and self.vfs.exists(encoded) | |
728 | if notload and b'r+' in mode and not self.vfs.stat(encoded).st_size: |
|
739 | if notload and b'r+' in mode and not self.vfs.stat(encoded).st_size: | |
729 | # when appending to an existing file, if the file has size zero, |
|
740 | # when appending to an existing file, if the file has size zero, | |
730 | # it should be considered as missing. Such zero-size files are |
|
741 | # it should be considered as missing. Such zero-size files are | |
731 | # the result of truncation when a transaction is aborted. |
|
742 | # the result of truncation when a transaction is aborted. | |
732 | notload = False |
|
743 | notload = False | |
733 | if not notload: |
|
744 | if not notload: | |
734 | self.fncache.add(path) |
|
745 | self.fncache.add(path) | |
735 | return self.vfs(encoded, mode, *args, **kw) |
|
746 | return self.vfs(encoded, mode, *args, **kw) | |
736 |
|
747 | |||
737 | def join(self, path): |
|
748 | def join(self, path): | |
738 | if path: |
|
749 | if path: | |
739 | return self.vfs.join(self.encode(path)) |
|
750 | return self.vfs.join(self.encode(path)) | |
740 | else: |
|
751 | else: | |
741 | return self.vfs.join(path) |
|
752 | return self.vfs.join(path) | |
742 |
|
753 | |||
743 | def register_file(self, path): |
|
754 | def register_file(self, path): | |
744 | """generic hook point to lets fncache steer its stew""" |
|
755 | """generic hook point to lets fncache steer its stew""" | |
745 | if path.startswith(b'data/') or path.startswith(b'meta/'): |
|
756 | if path.startswith(b'data/') or path.startswith(b'meta/'): | |
746 | self.fncache.add(path) |
|
757 | self.fncache.add(path) | |
747 |
|
758 | |||
748 |
|
759 | |||
749 | class fncachestore(basicstore): |
|
760 | class fncachestore(basicstore): | |
750 | def __init__(self, path, vfstype, dotencode): |
|
761 | def __init__(self, path, vfstype, dotencode): | |
751 | if dotencode: |
|
762 | if dotencode: | |
752 | encode = _pathencode |
|
763 | encode = _pathencode | |
753 | else: |
|
764 | else: | |
754 | encode = _plainhybridencode |
|
765 | encode = _plainhybridencode | |
755 | self.encode = encode |
|
766 | self.encode = encode | |
756 | vfs = vfstype(path + b'/store') |
|
767 | vfs = vfstype(path + b'/store') | |
757 | self.path = vfs.base |
|
768 | self.path = vfs.base | |
758 | self.pathsep = self.path + b'/' |
|
769 | self.pathsep = self.path + b'/' | |
759 | self.createmode = _calcmode(vfs) |
|
770 | self.createmode = _calcmode(vfs) | |
760 | vfs.createmode = self.createmode |
|
771 | vfs.createmode = self.createmode | |
761 | self.rawvfs = vfs |
|
772 | self.rawvfs = vfs | |
762 | fnc = fncache(vfs) |
|
773 | fnc = fncache(vfs) | |
763 | self.fncache = fnc |
|
774 | self.fncache = fnc | |
764 | self.vfs = _fncachevfs(vfs, fnc, encode) |
|
775 | self.vfs = _fncachevfs(vfs, fnc, encode) | |
765 | self.opener = self.vfs |
|
776 | self.opener = self.vfs | |
766 |
|
777 | |||
767 | def join(self, f): |
|
778 | def join(self, f): | |
768 | return self.pathsep + self.encode(f) |
|
779 | return self.pathsep + self.encode(f) | |
769 |
|
780 | |||
770 | def getsize(self, path): |
|
781 | def getsize(self, path): | |
771 | return self.rawvfs.stat(path).st_size |
|
782 | return self.rawvfs.stat(path).st_size | |
772 |
|
783 | |||
773 | def datafiles(self, matcher=None): |
|
784 | def datafiles(self, matcher=None, undecodable=None): | |
774 | for f in sorted(self.fncache): |
|
785 | for f in sorted(self.fncache): | |
775 | if not _matchtrackedpath(f, matcher): |
|
786 | if not _matchtrackedpath(f, matcher): | |
776 | continue |
|
787 | continue | |
777 | ef = self.encode(f) |
|
788 | ef = self.encode(f) | |
778 | try: |
|
789 | try: | |
779 | t = revlog_type(f) |
|
790 | t = revlog_type(f) | |
780 | assert t is not None, f |
|
791 | assert t is not None, f | |
781 | t |= FILEFLAGS_FILELOG |
|
792 | t |= FILEFLAGS_FILELOG | |
782 |
yield t, f, |
|
793 | yield t, f, self.getsize(ef) | |
783 | except OSError as err: |
|
794 | except OSError as err: | |
784 | if err.errno != errno.ENOENT: |
|
795 | if err.errno != errno.ENOENT: | |
785 | raise |
|
796 | raise | |
786 |
|
797 | |||
787 | def copylist(self): |
|
798 | def copylist(self): | |
788 | d = ( |
|
799 | d = ( | |
789 | b'bookmarks', |
|
800 | b'bookmarks', | |
790 | b'narrowspec', |
|
801 | b'narrowspec', | |
791 | b'data', |
|
802 | b'data', | |
792 | b'meta', |
|
803 | b'meta', | |
793 | b'dh', |
|
804 | b'dh', | |
794 | b'fncache', |
|
805 | b'fncache', | |
795 | b'phaseroots', |
|
806 | b'phaseroots', | |
796 | b'obsstore', |
|
807 | b'obsstore', | |
797 | b'00manifest.d', |
|
808 | b'00manifest.d', | |
798 | b'00manifest.i', |
|
809 | b'00manifest.i', | |
799 | b'00changelog.d', |
|
810 | b'00changelog.d', | |
800 | b'00changelog.i', |
|
811 | b'00changelog.i', | |
801 | b'requires', |
|
812 | b'requires', | |
802 | ) |
|
813 | ) | |
803 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in d] |
|
814 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in d] | |
804 |
|
815 | |||
805 | def write(self, tr): |
|
816 | def write(self, tr): | |
806 | self.fncache.write(tr) |
|
817 | self.fncache.write(tr) | |
807 |
|
818 | |||
808 | def invalidatecaches(self): |
|
819 | def invalidatecaches(self): | |
809 | self.fncache.entries = None |
|
820 | self.fncache.entries = None | |
810 | self.fncache.addls = set() |
|
821 | self.fncache.addls = set() | |
811 |
|
822 | |||
812 | def markremoved(self, fn): |
|
823 | def markremoved(self, fn): | |
813 | self.fncache.remove(fn) |
|
824 | self.fncache.remove(fn) | |
814 |
|
825 | |||
815 | def _exists(self, f): |
|
826 | def _exists(self, f): | |
816 | ef = self.encode(f) |
|
827 | ef = self.encode(f) | |
817 | try: |
|
828 | try: | |
818 | self.getsize(ef) |
|
829 | self.getsize(ef) | |
819 | return True |
|
830 | return True | |
820 | except OSError as err: |
|
831 | except OSError as err: | |
821 | if err.errno != errno.ENOENT: |
|
832 | if err.errno != errno.ENOENT: | |
822 | raise |
|
833 | raise | |
823 | # nonexistent entry |
|
834 | # nonexistent entry | |
824 | return False |
|
835 | return False | |
825 |
|
836 | |||
826 | def __contains__(self, path): |
|
837 | def __contains__(self, path): | |
827 | '''Checks if the store contains path''' |
|
838 | '''Checks if the store contains path''' | |
828 | path = b"/".join((b"data", path)) |
|
839 | path = b"/".join((b"data", path)) | |
829 | # check for files (exact match) |
|
840 | # check for files (exact match) | |
830 | e = path + b'.i' |
|
841 | e = path + b'.i' | |
831 | if e in self.fncache and self._exists(e): |
|
842 | if e in self.fncache and self._exists(e): | |
832 | return True |
|
843 | return True | |
833 | # now check for directories (prefix match) |
|
844 | # now check for directories (prefix match) | |
834 | if not path.endswith(b'/'): |
|
845 | if not path.endswith(b'/'): | |
835 | path += b'/' |
|
846 | path += b'/' | |
836 | for e in self.fncache: |
|
847 | for e in self.fncache: | |
837 | if e.startswith(path) and self._exists(e): |
|
848 | if e.startswith(path) and self._exists(e): | |
838 | return True |
|
849 | return True | |
839 | return False |
|
850 | return False |
@@ -1,918 +1,918 b'' | |||||
1 | # streamclone.py - producing and consuming streaming repository data |
|
1 | # streamclone.py - producing and consuming streaming repository data | |
2 | # |
|
2 | # | |
3 | # Copyright 2015 Gregory Szorc <gregory.szorc@gmail.com> |
|
3 | # Copyright 2015 Gregory Szorc <gregory.szorc@gmail.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 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import contextlib |
|
10 | import contextlib | |
11 | import errno |
|
11 | import errno | |
12 | import os |
|
12 | import os | |
13 | import struct |
|
13 | import struct | |
14 |
|
14 | |||
15 | from .i18n import _ |
|
15 | from .i18n import _ | |
16 | from .pycompat import open |
|
16 | from .pycompat import open | |
17 | from .interfaces import repository |
|
17 | from .interfaces import repository | |
18 | from . import ( |
|
18 | from . import ( | |
19 | bookmarks, |
|
19 | bookmarks, | |
20 | cacheutil, |
|
20 | cacheutil, | |
21 | error, |
|
21 | error, | |
22 | narrowspec, |
|
22 | narrowspec, | |
23 | phases, |
|
23 | phases, | |
24 | pycompat, |
|
24 | pycompat, | |
25 | requirements as requirementsmod, |
|
25 | requirements as requirementsmod, | |
26 | scmutil, |
|
26 | scmutil, | |
27 | store, |
|
27 | store, | |
28 | util, |
|
28 | util, | |
29 | ) |
|
29 | ) | |
30 | from .utils import ( |
|
30 | from .utils import ( | |
31 | stringutil, |
|
31 | stringutil, | |
32 | ) |
|
32 | ) | |
33 |
|
33 | |||
34 |
|
34 | |||
35 | def canperformstreamclone(pullop, bundle2=False): |
|
35 | def canperformstreamclone(pullop, bundle2=False): | |
36 | """Whether it is possible to perform a streaming clone as part of pull. |
|
36 | """Whether it is possible to perform a streaming clone as part of pull. | |
37 |
|
37 | |||
38 | ``bundle2`` will cause the function to consider stream clone through |
|
38 | ``bundle2`` will cause the function to consider stream clone through | |
39 | bundle2 and only through bundle2. |
|
39 | bundle2 and only through bundle2. | |
40 |
|
40 | |||
41 | Returns a tuple of (supported, requirements). ``supported`` is True if |
|
41 | Returns a tuple of (supported, requirements). ``supported`` is True if | |
42 | streaming clone is supported and False otherwise. ``requirements`` is |
|
42 | streaming clone is supported and False otherwise. ``requirements`` is | |
43 | a set of repo requirements from the remote, or ``None`` if stream clone |
|
43 | a set of repo requirements from the remote, or ``None`` if stream clone | |
44 | isn't supported. |
|
44 | isn't supported. | |
45 | """ |
|
45 | """ | |
46 | repo = pullop.repo |
|
46 | repo = pullop.repo | |
47 | remote = pullop.remote |
|
47 | remote = pullop.remote | |
48 |
|
48 | |||
49 | bundle2supported = False |
|
49 | bundle2supported = False | |
50 | if pullop.canusebundle2: |
|
50 | if pullop.canusebundle2: | |
51 | if b'v2' in pullop.remotebundle2caps.get(b'stream', []): |
|
51 | if b'v2' in pullop.remotebundle2caps.get(b'stream', []): | |
52 | bundle2supported = True |
|
52 | bundle2supported = True | |
53 | # else |
|
53 | # else | |
54 | # Server doesn't support bundle2 stream clone or doesn't support |
|
54 | # Server doesn't support bundle2 stream clone or doesn't support | |
55 | # the versions we support. Fall back and possibly allow legacy. |
|
55 | # the versions we support. Fall back and possibly allow legacy. | |
56 |
|
56 | |||
57 | # Ensures legacy code path uses available bundle2. |
|
57 | # Ensures legacy code path uses available bundle2. | |
58 | if bundle2supported and not bundle2: |
|
58 | if bundle2supported and not bundle2: | |
59 | return False, None |
|
59 | return False, None | |
60 | # Ensures bundle2 doesn't try to do a stream clone if it isn't supported. |
|
60 | # Ensures bundle2 doesn't try to do a stream clone if it isn't supported. | |
61 | elif bundle2 and not bundle2supported: |
|
61 | elif bundle2 and not bundle2supported: | |
62 | return False, None |
|
62 | return False, None | |
63 |
|
63 | |||
64 | # Streaming clone only works on empty repositories. |
|
64 | # Streaming clone only works on empty repositories. | |
65 | if len(repo): |
|
65 | if len(repo): | |
66 | return False, None |
|
66 | return False, None | |
67 |
|
67 | |||
68 | # Streaming clone only works if all data is being requested. |
|
68 | # Streaming clone only works if all data is being requested. | |
69 | if pullop.heads: |
|
69 | if pullop.heads: | |
70 | return False, None |
|
70 | return False, None | |
71 |
|
71 | |||
72 | streamrequested = pullop.streamclonerequested |
|
72 | streamrequested = pullop.streamclonerequested | |
73 |
|
73 | |||
74 | # If we don't have a preference, let the server decide for us. This |
|
74 | # If we don't have a preference, let the server decide for us. This | |
75 | # likely only comes into play in LANs. |
|
75 | # likely only comes into play in LANs. | |
76 | if streamrequested is None: |
|
76 | if streamrequested is None: | |
77 | # The server can advertise whether to prefer streaming clone. |
|
77 | # The server can advertise whether to prefer streaming clone. | |
78 | streamrequested = remote.capable(b'stream-preferred') |
|
78 | streamrequested = remote.capable(b'stream-preferred') | |
79 |
|
79 | |||
80 | if not streamrequested: |
|
80 | if not streamrequested: | |
81 | return False, None |
|
81 | return False, None | |
82 |
|
82 | |||
83 | # In order for stream clone to work, the client has to support all the |
|
83 | # In order for stream clone to work, the client has to support all the | |
84 | # requirements advertised by the server. |
|
84 | # requirements advertised by the server. | |
85 | # |
|
85 | # | |
86 | # The server advertises its requirements via the "stream" and "streamreqs" |
|
86 | # The server advertises its requirements via the "stream" and "streamreqs" | |
87 | # capability. "stream" (a value-less capability) is advertised if and only |
|
87 | # capability. "stream" (a value-less capability) is advertised if and only | |
88 | # if the only requirement is "revlogv1." Else, the "streamreqs" capability |
|
88 | # if the only requirement is "revlogv1." Else, the "streamreqs" capability | |
89 | # is advertised and contains a comma-delimited list of requirements. |
|
89 | # is advertised and contains a comma-delimited list of requirements. | |
90 | requirements = set() |
|
90 | requirements = set() | |
91 | if remote.capable(b'stream'): |
|
91 | if remote.capable(b'stream'): | |
92 | requirements.add(requirementsmod.REVLOGV1_REQUIREMENT) |
|
92 | requirements.add(requirementsmod.REVLOGV1_REQUIREMENT) | |
93 | else: |
|
93 | else: | |
94 | streamreqs = remote.capable(b'streamreqs') |
|
94 | streamreqs = remote.capable(b'streamreqs') | |
95 | # This is weird and shouldn't happen with modern servers. |
|
95 | # This is weird and shouldn't happen with modern servers. | |
96 | if not streamreqs: |
|
96 | if not streamreqs: | |
97 | pullop.repo.ui.warn( |
|
97 | pullop.repo.ui.warn( | |
98 | _( |
|
98 | _( | |
99 | b'warning: stream clone requested but server has them ' |
|
99 | b'warning: stream clone requested but server has them ' | |
100 | b'disabled\n' |
|
100 | b'disabled\n' | |
101 | ) |
|
101 | ) | |
102 | ) |
|
102 | ) | |
103 | return False, None |
|
103 | return False, None | |
104 |
|
104 | |||
105 | streamreqs = set(streamreqs.split(b',')) |
|
105 | streamreqs = set(streamreqs.split(b',')) | |
106 | # Server requires something we don't support. Bail. |
|
106 | # Server requires something we don't support. Bail. | |
107 | missingreqs = streamreqs - repo.supportedformats |
|
107 | missingreqs = streamreqs - repo.supportedformats | |
108 | if missingreqs: |
|
108 | if missingreqs: | |
109 | pullop.repo.ui.warn( |
|
109 | pullop.repo.ui.warn( | |
110 | _( |
|
110 | _( | |
111 | b'warning: stream clone requested but client is missing ' |
|
111 | b'warning: stream clone requested but client is missing ' | |
112 | b'requirements: %s\n' |
|
112 | b'requirements: %s\n' | |
113 | ) |
|
113 | ) | |
114 | % b', '.join(sorted(missingreqs)) |
|
114 | % b', '.join(sorted(missingreqs)) | |
115 | ) |
|
115 | ) | |
116 | pullop.repo.ui.warn( |
|
116 | pullop.repo.ui.warn( | |
117 | _( |
|
117 | _( | |
118 | b'(see https://www.mercurial-scm.org/wiki/MissingRequirement ' |
|
118 | b'(see https://www.mercurial-scm.org/wiki/MissingRequirement ' | |
119 | b'for more information)\n' |
|
119 | b'for more information)\n' | |
120 | ) |
|
120 | ) | |
121 | ) |
|
121 | ) | |
122 | return False, None |
|
122 | return False, None | |
123 | requirements = streamreqs |
|
123 | requirements = streamreqs | |
124 |
|
124 | |||
125 | return True, requirements |
|
125 | return True, requirements | |
126 |
|
126 | |||
127 |
|
127 | |||
128 | def maybeperformlegacystreamclone(pullop): |
|
128 | def maybeperformlegacystreamclone(pullop): | |
129 | """Possibly perform a legacy stream clone operation. |
|
129 | """Possibly perform a legacy stream clone operation. | |
130 |
|
130 | |||
131 | Legacy stream clones are performed as part of pull but before all other |
|
131 | Legacy stream clones are performed as part of pull but before all other | |
132 | operations. |
|
132 | operations. | |
133 |
|
133 | |||
134 | A legacy stream clone will not be performed if a bundle2 stream clone is |
|
134 | A legacy stream clone will not be performed if a bundle2 stream clone is | |
135 | supported. |
|
135 | supported. | |
136 | """ |
|
136 | """ | |
137 | from . import localrepo |
|
137 | from . import localrepo | |
138 |
|
138 | |||
139 | supported, requirements = canperformstreamclone(pullop) |
|
139 | supported, requirements = canperformstreamclone(pullop) | |
140 |
|
140 | |||
141 | if not supported: |
|
141 | if not supported: | |
142 | return |
|
142 | return | |
143 |
|
143 | |||
144 | repo = pullop.repo |
|
144 | repo = pullop.repo | |
145 | remote = pullop.remote |
|
145 | remote = pullop.remote | |
146 |
|
146 | |||
147 | # Save remote branchmap. We will use it later to speed up branchcache |
|
147 | # Save remote branchmap. We will use it later to speed up branchcache | |
148 | # creation. |
|
148 | # creation. | |
149 | rbranchmap = None |
|
149 | rbranchmap = None | |
150 | if remote.capable(b'branchmap'): |
|
150 | if remote.capable(b'branchmap'): | |
151 | with remote.commandexecutor() as e: |
|
151 | with remote.commandexecutor() as e: | |
152 | rbranchmap = e.callcommand(b'branchmap', {}).result() |
|
152 | rbranchmap = e.callcommand(b'branchmap', {}).result() | |
153 |
|
153 | |||
154 | repo.ui.status(_(b'streaming all changes\n')) |
|
154 | repo.ui.status(_(b'streaming all changes\n')) | |
155 |
|
155 | |||
156 | with remote.commandexecutor() as e: |
|
156 | with remote.commandexecutor() as e: | |
157 | fp = e.callcommand(b'stream_out', {}).result() |
|
157 | fp = e.callcommand(b'stream_out', {}).result() | |
158 |
|
158 | |||
159 | # TODO strictly speaking, this code should all be inside the context |
|
159 | # TODO strictly speaking, this code should all be inside the context | |
160 | # manager because the context manager is supposed to ensure all wire state |
|
160 | # manager because the context manager is supposed to ensure all wire state | |
161 | # is flushed when exiting. But the legacy peers don't do this, so it |
|
161 | # is flushed when exiting. But the legacy peers don't do this, so it | |
162 | # doesn't matter. |
|
162 | # doesn't matter. | |
163 | l = fp.readline() |
|
163 | l = fp.readline() | |
164 | try: |
|
164 | try: | |
165 | resp = int(l) |
|
165 | resp = int(l) | |
166 | except ValueError: |
|
166 | except ValueError: | |
167 | raise error.ResponseError( |
|
167 | raise error.ResponseError( | |
168 | _(b'unexpected response from remote server:'), l |
|
168 | _(b'unexpected response from remote server:'), l | |
169 | ) |
|
169 | ) | |
170 | if resp == 1: |
|
170 | if resp == 1: | |
171 | raise error.Abort(_(b'operation forbidden by server')) |
|
171 | raise error.Abort(_(b'operation forbidden by server')) | |
172 | elif resp == 2: |
|
172 | elif resp == 2: | |
173 | raise error.Abort(_(b'locking the remote repository failed')) |
|
173 | raise error.Abort(_(b'locking the remote repository failed')) | |
174 | elif resp != 0: |
|
174 | elif resp != 0: | |
175 | raise error.Abort(_(b'the server sent an unknown error code')) |
|
175 | raise error.Abort(_(b'the server sent an unknown error code')) | |
176 |
|
176 | |||
177 | l = fp.readline() |
|
177 | l = fp.readline() | |
178 | try: |
|
178 | try: | |
179 | filecount, bytecount = map(int, l.split(b' ', 1)) |
|
179 | filecount, bytecount = map(int, l.split(b' ', 1)) | |
180 | except (ValueError, TypeError): |
|
180 | except (ValueError, TypeError): | |
181 | raise error.ResponseError( |
|
181 | raise error.ResponseError( | |
182 | _(b'unexpected response from remote server:'), l |
|
182 | _(b'unexpected response from remote server:'), l | |
183 | ) |
|
183 | ) | |
184 |
|
184 | |||
185 | with repo.lock(): |
|
185 | with repo.lock(): | |
186 | consumev1(repo, fp, filecount, bytecount) |
|
186 | consumev1(repo, fp, filecount, bytecount) | |
187 |
|
187 | |||
188 | # new requirements = old non-format requirements + |
|
188 | # new requirements = old non-format requirements + | |
189 | # new format-related remote requirements |
|
189 | # new format-related remote requirements | |
190 | # requirements from the streamed-in repository |
|
190 | # requirements from the streamed-in repository | |
191 | repo.requirements = requirements | ( |
|
191 | repo.requirements = requirements | ( | |
192 | repo.requirements - repo.supportedformats |
|
192 | repo.requirements - repo.supportedformats | |
193 | ) |
|
193 | ) | |
194 | repo.svfs.options = localrepo.resolvestorevfsoptions( |
|
194 | repo.svfs.options = localrepo.resolvestorevfsoptions( | |
195 | repo.ui, repo.requirements, repo.features |
|
195 | repo.ui, repo.requirements, repo.features | |
196 | ) |
|
196 | ) | |
197 | scmutil.writereporequirements(repo) |
|
197 | scmutil.writereporequirements(repo) | |
198 |
|
198 | |||
199 | if rbranchmap: |
|
199 | if rbranchmap: | |
200 | repo._branchcaches.replace(repo, rbranchmap) |
|
200 | repo._branchcaches.replace(repo, rbranchmap) | |
201 |
|
201 | |||
202 | repo.invalidate() |
|
202 | repo.invalidate() | |
203 |
|
203 | |||
204 |
|
204 | |||
205 | def allowservergeneration(repo): |
|
205 | def allowservergeneration(repo): | |
206 | """Whether streaming clones are allowed from the server.""" |
|
206 | """Whether streaming clones are allowed from the server.""" | |
207 | if repository.REPO_FEATURE_STREAM_CLONE not in repo.features: |
|
207 | if repository.REPO_FEATURE_STREAM_CLONE not in repo.features: | |
208 | return False |
|
208 | return False | |
209 |
|
209 | |||
210 | if not repo.ui.configbool(b'server', b'uncompressed', untrusted=True): |
|
210 | if not repo.ui.configbool(b'server', b'uncompressed', untrusted=True): | |
211 | return False |
|
211 | return False | |
212 |
|
212 | |||
213 | # The way stream clone works makes it impossible to hide secret changesets. |
|
213 | # The way stream clone works makes it impossible to hide secret changesets. | |
214 | # So don't allow this by default. |
|
214 | # So don't allow this by default. | |
215 | secret = phases.hassecret(repo) |
|
215 | secret = phases.hassecret(repo) | |
216 | if secret: |
|
216 | if secret: | |
217 | return repo.ui.configbool(b'server', b'uncompressedallowsecret') |
|
217 | return repo.ui.configbool(b'server', b'uncompressedallowsecret') | |
218 |
|
218 | |||
219 | return True |
|
219 | return True | |
220 |
|
220 | |||
221 |
|
221 | |||
222 | # This is it's own function so extensions can override it. |
|
222 | # This is it's own function so extensions can override it. | |
223 | def _walkstreamfiles(repo, matcher=None): |
|
223 | def _walkstreamfiles(repo, matcher=None): | |
224 | return repo.store.walk(matcher) |
|
224 | return repo.store.walk(matcher) | |
225 |
|
225 | |||
226 |
|
226 | |||
227 | def generatev1(repo): |
|
227 | def generatev1(repo): | |
228 | """Emit content for version 1 of a streaming clone. |
|
228 | """Emit content for version 1 of a streaming clone. | |
229 |
|
229 | |||
230 | This returns a 3-tuple of (file count, byte size, data iterator). |
|
230 | This returns a 3-tuple of (file count, byte size, data iterator). | |
231 |
|
231 | |||
232 | The data iterator consists of N entries for each file being transferred. |
|
232 | The data iterator consists of N entries for each file being transferred. | |
233 | Each file entry starts as a line with the file name and integer size |
|
233 | Each file entry starts as a line with the file name and integer size | |
234 | delimited by a null byte. |
|
234 | delimited by a null byte. | |
235 |
|
235 | |||
236 | The raw file data follows. Following the raw file data is the next file |
|
236 | The raw file data follows. Following the raw file data is the next file | |
237 | entry, or EOF. |
|
237 | entry, or EOF. | |
238 |
|
238 | |||
239 | When used on the wire protocol, an additional line indicating protocol |
|
239 | When used on the wire protocol, an additional line indicating protocol | |
240 | success will be prepended to the stream. This function is not responsible |
|
240 | success will be prepended to the stream. This function is not responsible | |
241 | for adding it. |
|
241 | for adding it. | |
242 |
|
242 | |||
243 | This function will obtain a repository lock to ensure a consistent view of |
|
243 | This function will obtain a repository lock to ensure a consistent view of | |
244 | the store is captured. It therefore may raise LockError. |
|
244 | the store is captured. It therefore may raise LockError. | |
245 | """ |
|
245 | """ | |
246 | entries = [] |
|
246 | entries = [] | |
247 | total_bytes = 0 |
|
247 | total_bytes = 0 | |
248 | # Get consistent snapshot of repo, lock during scan. |
|
248 | # Get consistent snapshot of repo, lock during scan. | |
249 | with repo.lock(): |
|
249 | with repo.lock(): | |
250 | repo.ui.debug(b'scanning\n') |
|
250 | repo.ui.debug(b'scanning\n') | |
251 |
for file_type, name |
|
251 | for file_type, name, size in _walkstreamfiles(repo): | |
252 | if size: |
|
252 | if size: | |
253 | entries.append((name, size)) |
|
253 | entries.append((name, size)) | |
254 | total_bytes += size |
|
254 | total_bytes += size | |
255 | _test_sync_point_walk_1(repo) |
|
255 | _test_sync_point_walk_1(repo) | |
256 | _test_sync_point_walk_2(repo) |
|
256 | _test_sync_point_walk_2(repo) | |
257 |
|
257 | |||
258 | repo.ui.debug( |
|
258 | repo.ui.debug( | |
259 | b'%d files, %d bytes to transfer\n' % (len(entries), total_bytes) |
|
259 | b'%d files, %d bytes to transfer\n' % (len(entries), total_bytes) | |
260 | ) |
|
260 | ) | |
261 |
|
261 | |||
262 | svfs = repo.svfs |
|
262 | svfs = repo.svfs | |
263 | debugflag = repo.ui.debugflag |
|
263 | debugflag = repo.ui.debugflag | |
264 |
|
264 | |||
265 | def emitrevlogdata(): |
|
265 | def emitrevlogdata(): | |
266 | for name, size in entries: |
|
266 | for name, size in entries: | |
267 | if debugflag: |
|
267 | if debugflag: | |
268 | repo.ui.debug(b'sending %s (%d bytes)\n' % (name, size)) |
|
268 | repo.ui.debug(b'sending %s (%d bytes)\n' % (name, size)) | |
269 | # partially encode name over the wire for backwards compat |
|
269 | # partially encode name over the wire for backwards compat | |
270 | yield b'%s\0%d\n' % (store.encodedir(name), size) |
|
270 | yield b'%s\0%d\n' % (store.encodedir(name), size) | |
271 | # auditing at this stage is both pointless (paths are already |
|
271 | # auditing at this stage is both pointless (paths are already | |
272 | # trusted by the local repo) and expensive |
|
272 | # trusted by the local repo) and expensive | |
273 | with svfs(name, b'rb', auditpath=False) as fp: |
|
273 | with svfs(name, b'rb', auditpath=False) as fp: | |
274 | if size <= 65536: |
|
274 | if size <= 65536: | |
275 | yield fp.read(size) |
|
275 | yield fp.read(size) | |
276 | else: |
|
276 | else: | |
277 | for chunk in util.filechunkiter(fp, limit=size): |
|
277 | for chunk in util.filechunkiter(fp, limit=size): | |
278 | yield chunk |
|
278 | yield chunk | |
279 |
|
279 | |||
280 | return len(entries), total_bytes, emitrevlogdata() |
|
280 | return len(entries), total_bytes, emitrevlogdata() | |
281 |
|
281 | |||
282 |
|
282 | |||
283 | def generatev1wireproto(repo): |
|
283 | def generatev1wireproto(repo): | |
284 | """Emit content for version 1 of streaming clone suitable for the wire. |
|
284 | """Emit content for version 1 of streaming clone suitable for the wire. | |
285 |
|
285 | |||
286 | This is the data output from ``generatev1()`` with 2 header lines. The |
|
286 | This is the data output from ``generatev1()`` with 2 header lines. The | |
287 | first line indicates overall success. The 2nd contains the file count and |
|
287 | first line indicates overall success. The 2nd contains the file count and | |
288 | byte size of payload. |
|
288 | byte size of payload. | |
289 |
|
289 | |||
290 | The success line contains "0" for success, "1" for stream generation not |
|
290 | The success line contains "0" for success, "1" for stream generation not | |
291 | allowed, and "2" for error locking the repository (possibly indicating |
|
291 | allowed, and "2" for error locking the repository (possibly indicating | |
292 | a permissions error for the server process). |
|
292 | a permissions error for the server process). | |
293 | """ |
|
293 | """ | |
294 | if not allowservergeneration(repo): |
|
294 | if not allowservergeneration(repo): | |
295 | yield b'1\n' |
|
295 | yield b'1\n' | |
296 | return |
|
296 | return | |
297 |
|
297 | |||
298 | try: |
|
298 | try: | |
299 | filecount, bytecount, it = generatev1(repo) |
|
299 | filecount, bytecount, it = generatev1(repo) | |
300 | except error.LockError: |
|
300 | except error.LockError: | |
301 | yield b'2\n' |
|
301 | yield b'2\n' | |
302 | return |
|
302 | return | |
303 |
|
303 | |||
304 | # Indicates successful response. |
|
304 | # Indicates successful response. | |
305 | yield b'0\n' |
|
305 | yield b'0\n' | |
306 | yield b'%d %d\n' % (filecount, bytecount) |
|
306 | yield b'%d %d\n' % (filecount, bytecount) | |
307 | for chunk in it: |
|
307 | for chunk in it: | |
308 | yield chunk |
|
308 | yield chunk | |
309 |
|
309 | |||
310 |
|
310 | |||
311 | def generatebundlev1(repo, compression=b'UN'): |
|
311 | def generatebundlev1(repo, compression=b'UN'): | |
312 | """Emit content for version 1 of a stream clone bundle. |
|
312 | """Emit content for version 1 of a stream clone bundle. | |
313 |
|
313 | |||
314 | The first 4 bytes of the output ("HGS1") denote this as stream clone |
|
314 | The first 4 bytes of the output ("HGS1") denote this as stream clone | |
315 | bundle version 1. |
|
315 | bundle version 1. | |
316 |
|
316 | |||
317 | The next 2 bytes indicate the compression type. Only "UN" is currently |
|
317 | The next 2 bytes indicate the compression type. Only "UN" is currently | |
318 | supported. |
|
318 | supported. | |
319 |
|
319 | |||
320 | The next 16 bytes are two 64-bit big endian unsigned integers indicating |
|
320 | The next 16 bytes are two 64-bit big endian unsigned integers indicating | |
321 | file count and byte count, respectively. |
|
321 | file count and byte count, respectively. | |
322 |
|
322 | |||
323 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length |
|
323 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length | |
324 | of the requirements string, including a trailing \0. The following N bytes |
|
324 | of the requirements string, including a trailing \0. The following N bytes | |
325 | are the requirements string, which is ASCII containing a comma-delimited |
|
325 | are the requirements string, which is ASCII containing a comma-delimited | |
326 | list of repo requirements that are needed to support the data. |
|
326 | list of repo requirements that are needed to support the data. | |
327 |
|
327 | |||
328 | The remaining content is the output of ``generatev1()`` (which may be |
|
328 | The remaining content is the output of ``generatev1()`` (which may be | |
329 | compressed in the future). |
|
329 | compressed in the future). | |
330 |
|
330 | |||
331 | Returns a tuple of (requirements, data generator). |
|
331 | Returns a tuple of (requirements, data generator). | |
332 | """ |
|
332 | """ | |
333 | if compression != b'UN': |
|
333 | if compression != b'UN': | |
334 | raise ValueError(b'we do not support the compression argument yet') |
|
334 | raise ValueError(b'we do not support the compression argument yet') | |
335 |
|
335 | |||
336 | requirements = repo.requirements & repo.supportedformats |
|
336 | requirements = repo.requirements & repo.supportedformats | |
337 | requires = b','.join(sorted(requirements)) |
|
337 | requires = b','.join(sorted(requirements)) | |
338 |
|
338 | |||
339 | def gen(): |
|
339 | def gen(): | |
340 | yield b'HGS1' |
|
340 | yield b'HGS1' | |
341 | yield compression |
|
341 | yield compression | |
342 |
|
342 | |||
343 | filecount, bytecount, it = generatev1(repo) |
|
343 | filecount, bytecount, it = generatev1(repo) | |
344 | repo.ui.status( |
|
344 | repo.ui.status( | |
345 | _(b'writing %d bytes for %d files\n') % (bytecount, filecount) |
|
345 | _(b'writing %d bytes for %d files\n') % (bytecount, filecount) | |
346 | ) |
|
346 | ) | |
347 |
|
347 | |||
348 | yield struct.pack(b'>QQ', filecount, bytecount) |
|
348 | yield struct.pack(b'>QQ', filecount, bytecount) | |
349 | yield struct.pack(b'>H', len(requires) + 1) |
|
349 | yield struct.pack(b'>H', len(requires) + 1) | |
350 | yield requires + b'\0' |
|
350 | yield requires + b'\0' | |
351 |
|
351 | |||
352 | # This is where we'll add compression in the future. |
|
352 | # This is where we'll add compression in the future. | |
353 | assert compression == b'UN' |
|
353 | assert compression == b'UN' | |
354 |
|
354 | |||
355 | progress = repo.ui.makeprogress( |
|
355 | progress = repo.ui.makeprogress( | |
356 | _(b'bundle'), total=bytecount, unit=_(b'bytes') |
|
356 | _(b'bundle'), total=bytecount, unit=_(b'bytes') | |
357 | ) |
|
357 | ) | |
358 | progress.update(0) |
|
358 | progress.update(0) | |
359 |
|
359 | |||
360 | for chunk in it: |
|
360 | for chunk in it: | |
361 | progress.increment(step=len(chunk)) |
|
361 | progress.increment(step=len(chunk)) | |
362 | yield chunk |
|
362 | yield chunk | |
363 |
|
363 | |||
364 | progress.complete() |
|
364 | progress.complete() | |
365 |
|
365 | |||
366 | return requirements, gen() |
|
366 | return requirements, gen() | |
367 |
|
367 | |||
368 |
|
368 | |||
369 | def consumev1(repo, fp, filecount, bytecount): |
|
369 | def consumev1(repo, fp, filecount, bytecount): | |
370 | """Apply the contents from version 1 of a streaming clone file handle. |
|
370 | """Apply the contents from version 1 of a streaming clone file handle. | |
371 |
|
371 | |||
372 | This takes the output from "stream_out" and applies it to the specified |
|
372 | This takes the output from "stream_out" and applies it to the specified | |
373 | repository. |
|
373 | repository. | |
374 |
|
374 | |||
375 | Like "stream_out," the status line added by the wire protocol is not |
|
375 | Like "stream_out," the status line added by the wire protocol is not | |
376 | handled by this function. |
|
376 | handled by this function. | |
377 | """ |
|
377 | """ | |
378 | with repo.lock(): |
|
378 | with repo.lock(): | |
379 | repo.ui.status( |
|
379 | repo.ui.status( | |
380 | _(b'%d files to transfer, %s of data\n') |
|
380 | _(b'%d files to transfer, %s of data\n') | |
381 | % (filecount, util.bytecount(bytecount)) |
|
381 | % (filecount, util.bytecount(bytecount)) | |
382 | ) |
|
382 | ) | |
383 | progress = repo.ui.makeprogress( |
|
383 | progress = repo.ui.makeprogress( | |
384 | _(b'clone'), total=bytecount, unit=_(b'bytes') |
|
384 | _(b'clone'), total=bytecount, unit=_(b'bytes') | |
385 | ) |
|
385 | ) | |
386 | progress.update(0) |
|
386 | progress.update(0) | |
387 | start = util.timer() |
|
387 | start = util.timer() | |
388 |
|
388 | |||
389 | # TODO: get rid of (potential) inconsistency |
|
389 | # TODO: get rid of (potential) inconsistency | |
390 | # |
|
390 | # | |
391 | # If transaction is started and any @filecache property is |
|
391 | # If transaction is started and any @filecache property is | |
392 | # changed at this point, it causes inconsistency between |
|
392 | # changed at this point, it causes inconsistency between | |
393 | # in-memory cached property and streamclone-ed file on the |
|
393 | # in-memory cached property and streamclone-ed file on the | |
394 | # disk. Nested transaction prevents transaction scope "clone" |
|
394 | # disk. Nested transaction prevents transaction scope "clone" | |
395 | # below from writing in-memory changes out at the end of it, |
|
395 | # below from writing in-memory changes out at the end of it, | |
396 | # even though in-memory changes are discarded at the end of it |
|
396 | # even though in-memory changes are discarded at the end of it | |
397 | # regardless of transaction nesting. |
|
397 | # regardless of transaction nesting. | |
398 | # |
|
398 | # | |
399 | # But transaction nesting can't be simply prohibited, because |
|
399 | # But transaction nesting can't be simply prohibited, because | |
400 | # nesting occurs also in ordinary case (e.g. enabling |
|
400 | # nesting occurs also in ordinary case (e.g. enabling | |
401 | # clonebundles). |
|
401 | # clonebundles). | |
402 |
|
402 | |||
403 | with repo.transaction(b'clone'): |
|
403 | with repo.transaction(b'clone'): | |
404 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): |
|
404 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): | |
405 | for i in pycompat.xrange(filecount): |
|
405 | for i in pycompat.xrange(filecount): | |
406 | # XXX doesn't support '\n' or '\r' in filenames |
|
406 | # XXX doesn't support '\n' or '\r' in filenames | |
407 | l = fp.readline() |
|
407 | l = fp.readline() | |
408 | try: |
|
408 | try: | |
409 | name, size = l.split(b'\0', 1) |
|
409 | name, size = l.split(b'\0', 1) | |
410 | size = int(size) |
|
410 | size = int(size) | |
411 | except (ValueError, TypeError): |
|
411 | except (ValueError, TypeError): | |
412 | raise error.ResponseError( |
|
412 | raise error.ResponseError( | |
413 | _(b'unexpected response from remote server:'), l |
|
413 | _(b'unexpected response from remote server:'), l | |
414 | ) |
|
414 | ) | |
415 | if repo.ui.debugflag: |
|
415 | if repo.ui.debugflag: | |
416 | repo.ui.debug( |
|
416 | repo.ui.debug( | |
417 | b'adding %s (%s)\n' % (name, util.bytecount(size)) |
|
417 | b'adding %s (%s)\n' % (name, util.bytecount(size)) | |
418 | ) |
|
418 | ) | |
419 | # for backwards compat, name was partially encoded |
|
419 | # for backwards compat, name was partially encoded | |
420 | path = store.decodedir(name) |
|
420 | path = store.decodedir(name) | |
421 | with repo.svfs(path, b'w', backgroundclose=True) as ofp: |
|
421 | with repo.svfs(path, b'w', backgroundclose=True) as ofp: | |
422 | for chunk in util.filechunkiter(fp, limit=size): |
|
422 | for chunk in util.filechunkiter(fp, limit=size): | |
423 | progress.increment(step=len(chunk)) |
|
423 | progress.increment(step=len(chunk)) | |
424 | ofp.write(chunk) |
|
424 | ofp.write(chunk) | |
425 |
|
425 | |||
426 | # force @filecache properties to be reloaded from |
|
426 | # force @filecache properties to be reloaded from | |
427 | # streamclone-ed file at next access |
|
427 | # streamclone-ed file at next access | |
428 | repo.invalidate(clearfilecache=True) |
|
428 | repo.invalidate(clearfilecache=True) | |
429 |
|
429 | |||
430 | elapsed = util.timer() - start |
|
430 | elapsed = util.timer() - start | |
431 | if elapsed <= 0: |
|
431 | if elapsed <= 0: | |
432 | elapsed = 0.001 |
|
432 | elapsed = 0.001 | |
433 | progress.complete() |
|
433 | progress.complete() | |
434 | repo.ui.status( |
|
434 | repo.ui.status( | |
435 | _(b'transferred %s in %.1f seconds (%s/sec)\n') |
|
435 | _(b'transferred %s in %.1f seconds (%s/sec)\n') | |
436 | % ( |
|
436 | % ( | |
437 | util.bytecount(bytecount), |
|
437 | util.bytecount(bytecount), | |
438 | elapsed, |
|
438 | elapsed, | |
439 | util.bytecount(bytecount / elapsed), |
|
439 | util.bytecount(bytecount / elapsed), | |
440 | ) |
|
440 | ) | |
441 | ) |
|
441 | ) | |
442 |
|
442 | |||
443 |
|
443 | |||
444 | def readbundle1header(fp): |
|
444 | def readbundle1header(fp): | |
445 | compression = fp.read(2) |
|
445 | compression = fp.read(2) | |
446 | if compression != b'UN': |
|
446 | if compression != b'UN': | |
447 | raise error.Abort( |
|
447 | raise error.Abort( | |
448 | _( |
|
448 | _( | |
449 | b'only uncompressed stream clone bundles are ' |
|
449 | b'only uncompressed stream clone bundles are ' | |
450 | b'supported; got %s' |
|
450 | b'supported; got %s' | |
451 | ) |
|
451 | ) | |
452 | % compression |
|
452 | % compression | |
453 | ) |
|
453 | ) | |
454 |
|
454 | |||
455 | filecount, bytecount = struct.unpack(b'>QQ', fp.read(16)) |
|
455 | filecount, bytecount = struct.unpack(b'>QQ', fp.read(16)) | |
456 | requireslen = struct.unpack(b'>H', fp.read(2))[0] |
|
456 | requireslen = struct.unpack(b'>H', fp.read(2))[0] | |
457 | requires = fp.read(requireslen) |
|
457 | requires = fp.read(requireslen) | |
458 |
|
458 | |||
459 | if not requires.endswith(b'\0'): |
|
459 | if not requires.endswith(b'\0'): | |
460 | raise error.Abort( |
|
460 | raise error.Abort( | |
461 | _( |
|
461 | _( | |
462 | b'malformed stream clone bundle: ' |
|
462 | b'malformed stream clone bundle: ' | |
463 | b'requirements not properly encoded' |
|
463 | b'requirements not properly encoded' | |
464 | ) |
|
464 | ) | |
465 | ) |
|
465 | ) | |
466 |
|
466 | |||
467 | requirements = set(requires.rstrip(b'\0').split(b',')) |
|
467 | requirements = set(requires.rstrip(b'\0').split(b',')) | |
468 |
|
468 | |||
469 | return filecount, bytecount, requirements |
|
469 | return filecount, bytecount, requirements | |
470 |
|
470 | |||
471 |
|
471 | |||
472 | def applybundlev1(repo, fp): |
|
472 | def applybundlev1(repo, fp): | |
473 | """Apply the content from a stream clone bundle version 1. |
|
473 | """Apply the content from a stream clone bundle version 1. | |
474 |
|
474 | |||
475 | We assume the 4 byte header has been read and validated and the file handle |
|
475 | We assume the 4 byte header has been read and validated and the file handle | |
476 | is at the 2 byte compression identifier. |
|
476 | is at the 2 byte compression identifier. | |
477 | """ |
|
477 | """ | |
478 | if len(repo): |
|
478 | if len(repo): | |
479 | raise error.Abort( |
|
479 | raise error.Abort( | |
480 | _(b'cannot apply stream clone bundle on non-empty repo') |
|
480 | _(b'cannot apply stream clone bundle on non-empty repo') | |
481 | ) |
|
481 | ) | |
482 |
|
482 | |||
483 | filecount, bytecount, requirements = readbundle1header(fp) |
|
483 | filecount, bytecount, requirements = readbundle1header(fp) | |
484 | missingreqs = requirements - repo.supportedformats |
|
484 | missingreqs = requirements - repo.supportedformats | |
485 | if missingreqs: |
|
485 | if missingreqs: | |
486 | raise error.Abort( |
|
486 | raise error.Abort( | |
487 | _(b'unable to apply stream clone: unsupported format: %s') |
|
487 | _(b'unable to apply stream clone: unsupported format: %s') | |
488 | % b', '.join(sorted(missingreqs)) |
|
488 | % b', '.join(sorted(missingreqs)) | |
489 | ) |
|
489 | ) | |
490 |
|
490 | |||
491 | consumev1(repo, fp, filecount, bytecount) |
|
491 | consumev1(repo, fp, filecount, bytecount) | |
492 |
|
492 | |||
493 |
|
493 | |||
494 | class streamcloneapplier(object): |
|
494 | class streamcloneapplier(object): | |
495 | """Class to manage applying streaming clone bundles. |
|
495 | """Class to manage applying streaming clone bundles. | |
496 |
|
496 | |||
497 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle |
|
497 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle | |
498 | readers to perform bundle type-specific functionality. |
|
498 | readers to perform bundle type-specific functionality. | |
499 | """ |
|
499 | """ | |
500 |
|
500 | |||
501 | def __init__(self, fh): |
|
501 | def __init__(self, fh): | |
502 | self._fh = fh |
|
502 | self._fh = fh | |
503 |
|
503 | |||
504 | def apply(self, repo): |
|
504 | def apply(self, repo): | |
505 | return applybundlev1(repo, self._fh) |
|
505 | return applybundlev1(repo, self._fh) | |
506 |
|
506 | |||
507 |
|
507 | |||
508 | # type of file to stream |
|
508 | # type of file to stream | |
509 | _fileappend = 0 # append only file |
|
509 | _fileappend = 0 # append only file | |
510 | _filefull = 1 # full snapshot file |
|
510 | _filefull = 1 # full snapshot file | |
511 |
|
511 | |||
512 | # Source of the file |
|
512 | # Source of the file | |
513 | _srcstore = b's' # store (svfs) |
|
513 | _srcstore = b's' # store (svfs) | |
514 | _srccache = b'c' # cache (cache) |
|
514 | _srccache = b'c' # cache (cache) | |
515 |
|
515 | |||
516 | # This is it's own function so extensions can override it. |
|
516 | # This is it's own function so extensions can override it. | |
517 | def _walkstreamfullstorefiles(repo): |
|
517 | def _walkstreamfullstorefiles(repo): | |
518 | """list snapshot file from the store""" |
|
518 | """list snapshot file from the store""" | |
519 | fnames = [] |
|
519 | fnames = [] | |
520 | if not repo.publishing(): |
|
520 | if not repo.publishing(): | |
521 | fnames.append(b'phaseroots') |
|
521 | fnames.append(b'phaseroots') | |
522 | return fnames |
|
522 | return fnames | |
523 |
|
523 | |||
524 |
|
524 | |||
525 | def _filterfull(entry, copy, vfsmap): |
|
525 | def _filterfull(entry, copy, vfsmap): | |
526 | """actually copy the snapshot files""" |
|
526 | """actually copy the snapshot files""" | |
527 | src, name, ftype, data = entry |
|
527 | src, name, ftype, data = entry | |
528 | if ftype != _filefull: |
|
528 | if ftype != _filefull: | |
529 | return entry |
|
529 | return entry | |
530 | return (src, name, ftype, copy(vfsmap[src].join(name))) |
|
530 | return (src, name, ftype, copy(vfsmap[src].join(name))) | |
531 |
|
531 | |||
532 |
|
532 | |||
533 | @contextlib.contextmanager |
|
533 | @contextlib.contextmanager | |
534 | def maketempcopies(): |
|
534 | def maketempcopies(): | |
535 | """return a function to temporary copy file""" |
|
535 | """return a function to temporary copy file""" | |
536 | files = [] |
|
536 | files = [] | |
537 | try: |
|
537 | try: | |
538 |
|
538 | |||
539 | def copy(src): |
|
539 | def copy(src): | |
540 | fd, dst = pycompat.mkstemp() |
|
540 | fd, dst = pycompat.mkstemp() | |
541 | os.close(fd) |
|
541 | os.close(fd) | |
542 | files.append(dst) |
|
542 | files.append(dst) | |
543 | util.copyfiles(src, dst, hardlink=True) |
|
543 | util.copyfiles(src, dst, hardlink=True) | |
544 | return dst |
|
544 | return dst | |
545 |
|
545 | |||
546 | yield copy |
|
546 | yield copy | |
547 | finally: |
|
547 | finally: | |
548 | for tmp in files: |
|
548 | for tmp in files: | |
549 | util.tryunlink(tmp) |
|
549 | util.tryunlink(tmp) | |
550 |
|
550 | |||
551 |
|
551 | |||
552 | def _makemap(repo): |
|
552 | def _makemap(repo): | |
553 | """make a (src -> vfs) map for the repo""" |
|
553 | """make a (src -> vfs) map for the repo""" | |
554 | vfsmap = { |
|
554 | vfsmap = { | |
555 | _srcstore: repo.svfs, |
|
555 | _srcstore: repo.svfs, | |
556 | _srccache: repo.cachevfs, |
|
556 | _srccache: repo.cachevfs, | |
557 | } |
|
557 | } | |
558 | # we keep repo.vfs out of the on purpose, ther are too many danger there |
|
558 | # we keep repo.vfs out of the on purpose, ther are too many danger there | |
559 | # (eg: .hg/hgrc) |
|
559 | # (eg: .hg/hgrc) | |
560 | assert repo.vfs not in vfsmap.values() |
|
560 | assert repo.vfs not in vfsmap.values() | |
561 |
|
561 | |||
562 | return vfsmap |
|
562 | return vfsmap | |
563 |
|
563 | |||
564 |
|
564 | |||
565 | def _emit2(repo, entries, totalfilesize): |
|
565 | def _emit2(repo, entries, totalfilesize): | |
566 | """actually emit the stream bundle""" |
|
566 | """actually emit the stream bundle""" | |
567 | vfsmap = _makemap(repo) |
|
567 | vfsmap = _makemap(repo) | |
568 | # we keep repo.vfs out of the on purpose, ther are too many danger there |
|
568 | # we keep repo.vfs out of the on purpose, ther are too many danger there | |
569 | # (eg: .hg/hgrc), |
|
569 | # (eg: .hg/hgrc), | |
570 | # |
|
570 | # | |
571 | # this assert is duplicated (from _makemap) as author might think this is |
|
571 | # this assert is duplicated (from _makemap) as author might think this is | |
572 | # fine, while this is really not fine. |
|
572 | # fine, while this is really not fine. | |
573 | if repo.vfs in vfsmap.values(): |
|
573 | if repo.vfs in vfsmap.values(): | |
574 | raise error.ProgrammingError( |
|
574 | raise error.ProgrammingError( | |
575 | b'repo.vfs must not be added to vfsmap for security reasons' |
|
575 | b'repo.vfs must not be added to vfsmap for security reasons' | |
576 | ) |
|
576 | ) | |
577 |
|
577 | |||
578 | progress = repo.ui.makeprogress( |
|
578 | progress = repo.ui.makeprogress( | |
579 | _(b'bundle'), total=totalfilesize, unit=_(b'bytes') |
|
579 | _(b'bundle'), total=totalfilesize, unit=_(b'bytes') | |
580 | ) |
|
580 | ) | |
581 | progress.update(0) |
|
581 | progress.update(0) | |
582 | with maketempcopies() as copy, progress: |
|
582 | with maketempcopies() as copy, progress: | |
583 | # copy is delayed until we are in the try |
|
583 | # copy is delayed until we are in the try | |
584 | entries = [_filterfull(e, copy, vfsmap) for e in entries] |
|
584 | entries = [_filterfull(e, copy, vfsmap) for e in entries] | |
585 | yield None # this release the lock on the repository |
|
585 | yield None # this release the lock on the repository | |
586 | totalbytecount = 0 |
|
586 | totalbytecount = 0 | |
587 |
|
587 | |||
588 | for src, name, ftype, data in entries: |
|
588 | for src, name, ftype, data in entries: | |
589 | vfs = vfsmap[src] |
|
589 | vfs = vfsmap[src] | |
590 | yield src |
|
590 | yield src | |
591 | yield util.uvarintencode(len(name)) |
|
591 | yield util.uvarintencode(len(name)) | |
592 | if ftype == _fileappend: |
|
592 | if ftype == _fileappend: | |
593 | fp = vfs(name) |
|
593 | fp = vfs(name) | |
594 | size = data |
|
594 | size = data | |
595 | elif ftype == _filefull: |
|
595 | elif ftype == _filefull: | |
596 | fp = open(data, b'rb') |
|
596 | fp = open(data, b'rb') | |
597 | size = util.fstat(fp).st_size |
|
597 | size = util.fstat(fp).st_size | |
598 | bytecount = 0 |
|
598 | bytecount = 0 | |
599 | try: |
|
599 | try: | |
600 | yield util.uvarintencode(size) |
|
600 | yield util.uvarintencode(size) | |
601 | yield name |
|
601 | yield name | |
602 | if size <= 65536: |
|
602 | if size <= 65536: | |
603 | chunks = (fp.read(size),) |
|
603 | chunks = (fp.read(size),) | |
604 | else: |
|
604 | else: | |
605 | chunks = util.filechunkiter(fp, limit=size) |
|
605 | chunks = util.filechunkiter(fp, limit=size) | |
606 | for chunk in chunks: |
|
606 | for chunk in chunks: | |
607 | bytecount += len(chunk) |
|
607 | bytecount += len(chunk) | |
608 | totalbytecount += len(chunk) |
|
608 | totalbytecount += len(chunk) | |
609 | progress.update(totalbytecount) |
|
609 | progress.update(totalbytecount) | |
610 | yield chunk |
|
610 | yield chunk | |
611 | if bytecount != size: |
|
611 | if bytecount != size: | |
612 | # Would most likely be caused by a race due to `hg strip` or |
|
612 | # Would most likely be caused by a race due to `hg strip` or | |
613 | # a revlog split |
|
613 | # a revlog split | |
614 | raise error.Abort( |
|
614 | raise error.Abort( | |
615 | _( |
|
615 | _( | |
616 | b'clone could only read %d bytes from %s, but ' |
|
616 | b'clone could only read %d bytes from %s, but ' | |
617 | b'expected %d bytes' |
|
617 | b'expected %d bytes' | |
618 | ) |
|
618 | ) | |
619 | % (bytecount, name, size) |
|
619 | % (bytecount, name, size) | |
620 | ) |
|
620 | ) | |
621 | finally: |
|
621 | finally: | |
622 | fp.close() |
|
622 | fp.close() | |
623 |
|
623 | |||
624 |
|
624 | |||
625 | def _test_sync_point_walk_1(repo): |
|
625 | def _test_sync_point_walk_1(repo): | |
626 | """a function for synchronisation during tests""" |
|
626 | """a function for synchronisation during tests""" | |
627 |
|
627 | |||
628 |
|
628 | |||
629 | def _test_sync_point_walk_2(repo): |
|
629 | def _test_sync_point_walk_2(repo): | |
630 | """a function for synchronisation during tests""" |
|
630 | """a function for synchronisation during tests""" | |
631 |
|
631 | |||
632 |
|
632 | |||
633 | def _v2_walk(repo, includes, excludes, includeobsmarkers): |
|
633 | def _v2_walk(repo, includes, excludes, includeobsmarkers): | |
634 | """emit a seris of files information useful to clone a repo |
|
634 | """emit a seris of files information useful to clone a repo | |
635 |
|
635 | |||
636 | return (entries, totalfilesize) |
|
636 | return (entries, totalfilesize) | |
637 |
|
637 | |||
638 | entries is a list of tuple (vfs-key, file-path, file-type, size) |
|
638 | entries is a list of tuple (vfs-key, file-path, file-type, size) | |
639 |
|
639 | |||
640 | - `vfs-key`: is a key to the right vfs to write the file (see _makemap) |
|
640 | - `vfs-key`: is a key to the right vfs to write the file (see _makemap) | |
641 | - `name`: file path of the file to copy (to be feed to the vfss) |
|
641 | - `name`: file path of the file to copy (to be feed to the vfss) | |
642 | - `file-type`: do this file need to be copied with the source lock ? |
|
642 | - `file-type`: do this file need to be copied with the source lock ? | |
643 | - `size`: the size of the file (or None) |
|
643 | - `size`: the size of the file (or None) | |
644 | """ |
|
644 | """ | |
645 | assert repo._currentlock(repo._lockref) is not None |
|
645 | assert repo._currentlock(repo._lockref) is not None | |
646 | entries = [] |
|
646 | entries = [] | |
647 | totalfilesize = 0 |
|
647 | totalfilesize = 0 | |
648 |
|
648 | |||
649 | matcher = None |
|
649 | matcher = None | |
650 | if includes or excludes: |
|
650 | if includes or excludes: | |
651 | matcher = narrowspec.match(repo.root, includes, excludes) |
|
651 | matcher = narrowspec.match(repo.root, includes, excludes) | |
652 |
|
652 | |||
653 |
for rl_type, name |
|
653 | for rl_type, name, size in _walkstreamfiles(repo, matcher): | |
654 | if size: |
|
654 | if size: | |
655 | ft = _fileappend |
|
655 | ft = _fileappend | |
656 | if rl_type & store.FILEFLAGS_VOLATILE: |
|
656 | if rl_type & store.FILEFLAGS_VOLATILE: | |
657 | ft = _filefull |
|
657 | ft = _filefull | |
658 | entries.append((_srcstore, name, ft, size)) |
|
658 | entries.append((_srcstore, name, ft, size)) | |
659 | totalfilesize += size |
|
659 | totalfilesize += size | |
660 | for name in _walkstreamfullstorefiles(repo): |
|
660 | for name in _walkstreamfullstorefiles(repo): | |
661 | if repo.svfs.exists(name): |
|
661 | if repo.svfs.exists(name): | |
662 | totalfilesize += repo.svfs.lstat(name).st_size |
|
662 | totalfilesize += repo.svfs.lstat(name).st_size | |
663 | entries.append((_srcstore, name, _filefull, None)) |
|
663 | entries.append((_srcstore, name, _filefull, None)) | |
664 | if includeobsmarkers and repo.svfs.exists(b'obsstore'): |
|
664 | if includeobsmarkers and repo.svfs.exists(b'obsstore'): | |
665 | totalfilesize += repo.svfs.lstat(b'obsstore').st_size |
|
665 | totalfilesize += repo.svfs.lstat(b'obsstore').st_size | |
666 | entries.append((_srcstore, b'obsstore', _filefull, None)) |
|
666 | entries.append((_srcstore, b'obsstore', _filefull, None)) | |
667 | for name in cacheutil.cachetocopy(repo): |
|
667 | for name in cacheutil.cachetocopy(repo): | |
668 | if repo.cachevfs.exists(name): |
|
668 | if repo.cachevfs.exists(name): | |
669 | totalfilesize += repo.cachevfs.lstat(name).st_size |
|
669 | totalfilesize += repo.cachevfs.lstat(name).st_size | |
670 | entries.append((_srccache, name, _filefull, None)) |
|
670 | entries.append((_srccache, name, _filefull, None)) | |
671 | return entries, totalfilesize |
|
671 | return entries, totalfilesize | |
672 |
|
672 | |||
673 |
|
673 | |||
674 | def generatev2(repo, includes, excludes, includeobsmarkers): |
|
674 | def generatev2(repo, includes, excludes, includeobsmarkers): | |
675 | """Emit content for version 2 of a streaming clone. |
|
675 | """Emit content for version 2 of a streaming clone. | |
676 |
|
676 | |||
677 | the data stream consists the following entries: |
|
677 | the data stream consists the following entries: | |
678 | 1) A char representing the file destination (eg: store or cache) |
|
678 | 1) A char representing the file destination (eg: store or cache) | |
679 | 2) A varint containing the length of the filename |
|
679 | 2) A varint containing the length of the filename | |
680 | 3) A varint containing the length of file data |
|
680 | 3) A varint containing the length of file data | |
681 | 4) N bytes containing the filename (the internal, store-agnostic form) |
|
681 | 4) N bytes containing the filename (the internal, store-agnostic form) | |
682 | 5) N bytes containing the file data |
|
682 | 5) N bytes containing the file data | |
683 |
|
683 | |||
684 | Returns a 3-tuple of (file count, file size, data iterator). |
|
684 | Returns a 3-tuple of (file count, file size, data iterator). | |
685 | """ |
|
685 | """ | |
686 |
|
686 | |||
687 | with repo.lock(): |
|
687 | with repo.lock(): | |
688 |
|
688 | |||
689 | repo.ui.debug(b'scanning\n') |
|
689 | repo.ui.debug(b'scanning\n') | |
690 |
|
690 | |||
691 | entries, totalfilesize = _v2_walk( |
|
691 | entries, totalfilesize = _v2_walk( | |
692 | repo, |
|
692 | repo, | |
693 | includes=includes, |
|
693 | includes=includes, | |
694 | excludes=excludes, |
|
694 | excludes=excludes, | |
695 | includeobsmarkers=includeobsmarkers, |
|
695 | includeobsmarkers=includeobsmarkers, | |
696 | ) |
|
696 | ) | |
697 |
|
697 | |||
698 | chunks = _emit2(repo, entries, totalfilesize) |
|
698 | chunks = _emit2(repo, entries, totalfilesize) | |
699 | first = next(chunks) |
|
699 | first = next(chunks) | |
700 | assert first is None |
|
700 | assert first is None | |
701 | _test_sync_point_walk_1(repo) |
|
701 | _test_sync_point_walk_1(repo) | |
702 | _test_sync_point_walk_2(repo) |
|
702 | _test_sync_point_walk_2(repo) | |
703 |
|
703 | |||
704 | return len(entries), totalfilesize, chunks |
|
704 | return len(entries), totalfilesize, chunks | |
705 |
|
705 | |||
706 |
|
706 | |||
707 | @contextlib.contextmanager |
|
707 | @contextlib.contextmanager | |
708 | def nested(*ctxs): |
|
708 | def nested(*ctxs): | |
709 | this = ctxs[0] |
|
709 | this = ctxs[0] | |
710 | rest = ctxs[1:] |
|
710 | rest = ctxs[1:] | |
711 | with this: |
|
711 | with this: | |
712 | if rest: |
|
712 | if rest: | |
713 | with nested(*rest): |
|
713 | with nested(*rest): | |
714 | yield |
|
714 | yield | |
715 | else: |
|
715 | else: | |
716 | yield |
|
716 | yield | |
717 |
|
717 | |||
718 |
|
718 | |||
719 | def consumev2(repo, fp, filecount, filesize): |
|
719 | def consumev2(repo, fp, filecount, filesize): | |
720 | """Apply the contents from a version 2 streaming clone. |
|
720 | """Apply the contents from a version 2 streaming clone. | |
721 |
|
721 | |||
722 | Data is read from an object that only needs to provide a ``read(size)`` |
|
722 | Data is read from an object that only needs to provide a ``read(size)`` | |
723 | method. |
|
723 | method. | |
724 | """ |
|
724 | """ | |
725 | with repo.lock(): |
|
725 | with repo.lock(): | |
726 | repo.ui.status( |
|
726 | repo.ui.status( | |
727 | _(b'%d files to transfer, %s of data\n') |
|
727 | _(b'%d files to transfer, %s of data\n') | |
728 | % (filecount, util.bytecount(filesize)) |
|
728 | % (filecount, util.bytecount(filesize)) | |
729 | ) |
|
729 | ) | |
730 |
|
730 | |||
731 | start = util.timer() |
|
731 | start = util.timer() | |
732 | progress = repo.ui.makeprogress( |
|
732 | progress = repo.ui.makeprogress( | |
733 | _(b'clone'), total=filesize, unit=_(b'bytes') |
|
733 | _(b'clone'), total=filesize, unit=_(b'bytes') | |
734 | ) |
|
734 | ) | |
735 | progress.update(0) |
|
735 | progress.update(0) | |
736 |
|
736 | |||
737 | vfsmap = _makemap(repo) |
|
737 | vfsmap = _makemap(repo) | |
738 | # we keep repo.vfs out of the on purpose, ther are too many danger |
|
738 | # we keep repo.vfs out of the on purpose, ther are too many danger | |
739 | # there (eg: .hg/hgrc), |
|
739 | # there (eg: .hg/hgrc), | |
740 | # |
|
740 | # | |
741 | # this assert is duplicated (from _makemap) as author might think this |
|
741 | # this assert is duplicated (from _makemap) as author might think this | |
742 | # is fine, while this is really not fine. |
|
742 | # is fine, while this is really not fine. | |
743 | if repo.vfs in vfsmap.values(): |
|
743 | if repo.vfs in vfsmap.values(): | |
744 | raise error.ProgrammingError( |
|
744 | raise error.ProgrammingError( | |
745 | b'repo.vfs must not be added to vfsmap for security reasons' |
|
745 | b'repo.vfs must not be added to vfsmap for security reasons' | |
746 | ) |
|
746 | ) | |
747 |
|
747 | |||
748 | with repo.transaction(b'clone'): |
|
748 | with repo.transaction(b'clone'): | |
749 | ctxs = (vfs.backgroundclosing(repo.ui) for vfs in vfsmap.values()) |
|
749 | ctxs = (vfs.backgroundclosing(repo.ui) for vfs in vfsmap.values()) | |
750 | with nested(*ctxs): |
|
750 | with nested(*ctxs): | |
751 | for i in range(filecount): |
|
751 | for i in range(filecount): | |
752 | src = util.readexactly(fp, 1) |
|
752 | src = util.readexactly(fp, 1) | |
753 | vfs = vfsmap[src] |
|
753 | vfs = vfsmap[src] | |
754 | namelen = util.uvarintdecodestream(fp) |
|
754 | namelen = util.uvarintdecodestream(fp) | |
755 | datalen = util.uvarintdecodestream(fp) |
|
755 | datalen = util.uvarintdecodestream(fp) | |
756 |
|
756 | |||
757 | name = util.readexactly(fp, namelen) |
|
757 | name = util.readexactly(fp, namelen) | |
758 |
|
758 | |||
759 | if repo.ui.debugflag: |
|
759 | if repo.ui.debugflag: | |
760 | repo.ui.debug( |
|
760 | repo.ui.debug( | |
761 | b'adding [%s] %s (%s)\n' |
|
761 | b'adding [%s] %s (%s)\n' | |
762 | % (src, name, util.bytecount(datalen)) |
|
762 | % (src, name, util.bytecount(datalen)) | |
763 | ) |
|
763 | ) | |
764 |
|
764 | |||
765 | with vfs(name, b'w') as ofp: |
|
765 | with vfs(name, b'w') as ofp: | |
766 | for chunk in util.filechunkiter(fp, limit=datalen): |
|
766 | for chunk in util.filechunkiter(fp, limit=datalen): | |
767 | progress.increment(step=len(chunk)) |
|
767 | progress.increment(step=len(chunk)) | |
768 | ofp.write(chunk) |
|
768 | ofp.write(chunk) | |
769 |
|
769 | |||
770 | # force @filecache properties to be reloaded from |
|
770 | # force @filecache properties to be reloaded from | |
771 | # streamclone-ed file at next access |
|
771 | # streamclone-ed file at next access | |
772 | repo.invalidate(clearfilecache=True) |
|
772 | repo.invalidate(clearfilecache=True) | |
773 |
|
773 | |||
774 | elapsed = util.timer() - start |
|
774 | elapsed = util.timer() - start | |
775 | if elapsed <= 0: |
|
775 | if elapsed <= 0: | |
776 | elapsed = 0.001 |
|
776 | elapsed = 0.001 | |
777 | repo.ui.status( |
|
777 | repo.ui.status( | |
778 | _(b'transferred %s in %.1f seconds (%s/sec)\n') |
|
778 | _(b'transferred %s in %.1f seconds (%s/sec)\n') | |
779 | % ( |
|
779 | % ( | |
780 | util.bytecount(progress.pos), |
|
780 | util.bytecount(progress.pos), | |
781 | elapsed, |
|
781 | elapsed, | |
782 | util.bytecount(progress.pos / elapsed), |
|
782 | util.bytecount(progress.pos / elapsed), | |
783 | ) |
|
783 | ) | |
784 | ) |
|
784 | ) | |
785 | progress.complete() |
|
785 | progress.complete() | |
786 |
|
786 | |||
787 |
|
787 | |||
788 | def applybundlev2(repo, fp, filecount, filesize, requirements): |
|
788 | def applybundlev2(repo, fp, filecount, filesize, requirements): | |
789 | from . import localrepo |
|
789 | from . import localrepo | |
790 |
|
790 | |||
791 | missingreqs = [r for r in requirements if r not in repo.supported] |
|
791 | missingreqs = [r for r in requirements if r not in repo.supported] | |
792 | if missingreqs: |
|
792 | if missingreqs: | |
793 | raise error.Abort( |
|
793 | raise error.Abort( | |
794 | _(b'unable to apply stream clone: unsupported format: %s') |
|
794 | _(b'unable to apply stream clone: unsupported format: %s') | |
795 | % b', '.join(sorted(missingreqs)) |
|
795 | % b', '.join(sorted(missingreqs)) | |
796 | ) |
|
796 | ) | |
797 |
|
797 | |||
798 | consumev2(repo, fp, filecount, filesize) |
|
798 | consumev2(repo, fp, filecount, filesize) | |
799 |
|
799 | |||
800 | # new requirements = old non-format requirements + |
|
800 | # new requirements = old non-format requirements + | |
801 | # new format-related remote requirements |
|
801 | # new format-related remote requirements | |
802 | # requirements from the streamed-in repository |
|
802 | # requirements from the streamed-in repository | |
803 | repo.requirements = set(requirements) | ( |
|
803 | repo.requirements = set(requirements) | ( | |
804 | repo.requirements - repo.supportedformats |
|
804 | repo.requirements - repo.supportedformats | |
805 | ) |
|
805 | ) | |
806 | repo.svfs.options = localrepo.resolvestorevfsoptions( |
|
806 | repo.svfs.options = localrepo.resolvestorevfsoptions( | |
807 | repo.ui, repo.requirements, repo.features |
|
807 | repo.ui, repo.requirements, repo.features | |
808 | ) |
|
808 | ) | |
809 | scmutil.writereporequirements(repo) |
|
809 | scmutil.writereporequirements(repo) | |
810 |
|
810 | |||
811 |
|
811 | |||
812 | def _copy_files(src_vfs_map, dst_vfs_map, entries, progress): |
|
812 | def _copy_files(src_vfs_map, dst_vfs_map, entries, progress): | |
813 | hardlink = [True] |
|
813 | hardlink = [True] | |
814 |
|
814 | |||
815 | def copy_used(): |
|
815 | def copy_used(): | |
816 | hardlink[0] = False |
|
816 | hardlink[0] = False | |
817 | progress.topic = _(b'copying') |
|
817 | progress.topic = _(b'copying') | |
818 |
|
818 | |||
819 | for k, path, size in entries: |
|
819 | for k, path, size in entries: | |
820 | src_vfs = src_vfs_map[k] |
|
820 | src_vfs = src_vfs_map[k] | |
821 | dst_vfs = dst_vfs_map[k] |
|
821 | dst_vfs = dst_vfs_map[k] | |
822 | src_path = src_vfs.join(path) |
|
822 | src_path = src_vfs.join(path) | |
823 | dst_path = dst_vfs.join(path) |
|
823 | dst_path = dst_vfs.join(path) | |
824 | dirname = dst_vfs.dirname(path) |
|
824 | dirname = dst_vfs.dirname(path) | |
825 | if not dst_vfs.exists(dirname): |
|
825 | if not dst_vfs.exists(dirname): | |
826 | dst_vfs.makedirs(dirname) |
|
826 | dst_vfs.makedirs(dirname) | |
827 | dst_vfs.register_file(path) |
|
827 | dst_vfs.register_file(path) | |
828 | # XXX we could use the #nb_bytes argument. |
|
828 | # XXX we could use the #nb_bytes argument. | |
829 | util.copyfile( |
|
829 | util.copyfile( | |
830 | src_path, |
|
830 | src_path, | |
831 | dst_path, |
|
831 | dst_path, | |
832 | hardlink=hardlink[0], |
|
832 | hardlink=hardlink[0], | |
833 | no_hardlink_cb=copy_used, |
|
833 | no_hardlink_cb=copy_used, | |
834 | check_fs_hardlink=False, |
|
834 | check_fs_hardlink=False, | |
835 | ) |
|
835 | ) | |
836 | progress.increment() |
|
836 | progress.increment() | |
837 | return hardlink[0] |
|
837 | return hardlink[0] | |
838 |
|
838 | |||
839 |
|
839 | |||
840 | def local_copy(src_repo, dest_repo): |
|
840 | def local_copy(src_repo, dest_repo): | |
841 | """copy all content from one local repository to another |
|
841 | """copy all content from one local repository to another | |
842 |
|
842 | |||
843 | This is useful for local clone""" |
|
843 | This is useful for local clone""" | |
844 | src_store_requirements = { |
|
844 | src_store_requirements = { | |
845 | r |
|
845 | r | |
846 | for r in src_repo.requirements |
|
846 | for r in src_repo.requirements | |
847 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS |
|
847 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS | |
848 | } |
|
848 | } | |
849 | dest_store_requirements = { |
|
849 | dest_store_requirements = { | |
850 | r |
|
850 | r | |
851 | for r in dest_repo.requirements |
|
851 | for r in dest_repo.requirements | |
852 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS |
|
852 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS | |
853 | } |
|
853 | } | |
854 | assert src_store_requirements == dest_store_requirements |
|
854 | assert src_store_requirements == dest_store_requirements | |
855 |
|
855 | |||
856 | with dest_repo.lock(): |
|
856 | with dest_repo.lock(): | |
857 | with src_repo.lock(): |
|
857 | with src_repo.lock(): | |
858 |
|
858 | |||
859 | # bookmark is not integrated to the streaming as it might use the |
|
859 | # bookmark is not integrated to the streaming as it might use the | |
860 | # `repo.vfs` and they are too many sentitive data accessible |
|
860 | # `repo.vfs` and they are too many sentitive data accessible | |
861 | # through `repo.vfs` to expose it to streaming clone. |
|
861 | # through `repo.vfs` to expose it to streaming clone. | |
862 | src_book_vfs = bookmarks.bookmarksvfs(src_repo) |
|
862 | src_book_vfs = bookmarks.bookmarksvfs(src_repo) | |
863 | srcbookmarks = src_book_vfs.join(b'bookmarks') |
|
863 | srcbookmarks = src_book_vfs.join(b'bookmarks') | |
864 | bm_count = 0 |
|
864 | bm_count = 0 | |
865 | if os.path.exists(srcbookmarks): |
|
865 | if os.path.exists(srcbookmarks): | |
866 | bm_count = 1 |
|
866 | bm_count = 1 | |
867 |
|
867 | |||
868 | entries, totalfilesize = _v2_walk( |
|
868 | entries, totalfilesize = _v2_walk( | |
869 | src_repo, |
|
869 | src_repo, | |
870 | includes=None, |
|
870 | includes=None, | |
871 | excludes=None, |
|
871 | excludes=None, | |
872 | includeobsmarkers=True, |
|
872 | includeobsmarkers=True, | |
873 | ) |
|
873 | ) | |
874 | src_vfs_map = _makemap(src_repo) |
|
874 | src_vfs_map = _makemap(src_repo) | |
875 | dest_vfs_map = _makemap(dest_repo) |
|
875 | dest_vfs_map = _makemap(dest_repo) | |
876 | progress = src_repo.ui.makeprogress( |
|
876 | progress = src_repo.ui.makeprogress( | |
877 | topic=_(b'linking'), |
|
877 | topic=_(b'linking'), | |
878 | total=len(entries) + bm_count, |
|
878 | total=len(entries) + bm_count, | |
879 | unit=_(b'files'), |
|
879 | unit=_(b'files'), | |
880 | ) |
|
880 | ) | |
881 | # copy files |
|
881 | # copy files | |
882 | # |
|
882 | # | |
883 | # We could copy the full file while the source repository is locked |
|
883 | # We could copy the full file while the source repository is locked | |
884 | # and the other one without the lock. However, in the linking case, |
|
884 | # and the other one without the lock. However, in the linking case, | |
885 | # this would also requires checks that nobody is appending any data |
|
885 | # this would also requires checks that nobody is appending any data | |
886 | # to the files while we do the clone, so this is not done yet. We |
|
886 | # to the files while we do the clone, so this is not done yet. We | |
887 | # could do this blindly when copying files. |
|
887 | # could do this blindly when copying files. | |
888 | files = ((k, path, size) for k, path, ftype, size in entries) |
|
888 | files = ((k, path, size) for k, path, ftype, size in entries) | |
889 | hardlink = _copy_files(src_vfs_map, dest_vfs_map, files, progress) |
|
889 | hardlink = _copy_files(src_vfs_map, dest_vfs_map, files, progress) | |
890 |
|
890 | |||
891 | # copy bookmarks over |
|
891 | # copy bookmarks over | |
892 | if bm_count: |
|
892 | if bm_count: | |
893 | dst_book_vfs = bookmarks.bookmarksvfs(dest_repo) |
|
893 | dst_book_vfs = bookmarks.bookmarksvfs(dest_repo) | |
894 | dstbookmarks = dst_book_vfs.join(b'bookmarks') |
|
894 | dstbookmarks = dst_book_vfs.join(b'bookmarks') | |
895 | util.copyfile(srcbookmarks, dstbookmarks) |
|
895 | util.copyfile(srcbookmarks, dstbookmarks) | |
896 | progress.complete() |
|
896 | progress.complete() | |
897 | if hardlink: |
|
897 | if hardlink: | |
898 | msg = b'linked %d files\n' |
|
898 | msg = b'linked %d files\n' | |
899 | else: |
|
899 | else: | |
900 | msg = b'copied %d files\n' |
|
900 | msg = b'copied %d files\n' | |
901 | src_repo.ui.debug(msg % (len(entries) + bm_count)) |
|
901 | src_repo.ui.debug(msg % (len(entries) + bm_count)) | |
902 |
|
902 | |||
903 | with dest_repo.transaction(b"localclone") as tr: |
|
903 | with dest_repo.transaction(b"localclone") as tr: | |
904 | dest_repo.store.write(tr) |
|
904 | dest_repo.store.write(tr) | |
905 |
|
905 | |||
906 | # clean up transaction file as they do not make sense |
|
906 | # clean up transaction file as they do not make sense | |
907 | undo_files = [(dest_repo.svfs, b'undo.backupfiles')] |
|
907 | undo_files = [(dest_repo.svfs, b'undo.backupfiles')] | |
908 | undo_files.extend(dest_repo.undofiles()) |
|
908 | undo_files.extend(dest_repo.undofiles()) | |
909 | for undovfs, undofile in undo_files: |
|
909 | for undovfs, undofile in undo_files: | |
910 | try: |
|
910 | try: | |
911 | undovfs.unlink(undofile) |
|
911 | undovfs.unlink(undofile) | |
912 | except OSError as e: |
|
912 | except OSError as e: | |
913 | if e.errno != errno.ENOENT: |
|
913 | if e.errno != errno.ENOENT: | |
914 | msg = _(b'error removing %s: %s\n') |
|
914 | msg = _(b'error removing %s: %s\n') | |
915 | path = undovfs.join(undofile) |
|
915 | path = undovfs.join(undofile) | |
916 | e_msg = stringutil.forcebytestr(e) |
|
916 | e_msg = stringutil.forcebytestr(e) | |
917 | msg %= (path, e_msg) |
|
917 | msg %= (path, e_msg) | |
918 | dest_repo.ui.warn(msg) |
|
918 | dest_repo.ui.warn(msg) |
@@ -1,649 +1,649 b'' | |||||
1 | # upgrade.py - functions for in place upgrade of Mercurial repository |
|
1 | # upgrade.py - functions for in place upgrade of Mercurial repository | |
2 | # |
|
2 | # | |
3 | # Copyright (c) 2016-present, Gregory Szorc |
|
3 | # Copyright (c) 2016-present, Gregory Szorc | |
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 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import stat |
|
10 | import stat | |
11 |
|
11 | |||
12 | from ..i18n import _ |
|
12 | from ..i18n import _ | |
13 | from ..pycompat import getattr |
|
13 | from ..pycompat import getattr | |
14 | from .. import ( |
|
14 | from .. import ( | |
15 | changelog, |
|
15 | changelog, | |
16 | error, |
|
16 | error, | |
17 | filelog, |
|
17 | filelog, | |
18 | manifest, |
|
18 | manifest, | |
19 | metadata, |
|
19 | metadata, | |
20 | pycompat, |
|
20 | pycompat, | |
21 | requirements, |
|
21 | requirements, | |
22 | scmutil, |
|
22 | scmutil, | |
23 | store, |
|
23 | store, | |
24 | util, |
|
24 | util, | |
25 | vfs as vfsmod, |
|
25 | vfs as vfsmod, | |
26 | ) |
|
26 | ) | |
27 | from ..revlogutils import ( |
|
27 | from ..revlogutils import ( | |
28 | constants as revlogconst, |
|
28 | constants as revlogconst, | |
29 | flagutil, |
|
29 | flagutil, | |
30 | nodemap, |
|
30 | nodemap, | |
31 | sidedata as sidedatamod, |
|
31 | sidedata as sidedatamod, | |
32 | ) |
|
32 | ) | |
33 | from . import actions as upgrade_actions |
|
33 | from . import actions as upgrade_actions | |
34 |
|
34 | |||
35 |
|
35 | |||
36 | def get_sidedata_helpers(srcrepo, dstrepo): |
|
36 | def get_sidedata_helpers(srcrepo, dstrepo): | |
37 | use_w = srcrepo.ui.configbool(b'experimental', b'worker.repository-upgrade') |
|
37 | use_w = srcrepo.ui.configbool(b'experimental', b'worker.repository-upgrade') | |
38 | sequential = pycompat.iswindows or not use_w |
|
38 | sequential = pycompat.iswindows or not use_w | |
39 | if not sequential: |
|
39 | if not sequential: | |
40 | srcrepo.register_sidedata_computer( |
|
40 | srcrepo.register_sidedata_computer( | |
41 | revlogconst.KIND_CHANGELOG, |
|
41 | revlogconst.KIND_CHANGELOG, | |
42 | sidedatamod.SD_FILES, |
|
42 | sidedatamod.SD_FILES, | |
43 | (sidedatamod.SD_FILES,), |
|
43 | (sidedatamod.SD_FILES,), | |
44 | metadata._get_worker_sidedata_adder(srcrepo, dstrepo), |
|
44 | metadata._get_worker_sidedata_adder(srcrepo, dstrepo), | |
45 | flagutil.REVIDX_HASCOPIESINFO, |
|
45 | flagutil.REVIDX_HASCOPIESINFO, | |
46 | replace=True, |
|
46 | replace=True, | |
47 | ) |
|
47 | ) | |
48 | return sidedatamod.get_sidedata_helpers(srcrepo, dstrepo._wanted_sidedata) |
|
48 | return sidedatamod.get_sidedata_helpers(srcrepo, dstrepo._wanted_sidedata) | |
49 |
|
49 | |||
50 |
|
50 | |||
51 | def _revlogfrompath(repo, rl_type, path): |
|
51 | def _revlogfrompath(repo, rl_type, path): | |
52 | """Obtain a revlog from a repo path. |
|
52 | """Obtain a revlog from a repo path. | |
53 |
|
53 | |||
54 | An instance of the appropriate class is returned. |
|
54 | An instance of the appropriate class is returned. | |
55 | """ |
|
55 | """ | |
56 | if rl_type & store.FILEFLAGS_CHANGELOG: |
|
56 | if rl_type & store.FILEFLAGS_CHANGELOG: | |
57 | return changelog.changelog(repo.svfs) |
|
57 | return changelog.changelog(repo.svfs) | |
58 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: |
|
58 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: | |
59 | mandir = b'' |
|
59 | mandir = b'' | |
60 | if b'/' in path: |
|
60 | if b'/' in path: | |
61 | mandir = path.rsplit(b'/', 1)[0] |
|
61 | mandir = path.rsplit(b'/', 1)[0] | |
62 | return manifest.manifestrevlog( |
|
62 | return manifest.manifestrevlog( | |
63 | repo.nodeconstants, repo.svfs, tree=mandir |
|
63 | repo.nodeconstants, repo.svfs, tree=mandir | |
64 | ) |
|
64 | ) | |
65 | else: |
|
65 | else: | |
66 | # drop the extension and the `data/` prefix |
|
66 | # drop the extension and the `data/` prefix | |
67 | path_part = path.rsplit(b'.', 1)[0].split(b'/', 1) |
|
67 | path_part = path.rsplit(b'.', 1)[0].split(b'/', 1) | |
68 | if len(path_part) < 2: |
|
68 | if len(path_part) < 2: | |
69 | msg = _(b'cannot recognize revlog from filename: %s') |
|
69 | msg = _(b'cannot recognize revlog from filename: %s') | |
70 | msg %= path |
|
70 | msg %= path | |
71 | raise error.Abort(msg) |
|
71 | raise error.Abort(msg) | |
72 | path = path_part[1] |
|
72 | path = path_part[1] | |
73 | return filelog.filelog(repo.svfs, path) |
|
73 | return filelog.filelog(repo.svfs, path) | |
74 |
|
74 | |||
75 |
|
75 | |||
76 | def _copyrevlog(tr, destrepo, oldrl, rl_type, unencodedname): |
|
76 | def _copyrevlog(tr, destrepo, oldrl, rl_type, unencodedname): | |
77 | """copy all relevant files for `oldrl` into `destrepo` store |
|
77 | """copy all relevant files for `oldrl` into `destrepo` store | |
78 |
|
78 | |||
79 | Files are copied "as is" without any transformation. The copy is performed |
|
79 | Files are copied "as is" without any transformation. The copy is performed | |
80 | without extra checks. Callers are responsible for making sure the copied |
|
80 | without extra checks. Callers are responsible for making sure the copied | |
81 | content is compatible with format of the destination repository. |
|
81 | content is compatible with format of the destination repository. | |
82 | """ |
|
82 | """ | |
83 | oldrl = getattr(oldrl, '_revlog', oldrl) |
|
83 | oldrl = getattr(oldrl, '_revlog', oldrl) | |
84 | newrl = _revlogfrompath(destrepo, rl_type, unencodedname) |
|
84 | newrl = _revlogfrompath(destrepo, rl_type, unencodedname) | |
85 | newrl = getattr(newrl, '_revlog', newrl) |
|
85 | newrl = getattr(newrl, '_revlog', newrl) | |
86 |
|
86 | |||
87 | oldvfs = oldrl.opener |
|
87 | oldvfs = oldrl.opener | |
88 | newvfs = newrl.opener |
|
88 | newvfs = newrl.opener | |
89 | oldindex = oldvfs.join(oldrl._indexfile) |
|
89 | oldindex = oldvfs.join(oldrl._indexfile) | |
90 | newindex = newvfs.join(newrl._indexfile) |
|
90 | newindex = newvfs.join(newrl._indexfile) | |
91 | olddata = oldvfs.join(oldrl._datafile) |
|
91 | olddata = oldvfs.join(oldrl._datafile) | |
92 | newdata = newvfs.join(newrl._datafile) |
|
92 | newdata = newvfs.join(newrl._datafile) | |
93 |
|
93 | |||
94 | with newvfs(newrl._indexfile, b'w'): |
|
94 | with newvfs(newrl._indexfile, b'w'): | |
95 | pass # create all the directories |
|
95 | pass # create all the directories | |
96 |
|
96 | |||
97 | util.copyfile(oldindex, newindex) |
|
97 | util.copyfile(oldindex, newindex) | |
98 | copydata = oldrl.opener.exists(oldrl._datafile) |
|
98 | copydata = oldrl.opener.exists(oldrl._datafile) | |
99 | if copydata: |
|
99 | if copydata: | |
100 | util.copyfile(olddata, newdata) |
|
100 | util.copyfile(olddata, newdata) | |
101 |
|
101 | |||
102 | if rl_type & store.FILEFLAGS_FILELOG: |
|
102 | if rl_type & store.FILEFLAGS_FILELOG: | |
103 | destrepo.svfs.fncache.add(unencodedname) |
|
103 | destrepo.svfs.fncache.add(unencodedname) | |
104 | if copydata: |
|
104 | if copydata: | |
105 | destrepo.svfs.fncache.add(unencodedname[:-2] + b'.d') |
|
105 | destrepo.svfs.fncache.add(unencodedname[:-2] + b'.d') | |
106 |
|
106 | |||
107 |
|
107 | |||
108 | UPGRADE_CHANGELOG = b"changelog" |
|
108 | UPGRADE_CHANGELOG = b"changelog" | |
109 | UPGRADE_MANIFEST = b"manifest" |
|
109 | UPGRADE_MANIFEST = b"manifest" | |
110 | UPGRADE_FILELOGS = b"all-filelogs" |
|
110 | UPGRADE_FILELOGS = b"all-filelogs" | |
111 |
|
111 | |||
112 | UPGRADE_ALL_REVLOGS = frozenset( |
|
112 | UPGRADE_ALL_REVLOGS = frozenset( | |
113 | [UPGRADE_CHANGELOG, UPGRADE_MANIFEST, UPGRADE_FILELOGS] |
|
113 | [UPGRADE_CHANGELOG, UPGRADE_MANIFEST, UPGRADE_FILELOGS] | |
114 | ) |
|
114 | ) | |
115 |
|
115 | |||
116 |
|
116 | |||
117 | def matchrevlog(revlogfilter, rl_type): |
|
117 | def matchrevlog(revlogfilter, rl_type): | |
118 | """check if a revlog is selected for cloning. |
|
118 | """check if a revlog is selected for cloning. | |
119 |
|
119 | |||
120 | In other words, are there any updates which need to be done on revlog |
|
120 | In other words, are there any updates which need to be done on revlog | |
121 | or it can be blindly copied. |
|
121 | or it can be blindly copied. | |
122 |
|
122 | |||
123 | The store entry is checked against the passed filter""" |
|
123 | The store entry is checked against the passed filter""" | |
124 | if rl_type & store.FILEFLAGS_CHANGELOG: |
|
124 | if rl_type & store.FILEFLAGS_CHANGELOG: | |
125 | return UPGRADE_CHANGELOG in revlogfilter |
|
125 | return UPGRADE_CHANGELOG in revlogfilter | |
126 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: |
|
126 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: | |
127 | return UPGRADE_MANIFEST in revlogfilter |
|
127 | return UPGRADE_MANIFEST in revlogfilter | |
128 | assert rl_type & store.FILEFLAGS_FILELOG |
|
128 | assert rl_type & store.FILEFLAGS_FILELOG | |
129 | return UPGRADE_FILELOGS in revlogfilter |
|
129 | return UPGRADE_FILELOGS in revlogfilter | |
130 |
|
130 | |||
131 |
|
131 | |||
132 | def _perform_clone( |
|
132 | def _perform_clone( | |
133 | ui, |
|
133 | ui, | |
134 | dstrepo, |
|
134 | dstrepo, | |
135 | tr, |
|
135 | tr, | |
136 | old_revlog, |
|
136 | old_revlog, | |
137 | rl_type, |
|
137 | rl_type, | |
138 | unencoded, |
|
138 | unencoded, | |
139 | upgrade_op, |
|
139 | upgrade_op, | |
140 | sidedata_helpers, |
|
140 | sidedata_helpers, | |
141 | oncopiedrevision, |
|
141 | oncopiedrevision, | |
142 | ): |
|
142 | ): | |
143 | """returns the new revlog object created""" |
|
143 | """returns the new revlog object created""" | |
144 | newrl = None |
|
144 | newrl = None | |
145 | if matchrevlog(upgrade_op.revlogs_to_process, rl_type): |
|
145 | if matchrevlog(upgrade_op.revlogs_to_process, rl_type): | |
146 | ui.note( |
|
146 | ui.note( | |
147 | _(b'cloning %d revisions from %s\n') % (len(old_revlog), unencoded) |
|
147 | _(b'cloning %d revisions from %s\n') % (len(old_revlog), unencoded) | |
148 | ) |
|
148 | ) | |
149 | newrl = _revlogfrompath(dstrepo, rl_type, unencoded) |
|
149 | newrl = _revlogfrompath(dstrepo, rl_type, unencoded) | |
150 | old_revlog.clone( |
|
150 | old_revlog.clone( | |
151 | tr, |
|
151 | tr, | |
152 | newrl, |
|
152 | newrl, | |
153 | addrevisioncb=oncopiedrevision, |
|
153 | addrevisioncb=oncopiedrevision, | |
154 | deltareuse=upgrade_op.delta_reuse_mode, |
|
154 | deltareuse=upgrade_op.delta_reuse_mode, | |
155 | forcedeltabothparents=upgrade_op.force_re_delta_both_parents, |
|
155 | forcedeltabothparents=upgrade_op.force_re_delta_both_parents, | |
156 | sidedata_helpers=sidedata_helpers, |
|
156 | sidedata_helpers=sidedata_helpers, | |
157 | ) |
|
157 | ) | |
158 | else: |
|
158 | else: | |
159 | msg = _(b'blindly copying %s containing %i revisions\n') |
|
159 | msg = _(b'blindly copying %s containing %i revisions\n') | |
160 | ui.note(msg % (unencoded, len(old_revlog))) |
|
160 | ui.note(msg % (unencoded, len(old_revlog))) | |
161 | _copyrevlog(tr, dstrepo, old_revlog, rl_type, unencoded) |
|
161 | _copyrevlog(tr, dstrepo, old_revlog, rl_type, unencoded) | |
162 |
|
162 | |||
163 | newrl = _revlogfrompath(dstrepo, rl_type, unencoded) |
|
163 | newrl = _revlogfrompath(dstrepo, rl_type, unencoded) | |
164 | return newrl |
|
164 | return newrl | |
165 |
|
165 | |||
166 |
|
166 | |||
167 | def _clonerevlogs( |
|
167 | def _clonerevlogs( | |
168 | ui, |
|
168 | ui, | |
169 | srcrepo, |
|
169 | srcrepo, | |
170 | dstrepo, |
|
170 | dstrepo, | |
171 | tr, |
|
171 | tr, | |
172 | upgrade_op, |
|
172 | upgrade_op, | |
173 | ): |
|
173 | ): | |
174 | """Copy revlogs between 2 repos.""" |
|
174 | """Copy revlogs between 2 repos.""" | |
175 | revcount = 0 |
|
175 | revcount = 0 | |
176 | srcsize = 0 |
|
176 | srcsize = 0 | |
177 | srcrawsize = 0 |
|
177 | srcrawsize = 0 | |
178 | dstsize = 0 |
|
178 | dstsize = 0 | |
179 | fcount = 0 |
|
179 | fcount = 0 | |
180 | frevcount = 0 |
|
180 | frevcount = 0 | |
181 | fsrcsize = 0 |
|
181 | fsrcsize = 0 | |
182 | frawsize = 0 |
|
182 | frawsize = 0 | |
183 | fdstsize = 0 |
|
183 | fdstsize = 0 | |
184 | mcount = 0 |
|
184 | mcount = 0 | |
185 | mrevcount = 0 |
|
185 | mrevcount = 0 | |
186 | msrcsize = 0 |
|
186 | msrcsize = 0 | |
187 | mrawsize = 0 |
|
187 | mrawsize = 0 | |
188 | mdstsize = 0 |
|
188 | mdstsize = 0 | |
189 | crevcount = 0 |
|
189 | crevcount = 0 | |
190 | csrcsize = 0 |
|
190 | csrcsize = 0 | |
191 | crawsize = 0 |
|
191 | crawsize = 0 | |
192 | cdstsize = 0 |
|
192 | cdstsize = 0 | |
193 |
|
193 | |||
194 | alldatafiles = list(srcrepo.store.walk()) |
|
194 | alldatafiles = list(srcrepo.store.walk()) | |
195 | # mapping of data files which needs to be cloned |
|
195 | # mapping of data files which needs to be cloned | |
196 | # key is unencoded filename |
|
196 | # key is unencoded filename | |
197 | # value is revlog_object_from_srcrepo |
|
197 | # value is revlog_object_from_srcrepo | |
198 | manifests = {} |
|
198 | manifests = {} | |
199 | changelogs = {} |
|
199 | changelogs = {} | |
200 | filelogs = {} |
|
200 | filelogs = {} | |
201 |
|
201 | |||
202 | # Perform a pass to collect metadata. This validates we can open all |
|
202 | # Perform a pass to collect metadata. This validates we can open all | |
203 | # source files and allows a unified progress bar to be displayed. |
|
203 | # source files and allows a unified progress bar to be displayed. | |
204 |
for rl_type, unencoded, |
|
204 | for rl_type, unencoded, size in alldatafiles: | |
205 | if not rl_type & store.FILEFLAGS_REVLOG_MAIN: |
|
205 | if not rl_type & store.FILEFLAGS_REVLOG_MAIN: | |
206 | continue |
|
206 | continue | |
207 |
|
207 | |||
208 | # the store.walk function will wrongly pickup transaction backup and |
|
208 | # the store.walk function will wrongly pickup transaction backup and | |
209 | # get confused. As a quick fix for 5.9 release, we ignore those. |
|
209 | # get confused. As a quick fix for 5.9 release, we ignore those. | |
210 | # (this is not a module constants because it seems better to keep the |
|
210 | # (this is not a module constants because it seems better to keep the | |
211 | # hack together) |
|
211 | # hack together) | |
212 | skip_undo = ( |
|
212 | skip_undo = ( | |
213 | b'undo.backup.00changelog.i', |
|
213 | b'undo.backup.00changelog.i', | |
214 | b'undo.backup.00manifest.i', |
|
214 | b'undo.backup.00manifest.i', | |
215 | ) |
|
215 | ) | |
216 | if unencoded in skip_undo: |
|
216 | if unencoded in skip_undo: | |
217 | continue |
|
217 | continue | |
218 |
|
218 | |||
219 | rl = _revlogfrompath(srcrepo, rl_type, unencoded) |
|
219 | rl = _revlogfrompath(srcrepo, rl_type, unencoded) | |
220 |
|
220 | |||
221 | info = rl.storageinfo( |
|
221 | info = rl.storageinfo( | |
222 | exclusivefiles=True, |
|
222 | exclusivefiles=True, | |
223 | revisionscount=True, |
|
223 | revisionscount=True, | |
224 | trackedsize=True, |
|
224 | trackedsize=True, | |
225 | storedsize=True, |
|
225 | storedsize=True, | |
226 | ) |
|
226 | ) | |
227 |
|
227 | |||
228 | revcount += info[b'revisionscount'] or 0 |
|
228 | revcount += info[b'revisionscount'] or 0 | |
229 | datasize = info[b'storedsize'] or 0 |
|
229 | datasize = info[b'storedsize'] or 0 | |
230 | rawsize = info[b'trackedsize'] or 0 |
|
230 | rawsize = info[b'trackedsize'] or 0 | |
231 |
|
231 | |||
232 | srcsize += datasize |
|
232 | srcsize += datasize | |
233 | srcrawsize += rawsize |
|
233 | srcrawsize += rawsize | |
234 |
|
234 | |||
235 | # This is for the separate progress bars. |
|
235 | # This is for the separate progress bars. | |
236 | if rl_type & store.FILEFLAGS_CHANGELOG: |
|
236 | if rl_type & store.FILEFLAGS_CHANGELOG: | |
237 | changelogs[unencoded] = (rl_type, rl) |
|
237 | changelogs[unencoded] = (rl_type, rl) | |
238 | crevcount += len(rl) |
|
238 | crevcount += len(rl) | |
239 | csrcsize += datasize |
|
239 | csrcsize += datasize | |
240 | crawsize += rawsize |
|
240 | crawsize += rawsize | |
241 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: |
|
241 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: | |
242 | manifests[unencoded] = (rl_type, rl) |
|
242 | manifests[unencoded] = (rl_type, rl) | |
243 | mcount += 1 |
|
243 | mcount += 1 | |
244 | mrevcount += len(rl) |
|
244 | mrevcount += len(rl) | |
245 | msrcsize += datasize |
|
245 | msrcsize += datasize | |
246 | mrawsize += rawsize |
|
246 | mrawsize += rawsize | |
247 | elif rl_type & store.FILEFLAGS_FILELOG: |
|
247 | elif rl_type & store.FILEFLAGS_FILELOG: | |
248 | filelogs[unencoded] = (rl_type, rl) |
|
248 | filelogs[unencoded] = (rl_type, rl) | |
249 | fcount += 1 |
|
249 | fcount += 1 | |
250 | frevcount += len(rl) |
|
250 | frevcount += len(rl) | |
251 | fsrcsize += datasize |
|
251 | fsrcsize += datasize | |
252 | frawsize += rawsize |
|
252 | frawsize += rawsize | |
253 | else: |
|
253 | else: | |
254 | error.ProgrammingError(b'unknown revlog type') |
|
254 | error.ProgrammingError(b'unknown revlog type') | |
255 |
|
255 | |||
256 | if not revcount: |
|
256 | if not revcount: | |
257 | return |
|
257 | return | |
258 |
|
258 | |||
259 | ui.status( |
|
259 | ui.status( | |
260 | _( |
|
260 | _( | |
261 | b'migrating %d total revisions (%d in filelogs, %d in manifests, ' |
|
261 | b'migrating %d total revisions (%d in filelogs, %d in manifests, ' | |
262 | b'%d in changelog)\n' |
|
262 | b'%d in changelog)\n' | |
263 | ) |
|
263 | ) | |
264 | % (revcount, frevcount, mrevcount, crevcount) |
|
264 | % (revcount, frevcount, mrevcount, crevcount) | |
265 | ) |
|
265 | ) | |
266 | ui.status( |
|
266 | ui.status( | |
267 | _(b'migrating %s in store; %s tracked data\n') |
|
267 | _(b'migrating %s in store; %s tracked data\n') | |
268 | % ((util.bytecount(srcsize), util.bytecount(srcrawsize))) |
|
268 | % ((util.bytecount(srcsize), util.bytecount(srcrawsize))) | |
269 | ) |
|
269 | ) | |
270 |
|
270 | |||
271 | # Used to keep track of progress. |
|
271 | # Used to keep track of progress. | |
272 | progress = None |
|
272 | progress = None | |
273 |
|
273 | |||
274 | def oncopiedrevision(rl, rev, node): |
|
274 | def oncopiedrevision(rl, rev, node): | |
275 | progress.increment() |
|
275 | progress.increment() | |
276 |
|
276 | |||
277 | sidedata_helpers = get_sidedata_helpers(srcrepo, dstrepo) |
|
277 | sidedata_helpers = get_sidedata_helpers(srcrepo, dstrepo) | |
278 |
|
278 | |||
279 | # Migrating filelogs |
|
279 | # Migrating filelogs | |
280 | ui.status( |
|
280 | ui.status( | |
281 | _( |
|
281 | _( | |
282 | b'migrating %d filelogs containing %d revisions ' |
|
282 | b'migrating %d filelogs containing %d revisions ' | |
283 | b'(%s in store; %s tracked data)\n' |
|
283 | b'(%s in store; %s tracked data)\n' | |
284 | ) |
|
284 | ) | |
285 | % ( |
|
285 | % ( | |
286 | fcount, |
|
286 | fcount, | |
287 | frevcount, |
|
287 | frevcount, | |
288 | util.bytecount(fsrcsize), |
|
288 | util.bytecount(fsrcsize), | |
289 | util.bytecount(frawsize), |
|
289 | util.bytecount(frawsize), | |
290 | ) |
|
290 | ) | |
291 | ) |
|
291 | ) | |
292 | progress = srcrepo.ui.makeprogress(_(b'file revisions'), total=frevcount) |
|
292 | progress = srcrepo.ui.makeprogress(_(b'file revisions'), total=frevcount) | |
293 | for unencoded, (rl_type, oldrl) in sorted(filelogs.items()): |
|
293 | for unencoded, (rl_type, oldrl) in sorted(filelogs.items()): | |
294 | newrl = _perform_clone( |
|
294 | newrl = _perform_clone( | |
295 | ui, |
|
295 | ui, | |
296 | dstrepo, |
|
296 | dstrepo, | |
297 | tr, |
|
297 | tr, | |
298 | oldrl, |
|
298 | oldrl, | |
299 | rl_type, |
|
299 | rl_type, | |
300 | unencoded, |
|
300 | unencoded, | |
301 | upgrade_op, |
|
301 | upgrade_op, | |
302 | sidedata_helpers, |
|
302 | sidedata_helpers, | |
303 | oncopiedrevision, |
|
303 | oncopiedrevision, | |
304 | ) |
|
304 | ) | |
305 | info = newrl.storageinfo(storedsize=True) |
|
305 | info = newrl.storageinfo(storedsize=True) | |
306 | fdstsize += info[b'storedsize'] or 0 |
|
306 | fdstsize += info[b'storedsize'] or 0 | |
307 | ui.status( |
|
307 | ui.status( | |
308 | _( |
|
308 | _( | |
309 | b'finished migrating %d filelog revisions across %d ' |
|
309 | b'finished migrating %d filelog revisions across %d ' | |
310 | b'filelogs; change in size: %s\n' |
|
310 | b'filelogs; change in size: %s\n' | |
311 | ) |
|
311 | ) | |
312 | % (frevcount, fcount, util.bytecount(fdstsize - fsrcsize)) |
|
312 | % (frevcount, fcount, util.bytecount(fdstsize - fsrcsize)) | |
313 | ) |
|
313 | ) | |
314 |
|
314 | |||
315 | # Migrating manifests |
|
315 | # Migrating manifests | |
316 | ui.status( |
|
316 | ui.status( | |
317 | _( |
|
317 | _( | |
318 | b'migrating %d manifests containing %d revisions ' |
|
318 | b'migrating %d manifests containing %d revisions ' | |
319 | b'(%s in store; %s tracked data)\n' |
|
319 | b'(%s in store; %s tracked data)\n' | |
320 | ) |
|
320 | ) | |
321 | % ( |
|
321 | % ( | |
322 | mcount, |
|
322 | mcount, | |
323 | mrevcount, |
|
323 | mrevcount, | |
324 | util.bytecount(msrcsize), |
|
324 | util.bytecount(msrcsize), | |
325 | util.bytecount(mrawsize), |
|
325 | util.bytecount(mrawsize), | |
326 | ) |
|
326 | ) | |
327 | ) |
|
327 | ) | |
328 | if progress: |
|
328 | if progress: | |
329 | progress.complete() |
|
329 | progress.complete() | |
330 | progress = srcrepo.ui.makeprogress( |
|
330 | progress = srcrepo.ui.makeprogress( | |
331 | _(b'manifest revisions'), total=mrevcount |
|
331 | _(b'manifest revisions'), total=mrevcount | |
332 | ) |
|
332 | ) | |
333 | for unencoded, (rl_type, oldrl) in sorted(manifests.items()): |
|
333 | for unencoded, (rl_type, oldrl) in sorted(manifests.items()): | |
334 | newrl = _perform_clone( |
|
334 | newrl = _perform_clone( | |
335 | ui, |
|
335 | ui, | |
336 | dstrepo, |
|
336 | dstrepo, | |
337 | tr, |
|
337 | tr, | |
338 | oldrl, |
|
338 | oldrl, | |
339 | rl_type, |
|
339 | rl_type, | |
340 | unencoded, |
|
340 | unencoded, | |
341 | upgrade_op, |
|
341 | upgrade_op, | |
342 | sidedata_helpers, |
|
342 | sidedata_helpers, | |
343 | oncopiedrevision, |
|
343 | oncopiedrevision, | |
344 | ) |
|
344 | ) | |
345 | info = newrl.storageinfo(storedsize=True) |
|
345 | info = newrl.storageinfo(storedsize=True) | |
346 | mdstsize += info[b'storedsize'] or 0 |
|
346 | mdstsize += info[b'storedsize'] or 0 | |
347 | ui.status( |
|
347 | ui.status( | |
348 | _( |
|
348 | _( | |
349 | b'finished migrating %d manifest revisions across %d ' |
|
349 | b'finished migrating %d manifest revisions across %d ' | |
350 | b'manifests; change in size: %s\n' |
|
350 | b'manifests; change in size: %s\n' | |
351 | ) |
|
351 | ) | |
352 | % (mrevcount, mcount, util.bytecount(mdstsize - msrcsize)) |
|
352 | % (mrevcount, mcount, util.bytecount(mdstsize - msrcsize)) | |
353 | ) |
|
353 | ) | |
354 |
|
354 | |||
355 | # Migrating changelog |
|
355 | # Migrating changelog | |
356 | ui.status( |
|
356 | ui.status( | |
357 | _( |
|
357 | _( | |
358 | b'migrating changelog containing %d revisions ' |
|
358 | b'migrating changelog containing %d revisions ' | |
359 | b'(%s in store; %s tracked data)\n' |
|
359 | b'(%s in store; %s tracked data)\n' | |
360 | ) |
|
360 | ) | |
361 | % ( |
|
361 | % ( | |
362 | crevcount, |
|
362 | crevcount, | |
363 | util.bytecount(csrcsize), |
|
363 | util.bytecount(csrcsize), | |
364 | util.bytecount(crawsize), |
|
364 | util.bytecount(crawsize), | |
365 | ) |
|
365 | ) | |
366 | ) |
|
366 | ) | |
367 | if progress: |
|
367 | if progress: | |
368 | progress.complete() |
|
368 | progress.complete() | |
369 | progress = srcrepo.ui.makeprogress( |
|
369 | progress = srcrepo.ui.makeprogress( | |
370 | _(b'changelog revisions'), total=crevcount |
|
370 | _(b'changelog revisions'), total=crevcount | |
371 | ) |
|
371 | ) | |
372 | for unencoded, (rl_type, oldrl) in sorted(changelogs.items()): |
|
372 | for unencoded, (rl_type, oldrl) in sorted(changelogs.items()): | |
373 | newrl = _perform_clone( |
|
373 | newrl = _perform_clone( | |
374 | ui, |
|
374 | ui, | |
375 | dstrepo, |
|
375 | dstrepo, | |
376 | tr, |
|
376 | tr, | |
377 | oldrl, |
|
377 | oldrl, | |
378 | rl_type, |
|
378 | rl_type, | |
379 | unencoded, |
|
379 | unencoded, | |
380 | upgrade_op, |
|
380 | upgrade_op, | |
381 | sidedata_helpers, |
|
381 | sidedata_helpers, | |
382 | oncopiedrevision, |
|
382 | oncopiedrevision, | |
383 | ) |
|
383 | ) | |
384 | info = newrl.storageinfo(storedsize=True) |
|
384 | info = newrl.storageinfo(storedsize=True) | |
385 | cdstsize += info[b'storedsize'] or 0 |
|
385 | cdstsize += info[b'storedsize'] or 0 | |
386 | progress.complete() |
|
386 | progress.complete() | |
387 | ui.status( |
|
387 | ui.status( | |
388 | _( |
|
388 | _( | |
389 | b'finished migrating %d changelog revisions; change in size: ' |
|
389 | b'finished migrating %d changelog revisions; change in size: ' | |
390 | b'%s\n' |
|
390 | b'%s\n' | |
391 | ) |
|
391 | ) | |
392 | % (crevcount, util.bytecount(cdstsize - csrcsize)) |
|
392 | % (crevcount, util.bytecount(cdstsize - csrcsize)) | |
393 | ) |
|
393 | ) | |
394 |
|
394 | |||
395 | dstsize = fdstsize + mdstsize + cdstsize |
|
395 | dstsize = fdstsize + mdstsize + cdstsize | |
396 | ui.status( |
|
396 | ui.status( | |
397 | _( |
|
397 | _( | |
398 | b'finished migrating %d total revisions; total change in store ' |
|
398 | b'finished migrating %d total revisions; total change in store ' | |
399 | b'size: %s\n' |
|
399 | b'size: %s\n' | |
400 | ) |
|
400 | ) | |
401 | % (revcount, util.bytecount(dstsize - srcsize)) |
|
401 | % (revcount, util.bytecount(dstsize - srcsize)) | |
402 | ) |
|
402 | ) | |
403 |
|
403 | |||
404 |
|
404 | |||
405 | def _files_to_copy_post_revlog_clone(srcrepo): |
|
405 | def _files_to_copy_post_revlog_clone(srcrepo): | |
406 | """yields files which should be copied to destination after revlogs |
|
406 | """yields files which should be copied to destination after revlogs | |
407 | are cloned""" |
|
407 | are cloned""" | |
408 | for path, kind, st in sorted(srcrepo.store.vfs.readdir(b'', stat=True)): |
|
408 | for path, kind, st in sorted(srcrepo.store.vfs.readdir(b'', stat=True)): | |
409 | # don't copy revlogs as they are already cloned |
|
409 | # don't copy revlogs as they are already cloned | |
410 | if store.revlog_type(path) is not None: |
|
410 | if store.revlog_type(path) is not None: | |
411 | continue |
|
411 | continue | |
412 | # Skip transaction related files. |
|
412 | # Skip transaction related files. | |
413 | if path.startswith(b'undo'): |
|
413 | if path.startswith(b'undo'): | |
414 | continue |
|
414 | continue | |
415 | # Only copy regular files. |
|
415 | # Only copy regular files. | |
416 | if kind != stat.S_IFREG: |
|
416 | if kind != stat.S_IFREG: | |
417 | continue |
|
417 | continue | |
418 | # Skip other skipped files. |
|
418 | # Skip other skipped files. | |
419 | if path in (b'lock', b'fncache'): |
|
419 | if path in (b'lock', b'fncache'): | |
420 | continue |
|
420 | continue | |
421 | # TODO: should we skip cache too? |
|
421 | # TODO: should we skip cache too? | |
422 |
|
422 | |||
423 | yield path |
|
423 | yield path | |
424 |
|
424 | |||
425 |
|
425 | |||
426 | def _replacestores(currentrepo, upgradedrepo, backupvfs, upgrade_op): |
|
426 | def _replacestores(currentrepo, upgradedrepo, backupvfs, upgrade_op): | |
427 | """Replace the stores after current repository is upgraded |
|
427 | """Replace the stores after current repository is upgraded | |
428 |
|
428 | |||
429 | Creates a backup of current repository store at backup path |
|
429 | Creates a backup of current repository store at backup path | |
430 | Replaces upgraded store files in current repo from upgraded one |
|
430 | Replaces upgraded store files in current repo from upgraded one | |
431 |
|
431 | |||
432 | Arguments: |
|
432 | Arguments: | |
433 | currentrepo: repo object of current repository |
|
433 | currentrepo: repo object of current repository | |
434 | upgradedrepo: repo object of the upgraded data |
|
434 | upgradedrepo: repo object of the upgraded data | |
435 | backupvfs: vfs object for the backup path |
|
435 | backupvfs: vfs object for the backup path | |
436 | upgrade_op: upgrade operation object |
|
436 | upgrade_op: upgrade operation object | |
437 | to be used to decide what all is upgraded |
|
437 | to be used to decide what all is upgraded | |
438 | """ |
|
438 | """ | |
439 | # TODO: don't blindly rename everything in store |
|
439 | # TODO: don't blindly rename everything in store | |
440 | # There can be upgrades where store is not touched at all |
|
440 | # There can be upgrades where store is not touched at all | |
441 | if upgrade_op.backup_store: |
|
441 | if upgrade_op.backup_store: | |
442 | util.rename(currentrepo.spath, backupvfs.join(b'store')) |
|
442 | util.rename(currentrepo.spath, backupvfs.join(b'store')) | |
443 | else: |
|
443 | else: | |
444 | currentrepo.vfs.rmtree(b'store', forcibly=True) |
|
444 | currentrepo.vfs.rmtree(b'store', forcibly=True) | |
445 | util.rename(upgradedrepo.spath, currentrepo.spath) |
|
445 | util.rename(upgradedrepo.spath, currentrepo.spath) | |
446 |
|
446 | |||
447 |
|
447 | |||
448 | def finishdatamigration(ui, srcrepo, dstrepo, requirements): |
|
448 | def finishdatamigration(ui, srcrepo, dstrepo, requirements): | |
449 | """Hook point for extensions to perform additional actions during upgrade. |
|
449 | """Hook point for extensions to perform additional actions during upgrade. | |
450 |
|
450 | |||
451 | This function is called after revlogs and store files have been copied but |
|
451 | This function is called after revlogs and store files have been copied but | |
452 | before the new store is swapped into the original location. |
|
452 | before the new store is swapped into the original location. | |
453 | """ |
|
453 | """ | |
454 |
|
454 | |||
455 |
|
455 | |||
456 | def upgrade(ui, srcrepo, dstrepo, upgrade_op): |
|
456 | def upgrade(ui, srcrepo, dstrepo, upgrade_op): | |
457 | """Do the low-level work of upgrading a repository. |
|
457 | """Do the low-level work of upgrading a repository. | |
458 |
|
458 | |||
459 | The upgrade is effectively performed as a copy between a source |
|
459 | The upgrade is effectively performed as a copy between a source | |
460 | repository and a temporary destination repository. |
|
460 | repository and a temporary destination repository. | |
461 |
|
461 | |||
462 | The source repository is unmodified for as long as possible so the |
|
462 | The source repository is unmodified for as long as possible so the | |
463 | upgrade can abort at any time without causing loss of service for |
|
463 | upgrade can abort at any time without causing loss of service for | |
464 | readers and without corrupting the source repository. |
|
464 | readers and without corrupting the source repository. | |
465 | """ |
|
465 | """ | |
466 | assert srcrepo.currentwlock() |
|
466 | assert srcrepo.currentwlock() | |
467 | assert dstrepo.currentwlock() |
|
467 | assert dstrepo.currentwlock() | |
468 | backuppath = None |
|
468 | backuppath = None | |
469 | backupvfs = None |
|
469 | backupvfs = None | |
470 |
|
470 | |||
471 | ui.status( |
|
471 | ui.status( | |
472 | _( |
|
472 | _( | |
473 | b'(it is safe to interrupt this process any time before ' |
|
473 | b'(it is safe to interrupt this process any time before ' | |
474 | b'data migration completes)\n' |
|
474 | b'data migration completes)\n' | |
475 | ) |
|
475 | ) | |
476 | ) |
|
476 | ) | |
477 |
|
477 | |||
478 | if upgrade_actions.dirstatev2 in upgrade_op.upgrade_actions: |
|
478 | if upgrade_actions.dirstatev2 in upgrade_op.upgrade_actions: | |
479 | ui.status(_(b'upgrading to dirstate-v2 from v1\n')) |
|
479 | ui.status(_(b'upgrading to dirstate-v2 from v1\n')) | |
480 | upgrade_dirstate(ui, srcrepo, upgrade_op, b'v1', b'v2') |
|
480 | upgrade_dirstate(ui, srcrepo, upgrade_op, b'v1', b'v2') | |
481 | upgrade_op.upgrade_actions.remove(upgrade_actions.dirstatev2) |
|
481 | upgrade_op.upgrade_actions.remove(upgrade_actions.dirstatev2) | |
482 |
|
482 | |||
483 | if upgrade_actions.dirstatev2 in upgrade_op.removed_actions: |
|
483 | if upgrade_actions.dirstatev2 in upgrade_op.removed_actions: | |
484 | ui.status(_(b'downgrading from dirstate-v2 to v1\n')) |
|
484 | ui.status(_(b'downgrading from dirstate-v2 to v1\n')) | |
485 | upgrade_dirstate(ui, srcrepo, upgrade_op, b'v2', b'v1') |
|
485 | upgrade_dirstate(ui, srcrepo, upgrade_op, b'v2', b'v1') | |
486 | upgrade_op.removed_actions.remove(upgrade_actions.dirstatev2) |
|
486 | upgrade_op.removed_actions.remove(upgrade_actions.dirstatev2) | |
487 |
|
487 | |||
488 | if not (upgrade_op.upgrade_actions or upgrade_op.removed_actions): |
|
488 | if not (upgrade_op.upgrade_actions or upgrade_op.removed_actions): | |
489 | return |
|
489 | return | |
490 |
|
490 | |||
491 | if upgrade_op.requirements_only: |
|
491 | if upgrade_op.requirements_only: | |
492 | ui.status(_(b'upgrading repository requirements\n')) |
|
492 | ui.status(_(b'upgrading repository requirements\n')) | |
493 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
493 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
494 | # if there is only one action and that is persistent nodemap upgrade |
|
494 | # if there is only one action and that is persistent nodemap upgrade | |
495 | # directly write the nodemap file and update requirements instead of going |
|
495 | # directly write the nodemap file and update requirements instead of going | |
496 | # through the whole cloning process |
|
496 | # through the whole cloning process | |
497 | elif ( |
|
497 | elif ( | |
498 | len(upgrade_op.upgrade_actions) == 1 |
|
498 | len(upgrade_op.upgrade_actions) == 1 | |
499 | and b'persistent-nodemap' in upgrade_op.upgrade_actions_names |
|
499 | and b'persistent-nodemap' in upgrade_op.upgrade_actions_names | |
500 | and not upgrade_op.removed_actions |
|
500 | and not upgrade_op.removed_actions | |
501 | ): |
|
501 | ): | |
502 | ui.status( |
|
502 | ui.status( | |
503 | _(b'upgrading repository to use persistent nodemap feature\n') |
|
503 | _(b'upgrading repository to use persistent nodemap feature\n') | |
504 | ) |
|
504 | ) | |
505 | with srcrepo.transaction(b'upgrade') as tr: |
|
505 | with srcrepo.transaction(b'upgrade') as tr: | |
506 | unfi = srcrepo.unfiltered() |
|
506 | unfi = srcrepo.unfiltered() | |
507 | cl = unfi.changelog |
|
507 | cl = unfi.changelog | |
508 | nodemap.persist_nodemap(tr, cl, force=True) |
|
508 | nodemap.persist_nodemap(tr, cl, force=True) | |
509 | # we want to directly operate on the underlying revlog to force |
|
509 | # we want to directly operate on the underlying revlog to force | |
510 | # create a nodemap file. This is fine since this is upgrade code |
|
510 | # create a nodemap file. This is fine since this is upgrade code | |
511 | # and it heavily relies on repository being revlog based |
|
511 | # and it heavily relies on repository being revlog based | |
512 | # hence accessing private attributes can be justified |
|
512 | # hence accessing private attributes can be justified | |
513 | nodemap.persist_nodemap( |
|
513 | nodemap.persist_nodemap( | |
514 | tr, unfi.manifestlog._rootstore._revlog, force=True |
|
514 | tr, unfi.manifestlog._rootstore._revlog, force=True | |
515 | ) |
|
515 | ) | |
516 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
516 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
517 | elif ( |
|
517 | elif ( | |
518 | len(upgrade_op.removed_actions) == 1 |
|
518 | len(upgrade_op.removed_actions) == 1 | |
519 | and [ |
|
519 | and [ | |
520 | x |
|
520 | x | |
521 | for x in upgrade_op.removed_actions |
|
521 | for x in upgrade_op.removed_actions | |
522 | if x.name == b'persistent-nodemap' |
|
522 | if x.name == b'persistent-nodemap' | |
523 | ] |
|
523 | ] | |
524 | and not upgrade_op.upgrade_actions |
|
524 | and not upgrade_op.upgrade_actions | |
525 | ): |
|
525 | ): | |
526 | ui.status( |
|
526 | ui.status( | |
527 | _(b'downgrading repository to not use persistent nodemap feature\n') |
|
527 | _(b'downgrading repository to not use persistent nodemap feature\n') | |
528 | ) |
|
528 | ) | |
529 | with srcrepo.transaction(b'upgrade') as tr: |
|
529 | with srcrepo.transaction(b'upgrade') as tr: | |
530 | unfi = srcrepo.unfiltered() |
|
530 | unfi = srcrepo.unfiltered() | |
531 | cl = unfi.changelog |
|
531 | cl = unfi.changelog | |
532 | nodemap.delete_nodemap(tr, srcrepo, cl) |
|
532 | nodemap.delete_nodemap(tr, srcrepo, cl) | |
533 | # check comment 20 lines above for accessing private attributes |
|
533 | # check comment 20 lines above for accessing private attributes | |
534 | nodemap.delete_nodemap( |
|
534 | nodemap.delete_nodemap( | |
535 | tr, srcrepo, unfi.manifestlog._rootstore._revlog |
|
535 | tr, srcrepo, unfi.manifestlog._rootstore._revlog | |
536 | ) |
|
536 | ) | |
537 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
537 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
538 | else: |
|
538 | else: | |
539 | with dstrepo.transaction(b'upgrade') as tr: |
|
539 | with dstrepo.transaction(b'upgrade') as tr: | |
540 | _clonerevlogs( |
|
540 | _clonerevlogs( | |
541 | ui, |
|
541 | ui, | |
542 | srcrepo, |
|
542 | srcrepo, | |
543 | dstrepo, |
|
543 | dstrepo, | |
544 | tr, |
|
544 | tr, | |
545 | upgrade_op, |
|
545 | upgrade_op, | |
546 | ) |
|
546 | ) | |
547 |
|
547 | |||
548 | # Now copy other files in the store directory. |
|
548 | # Now copy other files in the store directory. | |
549 | for p in _files_to_copy_post_revlog_clone(srcrepo): |
|
549 | for p in _files_to_copy_post_revlog_clone(srcrepo): | |
550 | srcrepo.ui.status(_(b'copying %s\n') % p) |
|
550 | srcrepo.ui.status(_(b'copying %s\n') % p) | |
551 | src = srcrepo.store.rawvfs.join(p) |
|
551 | src = srcrepo.store.rawvfs.join(p) | |
552 | dst = dstrepo.store.rawvfs.join(p) |
|
552 | dst = dstrepo.store.rawvfs.join(p) | |
553 | util.copyfile(src, dst, copystat=True) |
|
553 | util.copyfile(src, dst, copystat=True) | |
554 |
|
554 | |||
555 | finishdatamigration(ui, srcrepo, dstrepo, requirements) |
|
555 | finishdatamigration(ui, srcrepo, dstrepo, requirements) | |
556 |
|
556 | |||
557 | ui.status(_(b'data fully upgraded in a temporary repository\n')) |
|
557 | ui.status(_(b'data fully upgraded in a temporary repository\n')) | |
558 |
|
558 | |||
559 | if upgrade_op.backup_store: |
|
559 | if upgrade_op.backup_store: | |
560 | backuppath = pycompat.mkdtemp( |
|
560 | backuppath = pycompat.mkdtemp( | |
561 | prefix=b'upgradebackup.', dir=srcrepo.path |
|
561 | prefix=b'upgradebackup.', dir=srcrepo.path | |
562 | ) |
|
562 | ) | |
563 | backupvfs = vfsmod.vfs(backuppath) |
|
563 | backupvfs = vfsmod.vfs(backuppath) | |
564 |
|
564 | |||
565 | # Make a backup of requires file first, as it is the first to be modified. |
|
565 | # Make a backup of requires file first, as it is the first to be modified. | |
566 | util.copyfile( |
|
566 | util.copyfile( | |
567 | srcrepo.vfs.join(b'requires'), backupvfs.join(b'requires') |
|
567 | srcrepo.vfs.join(b'requires'), backupvfs.join(b'requires') | |
568 | ) |
|
568 | ) | |
569 |
|
569 | |||
570 | # We install an arbitrary requirement that clients must not support |
|
570 | # We install an arbitrary requirement that clients must not support | |
571 | # as a mechanism to lock out new clients during the data swap. This is |
|
571 | # as a mechanism to lock out new clients during the data swap. This is | |
572 | # better than allowing a client to continue while the repository is in |
|
572 | # better than allowing a client to continue while the repository is in | |
573 | # an inconsistent state. |
|
573 | # an inconsistent state. | |
574 | ui.status( |
|
574 | ui.status( | |
575 | _( |
|
575 | _( | |
576 | b'marking source repository as being upgraded; clients will be ' |
|
576 | b'marking source repository as being upgraded; clients will be ' | |
577 | b'unable to read from repository\n' |
|
577 | b'unable to read from repository\n' | |
578 | ) |
|
578 | ) | |
579 | ) |
|
579 | ) | |
580 | scmutil.writereporequirements( |
|
580 | scmutil.writereporequirements( | |
581 | srcrepo, srcrepo.requirements | {b'upgradeinprogress'} |
|
581 | srcrepo, srcrepo.requirements | {b'upgradeinprogress'} | |
582 | ) |
|
582 | ) | |
583 |
|
583 | |||
584 | ui.status(_(b'starting in-place swap of repository data\n')) |
|
584 | ui.status(_(b'starting in-place swap of repository data\n')) | |
585 | if upgrade_op.backup_store: |
|
585 | if upgrade_op.backup_store: | |
586 | ui.status( |
|
586 | ui.status( | |
587 | _(b'replaced files will be backed up at %s\n') % backuppath |
|
587 | _(b'replaced files will be backed up at %s\n') % backuppath | |
588 | ) |
|
588 | ) | |
589 |
|
589 | |||
590 | # Now swap in the new store directory. Doing it as a rename should make |
|
590 | # Now swap in the new store directory. Doing it as a rename should make | |
591 | # the operation nearly instantaneous and atomic (at least in well-behaved |
|
591 | # the operation nearly instantaneous and atomic (at least in well-behaved | |
592 | # environments). |
|
592 | # environments). | |
593 | ui.status(_(b'replacing store...\n')) |
|
593 | ui.status(_(b'replacing store...\n')) | |
594 | tstart = util.timer() |
|
594 | tstart = util.timer() | |
595 | _replacestores(srcrepo, dstrepo, backupvfs, upgrade_op) |
|
595 | _replacestores(srcrepo, dstrepo, backupvfs, upgrade_op) | |
596 | elapsed = util.timer() - tstart |
|
596 | elapsed = util.timer() - tstart | |
597 | ui.status( |
|
597 | ui.status( | |
598 | _( |
|
598 | _( | |
599 | b'store replacement complete; repository was inconsistent for ' |
|
599 | b'store replacement complete; repository was inconsistent for ' | |
600 | b'%0.1fs\n' |
|
600 | b'%0.1fs\n' | |
601 | ) |
|
601 | ) | |
602 | % elapsed |
|
602 | % elapsed | |
603 | ) |
|
603 | ) | |
604 |
|
604 | |||
605 | # We first write the requirements file. Any new requirements will lock |
|
605 | # We first write the requirements file. Any new requirements will lock | |
606 | # out legacy clients. |
|
606 | # out legacy clients. | |
607 | ui.status( |
|
607 | ui.status( | |
608 | _( |
|
608 | _( | |
609 | b'finalizing requirements file and making repository readable ' |
|
609 | b'finalizing requirements file and making repository readable ' | |
610 | b'again\n' |
|
610 | b'again\n' | |
611 | ) |
|
611 | ) | |
612 | ) |
|
612 | ) | |
613 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
613 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
614 |
|
614 | |||
615 | if upgrade_op.backup_store: |
|
615 | if upgrade_op.backup_store: | |
616 | # The lock file from the old store won't be removed because nothing has a |
|
616 | # The lock file from the old store won't be removed because nothing has a | |
617 | # reference to its new location. So clean it up manually. Alternatively, we |
|
617 | # reference to its new location. So clean it up manually. Alternatively, we | |
618 | # could update srcrepo.svfs and other variables to point to the new |
|
618 | # could update srcrepo.svfs and other variables to point to the new | |
619 | # location. This is simpler. |
|
619 | # location. This is simpler. | |
620 | assert backupvfs is not None # help pytype |
|
620 | assert backupvfs is not None # help pytype | |
621 | backupvfs.unlink(b'store/lock') |
|
621 | backupvfs.unlink(b'store/lock') | |
622 |
|
622 | |||
623 | return backuppath |
|
623 | return backuppath | |
624 |
|
624 | |||
625 |
|
625 | |||
626 | def upgrade_dirstate(ui, srcrepo, upgrade_op, old, new): |
|
626 | def upgrade_dirstate(ui, srcrepo, upgrade_op, old, new): | |
627 | if upgrade_op.backup_store: |
|
627 | if upgrade_op.backup_store: | |
628 | backuppath = pycompat.mkdtemp( |
|
628 | backuppath = pycompat.mkdtemp( | |
629 | prefix=b'upgradebackup.', dir=srcrepo.path |
|
629 | prefix=b'upgradebackup.', dir=srcrepo.path | |
630 | ) |
|
630 | ) | |
631 | ui.status(_(b'replaced files will be backed up at %s\n') % backuppath) |
|
631 | ui.status(_(b'replaced files will be backed up at %s\n') % backuppath) | |
632 | backupvfs = vfsmod.vfs(backuppath) |
|
632 | backupvfs = vfsmod.vfs(backuppath) | |
633 | util.copyfile( |
|
633 | util.copyfile( | |
634 | srcrepo.vfs.join(b'requires'), backupvfs.join(b'requires') |
|
634 | srcrepo.vfs.join(b'requires'), backupvfs.join(b'requires') | |
635 | ) |
|
635 | ) | |
636 | util.copyfile( |
|
636 | util.copyfile( | |
637 | srcrepo.vfs.join(b'dirstate'), backupvfs.join(b'dirstate') |
|
637 | srcrepo.vfs.join(b'dirstate'), backupvfs.join(b'dirstate') | |
638 | ) |
|
638 | ) | |
639 |
|
639 | |||
640 | assert srcrepo.dirstate._use_dirstate_v2 == (old == b'v2') |
|
640 | assert srcrepo.dirstate._use_dirstate_v2 == (old == b'v2') | |
641 | srcrepo.dirstate._map._use_dirstate_tree = True |
|
641 | srcrepo.dirstate._map._use_dirstate_tree = True | |
642 | srcrepo.dirstate._map.preload() |
|
642 | srcrepo.dirstate._map.preload() | |
643 | srcrepo.dirstate._use_dirstate_v2 = new == b'v2' |
|
643 | srcrepo.dirstate._use_dirstate_v2 = new == b'v2' | |
644 | srcrepo.dirstate._map._use_dirstate_v2 = srcrepo.dirstate._use_dirstate_v2 |
|
644 | srcrepo.dirstate._map._use_dirstate_v2 = srcrepo.dirstate._use_dirstate_v2 | |
645 | srcrepo.dirstate._dirty = True |
|
645 | srcrepo.dirstate._dirty = True | |
646 | srcrepo.vfs.unlink(b'dirstate') |
|
646 | srcrepo.vfs.unlink(b'dirstate') | |
647 | srcrepo.dirstate.write(None) |
|
647 | srcrepo.dirstate.write(None) | |
648 |
|
648 | |||
649 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
649 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
@@ -1,586 +1,588 b'' | |||||
1 | # verify.py - repository integrity checking for Mercurial |
|
1 | # verify.py - repository integrity checking for Mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2006, 2007 Olivia Mackall <olivia@selenic.com> |
|
3 | # Copyright 2006, 2007 Olivia Mackall <olivia@selenic.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 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import os |
|
10 | import os | |
11 |
|
11 | |||
12 | from .i18n import _ |
|
12 | from .i18n import _ | |
13 | from .node import short |
|
13 | from .node import short | |
14 | from .utils import stringutil |
|
14 | from .utils import stringutil | |
15 |
|
15 | |||
16 | from . import ( |
|
16 | from . import ( | |
17 | error, |
|
17 | error, | |
18 | pycompat, |
|
18 | pycompat, | |
19 | revlog, |
|
19 | revlog, | |
20 | util, |
|
20 | util, | |
21 | ) |
|
21 | ) | |
22 |
|
22 | |||
23 | VERIFY_DEFAULT = 0 |
|
23 | VERIFY_DEFAULT = 0 | |
24 | VERIFY_FULL = 1 |
|
24 | VERIFY_FULL = 1 | |
25 |
|
25 | |||
26 |
|
26 | |||
27 | def verify(repo, level=None): |
|
27 | def verify(repo, level=None): | |
28 | with repo.lock(): |
|
28 | with repo.lock(): | |
29 | v = verifier(repo, level) |
|
29 | v = verifier(repo, level) | |
30 | return v.verify() |
|
30 | return v.verify() | |
31 |
|
31 | |||
32 |
|
32 | |||
33 | def _normpath(f): |
|
33 | def _normpath(f): | |
34 | # under hg < 2.4, convert didn't sanitize paths properly, so a |
|
34 | # under hg < 2.4, convert didn't sanitize paths properly, so a | |
35 | # converted repo may contain repeated slashes |
|
35 | # converted repo may contain repeated slashes | |
36 | while b'//' in f: |
|
36 | while b'//' in f: | |
37 | f = f.replace(b'//', b'/') |
|
37 | f = f.replace(b'//', b'/') | |
38 | return f |
|
38 | return f | |
39 |
|
39 | |||
40 |
|
40 | |||
41 | HINT_FNCACHE = _( |
|
41 | HINT_FNCACHE = _( | |
42 | b'hint: run "hg debugrebuildfncache" to recover from corrupt fncache\n' |
|
42 | b'hint: run "hg debugrebuildfncache" to recover from corrupt fncache\n' | |
43 | ) |
|
43 | ) | |
44 |
|
44 | |||
45 | WARN_PARENT_DIR_UNKNOWN_REV = _( |
|
45 | WARN_PARENT_DIR_UNKNOWN_REV = _( | |
46 | b"parent-directory manifest refers to unknown revision %s" |
|
46 | b"parent-directory manifest refers to unknown revision %s" | |
47 | ) |
|
47 | ) | |
48 |
|
48 | |||
49 | WARN_UNKNOWN_COPY_SOURCE = _( |
|
49 | WARN_UNKNOWN_COPY_SOURCE = _( | |
50 | b"warning: copy source of '%s' not in parents of %s" |
|
50 | b"warning: copy source of '%s' not in parents of %s" | |
51 | ) |
|
51 | ) | |
52 |
|
52 | |||
53 | WARN_NULLID_COPY_SOURCE = _( |
|
53 | WARN_NULLID_COPY_SOURCE = _( | |
54 | b"warning: %s@%s: copy source revision is nullid %s:%s\n" |
|
54 | b"warning: %s@%s: copy source revision is nullid %s:%s\n" | |
55 | ) |
|
55 | ) | |
56 |
|
56 | |||
57 |
|
57 | |||
58 | class verifier(object): |
|
58 | class verifier(object): | |
59 | def __init__(self, repo, level=None): |
|
59 | def __init__(self, repo, level=None): | |
60 | self.repo = repo.unfiltered() |
|
60 | self.repo = repo.unfiltered() | |
61 | self.ui = repo.ui |
|
61 | self.ui = repo.ui | |
62 | self.match = repo.narrowmatch() |
|
62 | self.match = repo.narrowmatch() | |
63 | if level is None: |
|
63 | if level is None: | |
64 | level = VERIFY_DEFAULT |
|
64 | level = VERIFY_DEFAULT | |
65 | self._level = level |
|
65 | self._level = level | |
66 | self.badrevs = set() |
|
66 | self.badrevs = set() | |
67 | self.errors = 0 |
|
67 | self.errors = 0 | |
68 | self.warnings = 0 |
|
68 | self.warnings = 0 | |
69 | self.havecl = len(repo.changelog) > 0 |
|
69 | self.havecl = len(repo.changelog) > 0 | |
70 | self.havemf = len(repo.manifestlog.getstorage(b'')) > 0 |
|
70 | self.havemf = len(repo.manifestlog.getstorage(b'')) > 0 | |
71 | self.revlogv1 = repo.changelog._format_version != revlog.REVLOGV0 |
|
71 | self.revlogv1 = repo.changelog._format_version != revlog.REVLOGV0 | |
72 | self.lrugetctx = util.lrucachefunc(repo.unfiltered().__getitem__) |
|
72 | self.lrugetctx = util.lrucachefunc(repo.unfiltered().__getitem__) | |
73 | self.refersmf = False |
|
73 | self.refersmf = False | |
74 | self.fncachewarned = False |
|
74 | self.fncachewarned = False | |
75 | # developer config: verify.skipflags |
|
75 | # developer config: verify.skipflags | |
76 | self.skipflags = repo.ui.configint(b'verify', b'skipflags') |
|
76 | self.skipflags = repo.ui.configint(b'verify', b'skipflags') | |
77 | self.warnorphanstorefiles = True |
|
77 | self.warnorphanstorefiles = True | |
78 |
|
78 | |||
79 | def _warn(self, msg): |
|
79 | def _warn(self, msg): | |
80 | """record a "warning" level issue""" |
|
80 | """record a "warning" level issue""" | |
81 | self.ui.warn(msg + b"\n") |
|
81 | self.ui.warn(msg + b"\n") | |
82 | self.warnings += 1 |
|
82 | self.warnings += 1 | |
83 |
|
83 | |||
84 | def _err(self, linkrev, msg, filename=None): |
|
84 | def _err(self, linkrev, msg, filename=None): | |
85 | """record a "error" level issue""" |
|
85 | """record a "error" level issue""" | |
86 | if linkrev is not None: |
|
86 | if linkrev is not None: | |
87 | self.badrevs.add(linkrev) |
|
87 | self.badrevs.add(linkrev) | |
88 | linkrev = b"%d" % linkrev |
|
88 | linkrev = b"%d" % linkrev | |
89 | else: |
|
89 | else: | |
90 | linkrev = b'?' |
|
90 | linkrev = b'?' | |
91 | msg = b"%s: %s" % (linkrev, msg) |
|
91 | msg = b"%s: %s" % (linkrev, msg) | |
92 | if filename: |
|
92 | if filename: | |
93 | msg = b"%s@%s" % (filename, msg) |
|
93 | msg = b"%s@%s" % (filename, msg) | |
94 | self.ui.warn(b" " + msg + b"\n") |
|
94 | self.ui.warn(b" " + msg + b"\n") | |
95 | self.errors += 1 |
|
95 | self.errors += 1 | |
96 |
|
96 | |||
97 | def _exc(self, linkrev, msg, inst, filename=None): |
|
97 | def _exc(self, linkrev, msg, inst, filename=None): | |
98 | """record exception raised during the verify process""" |
|
98 | """record exception raised during the verify process""" | |
99 | fmsg = stringutil.forcebytestr(inst) |
|
99 | fmsg = stringutil.forcebytestr(inst) | |
100 | if not fmsg: |
|
100 | if not fmsg: | |
101 | fmsg = pycompat.byterepr(inst) |
|
101 | fmsg = pycompat.byterepr(inst) | |
102 | self._err(linkrev, b"%s: %s" % (msg, fmsg), filename) |
|
102 | self._err(linkrev, b"%s: %s" % (msg, fmsg), filename) | |
103 |
|
103 | |||
104 | def _checkrevlog(self, obj, name, linkrev): |
|
104 | def _checkrevlog(self, obj, name, linkrev): | |
105 | """verify high level property of a revlog |
|
105 | """verify high level property of a revlog | |
106 |
|
106 | |||
107 | - revlog is present, |
|
107 | - revlog is present, | |
108 | - revlog is non-empty, |
|
108 | - revlog is non-empty, | |
109 | - sizes (index and data) are correct, |
|
109 | - sizes (index and data) are correct, | |
110 | - revlog's format version is correct. |
|
110 | - revlog's format version is correct. | |
111 | """ |
|
111 | """ | |
112 | if not len(obj) and (self.havecl or self.havemf): |
|
112 | if not len(obj) and (self.havecl or self.havemf): | |
113 | self._err(linkrev, _(b"empty or missing %s") % name) |
|
113 | self._err(linkrev, _(b"empty or missing %s") % name) | |
114 | return |
|
114 | return | |
115 |
|
115 | |||
116 | d = obj.checksize() |
|
116 | d = obj.checksize() | |
117 | if d[0]: |
|
117 | if d[0]: | |
118 | self._err(None, _(b"data length off by %d bytes") % d[0], name) |
|
118 | self._err(None, _(b"data length off by %d bytes") % d[0], name) | |
119 | if d[1]: |
|
119 | if d[1]: | |
120 | self._err(None, _(b"index contains %d extra bytes") % d[1], name) |
|
120 | self._err(None, _(b"index contains %d extra bytes") % d[1], name) | |
121 |
|
121 | |||
122 | if obj._format_version != revlog.REVLOGV0: |
|
122 | if obj._format_version != revlog.REVLOGV0: | |
123 | if not self.revlogv1: |
|
123 | if not self.revlogv1: | |
124 | self._warn(_(b"warning: `%s' uses revlog format 1") % name) |
|
124 | self._warn(_(b"warning: `%s' uses revlog format 1") % name) | |
125 | elif self.revlogv1: |
|
125 | elif self.revlogv1: | |
126 | self._warn(_(b"warning: `%s' uses revlog format 0") % name) |
|
126 | self._warn(_(b"warning: `%s' uses revlog format 0") % name) | |
127 |
|
127 | |||
128 | def _checkentry(self, obj, i, node, seen, linkrevs, f): |
|
128 | def _checkentry(self, obj, i, node, seen, linkrevs, f): | |
129 | """verify a single revlog entry |
|
129 | """verify a single revlog entry | |
130 |
|
130 | |||
131 | arguments are: |
|
131 | arguments are: | |
132 | - obj: the source revlog |
|
132 | - obj: the source revlog | |
133 | - i: the revision number |
|
133 | - i: the revision number | |
134 | - node: the revision node id |
|
134 | - node: the revision node id | |
135 | - seen: nodes previously seen for this revlog |
|
135 | - seen: nodes previously seen for this revlog | |
136 | - linkrevs: [changelog-revisions] introducing "node" |
|
136 | - linkrevs: [changelog-revisions] introducing "node" | |
137 | - f: string label ("changelog", "manifest", or filename) |
|
137 | - f: string label ("changelog", "manifest", or filename) | |
138 |
|
138 | |||
139 | Performs the following checks: |
|
139 | Performs the following checks: | |
140 | - linkrev points to an existing changelog revision, |
|
140 | - linkrev points to an existing changelog revision, | |
141 | - linkrev points to a changelog revision that introduces this revision, |
|
141 | - linkrev points to a changelog revision that introduces this revision, | |
142 | - linkrev points to the lowest of these changesets, |
|
142 | - linkrev points to the lowest of these changesets, | |
143 | - both parents exist in the revlog, |
|
143 | - both parents exist in the revlog, | |
144 | - the revision is not duplicated. |
|
144 | - the revision is not duplicated. | |
145 |
|
145 | |||
146 | Return the linkrev of the revision (or None for changelog's revisions). |
|
146 | Return the linkrev of the revision (or None for changelog's revisions). | |
147 | """ |
|
147 | """ | |
148 | lr = obj.linkrev(obj.rev(node)) |
|
148 | lr = obj.linkrev(obj.rev(node)) | |
149 | if lr < 0 or (self.havecl and lr not in linkrevs): |
|
149 | if lr < 0 or (self.havecl and lr not in linkrevs): | |
150 | if lr < 0 or lr >= len(self.repo.changelog): |
|
150 | if lr < 0 or lr >= len(self.repo.changelog): | |
151 | msg = _(b"rev %d points to nonexistent changeset %d") |
|
151 | msg = _(b"rev %d points to nonexistent changeset %d") | |
152 | else: |
|
152 | else: | |
153 | msg = _(b"rev %d points to unexpected changeset %d") |
|
153 | msg = _(b"rev %d points to unexpected changeset %d") | |
154 | self._err(None, msg % (i, lr), f) |
|
154 | self._err(None, msg % (i, lr), f) | |
155 | if linkrevs: |
|
155 | if linkrevs: | |
156 | if f and len(linkrevs) > 1: |
|
156 | if f and len(linkrevs) > 1: | |
157 | try: |
|
157 | try: | |
158 | # attempt to filter down to real linkrevs |
|
158 | # attempt to filter down to real linkrevs | |
159 | linkrevs = [] |
|
159 | linkrevs = [] | |
160 | for lr in linkrevs: |
|
160 | for lr in linkrevs: | |
161 | if self.lrugetctx(lr)[f].filenode() == node: |
|
161 | if self.lrugetctx(lr)[f].filenode() == node: | |
162 | linkrevs.append(lr) |
|
162 | linkrevs.append(lr) | |
163 | except Exception: |
|
163 | except Exception: | |
164 | pass |
|
164 | pass | |
165 | msg = _(b" (expected %s)") |
|
165 | msg = _(b" (expected %s)") | |
166 | msg %= b" ".join(map(pycompat.bytestr, linkrevs)) |
|
166 | msg %= b" ".join(map(pycompat.bytestr, linkrevs)) | |
167 | self._warn(msg) |
|
167 | self._warn(msg) | |
168 | lr = None # can't be trusted |
|
168 | lr = None # can't be trusted | |
169 |
|
169 | |||
170 | try: |
|
170 | try: | |
171 | p1, p2 = obj.parents(node) |
|
171 | p1, p2 = obj.parents(node) | |
172 | if p1 not in seen and p1 != self.repo.nullid: |
|
172 | if p1 not in seen and p1 != self.repo.nullid: | |
173 | msg = _(b"unknown parent 1 %s of %s") % (short(p1), short(node)) |
|
173 | msg = _(b"unknown parent 1 %s of %s") % (short(p1), short(node)) | |
174 | self._err(lr, msg, f) |
|
174 | self._err(lr, msg, f) | |
175 | if p2 not in seen and p2 != self.repo.nullid: |
|
175 | if p2 not in seen and p2 != self.repo.nullid: | |
176 | msg = _(b"unknown parent 2 %s of %s") % (short(p2), short(node)) |
|
176 | msg = _(b"unknown parent 2 %s of %s") % (short(p2), short(node)) | |
177 | self._err(lr, msg, f) |
|
177 | self._err(lr, msg, f) | |
178 | except Exception as inst: |
|
178 | except Exception as inst: | |
179 | self._exc(lr, _(b"checking parents of %s") % short(node), inst, f) |
|
179 | self._exc(lr, _(b"checking parents of %s") % short(node), inst, f) | |
180 |
|
180 | |||
181 | if node in seen: |
|
181 | if node in seen: | |
182 | self._err(lr, _(b"duplicate revision %d (%d)") % (i, seen[node]), f) |
|
182 | self._err(lr, _(b"duplicate revision %d (%d)") % (i, seen[node]), f) | |
183 | seen[node] = i |
|
183 | seen[node] = i | |
184 | return lr |
|
184 | return lr | |
185 |
|
185 | |||
186 | def verify(self): |
|
186 | def verify(self): | |
187 | """verify the content of the Mercurial repository |
|
187 | """verify the content of the Mercurial repository | |
188 |
|
188 | |||
189 | This method run all verifications, displaying issues as they are found. |
|
189 | This method run all verifications, displaying issues as they are found. | |
190 |
|
190 | |||
191 | return 1 if any error have been encountered, 0 otherwise.""" |
|
191 | return 1 if any error have been encountered, 0 otherwise.""" | |
192 | # initial validation and generic report |
|
192 | # initial validation and generic report | |
193 | repo = self.repo |
|
193 | repo = self.repo | |
194 | ui = repo.ui |
|
194 | ui = repo.ui | |
195 | if not repo.url().startswith(b'file:'): |
|
195 | if not repo.url().startswith(b'file:'): | |
196 | raise error.Abort(_(b"cannot verify bundle or remote repos")) |
|
196 | raise error.Abort(_(b"cannot verify bundle or remote repos")) | |
197 |
|
197 | |||
198 | if os.path.exists(repo.sjoin(b"journal")): |
|
198 | if os.path.exists(repo.sjoin(b"journal")): | |
199 | ui.warn(_(b"abandoned transaction found - run hg recover\n")) |
|
199 | ui.warn(_(b"abandoned transaction found - run hg recover\n")) | |
200 |
|
200 | |||
201 | if ui.verbose or not self.revlogv1: |
|
201 | if ui.verbose or not self.revlogv1: | |
202 | ui.status( |
|
202 | ui.status( | |
203 | _(b"repository uses revlog format %d\n") |
|
203 | _(b"repository uses revlog format %d\n") | |
204 | % (self.revlogv1 and 1 or 0) |
|
204 | % (self.revlogv1 and 1 or 0) | |
205 | ) |
|
205 | ) | |
206 |
|
206 | |||
207 | # data verification |
|
207 | # data verification | |
208 | mflinkrevs, filelinkrevs = self._verifychangelog() |
|
208 | mflinkrevs, filelinkrevs = self._verifychangelog() | |
209 | filenodes = self._verifymanifest(mflinkrevs) |
|
209 | filenodes = self._verifymanifest(mflinkrevs) | |
210 | del mflinkrevs |
|
210 | del mflinkrevs | |
211 | self._crosscheckfiles(filelinkrevs, filenodes) |
|
211 | self._crosscheckfiles(filelinkrevs, filenodes) | |
212 | totalfiles, filerevisions = self._verifyfiles(filenodes, filelinkrevs) |
|
212 | totalfiles, filerevisions = self._verifyfiles(filenodes, filelinkrevs) | |
213 |
|
213 | |||
214 | # final report |
|
214 | # final report | |
215 | ui.status( |
|
215 | ui.status( | |
216 | _(b"checked %d changesets with %d changes to %d files\n") |
|
216 | _(b"checked %d changesets with %d changes to %d files\n") | |
217 | % (len(repo.changelog), filerevisions, totalfiles) |
|
217 | % (len(repo.changelog), filerevisions, totalfiles) | |
218 | ) |
|
218 | ) | |
219 | if self.warnings: |
|
219 | if self.warnings: | |
220 | ui.warn(_(b"%d warnings encountered!\n") % self.warnings) |
|
220 | ui.warn(_(b"%d warnings encountered!\n") % self.warnings) | |
221 | if self.fncachewarned: |
|
221 | if self.fncachewarned: | |
222 | ui.warn(HINT_FNCACHE) |
|
222 | ui.warn(HINT_FNCACHE) | |
223 | if self.errors: |
|
223 | if self.errors: | |
224 | ui.warn(_(b"%d integrity errors encountered!\n") % self.errors) |
|
224 | ui.warn(_(b"%d integrity errors encountered!\n") % self.errors) | |
225 | if self.badrevs: |
|
225 | if self.badrevs: | |
226 | msg = _(b"(first damaged changeset appears to be %d)\n") |
|
226 | msg = _(b"(first damaged changeset appears to be %d)\n") | |
227 | msg %= min(self.badrevs) |
|
227 | msg %= min(self.badrevs) | |
228 | ui.warn(msg) |
|
228 | ui.warn(msg) | |
229 | return 1 |
|
229 | return 1 | |
230 | return 0 |
|
230 | return 0 | |
231 |
|
231 | |||
232 | def _verifychangelog(self): |
|
232 | def _verifychangelog(self): | |
233 | """verify the changelog of a repository |
|
233 | """verify the changelog of a repository | |
234 |
|
234 | |||
235 | The following checks are performed: |
|
235 | The following checks are performed: | |
236 | - all of `_checkrevlog` checks, |
|
236 | - all of `_checkrevlog` checks, | |
237 | - all of `_checkentry` checks (for each revisions), |
|
237 | - all of `_checkentry` checks (for each revisions), | |
238 | - each revision can be read. |
|
238 | - each revision can be read. | |
239 |
|
239 | |||
240 | The function returns some of the data observed in the changesets as a |
|
240 | The function returns some of the data observed in the changesets as a | |
241 | (mflinkrevs, filelinkrevs) tuples: |
|
241 | (mflinkrevs, filelinkrevs) tuples: | |
242 | - mflinkrevs: is a { manifest-node -> [changelog-rev] } mapping |
|
242 | - mflinkrevs: is a { manifest-node -> [changelog-rev] } mapping | |
243 | - filelinkrevs: is a { file-path -> [changelog-rev] } mapping |
|
243 | - filelinkrevs: is a { file-path -> [changelog-rev] } mapping | |
244 |
|
244 | |||
245 | If a matcher was specified, filelinkrevs will only contains matched |
|
245 | If a matcher was specified, filelinkrevs will only contains matched | |
246 | files. |
|
246 | files. | |
247 | """ |
|
247 | """ | |
248 | ui = self.ui |
|
248 | ui = self.ui | |
249 | repo = self.repo |
|
249 | repo = self.repo | |
250 | match = self.match |
|
250 | match = self.match | |
251 | cl = repo.changelog |
|
251 | cl = repo.changelog | |
252 |
|
252 | |||
253 | ui.status(_(b"checking changesets\n")) |
|
253 | ui.status(_(b"checking changesets\n")) | |
254 | mflinkrevs = {} |
|
254 | mflinkrevs = {} | |
255 | filelinkrevs = {} |
|
255 | filelinkrevs = {} | |
256 | seen = {} |
|
256 | seen = {} | |
257 | self._checkrevlog(cl, b"changelog", 0) |
|
257 | self._checkrevlog(cl, b"changelog", 0) | |
258 | progress = ui.makeprogress( |
|
258 | progress = ui.makeprogress( | |
259 | _(b'checking'), unit=_(b'changesets'), total=len(repo) |
|
259 | _(b'checking'), unit=_(b'changesets'), total=len(repo) | |
260 | ) |
|
260 | ) | |
261 | for i in repo: |
|
261 | for i in repo: | |
262 | progress.update(i) |
|
262 | progress.update(i) | |
263 | n = cl.node(i) |
|
263 | n = cl.node(i) | |
264 | self._checkentry(cl, i, n, seen, [i], b"changelog") |
|
264 | self._checkentry(cl, i, n, seen, [i], b"changelog") | |
265 |
|
265 | |||
266 | try: |
|
266 | try: | |
267 | changes = cl.read(n) |
|
267 | changes = cl.read(n) | |
268 | if changes[0] != self.repo.nullid: |
|
268 | if changes[0] != self.repo.nullid: | |
269 | mflinkrevs.setdefault(changes[0], []).append(i) |
|
269 | mflinkrevs.setdefault(changes[0], []).append(i) | |
270 | self.refersmf = True |
|
270 | self.refersmf = True | |
271 | for f in changes[3]: |
|
271 | for f in changes[3]: | |
272 | if match(f): |
|
272 | if match(f): | |
273 | filelinkrevs.setdefault(_normpath(f), []).append(i) |
|
273 | filelinkrevs.setdefault(_normpath(f), []).append(i) | |
274 | except Exception as inst: |
|
274 | except Exception as inst: | |
275 | self.refersmf = True |
|
275 | self.refersmf = True | |
276 | self._exc(i, _(b"unpacking changeset %s") % short(n), inst) |
|
276 | self._exc(i, _(b"unpacking changeset %s") % short(n), inst) | |
277 | progress.complete() |
|
277 | progress.complete() | |
278 | return mflinkrevs, filelinkrevs |
|
278 | return mflinkrevs, filelinkrevs | |
279 |
|
279 | |||
280 | def _verifymanifest( |
|
280 | def _verifymanifest( | |
281 | self, mflinkrevs, dir=b"", storefiles=None, subdirprogress=None |
|
281 | self, mflinkrevs, dir=b"", storefiles=None, subdirprogress=None | |
282 | ): |
|
282 | ): | |
283 | """verify the manifestlog content |
|
283 | """verify the manifestlog content | |
284 |
|
284 | |||
285 | Inputs: |
|
285 | Inputs: | |
286 | - mflinkrevs: a {manifest-node -> [changelog-revisions]} mapping |
|
286 | - mflinkrevs: a {manifest-node -> [changelog-revisions]} mapping | |
287 | - dir: a subdirectory to check (for tree manifest repo) |
|
287 | - dir: a subdirectory to check (for tree manifest repo) | |
288 | - storefiles: set of currently "orphan" files. |
|
288 | - storefiles: set of currently "orphan" files. | |
289 | - subdirprogress: a progress object |
|
289 | - subdirprogress: a progress object | |
290 |
|
290 | |||
291 | This function checks: |
|
291 | This function checks: | |
292 | * all of `_checkrevlog` checks (for all manifest related revlogs) |
|
292 | * all of `_checkrevlog` checks (for all manifest related revlogs) | |
293 | * all of `_checkentry` checks (for all manifest related revisions) |
|
293 | * all of `_checkentry` checks (for all manifest related revisions) | |
294 | * nodes for subdirectory exists in the sub-directory manifest |
|
294 | * nodes for subdirectory exists in the sub-directory manifest | |
295 | * each manifest entries have a file path |
|
295 | * each manifest entries have a file path | |
296 | * each manifest node refered in mflinkrevs exist in the manifest log |
|
296 | * each manifest node refered in mflinkrevs exist in the manifest log | |
297 |
|
297 | |||
298 | If tree manifest is in use and a matchers is specified, only the |
|
298 | If tree manifest is in use and a matchers is specified, only the | |
299 | sub-directories matching it will be verified. |
|
299 | sub-directories matching it will be verified. | |
300 |
|
300 | |||
301 | return a two level mapping: |
|
301 | return a two level mapping: | |
302 | {"path" -> { filenode -> changelog-revision}} |
|
302 | {"path" -> { filenode -> changelog-revision}} | |
303 |
|
303 | |||
304 | This mapping primarily contains entries for every files in the |
|
304 | This mapping primarily contains entries for every files in the | |
305 | repository. In addition, when tree-manifest is used, it also contains |
|
305 | repository. In addition, when tree-manifest is used, it also contains | |
306 | sub-directory entries. |
|
306 | sub-directory entries. | |
307 |
|
307 | |||
308 | If a matcher is provided, only matching paths will be included. |
|
308 | If a matcher is provided, only matching paths will be included. | |
309 | """ |
|
309 | """ | |
310 | repo = self.repo |
|
310 | repo = self.repo | |
311 | ui = self.ui |
|
311 | ui = self.ui | |
312 | match = self.match |
|
312 | match = self.match | |
313 | mfl = self.repo.manifestlog |
|
313 | mfl = self.repo.manifestlog | |
314 | mf = mfl.getstorage(dir) |
|
314 | mf = mfl.getstorage(dir) | |
315 |
|
315 | |||
316 | if not dir: |
|
316 | if not dir: | |
317 | self.ui.status(_(b"checking manifests\n")) |
|
317 | self.ui.status(_(b"checking manifests\n")) | |
318 |
|
318 | |||
319 | filenodes = {} |
|
319 | filenodes = {} | |
320 | subdirnodes = {} |
|
320 | subdirnodes = {} | |
321 | seen = {} |
|
321 | seen = {} | |
322 | label = b"manifest" |
|
322 | label = b"manifest" | |
323 | if dir: |
|
323 | if dir: | |
324 | label = dir |
|
324 | label = dir | |
325 | revlogfiles = mf.files() |
|
325 | revlogfiles = mf.files() | |
326 | storefiles.difference_update(revlogfiles) |
|
326 | storefiles.difference_update(revlogfiles) | |
327 | if subdirprogress: # should be true since we're in a subdirectory |
|
327 | if subdirprogress: # should be true since we're in a subdirectory | |
328 | subdirprogress.increment() |
|
328 | subdirprogress.increment() | |
329 | if self.refersmf: |
|
329 | if self.refersmf: | |
330 | # Do not check manifest if there are only changelog entries with |
|
330 | # Do not check manifest if there are only changelog entries with | |
331 | # null manifests. |
|
331 | # null manifests. | |
332 | self._checkrevlog(mf._revlog, label, 0) |
|
332 | self._checkrevlog(mf._revlog, label, 0) | |
333 | progress = ui.makeprogress( |
|
333 | progress = ui.makeprogress( | |
334 | _(b'checking'), unit=_(b'manifests'), total=len(mf) |
|
334 | _(b'checking'), unit=_(b'manifests'), total=len(mf) | |
335 | ) |
|
335 | ) | |
336 | for i in mf: |
|
336 | for i in mf: | |
337 | if not dir: |
|
337 | if not dir: | |
338 | progress.update(i) |
|
338 | progress.update(i) | |
339 | n = mf.node(i) |
|
339 | n = mf.node(i) | |
340 | lr = self._checkentry(mf, i, n, seen, mflinkrevs.get(n, []), label) |
|
340 | lr = self._checkentry(mf, i, n, seen, mflinkrevs.get(n, []), label) | |
341 | if n in mflinkrevs: |
|
341 | if n in mflinkrevs: | |
342 | del mflinkrevs[n] |
|
342 | del mflinkrevs[n] | |
343 | elif dir: |
|
343 | elif dir: | |
344 | msg = _(b"%s not in parent-directory manifest") % short(n) |
|
344 | msg = _(b"%s not in parent-directory manifest") % short(n) | |
345 | self._err(lr, msg, label) |
|
345 | self._err(lr, msg, label) | |
346 | else: |
|
346 | else: | |
347 | self._err(lr, _(b"%s not in changesets") % short(n), label) |
|
347 | self._err(lr, _(b"%s not in changesets") % short(n), label) | |
348 |
|
348 | |||
349 | try: |
|
349 | try: | |
350 | mfdelta = mfl.get(dir, n).readdelta(shallow=True) |
|
350 | mfdelta = mfl.get(dir, n).readdelta(shallow=True) | |
351 | for f, fn, fl in mfdelta.iterentries(): |
|
351 | for f, fn, fl in mfdelta.iterentries(): | |
352 | if not f: |
|
352 | if not f: | |
353 | self._err(lr, _(b"entry without name in manifest")) |
|
353 | self._err(lr, _(b"entry without name in manifest")) | |
354 | elif f == b"/dev/null": # ignore this in very old repos |
|
354 | elif f == b"/dev/null": # ignore this in very old repos | |
355 | continue |
|
355 | continue | |
356 | fullpath = dir + _normpath(f) |
|
356 | fullpath = dir + _normpath(f) | |
357 | if fl == b't': |
|
357 | if fl == b't': | |
358 | if not match.visitdir(fullpath): |
|
358 | if not match.visitdir(fullpath): | |
359 | continue |
|
359 | continue | |
360 | sdn = subdirnodes.setdefault(fullpath + b'/', {}) |
|
360 | sdn = subdirnodes.setdefault(fullpath + b'/', {}) | |
361 | sdn.setdefault(fn, []).append(lr) |
|
361 | sdn.setdefault(fn, []).append(lr) | |
362 | else: |
|
362 | else: | |
363 | if not match(fullpath): |
|
363 | if not match(fullpath): | |
364 | continue |
|
364 | continue | |
365 | filenodes.setdefault(fullpath, {}).setdefault(fn, lr) |
|
365 | filenodes.setdefault(fullpath, {}).setdefault(fn, lr) | |
366 | except Exception as inst: |
|
366 | except Exception as inst: | |
367 | self._exc(lr, _(b"reading delta %s") % short(n), inst, label) |
|
367 | self._exc(lr, _(b"reading delta %s") % short(n), inst, label) | |
368 | if self._level >= VERIFY_FULL: |
|
368 | if self._level >= VERIFY_FULL: | |
369 | try: |
|
369 | try: | |
370 | # Various issues can affect manifest. So we read each full |
|
370 | # Various issues can affect manifest. So we read each full | |
371 | # text from storage. This triggers the checks from the core |
|
371 | # text from storage. This triggers the checks from the core | |
372 | # code (eg: hash verification, filename are ordered, etc.) |
|
372 | # code (eg: hash verification, filename are ordered, etc.) | |
373 | mfdelta = mfl.get(dir, n).read() |
|
373 | mfdelta = mfl.get(dir, n).read() | |
374 | except Exception as inst: |
|
374 | except Exception as inst: | |
375 | msg = _(b"reading full manifest %s") % short(n) |
|
375 | msg = _(b"reading full manifest %s") % short(n) | |
376 | self._exc(lr, msg, inst, label) |
|
376 | self._exc(lr, msg, inst, label) | |
377 |
|
377 | |||
378 | if not dir: |
|
378 | if not dir: | |
379 | progress.complete() |
|
379 | progress.complete() | |
380 |
|
380 | |||
381 | if self.havemf: |
|
381 | if self.havemf: | |
382 | # since we delete entry in `mflinkrevs` during iteration, any |
|
382 | # since we delete entry in `mflinkrevs` during iteration, any | |
383 | # remaining entries are "missing". We need to issue errors for them. |
|
383 | # remaining entries are "missing". We need to issue errors for them. | |
384 | changesetpairs = [(c, m) for m in mflinkrevs for c in mflinkrevs[m]] |
|
384 | changesetpairs = [(c, m) for m in mflinkrevs for c in mflinkrevs[m]] | |
385 | for c, m in sorted(changesetpairs): |
|
385 | for c, m in sorted(changesetpairs): | |
386 | if dir: |
|
386 | if dir: | |
387 | self._err(c, WARN_PARENT_DIR_UNKNOWN_REV % short(m), label) |
|
387 | self._err(c, WARN_PARENT_DIR_UNKNOWN_REV % short(m), label) | |
388 | else: |
|
388 | else: | |
389 | msg = _(b"changeset refers to unknown revision %s") |
|
389 | msg = _(b"changeset refers to unknown revision %s") | |
390 | msg %= short(m) |
|
390 | msg %= short(m) | |
391 | self._err(c, msg, label) |
|
391 | self._err(c, msg, label) | |
392 |
|
392 | |||
393 | if not dir and subdirnodes: |
|
393 | if not dir and subdirnodes: | |
394 | self.ui.status(_(b"checking directory manifests\n")) |
|
394 | self.ui.status(_(b"checking directory manifests\n")) | |
395 | storefiles = set() |
|
395 | storefiles = set() | |
396 | subdirs = set() |
|
396 | subdirs = set() | |
397 | revlogv1 = self.revlogv1 |
|
397 | revlogv1 = self.revlogv1 | |
398 | for t, f, f2, size in repo.store.datafiles(): |
|
398 | undecodable = [] | |
399 | if not f: |
|
399 | for t, f, size in repo.store.datafiles(undecodable=undecodable): | |
400 | self._err(None, _(b"cannot decode filename '%s'") % f2) |
|
400 | if (size > 0 or not revlogv1) and f.startswith(b'meta/'): | |
401 | elif (size > 0 or not revlogv1) and f.startswith(b'meta/'): |
|
|||
402 | storefiles.add(_normpath(f)) |
|
401 | storefiles.add(_normpath(f)) | |
403 | subdirs.add(os.path.dirname(f)) |
|
402 | subdirs.add(os.path.dirname(f)) | |
|
403 | for f in undecodable: | |||
|
404 | self._err(None, _(b"cannot decode filename '%s'") % f) | |||
404 | subdirprogress = ui.makeprogress( |
|
405 | subdirprogress = ui.makeprogress( | |
405 | _(b'checking'), unit=_(b'manifests'), total=len(subdirs) |
|
406 | _(b'checking'), unit=_(b'manifests'), total=len(subdirs) | |
406 | ) |
|
407 | ) | |
407 |
|
408 | |||
408 | for subdir, linkrevs in pycompat.iteritems(subdirnodes): |
|
409 | for subdir, linkrevs in pycompat.iteritems(subdirnodes): | |
409 | subdirfilenodes = self._verifymanifest( |
|
410 | subdirfilenodes = self._verifymanifest( | |
410 | linkrevs, subdir, storefiles, subdirprogress |
|
411 | linkrevs, subdir, storefiles, subdirprogress | |
411 | ) |
|
412 | ) | |
412 | for f, onefilenodes in pycompat.iteritems(subdirfilenodes): |
|
413 | for f, onefilenodes in pycompat.iteritems(subdirfilenodes): | |
413 | filenodes.setdefault(f, {}).update(onefilenodes) |
|
414 | filenodes.setdefault(f, {}).update(onefilenodes) | |
414 |
|
415 | |||
415 | if not dir and subdirnodes: |
|
416 | if not dir and subdirnodes: | |
416 | assert subdirprogress is not None # help pytype |
|
417 | assert subdirprogress is not None # help pytype | |
417 | subdirprogress.complete() |
|
418 | subdirprogress.complete() | |
418 | if self.warnorphanstorefiles: |
|
419 | if self.warnorphanstorefiles: | |
419 | for f in sorted(storefiles): |
|
420 | for f in sorted(storefiles): | |
420 | self._warn(_(b"warning: orphan data file '%s'") % f) |
|
421 | self._warn(_(b"warning: orphan data file '%s'") % f) | |
421 |
|
422 | |||
422 | return filenodes |
|
423 | return filenodes | |
423 |
|
424 | |||
424 | def _crosscheckfiles(self, filelinkrevs, filenodes): |
|
425 | def _crosscheckfiles(self, filelinkrevs, filenodes): | |
425 | repo = self.repo |
|
426 | repo = self.repo | |
426 | ui = self.ui |
|
427 | ui = self.ui | |
427 | ui.status(_(b"crosschecking files in changesets and manifests\n")) |
|
428 | ui.status(_(b"crosschecking files in changesets and manifests\n")) | |
428 |
|
429 | |||
429 | total = len(filelinkrevs) + len(filenodes) |
|
430 | total = len(filelinkrevs) + len(filenodes) | |
430 | progress = ui.makeprogress( |
|
431 | progress = ui.makeprogress( | |
431 | _(b'crosschecking'), unit=_(b'files'), total=total |
|
432 | _(b'crosschecking'), unit=_(b'files'), total=total | |
432 | ) |
|
433 | ) | |
433 | if self.havemf: |
|
434 | if self.havemf: | |
434 | for f in sorted(filelinkrevs): |
|
435 | for f in sorted(filelinkrevs): | |
435 | progress.increment() |
|
436 | progress.increment() | |
436 | if f not in filenodes: |
|
437 | if f not in filenodes: | |
437 | lr = filelinkrevs[f][0] |
|
438 | lr = filelinkrevs[f][0] | |
438 | self._err(lr, _(b"in changeset but not in manifest"), f) |
|
439 | self._err(lr, _(b"in changeset but not in manifest"), f) | |
439 |
|
440 | |||
440 | if self.havecl: |
|
441 | if self.havecl: | |
441 | for f in sorted(filenodes): |
|
442 | for f in sorted(filenodes): | |
442 | progress.increment() |
|
443 | progress.increment() | |
443 | if f not in filelinkrevs: |
|
444 | if f not in filelinkrevs: | |
444 | try: |
|
445 | try: | |
445 | fl = repo.file(f) |
|
446 | fl = repo.file(f) | |
446 | lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]]) |
|
447 | lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]]) | |
447 | except Exception: |
|
448 | except Exception: | |
448 | lr = None |
|
449 | lr = None | |
449 | self._err(lr, _(b"in manifest but not in changeset"), f) |
|
450 | self._err(lr, _(b"in manifest but not in changeset"), f) | |
450 |
|
451 | |||
451 | progress.complete() |
|
452 | progress.complete() | |
452 |
|
453 | |||
453 | def _verifyfiles(self, filenodes, filelinkrevs): |
|
454 | def _verifyfiles(self, filenodes, filelinkrevs): | |
454 | repo = self.repo |
|
455 | repo = self.repo | |
455 | ui = self.ui |
|
456 | ui = self.ui | |
456 | lrugetctx = self.lrugetctx |
|
457 | lrugetctx = self.lrugetctx | |
457 | revlogv1 = self.revlogv1 |
|
458 | revlogv1 = self.revlogv1 | |
458 | havemf = self.havemf |
|
459 | havemf = self.havemf | |
459 | ui.status(_(b"checking files\n")) |
|
460 | ui.status(_(b"checking files\n")) | |
460 |
|
461 | |||
461 | storefiles = set() |
|
462 | storefiles = set() | |
462 | for rl_type, f, f2, size in repo.store.datafiles(): |
|
463 | undecodable = [] | |
463 | if not f: |
|
464 | for t, f, size in repo.store.datafiles(undecodable=undecodable): | |
464 | self._err(None, _(b"cannot decode filename '%s'") % f2) |
|
465 | if (size > 0 or not revlogv1) and f.startswith(b'data/'): | |
465 | elif (size > 0 or not revlogv1) and f.startswith(b'data/'): |
|
|||
466 | storefiles.add(_normpath(f)) |
|
466 | storefiles.add(_normpath(f)) | |
|
467 | for f in undecodable: | |||
|
468 | self._err(None, _(b"cannot decode filename '%s'") % f) | |||
467 |
|
469 | |||
468 | state = { |
|
470 | state = { | |
469 | # TODO this assumes revlog storage for changelog. |
|
471 | # TODO this assumes revlog storage for changelog. | |
470 | b'expectedversion': self.repo.changelog._format_version, |
|
472 | b'expectedversion': self.repo.changelog._format_version, | |
471 | b'skipflags': self.skipflags, |
|
473 | b'skipflags': self.skipflags, | |
472 | # experimental config: censor.policy |
|
474 | # experimental config: censor.policy | |
473 | b'erroroncensored': ui.config(b'censor', b'policy') == b'abort', |
|
475 | b'erroroncensored': ui.config(b'censor', b'policy') == b'abort', | |
474 | } |
|
476 | } | |
475 |
|
477 | |||
476 | files = sorted(set(filenodes) | set(filelinkrevs)) |
|
478 | files = sorted(set(filenodes) | set(filelinkrevs)) | |
477 | revisions = 0 |
|
479 | revisions = 0 | |
478 | progress = ui.makeprogress( |
|
480 | progress = ui.makeprogress( | |
479 | _(b'checking'), unit=_(b'files'), total=len(files) |
|
481 | _(b'checking'), unit=_(b'files'), total=len(files) | |
480 | ) |
|
482 | ) | |
481 | for i, f in enumerate(files): |
|
483 | for i, f in enumerate(files): | |
482 | progress.update(i, item=f) |
|
484 | progress.update(i, item=f) | |
483 | try: |
|
485 | try: | |
484 | linkrevs = filelinkrevs[f] |
|
486 | linkrevs = filelinkrevs[f] | |
485 | except KeyError: |
|
487 | except KeyError: | |
486 | # in manifest but not in changelog |
|
488 | # in manifest but not in changelog | |
487 | linkrevs = [] |
|
489 | linkrevs = [] | |
488 |
|
490 | |||
489 | if linkrevs: |
|
491 | if linkrevs: | |
490 | lr = linkrevs[0] |
|
492 | lr = linkrevs[0] | |
491 | else: |
|
493 | else: | |
492 | lr = None |
|
494 | lr = None | |
493 |
|
495 | |||
494 | try: |
|
496 | try: | |
495 | fl = repo.file(f) |
|
497 | fl = repo.file(f) | |
496 | except error.StorageError as e: |
|
498 | except error.StorageError as e: | |
497 | self._err(lr, _(b"broken revlog! (%s)") % e, f) |
|
499 | self._err(lr, _(b"broken revlog! (%s)") % e, f) | |
498 | continue |
|
500 | continue | |
499 |
|
501 | |||
500 | for ff in fl.files(): |
|
502 | for ff in fl.files(): | |
501 | try: |
|
503 | try: | |
502 | storefiles.remove(ff) |
|
504 | storefiles.remove(ff) | |
503 | except KeyError: |
|
505 | except KeyError: | |
504 | if self.warnorphanstorefiles: |
|
506 | if self.warnorphanstorefiles: | |
505 | msg = _(b" warning: revlog '%s' not in fncache!") |
|
507 | msg = _(b" warning: revlog '%s' not in fncache!") | |
506 | self._warn(msg % ff) |
|
508 | self._warn(msg % ff) | |
507 | self.fncachewarned = True |
|
509 | self.fncachewarned = True | |
508 |
|
510 | |||
509 | if not len(fl) and (self.havecl or self.havemf): |
|
511 | if not len(fl) and (self.havecl or self.havemf): | |
510 | self._err(lr, _(b"empty or missing %s") % f) |
|
512 | self._err(lr, _(b"empty or missing %s") % f) | |
511 | else: |
|
513 | else: | |
512 | # Guard against implementations not setting this. |
|
514 | # Guard against implementations not setting this. | |
513 | state[b'skipread'] = set() |
|
515 | state[b'skipread'] = set() | |
514 | state[b'safe_renamed'] = set() |
|
516 | state[b'safe_renamed'] = set() | |
515 |
|
517 | |||
516 | for problem in fl.verifyintegrity(state): |
|
518 | for problem in fl.verifyintegrity(state): | |
517 | if problem.node is not None: |
|
519 | if problem.node is not None: | |
518 | linkrev = fl.linkrev(fl.rev(problem.node)) |
|
520 | linkrev = fl.linkrev(fl.rev(problem.node)) | |
519 | else: |
|
521 | else: | |
520 | linkrev = None |
|
522 | linkrev = None | |
521 |
|
523 | |||
522 | if problem.warning: |
|
524 | if problem.warning: | |
523 | self._warn(problem.warning) |
|
525 | self._warn(problem.warning) | |
524 | elif problem.error: |
|
526 | elif problem.error: | |
525 | linkrev_msg = linkrev if linkrev is not None else lr |
|
527 | linkrev_msg = linkrev if linkrev is not None else lr | |
526 | self._err(linkrev_msg, problem.error, f) |
|
528 | self._err(linkrev_msg, problem.error, f) | |
527 | else: |
|
529 | else: | |
528 | raise error.ProgrammingError( |
|
530 | raise error.ProgrammingError( | |
529 | b'problem instance does not set warning or error ' |
|
531 | b'problem instance does not set warning or error ' | |
530 | b'attribute: %s' % problem.msg |
|
532 | b'attribute: %s' % problem.msg | |
531 | ) |
|
533 | ) | |
532 |
|
534 | |||
533 | seen = {} |
|
535 | seen = {} | |
534 | for i in fl: |
|
536 | for i in fl: | |
535 | revisions += 1 |
|
537 | revisions += 1 | |
536 | n = fl.node(i) |
|
538 | n = fl.node(i) | |
537 | lr = self._checkentry(fl, i, n, seen, linkrevs, f) |
|
539 | lr = self._checkentry(fl, i, n, seen, linkrevs, f) | |
538 | if f in filenodes: |
|
540 | if f in filenodes: | |
539 | if havemf and n not in filenodes[f]: |
|
541 | if havemf and n not in filenodes[f]: | |
540 | self._err(lr, _(b"%s not in manifests") % (short(n)), f) |
|
542 | self._err(lr, _(b"%s not in manifests") % (short(n)), f) | |
541 | else: |
|
543 | else: | |
542 | del filenodes[f][n] |
|
544 | del filenodes[f][n] | |
543 |
|
545 | |||
544 | if n in state[b'skipread'] and n not in state[b'safe_renamed']: |
|
546 | if n in state[b'skipread'] and n not in state[b'safe_renamed']: | |
545 | continue |
|
547 | continue | |
546 |
|
548 | |||
547 | # check renames |
|
549 | # check renames | |
548 | try: |
|
550 | try: | |
549 | # This requires resolving fulltext (at least on revlogs, |
|
551 | # This requires resolving fulltext (at least on revlogs, | |
550 | # though not with LFS revisions). We may want |
|
552 | # though not with LFS revisions). We may want | |
551 | # ``verifyintegrity()`` to pass a set of nodes with |
|
553 | # ``verifyintegrity()`` to pass a set of nodes with | |
552 | # rename metadata as an optimization. |
|
554 | # rename metadata as an optimization. | |
553 | rp = fl.renamed(n) |
|
555 | rp = fl.renamed(n) | |
554 | if rp: |
|
556 | if rp: | |
555 | if lr is not None and ui.verbose: |
|
557 | if lr is not None and ui.verbose: | |
556 | ctx = lrugetctx(lr) |
|
558 | ctx = lrugetctx(lr) | |
557 | if not any(rp[0] in pctx for pctx in ctx.parents()): |
|
559 | if not any(rp[0] in pctx for pctx in ctx.parents()): | |
558 | self._warn(WARN_UNKNOWN_COPY_SOURCE % (f, ctx)) |
|
560 | self._warn(WARN_UNKNOWN_COPY_SOURCE % (f, ctx)) | |
559 | fl2 = repo.file(rp[0]) |
|
561 | fl2 = repo.file(rp[0]) | |
560 | if not len(fl2): |
|
562 | if not len(fl2): | |
561 | m = _(b"empty or missing copy source revlog %s:%s") |
|
563 | m = _(b"empty or missing copy source revlog %s:%s") | |
562 | self._err(lr, m % (rp[0], short(rp[1])), f) |
|
564 | self._err(lr, m % (rp[0], short(rp[1])), f) | |
563 | elif rp[1] == self.repo.nullid: |
|
565 | elif rp[1] == self.repo.nullid: | |
564 | msg = WARN_NULLID_COPY_SOURCE |
|
566 | msg = WARN_NULLID_COPY_SOURCE | |
565 | msg %= (f, lr, rp[0], short(rp[1])) |
|
567 | msg %= (f, lr, rp[0], short(rp[1])) | |
566 | ui.note(msg) |
|
568 | ui.note(msg) | |
567 | else: |
|
569 | else: | |
568 | fl2.rev(rp[1]) |
|
570 | fl2.rev(rp[1]) | |
569 | except Exception as inst: |
|
571 | except Exception as inst: | |
570 | self._exc( |
|
572 | self._exc( | |
571 | lr, _(b"checking rename of %s") % short(n), inst, f |
|
573 | lr, _(b"checking rename of %s") % short(n), inst, f | |
572 | ) |
|
574 | ) | |
573 |
|
575 | |||
574 | # cross-check |
|
576 | # cross-check | |
575 | if f in filenodes: |
|
577 | if f in filenodes: | |
576 | fns = [(v, k) for k, v in pycompat.iteritems(filenodes[f])] |
|
578 | fns = [(v, k) for k, v in pycompat.iteritems(filenodes[f])] | |
577 | for lr, node in sorted(fns): |
|
579 | for lr, node in sorted(fns): | |
578 | msg = _(b"manifest refers to unknown revision %s") |
|
580 | msg = _(b"manifest refers to unknown revision %s") | |
579 | self._err(lr, msg % short(node), f) |
|
581 | self._err(lr, msg % short(node), f) | |
580 | progress.complete() |
|
582 | progress.complete() | |
581 |
|
583 | |||
582 | if self.warnorphanstorefiles: |
|
584 | if self.warnorphanstorefiles: | |
583 | for f in sorted(storefiles): |
|
585 | for f in sorted(storefiles): | |
584 | self._warn(_(b"warning: orphan data file '%s'") % f) |
|
586 | self._warn(_(b"warning: orphan data file '%s'") % f) | |
585 |
|
587 | |||
586 | return len(files), revisions |
|
588 | return len(files), revisions |
@@ -1,1613 +1,1613 b'' | |||||
1 | # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net> |
|
1 | # Copyright 21 May 2005 - (c) 2005 Jake Edge <jake@edge2.net> | |
2 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> |
|
2 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> | |
3 | # |
|
3 | # | |
4 | # This software may be used and distributed according to the terms of the |
|
4 | # This software may be used and distributed according to the terms of the | |
5 | # GNU General Public License version 2 or any later version. |
|
5 | # GNU General Public License version 2 or any later version. | |
6 |
|
6 | |||
7 | from __future__ import absolute_import |
|
7 | from __future__ import absolute_import | |
8 |
|
8 | |||
9 | import collections |
|
9 | import collections | |
10 | import contextlib |
|
10 | import contextlib | |
11 |
|
11 | |||
12 | from .i18n import _ |
|
12 | from .i18n import _ | |
13 | from .node import hex |
|
13 | from .node import hex | |
14 | from . import ( |
|
14 | from . import ( | |
15 | discovery, |
|
15 | discovery, | |
16 | encoding, |
|
16 | encoding, | |
17 | error, |
|
17 | error, | |
18 | match as matchmod, |
|
18 | match as matchmod, | |
19 | narrowspec, |
|
19 | narrowspec, | |
20 | pycompat, |
|
20 | pycompat, | |
21 | streamclone, |
|
21 | streamclone, | |
22 | templatefilters, |
|
22 | templatefilters, | |
23 | util, |
|
23 | util, | |
24 | wireprotoframing, |
|
24 | wireprotoframing, | |
25 | wireprototypes, |
|
25 | wireprototypes, | |
26 | ) |
|
26 | ) | |
27 | from .interfaces import util as interfaceutil |
|
27 | from .interfaces import util as interfaceutil | |
28 | from .utils import ( |
|
28 | from .utils import ( | |
29 | cborutil, |
|
29 | cborutil, | |
30 | hashutil, |
|
30 | hashutil, | |
31 | stringutil, |
|
31 | stringutil, | |
32 | ) |
|
32 | ) | |
33 |
|
33 | |||
34 | FRAMINGTYPE = b'application/mercurial-exp-framing-0006' |
|
34 | FRAMINGTYPE = b'application/mercurial-exp-framing-0006' | |
35 |
|
35 | |||
36 | HTTP_WIREPROTO_V2 = wireprototypes.HTTP_WIREPROTO_V2 |
|
36 | HTTP_WIREPROTO_V2 = wireprototypes.HTTP_WIREPROTO_V2 | |
37 |
|
37 | |||
38 | COMMANDS = wireprototypes.commanddict() |
|
38 | COMMANDS = wireprototypes.commanddict() | |
39 |
|
39 | |||
40 | # Value inserted into cache key computation function. Change the value to |
|
40 | # Value inserted into cache key computation function. Change the value to | |
41 | # force new cache keys for every command request. This should be done when |
|
41 | # force new cache keys for every command request. This should be done when | |
42 | # there is a change to how caching works, etc. |
|
42 | # there is a change to how caching works, etc. | |
43 | GLOBAL_CACHE_VERSION = 1 |
|
43 | GLOBAL_CACHE_VERSION = 1 | |
44 |
|
44 | |||
45 |
|
45 | |||
46 | def handlehttpv2request(rctx, req, res, checkperm, urlparts): |
|
46 | def handlehttpv2request(rctx, req, res, checkperm, urlparts): | |
47 | from .hgweb import common as hgwebcommon |
|
47 | from .hgweb import common as hgwebcommon | |
48 |
|
48 | |||
49 | # URL space looks like: <permissions>/<command>, where <permission> can |
|
49 | # URL space looks like: <permissions>/<command>, where <permission> can | |
50 | # be ``ro`` or ``rw`` to signal read-only or read-write, respectively. |
|
50 | # be ``ro`` or ``rw`` to signal read-only or read-write, respectively. | |
51 |
|
51 | |||
52 | # Root URL does nothing meaningful... yet. |
|
52 | # Root URL does nothing meaningful... yet. | |
53 | if not urlparts: |
|
53 | if not urlparts: | |
54 | res.status = b'200 OK' |
|
54 | res.status = b'200 OK' | |
55 | res.headers[b'Content-Type'] = b'text/plain' |
|
55 | res.headers[b'Content-Type'] = b'text/plain' | |
56 | res.setbodybytes(_(b'HTTP version 2 API handler')) |
|
56 | res.setbodybytes(_(b'HTTP version 2 API handler')) | |
57 | return |
|
57 | return | |
58 |
|
58 | |||
59 | if len(urlparts) == 1: |
|
59 | if len(urlparts) == 1: | |
60 | res.status = b'404 Not Found' |
|
60 | res.status = b'404 Not Found' | |
61 | res.headers[b'Content-Type'] = b'text/plain' |
|
61 | res.headers[b'Content-Type'] = b'text/plain' | |
62 | res.setbodybytes( |
|
62 | res.setbodybytes( | |
63 | _(b'do not know how to process %s\n') % req.dispatchpath |
|
63 | _(b'do not know how to process %s\n') % req.dispatchpath | |
64 | ) |
|
64 | ) | |
65 | return |
|
65 | return | |
66 |
|
66 | |||
67 | permission, command = urlparts[0:2] |
|
67 | permission, command = urlparts[0:2] | |
68 |
|
68 | |||
69 | if permission not in (b'ro', b'rw'): |
|
69 | if permission not in (b'ro', b'rw'): | |
70 | res.status = b'404 Not Found' |
|
70 | res.status = b'404 Not Found' | |
71 | res.headers[b'Content-Type'] = b'text/plain' |
|
71 | res.headers[b'Content-Type'] = b'text/plain' | |
72 | res.setbodybytes(_(b'unknown permission: %s') % permission) |
|
72 | res.setbodybytes(_(b'unknown permission: %s') % permission) | |
73 | return |
|
73 | return | |
74 |
|
74 | |||
75 | if req.method != b'POST': |
|
75 | if req.method != b'POST': | |
76 | res.status = b'405 Method Not Allowed' |
|
76 | res.status = b'405 Method Not Allowed' | |
77 | res.headers[b'Allow'] = b'POST' |
|
77 | res.headers[b'Allow'] = b'POST' | |
78 | res.setbodybytes(_(b'commands require POST requests')) |
|
78 | res.setbodybytes(_(b'commands require POST requests')) | |
79 | return |
|
79 | return | |
80 |
|
80 | |||
81 | # At some point we'll want to use our own API instead of recycling the |
|
81 | # At some point we'll want to use our own API instead of recycling the | |
82 | # behavior of version 1 of the wire protocol... |
|
82 | # behavior of version 1 of the wire protocol... | |
83 | # TODO return reasonable responses - not responses that overload the |
|
83 | # TODO return reasonable responses - not responses that overload the | |
84 | # HTTP status line message for error reporting. |
|
84 | # HTTP status line message for error reporting. | |
85 | try: |
|
85 | try: | |
86 | checkperm(rctx, req, b'pull' if permission == b'ro' else b'push') |
|
86 | checkperm(rctx, req, b'pull' if permission == b'ro' else b'push') | |
87 | except hgwebcommon.ErrorResponse as e: |
|
87 | except hgwebcommon.ErrorResponse as e: | |
88 | res.status = hgwebcommon.statusmessage( |
|
88 | res.status = hgwebcommon.statusmessage( | |
89 | e.code, stringutil.forcebytestr(e) |
|
89 | e.code, stringutil.forcebytestr(e) | |
90 | ) |
|
90 | ) | |
91 | for k, v in e.headers: |
|
91 | for k, v in e.headers: | |
92 | res.headers[k] = v |
|
92 | res.headers[k] = v | |
93 | res.setbodybytes(b'permission denied') |
|
93 | res.setbodybytes(b'permission denied') | |
94 | return |
|
94 | return | |
95 |
|
95 | |||
96 | # We have a special endpoint to reflect the request back at the client. |
|
96 | # We have a special endpoint to reflect the request back at the client. | |
97 | if command == b'debugreflect': |
|
97 | if command == b'debugreflect': | |
98 | _processhttpv2reflectrequest(rctx.repo.ui, rctx.repo, req, res) |
|
98 | _processhttpv2reflectrequest(rctx.repo.ui, rctx.repo, req, res) | |
99 | return |
|
99 | return | |
100 |
|
100 | |||
101 | # Extra commands that we handle that aren't really wire protocol |
|
101 | # Extra commands that we handle that aren't really wire protocol | |
102 | # commands. Think extra hard before making this hackery available to |
|
102 | # commands. Think extra hard before making this hackery available to | |
103 | # extension. |
|
103 | # extension. | |
104 | extracommands = {b'multirequest'} |
|
104 | extracommands = {b'multirequest'} | |
105 |
|
105 | |||
106 | if command not in COMMANDS and command not in extracommands: |
|
106 | if command not in COMMANDS and command not in extracommands: | |
107 | res.status = b'404 Not Found' |
|
107 | res.status = b'404 Not Found' | |
108 | res.headers[b'Content-Type'] = b'text/plain' |
|
108 | res.headers[b'Content-Type'] = b'text/plain' | |
109 | res.setbodybytes(_(b'unknown wire protocol command: %s\n') % command) |
|
109 | res.setbodybytes(_(b'unknown wire protocol command: %s\n') % command) | |
110 | return |
|
110 | return | |
111 |
|
111 | |||
112 | repo = rctx.repo |
|
112 | repo = rctx.repo | |
113 | ui = repo.ui |
|
113 | ui = repo.ui | |
114 |
|
114 | |||
115 | proto = httpv2protocolhandler(req, ui) |
|
115 | proto = httpv2protocolhandler(req, ui) | |
116 |
|
116 | |||
117 | if ( |
|
117 | if ( | |
118 | not COMMANDS.commandavailable(command, proto) |
|
118 | not COMMANDS.commandavailable(command, proto) | |
119 | and command not in extracommands |
|
119 | and command not in extracommands | |
120 | ): |
|
120 | ): | |
121 | res.status = b'404 Not Found' |
|
121 | res.status = b'404 Not Found' | |
122 | res.headers[b'Content-Type'] = b'text/plain' |
|
122 | res.headers[b'Content-Type'] = b'text/plain' | |
123 | res.setbodybytes(_(b'invalid wire protocol command: %s') % command) |
|
123 | res.setbodybytes(_(b'invalid wire protocol command: %s') % command) | |
124 | return |
|
124 | return | |
125 |
|
125 | |||
126 | # TODO consider cases where proxies may add additional Accept headers. |
|
126 | # TODO consider cases where proxies may add additional Accept headers. | |
127 | if req.headers.get(b'Accept') != FRAMINGTYPE: |
|
127 | if req.headers.get(b'Accept') != FRAMINGTYPE: | |
128 | res.status = b'406 Not Acceptable' |
|
128 | res.status = b'406 Not Acceptable' | |
129 | res.headers[b'Content-Type'] = b'text/plain' |
|
129 | res.headers[b'Content-Type'] = b'text/plain' | |
130 | res.setbodybytes( |
|
130 | res.setbodybytes( | |
131 | _(b'client MUST specify Accept header with value: %s\n') |
|
131 | _(b'client MUST specify Accept header with value: %s\n') | |
132 | % FRAMINGTYPE |
|
132 | % FRAMINGTYPE | |
133 | ) |
|
133 | ) | |
134 | return |
|
134 | return | |
135 |
|
135 | |||
136 | if req.headers.get(b'Content-Type') != FRAMINGTYPE: |
|
136 | if req.headers.get(b'Content-Type') != FRAMINGTYPE: | |
137 | res.status = b'415 Unsupported Media Type' |
|
137 | res.status = b'415 Unsupported Media Type' | |
138 | # TODO we should send a response with appropriate media type, |
|
138 | # TODO we should send a response with appropriate media type, | |
139 | # since client does Accept it. |
|
139 | # since client does Accept it. | |
140 | res.headers[b'Content-Type'] = b'text/plain' |
|
140 | res.headers[b'Content-Type'] = b'text/plain' | |
141 | res.setbodybytes( |
|
141 | res.setbodybytes( | |
142 | _(b'client MUST send Content-Type header with value: %s\n') |
|
142 | _(b'client MUST send Content-Type header with value: %s\n') | |
143 | % FRAMINGTYPE |
|
143 | % FRAMINGTYPE | |
144 | ) |
|
144 | ) | |
145 | return |
|
145 | return | |
146 |
|
146 | |||
147 | _processhttpv2request(ui, repo, req, res, permission, command, proto) |
|
147 | _processhttpv2request(ui, repo, req, res, permission, command, proto) | |
148 |
|
148 | |||
149 |
|
149 | |||
150 | def _processhttpv2reflectrequest(ui, repo, req, res): |
|
150 | def _processhttpv2reflectrequest(ui, repo, req, res): | |
151 | """Reads unified frame protocol request and dumps out state to client. |
|
151 | """Reads unified frame protocol request and dumps out state to client. | |
152 |
|
152 | |||
153 | This special endpoint can be used to help debug the wire protocol. |
|
153 | This special endpoint can be used to help debug the wire protocol. | |
154 |
|
154 | |||
155 | Instead of routing the request through the normal dispatch mechanism, |
|
155 | Instead of routing the request through the normal dispatch mechanism, | |
156 | we instead read all frames, decode them, and feed them into our state |
|
156 | we instead read all frames, decode them, and feed them into our state | |
157 | tracker. We then dump the log of all that activity back out to the |
|
157 | tracker. We then dump the log of all that activity back out to the | |
158 | client. |
|
158 | client. | |
159 | """ |
|
159 | """ | |
160 | # Reflection APIs have a history of being abused, accidentally disclosing |
|
160 | # Reflection APIs have a history of being abused, accidentally disclosing | |
161 | # sensitive data, etc. So we have a config knob. |
|
161 | # sensitive data, etc. So we have a config knob. | |
162 | if not ui.configbool(b'experimental', b'web.api.debugreflect'): |
|
162 | if not ui.configbool(b'experimental', b'web.api.debugreflect'): | |
163 | res.status = b'404 Not Found' |
|
163 | res.status = b'404 Not Found' | |
164 | res.headers[b'Content-Type'] = b'text/plain' |
|
164 | res.headers[b'Content-Type'] = b'text/plain' | |
165 | res.setbodybytes(_(b'debugreflect service not available')) |
|
165 | res.setbodybytes(_(b'debugreflect service not available')) | |
166 | return |
|
166 | return | |
167 |
|
167 | |||
168 | # We assume we have a unified framing protocol request body. |
|
168 | # We assume we have a unified framing protocol request body. | |
169 |
|
169 | |||
170 | reactor = wireprotoframing.serverreactor(ui) |
|
170 | reactor = wireprotoframing.serverreactor(ui) | |
171 | states = [] |
|
171 | states = [] | |
172 |
|
172 | |||
173 | while True: |
|
173 | while True: | |
174 | frame = wireprotoframing.readframe(req.bodyfh) |
|
174 | frame = wireprotoframing.readframe(req.bodyfh) | |
175 |
|
175 | |||
176 | if not frame: |
|
176 | if not frame: | |
177 | states.append(b'received: <no frame>') |
|
177 | states.append(b'received: <no frame>') | |
178 | break |
|
178 | break | |
179 |
|
179 | |||
180 | states.append( |
|
180 | states.append( | |
181 | b'received: %d %d %d %s' |
|
181 | b'received: %d %d %d %s' | |
182 | % (frame.typeid, frame.flags, frame.requestid, frame.payload) |
|
182 | % (frame.typeid, frame.flags, frame.requestid, frame.payload) | |
183 | ) |
|
183 | ) | |
184 |
|
184 | |||
185 | action, meta = reactor.onframerecv(frame) |
|
185 | action, meta = reactor.onframerecv(frame) | |
186 | states.append(templatefilters.json((action, meta))) |
|
186 | states.append(templatefilters.json((action, meta))) | |
187 |
|
187 | |||
188 | action, meta = reactor.oninputeof() |
|
188 | action, meta = reactor.oninputeof() | |
189 | meta[b'action'] = action |
|
189 | meta[b'action'] = action | |
190 | states.append(templatefilters.json(meta)) |
|
190 | states.append(templatefilters.json(meta)) | |
191 |
|
191 | |||
192 | res.status = b'200 OK' |
|
192 | res.status = b'200 OK' | |
193 | res.headers[b'Content-Type'] = b'text/plain' |
|
193 | res.headers[b'Content-Type'] = b'text/plain' | |
194 | res.setbodybytes(b'\n'.join(states)) |
|
194 | res.setbodybytes(b'\n'.join(states)) | |
195 |
|
195 | |||
196 |
|
196 | |||
197 | def _processhttpv2request(ui, repo, req, res, authedperm, reqcommand, proto): |
|
197 | def _processhttpv2request(ui, repo, req, res, authedperm, reqcommand, proto): | |
198 | """Post-validation handler for HTTPv2 requests. |
|
198 | """Post-validation handler for HTTPv2 requests. | |
199 |
|
199 | |||
200 | Called when the HTTP request contains unified frame-based protocol |
|
200 | Called when the HTTP request contains unified frame-based protocol | |
201 | frames for evaluation. |
|
201 | frames for evaluation. | |
202 | """ |
|
202 | """ | |
203 | # TODO Some HTTP clients are full duplex and can receive data before |
|
203 | # TODO Some HTTP clients are full duplex and can receive data before | |
204 | # the entire request is transmitted. Figure out a way to indicate support |
|
204 | # the entire request is transmitted. Figure out a way to indicate support | |
205 | # for that so we can opt into full duplex mode. |
|
205 | # for that so we can opt into full duplex mode. | |
206 | reactor = wireprotoframing.serverreactor(ui, deferoutput=True) |
|
206 | reactor = wireprotoframing.serverreactor(ui, deferoutput=True) | |
207 | seencommand = False |
|
207 | seencommand = False | |
208 |
|
208 | |||
209 | outstream = None |
|
209 | outstream = None | |
210 |
|
210 | |||
211 | while True: |
|
211 | while True: | |
212 | frame = wireprotoframing.readframe(req.bodyfh) |
|
212 | frame = wireprotoframing.readframe(req.bodyfh) | |
213 | if not frame: |
|
213 | if not frame: | |
214 | break |
|
214 | break | |
215 |
|
215 | |||
216 | action, meta = reactor.onframerecv(frame) |
|
216 | action, meta = reactor.onframerecv(frame) | |
217 |
|
217 | |||
218 | if action == b'wantframe': |
|
218 | if action == b'wantframe': | |
219 | # Need more data before we can do anything. |
|
219 | # Need more data before we can do anything. | |
220 | continue |
|
220 | continue | |
221 | elif action == b'runcommand': |
|
221 | elif action == b'runcommand': | |
222 | # Defer creating output stream because we need to wait for |
|
222 | # Defer creating output stream because we need to wait for | |
223 | # protocol settings frames so proper encoding can be applied. |
|
223 | # protocol settings frames so proper encoding can be applied. | |
224 | if not outstream: |
|
224 | if not outstream: | |
225 | outstream = reactor.makeoutputstream() |
|
225 | outstream = reactor.makeoutputstream() | |
226 |
|
226 | |||
227 | sentoutput = _httpv2runcommand( |
|
227 | sentoutput = _httpv2runcommand( | |
228 | ui, |
|
228 | ui, | |
229 | repo, |
|
229 | repo, | |
230 | req, |
|
230 | req, | |
231 | res, |
|
231 | res, | |
232 | authedperm, |
|
232 | authedperm, | |
233 | reqcommand, |
|
233 | reqcommand, | |
234 | reactor, |
|
234 | reactor, | |
235 | outstream, |
|
235 | outstream, | |
236 | meta, |
|
236 | meta, | |
237 | issubsequent=seencommand, |
|
237 | issubsequent=seencommand, | |
238 | ) |
|
238 | ) | |
239 |
|
239 | |||
240 | if sentoutput: |
|
240 | if sentoutput: | |
241 | return |
|
241 | return | |
242 |
|
242 | |||
243 | seencommand = True |
|
243 | seencommand = True | |
244 |
|
244 | |||
245 | elif action == b'error': |
|
245 | elif action == b'error': | |
246 | # TODO define proper error mechanism. |
|
246 | # TODO define proper error mechanism. | |
247 | res.status = b'200 OK' |
|
247 | res.status = b'200 OK' | |
248 | res.headers[b'Content-Type'] = b'text/plain' |
|
248 | res.headers[b'Content-Type'] = b'text/plain' | |
249 | res.setbodybytes(meta[b'message'] + b'\n') |
|
249 | res.setbodybytes(meta[b'message'] + b'\n') | |
250 | return |
|
250 | return | |
251 | else: |
|
251 | else: | |
252 | raise error.ProgrammingError( |
|
252 | raise error.ProgrammingError( | |
253 | b'unhandled action from frame processor: %s' % action |
|
253 | b'unhandled action from frame processor: %s' % action | |
254 | ) |
|
254 | ) | |
255 |
|
255 | |||
256 | action, meta = reactor.oninputeof() |
|
256 | action, meta = reactor.oninputeof() | |
257 | if action == b'sendframes': |
|
257 | if action == b'sendframes': | |
258 | # We assume we haven't started sending the response yet. If we're |
|
258 | # We assume we haven't started sending the response yet. If we're | |
259 | # wrong, the response type will raise an exception. |
|
259 | # wrong, the response type will raise an exception. | |
260 | res.status = b'200 OK' |
|
260 | res.status = b'200 OK' | |
261 | res.headers[b'Content-Type'] = FRAMINGTYPE |
|
261 | res.headers[b'Content-Type'] = FRAMINGTYPE | |
262 | res.setbodygen(meta[b'framegen']) |
|
262 | res.setbodygen(meta[b'framegen']) | |
263 | elif action == b'noop': |
|
263 | elif action == b'noop': | |
264 | pass |
|
264 | pass | |
265 | else: |
|
265 | else: | |
266 | raise error.ProgrammingError( |
|
266 | raise error.ProgrammingError( | |
267 | b'unhandled action from frame processor: %s' % action |
|
267 | b'unhandled action from frame processor: %s' % action | |
268 | ) |
|
268 | ) | |
269 |
|
269 | |||
270 |
|
270 | |||
271 | def _httpv2runcommand( |
|
271 | def _httpv2runcommand( | |
272 | ui, |
|
272 | ui, | |
273 | repo, |
|
273 | repo, | |
274 | req, |
|
274 | req, | |
275 | res, |
|
275 | res, | |
276 | authedperm, |
|
276 | authedperm, | |
277 | reqcommand, |
|
277 | reqcommand, | |
278 | reactor, |
|
278 | reactor, | |
279 | outstream, |
|
279 | outstream, | |
280 | command, |
|
280 | command, | |
281 | issubsequent, |
|
281 | issubsequent, | |
282 | ): |
|
282 | ): | |
283 | """Dispatch a wire protocol command made from HTTPv2 requests. |
|
283 | """Dispatch a wire protocol command made from HTTPv2 requests. | |
284 |
|
284 | |||
285 | The authenticated permission (``authedperm``) along with the original |
|
285 | The authenticated permission (``authedperm``) along with the original | |
286 | command from the URL (``reqcommand``) are passed in. |
|
286 | command from the URL (``reqcommand``) are passed in. | |
287 | """ |
|
287 | """ | |
288 | # We already validated that the session has permissions to perform the |
|
288 | # We already validated that the session has permissions to perform the | |
289 | # actions in ``authedperm``. In the unified frame protocol, the canonical |
|
289 | # actions in ``authedperm``. In the unified frame protocol, the canonical | |
290 | # command to run is expressed in a frame. However, the URL also requested |
|
290 | # command to run is expressed in a frame. However, the URL also requested | |
291 | # to run a specific command. We need to be careful that the command we |
|
291 | # to run a specific command. We need to be careful that the command we | |
292 | # run doesn't have permissions requirements greater than what was granted |
|
292 | # run doesn't have permissions requirements greater than what was granted | |
293 | # by ``authedperm``. |
|
293 | # by ``authedperm``. | |
294 | # |
|
294 | # | |
295 | # Our rule for this is we only allow one command per HTTP request and |
|
295 | # Our rule for this is we only allow one command per HTTP request and | |
296 | # that command must match the command in the URL. However, we make |
|
296 | # that command must match the command in the URL. However, we make | |
297 | # an exception for the ``multirequest`` URL. This URL is allowed to |
|
297 | # an exception for the ``multirequest`` URL. This URL is allowed to | |
298 | # execute multiple commands. We double check permissions of each command |
|
298 | # execute multiple commands. We double check permissions of each command | |
299 | # as it is invoked to ensure there is no privilege escalation. |
|
299 | # as it is invoked to ensure there is no privilege escalation. | |
300 | # TODO consider allowing multiple commands to regular command URLs |
|
300 | # TODO consider allowing multiple commands to regular command URLs | |
301 | # iff each command is the same. |
|
301 | # iff each command is the same. | |
302 |
|
302 | |||
303 | proto = httpv2protocolhandler(req, ui, args=command[b'args']) |
|
303 | proto = httpv2protocolhandler(req, ui, args=command[b'args']) | |
304 |
|
304 | |||
305 | if reqcommand == b'multirequest': |
|
305 | if reqcommand == b'multirequest': | |
306 | if not COMMANDS.commandavailable(command[b'command'], proto): |
|
306 | if not COMMANDS.commandavailable(command[b'command'], proto): | |
307 | # TODO proper error mechanism |
|
307 | # TODO proper error mechanism | |
308 | res.status = b'200 OK' |
|
308 | res.status = b'200 OK' | |
309 | res.headers[b'Content-Type'] = b'text/plain' |
|
309 | res.headers[b'Content-Type'] = b'text/plain' | |
310 | res.setbodybytes( |
|
310 | res.setbodybytes( | |
311 | _(b'wire protocol command not available: %s') |
|
311 | _(b'wire protocol command not available: %s') | |
312 | % command[b'command'] |
|
312 | % command[b'command'] | |
313 | ) |
|
313 | ) | |
314 | return True |
|
314 | return True | |
315 |
|
315 | |||
316 | # TODO don't use assert here, since it may be elided by -O. |
|
316 | # TODO don't use assert here, since it may be elided by -O. | |
317 | assert authedperm in (b'ro', b'rw') |
|
317 | assert authedperm in (b'ro', b'rw') | |
318 | wirecommand = COMMANDS[command[b'command']] |
|
318 | wirecommand = COMMANDS[command[b'command']] | |
319 | assert wirecommand.permission in (b'push', b'pull') |
|
319 | assert wirecommand.permission in (b'push', b'pull') | |
320 |
|
320 | |||
321 | if authedperm == b'ro' and wirecommand.permission != b'pull': |
|
321 | if authedperm == b'ro' and wirecommand.permission != b'pull': | |
322 | # TODO proper error mechanism |
|
322 | # TODO proper error mechanism | |
323 | res.status = b'403 Forbidden' |
|
323 | res.status = b'403 Forbidden' | |
324 | res.headers[b'Content-Type'] = b'text/plain' |
|
324 | res.headers[b'Content-Type'] = b'text/plain' | |
325 | res.setbodybytes( |
|
325 | res.setbodybytes( | |
326 | _(b'insufficient permissions to execute command: %s') |
|
326 | _(b'insufficient permissions to execute command: %s') | |
327 | % command[b'command'] |
|
327 | % command[b'command'] | |
328 | ) |
|
328 | ) | |
329 | return True |
|
329 | return True | |
330 |
|
330 | |||
331 | # TODO should we also call checkperm() here? Maybe not if we're going |
|
331 | # TODO should we also call checkperm() here? Maybe not if we're going | |
332 | # to overhaul that API. The granted scope from the URL check should |
|
332 | # to overhaul that API. The granted scope from the URL check should | |
333 | # be good enough. |
|
333 | # be good enough. | |
334 |
|
334 | |||
335 | else: |
|
335 | else: | |
336 | # Don't allow multiple commands outside of ``multirequest`` URL. |
|
336 | # Don't allow multiple commands outside of ``multirequest`` URL. | |
337 | if issubsequent: |
|
337 | if issubsequent: | |
338 | # TODO proper error mechanism |
|
338 | # TODO proper error mechanism | |
339 | res.status = b'200 OK' |
|
339 | res.status = b'200 OK' | |
340 | res.headers[b'Content-Type'] = b'text/plain' |
|
340 | res.headers[b'Content-Type'] = b'text/plain' | |
341 | res.setbodybytes( |
|
341 | res.setbodybytes( | |
342 | _(b'multiple commands cannot be issued to this URL') |
|
342 | _(b'multiple commands cannot be issued to this URL') | |
343 | ) |
|
343 | ) | |
344 | return True |
|
344 | return True | |
345 |
|
345 | |||
346 | if reqcommand != command[b'command']: |
|
346 | if reqcommand != command[b'command']: | |
347 | # TODO define proper error mechanism |
|
347 | # TODO define proper error mechanism | |
348 | res.status = b'200 OK' |
|
348 | res.status = b'200 OK' | |
349 | res.headers[b'Content-Type'] = b'text/plain' |
|
349 | res.headers[b'Content-Type'] = b'text/plain' | |
350 | res.setbodybytes(_(b'command in frame must match command in URL')) |
|
350 | res.setbodybytes(_(b'command in frame must match command in URL')) | |
351 | return True |
|
351 | return True | |
352 |
|
352 | |||
353 | res.status = b'200 OK' |
|
353 | res.status = b'200 OK' | |
354 | res.headers[b'Content-Type'] = FRAMINGTYPE |
|
354 | res.headers[b'Content-Type'] = FRAMINGTYPE | |
355 |
|
355 | |||
356 | try: |
|
356 | try: | |
357 | objs = dispatch(repo, proto, command[b'command'], command[b'redirect']) |
|
357 | objs = dispatch(repo, proto, command[b'command'], command[b'redirect']) | |
358 |
|
358 | |||
359 | action, meta = reactor.oncommandresponsereadyobjects( |
|
359 | action, meta = reactor.oncommandresponsereadyobjects( | |
360 | outstream, command[b'requestid'], objs |
|
360 | outstream, command[b'requestid'], objs | |
361 | ) |
|
361 | ) | |
362 |
|
362 | |||
363 | except error.WireprotoCommandError as e: |
|
363 | except error.WireprotoCommandError as e: | |
364 | action, meta = reactor.oncommanderror( |
|
364 | action, meta = reactor.oncommanderror( | |
365 | outstream, command[b'requestid'], e.message, e.messageargs |
|
365 | outstream, command[b'requestid'], e.message, e.messageargs | |
366 | ) |
|
366 | ) | |
367 |
|
367 | |||
368 | except Exception as e: |
|
368 | except Exception as e: | |
369 | action, meta = reactor.onservererror( |
|
369 | action, meta = reactor.onservererror( | |
370 | outstream, |
|
370 | outstream, | |
371 | command[b'requestid'], |
|
371 | command[b'requestid'], | |
372 | _(b'exception when invoking command: %s') |
|
372 | _(b'exception when invoking command: %s') | |
373 | % stringutil.forcebytestr(e), |
|
373 | % stringutil.forcebytestr(e), | |
374 | ) |
|
374 | ) | |
375 |
|
375 | |||
376 | if action == b'sendframes': |
|
376 | if action == b'sendframes': | |
377 | res.setbodygen(meta[b'framegen']) |
|
377 | res.setbodygen(meta[b'framegen']) | |
378 | return True |
|
378 | return True | |
379 | elif action == b'noop': |
|
379 | elif action == b'noop': | |
380 | return False |
|
380 | return False | |
381 | else: |
|
381 | else: | |
382 | raise error.ProgrammingError( |
|
382 | raise error.ProgrammingError( | |
383 | b'unhandled event from reactor: %s' % action |
|
383 | b'unhandled event from reactor: %s' % action | |
384 | ) |
|
384 | ) | |
385 |
|
385 | |||
386 |
|
386 | |||
387 | def getdispatchrepo(repo, proto, command): |
|
387 | def getdispatchrepo(repo, proto, command): | |
388 | viewconfig = repo.ui.config(b'server', b'view') |
|
388 | viewconfig = repo.ui.config(b'server', b'view') | |
389 | return repo.filtered(viewconfig) |
|
389 | return repo.filtered(viewconfig) | |
390 |
|
390 | |||
391 |
|
391 | |||
392 | def dispatch(repo, proto, command, redirect): |
|
392 | def dispatch(repo, proto, command, redirect): | |
393 | """Run a wire protocol command. |
|
393 | """Run a wire protocol command. | |
394 |
|
394 | |||
395 | Returns an iterable of objects that will be sent to the client. |
|
395 | Returns an iterable of objects that will be sent to the client. | |
396 | """ |
|
396 | """ | |
397 | repo = getdispatchrepo(repo, proto, command) |
|
397 | repo = getdispatchrepo(repo, proto, command) | |
398 |
|
398 | |||
399 | entry = COMMANDS[command] |
|
399 | entry = COMMANDS[command] | |
400 | func = entry.func |
|
400 | func = entry.func | |
401 | spec = entry.args |
|
401 | spec = entry.args | |
402 |
|
402 | |||
403 | args = proto.getargs(spec) |
|
403 | args = proto.getargs(spec) | |
404 |
|
404 | |||
405 | # There is some duplicate boilerplate code here for calling the command and |
|
405 | # There is some duplicate boilerplate code here for calling the command and | |
406 | # emitting objects. It is either that or a lot of indented code that looks |
|
406 | # emitting objects. It is either that or a lot of indented code that looks | |
407 | # like a pyramid (since there are a lot of code paths that result in not |
|
407 | # like a pyramid (since there are a lot of code paths that result in not | |
408 | # using the cacher). |
|
408 | # using the cacher). | |
409 | callcommand = lambda: func(repo, proto, **pycompat.strkwargs(args)) |
|
409 | callcommand = lambda: func(repo, proto, **pycompat.strkwargs(args)) | |
410 |
|
410 | |||
411 | # Request is not cacheable. Don't bother instantiating a cacher. |
|
411 | # Request is not cacheable. Don't bother instantiating a cacher. | |
412 | if not entry.cachekeyfn: |
|
412 | if not entry.cachekeyfn: | |
413 | for o in callcommand(): |
|
413 | for o in callcommand(): | |
414 | yield o |
|
414 | yield o | |
415 | return |
|
415 | return | |
416 |
|
416 | |||
417 | if redirect: |
|
417 | if redirect: | |
418 | redirecttargets = redirect[b'targets'] |
|
418 | redirecttargets = redirect[b'targets'] | |
419 | redirecthashes = redirect[b'hashes'] |
|
419 | redirecthashes = redirect[b'hashes'] | |
420 | else: |
|
420 | else: | |
421 | redirecttargets = [] |
|
421 | redirecttargets = [] | |
422 | redirecthashes = [] |
|
422 | redirecthashes = [] | |
423 |
|
423 | |||
424 | cacher = makeresponsecacher( |
|
424 | cacher = makeresponsecacher( | |
425 | repo, |
|
425 | repo, | |
426 | proto, |
|
426 | proto, | |
427 | command, |
|
427 | command, | |
428 | args, |
|
428 | args, | |
429 | cborutil.streamencode, |
|
429 | cborutil.streamencode, | |
430 | redirecttargets=redirecttargets, |
|
430 | redirecttargets=redirecttargets, | |
431 | redirecthashes=redirecthashes, |
|
431 | redirecthashes=redirecthashes, | |
432 | ) |
|
432 | ) | |
433 |
|
433 | |||
434 | # But we have no cacher. Do default handling. |
|
434 | # But we have no cacher. Do default handling. | |
435 | if not cacher: |
|
435 | if not cacher: | |
436 | for o in callcommand(): |
|
436 | for o in callcommand(): | |
437 | yield o |
|
437 | yield o | |
438 | return |
|
438 | return | |
439 |
|
439 | |||
440 | with cacher: |
|
440 | with cacher: | |
441 | cachekey = entry.cachekeyfn( |
|
441 | cachekey = entry.cachekeyfn( | |
442 | repo, proto, cacher, **pycompat.strkwargs(args) |
|
442 | repo, proto, cacher, **pycompat.strkwargs(args) | |
443 | ) |
|
443 | ) | |
444 |
|
444 | |||
445 | # No cache key or the cacher doesn't like it. Do default handling. |
|
445 | # No cache key or the cacher doesn't like it. Do default handling. | |
446 | if cachekey is None or not cacher.setcachekey(cachekey): |
|
446 | if cachekey is None or not cacher.setcachekey(cachekey): | |
447 | for o in callcommand(): |
|
447 | for o in callcommand(): | |
448 | yield o |
|
448 | yield o | |
449 | return |
|
449 | return | |
450 |
|
450 | |||
451 | # Serve it from the cache, if possible. |
|
451 | # Serve it from the cache, if possible. | |
452 | cached = cacher.lookup() |
|
452 | cached = cacher.lookup() | |
453 |
|
453 | |||
454 | if cached: |
|
454 | if cached: | |
455 | for o in cached[b'objs']: |
|
455 | for o in cached[b'objs']: | |
456 | yield o |
|
456 | yield o | |
457 | return |
|
457 | return | |
458 |
|
458 | |||
459 | # Else call the command and feed its output into the cacher, allowing |
|
459 | # Else call the command and feed its output into the cacher, allowing | |
460 | # the cacher to buffer/mutate objects as it desires. |
|
460 | # the cacher to buffer/mutate objects as it desires. | |
461 | for o in callcommand(): |
|
461 | for o in callcommand(): | |
462 | for o in cacher.onobject(o): |
|
462 | for o in cacher.onobject(o): | |
463 | yield o |
|
463 | yield o | |
464 |
|
464 | |||
465 | for o in cacher.onfinished(): |
|
465 | for o in cacher.onfinished(): | |
466 | yield o |
|
466 | yield o | |
467 |
|
467 | |||
468 |
|
468 | |||
469 | @interfaceutil.implementer(wireprototypes.baseprotocolhandler) |
|
469 | @interfaceutil.implementer(wireprototypes.baseprotocolhandler) | |
470 | class httpv2protocolhandler(object): |
|
470 | class httpv2protocolhandler(object): | |
471 | def __init__(self, req, ui, args=None): |
|
471 | def __init__(self, req, ui, args=None): | |
472 | self._req = req |
|
472 | self._req = req | |
473 | self._ui = ui |
|
473 | self._ui = ui | |
474 | self._args = args |
|
474 | self._args = args | |
475 |
|
475 | |||
476 | @property |
|
476 | @property | |
477 | def name(self): |
|
477 | def name(self): | |
478 | return HTTP_WIREPROTO_V2 |
|
478 | return HTTP_WIREPROTO_V2 | |
479 |
|
479 | |||
480 | def getargs(self, args): |
|
480 | def getargs(self, args): | |
481 | # First look for args that were passed but aren't registered on this |
|
481 | # First look for args that were passed but aren't registered on this | |
482 | # command. |
|
482 | # command. | |
483 | extra = set(self._args) - set(args) |
|
483 | extra = set(self._args) - set(args) | |
484 | if extra: |
|
484 | if extra: | |
485 | raise error.WireprotoCommandError( |
|
485 | raise error.WireprotoCommandError( | |
486 | b'unsupported argument to command: %s' |
|
486 | b'unsupported argument to command: %s' | |
487 | % b', '.join(sorted(extra)) |
|
487 | % b', '.join(sorted(extra)) | |
488 | ) |
|
488 | ) | |
489 |
|
489 | |||
490 | # And look for required arguments that are missing. |
|
490 | # And look for required arguments that are missing. | |
491 | missing = {a for a in args if args[a][b'required']} - set(self._args) |
|
491 | missing = {a for a in args if args[a][b'required']} - set(self._args) | |
492 |
|
492 | |||
493 | if missing: |
|
493 | if missing: | |
494 | raise error.WireprotoCommandError( |
|
494 | raise error.WireprotoCommandError( | |
495 | b'missing required arguments: %s' % b', '.join(sorted(missing)) |
|
495 | b'missing required arguments: %s' % b', '.join(sorted(missing)) | |
496 | ) |
|
496 | ) | |
497 |
|
497 | |||
498 | # Now derive the arguments to pass to the command, taking into |
|
498 | # Now derive the arguments to pass to the command, taking into | |
499 | # account the arguments specified by the client. |
|
499 | # account the arguments specified by the client. | |
500 | data = {} |
|
500 | data = {} | |
501 | for k, meta in sorted(args.items()): |
|
501 | for k, meta in sorted(args.items()): | |
502 | # This argument wasn't passed by the client. |
|
502 | # This argument wasn't passed by the client. | |
503 | if k not in self._args: |
|
503 | if k not in self._args: | |
504 | data[k] = meta[b'default']() |
|
504 | data[k] = meta[b'default']() | |
505 | continue |
|
505 | continue | |
506 |
|
506 | |||
507 | v = self._args[k] |
|
507 | v = self._args[k] | |
508 |
|
508 | |||
509 | # Sets may be expressed as lists. Silently normalize. |
|
509 | # Sets may be expressed as lists. Silently normalize. | |
510 | if meta[b'type'] == b'set' and isinstance(v, list): |
|
510 | if meta[b'type'] == b'set' and isinstance(v, list): | |
511 | v = set(v) |
|
511 | v = set(v) | |
512 |
|
512 | |||
513 | # TODO consider more/stronger type validation. |
|
513 | # TODO consider more/stronger type validation. | |
514 |
|
514 | |||
515 | data[k] = v |
|
515 | data[k] = v | |
516 |
|
516 | |||
517 | return data |
|
517 | return data | |
518 |
|
518 | |||
519 | def getprotocaps(self): |
|
519 | def getprotocaps(self): | |
520 | # Protocol capabilities are currently not implemented for HTTP V2. |
|
520 | # Protocol capabilities are currently not implemented for HTTP V2. | |
521 | return set() |
|
521 | return set() | |
522 |
|
522 | |||
523 | def getpayload(self): |
|
523 | def getpayload(self): | |
524 | raise NotImplementedError |
|
524 | raise NotImplementedError | |
525 |
|
525 | |||
526 | @contextlib.contextmanager |
|
526 | @contextlib.contextmanager | |
527 | def mayberedirectstdio(self): |
|
527 | def mayberedirectstdio(self): | |
528 | raise NotImplementedError |
|
528 | raise NotImplementedError | |
529 |
|
529 | |||
530 | def client(self): |
|
530 | def client(self): | |
531 | raise NotImplementedError |
|
531 | raise NotImplementedError | |
532 |
|
532 | |||
533 | def addcapabilities(self, repo, caps): |
|
533 | def addcapabilities(self, repo, caps): | |
534 | return caps |
|
534 | return caps | |
535 |
|
535 | |||
536 | def checkperm(self, perm): |
|
536 | def checkperm(self, perm): | |
537 | raise NotImplementedError |
|
537 | raise NotImplementedError | |
538 |
|
538 | |||
539 |
|
539 | |||
540 | def httpv2apidescriptor(req, repo): |
|
540 | def httpv2apidescriptor(req, repo): | |
541 | proto = httpv2protocolhandler(req, repo.ui) |
|
541 | proto = httpv2protocolhandler(req, repo.ui) | |
542 |
|
542 | |||
543 | return _capabilitiesv2(repo, proto) |
|
543 | return _capabilitiesv2(repo, proto) | |
544 |
|
544 | |||
545 |
|
545 | |||
546 | def _capabilitiesv2(repo, proto): |
|
546 | def _capabilitiesv2(repo, proto): | |
547 | """Obtain the set of capabilities for version 2 transports. |
|
547 | """Obtain the set of capabilities for version 2 transports. | |
548 |
|
548 | |||
549 | These capabilities are distinct from the capabilities for version 1 |
|
549 | These capabilities are distinct from the capabilities for version 1 | |
550 | transports. |
|
550 | transports. | |
551 | """ |
|
551 | """ | |
552 | caps = { |
|
552 | caps = { | |
553 | b'commands': {}, |
|
553 | b'commands': {}, | |
554 | b'framingmediatypes': [FRAMINGTYPE], |
|
554 | b'framingmediatypes': [FRAMINGTYPE], | |
555 | b'pathfilterprefixes': set(narrowspec.VALID_PREFIXES), |
|
555 | b'pathfilterprefixes': set(narrowspec.VALID_PREFIXES), | |
556 | } |
|
556 | } | |
557 |
|
557 | |||
558 | for command, entry in COMMANDS.items(): |
|
558 | for command, entry in COMMANDS.items(): | |
559 | args = {} |
|
559 | args = {} | |
560 |
|
560 | |||
561 | for arg, meta in entry.args.items(): |
|
561 | for arg, meta in entry.args.items(): | |
562 | args[arg] = { |
|
562 | args[arg] = { | |
563 | # TODO should this be a normalized type using CBOR's |
|
563 | # TODO should this be a normalized type using CBOR's | |
564 | # terminology? |
|
564 | # terminology? | |
565 | b'type': meta[b'type'], |
|
565 | b'type': meta[b'type'], | |
566 | b'required': meta[b'required'], |
|
566 | b'required': meta[b'required'], | |
567 | } |
|
567 | } | |
568 |
|
568 | |||
569 | if not meta[b'required']: |
|
569 | if not meta[b'required']: | |
570 | args[arg][b'default'] = meta[b'default']() |
|
570 | args[arg][b'default'] = meta[b'default']() | |
571 |
|
571 | |||
572 | if meta[b'validvalues']: |
|
572 | if meta[b'validvalues']: | |
573 | args[arg][b'validvalues'] = meta[b'validvalues'] |
|
573 | args[arg][b'validvalues'] = meta[b'validvalues'] | |
574 |
|
574 | |||
575 | # TODO this type of check should be defined in a per-command callback. |
|
575 | # TODO this type of check should be defined in a per-command callback. | |
576 | if ( |
|
576 | if ( | |
577 | command == b'rawstorefiledata' |
|
577 | command == b'rawstorefiledata' | |
578 | and not streamclone.allowservergeneration(repo) |
|
578 | and not streamclone.allowservergeneration(repo) | |
579 | ): |
|
579 | ): | |
580 | continue |
|
580 | continue | |
581 |
|
581 | |||
582 | caps[b'commands'][command] = { |
|
582 | caps[b'commands'][command] = { | |
583 | b'args': args, |
|
583 | b'args': args, | |
584 | b'permissions': [entry.permission], |
|
584 | b'permissions': [entry.permission], | |
585 | } |
|
585 | } | |
586 |
|
586 | |||
587 | if entry.extracapabilitiesfn: |
|
587 | if entry.extracapabilitiesfn: | |
588 | extracaps = entry.extracapabilitiesfn(repo, proto) |
|
588 | extracaps = entry.extracapabilitiesfn(repo, proto) | |
589 | caps[b'commands'][command].update(extracaps) |
|
589 | caps[b'commands'][command].update(extracaps) | |
590 |
|
590 | |||
591 | caps[b'rawrepoformats'] = sorted(repo.requirements & repo.supportedformats) |
|
591 | caps[b'rawrepoformats'] = sorted(repo.requirements & repo.supportedformats) | |
592 |
|
592 | |||
593 | targets = getadvertisedredirecttargets(repo, proto) |
|
593 | targets = getadvertisedredirecttargets(repo, proto) | |
594 | if targets: |
|
594 | if targets: | |
595 | caps[b'redirect'] = { |
|
595 | caps[b'redirect'] = { | |
596 | b'targets': [], |
|
596 | b'targets': [], | |
597 | b'hashes': [b'sha256', b'sha1'], |
|
597 | b'hashes': [b'sha256', b'sha1'], | |
598 | } |
|
598 | } | |
599 |
|
599 | |||
600 | for target in targets: |
|
600 | for target in targets: | |
601 | entry = { |
|
601 | entry = { | |
602 | b'name': target[b'name'], |
|
602 | b'name': target[b'name'], | |
603 | b'protocol': target[b'protocol'], |
|
603 | b'protocol': target[b'protocol'], | |
604 | b'uris': target[b'uris'], |
|
604 | b'uris': target[b'uris'], | |
605 | } |
|
605 | } | |
606 |
|
606 | |||
607 | for key in (b'snirequired', b'tlsversions'): |
|
607 | for key in (b'snirequired', b'tlsversions'): | |
608 | if key in target: |
|
608 | if key in target: | |
609 | entry[key] = target[key] |
|
609 | entry[key] = target[key] | |
610 |
|
610 | |||
611 | caps[b'redirect'][b'targets'].append(entry) |
|
611 | caps[b'redirect'][b'targets'].append(entry) | |
612 |
|
612 | |||
613 | return proto.addcapabilities(repo, caps) |
|
613 | return proto.addcapabilities(repo, caps) | |
614 |
|
614 | |||
615 |
|
615 | |||
616 | def getadvertisedredirecttargets(repo, proto): |
|
616 | def getadvertisedredirecttargets(repo, proto): | |
617 | """Obtain a list of content redirect targets. |
|
617 | """Obtain a list of content redirect targets. | |
618 |
|
618 | |||
619 | Returns a list containing potential redirect targets that will be |
|
619 | Returns a list containing potential redirect targets that will be | |
620 | advertised in capabilities data. Each dict MUST have the following |
|
620 | advertised in capabilities data. Each dict MUST have the following | |
621 | keys: |
|
621 | keys: | |
622 |
|
622 | |||
623 | name |
|
623 | name | |
624 | The name of this redirect target. This is the identifier clients use |
|
624 | The name of this redirect target. This is the identifier clients use | |
625 | to refer to a target. It is transferred as part of every command |
|
625 | to refer to a target. It is transferred as part of every command | |
626 | request. |
|
626 | request. | |
627 |
|
627 | |||
628 | protocol |
|
628 | protocol | |
629 | Network protocol used by this target. Typically this is the string |
|
629 | Network protocol used by this target. Typically this is the string | |
630 | in front of the ``://`` in a URL. e.g. ``https``. |
|
630 | in front of the ``://`` in a URL. e.g. ``https``. | |
631 |
|
631 | |||
632 | uris |
|
632 | uris | |
633 | List of representative URIs for this target. Clients can use the |
|
633 | List of representative URIs for this target. Clients can use the | |
634 | URIs to test parsing for compatibility or for ordering preference |
|
634 | URIs to test parsing for compatibility or for ordering preference | |
635 | for which target to use. |
|
635 | for which target to use. | |
636 |
|
636 | |||
637 | The following optional keys are recognized: |
|
637 | The following optional keys are recognized: | |
638 |
|
638 | |||
639 | snirequired |
|
639 | snirequired | |
640 | Bool indicating if Server Name Indication (SNI) is required to |
|
640 | Bool indicating if Server Name Indication (SNI) is required to | |
641 | connect to this target. |
|
641 | connect to this target. | |
642 |
|
642 | |||
643 | tlsversions |
|
643 | tlsversions | |
644 | List of bytes indicating which TLS versions are supported by this |
|
644 | List of bytes indicating which TLS versions are supported by this | |
645 | target. |
|
645 | target. | |
646 |
|
646 | |||
647 | By default, clients reflect the target order advertised by servers |
|
647 | By default, clients reflect the target order advertised by servers | |
648 | and servers will use the first client-advertised target when picking |
|
648 | and servers will use the first client-advertised target when picking | |
649 | a redirect target. So targets should be advertised in the order the |
|
649 | a redirect target. So targets should be advertised in the order the | |
650 | server prefers they be used. |
|
650 | server prefers they be used. | |
651 | """ |
|
651 | """ | |
652 | return [] |
|
652 | return [] | |
653 |
|
653 | |||
654 |
|
654 | |||
655 | def wireprotocommand( |
|
655 | def wireprotocommand( | |
656 | name, |
|
656 | name, | |
657 | args=None, |
|
657 | args=None, | |
658 | permission=b'push', |
|
658 | permission=b'push', | |
659 | cachekeyfn=None, |
|
659 | cachekeyfn=None, | |
660 | extracapabilitiesfn=None, |
|
660 | extracapabilitiesfn=None, | |
661 | ): |
|
661 | ): | |
662 | """Decorator to declare a wire protocol command. |
|
662 | """Decorator to declare a wire protocol command. | |
663 |
|
663 | |||
664 | ``name`` is the name of the wire protocol command being provided. |
|
664 | ``name`` is the name of the wire protocol command being provided. | |
665 |
|
665 | |||
666 | ``args`` is a dict defining arguments accepted by the command. Keys are |
|
666 | ``args`` is a dict defining arguments accepted by the command. Keys are | |
667 | the argument name. Values are dicts with the following keys: |
|
667 | the argument name. Values are dicts with the following keys: | |
668 |
|
668 | |||
669 | ``type`` |
|
669 | ``type`` | |
670 | The argument data type. Must be one of the following string |
|
670 | The argument data type. Must be one of the following string | |
671 | literals: ``bytes``, ``int``, ``list``, ``dict``, ``set``, |
|
671 | literals: ``bytes``, ``int``, ``list``, ``dict``, ``set``, | |
672 | or ``bool``. |
|
672 | or ``bool``. | |
673 |
|
673 | |||
674 | ``default`` |
|
674 | ``default`` | |
675 | A callable returning the default value for this argument. If not |
|
675 | A callable returning the default value for this argument. If not | |
676 | specified, ``None`` will be the default value. |
|
676 | specified, ``None`` will be the default value. | |
677 |
|
677 | |||
678 | ``example`` |
|
678 | ``example`` | |
679 | An example value for this argument. |
|
679 | An example value for this argument. | |
680 |
|
680 | |||
681 | ``validvalues`` |
|
681 | ``validvalues`` | |
682 | Set of recognized values for this argument. |
|
682 | Set of recognized values for this argument. | |
683 |
|
683 | |||
684 | ``permission`` defines the permission type needed to run this command. |
|
684 | ``permission`` defines the permission type needed to run this command. | |
685 | Can be ``push`` or ``pull``. These roughly map to read-write and read-only, |
|
685 | Can be ``push`` or ``pull``. These roughly map to read-write and read-only, | |
686 | respectively. Default is to assume command requires ``push`` permissions |
|
686 | respectively. Default is to assume command requires ``push`` permissions | |
687 | because otherwise commands not declaring their permissions could modify |
|
687 | because otherwise commands not declaring their permissions could modify | |
688 | a repository that is supposed to be read-only. |
|
688 | a repository that is supposed to be read-only. | |
689 |
|
689 | |||
690 | ``cachekeyfn`` defines an optional callable that can derive the |
|
690 | ``cachekeyfn`` defines an optional callable that can derive the | |
691 | cache key for this request. |
|
691 | cache key for this request. | |
692 |
|
692 | |||
693 | ``extracapabilitiesfn`` defines an optional callable that defines extra |
|
693 | ``extracapabilitiesfn`` defines an optional callable that defines extra | |
694 | command capabilities/parameters that are advertised next to the command |
|
694 | command capabilities/parameters that are advertised next to the command | |
695 | in the capabilities data structure describing the server. The callable |
|
695 | in the capabilities data structure describing the server. The callable | |
696 | receives as arguments the repository and protocol objects. It returns |
|
696 | receives as arguments the repository and protocol objects. It returns | |
697 | a dict of extra fields to add to the command descriptor. |
|
697 | a dict of extra fields to add to the command descriptor. | |
698 |
|
698 | |||
699 | Wire protocol commands are generators of objects to be serialized and |
|
699 | Wire protocol commands are generators of objects to be serialized and | |
700 | sent to the client. |
|
700 | sent to the client. | |
701 |
|
701 | |||
702 | If a command raises an uncaught exception, this will be translated into |
|
702 | If a command raises an uncaught exception, this will be translated into | |
703 | a command error. |
|
703 | a command error. | |
704 |
|
704 | |||
705 | All commands can opt in to being cacheable by defining a function |
|
705 | All commands can opt in to being cacheable by defining a function | |
706 | (``cachekeyfn``) that is called to derive a cache key. This function |
|
706 | (``cachekeyfn``) that is called to derive a cache key. This function | |
707 | receives the same arguments as the command itself plus a ``cacher`` |
|
707 | receives the same arguments as the command itself plus a ``cacher`` | |
708 | argument containing the active cacher for the request and returns a bytes |
|
708 | argument containing the active cacher for the request and returns a bytes | |
709 | containing the key in a cache the response to this command may be cached |
|
709 | containing the key in a cache the response to this command may be cached | |
710 | under. |
|
710 | under. | |
711 | """ |
|
711 | """ | |
712 | transports = { |
|
712 | transports = { | |
713 | k for k, v in wireprototypes.TRANSPORTS.items() if v[b'version'] == 2 |
|
713 | k for k, v in wireprototypes.TRANSPORTS.items() if v[b'version'] == 2 | |
714 | } |
|
714 | } | |
715 |
|
715 | |||
716 | if permission not in (b'push', b'pull'): |
|
716 | if permission not in (b'push', b'pull'): | |
717 | raise error.ProgrammingError( |
|
717 | raise error.ProgrammingError( | |
718 | b'invalid wire protocol permission; ' |
|
718 | b'invalid wire protocol permission; ' | |
719 | b'got %s; expected "push" or "pull"' % permission |
|
719 | b'got %s; expected "push" or "pull"' % permission | |
720 | ) |
|
720 | ) | |
721 |
|
721 | |||
722 | if args is None: |
|
722 | if args is None: | |
723 | args = {} |
|
723 | args = {} | |
724 |
|
724 | |||
725 | if not isinstance(args, dict): |
|
725 | if not isinstance(args, dict): | |
726 | raise error.ProgrammingError( |
|
726 | raise error.ProgrammingError( | |
727 | b'arguments for version 2 commands must be declared as dicts' |
|
727 | b'arguments for version 2 commands must be declared as dicts' | |
728 | ) |
|
728 | ) | |
729 |
|
729 | |||
730 | for arg, meta in args.items(): |
|
730 | for arg, meta in args.items(): | |
731 | if arg == b'*': |
|
731 | if arg == b'*': | |
732 | raise error.ProgrammingError( |
|
732 | raise error.ProgrammingError( | |
733 | b'* argument name not allowed on version 2 commands' |
|
733 | b'* argument name not allowed on version 2 commands' | |
734 | ) |
|
734 | ) | |
735 |
|
735 | |||
736 | if not isinstance(meta, dict): |
|
736 | if not isinstance(meta, dict): | |
737 | raise error.ProgrammingError( |
|
737 | raise error.ProgrammingError( | |
738 | b'arguments for version 2 commands ' |
|
738 | b'arguments for version 2 commands ' | |
739 | b'must declare metadata as a dict' |
|
739 | b'must declare metadata as a dict' | |
740 | ) |
|
740 | ) | |
741 |
|
741 | |||
742 | if b'type' not in meta: |
|
742 | if b'type' not in meta: | |
743 | raise error.ProgrammingError( |
|
743 | raise error.ProgrammingError( | |
744 | b'%s argument for command %s does not ' |
|
744 | b'%s argument for command %s does not ' | |
745 | b'declare type field' % (arg, name) |
|
745 | b'declare type field' % (arg, name) | |
746 | ) |
|
746 | ) | |
747 |
|
747 | |||
748 | if meta[b'type'] not in ( |
|
748 | if meta[b'type'] not in ( | |
749 | b'bytes', |
|
749 | b'bytes', | |
750 | b'int', |
|
750 | b'int', | |
751 | b'list', |
|
751 | b'list', | |
752 | b'dict', |
|
752 | b'dict', | |
753 | b'set', |
|
753 | b'set', | |
754 | b'bool', |
|
754 | b'bool', | |
755 | ): |
|
755 | ): | |
756 | raise error.ProgrammingError( |
|
756 | raise error.ProgrammingError( | |
757 | b'%s argument for command %s has ' |
|
757 | b'%s argument for command %s has ' | |
758 | b'illegal type: %s' % (arg, name, meta[b'type']) |
|
758 | b'illegal type: %s' % (arg, name, meta[b'type']) | |
759 | ) |
|
759 | ) | |
760 |
|
760 | |||
761 | if b'example' not in meta: |
|
761 | if b'example' not in meta: | |
762 | raise error.ProgrammingError( |
|
762 | raise error.ProgrammingError( | |
763 | b'%s argument for command %s does not ' |
|
763 | b'%s argument for command %s does not ' | |
764 | b'declare example field' % (arg, name) |
|
764 | b'declare example field' % (arg, name) | |
765 | ) |
|
765 | ) | |
766 |
|
766 | |||
767 | meta[b'required'] = b'default' not in meta |
|
767 | meta[b'required'] = b'default' not in meta | |
768 |
|
768 | |||
769 | meta.setdefault(b'default', lambda: None) |
|
769 | meta.setdefault(b'default', lambda: None) | |
770 | meta.setdefault(b'validvalues', None) |
|
770 | meta.setdefault(b'validvalues', None) | |
771 |
|
771 | |||
772 | def register(func): |
|
772 | def register(func): | |
773 | if name in COMMANDS: |
|
773 | if name in COMMANDS: | |
774 | raise error.ProgrammingError( |
|
774 | raise error.ProgrammingError( | |
775 | b'%s command already registered for version 2' % name |
|
775 | b'%s command already registered for version 2' % name | |
776 | ) |
|
776 | ) | |
777 |
|
777 | |||
778 | COMMANDS[name] = wireprototypes.commandentry( |
|
778 | COMMANDS[name] = wireprototypes.commandentry( | |
779 | func, |
|
779 | func, | |
780 | args=args, |
|
780 | args=args, | |
781 | transports=transports, |
|
781 | transports=transports, | |
782 | permission=permission, |
|
782 | permission=permission, | |
783 | cachekeyfn=cachekeyfn, |
|
783 | cachekeyfn=cachekeyfn, | |
784 | extracapabilitiesfn=extracapabilitiesfn, |
|
784 | extracapabilitiesfn=extracapabilitiesfn, | |
785 | ) |
|
785 | ) | |
786 |
|
786 | |||
787 | return func |
|
787 | return func | |
788 |
|
788 | |||
789 | return register |
|
789 | return register | |
790 |
|
790 | |||
791 |
|
791 | |||
792 | def makecommandcachekeyfn(command, localversion=None, allargs=False): |
|
792 | def makecommandcachekeyfn(command, localversion=None, allargs=False): | |
793 | """Construct a cache key derivation function with common features. |
|
793 | """Construct a cache key derivation function with common features. | |
794 |
|
794 | |||
795 | By default, the cache key is a hash of: |
|
795 | By default, the cache key is a hash of: | |
796 |
|
796 | |||
797 | * The command name. |
|
797 | * The command name. | |
798 | * A global cache version number. |
|
798 | * A global cache version number. | |
799 | * A local cache version number (passed via ``localversion``). |
|
799 | * A local cache version number (passed via ``localversion``). | |
800 | * All the arguments passed to the command. |
|
800 | * All the arguments passed to the command. | |
801 | * The media type used. |
|
801 | * The media type used. | |
802 | * Wire protocol version string. |
|
802 | * Wire protocol version string. | |
803 | * The repository path. |
|
803 | * The repository path. | |
804 | """ |
|
804 | """ | |
805 | if not allargs: |
|
805 | if not allargs: | |
806 | raise error.ProgrammingError( |
|
806 | raise error.ProgrammingError( | |
807 | b'only allargs=True is currently supported' |
|
807 | b'only allargs=True is currently supported' | |
808 | ) |
|
808 | ) | |
809 |
|
809 | |||
810 | if localversion is None: |
|
810 | if localversion is None: | |
811 | raise error.ProgrammingError(b'must set localversion argument value') |
|
811 | raise error.ProgrammingError(b'must set localversion argument value') | |
812 |
|
812 | |||
813 | def cachekeyfn(repo, proto, cacher, **args): |
|
813 | def cachekeyfn(repo, proto, cacher, **args): | |
814 | spec = COMMANDS[command] |
|
814 | spec = COMMANDS[command] | |
815 |
|
815 | |||
816 | # Commands that mutate the repo can not be cached. |
|
816 | # Commands that mutate the repo can not be cached. | |
817 | if spec.permission == b'push': |
|
817 | if spec.permission == b'push': | |
818 | return None |
|
818 | return None | |
819 |
|
819 | |||
820 | # TODO config option to disable caching. |
|
820 | # TODO config option to disable caching. | |
821 |
|
821 | |||
822 | # Our key derivation strategy is to construct a data structure |
|
822 | # Our key derivation strategy is to construct a data structure | |
823 | # holding everything that could influence cacheability and to hash |
|
823 | # holding everything that could influence cacheability and to hash | |
824 | # the CBOR representation of that. Using CBOR seems like it might |
|
824 | # the CBOR representation of that. Using CBOR seems like it might | |
825 | # be overkill. However, simpler hashing mechanisms are prone to |
|
825 | # be overkill. However, simpler hashing mechanisms are prone to | |
826 | # duplicate input issues. e.g. if you just concatenate two values, |
|
826 | # duplicate input issues. e.g. if you just concatenate two values, | |
827 | # "foo"+"bar" is identical to "fo"+"obar". Using CBOR provides |
|
827 | # "foo"+"bar" is identical to "fo"+"obar". Using CBOR provides | |
828 | # "padding" between values and prevents these problems. |
|
828 | # "padding" between values and prevents these problems. | |
829 |
|
829 | |||
830 | # Seed the hash with various data. |
|
830 | # Seed the hash with various data. | |
831 | state = { |
|
831 | state = { | |
832 | # To invalidate all cache keys. |
|
832 | # To invalidate all cache keys. | |
833 | b'globalversion': GLOBAL_CACHE_VERSION, |
|
833 | b'globalversion': GLOBAL_CACHE_VERSION, | |
834 | # More granular cache key invalidation. |
|
834 | # More granular cache key invalidation. | |
835 | b'localversion': localversion, |
|
835 | b'localversion': localversion, | |
836 | # Cache keys are segmented by command. |
|
836 | # Cache keys are segmented by command. | |
837 | b'command': command, |
|
837 | b'command': command, | |
838 | # Throw in the media type and API version strings so changes |
|
838 | # Throw in the media type and API version strings so changes | |
839 | # to exchange semantics invalid cache. |
|
839 | # to exchange semantics invalid cache. | |
840 | b'mediatype': FRAMINGTYPE, |
|
840 | b'mediatype': FRAMINGTYPE, | |
841 | b'version': HTTP_WIREPROTO_V2, |
|
841 | b'version': HTTP_WIREPROTO_V2, | |
842 | # So same requests for different repos don't share cache keys. |
|
842 | # So same requests for different repos don't share cache keys. | |
843 | b'repo': repo.root, |
|
843 | b'repo': repo.root, | |
844 | } |
|
844 | } | |
845 |
|
845 | |||
846 | # The arguments passed to us will have already been normalized. |
|
846 | # The arguments passed to us will have already been normalized. | |
847 | # Default values will be set, etc. This is important because it |
|
847 | # Default values will be set, etc. This is important because it | |
848 | # means that it doesn't matter if clients send an explicit argument |
|
848 | # means that it doesn't matter if clients send an explicit argument | |
849 | # or rely on the default value: it will all normalize to the same |
|
849 | # or rely on the default value: it will all normalize to the same | |
850 | # set of arguments on the server and therefore the same cache key. |
|
850 | # set of arguments on the server and therefore the same cache key. | |
851 | # |
|
851 | # | |
852 | # Arguments by their very nature must support being encoded to CBOR. |
|
852 | # Arguments by their very nature must support being encoded to CBOR. | |
853 | # And the CBOR encoder is deterministic. So we hash the arguments |
|
853 | # And the CBOR encoder is deterministic. So we hash the arguments | |
854 | # by feeding the CBOR of their representation into the hasher. |
|
854 | # by feeding the CBOR of their representation into the hasher. | |
855 | if allargs: |
|
855 | if allargs: | |
856 | state[b'args'] = pycompat.byteskwargs(args) |
|
856 | state[b'args'] = pycompat.byteskwargs(args) | |
857 |
|
857 | |||
858 | cacher.adjustcachekeystate(state) |
|
858 | cacher.adjustcachekeystate(state) | |
859 |
|
859 | |||
860 | hasher = hashutil.sha1() |
|
860 | hasher = hashutil.sha1() | |
861 | for chunk in cborutil.streamencode(state): |
|
861 | for chunk in cborutil.streamencode(state): | |
862 | hasher.update(chunk) |
|
862 | hasher.update(chunk) | |
863 |
|
863 | |||
864 | return pycompat.sysbytes(hasher.hexdigest()) |
|
864 | return pycompat.sysbytes(hasher.hexdigest()) | |
865 |
|
865 | |||
866 | return cachekeyfn |
|
866 | return cachekeyfn | |
867 |
|
867 | |||
868 |
|
868 | |||
869 | def makeresponsecacher( |
|
869 | def makeresponsecacher( | |
870 | repo, proto, command, args, objencoderfn, redirecttargets, redirecthashes |
|
870 | repo, proto, command, args, objencoderfn, redirecttargets, redirecthashes | |
871 | ): |
|
871 | ): | |
872 | """Construct a cacher for a cacheable command. |
|
872 | """Construct a cacher for a cacheable command. | |
873 |
|
873 | |||
874 | Returns an ``iwireprotocolcommandcacher`` instance. |
|
874 | Returns an ``iwireprotocolcommandcacher`` instance. | |
875 |
|
875 | |||
876 | Extensions can monkeypatch this function to provide custom caching |
|
876 | Extensions can monkeypatch this function to provide custom caching | |
877 | backends. |
|
877 | backends. | |
878 | """ |
|
878 | """ | |
879 | return None |
|
879 | return None | |
880 |
|
880 | |||
881 |
|
881 | |||
882 | def resolvenodes(repo, revisions): |
|
882 | def resolvenodes(repo, revisions): | |
883 | """Resolve nodes from a revisions specifier data structure.""" |
|
883 | """Resolve nodes from a revisions specifier data structure.""" | |
884 | cl = repo.changelog |
|
884 | cl = repo.changelog | |
885 | clhasnode = cl.hasnode |
|
885 | clhasnode = cl.hasnode | |
886 |
|
886 | |||
887 | seen = set() |
|
887 | seen = set() | |
888 | nodes = [] |
|
888 | nodes = [] | |
889 |
|
889 | |||
890 | if not isinstance(revisions, list): |
|
890 | if not isinstance(revisions, list): | |
891 | raise error.WireprotoCommandError( |
|
891 | raise error.WireprotoCommandError( | |
892 | b'revisions must be defined as an array' |
|
892 | b'revisions must be defined as an array' | |
893 | ) |
|
893 | ) | |
894 |
|
894 | |||
895 | for spec in revisions: |
|
895 | for spec in revisions: | |
896 | if b'type' not in spec: |
|
896 | if b'type' not in spec: | |
897 | raise error.WireprotoCommandError( |
|
897 | raise error.WireprotoCommandError( | |
898 | b'type key not present in revision specifier' |
|
898 | b'type key not present in revision specifier' | |
899 | ) |
|
899 | ) | |
900 |
|
900 | |||
901 | typ = spec[b'type'] |
|
901 | typ = spec[b'type'] | |
902 |
|
902 | |||
903 | if typ == b'changesetexplicit': |
|
903 | if typ == b'changesetexplicit': | |
904 | if b'nodes' not in spec: |
|
904 | if b'nodes' not in spec: | |
905 | raise error.WireprotoCommandError( |
|
905 | raise error.WireprotoCommandError( | |
906 | b'nodes key not present in changesetexplicit revision ' |
|
906 | b'nodes key not present in changesetexplicit revision ' | |
907 | b'specifier' |
|
907 | b'specifier' | |
908 | ) |
|
908 | ) | |
909 |
|
909 | |||
910 | for node in spec[b'nodes']: |
|
910 | for node in spec[b'nodes']: | |
911 | if node not in seen: |
|
911 | if node not in seen: | |
912 | nodes.append(node) |
|
912 | nodes.append(node) | |
913 | seen.add(node) |
|
913 | seen.add(node) | |
914 |
|
914 | |||
915 | elif typ == b'changesetexplicitdepth': |
|
915 | elif typ == b'changesetexplicitdepth': | |
916 | for key in (b'nodes', b'depth'): |
|
916 | for key in (b'nodes', b'depth'): | |
917 | if key not in spec: |
|
917 | if key not in spec: | |
918 | raise error.WireprotoCommandError( |
|
918 | raise error.WireprotoCommandError( | |
919 | b'%s key not present in changesetexplicitdepth revision ' |
|
919 | b'%s key not present in changesetexplicitdepth revision ' | |
920 | b'specifier', |
|
920 | b'specifier', | |
921 | (key,), |
|
921 | (key,), | |
922 | ) |
|
922 | ) | |
923 |
|
923 | |||
924 | for rev in repo.revs( |
|
924 | for rev in repo.revs( | |
925 | b'ancestors(%ln, %s)', spec[b'nodes'], spec[b'depth'] - 1 |
|
925 | b'ancestors(%ln, %s)', spec[b'nodes'], spec[b'depth'] - 1 | |
926 | ): |
|
926 | ): | |
927 | node = cl.node(rev) |
|
927 | node = cl.node(rev) | |
928 |
|
928 | |||
929 | if node not in seen: |
|
929 | if node not in seen: | |
930 | nodes.append(node) |
|
930 | nodes.append(node) | |
931 | seen.add(node) |
|
931 | seen.add(node) | |
932 |
|
932 | |||
933 | elif typ == b'changesetdagrange': |
|
933 | elif typ == b'changesetdagrange': | |
934 | for key in (b'roots', b'heads'): |
|
934 | for key in (b'roots', b'heads'): | |
935 | if key not in spec: |
|
935 | if key not in spec: | |
936 | raise error.WireprotoCommandError( |
|
936 | raise error.WireprotoCommandError( | |
937 | b'%s key not present in changesetdagrange revision ' |
|
937 | b'%s key not present in changesetdagrange revision ' | |
938 | b'specifier', |
|
938 | b'specifier', | |
939 | (key,), |
|
939 | (key,), | |
940 | ) |
|
940 | ) | |
941 |
|
941 | |||
942 | if not spec[b'heads']: |
|
942 | if not spec[b'heads']: | |
943 | raise error.WireprotoCommandError( |
|
943 | raise error.WireprotoCommandError( | |
944 | b'heads key in changesetdagrange cannot be empty' |
|
944 | b'heads key in changesetdagrange cannot be empty' | |
945 | ) |
|
945 | ) | |
946 |
|
946 | |||
947 | if spec[b'roots']: |
|
947 | if spec[b'roots']: | |
948 | common = [n for n in spec[b'roots'] if clhasnode(n)] |
|
948 | common = [n for n in spec[b'roots'] if clhasnode(n)] | |
949 | else: |
|
949 | else: | |
950 | common = [repo.nullid] |
|
950 | common = [repo.nullid] | |
951 |
|
951 | |||
952 | for n in discovery.outgoing(repo, common, spec[b'heads']).missing: |
|
952 | for n in discovery.outgoing(repo, common, spec[b'heads']).missing: | |
953 | if n not in seen: |
|
953 | if n not in seen: | |
954 | nodes.append(n) |
|
954 | nodes.append(n) | |
955 | seen.add(n) |
|
955 | seen.add(n) | |
956 |
|
956 | |||
957 | else: |
|
957 | else: | |
958 | raise error.WireprotoCommandError( |
|
958 | raise error.WireprotoCommandError( | |
959 | b'unknown revision specifier type: %s', (typ,) |
|
959 | b'unknown revision specifier type: %s', (typ,) | |
960 | ) |
|
960 | ) | |
961 |
|
961 | |||
962 | return nodes |
|
962 | return nodes | |
963 |
|
963 | |||
964 |
|
964 | |||
965 | @wireprotocommand(b'branchmap', permission=b'pull') |
|
965 | @wireprotocommand(b'branchmap', permission=b'pull') | |
966 | def branchmapv2(repo, proto): |
|
966 | def branchmapv2(repo, proto): | |
967 | yield { |
|
967 | yield { | |
968 | encoding.fromlocal(k): v |
|
968 | encoding.fromlocal(k): v | |
969 | for k, v in pycompat.iteritems(repo.branchmap()) |
|
969 | for k, v in pycompat.iteritems(repo.branchmap()) | |
970 | } |
|
970 | } | |
971 |
|
971 | |||
972 |
|
972 | |||
973 | @wireprotocommand(b'capabilities', permission=b'pull') |
|
973 | @wireprotocommand(b'capabilities', permission=b'pull') | |
974 | def capabilitiesv2(repo, proto): |
|
974 | def capabilitiesv2(repo, proto): | |
975 | yield _capabilitiesv2(repo, proto) |
|
975 | yield _capabilitiesv2(repo, proto) | |
976 |
|
976 | |||
977 |
|
977 | |||
978 | @wireprotocommand( |
|
978 | @wireprotocommand( | |
979 | b'changesetdata', |
|
979 | b'changesetdata', | |
980 | args={ |
|
980 | args={ | |
981 | b'revisions': { |
|
981 | b'revisions': { | |
982 | b'type': b'list', |
|
982 | b'type': b'list', | |
983 | b'example': [ |
|
983 | b'example': [ | |
984 | { |
|
984 | { | |
985 | b'type': b'changesetexplicit', |
|
985 | b'type': b'changesetexplicit', | |
986 | b'nodes': [b'abcdef...'], |
|
986 | b'nodes': [b'abcdef...'], | |
987 | } |
|
987 | } | |
988 | ], |
|
988 | ], | |
989 | }, |
|
989 | }, | |
990 | b'fields': { |
|
990 | b'fields': { | |
991 | b'type': b'set', |
|
991 | b'type': b'set', | |
992 | b'default': set, |
|
992 | b'default': set, | |
993 | b'example': {b'parents', b'revision'}, |
|
993 | b'example': {b'parents', b'revision'}, | |
994 | b'validvalues': {b'bookmarks', b'parents', b'phase', b'revision'}, |
|
994 | b'validvalues': {b'bookmarks', b'parents', b'phase', b'revision'}, | |
995 | }, |
|
995 | }, | |
996 | }, |
|
996 | }, | |
997 | permission=b'pull', |
|
997 | permission=b'pull', | |
998 | ) |
|
998 | ) | |
999 | def changesetdata(repo, proto, revisions, fields): |
|
999 | def changesetdata(repo, proto, revisions, fields): | |
1000 | # TODO look for unknown fields and abort when they can't be serviced. |
|
1000 | # TODO look for unknown fields and abort when they can't be serviced. | |
1001 | # This could probably be validated by dispatcher using validvalues. |
|
1001 | # This could probably be validated by dispatcher using validvalues. | |
1002 |
|
1002 | |||
1003 | cl = repo.changelog |
|
1003 | cl = repo.changelog | |
1004 | outgoing = resolvenodes(repo, revisions) |
|
1004 | outgoing = resolvenodes(repo, revisions) | |
1005 | publishing = repo.publishing() |
|
1005 | publishing = repo.publishing() | |
1006 |
|
1006 | |||
1007 | if outgoing: |
|
1007 | if outgoing: | |
1008 | repo.hook(b'preoutgoing', throw=True, source=b'serve') |
|
1008 | repo.hook(b'preoutgoing', throw=True, source=b'serve') | |
1009 |
|
1009 | |||
1010 | yield { |
|
1010 | yield { | |
1011 | b'totalitems': len(outgoing), |
|
1011 | b'totalitems': len(outgoing), | |
1012 | } |
|
1012 | } | |
1013 |
|
1013 | |||
1014 | # The phases of nodes already transferred to the client may have changed |
|
1014 | # The phases of nodes already transferred to the client may have changed | |
1015 | # since the client last requested data. We send phase-only records |
|
1015 | # since the client last requested data. We send phase-only records | |
1016 | # for these revisions, if requested. |
|
1016 | # for these revisions, if requested. | |
1017 | # TODO actually do this. We'll probably want to emit phase heads |
|
1017 | # TODO actually do this. We'll probably want to emit phase heads | |
1018 | # in the ancestry set of the outgoing revisions. This will ensure |
|
1018 | # in the ancestry set of the outgoing revisions. This will ensure | |
1019 | # that phase updates within that set are seen. |
|
1019 | # that phase updates within that set are seen. | |
1020 | if b'phase' in fields: |
|
1020 | if b'phase' in fields: | |
1021 | pass |
|
1021 | pass | |
1022 |
|
1022 | |||
1023 | nodebookmarks = {} |
|
1023 | nodebookmarks = {} | |
1024 | for mark, node in repo._bookmarks.items(): |
|
1024 | for mark, node in repo._bookmarks.items(): | |
1025 | nodebookmarks.setdefault(node, set()).add(mark) |
|
1025 | nodebookmarks.setdefault(node, set()).add(mark) | |
1026 |
|
1026 | |||
1027 | # It is already topologically sorted by revision number. |
|
1027 | # It is already topologically sorted by revision number. | |
1028 | for node in outgoing: |
|
1028 | for node in outgoing: | |
1029 | d = { |
|
1029 | d = { | |
1030 | b'node': node, |
|
1030 | b'node': node, | |
1031 | } |
|
1031 | } | |
1032 |
|
1032 | |||
1033 | if b'parents' in fields: |
|
1033 | if b'parents' in fields: | |
1034 | d[b'parents'] = cl.parents(node) |
|
1034 | d[b'parents'] = cl.parents(node) | |
1035 |
|
1035 | |||
1036 | if b'phase' in fields: |
|
1036 | if b'phase' in fields: | |
1037 | if publishing: |
|
1037 | if publishing: | |
1038 | d[b'phase'] = b'public' |
|
1038 | d[b'phase'] = b'public' | |
1039 | else: |
|
1039 | else: | |
1040 | ctx = repo[node] |
|
1040 | ctx = repo[node] | |
1041 | d[b'phase'] = ctx.phasestr() |
|
1041 | d[b'phase'] = ctx.phasestr() | |
1042 |
|
1042 | |||
1043 | if b'bookmarks' in fields and node in nodebookmarks: |
|
1043 | if b'bookmarks' in fields and node in nodebookmarks: | |
1044 | d[b'bookmarks'] = sorted(nodebookmarks[node]) |
|
1044 | d[b'bookmarks'] = sorted(nodebookmarks[node]) | |
1045 | del nodebookmarks[node] |
|
1045 | del nodebookmarks[node] | |
1046 |
|
1046 | |||
1047 | followingmeta = [] |
|
1047 | followingmeta = [] | |
1048 | followingdata = [] |
|
1048 | followingdata = [] | |
1049 |
|
1049 | |||
1050 | if b'revision' in fields: |
|
1050 | if b'revision' in fields: | |
1051 | revisiondata = cl.revision(node) |
|
1051 | revisiondata = cl.revision(node) | |
1052 | followingmeta.append((b'revision', len(revisiondata))) |
|
1052 | followingmeta.append((b'revision', len(revisiondata))) | |
1053 | followingdata.append(revisiondata) |
|
1053 | followingdata.append(revisiondata) | |
1054 |
|
1054 | |||
1055 | # TODO make it possible for extensions to wrap a function or register |
|
1055 | # TODO make it possible for extensions to wrap a function or register | |
1056 | # a handler to service custom fields. |
|
1056 | # a handler to service custom fields. | |
1057 |
|
1057 | |||
1058 | if followingmeta: |
|
1058 | if followingmeta: | |
1059 | d[b'fieldsfollowing'] = followingmeta |
|
1059 | d[b'fieldsfollowing'] = followingmeta | |
1060 |
|
1060 | |||
1061 | yield d |
|
1061 | yield d | |
1062 |
|
1062 | |||
1063 | for extra in followingdata: |
|
1063 | for extra in followingdata: | |
1064 | yield extra |
|
1064 | yield extra | |
1065 |
|
1065 | |||
1066 | # If requested, send bookmarks from nodes that didn't have revision |
|
1066 | # If requested, send bookmarks from nodes that didn't have revision | |
1067 | # data sent so receiver is aware of any bookmark updates. |
|
1067 | # data sent so receiver is aware of any bookmark updates. | |
1068 | if b'bookmarks' in fields: |
|
1068 | if b'bookmarks' in fields: | |
1069 | for node, marks in sorted(pycompat.iteritems(nodebookmarks)): |
|
1069 | for node, marks in sorted(pycompat.iteritems(nodebookmarks)): | |
1070 | yield { |
|
1070 | yield { | |
1071 | b'node': node, |
|
1071 | b'node': node, | |
1072 | b'bookmarks': sorted(marks), |
|
1072 | b'bookmarks': sorted(marks), | |
1073 | } |
|
1073 | } | |
1074 |
|
1074 | |||
1075 |
|
1075 | |||
1076 | class FileAccessError(Exception): |
|
1076 | class FileAccessError(Exception): | |
1077 | """Represents an error accessing a specific file.""" |
|
1077 | """Represents an error accessing a specific file.""" | |
1078 |
|
1078 | |||
1079 | def __init__(self, path, msg, args): |
|
1079 | def __init__(self, path, msg, args): | |
1080 | self.path = path |
|
1080 | self.path = path | |
1081 | self.msg = msg |
|
1081 | self.msg = msg | |
1082 | self.args = args |
|
1082 | self.args = args | |
1083 |
|
1083 | |||
1084 |
|
1084 | |||
1085 | def getfilestore(repo, proto, path): |
|
1085 | def getfilestore(repo, proto, path): | |
1086 | """Obtain a file storage object for use with wire protocol. |
|
1086 | """Obtain a file storage object for use with wire protocol. | |
1087 |
|
1087 | |||
1088 | Exists as a standalone function so extensions can monkeypatch to add |
|
1088 | Exists as a standalone function so extensions can monkeypatch to add | |
1089 | access control. |
|
1089 | access control. | |
1090 | """ |
|
1090 | """ | |
1091 | # This seems to work even if the file doesn't exist. So catch |
|
1091 | # This seems to work even if the file doesn't exist. So catch | |
1092 | # "empty" files and return an error. |
|
1092 | # "empty" files and return an error. | |
1093 | fl = repo.file(path) |
|
1093 | fl = repo.file(path) | |
1094 |
|
1094 | |||
1095 | if not len(fl): |
|
1095 | if not len(fl): | |
1096 | raise FileAccessError(path, b'unknown file: %s', (path,)) |
|
1096 | raise FileAccessError(path, b'unknown file: %s', (path,)) | |
1097 |
|
1097 | |||
1098 | return fl |
|
1098 | return fl | |
1099 |
|
1099 | |||
1100 |
|
1100 | |||
1101 | def emitfilerevisions(repo, path, revisions, linknodes, fields): |
|
1101 | def emitfilerevisions(repo, path, revisions, linknodes, fields): | |
1102 | for revision in revisions: |
|
1102 | for revision in revisions: | |
1103 | d = { |
|
1103 | d = { | |
1104 | b'node': revision.node, |
|
1104 | b'node': revision.node, | |
1105 | } |
|
1105 | } | |
1106 |
|
1106 | |||
1107 | if b'parents' in fields: |
|
1107 | if b'parents' in fields: | |
1108 | d[b'parents'] = [revision.p1node, revision.p2node] |
|
1108 | d[b'parents'] = [revision.p1node, revision.p2node] | |
1109 |
|
1109 | |||
1110 | if b'linknode' in fields: |
|
1110 | if b'linknode' in fields: | |
1111 | d[b'linknode'] = linknodes[revision.node] |
|
1111 | d[b'linknode'] = linknodes[revision.node] | |
1112 |
|
1112 | |||
1113 | followingmeta = [] |
|
1113 | followingmeta = [] | |
1114 | followingdata = [] |
|
1114 | followingdata = [] | |
1115 |
|
1115 | |||
1116 | if b'revision' in fields: |
|
1116 | if b'revision' in fields: | |
1117 | if revision.revision is not None: |
|
1117 | if revision.revision is not None: | |
1118 | followingmeta.append((b'revision', len(revision.revision))) |
|
1118 | followingmeta.append((b'revision', len(revision.revision))) | |
1119 | followingdata.append(revision.revision) |
|
1119 | followingdata.append(revision.revision) | |
1120 | else: |
|
1120 | else: | |
1121 | d[b'deltabasenode'] = revision.basenode |
|
1121 | d[b'deltabasenode'] = revision.basenode | |
1122 | followingmeta.append((b'delta', len(revision.delta))) |
|
1122 | followingmeta.append((b'delta', len(revision.delta))) | |
1123 | followingdata.append(revision.delta) |
|
1123 | followingdata.append(revision.delta) | |
1124 |
|
1124 | |||
1125 | if followingmeta: |
|
1125 | if followingmeta: | |
1126 | d[b'fieldsfollowing'] = followingmeta |
|
1126 | d[b'fieldsfollowing'] = followingmeta | |
1127 |
|
1127 | |||
1128 | yield d |
|
1128 | yield d | |
1129 |
|
1129 | |||
1130 | for extra in followingdata: |
|
1130 | for extra in followingdata: | |
1131 | yield extra |
|
1131 | yield extra | |
1132 |
|
1132 | |||
1133 |
|
1133 | |||
1134 | def makefilematcher(repo, pathfilter): |
|
1134 | def makefilematcher(repo, pathfilter): | |
1135 | """Construct a matcher from a path filter dict.""" |
|
1135 | """Construct a matcher from a path filter dict.""" | |
1136 |
|
1136 | |||
1137 | # Validate values. |
|
1137 | # Validate values. | |
1138 | if pathfilter: |
|
1138 | if pathfilter: | |
1139 | for key in (b'include', b'exclude'): |
|
1139 | for key in (b'include', b'exclude'): | |
1140 | for pattern in pathfilter.get(key, []): |
|
1140 | for pattern in pathfilter.get(key, []): | |
1141 | if not pattern.startswith((b'path:', b'rootfilesin:')): |
|
1141 | if not pattern.startswith((b'path:', b'rootfilesin:')): | |
1142 | raise error.WireprotoCommandError( |
|
1142 | raise error.WireprotoCommandError( | |
1143 | b'%s pattern must begin with `path:` or `rootfilesin:`; ' |
|
1143 | b'%s pattern must begin with `path:` or `rootfilesin:`; ' | |
1144 | b'got %s', |
|
1144 | b'got %s', | |
1145 | (key, pattern), |
|
1145 | (key, pattern), | |
1146 | ) |
|
1146 | ) | |
1147 |
|
1147 | |||
1148 | if pathfilter: |
|
1148 | if pathfilter: | |
1149 | matcher = matchmod.match( |
|
1149 | matcher = matchmod.match( | |
1150 | repo.root, |
|
1150 | repo.root, | |
1151 | b'', |
|
1151 | b'', | |
1152 | include=pathfilter.get(b'include', []), |
|
1152 | include=pathfilter.get(b'include', []), | |
1153 | exclude=pathfilter.get(b'exclude', []), |
|
1153 | exclude=pathfilter.get(b'exclude', []), | |
1154 | ) |
|
1154 | ) | |
1155 | else: |
|
1155 | else: | |
1156 | matcher = matchmod.match(repo.root, b'') |
|
1156 | matcher = matchmod.match(repo.root, b'') | |
1157 |
|
1157 | |||
1158 | # Requested patterns could include files not in the local store. So |
|
1158 | # Requested patterns could include files not in the local store. So | |
1159 | # filter those out. |
|
1159 | # filter those out. | |
1160 | return repo.narrowmatch(matcher) |
|
1160 | return repo.narrowmatch(matcher) | |
1161 |
|
1161 | |||
1162 |
|
1162 | |||
1163 | @wireprotocommand( |
|
1163 | @wireprotocommand( | |
1164 | b'filedata', |
|
1164 | b'filedata', | |
1165 | args={ |
|
1165 | args={ | |
1166 | b'haveparents': { |
|
1166 | b'haveparents': { | |
1167 | b'type': b'bool', |
|
1167 | b'type': b'bool', | |
1168 | b'default': lambda: False, |
|
1168 | b'default': lambda: False, | |
1169 | b'example': True, |
|
1169 | b'example': True, | |
1170 | }, |
|
1170 | }, | |
1171 | b'nodes': { |
|
1171 | b'nodes': { | |
1172 | b'type': b'list', |
|
1172 | b'type': b'list', | |
1173 | b'example': [b'0123456...'], |
|
1173 | b'example': [b'0123456...'], | |
1174 | }, |
|
1174 | }, | |
1175 | b'fields': { |
|
1175 | b'fields': { | |
1176 | b'type': b'set', |
|
1176 | b'type': b'set', | |
1177 | b'default': set, |
|
1177 | b'default': set, | |
1178 | b'example': {b'parents', b'revision'}, |
|
1178 | b'example': {b'parents', b'revision'}, | |
1179 | b'validvalues': {b'parents', b'revision', b'linknode'}, |
|
1179 | b'validvalues': {b'parents', b'revision', b'linknode'}, | |
1180 | }, |
|
1180 | }, | |
1181 | b'path': { |
|
1181 | b'path': { | |
1182 | b'type': b'bytes', |
|
1182 | b'type': b'bytes', | |
1183 | b'example': b'foo.txt', |
|
1183 | b'example': b'foo.txt', | |
1184 | }, |
|
1184 | }, | |
1185 | }, |
|
1185 | }, | |
1186 | permission=b'pull', |
|
1186 | permission=b'pull', | |
1187 | # TODO censoring a file revision won't invalidate the cache. |
|
1187 | # TODO censoring a file revision won't invalidate the cache. | |
1188 | # Figure out a way to take censoring into account when deriving |
|
1188 | # Figure out a way to take censoring into account when deriving | |
1189 | # the cache key. |
|
1189 | # the cache key. | |
1190 | cachekeyfn=makecommandcachekeyfn(b'filedata', 1, allargs=True), |
|
1190 | cachekeyfn=makecommandcachekeyfn(b'filedata', 1, allargs=True), | |
1191 | ) |
|
1191 | ) | |
1192 | def filedata(repo, proto, haveparents, nodes, fields, path): |
|
1192 | def filedata(repo, proto, haveparents, nodes, fields, path): | |
1193 | # TODO this API allows access to file revisions that are attached to |
|
1193 | # TODO this API allows access to file revisions that are attached to | |
1194 | # secret changesets. filesdata does not have this problem. Maybe this |
|
1194 | # secret changesets. filesdata does not have this problem. Maybe this | |
1195 | # API should be deleted? |
|
1195 | # API should be deleted? | |
1196 |
|
1196 | |||
1197 | try: |
|
1197 | try: | |
1198 | # Extensions may wish to access the protocol handler. |
|
1198 | # Extensions may wish to access the protocol handler. | |
1199 | store = getfilestore(repo, proto, path) |
|
1199 | store = getfilestore(repo, proto, path) | |
1200 | except FileAccessError as e: |
|
1200 | except FileAccessError as e: | |
1201 | raise error.WireprotoCommandError(e.msg, e.args) |
|
1201 | raise error.WireprotoCommandError(e.msg, e.args) | |
1202 |
|
1202 | |||
1203 | clnode = repo.changelog.node |
|
1203 | clnode = repo.changelog.node | |
1204 | linknodes = {} |
|
1204 | linknodes = {} | |
1205 |
|
1205 | |||
1206 | # Validate requested nodes. |
|
1206 | # Validate requested nodes. | |
1207 | for node in nodes: |
|
1207 | for node in nodes: | |
1208 | try: |
|
1208 | try: | |
1209 | store.rev(node) |
|
1209 | store.rev(node) | |
1210 | except error.LookupError: |
|
1210 | except error.LookupError: | |
1211 | raise error.WireprotoCommandError( |
|
1211 | raise error.WireprotoCommandError( | |
1212 | b'unknown file node: %s', (hex(node),) |
|
1212 | b'unknown file node: %s', (hex(node),) | |
1213 | ) |
|
1213 | ) | |
1214 |
|
1214 | |||
1215 | # TODO by creating the filectx against a specific file revision |
|
1215 | # TODO by creating the filectx against a specific file revision | |
1216 | # instead of changeset, linkrev() is always used. This is wrong for |
|
1216 | # instead of changeset, linkrev() is always used. This is wrong for | |
1217 | # cases where linkrev() may refer to a hidden changeset. But since this |
|
1217 | # cases where linkrev() may refer to a hidden changeset. But since this | |
1218 | # API doesn't know anything about changesets, we're not sure how to |
|
1218 | # API doesn't know anything about changesets, we're not sure how to | |
1219 | # disambiguate the linknode. Perhaps we should delete this API? |
|
1219 | # disambiguate the linknode. Perhaps we should delete this API? | |
1220 | fctx = repo.filectx(path, fileid=node) |
|
1220 | fctx = repo.filectx(path, fileid=node) | |
1221 | linknodes[node] = clnode(fctx.introrev()) |
|
1221 | linknodes[node] = clnode(fctx.introrev()) | |
1222 |
|
1222 | |||
1223 | revisions = store.emitrevisions( |
|
1223 | revisions = store.emitrevisions( | |
1224 | nodes, |
|
1224 | nodes, | |
1225 | revisiondata=b'revision' in fields, |
|
1225 | revisiondata=b'revision' in fields, | |
1226 | assumehaveparentrevisions=haveparents, |
|
1226 | assumehaveparentrevisions=haveparents, | |
1227 | ) |
|
1227 | ) | |
1228 |
|
1228 | |||
1229 | yield { |
|
1229 | yield { | |
1230 | b'totalitems': len(nodes), |
|
1230 | b'totalitems': len(nodes), | |
1231 | } |
|
1231 | } | |
1232 |
|
1232 | |||
1233 | for o in emitfilerevisions(repo, path, revisions, linknodes, fields): |
|
1233 | for o in emitfilerevisions(repo, path, revisions, linknodes, fields): | |
1234 | yield o |
|
1234 | yield o | |
1235 |
|
1235 | |||
1236 |
|
1236 | |||
1237 | def filesdatacapabilities(repo, proto): |
|
1237 | def filesdatacapabilities(repo, proto): | |
1238 | batchsize = repo.ui.configint( |
|
1238 | batchsize = repo.ui.configint( | |
1239 | b'experimental', b'server.filesdata.recommended-batch-size' |
|
1239 | b'experimental', b'server.filesdata.recommended-batch-size' | |
1240 | ) |
|
1240 | ) | |
1241 | return { |
|
1241 | return { | |
1242 | b'recommendedbatchsize': batchsize, |
|
1242 | b'recommendedbatchsize': batchsize, | |
1243 | } |
|
1243 | } | |
1244 |
|
1244 | |||
1245 |
|
1245 | |||
1246 | @wireprotocommand( |
|
1246 | @wireprotocommand( | |
1247 | b'filesdata', |
|
1247 | b'filesdata', | |
1248 | args={ |
|
1248 | args={ | |
1249 | b'haveparents': { |
|
1249 | b'haveparents': { | |
1250 | b'type': b'bool', |
|
1250 | b'type': b'bool', | |
1251 | b'default': lambda: False, |
|
1251 | b'default': lambda: False, | |
1252 | b'example': True, |
|
1252 | b'example': True, | |
1253 | }, |
|
1253 | }, | |
1254 | b'fields': { |
|
1254 | b'fields': { | |
1255 | b'type': b'set', |
|
1255 | b'type': b'set', | |
1256 | b'default': set, |
|
1256 | b'default': set, | |
1257 | b'example': {b'parents', b'revision'}, |
|
1257 | b'example': {b'parents', b'revision'}, | |
1258 | b'validvalues': { |
|
1258 | b'validvalues': { | |
1259 | b'firstchangeset', |
|
1259 | b'firstchangeset', | |
1260 | b'linknode', |
|
1260 | b'linknode', | |
1261 | b'parents', |
|
1261 | b'parents', | |
1262 | b'revision', |
|
1262 | b'revision', | |
1263 | }, |
|
1263 | }, | |
1264 | }, |
|
1264 | }, | |
1265 | b'pathfilter': { |
|
1265 | b'pathfilter': { | |
1266 | b'type': b'dict', |
|
1266 | b'type': b'dict', | |
1267 | b'default': lambda: None, |
|
1267 | b'default': lambda: None, | |
1268 | b'example': {b'include': [b'path:tests']}, |
|
1268 | b'example': {b'include': [b'path:tests']}, | |
1269 | }, |
|
1269 | }, | |
1270 | b'revisions': { |
|
1270 | b'revisions': { | |
1271 | b'type': b'list', |
|
1271 | b'type': b'list', | |
1272 | b'example': [ |
|
1272 | b'example': [ | |
1273 | { |
|
1273 | { | |
1274 | b'type': b'changesetexplicit', |
|
1274 | b'type': b'changesetexplicit', | |
1275 | b'nodes': [b'abcdef...'], |
|
1275 | b'nodes': [b'abcdef...'], | |
1276 | } |
|
1276 | } | |
1277 | ], |
|
1277 | ], | |
1278 | }, |
|
1278 | }, | |
1279 | }, |
|
1279 | }, | |
1280 | permission=b'pull', |
|
1280 | permission=b'pull', | |
1281 | # TODO censoring a file revision won't invalidate the cache. |
|
1281 | # TODO censoring a file revision won't invalidate the cache. | |
1282 | # Figure out a way to take censoring into account when deriving |
|
1282 | # Figure out a way to take censoring into account when deriving | |
1283 | # the cache key. |
|
1283 | # the cache key. | |
1284 | cachekeyfn=makecommandcachekeyfn(b'filesdata', 1, allargs=True), |
|
1284 | cachekeyfn=makecommandcachekeyfn(b'filesdata', 1, allargs=True), | |
1285 | extracapabilitiesfn=filesdatacapabilities, |
|
1285 | extracapabilitiesfn=filesdatacapabilities, | |
1286 | ) |
|
1286 | ) | |
1287 | def filesdata(repo, proto, haveparents, fields, pathfilter, revisions): |
|
1287 | def filesdata(repo, proto, haveparents, fields, pathfilter, revisions): | |
1288 | # TODO This should operate on a repo that exposes obsolete changesets. There |
|
1288 | # TODO This should operate on a repo that exposes obsolete changesets. There | |
1289 | # is a race between a client making a push that obsoletes a changeset and |
|
1289 | # is a race between a client making a push that obsoletes a changeset and | |
1290 | # another client fetching files data for that changeset. If a client has a |
|
1290 | # another client fetching files data for that changeset. If a client has a | |
1291 | # changeset, it should probably be allowed to access files data for that |
|
1291 | # changeset, it should probably be allowed to access files data for that | |
1292 | # changeset. |
|
1292 | # changeset. | |
1293 |
|
1293 | |||
1294 | outgoing = resolvenodes(repo, revisions) |
|
1294 | outgoing = resolvenodes(repo, revisions) | |
1295 | filematcher = makefilematcher(repo, pathfilter) |
|
1295 | filematcher = makefilematcher(repo, pathfilter) | |
1296 |
|
1296 | |||
1297 | # path -> {fnode: linknode} |
|
1297 | # path -> {fnode: linknode} | |
1298 | fnodes = collections.defaultdict(dict) |
|
1298 | fnodes = collections.defaultdict(dict) | |
1299 |
|
1299 | |||
1300 | # We collect the set of relevant file revisions by iterating the changeset |
|
1300 | # We collect the set of relevant file revisions by iterating the changeset | |
1301 | # revisions and either walking the set of files recorded in the changeset |
|
1301 | # revisions and either walking the set of files recorded in the changeset | |
1302 | # or by walking the manifest at that revision. There is probably room for a |
|
1302 | # or by walking the manifest at that revision. There is probably room for a | |
1303 | # storage-level API to request this data, as it can be expensive to compute |
|
1303 | # storage-level API to request this data, as it can be expensive to compute | |
1304 | # and would benefit from caching or alternate storage from what revlogs |
|
1304 | # and would benefit from caching or alternate storage from what revlogs | |
1305 | # provide. |
|
1305 | # provide. | |
1306 | for node in outgoing: |
|
1306 | for node in outgoing: | |
1307 | ctx = repo[node] |
|
1307 | ctx = repo[node] | |
1308 | mctx = ctx.manifestctx() |
|
1308 | mctx = ctx.manifestctx() | |
1309 | md = mctx.read() |
|
1309 | md = mctx.read() | |
1310 |
|
1310 | |||
1311 | if haveparents: |
|
1311 | if haveparents: | |
1312 | checkpaths = ctx.files() |
|
1312 | checkpaths = ctx.files() | |
1313 | else: |
|
1313 | else: | |
1314 | checkpaths = md.keys() |
|
1314 | checkpaths = md.keys() | |
1315 |
|
1315 | |||
1316 | for path in checkpaths: |
|
1316 | for path in checkpaths: | |
1317 | fnode = md[path] |
|
1317 | fnode = md[path] | |
1318 |
|
1318 | |||
1319 | if path in fnodes and fnode in fnodes[path]: |
|
1319 | if path in fnodes and fnode in fnodes[path]: | |
1320 | continue |
|
1320 | continue | |
1321 |
|
1321 | |||
1322 | if not filematcher(path): |
|
1322 | if not filematcher(path): | |
1323 | continue |
|
1323 | continue | |
1324 |
|
1324 | |||
1325 | fnodes[path].setdefault(fnode, node) |
|
1325 | fnodes[path].setdefault(fnode, node) | |
1326 |
|
1326 | |||
1327 | yield { |
|
1327 | yield { | |
1328 | b'totalpaths': len(fnodes), |
|
1328 | b'totalpaths': len(fnodes), | |
1329 | b'totalitems': sum(len(v) for v in fnodes.values()), |
|
1329 | b'totalitems': sum(len(v) for v in fnodes.values()), | |
1330 | } |
|
1330 | } | |
1331 |
|
1331 | |||
1332 | for path, filenodes in sorted(fnodes.items()): |
|
1332 | for path, filenodes in sorted(fnodes.items()): | |
1333 | try: |
|
1333 | try: | |
1334 | store = getfilestore(repo, proto, path) |
|
1334 | store = getfilestore(repo, proto, path) | |
1335 | except FileAccessError as e: |
|
1335 | except FileAccessError as e: | |
1336 | raise error.WireprotoCommandError(e.msg, e.args) |
|
1336 | raise error.WireprotoCommandError(e.msg, e.args) | |
1337 |
|
1337 | |||
1338 | yield { |
|
1338 | yield { | |
1339 | b'path': path, |
|
1339 | b'path': path, | |
1340 | b'totalitems': len(filenodes), |
|
1340 | b'totalitems': len(filenodes), | |
1341 | } |
|
1341 | } | |
1342 |
|
1342 | |||
1343 | revisions = store.emitrevisions( |
|
1343 | revisions = store.emitrevisions( | |
1344 | filenodes.keys(), |
|
1344 | filenodes.keys(), | |
1345 | revisiondata=b'revision' in fields, |
|
1345 | revisiondata=b'revision' in fields, | |
1346 | assumehaveparentrevisions=haveparents, |
|
1346 | assumehaveparentrevisions=haveparents, | |
1347 | ) |
|
1347 | ) | |
1348 |
|
1348 | |||
1349 | for o in emitfilerevisions(repo, path, revisions, filenodes, fields): |
|
1349 | for o in emitfilerevisions(repo, path, revisions, filenodes, fields): | |
1350 | yield o |
|
1350 | yield o | |
1351 |
|
1351 | |||
1352 |
|
1352 | |||
1353 | @wireprotocommand( |
|
1353 | @wireprotocommand( | |
1354 | b'heads', |
|
1354 | b'heads', | |
1355 | args={ |
|
1355 | args={ | |
1356 | b'publiconly': { |
|
1356 | b'publiconly': { | |
1357 | b'type': b'bool', |
|
1357 | b'type': b'bool', | |
1358 | b'default': lambda: False, |
|
1358 | b'default': lambda: False, | |
1359 | b'example': False, |
|
1359 | b'example': False, | |
1360 | }, |
|
1360 | }, | |
1361 | }, |
|
1361 | }, | |
1362 | permission=b'pull', |
|
1362 | permission=b'pull', | |
1363 | ) |
|
1363 | ) | |
1364 | def headsv2(repo, proto, publiconly): |
|
1364 | def headsv2(repo, proto, publiconly): | |
1365 | if publiconly: |
|
1365 | if publiconly: | |
1366 | repo = repo.filtered(b'immutable') |
|
1366 | repo = repo.filtered(b'immutable') | |
1367 |
|
1367 | |||
1368 | yield repo.heads() |
|
1368 | yield repo.heads() | |
1369 |
|
1369 | |||
1370 |
|
1370 | |||
1371 | @wireprotocommand( |
|
1371 | @wireprotocommand( | |
1372 | b'known', |
|
1372 | b'known', | |
1373 | args={ |
|
1373 | args={ | |
1374 | b'nodes': { |
|
1374 | b'nodes': { | |
1375 | b'type': b'list', |
|
1375 | b'type': b'list', | |
1376 | b'default': list, |
|
1376 | b'default': list, | |
1377 | b'example': [b'deadbeef'], |
|
1377 | b'example': [b'deadbeef'], | |
1378 | }, |
|
1378 | }, | |
1379 | }, |
|
1379 | }, | |
1380 | permission=b'pull', |
|
1380 | permission=b'pull', | |
1381 | ) |
|
1381 | ) | |
1382 | def knownv2(repo, proto, nodes): |
|
1382 | def knownv2(repo, proto, nodes): | |
1383 | result = b''.join(b'1' if n else b'0' for n in repo.known(nodes)) |
|
1383 | result = b''.join(b'1' if n else b'0' for n in repo.known(nodes)) | |
1384 | yield result |
|
1384 | yield result | |
1385 |
|
1385 | |||
1386 |
|
1386 | |||
1387 | @wireprotocommand( |
|
1387 | @wireprotocommand( | |
1388 | b'listkeys', |
|
1388 | b'listkeys', | |
1389 | args={ |
|
1389 | args={ | |
1390 | b'namespace': { |
|
1390 | b'namespace': { | |
1391 | b'type': b'bytes', |
|
1391 | b'type': b'bytes', | |
1392 | b'example': b'ns', |
|
1392 | b'example': b'ns', | |
1393 | }, |
|
1393 | }, | |
1394 | }, |
|
1394 | }, | |
1395 | permission=b'pull', |
|
1395 | permission=b'pull', | |
1396 | ) |
|
1396 | ) | |
1397 | def listkeysv2(repo, proto, namespace): |
|
1397 | def listkeysv2(repo, proto, namespace): | |
1398 | keys = repo.listkeys(encoding.tolocal(namespace)) |
|
1398 | keys = repo.listkeys(encoding.tolocal(namespace)) | |
1399 | keys = { |
|
1399 | keys = { | |
1400 | encoding.fromlocal(k): encoding.fromlocal(v) |
|
1400 | encoding.fromlocal(k): encoding.fromlocal(v) | |
1401 | for k, v in pycompat.iteritems(keys) |
|
1401 | for k, v in pycompat.iteritems(keys) | |
1402 | } |
|
1402 | } | |
1403 |
|
1403 | |||
1404 | yield keys |
|
1404 | yield keys | |
1405 |
|
1405 | |||
1406 |
|
1406 | |||
1407 | @wireprotocommand( |
|
1407 | @wireprotocommand( | |
1408 | b'lookup', |
|
1408 | b'lookup', | |
1409 | args={ |
|
1409 | args={ | |
1410 | b'key': { |
|
1410 | b'key': { | |
1411 | b'type': b'bytes', |
|
1411 | b'type': b'bytes', | |
1412 | b'example': b'foo', |
|
1412 | b'example': b'foo', | |
1413 | }, |
|
1413 | }, | |
1414 | }, |
|
1414 | }, | |
1415 | permission=b'pull', |
|
1415 | permission=b'pull', | |
1416 | ) |
|
1416 | ) | |
1417 | def lookupv2(repo, proto, key): |
|
1417 | def lookupv2(repo, proto, key): | |
1418 | key = encoding.tolocal(key) |
|
1418 | key = encoding.tolocal(key) | |
1419 |
|
1419 | |||
1420 | # TODO handle exception. |
|
1420 | # TODO handle exception. | |
1421 | node = repo.lookup(key) |
|
1421 | node = repo.lookup(key) | |
1422 |
|
1422 | |||
1423 | yield node |
|
1423 | yield node | |
1424 |
|
1424 | |||
1425 |
|
1425 | |||
1426 | def manifestdatacapabilities(repo, proto): |
|
1426 | def manifestdatacapabilities(repo, proto): | |
1427 | batchsize = repo.ui.configint( |
|
1427 | batchsize = repo.ui.configint( | |
1428 | b'experimental', b'server.manifestdata.recommended-batch-size' |
|
1428 | b'experimental', b'server.manifestdata.recommended-batch-size' | |
1429 | ) |
|
1429 | ) | |
1430 |
|
1430 | |||
1431 | return { |
|
1431 | return { | |
1432 | b'recommendedbatchsize': batchsize, |
|
1432 | b'recommendedbatchsize': batchsize, | |
1433 | } |
|
1433 | } | |
1434 |
|
1434 | |||
1435 |
|
1435 | |||
1436 | @wireprotocommand( |
|
1436 | @wireprotocommand( | |
1437 | b'manifestdata', |
|
1437 | b'manifestdata', | |
1438 | args={ |
|
1438 | args={ | |
1439 | b'nodes': { |
|
1439 | b'nodes': { | |
1440 | b'type': b'list', |
|
1440 | b'type': b'list', | |
1441 | b'example': [b'0123456...'], |
|
1441 | b'example': [b'0123456...'], | |
1442 | }, |
|
1442 | }, | |
1443 | b'haveparents': { |
|
1443 | b'haveparents': { | |
1444 | b'type': b'bool', |
|
1444 | b'type': b'bool', | |
1445 | b'default': lambda: False, |
|
1445 | b'default': lambda: False, | |
1446 | b'example': True, |
|
1446 | b'example': True, | |
1447 | }, |
|
1447 | }, | |
1448 | b'fields': { |
|
1448 | b'fields': { | |
1449 | b'type': b'set', |
|
1449 | b'type': b'set', | |
1450 | b'default': set, |
|
1450 | b'default': set, | |
1451 | b'example': {b'parents', b'revision'}, |
|
1451 | b'example': {b'parents', b'revision'}, | |
1452 | b'validvalues': {b'parents', b'revision'}, |
|
1452 | b'validvalues': {b'parents', b'revision'}, | |
1453 | }, |
|
1453 | }, | |
1454 | b'tree': { |
|
1454 | b'tree': { | |
1455 | b'type': b'bytes', |
|
1455 | b'type': b'bytes', | |
1456 | b'example': b'', |
|
1456 | b'example': b'', | |
1457 | }, |
|
1457 | }, | |
1458 | }, |
|
1458 | }, | |
1459 | permission=b'pull', |
|
1459 | permission=b'pull', | |
1460 | cachekeyfn=makecommandcachekeyfn(b'manifestdata', 1, allargs=True), |
|
1460 | cachekeyfn=makecommandcachekeyfn(b'manifestdata', 1, allargs=True), | |
1461 | extracapabilitiesfn=manifestdatacapabilities, |
|
1461 | extracapabilitiesfn=manifestdatacapabilities, | |
1462 | ) |
|
1462 | ) | |
1463 | def manifestdata(repo, proto, haveparents, nodes, fields, tree): |
|
1463 | def manifestdata(repo, proto, haveparents, nodes, fields, tree): | |
1464 | store = repo.manifestlog.getstorage(tree) |
|
1464 | store = repo.manifestlog.getstorage(tree) | |
1465 |
|
1465 | |||
1466 | # Validate the node is known and abort on unknown revisions. |
|
1466 | # Validate the node is known and abort on unknown revisions. | |
1467 | for node in nodes: |
|
1467 | for node in nodes: | |
1468 | try: |
|
1468 | try: | |
1469 | store.rev(node) |
|
1469 | store.rev(node) | |
1470 | except error.LookupError: |
|
1470 | except error.LookupError: | |
1471 | raise error.WireprotoCommandError(b'unknown node: %s', (node,)) |
|
1471 | raise error.WireprotoCommandError(b'unknown node: %s', (node,)) | |
1472 |
|
1472 | |||
1473 | revisions = store.emitrevisions( |
|
1473 | revisions = store.emitrevisions( | |
1474 | nodes, |
|
1474 | nodes, | |
1475 | revisiondata=b'revision' in fields, |
|
1475 | revisiondata=b'revision' in fields, | |
1476 | assumehaveparentrevisions=haveparents, |
|
1476 | assumehaveparentrevisions=haveparents, | |
1477 | ) |
|
1477 | ) | |
1478 |
|
1478 | |||
1479 | yield { |
|
1479 | yield { | |
1480 | b'totalitems': len(nodes), |
|
1480 | b'totalitems': len(nodes), | |
1481 | } |
|
1481 | } | |
1482 |
|
1482 | |||
1483 | for revision in revisions: |
|
1483 | for revision in revisions: | |
1484 | d = { |
|
1484 | d = { | |
1485 | b'node': revision.node, |
|
1485 | b'node': revision.node, | |
1486 | } |
|
1486 | } | |
1487 |
|
1487 | |||
1488 | if b'parents' in fields: |
|
1488 | if b'parents' in fields: | |
1489 | d[b'parents'] = [revision.p1node, revision.p2node] |
|
1489 | d[b'parents'] = [revision.p1node, revision.p2node] | |
1490 |
|
1490 | |||
1491 | followingmeta = [] |
|
1491 | followingmeta = [] | |
1492 | followingdata = [] |
|
1492 | followingdata = [] | |
1493 |
|
1493 | |||
1494 | if b'revision' in fields: |
|
1494 | if b'revision' in fields: | |
1495 | if revision.revision is not None: |
|
1495 | if revision.revision is not None: | |
1496 | followingmeta.append((b'revision', len(revision.revision))) |
|
1496 | followingmeta.append((b'revision', len(revision.revision))) | |
1497 | followingdata.append(revision.revision) |
|
1497 | followingdata.append(revision.revision) | |
1498 | else: |
|
1498 | else: | |
1499 | d[b'deltabasenode'] = revision.basenode |
|
1499 | d[b'deltabasenode'] = revision.basenode | |
1500 | followingmeta.append((b'delta', len(revision.delta))) |
|
1500 | followingmeta.append((b'delta', len(revision.delta))) | |
1501 | followingdata.append(revision.delta) |
|
1501 | followingdata.append(revision.delta) | |
1502 |
|
1502 | |||
1503 | if followingmeta: |
|
1503 | if followingmeta: | |
1504 | d[b'fieldsfollowing'] = followingmeta |
|
1504 | d[b'fieldsfollowing'] = followingmeta | |
1505 |
|
1505 | |||
1506 | yield d |
|
1506 | yield d | |
1507 |
|
1507 | |||
1508 | for extra in followingdata: |
|
1508 | for extra in followingdata: | |
1509 | yield extra |
|
1509 | yield extra | |
1510 |
|
1510 | |||
1511 |
|
1511 | |||
1512 | @wireprotocommand( |
|
1512 | @wireprotocommand( | |
1513 | b'pushkey', |
|
1513 | b'pushkey', | |
1514 | args={ |
|
1514 | args={ | |
1515 | b'namespace': { |
|
1515 | b'namespace': { | |
1516 | b'type': b'bytes', |
|
1516 | b'type': b'bytes', | |
1517 | b'example': b'ns', |
|
1517 | b'example': b'ns', | |
1518 | }, |
|
1518 | }, | |
1519 | b'key': { |
|
1519 | b'key': { | |
1520 | b'type': b'bytes', |
|
1520 | b'type': b'bytes', | |
1521 | b'example': b'key', |
|
1521 | b'example': b'key', | |
1522 | }, |
|
1522 | }, | |
1523 | b'old': { |
|
1523 | b'old': { | |
1524 | b'type': b'bytes', |
|
1524 | b'type': b'bytes', | |
1525 | b'example': b'old', |
|
1525 | b'example': b'old', | |
1526 | }, |
|
1526 | }, | |
1527 | b'new': { |
|
1527 | b'new': { | |
1528 | b'type': b'bytes', |
|
1528 | b'type': b'bytes', | |
1529 | b'example': b'new', |
|
1529 | b'example': b'new', | |
1530 | }, |
|
1530 | }, | |
1531 | }, |
|
1531 | }, | |
1532 | permission=b'push', |
|
1532 | permission=b'push', | |
1533 | ) |
|
1533 | ) | |
1534 | def pushkeyv2(repo, proto, namespace, key, old, new): |
|
1534 | def pushkeyv2(repo, proto, namespace, key, old, new): | |
1535 | # TODO handle ui output redirection |
|
1535 | # TODO handle ui output redirection | |
1536 | yield repo.pushkey( |
|
1536 | yield repo.pushkey( | |
1537 | encoding.tolocal(namespace), |
|
1537 | encoding.tolocal(namespace), | |
1538 | encoding.tolocal(key), |
|
1538 | encoding.tolocal(key), | |
1539 | encoding.tolocal(old), |
|
1539 | encoding.tolocal(old), | |
1540 | encoding.tolocal(new), |
|
1540 | encoding.tolocal(new), | |
1541 | ) |
|
1541 | ) | |
1542 |
|
1542 | |||
1543 |
|
1543 | |||
1544 | @wireprotocommand( |
|
1544 | @wireprotocommand( | |
1545 | b'rawstorefiledata', |
|
1545 | b'rawstorefiledata', | |
1546 | args={ |
|
1546 | args={ | |
1547 | b'files': { |
|
1547 | b'files': { | |
1548 | b'type': b'list', |
|
1548 | b'type': b'list', | |
1549 | b'example': [b'changelog', b'manifestlog'], |
|
1549 | b'example': [b'changelog', b'manifestlog'], | |
1550 | }, |
|
1550 | }, | |
1551 | b'pathfilter': { |
|
1551 | b'pathfilter': { | |
1552 | b'type': b'list', |
|
1552 | b'type': b'list', | |
1553 | b'default': lambda: None, |
|
1553 | b'default': lambda: None, | |
1554 | b'example': {b'include': [b'path:tests']}, |
|
1554 | b'example': {b'include': [b'path:tests']}, | |
1555 | }, |
|
1555 | }, | |
1556 | }, |
|
1556 | }, | |
1557 | permission=b'pull', |
|
1557 | permission=b'pull', | |
1558 | ) |
|
1558 | ) | |
1559 | def rawstorefiledata(repo, proto, files, pathfilter): |
|
1559 | def rawstorefiledata(repo, proto, files, pathfilter): | |
1560 | if not streamclone.allowservergeneration(repo): |
|
1560 | if not streamclone.allowservergeneration(repo): | |
1561 | raise error.WireprotoCommandError(b'stream clone is disabled') |
|
1561 | raise error.WireprotoCommandError(b'stream clone is disabled') | |
1562 |
|
1562 | |||
1563 | # TODO support dynamically advertising what store files "sets" are |
|
1563 | # TODO support dynamically advertising what store files "sets" are | |
1564 | # available. For now, we support changelog, manifestlog, and files. |
|
1564 | # available. For now, we support changelog, manifestlog, and files. | |
1565 | files = set(files) |
|
1565 | files = set(files) | |
1566 | allowedfiles = {b'changelog', b'manifestlog'} |
|
1566 | allowedfiles = {b'changelog', b'manifestlog'} | |
1567 |
|
1567 | |||
1568 | unsupported = files - allowedfiles |
|
1568 | unsupported = files - allowedfiles | |
1569 | if unsupported: |
|
1569 | if unsupported: | |
1570 | raise error.WireprotoCommandError( |
|
1570 | raise error.WireprotoCommandError( | |
1571 | b'unknown file type: %s', (b', '.join(sorted(unsupported)),) |
|
1571 | b'unknown file type: %s', (b', '.join(sorted(unsupported)),) | |
1572 | ) |
|
1572 | ) | |
1573 |
|
1573 | |||
1574 | with repo.lock(): |
|
1574 | with repo.lock(): | |
1575 | topfiles = list(repo.store.topfiles()) |
|
1575 | topfiles = list(repo.store.topfiles()) | |
1576 |
|
1576 | |||
1577 | sendfiles = [] |
|
1577 | sendfiles = [] | |
1578 | totalsize = 0 |
|
1578 | totalsize = 0 | |
1579 |
|
1579 | |||
1580 | # TODO this is a bunch of storage layer interface abstractions because |
|
1580 | # TODO this is a bunch of storage layer interface abstractions because | |
1581 | # it assumes revlogs. |
|
1581 | # it assumes revlogs. | |
1582 |
for rl_type, name, |
|
1582 | for rl_type, name, size in topfiles: | |
1583 | # XXX use the `rl_type` for that |
|
1583 | # XXX use the `rl_type` for that | |
1584 | if b'changelog' in files and name.startswith(b'00changelog'): |
|
1584 | if b'changelog' in files and name.startswith(b'00changelog'): | |
1585 | pass |
|
1585 | pass | |
1586 | elif b'manifestlog' in files and name.startswith(b'00manifest'): |
|
1586 | elif b'manifestlog' in files and name.startswith(b'00manifest'): | |
1587 | pass |
|
1587 | pass | |
1588 | else: |
|
1588 | else: | |
1589 | continue |
|
1589 | continue | |
1590 |
|
1590 | |||
1591 | sendfiles.append((b'store', name, size)) |
|
1591 | sendfiles.append((b'store', name, size)) | |
1592 | totalsize += size |
|
1592 | totalsize += size | |
1593 |
|
1593 | |||
1594 | yield { |
|
1594 | yield { | |
1595 | b'filecount': len(sendfiles), |
|
1595 | b'filecount': len(sendfiles), | |
1596 | b'totalsize': totalsize, |
|
1596 | b'totalsize': totalsize, | |
1597 | } |
|
1597 | } | |
1598 |
|
1598 | |||
1599 | for location, name, size in sendfiles: |
|
1599 | for location, name, size in sendfiles: | |
1600 | yield { |
|
1600 | yield { | |
1601 | b'location': location, |
|
1601 | b'location': location, | |
1602 | b'path': name, |
|
1602 | b'path': name, | |
1603 | b'size': size, |
|
1603 | b'size': size, | |
1604 | } |
|
1604 | } | |
1605 |
|
1605 | |||
1606 | # We have to use a closure for this to ensure the context manager is |
|
1606 | # We have to use a closure for this to ensure the context manager is | |
1607 | # closed only after sending the final chunk. |
|
1607 | # closed only after sending the final chunk. | |
1608 | def getfiledata(): |
|
1608 | def getfiledata(): | |
1609 | with repo.svfs(name, b'rb', auditpath=False) as fh: |
|
1609 | with repo.svfs(name, b'rb', auditpath=False) as fh: | |
1610 | for chunk in util.filechunkiter(fh, limit=size): |
|
1610 | for chunk in util.filechunkiter(fh, limit=size): | |
1611 | yield chunk |
|
1611 | yield chunk | |
1612 |
|
1612 | |||
1613 | yield wireprototypes.indefinitebytestringresponse(getfiledata()) |
|
1613 | yield wireprototypes.indefinitebytestringresponse(getfiledata()) |
@@ -1,741 +1,745 b'' | |||||
1 | # simplestorerepo.py - Extension that swaps in alternate repository storage. |
|
1 | # simplestorerepo.py - Extension that swaps in alternate repository storage. | |
2 | # |
|
2 | # | |
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.com> |
|
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.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 | # To use this with the test suite: |
|
8 | # To use this with the test suite: | |
9 | # |
|
9 | # | |
10 | # $ HGREPOFEATURES="simplestore" ./run-tests.py \ |
|
10 | # $ HGREPOFEATURES="simplestore" ./run-tests.py \ | |
11 | # --extra-config-opt extensions.simplestore=`pwd`/simplestorerepo.py |
|
11 | # --extra-config-opt extensions.simplestore=`pwd`/simplestorerepo.py | |
12 |
|
12 | |||
13 | from __future__ import absolute_import |
|
13 | from __future__ import absolute_import | |
14 |
|
14 | |||
15 | import stat |
|
15 | import stat | |
16 |
|
16 | |||
17 | from mercurial.i18n import _ |
|
17 | from mercurial.i18n import _ | |
18 | from mercurial.node import ( |
|
18 | from mercurial.node import ( | |
19 | bin, |
|
19 | bin, | |
20 | hex, |
|
20 | hex, | |
21 | nullrev, |
|
21 | nullrev, | |
22 | ) |
|
22 | ) | |
23 | from mercurial.thirdparty import attr |
|
23 | from mercurial.thirdparty import attr | |
24 | from mercurial import ( |
|
24 | from mercurial import ( | |
25 | ancestor, |
|
25 | ancestor, | |
26 | bundlerepo, |
|
26 | bundlerepo, | |
27 | error, |
|
27 | error, | |
28 | extensions, |
|
28 | extensions, | |
29 | localrepo, |
|
29 | localrepo, | |
30 | mdiff, |
|
30 | mdiff, | |
31 | pycompat, |
|
31 | pycompat, | |
32 | revlog, |
|
32 | revlog, | |
33 | store, |
|
33 | store, | |
34 | verify, |
|
34 | verify, | |
35 | ) |
|
35 | ) | |
36 | from mercurial.interfaces import ( |
|
36 | from mercurial.interfaces import ( | |
37 | repository, |
|
37 | repository, | |
38 | util as interfaceutil, |
|
38 | util as interfaceutil, | |
39 | ) |
|
39 | ) | |
40 | from mercurial.utils import ( |
|
40 | from mercurial.utils import ( | |
41 | cborutil, |
|
41 | cborutil, | |
42 | storageutil, |
|
42 | storageutil, | |
43 | ) |
|
43 | ) | |
44 | from mercurial.revlogutils import flagutil |
|
44 | from mercurial.revlogutils import flagutil | |
45 |
|
45 | |||
46 | # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for |
|
46 | # Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for | |
47 | # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should |
|
47 | # extensions which SHIP WITH MERCURIAL. Non-mainline extensions should | |
48 | # be specifying the version(s) of Mercurial they are tested with, or |
|
48 | # be specifying the version(s) of Mercurial they are tested with, or | |
49 | # leave the attribute unspecified. |
|
49 | # leave the attribute unspecified. | |
50 | testedwith = b'ships-with-hg-core' |
|
50 | testedwith = b'ships-with-hg-core' | |
51 |
|
51 | |||
52 | REQUIREMENT = b'testonly-simplestore' |
|
52 | REQUIREMENT = b'testonly-simplestore' | |
53 |
|
53 | |||
54 |
|
54 | |||
55 | def validatenode(node): |
|
55 | def validatenode(node): | |
56 | if isinstance(node, int): |
|
56 | if isinstance(node, int): | |
57 | raise ValueError('expected node; got int') |
|
57 | raise ValueError('expected node; got int') | |
58 |
|
58 | |||
59 | if len(node) != 20: |
|
59 | if len(node) != 20: | |
60 | raise ValueError('expected 20 byte node') |
|
60 | raise ValueError('expected 20 byte node') | |
61 |
|
61 | |||
62 |
|
62 | |||
63 | def validaterev(rev): |
|
63 | def validaterev(rev): | |
64 | if not isinstance(rev, int): |
|
64 | if not isinstance(rev, int): | |
65 | raise ValueError('expected int') |
|
65 | raise ValueError('expected int') | |
66 |
|
66 | |||
67 |
|
67 | |||
68 | class simplestoreerror(error.StorageError): |
|
68 | class simplestoreerror(error.StorageError): | |
69 | pass |
|
69 | pass | |
70 |
|
70 | |||
71 |
|
71 | |||
72 | @interfaceutil.implementer(repository.irevisiondelta) |
|
72 | @interfaceutil.implementer(repository.irevisiondelta) | |
73 | @attr.s(slots=True) |
|
73 | @attr.s(slots=True) | |
74 | class simplestorerevisiondelta(object): |
|
74 | class simplestorerevisiondelta(object): | |
75 | node = attr.ib() |
|
75 | node = attr.ib() | |
76 | p1node = attr.ib() |
|
76 | p1node = attr.ib() | |
77 | p2node = attr.ib() |
|
77 | p2node = attr.ib() | |
78 | basenode = attr.ib() |
|
78 | basenode = attr.ib() | |
79 | flags = attr.ib() |
|
79 | flags = attr.ib() | |
80 | baserevisionsize = attr.ib() |
|
80 | baserevisionsize = attr.ib() | |
81 | revision = attr.ib() |
|
81 | revision = attr.ib() | |
82 | delta = attr.ib() |
|
82 | delta = attr.ib() | |
83 | linknode = attr.ib(default=None) |
|
83 | linknode = attr.ib(default=None) | |
84 |
|
84 | |||
85 |
|
85 | |||
86 | @interfaceutil.implementer(repository.iverifyproblem) |
|
86 | @interfaceutil.implementer(repository.iverifyproblem) | |
87 | @attr.s(frozen=True) |
|
87 | @attr.s(frozen=True) | |
88 | class simplefilestoreproblem(object): |
|
88 | class simplefilestoreproblem(object): | |
89 | warning = attr.ib(default=None) |
|
89 | warning = attr.ib(default=None) | |
90 | error = attr.ib(default=None) |
|
90 | error = attr.ib(default=None) | |
91 | node = attr.ib(default=None) |
|
91 | node = attr.ib(default=None) | |
92 |
|
92 | |||
93 |
|
93 | |||
94 | @interfaceutil.implementer(repository.ifilestorage) |
|
94 | @interfaceutil.implementer(repository.ifilestorage) | |
95 | class filestorage(object): |
|
95 | class filestorage(object): | |
96 | """Implements storage for a tracked path. |
|
96 | """Implements storage for a tracked path. | |
97 |
|
97 | |||
98 | Data is stored in the VFS in a directory corresponding to the tracked |
|
98 | Data is stored in the VFS in a directory corresponding to the tracked | |
99 | path. |
|
99 | path. | |
100 |
|
100 | |||
101 | Index data is stored in an ``index`` file using CBOR. |
|
101 | Index data is stored in an ``index`` file using CBOR. | |
102 |
|
102 | |||
103 | Fulltext data is stored in files having names of the node. |
|
103 | Fulltext data is stored in files having names of the node. | |
104 | """ |
|
104 | """ | |
105 |
|
105 | |||
106 | _flagserrorclass = simplestoreerror |
|
106 | _flagserrorclass = simplestoreerror | |
107 |
|
107 | |||
108 | def __init__(self, repo, svfs, path): |
|
108 | def __init__(self, repo, svfs, path): | |
109 | self.nullid = repo.nullid |
|
109 | self.nullid = repo.nullid | |
110 | self._repo = repo |
|
110 | self._repo = repo | |
111 | self._svfs = svfs |
|
111 | self._svfs = svfs | |
112 | self._path = path |
|
112 | self._path = path | |
113 |
|
113 | |||
114 | self._storepath = b'/'.join([b'data', path]) |
|
114 | self._storepath = b'/'.join([b'data', path]) | |
115 | self._indexpath = b'/'.join([self._storepath, b'index']) |
|
115 | self._indexpath = b'/'.join([self._storepath, b'index']) | |
116 |
|
116 | |||
117 | indexdata = self._svfs.tryread(self._indexpath) |
|
117 | indexdata = self._svfs.tryread(self._indexpath) | |
118 | if indexdata: |
|
118 | if indexdata: | |
119 | indexdata = cborutil.decodeall(indexdata) |
|
119 | indexdata = cborutil.decodeall(indexdata) | |
120 |
|
120 | |||
121 | self._indexdata = indexdata or [] |
|
121 | self._indexdata = indexdata or [] | |
122 | self._indexbynode = {} |
|
122 | self._indexbynode = {} | |
123 | self._indexbyrev = {} |
|
123 | self._indexbyrev = {} | |
124 | self._index = [] |
|
124 | self._index = [] | |
125 | self._refreshindex() |
|
125 | self._refreshindex() | |
126 |
|
126 | |||
127 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
127 | self._flagprocessors = dict(flagutil.flagprocessors) | |
128 |
|
128 | |||
129 | def _refreshindex(self): |
|
129 | def _refreshindex(self): | |
130 | self._indexbynode.clear() |
|
130 | self._indexbynode.clear() | |
131 | self._indexbyrev.clear() |
|
131 | self._indexbyrev.clear() | |
132 | self._index = [] |
|
132 | self._index = [] | |
133 |
|
133 | |||
134 | for i, entry in enumerate(self._indexdata): |
|
134 | for i, entry in enumerate(self._indexdata): | |
135 | self._indexbynode[entry[b'node']] = entry |
|
135 | self._indexbynode[entry[b'node']] = entry | |
136 | self._indexbyrev[i] = entry |
|
136 | self._indexbyrev[i] = entry | |
137 |
|
137 | |||
138 | self._indexbynode[self._repo.nullid] = { |
|
138 | self._indexbynode[self._repo.nullid] = { | |
139 | b'node': self._repo.nullid, |
|
139 | b'node': self._repo.nullid, | |
140 | b'p1': self._repo.nullid, |
|
140 | b'p1': self._repo.nullid, | |
141 | b'p2': self._repo.nullid, |
|
141 | b'p2': self._repo.nullid, | |
142 | b'linkrev': nullrev, |
|
142 | b'linkrev': nullrev, | |
143 | b'flags': 0, |
|
143 | b'flags': 0, | |
144 | } |
|
144 | } | |
145 |
|
145 | |||
146 | self._indexbyrev[nullrev] = { |
|
146 | self._indexbyrev[nullrev] = { | |
147 | b'node': self._repo.nullid, |
|
147 | b'node': self._repo.nullid, | |
148 | b'p1': self._repo.nullid, |
|
148 | b'p1': self._repo.nullid, | |
149 | b'p2': self._repo.nullid, |
|
149 | b'p2': self._repo.nullid, | |
150 | b'linkrev': nullrev, |
|
150 | b'linkrev': nullrev, | |
151 | b'flags': 0, |
|
151 | b'flags': 0, | |
152 | } |
|
152 | } | |
153 |
|
153 | |||
154 | for i, entry in enumerate(self._indexdata): |
|
154 | for i, entry in enumerate(self._indexdata): | |
155 | p1rev, p2rev = self.parentrevs(self.rev(entry[b'node'])) |
|
155 | p1rev, p2rev = self.parentrevs(self.rev(entry[b'node'])) | |
156 |
|
156 | |||
157 | # start, length, rawsize, chainbase, linkrev, p1, p2, node |
|
157 | # start, length, rawsize, chainbase, linkrev, p1, p2, node | |
158 | self._index.append( |
|
158 | self._index.append( | |
159 | (0, 0, 0, -1, entry[b'linkrev'], p1rev, p2rev, entry[b'node']) |
|
159 | (0, 0, 0, -1, entry[b'linkrev'], p1rev, p2rev, entry[b'node']) | |
160 | ) |
|
160 | ) | |
161 |
|
161 | |||
162 | self._index.append((0, 0, 0, -1, -1, -1, -1, self._repo.nullid)) |
|
162 | self._index.append((0, 0, 0, -1, -1, -1, -1, self._repo.nullid)) | |
163 |
|
163 | |||
164 | def __len__(self): |
|
164 | def __len__(self): | |
165 | return len(self._indexdata) |
|
165 | return len(self._indexdata) | |
166 |
|
166 | |||
167 | def __iter__(self): |
|
167 | def __iter__(self): | |
168 | return iter(range(len(self))) |
|
168 | return iter(range(len(self))) | |
169 |
|
169 | |||
170 | def revs(self, start=0, stop=None): |
|
170 | def revs(self, start=0, stop=None): | |
171 | step = 1 |
|
171 | step = 1 | |
172 | if stop is not None: |
|
172 | if stop is not None: | |
173 | if start > stop: |
|
173 | if start > stop: | |
174 | step = -1 |
|
174 | step = -1 | |
175 |
|
175 | |||
176 | stop += step |
|
176 | stop += step | |
177 | else: |
|
177 | else: | |
178 | stop = len(self) |
|
178 | stop = len(self) | |
179 |
|
179 | |||
180 | return range(start, stop, step) |
|
180 | return range(start, stop, step) | |
181 |
|
181 | |||
182 | def parents(self, node): |
|
182 | def parents(self, node): | |
183 | validatenode(node) |
|
183 | validatenode(node) | |
184 |
|
184 | |||
185 | if node not in self._indexbynode: |
|
185 | if node not in self._indexbynode: | |
186 | raise KeyError('unknown node') |
|
186 | raise KeyError('unknown node') | |
187 |
|
187 | |||
188 | entry = self._indexbynode[node] |
|
188 | entry = self._indexbynode[node] | |
189 |
|
189 | |||
190 | return entry[b'p1'], entry[b'p2'] |
|
190 | return entry[b'p1'], entry[b'p2'] | |
191 |
|
191 | |||
192 | def parentrevs(self, rev): |
|
192 | def parentrevs(self, rev): | |
193 | p1, p2 = self.parents(self._indexbyrev[rev][b'node']) |
|
193 | p1, p2 = self.parents(self._indexbyrev[rev][b'node']) | |
194 | return self.rev(p1), self.rev(p2) |
|
194 | return self.rev(p1), self.rev(p2) | |
195 |
|
195 | |||
196 | def rev(self, node): |
|
196 | def rev(self, node): | |
197 | validatenode(node) |
|
197 | validatenode(node) | |
198 |
|
198 | |||
199 | try: |
|
199 | try: | |
200 | self._indexbynode[node] |
|
200 | self._indexbynode[node] | |
201 | except KeyError: |
|
201 | except KeyError: | |
202 | raise error.LookupError(node, self._indexpath, _('no node')) |
|
202 | raise error.LookupError(node, self._indexpath, _('no node')) | |
203 |
|
203 | |||
204 | for rev, entry in self._indexbyrev.items(): |
|
204 | for rev, entry in self._indexbyrev.items(): | |
205 | if entry[b'node'] == node: |
|
205 | if entry[b'node'] == node: | |
206 | return rev |
|
206 | return rev | |
207 |
|
207 | |||
208 | raise error.ProgrammingError(b'this should not occur') |
|
208 | raise error.ProgrammingError(b'this should not occur') | |
209 |
|
209 | |||
210 | def node(self, rev): |
|
210 | def node(self, rev): | |
211 | validaterev(rev) |
|
211 | validaterev(rev) | |
212 |
|
212 | |||
213 | return self._indexbyrev[rev][b'node'] |
|
213 | return self._indexbyrev[rev][b'node'] | |
214 |
|
214 | |||
215 | def hasnode(self, node): |
|
215 | def hasnode(self, node): | |
216 | validatenode(node) |
|
216 | validatenode(node) | |
217 | return node in self._indexbynode |
|
217 | return node in self._indexbynode | |
218 |
|
218 | |||
219 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
219 | def censorrevision(self, tr, censornode, tombstone=b''): | |
220 | raise NotImplementedError('TODO') |
|
220 | raise NotImplementedError('TODO') | |
221 |
|
221 | |||
222 | def lookup(self, node): |
|
222 | def lookup(self, node): | |
223 | if isinstance(node, int): |
|
223 | if isinstance(node, int): | |
224 | return self.node(node) |
|
224 | return self.node(node) | |
225 |
|
225 | |||
226 | if len(node) == 20: |
|
226 | if len(node) == 20: | |
227 | self.rev(node) |
|
227 | self.rev(node) | |
228 | return node |
|
228 | return node | |
229 |
|
229 | |||
230 | try: |
|
230 | try: | |
231 | rev = int(node) |
|
231 | rev = int(node) | |
232 | if '%d' % rev != node: |
|
232 | if '%d' % rev != node: | |
233 | raise ValueError |
|
233 | raise ValueError | |
234 |
|
234 | |||
235 | if rev < 0: |
|
235 | if rev < 0: | |
236 | rev = len(self) + rev |
|
236 | rev = len(self) + rev | |
237 | if rev < 0 or rev >= len(self): |
|
237 | if rev < 0 or rev >= len(self): | |
238 | raise ValueError |
|
238 | raise ValueError | |
239 |
|
239 | |||
240 | return self.node(rev) |
|
240 | return self.node(rev) | |
241 | except (ValueError, OverflowError): |
|
241 | except (ValueError, OverflowError): | |
242 | pass |
|
242 | pass | |
243 |
|
243 | |||
244 | if len(node) == 40: |
|
244 | if len(node) == 40: | |
245 | try: |
|
245 | try: | |
246 | rawnode = bin(node) |
|
246 | rawnode = bin(node) | |
247 | self.rev(rawnode) |
|
247 | self.rev(rawnode) | |
248 | return rawnode |
|
248 | return rawnode | |
249 | except TypeError: |
|
249 | except TypeError: | |
250 | pass |
|
250 | pass | |
251 |
|
251 | |||
252 | raise error.LookupError(node, self._path, _('invalid lookup input')) |
|
252 | raise error.LookupError(node, self._path, _('invalid lookup input')) | |
253 |
|
253 | |||
254 | def linkrev(self, rev): |
|
254 | def linkrev(self, rev): | |
255 | validaterev(rev) |
|
255 | validaterev(rev) | |
256 |
|
256 | |||
257 | return self._indexbyrev[rev][b'linkrev'] |
|
257 | return self._indexbyrev[rev][b'linkrev'] | |
258 |
|
258 | |||
259 | def _flags(self, rev): |
|
259 | def _flags(self, rev): | |
260 | validaterev(rev) |
|
260 | validaterev(rev) | |
261 |
|
261 | |||
262 | return self._indexbyrev[rev][b'flags'] |
|
262 | return self._indexbyrev[rev][b'flags'] | |
263 |
|
263 | |||
264 | def _candelta(self, baserev, rev): |
|
264 | def _candelta(self, baserev, rev): | |
265 | validaterev(baserev) |
|
265 | validaterev(baserev) | |
266 | validaterev(rev) |
|
266 | validaterev(rev) | |
267 |
|
267 | |||
268 | if (self._flags(baserev) & revlog.REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
268 | if (self._flags(baserev) & revlog.REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
269 | self._flags(rev) & revlog.REVIDX_RAWTEXT_CHANGING_FLAGS |
|
269 | self._flags(rev) & revlog.REVIDX_RAWTEXT_CHANGING_FLAGS | |
270 | ): |
|
270 | ): | |
271 | return False |
|
271 | return False | |
272 |
|
272 | |||
273 | return True |
|
273 | return True | |
274 |
|
274 | |||
275 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
275 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
276 | if p1 is None and p2 is None: |
|
276 | if p1 is None and p2 is None: | |
277 | p1, p2 = self.parents(node) |
|
277 | p1, p2 = self.parents(node) | |
278 | if node != storageutil.hashrevisionsha1(text, p1, p2): |
|
278 | if node != storageutil.hashrevisionsha1(text, p1, p2): | |
279 | raise simplestoreerror( |
|
279 | raise simplestoreerror( | |
280 | _("integrity check failed on %s") % self._path |
|
280 | _("integrity check failed on %s") % self._path | |
281 | ) |
|
281 | ) | |
282 |
|
282 | |||
283 | def revision(self, nodeorrev, raw=False): |
|
283 | def revision(self, nodeorrev, raw=False): | |
284 | if isinstance(nodeorrev, int): |
|
284 | if isinstance(nodeorrev, int): | |
285 | node = self.node(nodeorrev) |
|
285 | node = self.node(nodeorrev) | |
286 | else: |
|
286 | else: | |
287 | node = nodeorrev |
|
287 | node = nodeorrev | |
288 | validatenode(node) |
|
288 | validatenode(node) | |
289 |
|
289 | |||
290 | if node == self._repo.nullid: |
|
290 | if node == self._repo.nullid: | |
291 | return b'' |
|
291 | return b'' | |
292 |
|
292 | |||
293 | rev = self.rev(node) |
|
293 | rev = self.rev(node) | |
294 | flags = self._flags(rev) |
|
294 | flags = self._flags(rev) | |
295 |
|
295 | |||
296 | path = b'/'.join([self._storepath, hex(node)]) |
|
296 | path = b'/'.join([self._storepath, hex(node)]) | |
297 | rawtext = self._svfs.read(path) |
|
297 | rawtext = self._svfs.read(path) | |
298 |
|
298 | |||
299 | if raw: |
|
299 | if raw: | |
300 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
300 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
301 | text = rawtext |
|
301 | text = rawtext | |
302 | else: |
|
302 | else: | |
303 | r = flagutil.processflagsread(self, rawtext, flags) |
|
303 | r = flagutil.processflagsread(self, rawtext, flags) | |
304 | text, validatehash = r |
|
304 | text, validatehash = r | |
305 | if validatehash: |
|
305 | if validatehash: | |
306 | self.checkhash(text, node, rev=rev) |
|
306 | self.checkhash(text, node, rev=rev) | |
307 |
|
307 | |||
308 | return text |
|
308 | return text | |
309 |
|
309 | |||
310 | def rawdata(self, nodeorrev): |
|
310 | def rawdata(self, nodeorrev): | |
311 | return self.revision(raw=True) |
|
311 | return self.revision(raw=True) | |
312 |
|
312 | |||
313 | def read(self, node): |
|
313 | def read(self, node): | |
314 | validatenode(node) |
|
314 | validatenode(node) | |
315 |
|
315 | |||
316 | revision = self.revision(node) |
|
316 | revision = self.revision(node) | |
317 |
|
317 | |||
318 | if not revision.startswith(b'\1\n'): |
|
318 | if not revision.startswith(b'\1\n'): | |
319 | return revision |
|
319 | return revision | |
320 |
|
320 | |||
321 | start = revision.index(b'\1\n', 2) |
|
321 | start = revision.index(b'\1\n', 2) | |
322 | return revision[start + 2 :] |
|
322 | return revision[start + 2 :] | |
323 |
|
323 | |||
324 | def renamed(self, node): |
|
324 | def renamed(self, node): | |
325 | validatenode(node) |
|
325 | validatenode(node) | |
326 |
|
326 | |||
327 | if self.parents(node)[0] != self._repo.nullid: |
|
327 | if self.parents(node)[0] != self._repo.nullid: | |
328 | return False |
|
328 | return False | |
329 |
|
329 | |||
330 | fulltext = self.revision(node) |
|
330 | fulltext = self.revision(node) | |
331 | m = storageutil.parsemeta(fulltext)[0] |
|
331 | m = storageutil.parsemeta(fulltext)[0] | |
332 |
|
332 | |||
333 | if m and 'copy' in m: |
|
333 | if m and 'copy' in m: | |
334 | return m['copy'], bin(m['copyrev']) |
|
334 | return m['copy'], bin(m['copyrev']) | |
335 |
|
335 | |||
336 | return False |
|
336 | return False | |
337 |
|
337 | |||
338 | def cmp(self, node, text): |
|
338 | def cmp(self, node, text): | |
339 | validatenode(node) |
|
339 | validatenode(node) | |
340 |
|
340 | |||
341 | t = text |
|
341 | t = text | |
342 |
|
342 | |||
343 | if text.startswith(b'\1\n'): |
|
343 | if text.startswith(b'\1\n'): | |
344 | t = b'\1\n\1\n' + text |
|
344 | t = b'\1\n\1\n' + text | |
345 |
|
345 | |||
346 | p1, p2 = self.parents(node) |
|
346 | p1, p2 = self.parents(node) | |
347 |
|
347 | |||
348 | if storageutil.hashrevisionsha1(t, p1, p2) == node: |
|
348 | if storageutil.hashrevisionsha1(t, p1, p2) == node: | |
349 | return False |
|
349 | return False | |
350 |
|
350 | |||
351 | if self.iscensored(self.rev(node)): |
|
351 | if self.iscensored(self.rev(node)): | |
352 | return text != b'' |
|
352 | return text != b'' | |
353 |
|
353 | |||
354 | if self.renamed(node): |
|
354 | if self.renamed(node): | |
355 | t2 = self.read(node) |
|
355 | t2 = self.read(node) | |
356 | return t2 != text |
|
356 | return t2 != text | |
357 |
|
357 | |||
358 | return True |
|
358 | return True | |
359 |
|
359 | |||
360 | def size(self, rev): |
|
360 | def size(self, rev): | |
361 | validaterev(rev) |
|
361 | validaterev(rev) | |
362 |
|
362 | |||
363 | node = self._indexbyrev[rev][b'node'] |
|
363 | node = self._indexbyrev[rev][b'node'] | |
364 |
|
364 | |||
365 | if self.renamed(node): |
|
365 | if self.renamed(node): | |
366 | return len(self.read(node)) |
|
366 | return len(self.read(node)) | |
367 |
|
367 | |||
368 | if self.iscensored(rev): |
|
368 | if self.iscensored(rev): | |
369 | return 0 |
|
369 | return 0 | |
370 |
|
370 | |||
371 | return len(self.revision(node)) |
|
371 | return len(self.revision(node)) | |
372 |
|
372 | |||
373 | def iscensored(self, rev): |
|
373 | def iscensored(self, rev): | |
374 | validaterev(rev) |
|
374 | validaterev(rev) | |
375 |
|
375 | |||
376 | return self._flags(rev) & repository.REVISION_FLAG_CENSORED |
|
376 | return self._flags(rev) & repository.REVISION_FLAG_CENSORED | |
377 |
|
377 | |||
378 | def commonancestorsheads(self, a, b): |
|
378 | def commonancestorsheads(self, a, b): | |
379 | validatenode(a) |
|
379 | validatenode(a) | |
380 | validatenode(b) |
|
380 | validatenode(b) | |
381 |
|
381 | |||
382 | a = self.rev(a) |
|
382 | a = self.rev(a) | |
383 | b = self.rev(b) |
|
383 | b = self.rev(b) | |
384 |
|
384 | |||
385 | ancestors = ancestor.commonancestorsheads(self.parentrevs, a, b) |
|
385 | ancestors = ancestor.commonancestorsheads(self.parentrevs, a, b) | |
386 | return pycompat.maplist(self.node, ancestors) |
|
386 | return pycompat.maplist(self.node, ancestors) | |
387 |
|
387 | |||
388 | def descendants(self, revs): |
|
388 | def descendants(self, revs): | |
389 | # This is a copy of revlog.descendants() |
|
389 | # This is a copy of revlog.descendants() | |
390 | first = min(revs) |
|
390 | first = min(revs) | |
391 | if first == nullrev: |
|
391 | if first == nullrev: | |
392 | for i in self: |
|
392 | for i in self: | |
393 | yield i |
|
393 | yield i | |
394 | return |
|
394 | return | |
395 |
|
395 | |||
396 | seen = set(revs) |
|
396 | seen = set(revs) | |
397 | for i in self.revs(start=first + 1): |
|
397 | for i in self.revs(start=first + 1): | |
398 | for x in self.parentrevs(i): |
|
398 | for x in self.parentrevs(i): | |
399 | if x != nullrev and x in seen: |
|
399 | if x != nullrev and x in seen: | |
400 | seen.add(i) |
|
400 | seen.add(i) | |
401 | yield i |
|
401 | yield i | |
402 | break |
|
402 | break | |
403 |
|
403 | |||
404 | # Required by verify. |
|
404 | # Required by verify. | |
405 | def files(self): |
|
405 | def files(self): | |
406 | entries = self._svfs.listdir(self._storepath) |
|
406 | entries = self._svfs.listdir(self._storepath) | |
407 |
|
407 | |||
408 | # Strip out undo.backup.* files created as part of transaction |
|
408 | # Strip out undo.backup.* files created as part of transaction | |
409 | # recording. |
|
409 | # recording. | |
410 | entries = [f for f in entries if not f.startswith('undo.backup.')] |
|
410 | entries = [f for f in entries if not f.startswith('undo.backup.')] | |
411 |
|
411 | |||
412 | return [b'/'.join((self._storepath, f)) for f in entries] |
|
412 | return [b'/'.join((self._storepath, f)) for f in entries] | |
413 |
|
413 | |||
414 | def storageinfo( |
|
414 | def storageinfo( | |
415 | self, |
|
415 | self, | |
416 | exclusivefiles=False, |
|
416 | exclusivefiles=False, | |
417 | sharedfiles=False, |
|
417 | sharedfiles=False, | |
418 | revisionscount=False, |
|
418 | revisionscount=False, | |
419 | trackedsize=False, |
|
419 | trackedsize=False, | |
420 | storedsize=False, |
|
420 | storedsize=False, | |
421 | ): |
|
421 | ): | |
422 | # TODO do a real implementation of this |
|
422 | # TODO do a real implementation of this | |
423 | return { |
|
423 | return { | |
424 | 'exclusivefiles': [], |
|
424 | 'exclusivefiles': [], | |
425 | 'sharedfiles': [], |
|
425 | 'sharedfiles': [], | |
426 | 'revisionscount': len(self), |
|
426 | 'revisionscount': len(self), | |
427 | 'trackedsize': 0, |
|
427 | 'trackedsize': 0, | |
428 | 'storedsize': None, |
|
428 | 'storedsize': None, | |
429 | } |
|
429 | } | |
430 |
|
430 | |||
431 | def verifyintegrity(self, state): |
|
431 | def verifyintegrity(self, state): | |
432 | state['skipread'] = set() |
|
432 | state['skipread'] = set() | |
433 | for rev in self: |
|
433 | for rev in self: | |
434 | node = self.node(rev) |
|
434 | node = self.node(rev) | |
435 | try: |
|
435 | try: | |
436 | self.revision(node) |
|
436 | self.revision(node) | |
437 | except Exception as e: |
|
437 | except Exception as e: | |
438 | yield simplefilestoreproblem( |
|
438 | yield simplefilestoreproblem( | |
439 | error='unpacking %s: %s' % (node, e), node=node |
|
439 | error='unpacking %s: %s' % (node, e), node=node | |
440 | ) |
|
440 | ) | |
441 | state['skipread'].add(node) |
|
441 | state['skipread'].add(node) | |
442 |
|
442 | |||
443 | def emitrevisions( |
|
443 | def emitrevisions( | |
444 | self, |
|
444 | self, | |
445 | nodes, |
|
445 | nodes, | |
446 | nodesorder=None, |
|
446 | nodesorder=None, | |
447 | revisiondata=False, |
|
447 | revisiondata=False, | |
448 | assumehaveparentrevisions=False, |
|
448 | assumehaveparentrevisions=False, | |
449 | deltamode=repository.CG_DELTAMODE_STD, |
|
449 | deltamode=repository.CG_DELTAMODE_STD, | |
450 | sidedata_helpers=None, |
|
450 | sidedata_helpers=None, | |
451 | ): |
|
451 | ): | |
452 | # TODO this will probably break on some ordering options. |
|
452 | # TODO this will probably break on some ordering options. | |
453 | nodes = [n for n in nodes if n != self._repo.nullid] |
|
453 | nodes = [n for n in nodes if n != self._repo.nullid] | |
454 | if not nodes: |
|
454 | if not nodes: | |
455 | return |
|
455 | return | |
456 | for delta in storageutil.emitrevisions( |
|
456 | for delta in storageutil.emitrevisions( | |
457 | self, |
|
457 | self, | |
458 | nodes, |
|
458 | nodes, | |
459 | nodesorder, |
|
459 | nodesorder, | |
460 | simplestorerevisiondelta, |
|
460 | simplestorerevisiondelta, | |
461 | revisiondata=revisiondata, |
|
461 | revisiondata=revisiondata, | |
462 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
462 | assumehaveparentrevisions=assumehaveparentrevisions, | |
463 | deltamode=deltamode, |
|
463 | deltamode=deltamode, | |
464 | sidedata_helpers=sidedata_helpers, |
|
464 | sidedata_helpers=sidedata_helpers, | |
465 | ): |
|
465 | ): | |
466 | yield delta |
|
466 | yield delta | |
467 |
|
467 | |||
468 | def add(self, text, meta, transaction, linkrev, p1, p2): |
|
468 | def add(self, text, meta, transaction, linkrev, p1, p2): | |
469 | if meta or text.startswith(b'\1\n'): |
|
469 | if meta or text.startswith(b'\1\n'): | |
470 | text = storageutil.packmeta(meta, text) |
|
470 | text = storageutil.packmeta(meta, text) | |
471 |
|
471 | |||
472 | return self.addrevision(text, transaction, linkrev, p1, p2) |
|
472 | return self.addrevision(text, transaction, linkrev, p1, p2) | |
473 |
|
473 | |||
474 | def addrevision( |
|
474 | def addrevision( | |
475 | self, |
|
475 | self, | |
476 | text, |
|
476 | text, | |
477 | transaction, |
|
477 | transaction, | |
478 | linkrev, |
|
478 | linkrev, | |
479 | p1, |
|
479 | p1, | |
480 | p2, |
|
480 | p2, | |
481 | node=None, |
|
481 | node=None, | |
482 | flags=revlog.REVIDX_DEFAULT_FLAGS, |
|
482 | flags=revlog.REVIDX_DEFAULT_FLAGS, | |
483 | cachedelta=None, |
|
483 | cachedelta=None, | |
484 | ): |
|
484 | ): | |
485 | validatenode(p1) |
|
485 | validatenode(p1) | |
486 | validatenode(p2) |
|
486 | validatenode(p2) | |
487 |
|
487 | |||
488 | if flags: |
|
488 | if flags: | |
489 | node = node or storageutil.hashrevisionsha1(text, p1, p2) |
|
489 | node = node or storageutil.hashrevisionsha1(text, p1, p2) | |
490 |
|
490 | |||
491 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) |
|
491 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) | |
492 |
|
492 | |||
493 | node = node or storageutil.hashrevisionsha1(text, p1, p2) |
|
493 | node = node or storageutil.hashrevisionsha1(text, p1, p2) | |
494 |
|
494 | |||
495 | if node in self._indexbynode: |
|
495 | if node in self._indexbynode: | |
496 | return node |
|
496 | return node | |
497 |
|
497 | |||
498 | if validatehash: |
|
498 | if validatehash: | |
499 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
499 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
500 |
|
500 | |||
501 | return self._addrawrevision( |
|
501 | return self._addrawrevision( | |
502 | node, rawtext, transaction, linkrev, p1, p2, flags |
|
502 | node, rawtext, transaction, linkrev, p1, p2, flags | |
503 | ) |
|
503 | ) | |
504 |
|
504 | |||
505 | def _addrawrevision(self, node, rawtext, transaction, link, p1, p2, flags): |
|
505 | def _addrawrevision(self, node, rawtext, transaction, link, p1, p2, flags): | |
506 | transaction.addbackup(self._indexpath) |
|
506 | transaction.addbackup(self._indexpath) | |
507 |
|
507 | |||
508 | path = b'/'.join([self._storepath, hex(node)]) |
|
508 | path = b'/'.join([self._storepath, hex(node)]) | |
509 |
|
509 | |||
510 | self._svfs.write(path, rawtext) |
|
510 | self._svfs.write(path, rawtext) | |
511 |
|
511 | |||
512 | self._indexdata.append( |
|
512 | self._indexdata.append( | |
513 | { |
|
513 | { | |
514 | b'node': node, |
|
514 | b'node': node, | |
515 | b'p1': p1, |
|
515 | b'p1': p1, | |
516 | b'p2': p2, |
|
516 | b'p2': p2, | |
517 | b'linkrev': link, |
|
517 | b'linkrev': link, | |
518 | b'flags': flags, |
|
518 | b'flags': flags, | |
519 | } |
|
519 | } | |
520 | ) |
|
520 | ) | |
521 |
|
521 | |||
522 | self._reflectindexupdate() |
|
522 | self._reflectindexupdate() | |
523 |
|
523 | |||
524 | return node |
|
524 | return node | |
525 |
|
525 | |||
526 | def _reflectindexupdate(self): |
|
526 | def _reflectindexupdate(self): | |
527 | self._refreshindex() |
|
527 | self._refreshindex() | |
528 | self._svfs.write( |
|
528 | self._svfs.write( | |
529 | self._indexpath, ''.join(cborutil.streamencode(self._indexdata)) |
|
529 | self._indexpath, ''.join(cborutil.streamencode(self._indexdata)) | |
530 | ) |
|
530 | ) | |
531 |
|
531 | |||
532 | def addgroup( |
|
532 | def addgroup( | |
533 | self, |
|
533 | self, | |
534 | deltas, |
|
534 | deltas, | |
535 | linkmapper, |
|
535 | linkmapper, | |
536 | transaction, |
|
536 | transaction, | |
537 | addrevisioncb=None, |
|
537 | addrevisioncb=None, | |
538 | duplicaterevisioncb=None, |
|
538 | duplicaterevisioncb=None, | |
539 | maybemissingparents=False, |
|
539 | maybemissingparents=False, | |
540 | ): |
|
540 | ): | |
541 | if maybemissingparents: |
|
541 | if maybemissingparents: | |
542 | raise error.Abort( |
|
542 | raise error.Abort( | |
543 | _('simple store does not support missing parents ' 'write mode') |
|
543 | _('simple store does not support missing parents ' 'write mode') | |
544 | ) |
|
544 | ) | |
545 |
|
545 | |||
546 | empty = True |
|
546 | empty = True | |
547 |
|
547 | |||
548 | transaction.addbackup(self._indexpath) |
|
548 | transaction.addbackup(self._indexpath) | |
549 |
|
549 | |||
550 | for node, p1, p2, linknode, deltabase, delta, flags in deltas: |
|
550 | for node, p1, p2, linknode, deltabase, delta, flags in deltas: | |
551 | linkrev = linkmapper(linknode) |
|
551 | linkrev = linkmapper(linknode) | |
552 | flags = flags or revlog.REVIDX_DEFAULT_FLAGS |
|
552 | flags = flags or revlog.REVIDX_DEFAULT_FLAGS | |
553 |
|
553 | |||
554 | if node in self._indexbynode: |
|
554 | if node in self._indexbynode: | |
555 | if duplicaterevisioncb: |
|
555 | if duplicaterevisioncb: | |
556 | duplicaterevisioncb(self, self.rev(node)) |
|
556 | duplicaterevisioncb(self, self.rev(node)) | |
557 | empty = False |
|
557 | empty = False | |
558 | continue |
|
558 | continue | |
559 |
|
559 | |||
560 | # Need to resolve the fulltext from the delta base. |
|
560 | # Need to resolve the fulltext from the delta base. | |
561 | if deltabase == self._repo.nullid: |
|
561 | if deltabase == self._repo.nullid: | |
562 | text = mdiff.patch(b'', delta) |
|
562 | text = mdiff.patch(b'', delta) | |
563 | else: |
|
563 | else: | |
564 | text = mdiff.patch(self.revision(deltabase), delta) |
|
564 | text = mdiff.patch(self.revision(deltabase), delta) | |
565 |
|
565 | |||
566 | rev = self._addrawrevision( |
|
566 | rev = self._addrawrevision( | |
567 | node, text, transaction, linkrev, p1, p2, flags |
|
567 | node, text, transaction, linkrev, p1, p2, flags | |
568 | ) |
|
568 | ) | |
569 |
|
569 | |||
570 | if addrevisioncb: |
|
570 | if addrevisioncb: | |
571 | addrevisioncb(self, rev) |
|
571 | addrevisioncb(self, rev) | |
572 | empty = False |
|
572 | empty = False | |
573 | return not empty |
|
573 | return not empty | |
574 |
|
574 | |||
575 | def _headrevs(self): |
|
575 | def _headrevs(self): | |
576 | # Assume all revisions are heads by default. |
|
576 | # Assume all revisions are heads by default. | |
577 | revishead = {rev: True for rev in self._indexbyrev} |
|
577 | revishead = {rev: True for rev in self._indexbyrev} | |
578 |
|
578 | |||
579 | for rev, entry in self._indexbyrev.items(): |
|
579 | for rev, entry in self._indexbyrev.items(): | |
580 | # Unset head flag for all seen parents. |
|
580 | # Unset head flag for all seen parents. | |
581 | revishead[self.rev(entry[b'p1'])] = False |
|
581 | revishead[self.rev(entry[b'p1'])] = False | |
582 | revishead[self.rev(entry[b'p2'])] = False |
|
582 | revishead[self.rev(entry[b'p2'])] = False | |
583 |
|
583 | |||
584 | return [rev for rev, ishead in sorted(revishead.items()) if ishead] |
|
584 | return [rev for rev, ishead in sorted(revishead.items()) if ishead] | |
585 |
|
585 | |||
586 | def heads(self, start=None, stop=None): |
|
586 | def heads(self, start=None, stop=None): | |
587 | # This is copied from revlog.py. |
|
587 | # This is copied from revlog.py. | |
588 | if start is None and stop is None: |
|
588 | if start is None and stop is None: | |
589 | if not len(self): |
|
589 | if not len(self): | |
590 | return [self._repo.nullid] |
|
590 | return [self._repo.nullid] | |
591 | return [self.node(r) for r in self._headrevs()] |
|
591 | return [self.node(r) for r in self._headrevs()] | |
592 |
|
592 | |||
593 | if start is None: |
|
593 | if start is None: | |
594 | start = self._repo.nullid |
|
594 | start = self._repo.nullid | |
595 | if stop is None: |
|
595 | if stop is None: | |
596 | stop = [] |
|
596 | stop = [] | |
597 | stoprevs = {self.rev(n) for n in stop} |
|
597 | stoprevs = {self.rev(n) for n in stop} | |
598 | startrev = self.rev(start) |
|
598 | startrev = self.rev(start) | |
599 | reachable = {startrev} |
|
599 | reachable = {startrev} | |
600 | heads = {startrev} |
|
600 | heads = {startrev} | |
601 |
|
601 | |||
602 | parentrevs = self.parentrevs |
|
602 | parentrevs = self.parentrevs | |
603 | for r in self.revs(start=startrev + 1): |
|
603 | for r in self.revs(start=startrev + 1): | |
604 | for p in parentrevs(r): |
|
604 | for p in parentrevs(r): | |
605 | if p in reachable: |
|
605 | if p in reachable: | |
606 | if r not in stoprevs: |
|
606 | if r not in stoprevs: | |
607 | reachable.add(r) |
|
607 | reachable.add(r) | |
608 | heads.add(r) |
|
608 | heads.add(r) | |
609 | if p in heads and p not in stoprevs: |
|
609 | if p in heads and p not in stoprevs: | |
610 | heads.remove(p) |
|
610 | heads.remove(p) | |
611 |
|
611 | |||
612 | return [self.node(r) for r in heads] |
|
612 | return [self.node(r) for r in heads] | |
613 |
|
613 | |||
614 | def children(self, node): |
|
614 | def children(self, node): | |
615 | validatenode(node) |
|
615 | validatenode(node) | |
616 |
|
616 | |||
617 | # This is a copy of revlog.children(). |
|
617 | # This is a copy of revlog.children(). | |
618 | c = [] |
|
618 | c = [] | |
619 | p = self.rev(node) |
|
619 | p = self.rev(node) | |
620 | for r in self.revs(start=p + 1): |
|
620 | for r in self.revs(start=p + 1): | |
621 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
621 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
622 | if prevs: |
|
622 | if prevs: | |
623 | for pr in prevs: |
|
623 | for pr in prevs: | |
624 | if pr == p: |
|
624 | if pr == p: | |
625 | c.append(self.node(r)) |
|
625 | c.append(self.node(r)) | |
626 | elif p == nullrev: |
|
626 | elif p == nullrev: | |
627 | c.append(self.node(r)) |
|
627 | c.append(self.node(r)) | |
628 | return c |
|
628 | return c | |
629 |
|
629 | |||
630 | def getstrippoint(self, minlink): |
|
630 | def getstrippoint(self, minlink): | |
631 | return storageutil.resolvestripinfo( |
|
631 | return storageutil.resolvestripinfo( | |
632 | minlink, |
|
632 | minlink, | |
633 | len(self) - 1, |
|
633 | len(self) - 1, | |
634 | self._headrevs(), |
|
634 | self._headrevs(), | |
635 | self.linkrev, |
|
635 | self.linkrev, | |
636 | self.parentrevs, |
|
636 | self.parentrevs, | |
637 | ) |
|
637 | ) | |
638 |
|
638 | |||
639 | def strip(self, minlink, transaction): |
|
639 | def strip(self, minlink, transaction): | |
640 | if not len(self): |
|
640 | if not len(self): | |
641 | return |
|
641 | return | |
642 |
|
642 | |||
643 | rev, _ignored = self.getstrippoint(minlink) |
|
643 | rev, _ignored = self.getstrippoint(minlink) | |
644 | if rev == len(self): |
|
644 | if rev == len(self): | |
645 | return |
|
645 | return | |
646 |
|
646 | |||
647 | # Purge index data starting at the requested revision. |
|
647 | # Purge index data starting at the requested revision. | |
648 | self._indexdata[rev:] = [] |
|
648 | self._indexdata[rev:] = [] | |
649 | self._reflectindexupdate() |
|
649 | self._reflectindexupdate() | |
650 |
|
650 | |||
651 |
|
651 | |||
652 | def issimplestorefile(f, kind, st): |
|
652 | def issimplestorefile(f, kind, st): | |
653 | if kind != stat.S_IFREG: |
|
653 | if kind != stat.S_IFREG: | |
654 | return False |
|
654 | return False | |
655 |
|
655 | |||
656 | if store.isrevlog(f, kind, st): |
|
656 | if store.isrevlog(f, kind, st): | |
657 | return False |
|
657 | return False | |
658 |
|
658 | |||
659 | # Ignore transaction undo files. |
|
659 | # Ignore transaction undo files. | |
660 | if f.startswith('undo.'): |
|
660 | if f.startswith('undo.'): | |
661 | return False |
|
661 | return False | |
662 |
|
662 | |||
663 | # Otherwise assume it belongs to the simple store. |
|
663 | # Otherwise assume it belongs to the simple store. | |
664 | return True |
|
664 | return True | |
665 |
|
665 | |||
666 |
|
666 | |||
667 | class simplestore(store.encodedstore): |
|
667 | class simplestore(store.encodedstore): | |
668 | def datafiles(self): |
|
668 | def datafiles(self, undecodable=None): | |
669 | for x in super(simplestore, self).datafiles(): |
|
669 | for x in super(simplestore, self).datafiles(): | |
670 | yield x |
|
670 | yield x | |
671 |
|
671 | |||
672 | # Supplement with non-revlog files. |
|
672 | # Supplement with non-revlog files. | |
673 | extrafiles = self._walk('data', True, filefilter=issimplestorefile) |
|
673 | extrafiles = self._walk('data', True, filefilter=issimplestorefile) | |
674 |
|
674 | |||
675 |
for |
|
675 | for f1, size in extrafiles: | |
676 | try: |
|
676 | try: | |
677 |
|
|
677 | f2 = store.decodefilename(f1) | |
678 | except KeyError: |
|
678 | except KeyError: | |
679 |
|
|
679 | if undecodable is None: | |
|
680 | raise error.StorageError(b'undecodable revlog name %s' % f1) | |||
|
681 | else: | |||
|
682 | undecodable.append(f1) | |||
|
683 | continue | |||
680 |
|
684 | |||
681 |
yield |
|
685 | yield f2, size | |
682 |
|
686 | |||
683 |
|
687 | |||
684 | def reposetup(ui, repo): |
|
688 | def reposetup(ui, repo): | |
685 | if not repo.local(): |
|
689 | if not repo.local(): | |
686 | return |
|
690 | return | |
687 |
|
691 | |||
688 | if isinstance(repo, bundlerepo.bundlerepository): |
|
692 | if isinstance(repo, bundlerepo.bundlerepository): | |
689 | raise error.Abort(_('cannot use simple store with bundlerepo')) |
|
693 | raise error.Abort(_('cannot use simple store with bundlerepo')) | |
690 |
|
694 | |||
691 | class simplestorerepo(repo.__class__): |
|
695 | class simplestorerepo(repo.__class__): | |
692 | def file(self, f): |
|
696 | def file(self, f): | |
693 | return filestorage(repo, self.svfs, f) |
|
697 | return filestorage(repo, self.svfs, f) | |
694 |
|
698 | |||
695 | repo.__class__ = simplestorerepo |
|
699 | repo.__class__ = simplestorerepo | |
696 |
|
700 | |||
697 |
|
701 | |||
698 | def featuresetup(ui, supported): |
|
702 | def featuresetup(ui, supported): | |
699 | supported.add(REQUIREMENT) |
|
703 | supported.add(REQUIREMENT) | |
700 |
|
704 | |||
701 |
|
705 | |||
702 | def newreporequirements(orig, ui, createopts): |
|
706 | def newreporequirements(orig, ui, createopts): | |
703 | """Modifies default requirements for new repos to use the simple store.""" |
|
707 | """Modifies default requirements for new repos to use the simple store.""" | |
704 | requirements = orig(ui, createopts) |
|
708 | requirements = orig(ui, createopts) | |
705 |
|
709 | |||
706 | # These requirements are only used to affect creation of the store |
|
710 | # These requirements are only used to affect creation of the store | |
707 | # object. We have our own store. So we can remove them. |
|
711 | # object. We have our own store. So we can remove them. | |
708 | # TODO do this once we feel like taking the test hit. |
|
712 | # TODO do this once we feel like taking the test hit. | |
709 | # if 'fncache' in requirements: |
|
713 | # if 'fncache' in requirements: | |
710 | # requirements.remove('fncache') |
|
714 | # requirements.remove('fncache') | |
711 | # if 'dotencode' in requirements: |
|
715 | # if 'dotencode' in requirements: | |
712 | # requirements.remove('dotencode') |
|
716 | # requirements.remove('dotencode') | |
713 |
|
717 | |||
714 | requirements.add(REQUIREMENT) |
|
718 | requirements.add(REQUIREMENT) | |
715 |
|
719 | |||
716 | return requirements |
|
720 | return requirements | |
717 |
|
721 | |||
718 |
|
722 | |||
719 | def makestore(orig, requirements, path, vfstype): |
|
723 | def makestore(orig, requirements, path, vfstype): | |
720 | if REQUIREMENT not in requirements: |
|
724 | if REQUIREMENT not in requirements: | |
721 | return orig(requirements, path, vfstype) |
|
725 | return orig(requirements, path, vfstype) | |
722 |
|
726 | |||
723 | return simplestore(path, vfstype) |
|
727 | return simplestore(path, vfstype) | |
724 |
|
728 | |||
725 |
|
729 | |||
726 | def verifierinit(orig, self, *args, **kwargs): |
|
730 | def verifierinit(orig, self, *args, **kwargs): | |
727 | orig(self, *args, **kwargs) |
|
731 | orig(self, *args, **kwargs) | |
728 |
|
732 | |||
729 | # We don't care that files in the store don't align with what is |
|
733 | # We don't care that files in the store don't align with what is | |
730 | # advertised. So suppress these warnings. |
|
734 | # advertised. So suppress these warnings. | |
731 | self.warnorphanstorefiles = False |
|
735 | self.warnorphanstorefiles = False | |
732 |
|
736 | |||
733 |
|
737 | |||
734 | def extsetup(ui): |
|
738 | def extsetup(ui): | |
735 | localrepo.featuresetupfuncs.add(featuresetup) |
|
739 | localrepo.featuresetupfuncs.add(featuresetup) | |
736 |
|
740 | |||
737 | extensions.wrapfunction( |
|
741 | extensions.wrapfunction( | |
738 | localrepo, 'newreporequirements', newreporequirements |
|
742 | localrepo, 'newreporequirements', newreporequirements | |
739 | ) |
|
743 | ) | |
740 | extensions.wrapfunction(localrepo, 'makestore', makestore) |
|
744 | extensions.wrapfunction(localrepo, 'makestore', makestore) | |
741 | extensions.wrapfunction(verify.verifier, '__init__', verifierinit) |
|
745 | extensions.wrapfunction(verify.verifier, '__init__', verifierinit) |
General Comments 0
You need to be logged in to leave comments.
Login now