Show More
@@ -1,823 +1,824 b'' | |||||
1 | # Copyright 2009-2010 Gregory P. Ward |
|
1 | # Copyright 2009-2010 Gregory P. Ward | |
2 | # Copyright 2009-2010 Intelerad Medical Systems Incorporated |
|
2 | # Copyright 2009-2010 Intelerad Medical Systems Incorporated | |
3 | # Copyright 2010-2011 Fog Creek Software |
|
3 | # Copyright 2010-2011 Fog Creek Software | |
4 | # Copyright 2010-2011 Unity Technologies |
|
4 | # Copyright 2010-2011 Unity Technologies | |
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 | '''largefiles utility code: must not import other modules in this package.''' |
|
9 | '''largefiles utility code: must not import other modules in this package.''' | |
10 |
|
10 | |||
11 | import contextlib |
|
11 | import contextlib | |
12 | import copy |
|
12 | import copy | |
13 | import os |
|
13 | import os | |
14 | import stat |
|
14 | import stat | |
15 |
|
15 | |||
16 | from mercurial.i18n import _ |
|
16 | from mercurial.i18n import _ | |
17 | from mercurial.node import hex |
|
17 | from mercurial.node import hex | |
18 | from mercurial.pycompat import open |
|
18 | from mercurial.pycompat import open | |
19 |
|
19 | |||
20 | from mercurial import ( |
|
20 | from mercurial import ( | |
21 | dirstate, |
|
21 | dirstate, | |
22 | encoding, |
|
22 | encoding, | |
23 | error, |
|
23 | error, | |
24 | httpconnection, |
|
24 | httpconnection, | |
25 | match as matchmod, |
|
25 | match as matchmod, | |
26 | pycompat, |
|
26 | pycompat, | |
27 | requirements, |
|
27 | requirements, | |
28 | scmutil, |
|
28 | scmutil, | |
29 | sparse, |
|
29 | sparse, | |
30 | util, |
|
30 | util, | |
31 | vfs as vfsmod, |
|
31 | vfs as vfsmod, | |
32 | ) |
|
32 | ) | |
33 | from mercurial.utils import hashutil |
|
33 | from mercurial.utils import hashutil | |
34 | from mercurial.dirstateutils import timestamp |
|
34 | from mercurial.dirstateutils import timestamp | |
35 |
|
35 | |||
36 | shortname = b'.hglf' |
|
36 | shortname = b'.hglf' | |
37 | shortnameslash = shortname + b'/' |
|
37 | shortnameslash = shortname + b'/' | |
38 | longname = b'largefiles' |
|
38 | longname = b'largefiles' | |
39 |
|
39 | |||
40 | # -- Private worker functions ------------------------------------------ |
|
40 | # -- Private worker functions ------------------------------------------ | |
41 |
|
41 | |||
42 |
|
42 | |||
43 | @contextlib.contextmanager |
|
43 | @contextlib.contextmanager | |
44 | def lfstatus(repo, value=True): |
|
44 | def lfstatus(repo, value=True): | |
45 | oldvalue = getattr(repo, 'lfstatus', False) |
|
45 | oldvalue = getattr(repo, 'lfstatus', False) | |
46 | repo.lfstatus = value |
|
46 | repo.lfstatus = value | |
47 | try: |
|
47 | try: | |
48 | yield |
|
48 | yield | |
49 | finally: |
|
49 | finally: | |
50 | repo.lfstatus = oldvalue |
|
50 | repo.lfstatus = oldvalue | |
51 |
|
51 | |||
52 |
|
52 | |||
53 | def getminsize(ui, assumelfiles, opt, default=10): |
|
53 | def getminsize(ui, assumelfiles, opt, default=10): | |
54 | lfsize = opt |
|
54 | lfsize = opt | |
55 | if not lfsize and assumelfiles: |
|
55 | if not lfsize and assumelfiles: | |
56 | lfsize = ui.config(longname, b'minsize', default=default) |
|
56 | lfsize = ui.config(longname, b'minsize', default=default) | |
57 | if lfsize: |
|
57 | if lfsize: | |
58 | try: |
|
58 | try: | |
59 | lfsize = float(lfsize) |
|
59 | lfsize = float(lfsize) | |
60 | except ValueError: |
|
60 | except ValueError: | |
61 | raise error.Abort( |
|
61 | raise error.Abort( | |
62 | _(b'largefiles: size must be number (not %s)\n') % lfsize |
|
62 | _(b'largefiles: size must be number (not %s)\n') % lfsize | |
63 | ) |
|
63 | ) | |
64 | if lfsize is None: |
|
64 | if lfsize is None: | |
65 | raise error.Abort(_(b'minimum size for largefiles must be specified')) |
|
65 | raise error.Abort(_(b'minimum size for largefiles must be specified')) | |
66 | return lfsize |
|
66 | return lfsize | |
67 |
|
67 | |||
68 |
|
68 | |||
69 | def link(src, dest): |
|
69 | def link(src, dest): | |
70 | """Try to create hardlink - if that fails, efficiently make a copy.""" |
|
70 | """Try to create hardlink - if that fails, efficiently make a copy.""" | |
71 | util.makedirs(os.path.dirname(dest)) |
|
71 | util.makedirs(os.path.dirname(dest)) | |
72 | try: |
|
72 | try: | |
73 | util.oslink(src, dest) |
|
73 | util.oslink(src, dest) | |
74 | except OSError: |
|
74 | except OSError: | |
75 | # if hardlinks fail, fallback on atomic copy |
|
75 | # if hardlinks fail, fallback on atomic copy | |
76 | with open(src, b'rb') as srcf, util.atomictempfile(dest) as dstf: |
|
76 | with open(src, b'rb') as srcf, util.atomictempfile(dest) as dstf: | |
77 | for chunk in util.filechunkiter(srcf): |
|
77 | for chunk in util.filechunkiter(srcf): | |
78 | dstf.write(chunk) |
|
78 | dstf.write(chunk) | |
79 | os.chmod(dest, os.stat(src).st_mode) |
|
79 | os.chmod(dest, os.stat(src).st_mode) | |
80 |
|
80 | |||
81 |
|
81 | |||
82 | def usercachepath(ui, hash): |
|
82 | def usercachepath(ui, hash): | |
83 | """Return the correct location in the "global" largefiles cache for a file |
|
83 | """Return the correct location in the "global" largefiles cache for a file | |
84 | with the given hash. |
|
84 | with the given hash. | |
85 | This cache is used for sharing of largefiles across repositories - both |
|
85 | This cache is used for sharing of largefiles across repositories - both | |
86 | to preserve download bandwidth and storage space.""" |
|
86 | to preserve download bandwidth and storage space.""" | |
87 | return os.path.join(_usercachedir(ui), hash) |
|
87 | return os.path.join(_usercachedir(ui), hash) | |
88 |
|
88 | |||
89 |
|
89 | |||
90 | def _usercachedir(ui, name=longname): |
|
90 | def _usercachedir(ui, name=longname): | |
91 | '''Return the location of the "global" largefiles cache.''' |
|
91 | '''Return the location of the "global" largefiles cache.''' | |
92 | path = ui.configpath(name, b'usercache') |
|
92 | path = ui.configpath(name, b'usercache') | |
93 | if path: |
|
93 | if path: | |
94 | return path |
|
94 | return path | |
95 |
|
95 | |||
96 | hint = None |
|
96 | hint = None | |
97 |
|
97 | |||
98 | if pycompat.iswindows: |
|
98 | if pycompat.iswindows: | |
99 | appdata = encoding.environ.get( |
|
99 | appdata = encoding.environ.get( | |
100 | b'LOCALAPPDATA', encoding.environ.get(b'APPDATA') |
|
100 | b'LOCALAPPDATA', encoding.environ.get(b'APPDATA') | |
101 | ) |
|
101 | ) | |
102 | if appdata: |
|
102 | if appdata: | |
103 | return os.path.join(appdata, name) |
|
103 | return os.path.join(appdata, name) | |
104 |
|
104 | |||
105 | hint = _(b"define %s or %s in the environment, or set %s.usercache") % ( |
|
105 | hint = _(b"define %s or %s in the environment, or set %s.usercache") % ( | |
106 | b"LOCALAPPDATA", |
|
106 | b"LOCALAPPDATA", | |
107 | b"APPDATA", |
|
107 | b"APPDATA", | |
108 | name, |
|
108 | name, | |
109 | ) |
|
109 | ) | |
110 | elif pycompat.isdarwin: |
|
110 | elif pycompat.isdarwin: | |
111 | home = encoding.environ.get(b'HOME') |
|
111 | home = encoding.environ.get(b'HOME') | |
112 | if home: |
|
112 | if home: | |
113 | return os.path.join(home, b'Library', b'Caches', name) |
|
113 | return os.path.join(home, b'Library', b'Caches', name) | |
114 |
|
114 | |||
115 | hint = _(b"define %s in the environment, or set %s.usercache") % ( |
|
115 | hint = _(b"define %s in the environment, or set %s.usercache") % ( | |
116 | b"HOME", |
|
116 | b"HOME", | |
117 | name, |
|
117 | name, | |
118 | ) |
|
118 | ) | |
119 | elif pycompat.isposix: |
|
119 | elif pycompat.isposix: | |
120 | path = encoding.environ.get(b'XDG_CACHE_HOME') |
|
120 | path = encoding.environ.get(b'XDG_CACHE_HOME') | |
121 | if path: |
|
121 | if path: | |
122 | return os.path.join(path, name) |
|
122 | return os.path.join(path, name) | |
123 | home = encoding.environ.get(b'HOME') |
|
123 | home = encoding.environ.get(b'HOME') | |
124 | if home: |
|
124 | if home: | |
125 | return os.path.join(home, b'.cache', name) |
|
125 | return os.path.join(home, b'.cache', name) | |
126 |
|
126 | |||
127 | hint = _(b"define %s or %s in the environment, or set %s.usercache") % ( |
|
127 | hint = _(b"define %s or %s in the environment, or set %s.usercache") % ( | |
128 | b"XDG_CACHE_HOME", |
|
128 | b"XDG_CACHE_HOME", | |
129 | b"HOME", |
|
129 | b"HOME", | |
130 | name, |
|
130 | name, | |
131 | ) |
|
131 | ) | |
132 | else: |
|
132 | else: | |
133 | raise error.Abort( |
|
133 | raise error.Abort( | |
134 | _(b'unknown operating system: %s\n') % pycompat.osname |
|
134 | _(b'unknown operating system: %s\n') % pycompat.osname | |
135 | ) |
|
135 | ) | |
136 |
|
136 | |||
137 | raise error.Abort(_(b'unknown %s usercache location') % name, hint=hint) |
|
137 | raise error.Abort(_(b'unknown %s usercache location') % name, hint=hint) | |
138 |
|
138 | |||
139 |
|
139 | |||
140 | def inusercache(ui, hash): |
|
140 | def inusercache(ui, hash): | |
141 | path = usercachepath(ui, hash) |
|
141 | path = usercachepath(ui, hash) | |
142 | return os.path.exists(path) |
|
142 | return os.path.exists(path) | |
143 |
|
143 | |||
144 |
|
144 | |||
145 | def findfile(repo, hash): |
|
145 | def findfile(repo, hash): | |
146 | """Return store path of the largefile with the specified hash. |
|
146 | """Return store path of the largefile with the specified hash. | |
147 | As a side effect, the file might be linked from user cache. |
|
147 | As a side effect, the file might be linked from user cache. | |
148 | Return None if the file can't be found locally.""" |
|
148 | Return None if the file can't be found locally.""" | |
149 | path, exists = findstorepath(repo, hash) |
|
149 | path, exists = findstorepath(repo, hash) | |
150 | if exists: |
|
150 | if exists: | |
151 | repo.ui.note(_(b'found %s in store\n') % hash) |
|
151 | repo.ui.note(_(b'found %s in store\n') % hash) | |
152 | return path |
|
152 | return path | |
153 | elif inusercache(repo.ui, hash): |
|
153 | elif inusercache(repo.ui, hash): | |
154 | repo.ui.note(_(b'found %s in system cache\n') % hash) |
|
154 | repo.ui.note(_(b'found %s in system cache\n') % hash) | |
155 | path = storepath(repo, hash) |
|
155 | path = storepath(repo, hash) | |
156 | link(usercachepath(repo.ui, hash), path) |
|
156 | link(usercachepath(repo.ui, hash), path) | |
157 | return path |
|
157 | return path | |
158 | return None |
|
158 | return None | |
159 |
|
159 | |||
160 |
|
160 | |||
161 | class largefilesdirstate(dirstate.dirstate): |
|
161 | class largefilesdirstate(dirstate.dirstate): | |
162 | _large_file_dirstate = True |
|
162 | _large_file_dirstate = True | |
163 | _tr_key_suffix = b'-large-files' |
|
163 | _tr_key_suffix = b'-large-files' | |
164 |
|
164 | |||
165 | def __getitem__(self, key): |
|
165 | def __getitem__(self, key): | |
166 | return super(largefilesdirstate, self).__getitem__(unixpath(key)) |
|
166 | return super(largefilesdirstate, self).__getitem__(unixpath(key)) | |
167 |
|
167 | |||
168 | def set_tracked(self, f): |
|
168 | def set_tracked(self, f): | |
169 | return super(largefilesdirstate, self).set_tracked(unixpath(f)) |
|
169 | return super(largefilesdirstate, self).set_tracked(unixpath(f)) | |
170 |
|
170 | |||
171 | def set_untracked(self, f): |
|
171 | def set_untracked(self, f): | |
172 | return super(largefilesdirstate, self).set_untracked(unixpath(f)) |
|
172 | return super(largefilesdirstate, self).set_untracked(unixpath(f)) | |
173 |
|
173 | |||
174 | def normal(self, f, parentfiledata=None): |
|
174 | def normal(self, f, parentfiledata=None): | |
175 | # not sure if we should pass the `parentfiledata` down or throw it |
|
175 | # not sure if we should pass the `parentfiledata` down or throw it | |
176 | # away. So throwing it away to stay on the safe side. |
|
176 | # away. So throwing it away to stay on the safe side. | |
177 | return super(largefilesdirstate, self).normal(unixpath(f)) |
|
177 | return super(largefilesdirstate, self).normal(unixpath(f)) | |
178 |
|
178 | |||
179 | def remove(self, f): |
|
179 | def remove(self, f): | |
180 | return super(largefilesdirstate, self).remove(unixpath(f)) |
|
180 | return super(largefilesdirstate, self).remove(unixpath(f)) | |
181 |
|
181 | |||
182 | def add(self, f): |
|
182 | def add(self, f): | |
183 | return super(largefilesdirstate, self).add(unixpath(f)) |
|
183 | return super(largefilesdirstate, self).add(unixpath(f)) | |
184 |
|
184 | |||
185 | def drop(self, f): |
|
185 | def drop(self, f): | |
186 | return super(largefilesdirstate, self).drop(unixpath(f)) |
|
186 | return super(largefilesdirstate, self).drop(unixpath(f)) | |
187 |
|
187 | |||
188 | def forget(self, f): |
|
188 | def forget(self, f): | |
189 | return super(largefilesdirstate, self).forget(unixpath(f)) |
|
189 | return super(largefilesdirstate, self).forget(unixpath(f)) | |
190 |
|
190 | |||
191 | def normallookup(self, f): |
|
191 | def normallookup(self, f): | |
192 | return super(largefilesdirstate, self).normallookup(unixpath(f)) |
|
192 | return super(largefilesdirstate, self).normallookup(unixpath(f)) | |
193 |
|
193 | |||
194 | def _ignore(self, f): |
|
194 | def _ignore(self, f): | |
195 | return False |
|
195 | return False | |
196 |
|
196 | |||
197 | def write(self, tr): |
|
197 | def write(self, tr): | |
198 | # (1) disable PENDING mode always |
|
198 | # (1) disable PENDING mode always | |
199 | # (lfdirstate isn't yet managed as a part of the transaction) |
|
199 | # (lfdirstate isn't yet managed as a part of the transaction) | |
200 | # (2) avoid develwarn 'use dirstate.write with ....' |
|
200 | # (2) avoid develwarn 'use dirstate.write with ....' | |
201 | if tr: |
|
201 | if tr: | |
202 | tr.addbackup(b'largefiles/dirstate', location=b'plain') |
|
202 | tr.addbackup(b'largefiles/dirstate', location=b'plain') | |
203 | super(largefilesdirstate, self).write(None) |
|
203 | super(largefilesdirstate, self).write(None) | |
204 |
|
204 | |||
205 |
|
205 | |||
206 | def openlfdirstate(ui, repo, create=True): |
|
206 | def openlfdirstate(ui, repo, create=True): | |
207 | """ |
|
207 | """ | |
208 | Return a dirstate object that tracks largefiles: i.e. its root is |
|
208 | Return a dirstate object that tracks largefiles: i.e. its root is | |
209 | the repo root, but it is saved in .hg/largefiles/dirstate. |
|
209 | the repo root, but it is saved in .hg/largefiles/dirstate. | |
210 |
|
210 | |||
211 | If a dirstate object already exists and is being used for a 'changing_*' |
|
211 | If a dirstate object already exists and is being used for a 'changing_*' | |
212 | context, it will be returned. |
|
212 | context, it will be returned. | |
213 | """ |
|
213 | """ | |
214 | sub_dirstate = getattr(repo.dirstate, '_sub_dirstate', None) |
|
214 | sub_dirstate = getattr(repo.dirstate, '_sub_dirstate', None) | |
215 | if sub_dirstate is not None: |
|
215 | if sub_dirstate is not None: | |
216 | return sub_dirstate |
|
216 | return sub_dirstate | |
217 | vfs = repo.vfs |
|
217 | vfs = repo.vfs | |
218 | lfstoredir = longname |
|
218 | lfstoredir = longname | |
219 | opener = vfsmod.vfs(vfs.join(lfstoredir)) |
|
219 | opener = vfsmod.vfs(vfs.join(lfstoredir)) | |
220 | use_dirstate_v2 = requirements.DIRSTATE_V2_REQUIREMENT in repo.requirements |
|
220 | use_dirstate_v2 = requirements.DIRSTATE_V2_REQUIREMENT in repo.requirements | |
221 | lfdirstate = largefilesdirstate( |
|
221 | lfdirstate = largefilesdirstate( | |
222 | opener, |
|
222 | opener, | |
223 | ui, |
|
223 | ui, | |
224 | repo.root, |
|
224 | repo.root, | |
225 | repo.dirstate._validate, |
|
225 | repo.dirstate._validate, | |
226 | lambda: sparse.matcher(repo), |
|
226 | lambda: sparse.matcher(repo), | |
227 | repo.nodeconstants, |
|
227 | repo.nodeconstants, | |
228 | use_dirstate_v2, |
|
228 | use_dirstate_v2, | |
229 | ) |
|
229 | ) | |
230 |
|
230 | |||
231 | # If the largefiles dirstate does not exist, populate and create |
|
231 | # If the largefiles dirstate does not exist, populate and create | |
232 | # it. This ensures that we create it on the first meaningful |
|
232 | # it. This ensures that we create it on the first meaningful | |
233 | # largefiles operation in a new clone. |
|
233 | # largefiles operation in a new clone. | |
234 | if create and not vfs.exists(vfs.join(lfstoredir, b'dirstate')): |
|
234 | if create and not vfs.exists(vfs.join(lfstoredir, b'dirstate')): | |
235 | try: |
|
235 | try: | |
236 | with repo.wlock(wait=False), lfdirstate.changing_files(repo): |
|
236 | with repo.wlock(wait=False), lfdirstate.changing_files(repo): | |
237 | matcher = getstandinmatcher(repo) |
|
237 | matcher = getstandinmatcher(repo) | |
238 | standins = repo.dirstate.walk( |
|
238 | standins = repo.dirstate.walk( | |
239 | matcher, subrepos=[], unknown=False, ignored=False |
|
239 | matcher, subrepos=[], unknown=False, ignored=False | |
240 | ) |
|
240 | ) | |
241 |
|
241 | |||
242 | if len(standins) > 0: |
|
242 | if len(standins) > 0: | |
243 | vfs.makedirs(lfstoredir) |
|
243 | vfs.makedirs(lfstoredir) | |
244 |
|
244 | |||
245 | for standin in standins: |
|
245 | for standin in standins: | |
246 | lfile = splitstandin(standin) |
|
246 | lfile = splitstandin(standin) | |
247 | lfdirstate.hacky_extension_update_file( |
|
247 | lfdirstate.hacky_extension_update_file( | |
248 | lfile, |
|
248 | lfile, | |
249 | p1_tracked=True, |
|
249 | p1_tracked=True, | |
250 | wc_tracked=True, |
|
250 | wc_tracked=True, | |
251 | possibly_dirty=True, |
|
251 | possibly_dirty=True, | |
252 | ) |
|
252 | ) | |
253 | except error.LockError: |
|
253 | except error.LockError: | |
254 | # Assume that whatever was holding the lock was important. |
|
254 | # Assume that whatever was holding the lock was important. | |
255 | # If we were doing something important, we would already have |
|
255 | # If we were doing something important, we would already have | |
256 | # either the lock or a largefile dirstate. |
|
256 | # either the lock or a largefile dirstate. | |
257 | pass |
|
257 | pass | |
258 | return lfdirstate |
|
258 | return lfdirstate | |
259 |
|
259 | |||
260 |
|
260 | |||
261 | def lfdirstatestatus(lfdirstate, repo): |
|
261 | def lfdirstatestatus(lfdirstate, repo): | |
262 | pctx = repo[b'.'] |
|
262 | pctx = repo[b'.'] | |
263 | match = matchmod.always() |
|
263 | match = matchmod.always() | |
264 | unsure, s, mtime_boundary = lfdirstate.status( |
|
264 | unsure, s, mtime_boundary = lfdirstate.status( | |
265 | match, subrepos=[], ignored=False, clean=False, unknown=False |
|
265 | match, subrepos=[], ignored=False, clean=False, unknown=False | |
266 | ) |
|
266 | ) | |
267 | modified, clean = s.modified, s.clean |
|
267 | modified, clean = s.modified, s.clean | |
268 | wctx = repo[None] |
|
268 | wctx = repo[None] | |
269 | for lfile in unsure: |
|
269 | for lfile in unsure: | |
270 | try: |
|
270 | try: | |
271 | fctx = pctx[standin(lfile)] |
|
271 | fctx = pctx[standin(lfile)] | |
272 | except LookupError: |
|
272 | except LookupError: | |
273 | fctx = None |
|
273 | fctx = None | |
274 | if not fctx or readasstandin(fctx) != hashfile(repo.wjoin(lfile)): |
|
274 | if not fctx or readasstandin(fctx) != hashfile(repo.wjoin(lfile)): | |
275 | modified.append(lfile) |
|
275 | modified.append(lfile) | |
276 | else: |
|
276 | else: | |
277 | clean.append(lfile) |
|
277 | clean.append(lfile) | |
278 | st = wctx[lfile].lstat() |
|
278 | st = wctx[lfile].lstat() | |
279 | mode = st.st_mode |
|
279 | mode = st.st_mode | |
280 | size = st.st_size |
|
280 | size = st.st_size | |
281 | mtime = timestamp.reliable_mtime_of(st, mtime_boundary) |
|
281 | mtime = timestamp.reliable_mtime_of(st, mtime_boundary) | |
282 | if mtime is not None: |
|
282 | if mtime is not None: | |
283 | cache_data = (mode, size, mtime) |
|
283 | cache_data = (mode, size, mtime) | |
284 | lfdirstate.set_clean(lfile, cache_data) |
|
284 | lfdirstate.set_clean(lfile, cache_data) | |
285 | return s |
|
285 | return s | |
286 |
|
286 | |||
287 |
|
287 | |||
288 | def listlfiles(repo, rev=None, matcher=None): |
|
288 | def listlfiles(repo, rev=None, matcher=None): | |
289 | """return a list of largefiles in the working copy or the |
|
289 | """return a list of largefiles in the working copy or the | |
290 | specified changeset""" |
|
290 | specified changeset""" | |
291 |
|
291 | |||
292 | if matcher is None: |
|
292 | if matcher is None: | |
293 | matcher = getstandinmatcher(repo) |
|
293 | matcher = getstandinmatcher(repo) | |
294 |
|
294 | |||
295 | # ignore unknown files in working directory |
|
295 | # ignore unknown files in working directory | |
296 | return [ |
|
296 | return [ | |
297 | splitstandin(f) |
|
297 | splitstandin(f) | |
298 | for f in repo[rev].walk(matcher) |
|
298 | for f in repo[rev].walk(matcher) | |
299 | if rev is not None or repo.dirstate.get_entry(f).any_tracked |
|
299 | if rev is not None or repo.dirstate.get_entry(f).any_tracked | |
300 | ] |
|
300 | ] | |
301 |
|
301 | |||
302 |
|
302 | |||
303 | def instore(repo, hash, forcelocal=False): |
|
303 | def instore(repo, hash, forcelocal=False): | |
304 | '''Return true if a largefile with the given hash exists in the store''' |
|
304 | '''Return true if a largefile with the given hash exists in the store''' | |
305 | return os.path.exists(storepath(repo, hash, forcelocal)) |
|
305 | return os.path.exists(storepath(repo, hash, forcelocal)) | |
306 |
|
306 | |||
307 |
|
307 | |||
308 | def storepath(repo, hash, forcelocal=False): |
|
308 | def storepath(repo, hash, forcelocal=False): | |
309 | """Return the correct location in the repository largefiles store for a |
|
309 | """Return the correct location in the repository largefiles store for a | |
310 | file with the given hash.""" |
|
310 | file with the given hash.""" | |
311 | if not forcelocal and repo.shared(): |
|
311 | if not forcelocal and repo.shared(): | |
312 | return repo.vfs.reljoin(repo.sharedpath, longname, hash) |
|
312 | return repo.vfs.reljoin(repo.sharedpath, longname, hash) | |
313 | return repo.vfs.join(longname, hash) |
|
313 | return repo.vfs.join(longname, hash) | |
314 |
|
314 | |||
315 |
|
315 | |||
316 | def findstorepath(repo, hash): |
|
316 | def findstorepath(repo, hash): | |
317 | """Search through the local store path(s) to find the file for the given |
|
317 | """Search through the local store path(s) to find the file for the given | |
318 | hash. If the file is not found, its path in the primary store is returned. |
|
318 | hash. If the file is not found, its path in the primary store is returned. | |
319 | The return value is a tuple of (path, exists(path)). |
|
319 | The return value is a tuple of (path, exists(path)). | |
320 | """ |
|
320 | """ | |
321 | # For shared repos, the primary store is in the share source. But for |
|
321 | # For shared repos, the primary store is in the share source. But for | |
322 | # backward compatibility, force a lookup in the local store if it wasn't |
|
322 | # backward compatibility, force a lookup in the local store if it wasn't | |
323 | # found in the share source. |
|
323 | # found in the share source. | |
324 | path = storepath(repo, hash, False) |
|
324 | path = storepath(repo, hash, False) | |
325 |
|
325 | |||
326 | if instore(repo, hash): |
|
326 | if instore(repo, hash): | |
327 | return (path, True) |
|
327 | return (path, True) | |
328 | elif repo.shared() and instore(repo, hash, True): |
|
328 | elif repo.shared() and instore(repo, hash, True): | |
329 | return storepath(repo, hash, True), True |
|
329 | return storepath(repo, hash, True), True | |
330 |
|
330 | |||
331 | return (path, False) |
|
331 | return (path, False) | |
332 |
|
332 | |||
333 |
|
333 | |||
334 | def copyfromcache(repo, hash, filename): |
|
334 | def copyfromcache(repo, hash, filename): | |
335 | """Copy the specified largefile from the repo or system cache to |
|
335 | """Copy the specified largefile from the repo or system cache to | |
336 | filename in the repository. Return true on success or false if the |
|
336 | filename in the repository. Return true on success or false if the | |
337 | file was not found in either cache (which should not happened: |
|
337 | file was not found in either cache (which should not happened: | |
338 | this is meant to be called only after ensuring that the needed |
|
338 | this is meant to be called only after ensuring that the needed | |
339 | largefile exists in the cache).""" |
|
339 | largefile exists in the cache).""" | |
340 | wvfs = repo.wvfs |
|
340 | wvfs = repo.wvfs | |
341 | path = findfile(repo, hash) |
|
341 | path = findfile(repo, hash) | |
342 | if path is None: |
|
342 | if path is None: | |
343 | return False |
|
343 | return False | |
344 | wvfs.makedirs(wvfs.dirname(wvfs.join(filename))) |
|
344 | wvfs.makedirs(wvfs.dirname(wvfs.join(filename))) | |
345 | # The write may fail before the file is fully written, but we |
|
345 | # The write may fail before the file is fully written, but we | |
346 | # don't use atomic writes in the working copy. |
|
346 | # don't use atomic writes in the working copy. | |
347 | with open(path, b'rb') as srcfd, wvfs(filename, b'wb') as destfd: |
|
347 | with open(path, b'rb') as srcfd, wvfs(filename, b'wb') as destfd: | |
348 | gothash = copyandhash(util.filechunkiter(srcfd), destfd) |
|
348 | gothash = copyandhash(util.filechunkiter(srcfd), destfd) | |
349 | if gothash != hash: |
|
349 | if gothash != hash: | |
350 | repo.ui.warn( |
|
350 | repo.ui.warn( | |
351 | _(b'%s: data corruption in %s with hash %s\n') |
|
351 | _(b'%s: data corruption in %s with hash %s\n') | |
352 | % (filename, path, gothash) |
|
352 | % (filename, path, gothash) | |
353 | ) |
|
353 | ) | |
354 | wvfs.unlink(filename) |
|
354 | wvfs.unlink(filename) | |
355 | return False |
|
355 | return False | |
356 | return True |
|
356 | return True | |
357 |
|
357 | |||
358 |
|
358 | |||
359 | def copytostore(repo, ctx, file, fstandin): |
|
359 | def copytostore(repo, ctx, file, fstandin): | |
360 | wvfs = repo.wvfs |
|
360 | wvfs = repo.wvfs | |
361 | hash = readasstandin(ctx[fstandin]) |
|
361 | hash = readasstandin(ctx[fstandin]) | |
362 | if instore(repo, hash): |
|
362 | if instore(repo, hash): | |
363 | return |
|
363 | return | |
364 | if wvfs.exists(file): |
|
364 | if wvfs.exists(file): | |
365 | copytostoreabsolute(repo, wvfs.join(file), hash) |
|
365 | copytostoreabsolute(repo, wvfs.join(file), hash) | |
366 | else: |
|
366 | else: | |
367 | repo.ui.warn( |
|
367 | repo.ui.warn( | |
368 | _(b"%s: largefile %s not available from local store\n") |
|
368 | _(b"%s: largefile %s not available from local store\n") | |
369 | % (file, hash) |
|
369 | % (file, hash) | |
370 | ) |
|
370 | ) | |
371 |
|
371 | |||
372 |
|
372 | |||
373 | def copyalltostore(repo, node): |
|
373 | def copyalltostore(repo, node): | |
374 | '''Copy all largefiles in a given revision to the store''' |
|
374 | '''Copy all largefiles in a given revision to the store''' | |
375 |
|
375 | |||
376 | ctx = repo[node] |
|
376 | ctx = repo[node] | |
377 | for filename in ctx.files(): |
|
377 | for filename in ctx.files(): | |
378 | realfile = splitstandin(filename) |
|
378 | realfile = splitstandin(filename) | |
379 | if realfile is not None and filename in ctx.manifest(): |
|
379 | if realfile is not None and filename in ctx.manifest(): | |
380 | copytostore(repo, ctx, realfile, filename) |
|
380 | copytostore(repo, ctx, realfile, filename) | |
381 |
|
381 | |||
382 |
|
382 | |||
383 | def copytostoreabsolute(repo, file, hash): |
|
383 | def copytostoreabsolute(repo, file, hash): | |
384 | if inusercache(repo.ui, hash): |
|
384 | if inusercache(repo.ui, hash): | |
385 | link(usercachepath(repo.ui, hash), storepath(repo, hash)) |
|
385 | link(usercachepath(repo.ui, hash), storepath(repo, hash)) | |
386 | else: |
|
386 | else: | |
387 | util.makedirs(os.path.dirname(storepath(repo, hash))) |
|
387 | util.makedirs(os.path.dirname(storepath(repo, hash))) | |
388 | with open(file, b'rb') as srcf: |
|
388 | with open(file, b'rb') as srcf: | |
389 | with util.atomictempfile( |
|
389 | with util.atomictempfile( | |
390 | storepath(repo, hash), createmode=repo.store.createmode |
|
390 | storepath(repo, hash), createmode=repo.store.createmode | |
391 | ) as dstf: |
|
391 | ) as dstf: | |
392 | for chunk in util.filechunkiter(srcf): |
|
392 | for chunk in util.filechunkiter(srcf): | |
393 | dstf.write(chunk) |
|
393 | dstf.write(chunk) | |
394 | linktousercache(repo, hash) |
|
394 | linktousercache(repo, hash) | |
395 |
|
395 | |||
396 |
|
396 | |||
397 | def linktousercache(repo, hash): |
|
397 | def linktousercache(repo, hash): | |
398 | """Link / copy the largefile with the specified hash from the store |
|
398 | """Link / copy the largefile with the specified hash from the store | |
399 | to the cache.""" |
|
399 | to the cache.""" | |
400 | path = usercachepath(repo.ui, hash) |
|
400 | path = usercachepath(repo.ui, hash) | |
401 | link(storepath(repo, hash), path) |
|
401 | link(storepath(repo, hash), path) | |
402 |
|
402 | |||
403 |
|
403 | |||
404 | def getstandinmatcher(repo, rmatcher=None): |
|
404 | def getstandinmatcher(repo, rmatcher=None): | |
405 | '''Return a match object that applies rmatcher to the standin directory''' |
|
405 | '''Return a match object that applies rmatcher to the standin directory''' | |
406 | wvfs = repo.wvfs |
|
406 | wvfs = repo.wvfs | |
407 | standindir = shortname |
|
407 | standindir = shortname | |
408 |
|
408 | |||
409 | # no warnings about missing files or directories |
|
409 | # no warnings about missing files or directories | |
410 | badfn = lambda f, msg: None |
|
410 | badfn = lambda f, msg: None | |
411 |
|
411 | |||
412 | if rmatcher and not rmatcher.always(): |
|
412 | if rmatcher and not rmatcher.always(): | |
413 | pats = [wvfs.join(standindir, pat) for pat in rmatcher.files()] |
|
413 | pats = [wvfs.join(standindir, pat) for pat in rmatcher.files()] | |
414 | if not pats: |
|
414 | if not pats: | |
415 | pats = [wvfs.join(standindir)] |
|
415 | pats = [wvfs.join(standindir)] | |
416 | match = scmutil.match(repo[None], pats, badfn=badfn) |
|
416 | match = scmutil.match(repo[None], pats, badfn=badfn) | |
417 | else: |
|
417 | else: | |
418 | # no patterns: relative to repo root |
|
418 | # no patterns: relative to repo root | |
419 | match = scmutil.match(repo[None], [wvfs.join(standindir)], badfn=badfn) |
|
419 | match = scmutil.match(repo[None], [wvfs.join(standindir)], badfn=badfn) | |
420 | return match |
|
420 | return match | |
421 |
|
421 | |||
422 |
|
422 | |||
423 | def composestandinmatcher(repo, rmatcher): |
|
423 | def composestandinmatcher(repo, rmatcher): | |
424 | """Return a matcher that accepts standins corresponding to the |
|
424 | """Return a matcher that accepts standins corresponding to the | |
425 | files accepted by rmatcher. Pass the list of files in the matcher |
|
425 | files accepted by rmatcher. Pass the list of files in the matcher | |
426 | as the paths specified by the user.""" |
|
426 | as the paths specified by the user.""" | |
427 | smatcher = getstandinmatcher(repo, rmatcher) |
|
427 | smatcher = getstandinmatcher(repo, rmatcher) | |
428 | isstandin = smatcher.matchfn |
|
428 | isstandin = smatcher.matchfn | |
429 |
|
429 | |||
430 | def composedmatchfn(f): |
|
430 | def composedmatchfn(f): | |
431 | return isstandin(f) and rmatcher.matchfn(splitstandin(f)) |
|
431 | return isstandin(f) and rmatcher.matchfn(splitstandin(f)) | |
432 |
|
432 | |||
433 | smatcher.matchfn = composedmatchfn |
|
433 | smatcher.matchfn = composedmatchfn | |
434 |
|
434 | |||
435 | return smatcher |
|
435 | return smatcher | |
436 |
|
436 | |||
437 |
|
437 | |||
438 | def standin(filename): |
|
438 | def standin(filename): | |
439 | """Return the repo-relative path to the standin for the specified big |
|
439 | """Return the repo-relative path to the standin for the specified big | |
440 | file.""" |
|
440 | file.""" | |
441 | # Notes: |
|
441 | # Notes: | |
442 | # 1) Some callers want an absolute path, but for instance addlargefiles |
|
442 | # 1) Some callers want an absolute path, but for instance addlargefiles | |
443 | # needs it repo-relative so it can be passed to repo[None].add(). So |
|
443 | # needs it repo-relative so it can be passed to repo[None].add(). So | |
444 | # leave it up to the caller to use repo.wjoin() to get an absolute path. |
|
444 | # leave it up to the caller to use repo.wjoin() to get an absolute path. | |
445 | # 2) Join with '/' because that's what dirstate always uses, even on |
|
445 | # 2) Join with '/' because that's what dirstate always uses, even on | |
446 | # Windows. Change existing separator to '/' first in case we are |
|
446 | # Windows. Change existing separator to '/' first in case we are | |
447 | # passed filenames from an external source (like the command line). |
|
447 | # passed filenames from an external source (like the command line). | |
448 | return shortnameslash + util.pconvert(filename) |
|
448 | return shortnameslash + util.pconvert(filename) | |
449 |
|
449 | |||
450 |
|
450 | |||
451 | def isstandin(filename): |
|
451 | def isstandin(filename): | |
452 | """Return true if filename is a big file standin. filename must be |
|
452 | """Return true if filename is a big file standin. filename must be | |
453 | in Mercurial's internal form (slash-separated).""" |
|
453 | in Mercurial's internal form (slash-separated).""" | |
454 | return filename.startswith(shortnameslash) |
|
454 | return filename.startswith(shortnameslash) | |
455 |
|
455 | |||
456 |
|
456 | |||
457 | def splitstandin(filename): |
|
457 | def splitstandin(filename): | |
458 | # Split on / because that's what dirstate always uses, even on Windows. |
|
458 | # Split on / because that's what dirstate always uses, even on Windows. | |
459 | # Change local separator to / first just in case we are passed filenames |
|
459 | # Change local separator to / first just in case we are passed filenames | |
460 | # from an external source (like the command line). |
|
460 | # from an external source (like the command line). | |
461 | bits = util.pconvert(filename).split(b'/', 1) |
|
461 | bits = util.pconvert(filename).split(b'/', 1) | |
462 | if len(bits) == 2 and bits[0] == shortname: |
|
462 | if len(bits) == 2 and bits[0] == shortname: | |
463 | return bits[1] |
|
463 | return bits[1] | |
464 | else: |
|
464 | else: | |
465 | return None |
|
465 | return None | |
466 |
|
466 | |||
467 |
|
467 | |||
468 | def updatestandin(repo, lfile, standin): |
|
468 | def updatestandin(repo, lfile, standin): | |
469 | """Re-calculate hash value of lfile and write it into standin |
|
469 | """Re-calculate hash value of lfile and write it into standin | |
470 |
|
470 | |||
471 | This assumes that "lfutil.standin(lfile) == standin", for efficiency. |
|
471 | This assumes that "lfutil.standin(lfile) == standin", for efficiency. | |
472 | """ |
|
472 | """ | |
473 | file = repo.wjoin(lfile) |
|
473 | file = repo.wjoin(lfile) | |
474 | if repo.wvfs.exists(lfile): |
|
474 | if repo.wvfs.exists(lfile): | |
475 | hash = hashfile(file) |
|
475 | hash = hashfile(file) | |
476 | executable = getexecutable(file) |
|
476 | executable = getexecutable(file) | |
477 | writestandin(repo, standin, hash, executable) |
|
477 | writestandin(repo, standin, hash, executable) | |
478 | else: |
|
478 | else: | |
479 | raise error.Abort(_(b'%s: file not found!') % lfile) |
|
479 | raise error.Abort(_(b'%s: file not found!') % lfile) | |
480 |
|
480 | |||
481 |
|
481 | |||
482 | def readasstandin(fctx): |
|
482 | def readasstandin(fctx): | |
483 | """read hex hash from given filectx of standin file |
|
483 | """read hex hash from given filectx of standin file | |
484 |
|
484 | |||
485 | This encapsulates how "standin" data is stored into storage layer.""" |
|
485 | This encapsulates how "standin" data is stored into storage layer.""" | |
486 | return fctx.data().strip() |
|
486 | return fctx.data().strip() | |
487 |
|
487 | |||
488 |
|
488 | |||
489 | def writestandin(repo, standin, hash, executable): |
|
489 | def writestandin(repo, standin, hash, executable): | |
490 | '''write hash to <repo.root>/<standin>''' |
|
490 | '''write hash to <repo.root>/<standin>''' | |
491 | repo.wwrite(standin, hash + b'\n', executable and b'x' or b'') |
|
491 | repo.wwrite(standin, hash + b'\n', executable and b'x' or b'') | |
492 |
|
492 | |||
493 |
|
493 | |||
494 | def copyandhash(instream, outfile): |
|
494 | def copyandhash(instream, outfile): | |
495 | """Read bytes from instream (iterable) and write them to outfile, |
|
495 | """Read bytes from instream (iterable) and write them to outfile, | |
496 | computing the SHA-1 hash of the data along the way. Return the hash.""" |
|
496 | computing the SHA-1 hash of the data along the way. Return the hash.""" | |
497 | hasher = hashutil.sha1(b'') |
|
497 | hasher = hashutil.sha1(b'') | |
498 | for data in instream: |
|
498 | for data in instream: | |
499 | hasher.update(data) |
|
499 | hasher.update(data) | |
500 | outfile.write(data) |
|
500 | outfile.write(data) | |
501 | return hex(hasher.digest()) |
|
501 | return hex(hasher.digest()) | |
502 |
|
502 | |||
503 |
|
503 | |||
504 | def hashfile(file): |
|
504 | def hashfile(file): | |
505 | if not os.path.exists(file): |
|
505 | if not os.path.exists(file): | |
506 | return b'' |
|
506 | return b'' | |
507 | with open(file, b'rb') as fd: |
|
507 | with open(file, b'rb') as fd: | |
508 | return hexsha1(fd) |
|
508 | return hexsha1(fd) | |
509 |
|
509 | |||
510 |
|
510 | |||
511 | def getexecutable(filename): |
|
511 | def getexecutable(filename): | |
512 | mode = os.stat(filename).st_mode |
|
512 | mode = os.stat(filename).st_mode | |
513 | return ( |
|
513 | return ( | |
514 | (mode & stat.S_IXUSR) |
|
514 | (mode & stat.S_IXUSR) | |
515 | and (mode & stat.S_IXGRP) |
|
515 | and (mode & stat.S_IXGRP) | |
516 | and (mode & stat.S_IXOTH) |
|
516 | and (mode & stat.S_IXOTH) | |
517 | ) |
|
517 | ) | |
518 |
|
518 | |||
519 |
|
519 | |||
520 | def urljoin(first, second, *arg): |
|
520 | def urljoin(first, second, *arg): | |
521 | def join(left, right): |
|
521 | def join(left, right): | |
522 | if not left.endswith(b'/'): |
|
522 | if not left.endswith(b'/'): | |
523 | left += b'/' |
|
523 | left += b'/' | |
524 | if right.startswith(b'/'): |
|
524 | if right.startswith(b'/'): | |
525 | right = right[1:] |
|
525 | right = right[1:] | |
526 | return left + right |
|
526 | return left + right | |
527 |
|
527 | |||
528 | url = join(first, second) |
|
528 | url = join(first, second) | |
529 | for a in arg: |
|
529 | for a in arg: | |
530 | url = join(url, a) |
|
530 | url = join(url, a) | |
531 | return url |
|
531 | return url | |
532 |
|
532 | |||
533 |
|
533 | |||
534 | def hexsha1(fileobj): |
|
534 | def hexsha1(fileobj): | |
535 | """hexsha1 returns the hex-encoded sha1 sum of the data in the file-like |
|
535 | """hexsha1 returns the hex-encoded sha1 sum of the data in the file-like | |
536 | object data""" |
|
536 | object data""" | |
537 | h = hashutil.sha1() |
|
537 | h = hashutil.sha1() | |
538 | for chunk in util.filechunkiter(fileobj): |
|
538 | for chunk in util.filechunkiter(fileobj): | |
539 | h.update(chunk) |
|
539 | h.update(chunk) | |
540 | return hex(h.digest()) |
|
540 | return hex(h.digest()) | |
541 |
|
541 | |||
542 |
|
542 | |||
543 | def httpsendfile(ui, filename): |
|
543 | def httpsendfile(ui, filename): | |
544 | return httpconnection.httpsendfile(ui, filename, b'rb') |
|
544 | return httpconnection.httpsendfile(ui, filename, b'rb') | |
545 |
|
545 | |||
546 |
|
546 | |||
547 | def unixpath(path): |
|
547 | def unixpath(path): | |
548 | '''Return a version of path normalized for use with the lfdirstate.''' |
|
548 | '''Return a version of path normalized for use with the lfdirstate.''' | |
549 | return util.pconvert(os.path.normpath(path)) |
|
549 | return util.pconvert(os.path.normpath(path)) | |
550 |
|
550 | |||
551 |
|
551 | |||
552 | def islfilesrepo(repo): |
|
552 | def islfilesrepo(repo): | |
553 | '''Return true if the repo is a largefile repo.''' |
|
553 | '''Return true if the repo is a largefile repo.''' | |
554 | if b'largefiles' in repo.requirements and any( |
|
554 | if b'largefiles' in repo.requirements and any( | |
555 | shortnameslash in f[1] for f in repo.store.datafiles() |
|
555 | shortnameslash in entry.unencoded_path | |
|
556 | for entry in repo.store.datafiles() | |||
556 | ): |
|
557 | ): | |
557 | return True |
|
558 | return True | |
558 |
|
559 | |||
559 | return any(openlfdirstate(repo.ui, repo, False)) |
|
560 | return any(openlfdirstate(repo.ui, repo, False)) | |
560 |
|
561 | |||
561 |
|
562 | |||
562 | class storeprotonotcapable(Exception): |
|
563 | class storeprotonotcapable(Exception): | |
563 | def __init__(self, storetypes): |
|
564 | def __init__(self, storetypes): | |
564 | self.storetypes = storetypes |
|
565 | self.storetypes = storetypes | |
565 |
|
566 | |||
566 |
|
567 | |||
567 | def getstandinsstate(repo): |
|
568 | def getstandinsstate(repo): | |
568 | standins = [] |
|
569 | standins = [] | |
569 | matcher = getstandinmatcher(repo) |
|
570 | matcher = getstandinmatcher(repo) | |
570 | wctx = repo[None] |
|
571 | wctx = repo[None] | |
571 | for standin in repo.dirstate.walk( |
|
572 | for standin in repo.dirstate.walk( | |
572 | matcher, subrepos=[], unknown=False, ignored=False |
|
573 | matcher, subrepos=[], unknown=False, ignored=False | |
573 | ): |
|
574 | ): | |
574 | lfile = splitstandin(standin) |
|
575 | lfile = splitstandin(standin) | |
575 | try: |
|
576 | try: | |
576 | hash = readasstandin(wctx[standin]) |
|
577 | hash = readasstandin(wctx[standin]) | |
577 | except IOError: |
|
578 | except IOError: | |
578 | hash = None |
|
579 | hash = None | |
579 | standins.append((lfile, hash)) |
|
580 | standins.append((lfile, hash)) | |
580 | return standins |
|
581 | return standins | |
581 |
|
582 | |||
582 |
|
583 | |||
583 | def synclfdirstate(repo, lfdirstate, lfile, normallookup): |
|
584 | def synclfdirstate(repo, lfdirstate, lfile, normallookup): | |
584 | lfstandin = standin(lfile) |
|
585 | lfstandin = standin(lfile) | |
585 | if lfstandin not in repo.dirstate: |
|
586 | if lfstandin not in repo.dirstate: | |
586 | lfdirstate.hacky_extension_update_file( |
|
587 | lfdirstate.hacky_extension_update_file( | |
587 | lfile, |
|
588 | lfile, | |
588 | p1_tracked=False, |
|
589 | p1_tracked=False, | |
589 | wc_tracked=False, |
|
590 | wc_tracked=False, | |
590 | ) |
|
591 | ) | |
591 | else: |
|
592 | else: | |
592 | entry = repo.dirstate.get_entry(lfstandin) |
|
593 | entry = repo.dirstate.get_entry(lfstandin) | |
593 | lfdirstate.hacky_extension_update_file( |
|
594 | lfdirstate.hacky_extension_update_file( | |
594 | lfile, |
|
595 | lfile, | |
595 | wc_tracked=entry.tracked, |
|
596 | wc_tracked=entry.tracked, | |
596 | p1_tracked=entry.p1_tracked, |
|
597 | p1_tracked=entry.p1_tracked, | |
597 | p2_info=entry.p2_info, |
|
598 | p2_info=entry.p2_info, | |
598 | possibly_dirty=True, |
|
599 | possibly_dirty=True, | |
599 | ) |
|
600 | ) | |
600 |
|
601 | |||
601 |
|
602 | |||
602 | def markcommitted(orig, ctx, node): |
|
603 | def markcommitted(orig, ctx, node): | |
603 | repo = ctx.repo() |
|
604 | repo = ctx.repo() | |
604 |
|
605 | |||
605 | with repo.dirstate.changing_parents(repo): |
|
606 | with repo.dirstate.changing_parents(repo): | |
606 | orig(node) |
|
607 | orig(node) | |
607 |
|
608 | |||
608 | # ATTENTION: "ctx.files()" may differ from "repo[node].files()" |
|
609 | # ATTENTION: "ctx.files()" may differ from "repo[node].files()" | |
609 | # because files coming from the 2nd parent are omitted in the latter. |
|
610 | # because files coming from the 2nd parent are omitted in the latter. | |
610 | # |
|
611 | # | |
611 | # The former should be used to get targets of "synclfdirstate", |
|
612 | # The former should be used to get targets of "synclfdirstate", | |
612 | # because such files: |
|
613 | # because such files: | |
613 | # - are marked as "a" by "patch.patch()" (e.g. via transplant), and |
|
614 | # - are marked as "a" by "patch.patch()" (e.g. via transplant), and | |
614 | # - have to be marked as "n" after commit, but |
|
615 | # - have to be marked as "n" after commit, but | |
615 | # - aren't listed in "repo[node].files()" |
|
616 | # - aren't listed in "repo[node].files()" | |
616 |
|
617 | |||
617 | lfdirstate = openlfdirstate(repo.ui, repo) |
|
618 | lfdirstate = openlfdirstate(repo.ui, repo) | |
618 | for f in ctx.files(): |
|
619 | for f in ctx.files(): | |
619 | lfile = splitstandin(f) |
|
620 | lfile = splitstandin(f) | |
620 | if lfile is not None: |
|
621 | if lfile is not None: | |
621 | synclfdirstate(repo, lfdirstate, lfile, False) |
|
622 | synclfdirstate(repo, lfdirstate, lfile, False) | |
622 |
|
623 | |||
623 | # As part of committing, copy all of the largefiles into the cache. |
|
624 | # As part of committing, copy all of the largefiles into the cache. | |
624 | # |
|
625 | # | |
625 | # Using "node" instead of "ctx" implies additional "repo[node]" |
|
626 | # Using "node" instead of "ctx" implies additional "repo[node]" | |
626 | # lookup while copyalltostore(), but can omit redundant check for |
|
627 | # lookup while copyalltostore(), but can omit redundant check for | |
627 | # files comming from the 2nd parent, which should exist in store |
|
628 | # files comming from the 2nd parent, which should exist in store | |
628 | # at merging. |
|
629 | # at merging. | |
629 | copyalltostore(repo, node) |
|
630 | copyalltostore(repo, node) | |
630 |
|
631 | |||
631 |
|
632 | |||
632 | def getlfilestoupdate(oldstandins, newstandins): |
|
633 | def getlfilestoupdate(oldstandins, newstandins): | |
633 | changedstandins = set(oldstandins).symmetric_difference(set(newstandins)) |
|
634 | changedstandins = set(oldstandins).symmetric_difference(set(newstandins)) | |
634 | filelist = [] |
|
635 | filelist = [] | |
635 | for f in changedstandins: |
|
636 | for f in changedstandins: | |
636 | if f[0] not in filelist: |
|
637 | if f[0] not in filelist: | |
637 | filelist.append(f[0]) |
|
638 | filelist.append(f[0]) | |
638 | return filelist |
|
639 | return filelist | |
639 |
|
640 | |||
640 |
|
641 | |||
641 | def getlfilestoupload(repo, missing, addfunc): |
|
642 | def getlfilestoupload(repo, missing, addfunc): | |
642 | makeprogress = repo.ui.makeprogress |
|
643 | makeprogress = repo.ui.makeprogress | |
643 | with makeprogress( |
|
644 | with makeprogress( | |
644 | _(b'finding outgoing largefiles'), |
|
645 | _(b'finding outgoing largefiles'), | |
645 | unit=_(b'revisions'), |
|
646 | unit=_(b'revisions'), | |
646 | total=len(missing), |
|
647 | total=len(missing), | |
647 | ) as progress: |
|
648 | ) as progress: | |
648 | for i, n in enumerate(missing): |
|
649 | for i, n in enumerate(missing): | |
649 | progress.update(i) |
|
650 | progress.update(i) | |
650 | parents = [p for p in repo[n].parents() if p != repo.nullid] |
|
651 | parents = [p for p in repo[n].parents() if p != repo.nullid] | |
651 |
|
652 | |||
652 | with lfstatus(repo, value=False): |
|
653 | with lfstatus(repo, value=False): | |
653 | ctx = repo[n] |
|
654 | ctx = repo[n] | |
654 |
|
655 | |||
655 | files = set(ctx.files()) |
|
656 | files = set(ctx.files()) | |
656 | if len(parents) == 2: |
|
657 | if len(parents) == 2: | |
657 | mc = ctx.manifest() |
|
658 | mc = ctx.manifest() | |
658 | mp1 = ctx.p1().manifest() |
|
659 | mp1 = ctx.p1().manifest() | |
659 | mp2 = ctx.p2().manifest() |
|
660 | mp2 = ctx.p2().manifest() | |
660 | for f in mp1: |
|
661 | for f in mp1: | |
661 | if f not in mc: |
|
662 | if f not in mc: | |
662 | files.add(f) |
|
663 | files.add(f) | |
663 | for f in mp2: |
|
664 | for f in mp2: | |
664 | if f not in mc: |
|
665 | if f not in mc: | |
665 | files.add(f) |
|
666 | files.add(f) | |
666 | for f in mc: |
|
667 | for f in mc: | |
667 | if mc[f] != mp1.get(f, None) or mc[f] != mp2.get(f, None): |
|
668 | if mc[f] != mp1.get(f, None) or mc[f] != mp2.get(f, None): | |
668 | files.add(f) |
|
669 | files.add(f) | |
669 | for fn in files: |
|
670 | for fn in files: | |
670 | if isstandin(fn) and fn in ctx: |
|
671 | if isstandin(fn) and fn in ctx: | |
671 | addfunc(fn, readasstandin(ctx[fn])) |
|
672 | addfunc(fn, readasstandin(ctx[fn])) | |
672 |
|
673 | |||
673 |
|
674 | |||
674 | def updatestandinsbymatch(repo, match): |
|
675 | def updatestandinsbymatch(repo, match): | |
675 | """Update standins in the working directory according to specified match |
|
676 | """Update standins in the working directory according to specified match | |
676 |
|
677 | |||
677 | This returns (possibly modified) ``match`` object to be used for |
|
678 | This returns (possibly modified) ``match`` object to be used for | |
678 | subsequent commit process. |
|
679 | subsequent commit process. | |
679 | """ |
|
680 | """ | |
680 |
|
681 | |||
681 | ui = repo.ui |
|
682 | ui = repo.ui | |
682 |
|
683 | |||
683 | # Case 1: user calls commit with no specific files or |
|
684 | # Case 1: user calls commit with no specific files or | |
684 | # include/exclude patterns: refresh and commit all files that |
|
685 | # include/exclude patterns: refresh and commit all files that | |
685 | # are "dirty". |
|
686 | # are "dirty". | |
686 | if match is None or match.always(): |
|
687 | if match is None or match.always(): | |
687 | # Spend a bit of time here to get a list of files we know |
|
688 | # Spend a bit of time here to get a list of files we know | |
688 | # are modified so we can compare only against those. |
|
689 | # are modified so we can compare only against those. | |
689 | # It can cost a lot of time (several seconds) |
|
690 | # It can cost a lot of time (several seconds) | |
690 | # otherwise to update all standins if the largefiles are |
|
691 | # otherwise to update all standins if the largefiles are | |
691 | # large. |
|
692 | # large. | |
692 | dirtymatch = matchmod.always() |
|
693 | dirtymatch = matchmod.always() | |
693 | with repo.dirstate.running_status(repo): |
|
694 | with repo.dirstate.running_status(repo): | |
694 | lfdirstate = openlfdirstate(ui, repo) |
|
695 | lfdirstate = openlfdirstate(ui, repo) | |
695 | unsure, s, mtime_boundary = lfdirstate.status( |
|
696 | unsure, s, mtime_boundary = lfdirstate.status( | |
696 | dirtymatch, |
|
697 | dirtymatch, | |
697 | subrepos=[], |
|
698 | subrepos=[], | |
698 | ignored=False, |
|
699 | ignored=False, | |
699 | clean=False, |
|
700 | clean=False, | |
700 | unknown=False, |
|
701 | unknown=False, | |
701 | ) |
|
702 | ) | |
702 | modifiedfiles = unsure + s.modified + s.added + s.removed |
|
703 | modifiedfiles = unsure + s.modified + s.added + s.removed | |
703 | lfiles = listlfiles(repo) |
|
704 | lfiles = listlfiles(repo) | |
704 | # this only loops through largefiles that exist (not |
|
705 | # this only loops through largefiles that exist (not | |
705 | # removed/renamed) |
|
706 | # removed/renamed) | |
706 | for lfile in lfiles: |
|
707 | for lfile in lfiles: | |
707 | if lfile in modifiedfiles: |
|
708 | if lfile in modifiedfiles: | |
708 | fstandin = standin(lfile) |
|
709 | fstandin = standin(lfile) | |
709 | if repo.wvfs.exists(fstandin): |
|
710 | if repo.wvfs.exists(fstandin): | |
710 | # this handles the case where a rebase is being |
|
711 | # this handles the case where a rebase is being | |
711 | # performed and the working copy is not updated |
|
712 | # performed and the working copy is not updated | |
712 | # yet. |
|
713 | # yet. | |
713 | if repo.wvfs.exists(lfile): |
|
714 | if repo.wvfs.exists(lfile): | |
714 | updatestandin(repo, lfile, fstandin) |
|
715 | updatestandin(repo, lfile, fstandin) | |
715 |
|
716 | |||
716 | return match |
|
717 | return match | |
717 |
|
718 | |||
718 | lfiles = listlfiles(repo) |
|
719 | lfiles = listlfiles(repo) | |
719 | match._files = repo._subdirlfs(match.files(), lfiles) |
|
720 | match._files = repo._subdirlfs(match.files(), lfiles) | |
720 |
|
721 | |||
721 | # Case 2: user calls commit with specified patterns: refresh |
|
722 | # Case 2: user calls commit with specified patterns: refresh | |
722 | # any matching big files. |
|
723 | # any matching big files. | |
723 | smatcher = composestandinmatcher(repo, match) |
|
724 | smatcher = composestandinmatcher(repo, match) | |
724 | standins = repo.dirstate.walk( |
|
725 | standins = repo.dirstate.walk( | |
725 | smatcher, subrepos=[], unknown=False, ignored=False |
|
726 | smatcher, subrepos=[], unknown=False, ignored=False | |
726 | ) |
|
727 | ) | |
727 |
|
728 | |||
728 | # No matching big files: get out of the way and pass control to |
|
729 | # No matching big files: get out of the way and pass control to | |
729 | # the usual commit() method. |
|
730 | # the usual commit() method. | |
730 | if not standins: |
|
731 | if not standins: | |
731 | return match |
|
732 | return match | |
732 |
|
733 | |||
733 | # Refresh all matching big files. It's possible that the |
|
734 | # Refresh all matching big files. It's possible that the | |
734 | # commit will end up failing, in which case the big files will |
|
735 | # commit will end up failing, in which case the big files will | |
735 | # stay refreshed. No harm done: the user modified them and |
|
736 | # stay refreshed. No harm done: the user modified them and | |
736 | # asked to commit them, so sooner or later we're going to |
|
737 | # asked to commit them, so sooner or later we're going to | |
737 | # refresh the standins. Might as well leave them refreshed. |
|
738 | # refresh the standins. Might as well leave them refreshed. | |
738 | lfdirstate = openlfdirstate(ui, repo) |
|
739 | lfdirstate = openlfdirstate(ui, repo) | |
739 | for fstandin in standins: |
|
740 | for fstandin in standins: | |
740 | lfile = splitstandin(fstandin) |
|
741 | lfile = splitstandin(fstandin) | |
741 | if lfdirstate.get_entry(lfile).tracked: |
|
742 | if lfdirstate.get_entry(lfile).tracked: | |
742 | updatestandin(repo, lfile, fstandin) |
|
743 | updatestandin(repo, lfile, fstandin) | |
743 |
|
744 | |||
744 | # Cook up a new matcher that only matches regular files or |
|
745 | # Cook up a new matcher that only matches regular files or | |
745 | # standins corresponding to the big files requested by the |
|
746 | # standins corresponding to the big files requested by the | |
746 | # user. Have to modify _files to prevent commit() from |
|
747 | # user. Have to modify _files to prevent commit() from | |
747 | # complaining "not tracked" for big files. |
|
748 | # complaining "not tracked" for big files. | |
748 | match = copy.copy(match) |
|
749 | match = copy.copy(match) | |
749 | origmatchfn = match.matchfn |
|
750 | origmatchfn = match.matchfn | |
750 |
|
751 | |||
751 | # Check both the list of largefiles and the list of |
|
752 | # Check both the list of largefiles and the list of | |
752 | # standins because if a largefile was removed, it |
|
753 | # standins because if a largefile was removed, it | |
753 | # won't be in the list of largefiles at this point |
|
754 | # won't be in the list of largefiles at this point | |
754 | match._files += sorted(standins) |
|
755 | match._files += sorted(standins) | |
755 |
|
756 | |||
756 | actualfiles = [] |
|
757 | actualfiles = [] | |
757 | for f in match._files: |
|
758 | for f in match._files: | |
758 | fstandin = standin(f) |
|
759 | fstandin = standin(f) | |
759 |
|
760 | |||
760 | # For largefiles, only one of the normal and standin should be |
|
761 | # For largefiles, only one of the normal and standin should be | |
761 | # committed (except if one of them is a remove). In the case of a |
|
762 | # committed (except if one of them is a remove). In the case of a | |
762 | # standin removal, drop the normal file if it is unknown to dirstate. |
|
763 | # standin removal, drop the normal file if it is unknown to dirstate. | |
763 | # Thus, skip plain largefile names but keep the standin. |
|
764 | # Thus, skip plain largefile names but keep the standin. | |
764 | if f in lfiles or fstandin in standins: |
|
765 | if f in lfiles or fstandin in standins: | |
765 | if not repo.dirstate.get_entry(fstandin).removed: |
|
766 | if not repo.dirstate.get_entry(fstandin).removed: | |
766 | if not repo.dirstate.get_entry(f).removed: |
|
767 | if not repo.dirstate.get_entry(f).removed: | |
767 | continue |
|
768 | continue | |
768 | elif not repo.dirstate.get_entry(f).any_tracked: |
|
769 | elif not repo.dirstate.get_entry(f).any_tracked: | |
769 | continue |
|
770 | continue | |
770 |
|
771 | |||
771 | actualfiles.append(f) |
|
772 | actualfiles.append(f) | |
772 | match._files = actualfiles |
|
773 | match._files = actualfiles | |
773 |
|
774 | |||
774 | def matchfn(f): |
|
775 | def matchfn(f): | |
775 | if origmatchfn(f): |
|
776 | if origmatchfn(f): | |
776 | return f not in lfiles |
|
777 | return f not in lfiles | |
777 | else: |
|
778 | else: | |
778 | return f in standins |
|
779 | return f in standins | |
779 |
|
780 | |||
780 | match.matchfn = matchfn |
|
781 | match.matchfn = matchfn | |
781 |
|
782 | |||
782 | return match |
|
783 | return match | |
783 |
|
784 | |||
784 |
|
785 | |||
785 | class automatedcommithook: |
|
786 | class automatedcommithook: | |
786 | """Stateful hook to update standins at the 1st commit of resuming |
|
787 | """Stateful hook to update standins at the 1st commit of resuming | |
787 |
|
788 | |||
788 | For efficiency, updating standins in the working directory should |
|
789 | For efficiency, updating standins in the working directory should | |
789 | be avoided while automated committing (like rebase, transplant and |
|
790 | be avoided while automated committing (like rebase, transplant and | |
790 | so on), because they should be updated before committing. |
|
791 | so on), because they should be updated before committing. | |
791 |
|
792 | |||
792 | But the 1st commit of resuming automated committing (e.g. ``rebase |
|
793 | But the 1st commit of resuming automated committing (e.g. ``rebase | |
793 | --continue``) should update them, because largefiles may be |
|
794 | --continue``) should update them, because largefiles may be | |
794 | modified manually. |
|
795 | modified manually. | |
795 | """ |
|
796 | """ | |
796 |
|
797 | |||
797 | def __init__(self, resuming): |
|
798 | def __init__(self, resuming): | |
798 | self.resuming = resuming |
|
799 | self.resuming = resuming | |
799 |
|
800 | |||
800 | def __call__(self, repo, match): |
|
801 | def __call__(self, repo, match): | |
801 | if self.resuming: |
|
802 | if self.resuming: | |
802 | self.resuming = False # avoids updating at subsequent commits |
|
803 | self.resuming = False # avoids updating at subsequent commits | |
803 | return updatestandinsbymatch(repo, match) |
|
804 | return updatestandinsbymatch(repo, match) | |
804 | else: |
|
805 | else: | |
805 | return match |
|
806 | return match | |
806 |
|
807 | |||
807 |
|
808 | |||
808 | def getstatuswriter(ui, repo, forcibly=None): |
|
809 | def getstatuswriter(ui, repo, forcibly=None): | |
809 | """Return the function to write largefiles specific status out |
|
810 | """Return the function to write largefiles specific status out | |
810 |
|
811 | |||
811 | If ``forcibly`` is ``None``, this returns the last element of |
|
812 | If ``forcibly`` is ``None``, this returns the last element of | |
812 | ``repo._lfstatuswriters`` as "default" writer function. |
|
813 | ``repo._lfstatuswriters`` as "default" writer function. | |
813 |
|
814 | |||
814 | Otherwise, this returns the function to always write out (or |
|
815 | Otherwise, this returns the function to always write out (or | |
815 | ignore if ``not forcibly``) status. |
|
816 | ignore if ``not forcibly``) status. | |
816 | """ |
|
817 | """ | |
817 | if forcibly is None and util.safehasattr(repo, b'_largefilesenabled'): |
|
818 | if forcibly is None and util.safehasattr(repo, b'_largefilesenabled'): | |
818 | return repo._lfstatuswriters[-1] |
|
819 | return repo._lfstatuswriters[-1] | |
819 | else: |
|
820 | else: | |
820 | if forcibly: |
|
821 | if forcibly: | |
821 | return ui.status # forcibly WRITE OUT |
|
822 | return ui.status # forcibly WRITE OUT | |
822 | else: |
|
823 | else: | |
823 | return lambda *msg, **opts: None # forcibly IGNORE |
|
824 | return lambda *msg, **opts: None # forcibly IGNORE |
@@ -1,469 +1,470 b'' | |||||
1 | # Copyright 2009-2010 Gregory P. Ward |
|
1 | # Copyright 2009-2010 Gregory P. Ward | |
2 | # Copyright 2009-2010 Intelerad Medical Systems Incorporated |
|
2 | # Copyright 2009-2010 Intelerad Medical Systems Incorporated | |
3 | # Copyright 2010-2011 Fog Creek Software |
|
3 | # Copyright 2010-2011 Fog Creek Software | |
4 | # Copyright 2010-2011 Unity Technologies |
|
4 | # Copyright 2010-2011 Unity Technologies | |
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 | '''setup for largefiles repositories: reposetup''' |
|
9 | '''setup for largefiles repositories: reposetup''' | |
10 |
|
10 | |||
11 | import copy |
|
11 | import copy | |
12 |
|
12 | |||
13 | from mercurial.i18n import _ |
|
13 | from mercurial.i18n import _ | |
14 |
|
14 | |||
15 | from mercurial import ( |
|
15 | from mercurial import ( | |
16 | error, |
|
16 | error, | |
17 | extensions, |
|
17 | extensions, | |
18 | localrepo, |
|
18 | localrepo, | |
19 | match as matchmod, |
|
19 | match as matchmod, | |
20 | scmutil, |
|
20 | scmutil, | |
21 | util, |
|
21 | util, | |
22 | ) |
|
22 | ) | |
23 |
|
23 | |||
24 | from mercurial.dirstateutils import timestamp |
|
24 | from mercurial.dirstateutils import timestamp | |
25 |
|
25 | |||
26 | from . import ( |
|
26 | from . import ( | |
27 | lfcommands, |
|
27 | lfcommands, | |
28 | lfutil, |
|
28 | lfutil, | |
29 | ) |
|
29 | ) | |
30 |
|
30 | |||
31 |
|
31 | |||
32 | def reposetup(ui, repo): |
|
32 | def reposetup(ui, repo): | |
33 | # wire repositories should be given new wireproto functions |
|
33 | # wire repositories should be given new wireproto functions | |
34 | # by "proto.wirereposetup()" via "hg.wirepeersetupfuncs" |
|
34 | # by "proto.wirereposetup()" via "hg.wirepeersetupfuncs" | |
35 | if not repo.local(): |
|
35 | if not repo.local(): | |
36 | return |
|
36 | return | |
37 |
|
37 | |||
38 | class lfilesrepo(repo.__class__): |
|
38 | class lfilesrepo(repo.__class__): | |
39 | # the mark to examine whether "repo" object enables largefiles or not |
|
39 | # the mark to examine whether "repo" object enables largefiles or not | |
40 | _largefilesenabled = True |
|
40 | _largefilesenabled = True | |
41 |
|
41 | |||
42 | lfstatus = False |
|
42 | lfstatus = False | |
43 |
|
43 | |||
44 | # When lfstatus is set, return a context that gives the names |
|
44 | # When lfstatus is set, return a context that gives the names | |
45 | # of largefiles instead of their corresponding standins and |
|
45 | # of largefiles instead of their corresponding standins and | |
46 | # identifies the largefiles as always binary, regardless of |
|
46 | # identifies the largefiles as always binary, regardless of | |
47 | # their actual contents. |
|
47 | # their actual contents. | |
48 | def __getitem__(self, changeid): |
|
48 | def __getitem__(self, changeid): | |
49 | ctx = super(lfilesrepo, self).__getitem__(changeid) |
|
49 | ctx = super(lfilesrepo, self).__getitem__(changeid) | |
50 | if self.lfstatus: |
|
50 | if self.lfstatus: | |
51 |
|
51 | |||
52 | def files(orig): |
|
52 | def files(orig): | |
53 | filenames = orig() |
|
53 | filenames = orig() | |
54 | return [lfutil.splitstandin(f) or f for f in filenames] |
|
54 | return [lfutil.splitstandin(f) or f for f in filenames] | |
55 |
|
55 | |||
56 | extensions.wrapfunction(ctx, 'files', files) |
|
56 | extensions.wrapfunction(ctx, 'files', files) | |
57 |
|
57 | |||
58 | def manifest(orig): |
|
58 | def manifest(orig): | |
59 | man1 = orig() |
|
59 | man1 = orig() | |
60 |
|
60 | |||
61 | class lfilesmanifest(man1.__class__): |
|
61 | class lfilesmanifest(man1.__class__): | |
62 | def __contains__(self, filename): |
|
62 | def __contains__(self, filename): | |
63 | orig = super(lfilesmanifest, self).__contains__ |
|
63 | orig = super(lfilesmanifest, self).__contains__ | |
64 | return orig(filename) or orig( |
|
64 | return orig(filename) or orig( | |
65 | lfutil.standin(filename) |
|
65 | lfutil.standin(filename) | |
66 | ) |
|
66 | ) | |
67 |
|
67 | |||
68 | man1.__class__ = lfilesmanifest |
|
68 | man1.__class__ = lfilesmanifest | |
69 | return man1 |
|
69 | return man1 | |
70 |
|
70 | |||
71 | extensions.wrapfunction(ctx, 'manifest', manifest) |
|
71 | extensions.wrapfunction(ctx, 'manifest', manifest) | |
72 |
|
72 | |||
73 | def filectx(orig, path, fileid=None, filelog=None): |
|
73 | def filectx(orig, path, fileid=None, filelog=None): | |
74 | try: |
|
74 | try: | |
75 | if filelog is not None: |
|
75 | if filelog is not None: | |
76 | result = orig(path, fileid, filelog) |
|
76 | result = orig(path, fileid, filelog) | |
77 | else: |
|
77 | else: | |
78 | result = orig(path, fileid) |
|
78 | result = orig(path, fileid) | |
79 | except error.LookupError: |
|
79 | except error.LookupError: | |
80 | # Adding a null character will cause Mercurial to |
|
80 | # Adding a null character will cause Mercurial to | |
81 | # identify this as a binary file. |
|
81 | # identify this as a binary file. | |
82 | if filelog is not None: |
|
82 | if filelog is not None: | |
83 | result = orig(lfutil.standin(path), fileid, filelog) |
|
83 | result = orig(lfutil.standin(path), fileid, filelog) | |
84 | else: |
|
84 | else: | |
85 | result = orig(lfutil.standin(path), fileid) |
|
85 | result = orig(lfutil.standin(path), fileid) | |
86 | olddata = result.data |
|
86 | olddata = result.data | |
87 | result.data = lambda: olddata() + b'\0' |
|
87 | result.data = lambda: olddata() + b'\0' | |
88 | return result |
|
88 | return result | |
89 |
|
89 | |||
90 | extensions.wrapfunction(ctx, 'filectx', filectx) |
|
90 | extensions.wrapfunction(ctx, 'filectx', filectx) | |
91 |
|
91 | |||
92 | return ctx |
|
92 | return ctx | |
93 |
|
93 | |||
94 | # Figure out the status of big files and insert them into the |
|
94 | # Figure out the status of big files and insert them into the | |
95 | # appropriate list in the result. Also removes standin files |
|
95 | # appropriate list in the result. Also removes standin files | |
96 | # from the listing. Revert to the original status if |
|
96 | # from the listing. Revert to the original status if | |
97 | # self.lfstatus is False. |
|
97 | # self.lfstatus is False. | |
98 | # XXX large file status is buggy when used on repo proxy. |
|
98 | # XXX large file status is buggy when used on repo proxy. | |
99 | # XXX this needs to be investigated. |
|
99 | # XXX this needs to be investigated. | |
100 | @localrepo.unfilteredmethod |
|
100 | @localrepo.unfilteredmethod | |
101 | def status( |
|
101 | def status( | |
102 | self, |
|
102 | self, | |
103 | node1=b'.', |
|
103 | node1=b'.', | |
104 | node2=None, |
|
104 | node2=None, | |
105 | match=None, |
|
105 | match=None, | |
106 | ignored=False, |
|
106 | ignored=False, | |
107 | clean=False, |
|
107 | clean=False, | |
108 | unknown=False, |
|
108 | unknown=False, | |
109 | listsubrepos=False, |
|
109 | listsubrepos=False, | |
110 | ): |
|
110 | ): | |
111 | listignored, listclean, listunknown = ignored, clean, unknown |
|
111 | listignored, listclean, listunknown = ignored, clean, unknown | |
112 | orig = super(lfilesrepo, self).status |
|
112 | orig = super(lfilesrepo, self).status | |
113 | if not self.lfstatus: |
|
113 | if not self.lfstatus: | |
114 | return orig( |
|
114 | return orig( | |
115 | node1, |
|
115 | node1, | |
116 | node2, |
|
116 | node2, | |
117 | match, |
|
117 | match, | |
118 | listignored, |
|
118 | listignored, | |
119 | listclean, |
|
119 | listclean, | |
120 | listunknown, |
|
120 | listunknown, | |
121 | listsubrepos, |
|
121 | listsubrepos, | |
122 | ) |
|
122 | ) | |
123 |
|
123 | |||
124 | # some calls in this function rely on the old version of status |
|
124 | # some calls in this function rely on the old version of status | |
125 | self.lfstatus = False |
|
125 | self.lfstatus = False | |
126 | ctx1 = self[node1] |
|
126 | ctx1 = self[node1] | |
127 | ctx2 = self[node2] |
|
127 | ctx2 = self[node2] | |
128 | working = ctx2.rev() is None |
|
128 | working = ctx2.rev() is None | |
129 | parentworking = working and ctx1 == self[b'.'] |
|
129 | parentworking = working and ctx1 == self[b'.'] | |
130 |
|
130 | |||
131 | if match is None: |
|
131 | if match is None: | |
132 | match = matchmod.always() |
|
132 | match = matchmod.always() | |
133 |
|
133 | |||
134 | try: |
|
134 | try: | |
135 | # updating the dirstate is optional |
|
135 | # updating the dirstate is optional | |
136 | # so we don't wait on the lock |
|
136 | # so we don't wait on the lock | |
137 | wlock = self.wlock(False) |
|
137 | wlock = self.wlock(False) | |
138 | gotlock = True |
|
138 | gotlock = True | |
139 | except error.LockError: |
|
139 | except error.LockError: | |
140 | wlock = util.nullcontextmanager() |
|
140 | wlock = util.nullcontextmanager() | |
141 | gotlock = False |
|
141 | gotlock = False | |
142 | with wlock, self.dirstate.running_status(self): |
|
142 | with wlock, self.dirstate.running_status(self): | |
143 |
|
143 | |||
144 | # First check if paths or patterns were specified on the |
|
144 | # First check if paths or patterns were specified on the | |
145 | # command line. If there were, and they don't match any |
|
145 | # command line. If there were, and they don't match any | |
146 | # largefiles, we should just bail here and let super |
|
146 | # largefiles, we should just bail here and let super | |
147 | # handle it -- thus gaining a big performance boost. |
|
147 | # handle it -- thus gaining a big performance boost. | |
148 | lfdirstate = lfutil.openlfdirstate(ui, self) |
|
148 | lfdirstate = lfutil.openlfdirstate(ui, self) | |
149 | if not match.always(): |
|
149 | if not match.always(): | |
150 | for f in lfdirstate: |
|
150 | for f in lfdirstate: | |
151 | if match(f): |
|
151 | if match(f): | |
152 | break |
|
152 | break | |
153 | else: |
|
153 | else: | |
154 | return orig( |
|
154 | return orig( | |
155 | node1, |
|
155 | node1, | |
156 | node2, |
|
156 | node2, | |
157 | match, |
|
157 | match, | |
158 | listignored, |
|
158 | listignored, | |
159 | listclean, |
|
159 | listclean, | |
160 | listunknown, |
|
160 | listunknown, | |
161 | listsubrepos, |
|
161 | listsubrepos, | |
162 | ) |
|
162 | ) | |
163 |
|
163 | |||
164 | # Create a copy of match that matches standins instead |
|
164 | # Create a copy of match that matches standins instead | |
165 | # of largefiles. |
|
165 | # of largefiles. | |
166 | def tostandins(files): |
|
166 | def tostandins(files): | |
167 | if not working: |
|
167 | if not working: | |
168 | return files |
|
168 | return files | |
169 | newfiles = [] |
|
169 | newfiles = [] | |
170 | dirstate = self.dirstate |
|
170 | dirstate = self.dirstate | |
171 | for f in files: |
|
171 | for f in files: | |
172 | sf = lfutil.standin(f) |
|
172 | sf = lfutil.standin(f) | |
173 | if sf in dirstate: |
|
173 | if sf in dirstate: | |
174 | newfiles.append(sf) |
|
174 | newfiles.append(sf) | |
175 | elif dirstate.hasdir(sf): |
|
175 | elif dirstate.hasdir(sf): | |
176 | # Directory entries could be regular or |
|
176 | # Directory entries could be regular or | |
177 | # standin, check both |
|
177 | # standin, check both | |
178 | newfiles.extend((f, sf)) |
|
178 | newfiles.extend((f, sf)) | |
179 | else: |
|
179 | else: | |
180 | newfiles.append(f) |
|
180 | newfiles.append(f) | |
181 | return newfiles |
|
181 | return newfiles | |
182 |
|
182 | |||
183 | m = copy.copy(match) |
|
183 | m = copy.copy(match) | |
184 | m._files = tostandins(m._files) |
|
184 | m._files = tostandins(m._files) | |
185 |
|
185 | |||
186 | result = orig( |
|
186 | result = orig( | |
187 | node1, node2, m, ignored, clean, unknown, listsubrepos |
|
187 | node1, node2, m, ignored, clean, unknown, listsubrepos | |
188 | ) |
|
188 | ) | |
189 | if working: |
|
189 | if working: | |
190 |
|
190 | |||
191 | def sfindirstate(f): |
|
191 | def sfindirstate(f): | |
192 | sf = lfutil.standin(f) |
|
192 | sf = lfutil.standin(f) | |
193 | dirstate = self.dirstate |
|
193 | dirstate = self.dirstate | |
194 | return sf in dirstate or dirstate.hasdir(sf) |
|
194 | return sf in dirstate or dirstate.hasdir(sf) | |
195 |
|
195 | |||
196 | match._files = [f for f in match._files if sfindirstate(f)] |
|
196 | match._files = [f for f in match._files if sfindirstate(f)] | |
197 | # Don't waste time getting the ignored and unknown |
|
197 | # Don't waste time getting the ignored and unknown | |
198 | # files from lfdirstate |
|
198 | # files from lfdirstate | |
199 | unsure, s, mtime_boundary = lfdirstate.status( |
|
199 | unsure, s, mtime_boundary = lfdirstate.status( | |
200 | match, |
|
200 | match, | |
201 | subrepos=[], |
|
201 | subrepos=[], | |
202 | ignored=False, |
|
202 | ignored=False, | |
203 | clean=listclean, |
|
203 | clean=listclean, | |
204 | unknown=False, |
|
204 | unknown=False, | |
205 | ) |
|
205 | ) | |
206 | (modified, added, removed, deleted, clean) = ( |
|
206 | (modified, added, removed, deleted, clean) = ( | |
207 | s.modified, |
|
207 | s.modified, | |
208 | s.added, |
|
208 | s.added, | |
209 | s.removed, |
|
209 | s.removed, | |
210 | s.deleted, |
|
210 | s.deleted, | |
211 | s.clean, |
|
211 | s.clean, | |
212 | ) |
|
212 | ) | |
213 | if parentworking: |
|
213 | if parentworking: | |
214 | wctx = repo[None] |
|
214 | wctx = repo[None] | |
215 | for lfile in unsure: |
|
215 | for lfile in unsure: | |
216 | standin = lfutil.standin(lfile) |
|
216 | standin = lfutil.standin(lfile) | |
217 | if standin not in ctx1: |
|
217 | if standin not in ctx1: | |
218 | # from second parent |
|
218 | # from second parent | |
219 | modified.append(lfile) |
|
219 | modified.append(lfile) | |
220 | elif lfutil.readasstandin( |
|
220 | elif lfutil.readasstandin( | |
221 | ctx1[standin] |
|
221 | ctx1[standin] | |
222 | ) != lfutil.hashfile(self.wjoin(lfile)): |
|
222 | ) != lfutil.hashfile(self.wjoin(lfile)): | |
223 | modified.append(lfile) |
|
223 | modified.append(lfile) | |
224 | else: |
|
224 | else: | |
225 | if listclean: |
|
225 | if listclean: | |
226 | clean.append(lfile) |
|
226 | clean.append(lfile) | |
227 | s = wctx[lfile].lstat() |
|
227 | s = wctx[lfile].lstat() | |
228 | mode = s.st_mode |
|
228 | mode = s.st_mode | |
229 | size = s.st_size |
|
229 | size = s.st_size | |
230 | mtime = timestamp.reliable_mtime_of( |
|
230 | mtime = timestamp.reliable_mtime_of( | |
231 | s, mtime_boundary |
|
231 | s, mtime_boundary | |
232 | ) |
|
232 | ) | |
233 | if mtime is not None: |
|
233 | if mtime is not None: | |
234 | cache_data = (mode, size, mtime) |
|
234 | cache_data = (mode, size, mtime) | |
235 | lfdirstate.set_clean(lfile, cache_data) |
|
235 | lfdirstate.set_clean(lfile, cache_data) | |
236 | else: |
|
236 | else: | |
237 | tocheck = unsure + modified + added + clean |
|
237 | tocheck = unsure + modified + added + clean | |
238 | modified, added, clean = [], [], [] |
|
238 | modified, added, clean = [], [], [] | |
239 | checkexec = self.dirstate._checkexec |
|
239 | checkexec = self.dirstate._checkexec | |
240 |
|
240 | |||
241 | for lfile in tocheck: |
|
241 | for lfile in tocheck: | |
242 | standin = lfutil.standin(lfile) |
|
242 | standin = lfutil.standin(lfile) | |
243 | if standin in ctx1: |
|
243 | if standin in ctx1: | |
244 | abslfile = self.wjoin(lfile) |
|
244 | abslfile = self.wjoin(lfile) | |
245 | if ( |
|
245 | if ( | |
246 | lfutil.readasstandin(ctx1[standin]) |
|
246 | lfutil.readasstandin(ctx1[standin]) | |
247 | != lfutil.hashfile(abslfile) |
|
247 | != lfutil.hashfile(abslfile) | |
248 | ) or ( |
|
248 | ) or ( | |
249 | checkexec |
|
249 | checkexec | |
250 | and (b'x' in ctx1.flags(standin)) |
|
250 | and (b'x' in ctx1.flags(standin)) | |
251 | != bool(lfutil.getexecutable(abslfile)) |
|
251 | != bool(lfutil.getexecutable(abslfile)) | |
252 | ): |
|
252 | ): | |
253 | modified.append(lfile) |
|
253 | modified.append(lfile) | |
254 | elif listclean: |
|
254 | elif listclean: | |
255 | clean.append(lfile) |
|
255 | clean.append(lfile) | |
256 | else: |
|
256 | else: | |
257 | added.append(lfile) |
|
257 | added.append(lfile) | |
258 |
|
258 | |||
259 | # at this point, 'removed' contains largefiles |
|
259 | # at this point, 'removed' contains largefiles | |
260 | # marked as 'R' in the working context. |
|
260 | # marked as 'R' in the working context. | |
261 | # then, largefiles not managed also in the target |
|
261 | # then, largefiles not managed also in the target | |
262 | # context should be excluded from 'removed'. |
|
262 | # context should be excluded from 'removed'. | |
263 | removed = [ |
|
263 | removed = [ | |
264 | lfile |
|
264 | lfile | |
265 | for lfile in removed |
|
265 | for lfile in removed | |
266 | if lfutil.standin(lfile) in ctx1 |
|
266 | if lfutil.standin(lfile) in ctx1 | |
267 | ] |
|
267 | ] | |
268 |
|
268 | |||
269 | # Standins no longer found in lfdirstate have been deleted |
|
269 | # Standins no longer found in lfdirstate have been deleted | |
270 | for standin in ctx1.walk(lfutil.getstandinmatcher(self)): |
|
270 | for standin in ctx1.walk(lfutil.getstandinmatcher(self)): | |
271 | lfile = lfutil.splitstandin(standin) |
|
271 | lfile = lfutil.splitstandin(standin) | |
272 | if not match(lfile): |
|
272 | if not match(lfile): | |
273 | continue |
|
273 | continue | |
274 | if lfile not in lfdirstate: |
|
274 | if lfile not in lfdirstate: | |
275 | deleted.append(lfile) |
|
275 | deleted.append(lfile) | |
276 | # Sync "largefile has been removed" back to the |
|
276 | # Sync "largefile has been removed" back to the | |
277 | # standin. Removing a file as a side effect of |
|
277 | # standin. Removing a file as a side effect of | |
278 | # running status is gross, but the alternatives (if |
|
278 | # running status is gross, but the alternatives (if | |
279 | # any) are worse. |
|
279 | # any) are worse. | |
280 | self.wvfs.unlinkpath(standin, ignoremissing=True) |
|
280 | self.wvfs.unlinkpath(standin, ignoremissing=True) | |
281 |
|
281 | |||
282 | # Filter result lists |
|
282 | # Filter result lists | |
283 | result = list(result) |
|
283 | result = list(result) | |
284 |
|
284 | |||
285 | # Largefiles are not really removed when they're |
|
285 | # Largefiles are not really removed when they're | |
286 | # still in the normal dirstate. Likewise, normal |
|
286 | # still in the normal dirstate. Likewise, normal | |
287 | # files are not really removed if they are still in |
|
287 | # files are not really removed if they are still in | |
288 | # lfdirstate. This happens in merges where files |
|
288 | # lfdirstate. This happens in merges where files | |
289 | # change type. |
|
289 | # change type. | |
290 | removed = [f for f in removed if f not in self.dirstate] |
|
290 | removed = [f for f in removed if f not in self.dirstate] | |
291 | result[2] = [f for f in result[2] if f not in lfdirstate] |
|
291 | result[2] = [f for f in result[2] if f not in lfdirstate] | |
292 |
|
292 | |||
293 | lfiles = set(lfdirstate) |
|
293 | lfiles = set(lfdirstate) | |
294 | # Unknown files |
|
294 | # Unknown files | |
295 | result[4] = set(result[4]).difference(lfiles) |
|
295 | result[4] = set(result[4]).difference(lfiles) | |
296 | # Ignored files |
|
296 | # Ignored files | |
297 | result[5] = set(result[5]).difference(lfiles) |
|
297 | result[5] = set(result[5]).difference(lfiles) | |
298 | # combine normal files and largefiles |
|
298 | # combine normal files and largefiles | |
299 | normals = [ |
|
299 | normals = [ | |
300 | [fn for fn in filelist if not lfutil.isstandin(fn)] |
|
300 | [fn for fn in filelist if not lfutil.isstandin(fn)] | |
301 | for filelist in result |
|
301 | for filelist in result | |
302 | ] |
|
302 | ] | |
303 | lfstatus = ( |
|
303 | lfstatus = ( | |
304 | modified, |
|
304 | modified, | |
305 | added, |
|
305 | added, | |
306 | removed, |
|
306 | removed, | |
307 | deleted, |
|
307 | deleted, | |
308 | [], |
|
308 | [], | |
309 | [], |
|
309 | [], | |
310 | clean, |
|
310 | clean, | |
311 | ) |
|
311 | ) | |
312 | result = [ |
|
312 | result = [ | |
313 | sorted(list1 + list2) |
|
313 | sorted(list1 + list2) | |
314 | for (list1, list2) in zip(normals, lfstatus) |
|
314 | for (list1, list2) in zip(normals, lfstatus) | |
315 | ] |
|
315 | ] | |
316 | else: # not against working directory |
|
316 | else: # not against working directory | |
317 | result = [ |
|
317 | result = [ | |
318 | [lfutil.splitstandin(f) or f for f in items] |
|
318 | [lfutil.splitstandin(f) or f for f in items] | |
319 | for items in result |
|
319 | for items in result | |
320 | ] |
|
320 | ] | |
321 |
|
321 | |||
322 | if gotlock: |
|
322 | if gotlock: | |
323 | lfdirstate.write(self.currenttransaction()) |
|
323 | lfdirstate.write(self.currenttransaction()) | |
324 | else: |
|
324 | else: | |
325 | lfdirstate.invalidate() |
|
325 | lfdirstate.invalidate() | |
326 |
|
326 | |||
327 | self.lfstatus = True |
|
327 | self.lfstatus = True | |
328 | return scmutil.status(*result) |
|
328 | return scmutil.status(*result) | |
329 |
|
329 | |||
330 | def commitctx(self, ctx, *args, **kwargs): |
|
330 | def commitctx(self, ctx, *args, **kwargs): | |
331 | node = super(lfilesrepo, self).commitctx(ctx, *args, **kwargs) |
|
331 | node = super(lfilesrepo, self).commitctx(ctx, *args, **kwargs) | |
332 |
|
332 | |||
333 | class lfilesctx(ctx.__class__): |
|
333 | class lfilesctx(ctx.__class__): | |
334 | def markcommitted(self, node): |
|
334 | def markcommitted(self, node): | |
335 | orig = super(lfilesctx, self).markcommitted |
|
335 | orig = super(lfilesctx, self).markcommitted | |
336 | return lfutil.markcommitted(orig, self, node) |
|
336 | return lfutil.markcommitted(orig, self, node) | |
337 |
|
337 | |||
338 | ctx.__class__ = lfilesctx |
|
338 | ctx.__class__ = lfilesctx | |
339 | return node |
|
339 | return node | |
340 |
|
340 | |||
341 | # Before commit, largefile standins have not had their |
|
341 | # Before commit, largefile standins have not had their | |
342 | # contents updated to reflect the hash of their largefile. |
|
342 | # contents updated to reflect the hash of their largefile. | |
343 | # Do that here. |
|
343 | # Do that here. | |
344 | def commit( |
|
344 | def commit( | |
345 | self, |
|
345 | self, | |
346 | text=b"", |
|
346 | text=b"", | |
347 | user=None, |
|
347 | user=None, | |
348 | date=None, |
|
348 | date=None, | |
349 | match=None, |
|
349 | match=None, | |
350 | force=False, |
|
350 | force=False, | |
351 | editor=False, |
|
351 | editor=False, | |
352 | extra=None, |
|
352 | extra=None, | |
353 | ): |
|
353 | ): | |
354 | if extra is None: |
|
354 | if extra is None: | |
355 | extra = {} |
|
355 | extra = {} | |
356 | orig = super(lfilesrepo, self).commit |
|
356 | orig = super(lfilesrepo, self).commit | |
357 |
|
357 | |||
358 | with self.wlock(): |
|
358 | with self.wlock(): | |
359 | lfcommithook = self._lfcommithooks[-1] |
|
359 | lfcommithook = self._lfcommithooks[-1] | |
360 | match = lfcommithook(self, match) |
|
360 | match = lfcommithook(self, match) | |
361 | result = orig( |
|
361 | result = orig( | |
362 | text=text, |
|
362 | text=text, | |
363 | user=user, |
|
363 | user=user, | |
364 | date=date, |
|
364 | date=date, | |
365 | match=match, |
|
365 | match=match, | |
366 | force=force, |
|
366 | force=force, | |
367 | editor=editor, |
|
367 | editor=editor, | |
368 | extra=extra, |
|
368 | extra=extra, | |
369 | ) |
|
369 | ) | |
370 | return result |
|
370 | return result | |
371 |
|
371 | |||
372 | # TODO: _subdirlfs should be moved into "lfutil.py", because |
|
372 | # TODO: _subdirlfs should be moved into "lfutil.py", because | |
373 | # it is referred only from "lfutil.updatestandinsbymatch" |
|
373 | # it is referred only from "lfutil.updatestandinsbymatch" | |
374 | def _subdirlfs(self, files, lfiles): |
|
374 | def _subdirlfs(self, files, lfiles): | |
375 | """ |
|
375 | """ | |
376 | Adjust matched file list |
|
376 | Adjust matched file list | |
377 | If we pass a directory to commit whose only committable files |
|
377 | If we pass a directory to commit whose only committable files | |
378 | are largefiles, the core commit code aborts before finding |
|
378 | are largefiles, the core commit code aborts before finding | |
379 | the largefiles. |
|
379 | the largefiles. | |
380 | So we do the following: |
|
380 | So we do the following: | |
381 | For directories that only have largefiles as matches, |
|
381 | For directories that only have largefiles as matches, | |
382 | we explicitly add the largefiles to the match list and remove |
|
382 | we explicitly add the largefiles to the match list and remove | |
383 | the directory. |
|
383 | the directory. | |
384 | In other cases, we leave the match list unmodified. |
|
384 | In other cases, we leave the match list unmodified. | |
385 | """ |
|
385 | """ | |
386 | actualfiles = [] |
|
386 | actualfiles = [] | |
387 | dirs = [] |
|
387 | dirs = [] | |
388 | regulars = [] |
|
388 | regulars = [] | |
389 |
|
389 | |||
390 | for f in files: |
|
390 | for f in files: | |
391 | if lfutil.isstandin(f + b'/'): |
|
391 | if lfutil.isstandin(f + b'/'): | |
392 | raise error.Abort( |
|
392 | raise error.Abort( | |
393 | _(b'file "%s" is a largefile standin') % f, |
|
393 | _(b'file "%s" is a largefile standin') % f, | |
394 | hint=b'commit the largefile itself instead', |
|
394 | hint=b'commit the largefile itself instead', | |
395 | ) |
|
395 | ) | |
396 | # Scan directories |
|
396 | # Scan directories | |
397 | if self.wvfs.isdir(f): |
|
397 | if self.wvfs.isdir(f): | |
398 | dirs.append(f) |
|
398 | dirs.append(f) | |
399 | else: |
|
399 | else: | |
400 | regulars.append(f) |
|
400 | regulars.append(f) | |
401 |
|
401 | |||
402 | for f in dirs: |
|
402 | for f in dirs: | |
403 | matcheddir = False |
|
403 | matcheddir = False | |
404 | d = self.dirstate.normalize(f) + b'/' |
|
404 | d = self.dirstate.normalize(f) + b'/' | |
405 | # Check for matched normal files |
|
405 | # Check for matched normal files | |
406 | for mf in regulars: |
|
406 | for mf in regulars: | |
407 | if self.dirstate.normalize(mf).startswith(d): |
|
407 | if self.dirstate.normalize(mf).startswith(d): | |
408 | actualfiles.append(f) |
|
408 | actualfiles.append(f) | |
409 | matcheddir = True |
|
409 | matcheddir = True | |
410 | break |
|
410 | break | |
411 | if not matcheddir: |
|
411 | if not matcheddir: | |
412 | # If no normal match, manually append |
|
412 | # If no normal match, manually append | |
413 | # any matching largefiles |
|
413 | # any matching largefiles | |
414 | for lf in lfiles: |
|
414 | for lf in lfiles: | |
415 | if self.dirstate.normalize(lf).startswith(d): |
|
415 | if self.dirstate.normalize(lf).startswith(d): | |
416 | actualfiles.append(lf) |
|
416 | actualfiles.append(lf) | |
417 | if not matcheddir: |
|
417 | if not matcheddir: | |
418 | # There may still be normal files in the dir, so |
|
418 | # There may still be normal files in the dir, so | |
419 | # add a directory to the list, which |
|
419 | # add a directory to the list, which | |
420 | # forces status/dirstate to walk all files and |
|
420 | # forces status/dirstate to walk all files and | |
421 | # call the match function on the matcher, even |
|
421 | # call the match function on the matcher, even | |
422 | # on case sensitive filesystems. |
|
422 | # on case sensitive filesystems. | |
423 | actualfiles.append(b'.') |
|
423 | actualfiles.append(b'.') | |
424 | matcheddir = True |
|
424 | matcheddir = True | |
425 | # Nothing in dir, so readd it |
|
425 | # Nothing in dir, so readd it | |
426 | # and let commit reject it |
|
426 | # and let commit reject it | |
427 | if not matcheddir: |
|
427 | if not matcheddir: | |
428 | actualfiles.append(f) |
|
428 | actualfiles.append(f) | |
429 |
|
429 | |||
430 | # Always add normal files |
|
430 | # Always add normal files | |
431 | actualfiles += regulars |
|
431 | actualfiles += regulars | |
432 | return actualfiles |
|
432 | return actualfiles | |
433 |
|
433 | |||
434 | repo.__class__ = lfilesrepo |
|
434 | repo.__class__ = lfilesrepo | |
435 |
|
435 | |||
436 | # stack of hooks being executed before committing. |
|
436 | # stack of hooks being executed before committing. | |
437 | # only last element ("_lfcommithooks[-1]") is used for each committing. |
|
437 | # only last element ("_lfcommithooks[-1]") is used for each committing. | |
438 | repo._lfcommithooks = [lfutil.updatestandinsbymatch] |
|
438 | repo._lfcommithooks = [lfutil.updatestandinsbymatch] | |
439 |
|
439 | |||
440 | # Stack of status writer functions taking "*msg, **opts" arguments |
|
440 | # Stack of status writer functions taking "*msg, **opts" arguments | |
441 | # like "ui.status()". Only last element ("_lfstatuswriters[-1]") |
|
441 | # like "ui.status()". Only last element ("_lfstatuswriters[-1]") | |
442 | # is used to write status out. |
|
442 | # is used to write status out. | |
443 | repo._lfstatuswriters = [ui.status] |
|
443 | repo._lfstatuswriters = [ui.status] | |
444 |
|
444 | |||
445 | def prepushoutgoinghook(pushop): |
|
445 | def prepushoutgoinghook(pushop): | |
446 | """Push largefiles for pushop before pushing revisions.""" |
|
446 | """Push largefiles for pushop before pushing revisions.""" | |
447 | lfrevs = pushop.lfrevs |
|
447 | lfrevs = pushop.lfrevs | |
448 | if lfrevs is None: |
|
448 | if lfrevs is None: | |
449 | lfrevs = pushop.outgoing.missing |
|
449 | lfrevs = pushop.outgoing.missing | |
450 | if lfrevs: |
|
450 | if lfrevs: | |
451 | toupload = set() |
|
451 | toupload = set() | |
452 | addfunc = lambda fn, lfhash: toupload.add(lfhash) |
|
452 | addfunc = lambda fn, lfhash: toupload.add(lfhash) | |
453 | lfutil.getlfilestoupload(pushop.repo, lfrevs, addfunc) |
|
453 | lfutil.getlfilestoupload(pushop.repo, lfrevs, addfunc) | |
454 | lfcommands.uploadlfiles(ui, pushop.repo, pushop.remote, toupload) |
|
454 | lfcommands.uploadlfiles(ui, pushop.repo, pushop.remote, toupload) | |
455 |
|
455 | |||
456 | repo.prepushoutgoinghooks.add(b"largefiles", prepushoutgoinghook) |
|
456 | repo.prepushoutgoinghooks.add(b"largefiles", prepushoutgoinghook) | |
457 |
|
457 | |||
458 | def checkrequireslfiles(ui, repo, **kwargs): |
|
458 | def checkrequireslfiles(ui, repo, **kwargs): | |
459 | with repo.lock(): |
|
459 | with repo.lock(): | |
460 | if b'largefiles' not in repo.requirements and any( |
|
460 | if b'largefiles' not in repo.requirements and any( | |
461 |
lfutil.shortname + b'/' in |
|
461 | lfutil.shortname + b'/' in entry.unencoded_path | |
|
462 | for entry in repo.store.datafiles() | |||
462 | ): |
|
463 | ): | |
463 | repo.requirements.add(b'largefiles') |
|
464 | repo.requirements.add(b'largefiles') | |
464 | scmutil.writereporequirements(repo) |
|
465 | scmutil.writereporequirements(repo) | |
465 |
|
466 | |||
466 | ui.setconfig( |
|
467 | ui.setconfig( | |
467 | b'hooks', b'changegroup.lfiles', checkrequireslfiles, b'largefiles' |
|
468 | b'hooks', b'changegroup.lfiles', checkrequireslfiles, b'largefiles' | |
468 | ) |
|
469 | ) | |
469 | ui.setconfig(b'hooks', b'commit.lfiles', checkrequireslfiles, b'largefiles') |
|
470 | ui.setconfig(b'hooks', b'commit.lfiles', checkrequireslfiles, b'largefiles') |
@@ -1,695 +1,696 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 |
|
7 | |||
8 | import itertools |
|
8 | import itertools | |
9 | import os |
|
9 | import os | |
10 |
|
10 | |||
11 | from mercurial.i18n import _ |
|
11 | from mercurial.i18n import _ | |
12 | from mercurial.node import ( |
|
12 | from mercurial.node import ( | |
13 | hex, |
|
13 | hex, | |
14 | short, |
|
14 | short, | |
15 | ) |
|
15 | ) | |
16 | from mercurial import ( |
|
16 | from mercurial import ( | |
17 | bundle2, |
|
17 | bundle2, | |
18 | cmdutil, |
|
18 | cmdutil, | |
19 | commands, |
|
19 | commands, | |
20 | discovery, |
|
20 | discovery, | |
21 | encoding, |
|
21 | encoding, | |
22 | error, |
|
22 | error, | |
23 | exchange, |
|
23 | exchange, | |
24 | extensions, |
|
24 | extensions, | |
25 | hg, |
|
25 | hg, | |
26 | narrowspec, |
|
26 | narrowspec, | |
27 | pathutil, |
|
27 | pathutil, | |
28 | pycompat, |
|
28 | pycompat, | |
29 | registrar, |
|
29 | registrar, | |
30 | repair, |
|
30 | repair, | |
31 | repoview, |
|
31 | repoview, | |
32 | requirements, |
|
32 | requirements, | |
33 | sparse, |
|
33 | sparse, | |
34 | util, |
|
34 | util, | |
35 | wireprototypes, |
|
35 | wireprototypes, | |
36 | ) |
|
36 | ) | |
37 | from mercurial.utils import ( |
|
37 | from mercurial.utils import ( | |
38 | urlutil, |
|
38 | urlutil, | |
39 | ) |
|
39 | ) | |
40 |
|
40 | |||
41 | table = {} |
|
41 | table = {} | |
42 | command = registrar.command(table) |
|
42 | command = registrar.command(table) | |
43 |
|
43 | |||
44 |
|
44 | |||
45 | def setup(): |
|
45 | def setup(): | |
46 | """Wraps user-facing mercurial commands with narrow-aware versions.""" |
|
46 | """Wraps user-facing mercurial commands with narrow-aware versions.""" | |
47 |
|
47 | |||
48 | entry = extensions.wrapcommand(commands.table, b'clone', clonenarrowcmd) |
|
48 | entry = extensions.wrapcommand(commands.table, b'clone', clonenarrowcmd) | |
49 | entry[1].append( |
|
49 | entry[1].append( | |
50 | (b'', b'narrow', None, _(b"create a narrow clone of select files")) |
|
50 | (b'', b'narrow', None, _(b"create a narrow clone of select files")) | |
51 | ) |
|
51 | ) | |
52 | entry[1].append( |
|
52 | entry[1].append( | |
53 | ( |
|
53 | ( | |
54 | b'', |
|
54 | b'', | |
55 | b'depth', |
|
55 | b'depth', | |
56 | b'', |
|
56 | b'', | |
57 | _(b"limit the history fetched by distance from heads"), |
|
57 | _(b"limit the history fetched by distance from heads"), | |
58 | ) |
|
58 | ) | |
59 | ) |
|
59 | ) | |
60 | entry[1].append((b'', b'narrowspec', b'', _(b"read narrowspecs from file"))) |
|
60 | entry[1].append((b'', b'narrowspec', b'', _(b"read narrowspecs from file"))) | |
61 | # TODO(durin42): unify sparse/narrow --include/--exclude logic a bit |
|
61 | # TODO(durin42): unify sparse/narrow --include/--exclude logic a bit | |
62 | if b'sparse' not in extensions.enabled(): |
|
62 | if b'sparse' not in extensions.enabled(): | |
63 | entry[1].append( |
|
63 | entry[1].append( | |
64 | (b'', b'include', [], _(b"specifically fetch this file/directory")) |
|
64 | (b'', b'include', [], _(b"specifically fetch this file/directory")) | |
65 | ) |
|
65 | ) | |
66 | entry[1].append( |
|
66 | entry[1].append( | |
67 | ( |
|
67 | ( | |
68 | b'', |
|
68 | b'', | |
69 | b'exclude', |
|
69 | b'exclude', | |
70 | [], |
|
70 | [], | |
71 | _(b"do not fetch this file/directory, even if included"), |
|
71 | _(b"do not fetch this file/directory, even if included"), | |
72 | ) |
|
72 | ) | |
73 | ) |
|
73 | ) | |
74 |
|
74 | |||
75 | entry = extensions.wrapcommand(commands.table, b'pull', pullnarrowcmd) |
|
75 | entry = extensions.wrapcommand(commands.table, b'pull', pullnarrowcmd) | |
76 | entry[1].append( |
|
76 | entry[1].append( | |
77 | ( |
|
77 | ( | |
78 | b'', |
|
78 | b'', | |
79 | b'depth', |
|
79 | b'depth', | |
80 | b'', |
|
80 | b'', | |
81 | _(b"limit the history fetched by distance from heads"), |
|
81 | _(b"limit the history fetched by distance from heads"), | |
82 | ) |
|
82 | ) | |
83 | ) |
|
83 | ) | |
84 |
|
84 | |||
85 | extensions.wrapcommand(commands.table, b'archive', archivenarrowcmd) |
|
85 | extensions.wrapcommand(commands.table, b'archive', archivenarrowcmd) | |
86 |
|
86 | |||
87 |
|
87 | |||
88 | def clonenarrowcmd(orig, ui, repo, *args, **opts): |
|
88 | def clonenarrowcmd(orig, ui, repo, *args, **opts): | |
89 | """Wraps clone command, so 'hg clone' first wraps localrepo.clone().""" |
|
89 | """Wraps clone command, so 'hg clone' first wraps localrepo.clone().""" | |
90 | opts = pycompat.byteskwargs(opts) |
|
90 | opts = pycompat.byteskwargs(opts) | |
91 | wrappedextraprepare = util.nullcontextmanager() |
|
91 | wrappedextraprepare = util.nullcontextmanager() | |
92 | narrowspecfile = opts[b'narrowspec'] |
|
92 | narrowspecfile = opts[b'narrowspec'] | |
93 |
|
93 | |||
94 | if narrowspecfile: |
|
94 | if narrowspecfile: | |
95 | filepath = os.path.join(encoding.getcwd(), narrowspecfile) |
|
95 | filepath = os.path.join(encoding.getcwd(), narrowspecfile) | |
96 | ui.status(_(b"reading narrowspec from '%s'\n") % filepath) |
|
96 | ui.status(_(b"reading narrowspec from '%s'\n") % filepath) | |
97 | try: |
|
97 | try: | |
98 | fdata = util.readfile(filepath) |
|
98 | fdata = util.readfile(filepath) | |
99 | except IOError as inst: |
|
99 | except IOError as inst: | |
100 | raise error.Abort( |
|
100 | raise error.Abort( | |
101 | _(b"cannot read narrowspecs from '%s': %s") |
|
101 | _(b"cannot read narrowspecs from '%s': %s") | |
102 | % (filepath, encoding.strtolocal(inst.strerror)) |
|
102 | % (filepath, encoding.strtolocal(inst.strerror)) | |
103 | ) |
|
103 | ) | |
104 |
|
104 | |||
105 | includes, excludes, profiles = sparse.parseconfig(ui, fdata, b'narrow') |
|
105 | includes, excludes, profiles = sparse.parseconfig(ui, fdata, b'narrow') | |
106 | if profiles: |
|
106 | if profiles: | |
107 | raise error.ConfigError( |
|
107 | raise error.ConfigError( | |
108 | _( |
|
108 | _( | |
109 | b"cannot specify other files using '%include' in" |
|
109 | b"cannot specify other files using '%include' in" | |
110 | b" narrowspec" |
|
110 | b" narrowspec" | |
111 | ) |
|
111 | ) | |
112 | ) |
|
112 | ) | |
113 |
|
113 | |||
114 | narrowspec.validatepatterns(includes) |
|
114 | narrowspec.validatepatterns(includes) | |
115 | narrowspec.validatepatterns(excludes) |
|
115 | narrowspec.validatepatterns(excludes) | |
116 |
|
116 | |||
117 | # narrowspec is passed so we should assume that user wants narrow clone |
|
117 | # narrowspec is passed so we should assume that user wants narrow clone | |
118 | opts[b'narrow'] = True |
|
118 | opts[b'narrow'] = True | |
119 | opts[b'include'].extend(includes) |
|
119 | opts[b'include'].extend(includes) | |
120 | opts[b'exclude'].extend(excludes) |
|
120 | opts[b'exclude'].extend(excludes) | |
121 |
|
121 | |||
122 | if opts[b'narrow']: |
|
122 | if opts[b'narrow']: | |
123 |
|
123 | |||
124 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): |
|
124 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): | |
125 | orig(pullop, kwargs) |
|
125 | orig(pullop, kwargs) | |
126 |
|
126 | |||
127 | if opts.get(b'depth'): |
|
127 | if opts.get(b'depth'): | |
128 | kwargs[b'depth'] = opts[b'depth'] |
|
128 | kwargs[b'depth'] = opts[b'depth'] | |
129 |
|
129 | |||
130 | wrappedextraprepare = extensions.wrappedfunction( |
|
130 | wrappedextraprepare = extensions.wrappedfunction( | |
131 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen |
|
131 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen | |
132 | ) |
|
132 | ) | |
133 |
|
133 | |||
134 | with wrappedextraprepare: |
|
134 | with wrappedextraprepare: | |
135 | return orig(ui, repo, *args, **pycompat.strkwargs(opts)) |
|
135 | return orig(ui, repo, *args, **pycompat.strkwargs(opts)) | |
136 |
|
136 | |||
137 |
|
137 | |||
138 | def pullnarrowcmd(orig, ui, repo, *args, **opts): |
|
138 | def pullnarrowcmd(orig, ui, repo, *args, **opts): | |
139 | """Wraps pull command to allow modifying narrow spec.""" |
|
139 | """Wraps pull command to allow modifying narrow spec.""" | |
140 | wrappedextraprepare = util.nullcontextmanager() |
|
140 | wrappedextraprepare = util.nullcontextmanager() | |
141 | if requirements.NARROW_REQUIREMENT in repo.requirements: |
|
141 | if requirements.NARROW_REQUIREMENT in repo.requirements: | |
142 |
|
142 | |||
143 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): |
|
143 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): | |
144 | orig(pullop, kwargs) |
|
144 | orig(pullop, kwargs) | |
145 | if opts.get('depth'): |
|
145 | if opts.get('depth'): | |
146 | kwargs[b'depth'] = opts['depth'] |
|
146 | kwargs[b'depth'] = opts['depth'] | |
147 |
|
147 | |||
148 | wrappedextraprepare = extensions.wrappedfunction( |
|
148 | wrappedextraprepare = extensions.wrappedfunction( | |
149 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen |
|
149 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen | |
150 | ) |
|
150 | ) | |
151 |
|
151 | |||
152 | with wrappedextraprepare: |
|
152 | with wrappedextraprepare: | |
153 | return orig(ui, repo, *args, **opts) |
|
153 | return orig(ui, repo, *args, **opts) | |
154 |
|
154 | |||
155 |
|
155 | |||
156 | def archivenarrowcmd(orig, ui, repo, *args, **opts): |
|
156 | def archivenarrowcmd(orig, ui, repo, *args, **opts): | |
157 | """Wraps archive command to narrow the default includes.""" |
|
157 | """Wraps archive command to narrow the default includes.""" | |
158 | if requirements.NARROW_REQUIREMENT in repo.requirements: |
|
158 | if requirements.NARROW_REQUIREMENT in repo.requirements: | |
159 | repo_includes, repo_excludes = repo.narrowpats |
|
159 | repo_includes, repo_excludes = repo.narrowpats | |
160 | includes = set(opts.get('include', [])) |
|
160 | includes = set(opts.get('include', [])) | |
161 | excludes = set(opts.get('exclude', [])) |
|
161 | excludes = set(opts.get('exclude', [])) | |
162 | includes, excludes, unused_invalid = narrowspec.restrictpatterns( |
|
162 | includes, excludes, unused_invalid = narrowspec.restrictpatterns( | |
163 | includes, excludes, repo_includes, repo_excludes |
|
163 | includes, excludes, repo_includes, repo_excludes | |
164 | ) |
|
164 | ) | |
165 | if includes: |
|
165 | if includes: | |
166 | opts['include'] = includes |
|
166 | opts['include'] = includes | |
167 | if excludes: |
|
167 | if excludes: | |
168 | opts['exclude'] = excludes |
|
168 | opts['exclude'] = excludes | |
169 | return orig(ui, repo, *args, **opts) |
|
169 | return orig(ui, repo, *args, **opts) | |
170 |
|
170 | |||
171 |
|
171 | |||
172 | def pullbundle2extraprepare(orig, pullop, kwargs): |
|
172 | def pullbundle2extraprepare(orig, pullop, kwargs): | |
173 | repo = pullop.repo |
|
173 | repo = pullop.repo | |
174 | if requirements.NARROW_REQUIREMENT not in repo.requirements: |
|
174 | if requirements.NARROW_REQUIREMENT not in repo.requirements: | |
175 | return orig(pullop, kwargs) |
|
175 | return orig(pullop, kwargs) | |
176 |
|
176 | |||
177 | if wireprototypes.NARROWCAP not in pullop.remote.capabilities(): |
|
177 | if wireprototypes.NARROWCAP not in pullop.remote.capabilities(): | |
178 | raise error.Abort(_(b"server does not support narrow clones")) |
|
178 | raise error.Abort(_(b"server does not support narrow clones")) | |
179 | orig(pullop, kwargs) |
|
179 | orig(pullop, kwargs) | |
180 | kwargs[b'narrow'] = True |
|
180 | kwargs[b'narrow'] = True | |
181 | include, exclude = repo.narrowpats |
|
181 | include, exclude = repo.narrowpats | |
182 | kwargs[b'oldincludepats'] = include |
|
182 | kwargs[b'oldincludepats'] = include | |
183 | kwargs[b'oldexcludepats'] = exclude |
|
183 | kwargs[b'oldexcludepats'] = exclude | |
184 | if include: |
|
184 | if include: | |
185 | kwargs[b'includepats'] = include |
|
185 | kwargs[b'includepats'] = include | |
186 | if exclude: |
|
186 | if exclude: | |
187 | kwargs[b'excludepats'] = exclude |
|
187 | kwargs[b'excludepats'] = exclude | |
188 | # calculate known nodes only in ellipses cases because in non-ellipses cases |
|
188 | # calculate known nodes only in ellipses cases because in non-ellipses cases | |
189 | # we have all the nodes |
|
189 | # we have all the nodes | |
190 | if wireprototypes.ELLIPSESCAP1 in pullop.remote.capabilities(): |
|
190 | if wireprototypes.ELLIPSESCAP1 in pullop.remote.capabilities(): | |
191 | kwargs[b'known'] = [ |
|
191 | kwargs[b'known'] = [ | |
192 | hex(ctx.node()) |
|
192 | hex(ctx.node()) | |
193 | for ctx in repo.set(b'::%ln', pullop.common) |
|
193 | for ctx in repo.set(b'::%ln', pullop.common) | |
194 | if ctx.node() != repo.nullid |
|
194 | if ctx.node() != repo.nullid | |
195 | ] |
|
195 | ] | |
196 | if not kwargs[b'known']: |
|
196 | if not kwargs[b'known']: | |
197 | # Mercurial serializes an empty list as '' and deserializes it as |
|
197 | # Mercurial serializes an empty list as '' and deserializes it as | |
198 | # [''], so delete it instead to avoid handling the empty string on |
|
198 | # [''], so delete it instead to avoid handling the empty string on | |
199 | # the server. |
|
199 | # the server. | |
200 | del kwargs[b'known'] |
|
200 | del kwargs[b'known'] | |
201 |
|
201 | |||
202 |
|
202 | |||
203 | extensions.wrapfunction( |
|
203 | extensions.wrapfunction( | |
204 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare |
|
204 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare | |
205 | ) |
|
205 | ) | |
206 |
|
206 | |||
207 |
|
207 | |||
208 | def _narrow( |
|
208 | def _narrow( | |
209 | ui, |
|
209 | ui, | |
210 | repo, |
|
210 | repo, | |
211 | remote, |
|
211 | remote, | |
212 | commoninc, |
|
212 | commoninc, | |
213 | oldincludes, |
|
213 | oldincludes, | |
214 | oldexcludes, |
|
214 | oldexcludes, | |
215 | newincludes, |
|
215 | newincludes, | |
216 | newexcludes, |
|
216 | newexcludes, | |
217 | force, |
|
217 | force, | |
218 | backup, |
|
218 | backup, | |
219 | ): |
|
219 | ): | |
220 | oldmatch = narrowspec.match(repo.root, oldincludes, oldexcludes) |
|
220 | oldmatch = narrowspec.match(repo.root, oldincludes, oldexcludes) | |
221 | newmatch = narrowspec.match(repo.root, newincludes, newexcludes) |
|
221 | newmatch = narrowspec.match(repo.root, newincludes, newexcludes) | |
222 |
|
222 | |||
223 | # This is essentially doing "hg outgoing" to find all local-only |
|
223 | # This is essentially doing "hg outgoing" to find all local-only | |
224 | # commits. We will then check that the local-only commits don't |
|
224 | # commits. We will then check that the local-only commits don't | |
225 | # have any changes to files that will be untracked. |
|
225 | # have any changes to files that will be untracked. | |
226 | unfi = repo.unfiltered() |
|
226 | unfi = repo.unfiltered() | |
227 | outgoing = discovery.findcommonoutgoing(unfi, remote, commoninc=commoninc) |
|
227 | outgoing = discovery.findcommonoutgoing(unfi, remote, commoninc=commoninc) | |
228 | ui.status(_(b'looking for local changes to affected paths\n')) |
|
228 | ui.status(_(b'looking for local changes to affected paths\n')) | |
229 | progress = ui.makeprogress( |
|
229 | progress = ui.makeprogress( | |
230 | topic=_(b'changesets'), |
|
230 | topic=_(b'changesets'), | |
231 | unit=_(b'changesets'), |
|
231 | unit=_(b'changesets'), | |
232 | total=len(outgoing.missing) + len(outgoing.excluded), |
|
232 | total=len(outgoing.missing) + len(outgoing.excluded), | |
233 | ) |
|
233 | ) | |
234 | localnodes = [] |
|
234 | localnodes = [] | |
235 | with progress: |
|
235 | with progress: | |
236 | for n in itertools.chain(outgoing.missing, outgoing.excluded): |
|
236 | for n in itertools.chain(outgoing.missing, outgoing.excluded): | |
237 | progress.increment() |
|
237 | progress.increment() | |
238 | if any(oldmatch(f) and not newmatch(f) for f in unfi[n].files()): |
|
238 | if any(oldmatch(f) and not newmatch(f) for f in unfi[n].files()): | |
239 | localnodes.append(n) |
|
239 | localnodes.append(n) | |
240 | revstostrip = unfi.revs(b'descendants(%ln)', localnodes) |
|
240 | revstostrip = unfi.revs(b'descendants(%ln)', localnodes) | |
241 | hiddenrevs = repoview.filterrevs(repo, b'visible') |
|
241 | hiddenrevs = repoview.filterrevs(repo, b'visible') | |
242 | visibletostrip = list( |
|
242 | visibletostrip = list( | |
243 | repo.changelog.node(r) for r in (revstostrip - hiddenrevs) |
|
243 | repo.changelog.node(r) for r in (revstostrip - hiddenrevs) | |
244 | ) |
|
244 | ) | |
245 | if visibletostrip: |
|
245 | if visibletostrip: | |
246 | ui.status( |
|
246 | ui.status( | |
247 | _( |
|
247 | _( | |
248 | b'The following changeset(s) or their ancestors have ' |
|
248 | b'The following changeset(s) or their ancestors have ' | |
249 | b'local changes not on the remote:\n' |
|
249 | b'local changes not on the remote:\n' | |
250 | ) |
|
250 | ) | |
251 | ) |
|
251 | ) | |
252 | maxnodes = 10 |
|
252 | maxnodes = 10 | |
253 | if ui.verbose or len(visibletostrip) <= maxnodes: |
|
253 | if ui.verbose or len(visibletostrip) <= maxnodes: | |
254 | for n in visibletostrip: |
|
254 | for n in visibletostrip: | |
255 | ui.status(b'%s\n' % short(n)) |
|
255 | ui.status(b'%s\n' % short(n)) | |
256 | else: |
|
256 | else: | |
257 | for n in visibletostrip[:maxnodes]: |
|
257 | for n in visibletostrip[:maxnodes]: | |
258 | ui.status(b'%s\n' % short(n)) |
|
258 | ui.status(b'%s\n' % short(n)) | |
259 | ui.status( |
|
259 | ui.status( | |
260 | _(b'...and %d more, use --verbose to list all\n') |
|
260 | _(b'...and %d more, use --verbose to list all\n') | |
261 | % (len(visibletostrip) - maxnodes) |
|
261 | % (len(visibletostrip) - maxnodes) | |
262 | ) |
|
262 | ) | |
263 | if not force: |
|
263 | if not force: | |
264 | raise error.StateError( |
|
264 | raise error.StateError( | |
265 | _(b'local changes found'), |
|
265 | _(b'local changes found'), | |
266 | hint=_(b'use --force-delete-local-changes to ignore'), |
|
266 | hint=_(b'use --force-delete-local-changes to ignore'), | |
267 | ) |
|
267 | ) | |
268 |
|
268 | |||
269 | with ui.uninterruptible(): |
|
269 | with ui.uninterruptible(): | |
270 | if revstostrip: |
|
270 | if revstostrip: | |
271 | tostrip = [unfi.changelog.node(r) for r in revstostrip] |
|
271 | tostrip = [unfi.changelog.node(r) for r in revstostrip] | |
272 | if repo[b'.'].node() in tostrip: |
|
272 | if repo[b'.'].node() in tostrip: | |
273 | # stripping working copy, so move to a different commit first |
|
273 | # stripping working copy, so move to a different commit first | |
274 | urev = max( |
|
274 | urev = max( | |
275 | repo.revs( |
|
275 | repo.revs( | |
276 | b'(::%n) - %ln + null', |
|
276 | b'(::%n) - %ln + null', | |
277 | repo[b'.'].node(), |
|
277 | repo[b'.'].node(), | |
278 | visibletostrip, |
|
278 | visibletostrip, | |
279 | ) |
|
279 | ) | |
280 | ) |
|
280 | ) | |
281 | hg.clean(repo, urev) |
|
281 | hg.clean(repo, urev) | |
282 | overrides = {(b'devel', b'strip-obsmarkers'): False} |
|
282 | overrides = {(b'devel', b'strip-obsmarkers'): False} | |
283 | if backup: |
|
283 | if backup: | |
284 | ui.status(_(b'moving unwanted changesets to backup\n')) |
|
284 | ui.status(_(b'moving unwanted changesets to backup\n')) | |
285 | else: |
|
285 | else: | |
286 | ui.status(_(b'deleting unwanted changesets\n')) |
|
286 | ui.status(_(b'deleting unwanted changesets\n')) | |
287 | with ui.configoverride(overrides, b'narrow'): |
|
287 | with ui.configoverride(overrides, b'narrow'): | |
288 | repair.strip(ui, unfi, tostrip, topic=b'narrow', backup=backup) |
|
288 | repair.strip(ui, unfi, tostrip, topic=b'narrow', backup=backup) | |
289 |
|
289 | |||
290 | todelete = [] |
|
290 | todelete = [] | |
291 |
for |
|
291 | for entry in repo.store.datafiles(): | |
|
292 | f = entry.unencoded_path | |||
292 | if f.startswith(b'data/'): |
|
293 | if f.startswith(b'data/'): | |
293 | file = f[5:-2] |
|
294 | file = f[5:-2] | |
294 | if not newmatch(file): |
|
295 | if not newmatch(file): | |
295 | todelete.append(f) |
|
296 | todelete.append(f) | |
296 | elif f.startswith(b'meta/'): |
|
297 | elif f.startswith(b'meta/'): | |
297 | dir = f[5:-13] |
|
298 | dir = f[5:-13] | |
298 | dirs = sorted(pathutil.dirs({dir})) + [dir] |
|
299 | dirs = sorted(pathutil.dirs({dir})) + [dir] | |
299 | include = True |
|
300 | include = True | |
300 | for d in dirs: |
|
301 | for d in dirs: | |
301 | visit = newmatch.visitdir(d) |
|
302 | visit = newmatch.visitdir(d) | |
302 | if not visit: |
|
303 | if not visit: | |
303 | include = False |
|
304 | include = False | |
304 | break |
|
305 | break | |
305 | if visit == b'all': |
|
306 | if visit == b'all': | |
306 | break |
|
307 | break | |
307 | if not include: |
|
308 | if not include: | |
308 | todelete.append(f) |
|
309 | todelete.append(f) | |
309 |
|
310 | |||
310 | repo.destroying() |
|
311 | repo.destroying() | |
311 |
|
312 | |||
312 | with repo.transaction(b'narrowing'): |
|
313 | with repo.transaction(b'narrowing'): | |
313 | # Update narrowspec before removing revlogs, so repo won't be |
|
314 | # Update narrowspec before removing revlogs, so repo won't be | |
314 | # corrupt in case of crash |
|
315 | # corrupt in case of crash | |
315 | repo.setnarrowpats(newincludes, newexcludes) |
|
316 | repo.setnarrowpats(newincludes, newexcludes) | |
316 |
|
317 | |||
317 | for f in todelete: |
|
318 | for f in todelete: | |
318 | ui.status(_(b'deleting %s\n') % f) |
|
319 | ui.status(_(b'deleting %s\n') % f) | |
319 | util.unlinkpath(repo.svfs.join(f)) |
|
320 | util.unlinkpath(repo.svfs.join(f)) | |
320 | repo.store.markremoved(f) |
|
321 | repo.store.markremoved(f) | |
321 |
|
322 | |||
322 | ui.status(_(b'deleting unwanted files from working copy\n')) |
|
323 | ui.status(_(b'deleting unwanted files from working copy\n')) | |
323 | with repo.dirstate.changing_parents(repo): |
|
324 | with repo.dirstate.changing_parents(repo): | |
324 | narrowspec.updateworkingcopy(repo, assumeclean=True) |
|
325 | narrowspec.updateworkingcopy(repo, assumeclean=True) | |
325 | narrowspec.copytoworkingcopy(repo) |
|
326 | narrowspec.copytoworkingcopy(repo) | |
326 |
|
327 | |||
327 | repo.destroyed() |
|
328 | repo.destroyed() | |
328 |
|
329 | |||
329 |
|
330 | |||
330 | def _widen( |
|
331 | def _widen( | |
331 | ui, |
|
332 | ui, | |
332 | repo, |
|
333 | repo, | |
333 | remote, |
|
334 | remote, | |
334 | commoninc, |
|
335 | commoninc, | |
335 | oldincludes, |
|
336 | oldincludes, | |
336 | oldexcludes, |
|
337 | oldexcludes, | |
337 | newincludes, |
|
338 | newincludes, | |
338 | newexcludes, |
|
339 | newexcludes, | |
339 | ): |
|
340 | ): | |
340 | # 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 | |
341 | # 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 | |
342 | # 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 | |
343 | # 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. | |
344 | # Theoretically a non-ellipses repo should be able to use narrow |
|
345 | # Theoretically a non-ellipses repo should be able to use narrow | |
345 | # functionality from an ellipses enabled server |
|
346 | # functionality from an ellipses enabled server | |
346 | remotecap = remote.capabilities() |
|
347 | remotecap = remote.capabilities() | |
347 | ellipsesremote = any( |
|
348 | ellipsesremote = any( | |
348 | cap in remotecap for cap in wireprototypes.SUPPORTED_ELLIPSESCAP |
|
349 | cap in remotecap for cap in wireprototypes.SUPPORTED_ELLIPSESCAP | |
349 | ) |
|
350 | ) | |
350 |
|
351 | |||
351 | # 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 | |
352 | # ellipses capabilities |
|
353 | # ellipses capabilities | |
353 | isoldellipses = ( |
|
354 | isoldellipses = ( | |
354 | ellipsesremote |
|
355 | ellipsesremote | |
355 | and wireprototypes.ELLIPSESCAP1 in remotecap |
|
356 | and wireprototypes.ELLIPSESCAP1 in remotecap | |
356 | and wireprototypes.ELLIPSESCAP not in remotecap |
|
357 | and wireprototypes.ELLIPSESCAP not in remotecap | |
357 | ) |
|
358 | ) | |
358 |
|
359 | |||
359 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): |
|
360 | def pullbundle2extraprepare_widen(orig, pullop, kwargs): | |
360 | orig(pullop, kwargs) |
|
361 | orig(pullop, kwargs) | |
361 | # The old{in,ex}cludepats have already been set by orig() |
|
362 | # The old{in,ex}cludepats have already been set by orig() | |
362 | kwargs[b'includepats'] = newincludes |
|
363 | kwargs[b'includepats'] = newincludes | |
363 | kwargs[b'excludepats'] = newexcludes |
|
364 | kwargs[b'excludepats'] = newexcludes | |
364 |
|
365 | |||
365 | wrappedextraprepare = extensions.wrappedfunction( |
|
366 | wrappedextraprepare = extensions.wrappedfunction( | |
366 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen |
|
367 | exchange, b'_pullbundle2extraprepare', pullbundle2extraprepare_widen | |
367 | ) |
|
368 | ) | |
368 |
|
369 | |||
369 | # define a function that narrowbundle2 can call after creating the |
|
370 | # define a function that narrowbundle2 can call after creating the | |
370 | # backup bundle, but before applying the bundle from the server |
|
371 | # backup bundle, but before applying the bundle from the server | |
371 | def setnewnarrowpats(): |
|
372 | def setnewnarrowpats(): | |
372 | repo.setnarrowpats(newincludes, newexcludes) |
|
373 | repo.setnarrowpats(newincludes, newexcludes) | |
373 |
|
374 | |||
374 | repo.setnewnarrowpats = setnewnarrowpats |
|
375 | repo.setnewnarrowpats = setnewnarrowpats | |
375 | # silence the devel-warning of applying an empty changegroup |
|
376 | # silence the devel-warning of applying an empty changegroup | |
376 | overrides = {(b'devel', b'all-warnings'): False} |
|
377 | overrides = {(b'devel', b'all-warnings'): False} | |
377 |
|
378 | |||
378 | common = commoninc[0] |
|
379 | common = commoninc[0] | |
379 | with ui.uninterruptible(): |
|
380 | with ui.uninterruptible(): | |
380 | if ellipsesremote: |
|
381 | if ellipsesremote: | |
381 | ds = repo.dirstate |
|
382 | ds = repo.dirstate | |
382 | p1, p2 = ds.p1(), ds.p2() |
|
383 | p1, p2 = ds.p1(), ds.p2() | |
383 | with ds.changing_parents(repo): |
|
384 | with ds.changing_parents(repo): | |
384 | ds.setparents(repo.nullid, repo.nullid) |
|
385 | ds.setparents(repo.nullid, repo.nullid) | |
385 | if isoldellipses: |
|
386 | if isoldellipses: | |
386 | with wrappedextraprepare: |
|
387 | with wrappedextraprepare: | |
387 | exchange.pull(repo, remote, heads=common) |
|
388 | exchange.pull(repo, remote, heads=common) | |
388 | else: |
|
389 | else: | |
389 | known = [] |
|
390 | known = [] | |
390 | if ellipsesremote: |
|
391 | if ellipsesremote: | |
391 | known = [ |
|
392 | known = [ | |
392 | ctx.node() |
|
393 | ctx.node() | |
393 | for ctx in repo.set(b'::%ln', common) |
|
394 | for ctx in repo.set(b'::%ln', common) | |
394 | if ctx.node() != repo.nullid |
|
395 | if ctx.node() != repo.nullid | |
395 | ] |
|
396 | ] | |
396 | with remote.commandexecutor() as e: |
|
397 | with remote.commandexecutor() as e: | |
397 | bundle = e.callcommand( |
|
398 | bundle = e.callcommand( | |
398 | b'narrow_widen', |
|
399 | b'narrow_widen', | |
399 | { |
|
400 | { | |
400 | b'oldincludes': oldincludes, |
|
401 | b'oldincludes': oldincludes, | |
401 | b'oldexcludes': oldexcludes, |
|
402 | b'oldexcludes': oldexcludes, | |
402 | b'newincludes': newincludes, |
|
403 | b'newincludes': newincludes, | |
403 | b'newexcludes': newexcludes, |
|
404 | b'newexcludes': newexcludes, | |
404 | b'cgversion': b'03', |
|
405 | b'cgversion': b'03', | |
405 | b'commonheads': common, |
|
406 | b'commonheads': common, | |
406 | b'known': known, |
|
407 | b'known': known, | |
407 | b'ellipses': ellipsesremote, |
|
408 | b'ellipses': ellipsesremote, | |
408 | }, |
|
409 | }, | |
409 | ).result() |
|
410 | ).result() | |
410 |
|
411 | |||
411 | trmanager = exchange.transactionmanager( |
|
412 | trmanager = exchange.transactionmanager( | |
412 | repo, b'widen', remote.url() |
|
413 | repo, b'widen', remote.url() | |
413 | ) |
|
414 | ) | |
414 | with trmanager, repo.ui.configoverride(overrides, b'widen'): |
|
415 | with trmanager, repo.ui.configoverride(overrides, b'widen'): | |
415 | op = bundle2.bundleoperation( |
|
416 | op = bundle2.bundleoperation( | |
416 | repo, trmanager.transaction, source=b'widen' |
|
417 | repo, trmanager.transaction, source=b'widen' | |
417 | ) |
|
418 | ) | |
418 | # TODO: we should catch error.Abort here |
|
419 | # TODO: we should catch error.Abort here | |
419 | bundle2.processbundle(repo, bundle, op=op, remote=remote) |
|
420 | bundle2.processbundle(repo, bundle, op=op, remote=remote) | |
420 |
|
421 | |||
421 | if ellipsesremote: |
|
422 | if ellipsesremote: | |
422 | with ds.changing_parents(repo): |
|
423 | with ds.changing_parents(repo): | |
423 | ds.setparents(p1, p2) |
|
424 | ds.setparents(p1, p2) | |
424 |
|
425 | |||
425 | with repo.transaction(b'widening'), repo.dirstate.changing_parents( |
|
426 | with repo.transaction(b'widening'), repo.dirstate.changing_parents( | |
426 | repo |
|
427 | repo | |
427 | ): |
|
428 | ): | |
428 | repo.setnewnarrowpats() |
|
429 | repo.setnewnarrowpats() | |
429 | narrowspec.updateworkingcopy(repo) |
|
430 | narrowspec.updateworkingcopy(repo) | |
430 | narrowspec.copytoworkingcopy(repo) |
|
431 | narrowspec.copytoworkingcopy(repo) | |
431 |
|
432 | |||
432 |
|
433 | |||
433 | # TODO(rdamazio): Make new matcher format and update description |
|
434 | # TODO(rdamazio): Make new matcher format and update description | |
434 | @command( |
|
435 | @command( | |
435 | b'tracked', |
|
436 | b'tracked', | |
436 | [ |
|
437 | [ | |
437 | (b'', b'addinclude', [], _(b'new paths to include')), |
|
438 | (b'', b'addinclude', [], _(b'new paths to include')), | |
438 | (b'', b'removeinclude', [], _(b'old paths to no longer include')), |
|
439 | (b'', b'removeinclude', [], _(b'old paths to no longer include')), | |
439 | ( |
|
440 | ( | |
440 | b'', |
|
441 | b'', | |
441 | b'auto-remove-includes', |
|
442 | b'auto-remove-includes', | |
442 | False, |
|
443 | False, | |
443 | _(b'automatically choose unused includes to remove'), |
|
444 | _(b'automatically choose unused includes to remove'), | |
444 | ), |
|
445 | ), | |
445 | (b'', b'addexclude', [], _(b'new paths to exclude')), |
|
446 | (b'', b'addexclude', [], _(b'new paths to exclude')), | |
446 | (b'', b'import-rules', b'', _(b'import narrowspecs from a file')), |
|
447 | (b'', b'import-rules', b'', _(b'import narrowspecs from a file')), | |
447 | (b'', b'removeexclude', [], _(b'old paths to no longer exclude')), |
|
448 | (b'', b'removeexclude', [], _(b'old paths to no longer exclude')), | |
448 | ( |
|
449 | ( | |
449 | b'', |
|
450 | b'', | |
450 | b'clear', |
|
451 | b'clear', | |
451 | False, |
|
452 | False, | |
452 | _(b'whether to replace the existing narrowspec'), |
|
453 | _(b'whether to replace the existing narrowspec'), | |
453 | ), |
|
454 | ), | |
454 | ( |
|
455 | ( | |
455 | b'', |
|
456 | b'', | |
456 | b'force-delete-local-changes', |
|
457 | b'force-delete-local-changes', | |
457 | False, |
|
458 | False, | |
458 | _(b'forces deletion of local changes when narrowing'), |
|
459 | _(b'forces deletion of local changes when narrowing'), | |
459 | ), |
|
460 | ), | |
460 | ( |
|
461 | ( | |
461 | b'', |
|
462 | b'', | |
462 | b'backup', |
|
463 | b'backup', | |
463 | True, |
|
464 | True, | |
464 | _(b'back up local changes when narrowing'), |
|
465 | _(b'back up local changes when narrowing'), | |
465 | ), |
|
466 | ), | |
466 | ( |
|
467 | ( | |
467 | b'', |
|
468 | b'', | |
468 | b'update-working-copy', |
|
469 | b'update-working-copy', | |
469 | False, |
|
470 | False, | |
470 | _(b'update working copy when the store has changed'), |
|
471 | _(b'update working copy when the store has changed'), | |
471 | ), |
|
472 | ), | |
472 | ] |
|
473 | ] | |
473 | + commands.remoteopts, |
|
474 | + commands.remoteopts, | |
474 | _(b'[OPTIONS]... [REMOTE]'), |
|
475 | _(b'[OPTIONS]... [REMOTE]'), | |
475 | inferrepo=True, |
|
476 | inferrepo=True, | |
476 | helpcategory=command.CATEGORY_MAINTENANCE, |
|
477 | helpcategory=command.CATEGORY_MAINTENANCE, | |
477 | ) |
|
478 | ) | |
478 | def trackedcmd(ui, repo, remotepath=None, *pats, **opts): |
|
479 | def trackedcmd(ui, repo, remotepath=None, *pats, **opts): | |
479 | """show or change the current narrowspec |
|
480 | """show or change the current narrowspec | |
480 |
|
481 | |||
481 | With no argument, shows the current narrowspec entries, one per line. Each |
|
482 | With no argument, shows the current narrowspec entries, one per line. Each | |
482 | line will be prefixed with 'I' or 'X' for included or excluded patterns, |
|
483 | line will be prefixed with 'I' or 'X' for included or excluded patterns, | |
483 | respectively. |
|
484 | respectively. | |
484 |
|
485 | |||
485 | The narrowspec is comprised of expressions to match remote files and/or |
|
486 | The narrowspec is comprised of expressions to match remote files and/or | |
486 | directories that should be pulled into your client. |
|
487 | directories that should be pulled into your client. | |
487 | The narrowspec has *include* and *exclude* expressions, with excludes always |
|
488 | The narrowspec has *include* and *exclude* expressions, with excludes always | |
488 | trumping includes: that is, if a file matches an exclude expression, it will |
|
489 | trumping includes: that is, if a file matches an exclude expression, it will | |
489 | be excluded even if it also matches an include expression. |
|
490 | be excluded even if it also matches an include expression. | |
490 | Excluding files that were never included has no effect. |
|
491 | Excluding files that were never included has no effect. | |
491 |
|
492 | |||
492 | Each included or excluded entry is in the format described by |
|
493 | Each included or excluded entry is in the format described by | |
493 | 'hg help patterns'. |
|
494 | 'hg help patterns'. | |
494 |
|
495 | |||
495 | The options allow you to add or remove included and excluded expressions. |
|
496 | The options allow you to add or remove included and excluded expressions. | |
496 |
|
497 | |||
497 | If --clear is specified, then all previous includes and excludes are DROPPED |
|
498 | If --clear is specified, then all previous includes and excludes are DROPPED | |
498 | and replaced by the new ones specified to --addinclude and --addexclude. |
|
499 | and replaced by the new ones specified to --addinclude and --addexclude. | |
499 | If --clear is specified without any further options, the narrowspec will be |
|
500 | If --clear is specified without any further options, the narrowspec will be | |
500 | empty and will not match any files. |
|
501 | empty and will not match any files. | |
501 |
|
502 | |||
502 | If --auto-remove-includes is specified, then those includes that don't match |
|
503 | If --auto-remove-includes is specified, then those includes that don't match | |
503 | any files modified by currently visible local commits (those not shared by |
|
504 | any files modified by currently visible local commits (those not shared by | |
504 | the remote) will be added to the set of explicitly specified includes to |
|
505 | the remote) will be added to the set of explicitly specified includes to | |
505 | remove. |
|
506 | remove. | |
506 |
|
507 | |||
507 | --import-rules accepts a path to a file containing rules, allowing you to |
|
508 | --import-rules accepts a path to a file containing rules, allowing you to | |
508 | add --addinclude, --addexclude rules in bulk. Like the other include and |
|
509 | add --addinclude, --addexclude rules in bulk. Like the other include and | |
509 | exclude switches, the changes are applied immediately. |
|
510 | exclude switches, the changes are applied immediately. | |
510 | """ |
|
511 | """ | |
511 | opts = pycompat.byteskwargs(opts) |
|
512 | opts = pycompat.byteskwargs(opts) | |
512 | if requirements.NARROW_REQUIREMENT not in repo.requirements: |
|
513 | if requirements.NARROW_REQUIREMENT not in repo.requirements: | |
513 | raise error.InputError( |
|
514 | raise error.InputError( | |
514 | _( |
|
515 | _( | |
515 | b'the tracked command is only supported on ' |
|
516 | b'the tracked command is only supported on ' | |
516 | b'repositories cloned with --narrow' |
|
517 | b'repositories cloned with --narrow' | |
517 | ) |
|
518 | ) | |
518 | ) |
|
519 | ) | |
519 |
|
520 | |||
520 | # Before supporting, decide whether it "hg tracked --clear" should mean |
|
521 | # Before supporting, decide whether it "hg tracked --clear" should mean | |
521 | # tracking no paths or all paths. |
|
522 | # tracking no paths or all paths. | |
522 | if opts[b'clear']: |
|
523 | if opts[b'clear']: | |
523 | raise error.InputError(_(b'the --clear option is not yet supported')) |
|
524 | raise error.InputError(_(b'the --clear option is not yet supported')) | |
524 |
|
525 | |||
525 | # import rules from a file |
|
526 | # import rules from a file | |
526 | newrules = opts.get(b'import_rules') |
|
527 | newrules = opts.get(b'import_rules') | |
527 | if newrules: |
|
528 | if newrules: | |
528 | try: |
|
529 | try: | |
529 | filepath = os.path.join(encoding.getcwd(), newrules) |
|
530 | filepath = os.path.join(encoding.getcwd(), newrules) | |
530 | fdata = util.readfile(filepath) |
|
531 | fdata = util.readfile(filepath) | |
531 | except IOError as inst: |
|
532 | except IOError as inst: | |
532 | raise error.StorageError( |
|
533 | raise error.StorageError( | |
533 | _(b"cannot read narrowspecs from '%s': %s") |
|
534 | _(b"cannot read narrowspecs from '%s': %s") | |
534 | % (filepath, encoding.strtolocal(inst.strerror)) |
|
535 | % (filepath, encoding.strtolocal(inst.strerror)) | |
535 | ) |
|
536 | ) | |
536 | includepats, excludepats, profiles = sparse.parseconfig( |
|
537 | includepats, excludepats, profiles = sparse.parseconfig( | |
537 | ui, fdata, b'narrow' |
|
538 | ui, fdata, b'narrow' | |
538 | ) |
|
539 | ) | |
539 | if profiles: |
|
540 | if profiles: | |
540 | raise error.InputError( |
|
541 | raise error.InputError( | |
541 | _( |
|
542 | _( | |
542 | b"including other spec files using '%include' " |
|
543 | b"including other spec files using '%include' " | |
543 | b"is not supported in narrowspec" |
|
544 | b"is not supported in narrowspec" | |
544 | ) |
|
545 | ) | |
545 | ) |
|
546 | ) | |
546 | opts[b'addinclude'].extend(includepats) |
|
547 | opts[b'addinclude'].extend(includepats) | |
547 | opts[b'addexclude'].extend(excludepats) |
|
548 | opts[b'addexclude'].extend(excludepats) | |
548 |
|
549 | |||
549 | addedincludes = narrowspec.parsepatterns(opts[b'addinclude']) |
|
550 | addedincludes = narrowspec.parsepatterns(opts[b'addinclude']) | |
550 | removedincludes = narrowspec.parsepatterns(opts[b'removeinclude']) |
|
551 | removedincludes = narrowspec.parsepatterns(opts[b'removeinclude']) | |
551 | addedexcludes = narrowspec.parsepatterns(opts[b'addexclude']) |
|
552 | addedexcludes = narrowspec.parsepatterns(opts[b'addexclude']) | |
552 | removedexcludes = narrowspec.parsepatterns(opts[b'removeexclude']) |
|
553 | removedexcludes = narrowspec.parsepatterns(opts[b'removeexclude']) | |
553 | autoremoveincludes = opts[b'auto_remove_includes'] |
|
554 | autoremoveincludes = opts[b'auto_remove_includes'] | |
554 |
|
555 | |||
555 | update_working_copy = opts[b'update_working_copy'] |
|
556 | update_working_copy = opts[b'update_working_copy'] | |
556 | only_show = not ( |
|
557 | only_show = not ( | |
557 | addedincludes |
|
558 | addedincludes | |
558 | or removedincludes |
|
559 | or removedincludes | |
559 | or addedexcludes |
|
560 | or addedexcludes | |
560 | or removedexcludes |
|
561 | or removedexcludes | |
561 | or newrules |
|
562 | or newrules | |
562 | or autoremoveincludes |
|
563 | or autoremoveincludes | |
563 | or update_working_copy |
|
564 | or update_working_copy | |
564 | ) |
|
565 | ) | |
565 |
|
566 | |||
566 | # Only print the current narrowspec. |
|
567 | # Only print the current narrowspec. | |
567 | if only_show: |
|
568 | if only_show: | |
568 | oldincludes, oldexcludes = repo.narrowpats |
|
569 | oldincludes, oldexcludes = repo.narrowpats | |
569 | ui.pager(b'tracked') |
|
570 | ui.pager(b'tracked') | |
570 | fm = ui.formatter(b'narrow', opts) |
|
571 | fm = ui.formatter(b'narrow', opts) | |
571 | for i in sorted(oldincludes): |
|
572 | for i in sorted(oldincludes): | |
572 | fm.startitem() |
|
573 | fm.startitem() | |
573 | fm.write(b'status', b'%s ', b'I', label=b'narrow.included') |
|
574 | fm.write(b'status', b'%s ', b'I', label=b'narrow.included') | |
574 | fm.write(b'pat', b'%s\n', i, label=b'narrow.included') |
|
575 | fm.write(b'pat', b'%s\n', i, label=b'narrow.included') | |
575 | for i in sorted(oldexcludes): |
|
576 | for i in sorted(oldexcludes): | |
576 | fm.startitem() |
|
577 | fm.startitem() | |
577 | fm.write(b'status', b'%s ', b'X', label=b'narrow.excluded') |
|
578 | fm.write(b'status', b'%s ', b'X', label=b'narrow.excluded') | |
578 | fm.write(b'pat', b'%s\n', i, label=b'narrow.excluded') |
|
579 | fm.write(b'pat', b'%s\n', i, label=b'narrow.excluded') | |
579 | fm.end() |
|
580 | fm.end() | |
580 | return 0 |
|
581 | return 0 | |
581 |
|
582 | |||
582 | with repo.wlock(), repo.lock(): |
|
583 | with repo.wlock(), repo.lock(): | |
583 | oldincludes, oldexcludes = repo.narrowpats |
|
584 | oldincludes, oldexcludes = repo.narrowpats | |
584 |
|
585 | |||
585 | # filter the user passed additions and deletions into actual additions and |
|
586 | # filter the user passed additions and deletions into actual additions and | |
586 | # deletions of excludes and includes |
|
587 | # deletions of excludes and includes | |
587 | addedincludes -= oldincludes |
|
588 | addedincludes -= oldincludes | |
588 | removedincludes &= oldincludes |
|
589 | removedincludes &= oldincludes | |
589 | addedexcludes -= oldexcludes |
|
590 | addedexcludes -= oldexcludes | |
590 | removedexcludes &= oldexcludes |
|
591 | removedexcludes &= oldexcludes | |
591 |
|
592 | |||
592 | widening = addedincludes or removedexcludes |
|
593 | widening = addedincludes or removedexcludes | |
593 | narrowing = removedincludes or addedexcludes |
|
594 | narrowing = removedincludes or addedexcludes | |
594 |
|
595 | |||
595 | if update_working_copy: |
|
596 | if update_working_copy: | |
596 | with repo.transaction(b'narrow-wc'), repo.dirstate.changing_parents( |
|
597 | with repo.transaction(b'narrow-wc'), repo.dirstate.changing_parents( | |
597 | repo |
|
598 | repo | |
598 | ): |
|
599 | ): | |
599 | narrowspec.updateworkingcopy(repo) |
|
600 | narrowspec.updateworkingcopy(repo) | |
600 | narrowspec.copytoworkingcopy(repo) |
|
601 | narrowspec.copytoworkingcopy(repo) | |
601 | return 0 |
|
602 | return 0 | |
602 |
|
603 | |||
603 | if not (widening or narrowing or autoremoveincludes): |
|
604 | if not (widening or narrowing or autoremoveincludes): | |
604 | ui.status(_(b"nothing to widen or narrow\n")) |
|
605 | ui.status(_(b"nothing to widen or narrow\n")) | |
605 | return 0 |
|
606 | return 0 | |
606 |
|
607 | |||
607 | cmdutil.bailifchanged(repo) |
|
608 | cmdutil.bailifchanged(repo) | |
608 |
|
609 | |||
609 | # Find the revisions we have in common with the remote. These will |
|
610 | # Find the revisions we have in common with the remote. These will | |
610 | # be used for finding local-only changes for narrowing. They will |
|
611 | # be used for finding local-only changes for narrowing. They will | |
611 | # also define the set of revisions to update for widening. |
|
612 | # also define the set of revisions to update for widening. | |
612 | path = urlutil.get_unique_pull_path_obj(b'tracked', ui, remotepath) |
|
613 | path = urlutil.get_unique_pull_path_obj(b'tracked', ui, remotepath) | |
613 | ui.status(_(b'comparing with %s\n') % urlutil.hidepassword(path.loc)) |
|
614 | ui.status(_(b'comparing with %s\n') % urlutil.hidepassword(path.loc)) | |
614 | remote = hg.peer(repo, opts, path) |
|
615 | remote = hg.peer(repo, opts, path) | |
615 |
|
616 | |||
616 | try: |
|
617 | try: | |
617 | # check narrow support before doing anything if widening needs to be |
|
618 | # check narrow support before doing anything if widening needs to be | |
618 | # performed. In future we should also abort if client is ellipses and |
|
619 | # performed. In future we should also abort if client is ellipses and | |
619 | # server does not support ellipses |
|
620 | # server does not support ellipses | |
620 | if ( |
|
621 | if ( | |
621 | widening |
|
622 | widening | |
622 | and wireprototypes.NARROWCAP not in remote.capabilities() |
|
623 | and wireprototypes.NARROWCAP not in remote.capabilities() | |
623 | ): |
|
624 | ): | |
624 | raise error.Abort(_(b"server does not support narrow clones")) |
|
625 | raise error.Abort(_(b"server does not support narrow clones")) | |
625 |
|
626 | |||
626 | commoninc = discovery.findcommonincoming(repo, remote) |
|
627 | commoninc = discovery.findcommonincoming(repo, remote) | |
627 |
|
628 | |||
628 | if autoremoveincludes: |
|
629 | if autoremoveincludes: | |
629 | outgoing = discovery.findcommonoutgoing( |
|
630 | outgoing = discovery.findcommonoutgoing( | |
630 | repo, remote, commoninc=commoninc |
|
631 | repo, remote, commoninc=commoninc | |
631 | ) |
|
632 | ) | |
632 | ui.status(_(b'looking for unused includes to remove\n')) |
|
633 | ui.status(_(b'looking for unused includes to remove\n')) | |
633 | localfiles = set() |
|
634 | localfiles = set() | |
634 | for n in itertools.chain(outgoing.missing, outgoing.excluded): |
|
635 | for n in itertools.chain(outgoing.missing, outgoing.excluded): | |
635 | localfiles.update(repo[n].files()) |
|
636 | localfiles.update(repo[n].files()) | |
636 | suggestedremovals = [] |
|
637 | suggestedremovals = [] | |
637 | for include in sorted(oldincludes): |
|
638 | for include in sorted(oldincludes): | |
638 | match = narrowspec.match(repo.root, [include], oldexcludes) |
|
639 | match = narrowspec.match(repo.root, [include], oldexcludes) | |
639 | if not any(match(f) for f in localfiles): |
|
640 | if not any(match(f) for f in localfiles): | |
640 | suggestedremovals.append(include) |
|
641 | suggestedremovals.append(include) | |
641 | if suggestedremovals: |
|
642 | if suggestedremovals: | |
642 | for s in suggestedremovals: |
|
643 | for s in suggestedremovals: | |
643 | ui.status(b'%s\n' % s) |
|
644 | ui.status(b'%s\n' % s) | |
644 | if ( |
|
645 | if ( | |
645 | ui.promptchoice( |
|
646 | ui.promptchoice( | |
646 | _( |
|
647 | _( | |
647 | b'remove these unused includes (yn)?' |
|
648 | b'remove these unused includes (yn)?' | |
648 | b'$$ &Yes $$ &No' |
|
649 | b'$$ &Yes $$ &No' | |
649 | ) |
|
650 | ) | |
650 | ) |
|
651 | ) | |
651 | == 0 |
|
652 | == 0 | |
652 | ): |
|
653 | ): | |
653 | removedincludes.update(suggestedremovals) |
|
654 | removedincludes.update(suggestedremovals) | |
654 | narrowing = True |
|
655 | narrowing = True | |
655 | else: |
|
656 | else: | |
656 | ui.status(_(b'found no unused includes\n')) |
|
657 | ui.status(_(b'found no unused includes\n')) | |
657 |
|
658 | |||
658 | if narrowing: |
|
659 | if narrowing: | |
659 | newincludes = oldincludes - removedincludes |
|
660 | newincludes = oldincludes - removedincludes | |
660 | newexcludes = oldexcludes | addedexcludes |
|
661 | newexcludes = oldexcludes | addedexcludes | |
661 | _narrow( |
|
662 | _narrow( | |
662 | ui, |
|
663 | ui, | |
663 | repo, |
|
664 | repo, | |
664 | remote, |
|
665 | remote, | |
665 | commoninc, |
|
666 | commoninc, | |
666 | oldincludes, |
|
667 | oldincludes, | |
667 | oldexcludes, |
|
668 | oldexcludes, | |
668 | newincludes, |
|
669 | newincludes, | |
669 | newexcludes, |
|
670 | newexcludes, | |
670 | opts[b'force_delete_local_changes'], |
|
671 | opts[b'force_delete_local_changes'], | |
671 | opts[b'backup'], |
|
672 | opts[b'backup'], | |
672 | ) |
|
673 | ) | |
673 | # _narrow() updated the narrowspec and _widen() below needs to |
|
674 | # _narrow() updated the narrowspec and _widen() below needs to | |
674 | # use the updated values as its base (otherwise removed includes |
|
675 | # use the updated values as its base (otherwise removed includes | |
675 | # and addedexcludes will be lost in the resulting narrowspec) |
|
676 | # and addedexcludes will be lost in the resulting narrowspec) | |
676 | oldincludes = newincludes |
|
677 | oldincludes = newincludes | |
677 | oldexcludes = newexcludes |
|
678 | oldexcludes = newexcludes | |
678 |
|
679 | |||
679 | if widening: |
|
680 | if widening: | |
680 | newincludes = oldincludes | addedincludes |
|
681 | newincludes = oldincludes | addedincludes | |
681 | newexcludes = oldexcludes - removedexcludes |
|
682 | newexcludes = oldexcludes - removedexcludes | |
682 | _widen( |
|
683 | _widen( | |
683 | ui, |
|
684 | ui, | |
684 | repo, |
|
685 | repo, | |
685 | remote, |
|
686 | remote, | |
686 | commoninc, |
|
687 | commoninc, | |
687 | oldincludes, |
|
688 | oldincludes, | |
688 | oldexcludes, |
|
689 | oldexcludes, | |
689 | newincludes, |
|
690 | newincludes, | |
690 | newexcludes, |
|
691 | newexcludes, | |
691 | ) |
|
692 | ) | |
692 | finally: |
|
693 | finally: | |
693 | remote.close() |
|
694 | remote.close() | |
694 |
|
695 | |||
695 | return 0 |
|
696 | return 0 |
@@ -1,437 +1,446 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 |
|
7 | |||
8 | import os |
|
8 | import os | |
9 | import stat |
|
9 | import stat | |
10 | import time |
|
10 | import time | |
11 | import zlib |
|
11 | import zlib | |
12 |
|
12 | |||
13 | from mercurial.i18n import _ |
|
13 | from mercurial.i18n import _ | |
14 | from mercurial.node import bin, hex |
|
14 | from mercurial.node import bin, hex | |
15 | from mercurial.pycompat import open |
|
15 | from mercurial.pycompat import open | |
16 | from mercurial import ( |
|
16 | from mercurial import ( | |
17 | changegroup, |
|
17 | changegroup, | |
18 | changelog, |
|
18 | changelog, | |
19 | context, |
|
19 | context, | |
20 | error, |
|
20 | error, | |
21 | extensions, |
|
21 | extensions, | |
22 | match, |
|
22 | match, | |
23 | scmutil, |
|
23 | scmutil, | |
24 | store, |
|
24 | store, | |
25 | streamclone, |
|
25 | streamclone, | |
26 | util, |
|
26 | util, | |
27 | wireprotoserver, |
|
27 | wireprotoserver, | |
28 | wireprototypes, |
|
28 | wireprototypes, | |
29 | wireprotov1server, |
|
29 | wireprotov1server, | |
30 | ) |
|
30 | ) | |
31 | from . import ( |
|
31 | from . import ( | |
32 | constants, |
|
32 | constants, | |
33 | shallowutil, |
|
33 | shallowutil, | |
34 | ) |
|
34 | ) | |
35 |
|
35 | |||
36 | _sshv1server = wireprotoserver.sshv1protocolhandler |
|
36 | _sshv1server = wireprotoserver.sshv1protocolhandler | |
37 |
|
37 | |||
38 |
|
38 | |||
39 | def setupserver(ui, repo): |
|
39 | def setupserver(ui, repo): | |
40 | """Sets up a normal Mercurial repo so it can serve files to shallow repos.""" |
|
40 | """Sets up a normal Mercurial repo so it can serve files to shallow repos.""" | |
41 | onetimesetup(ui) |
|
41 | onetimesetup(ui) | |
42 |
|
42 | |||
43 | # don't send files to shallow clients during pulls |
|
43 | # don't send files to shallow clients during pulls | |
44 | def generatefiles( |
|
44 | def generatefiles( | |
45 | orig, self, changedfiles, linknodes, commonrevs, source, *args, **kwargs |
|
45 | orig, self, changedfiles, linknodes, commonrevs, source, *args, **kwargs | |
46 | ): |
|
46 | ): | |
47 | caps = self._bundlecaps or [] |
|
47 | caps = self._bundlecaps or [] | |
48 | if constants.BUNDLE2_CAPABLITY in caps: |
|
48 | if constants.BUNDLE2_CAPABLITY in caps: | |
49 | # only send files that don't match the specified patterns |
|
49 | # only send files that don't match the specified patterns | |
50 | includepattern = None |
|
50 | includepattern = None | |
51 | excludepattern = None |
|
51 | excludepattern = None | |
52 | for cap in self._bundlecaps or []: |
|
52 | for cap in self._bundlecaps or []: | |
53 | if cap.startswith(b"includepattern="): |
|
53 | if cap.startswith(b"includepattern="): | |
54 | includepattern = cap[len(b"includepattern=") :].split(b'\0') |
|
54 | includepattern = cap[len(b"includepattern=") :].split(b'\0') | |
55 | elif cap.startswith(b"excludepattern="): |
|
55 | elif cap.startswith(b"excludepattern="): | |
56 | excludepattern = cap[len(b"excludepattern=") :].split(b'\0') |
|
56 | excludepattern = cap[len(b"excludepattern=") :].split(b'\0') | |
57 |
|
57 | |||
58 | m = match.always() |
|
58 | m = match.always() | |
59 | if includepattern or excludepattern: |
|
59 | if includepattern or excludepattern: | |
60 | m = match.match( |
|
60 | m = match.match( | |
61 | repo.root, b'', None, includepattern, excludepattern |
|
61 | repo.root, b'', None, includepattern, excludepattern | |
62 | ) |
|
62 | ) | |
63 |
|
63 | |||
64 | changedfiles = list([f for f in changedfiles if not m(f)]) |
|
64 | changedfiles = list([f for f in changedfiles if not m(f)]) | |
65 | return orig( |
|
65 | return orig( | |
66 | self, changedfiles, linknodes, commonrevs, source, *args, **kwargs |
|
66 | self, changedfiles, linknodes, commonrevs, source, *args, **kwargs | |
67 | ) |
|
67 | ) | |
68 |
|
68 | |||
69 | extensions.wrapfunction( |
|
69 | extensions.wrapfunction( | |
70 | changegroup.cgpacker, b'generatefiles', generatefiles |
|
70 | changegroup.cgpacker, b'generatefiles', generatefiles | |
71 | ) |
|
71 | ) | |
72 |
|
72 | |||
73 |
|
73 | |||
74 | onetime = False |
|
74 | onetime = False | |
75 |
|
75 | |||
76 |
|
76 | |||
77 | def onetimesetup(ui): |
|
77 | def onetimesetup(ui): | |
78 | """Configures the wireprotocol for both clients and servers.""" |
|
78 | """Configures the wireprotocol for both clients and servers.""" | |
79 | global onetime |
|
79 | global onetime | |
80 | if onetime: |
|
80 | if onetime: | |
81 | return |
|
81 | return | |
82 | onetime = True |
|
82 | onetime = True | |
83 |
|
83 | |||
84 | # support file content requests |
|
84 | # support file content requests | |
85 | wireprotov1server.wireprotocommand( |
|
85 | wireprotov1server.wireprotocommand( | |
86 | b'x_rfl_getflogheads', b'path', permission=b'pull' |
|
86 | b'x_rfl_getflogheads', b'path', permission=b'pull' | |
87 | )(getflogheads) |
|
87 | )(getflogheads) | |
88 | wireprotov1server.wireprotocommand( |
|
88 | wireprotov1server.wireprotocommand( | |
89 | b'x_rfl_getfiles', b'', permission=b'pull' |
|
89 | b'x_rfl_getfiles', b'', permission=b'pull' | |
90 | )(getfiles) |
|
90 | )(getfiles) | |
91 | wireprotov1server.wireprotocommand( |
|
91 | wireprotov1server.wireprotocommand( | |
92 | b'x_rfl_getfile', b'file node', permission=b'pull' |
|
92 | b'x_rfl_getfile', b'file node', permission=b'pull' | |
93 | )(getfile) |
|
93 | )(getfile) | |
94 |
|
94 | |||
95 | class streamstate: |
|
95 | class streamstate: | |
96 | match = None |
|
96 | match = None | |
97 | shallowremote = False |
|
97 | shallowremote = False | |
98 | noflatmf = False |
|
98 | noflatmf = False | |
99 |
|
99 | |||
100 | state = streamstate() |
|
100 | state = streamstate() | |
101 |
|
101 | |||
102 | def stream_out_shallow(repo, proto, other): |
|
102 | def stream_out_shallow(repo, proto, other): | |
103 | includepattern = None |
|
103 | includepattern = None | |
104 | excludepattern = None |
|
104 | excludepattern = None | |
105 | raw = other.get(b'includepattern') |
|
105 | raw = other.get(b'includepattern') | |
106 | if raw: |
|
106 | if raw: | |
107 | includepattern = raw.split(b'\0') |
|
107 | includepattern = raw.split(b'\0') | |
108 | raw = other.get(b'excludepattern') |
|
108 | raw = other.get(b'excludepattern') | |
109 | if raw: |
|
109 | if raw: | |
110 | excludepattern = raw.split(b'\0') |
|
110 | excludepattern = raw.split(b'\0') | |
111 |
|
111 | |||
112 | oldshallow = state.shallowremote |
|
112 | oldshallow = state.shallowremote | |
113 | oldmatch = state.match |
|
113 | oldmatch = state.match | |
114 | oldnoflatmf = state.noflatmf |
|
114 | oldnoflatmf = state.noflatmf | |
115 | try: |
|
115 | try: | |
116 | state.shallowremote = True |
|
116 | state.shallowremote = True | |
117 | state.match = match.always() |
|
117 | state.match = match.always() | |
118 | state.noflatmf = other.get(b'noflatmanifest') == b'True' |
|
118 | state.noflatmf = other.get(b'noflatmanifest') == b'True' | |
119 | if includepattern or excludepattern: |
|
119 | if includepattern or excludepattern: | |
120 | state.match = match.match( |
|
120 | state.match = match.match( | |
121 | repo.root, b'', None, includepattern, excludepattern |
|
121 | repo.root, b'', None, includepattern, excludepattern | |
122 | ) |
|
122 | ) | |
123 | streamres = wireprotov1server.stream(repo, proto) |
|
123 | streamres = wireprotov1server.stream(repo, proto) | |
124 |
|
124 | |||
125 | # Force the first value to execute, so the file list is computed |
|
125 | # Force the first value to execute, so the file list is computed | |
126 | # within the try/finally scope |
|
126 | # within the try/finally scope | |
127 | first = next(streamres.gen) |
|
127 | first = next(streamres.gen) | |
128 | second = next(streamres.gen) |
|
128 | second = next(streamres.gen) | |
129 |
|
129 | |||
130 | def gen(): |
|
130 | def gen(): | |
131 | yield first |
|
131 | yield first | |
132 | yield second |
|
132 | yield second | |
133 | for value in streamres.gen: |
|
133 | for value in streamres.gen: | |
134 | yield value |
|
134 | yield value | |
135 |
|
135 | |||
136 | return wireprototypes.streamres(gen()) |
|
136 | return wireprototypes.streamres(gen()) | |
137 | finally: |
|
137 | finally: | |
138 | state.shallowremote = oldshallow |
|
138 | state.shallowremote = oldshallow | |
139 | state.match = oldmatch |
|
139 | state.match = oldmatch | |
140 | state.noflatmf = oldnoflatmf |
|
140 | state.noflatmf = oldnoflatmf | |
141 |
|
141 | |||
142 | wireprotov1server.commands[b'stream_out_shallow'] = ( |
|
142 | wireprotov1server.commands[b'stream_out_shallow'] = ( | |
143 | stream_out_shallow, |
|
143 | stream_out_shallow, | |
144 | b'*', |
|
144 | b'*', | |
145 | ) |
|
145 | ) | |
146 |
|
146 | |||
147 | # don't clone filelogs to shallow clients |
|
147 | # don't clone filelogs to shallow clients | |
148 | def _walkstreamfiles(orig, repo, matcher=None): |
|
148 | def _walkstreamfiles(orig, repo, matcher=None): | |
149 | if state.shallowremote: |
|
149 | if state.shallowremote: | |
150 | # if we are shallow ourselves, stream our local commits |
|
150 | # if we are shallow ourselves, stream our local commits | |
151 | if shallowutil.isenabled(repo): |
|
151 | if shallowutil.isenabled(repo): | |
152 | striplen = len(repo.store.path) + 1 |
|
152 | striplen = len(repo.store.path) + 1 | |
153 | readdir = repo.store.rawvfs.readdir |
|
153 | readdir = repo.store.rawvfs.readdir | |
154 | visit = [os.path.join(repo.store.path, b'data')] |
|
154 | visit = [os.path.join(repo.store.path, b'data')] | |
155 | while visit: |
|
155 | while visit: | |
156 | p = visit.pop() |
|
156 | p = visit.pop() | |
157 | for f, kind, st in readdir(p, stat=True): |
|
157 | for f, kind, st in readdir(p, stat=True): | |
158 | fp = p + b'/' + f |
|
158 | fp = p + b'/' + f | |
159 | if kind == stat.S_IFREG: |
|
159 | if kind == stat.S_IFREG: | |
160 | if not fp.endswith(b'.i') and not fp.endswith( |
|
160 | if not fp.endswith(b'.i') and not fp.endswith( | |
161 | b'.d' |
|
161 | b'.d' | |
162 | ): |
|
162 | ): | |
163 | n = util.pconvert(fp[striplen:]) |
|
163 | n = util.pconvert(fp[striplen:]) | |
164 | d = store.decodedir(n) |
|
164 | d = store.decodedir(n) | |
165 |
|
|
165 | yield store.StoreEntry( | |
166 |
|
|
166 | unencoded_path=d, | |
|
167 | is_revlog=True, | |||
|
168 | revlog_type=None, | |||
|
169 | is_revlog_main=False, | |||
|
170 | is_volatile=False, | |||
|
171 | file_size=st.st_size, | |||
|
172 | ) | |||
|
173 | ||||
167 | if kind == stat.S_IFDIR: |
|
174 | if kind == stat.S_IFDIR: | |
168 | visit.append(fp) |
|
175 | visit.append(fp) | |
169 |
|
176 | |||
170 | if scmutil.istreemanifest(repo): |
|
177 | if scmutil.istreemanifest(repo): | |
171 |
for |
|
178 | for entry in repo.store.datafiles(): | |
|
179 | u = entry.unencoded_path | |||
172 | if u.startswith(b'meta/') and ( |
|
180 | if u.startswith(b'meta/') and ( | |
173 | u.endswith(b'.i') or u.endswith(b'.d') |
|
181 | u.endswith(b'.i') or u.endswith(b'.d') | |
174 | ): |
|
182 | ): | |
175 |
yield |
|
183 | yield entry | |
176 |
|
184 | |||
177 | # Return .d and .i files that do not match the shallow pattern |
|
185 | # Return .d and .i files that do not match the shallow pattern | |
178 | match = state.match |
|
186 | match = state.match | |
179 | if match and not match.always(): |
|
187 | if match and not match.always(): | |
180 |
for |
|
188 | for entry in repo.store.datafiles(): | |
|
189 | u = entry.unencoded_path | |||
181 | f = u[5:-2] # trim data/... and .i/.d |
|
190 | f = u[5:-2] # trim data/... and .i/.d | |
182 | if not state.match(f): |
|
191 | if not state.match(f): | |
183 |
yield |
|
192 | yield entry | |
184 |
|
193 | |||
185 | for x in repo.store.topfiles(): |
|
194 | for x in repo.store.topfiles(): | |
186 | if state.noflatmf and x[1][:11] == b'00manifest.': |
|
195 | if state.noflatmf and x[1][:11] == b'00manifest.': | |
187 | continue |
|
196 | continue | |
188 | yield x |
|
197 | yield x | |
189 |
|
198 | |||
190 | elif shallowutil.isenabled(repo): |
|
199 | elif shallowutil.isenabled(repo): | |
191 | # don't allow cloning from a shallow repo to a full repo |
|
200 | # don't allow cloning from a shallow repo to a full repo | |
192 | # since it would require fetching every version of every |
|
201 | # since it would require fetching every version of every | |
193 | # file in order to create the revlogs. |
|
202 | # file in order to create the revlogs. | |
194 | raise error.Abort( |
|
203 | raise error.Abort( | |
195 | _(b"Cannot clone from a shallow repo to a full repo.") |
|
204 | _(b"Cannot clone from a shallow repo to a full repo.") | |
196 | ) |
|
205 | ) | |
197 | else: |
|
206 | else: | |
198 | for x in orig(repo, matcher): |
|
207 | for x in orig(repo, matcher): | |
199 | yield x |
|
208 | yield x | |
200 |
|
209 | |||
201 | extensions.wrapfunction(streamclone, b'_walkstreamfiles', _walkstreamfiles) |
|
210 | extensions.wrapfunction(streamclone, b'_walkstreamfiles', _walkstreamfiles) | |
202 |
|
211 | |||
203 | # expose remotefilelog capabilities |
|
212 | # expose remotefilelog capabilities | |
204 | def _capabilities(orig, repo, proto): |
|
213 | def _capabilities(orig, repo, proto): | |
205 | caps = orig(repo, proto) |
|
214 | caps = orig(repo, proto) | |
206 | if shallowutil.isenabled(repo) or ui.configbool( |
|
215 | if shallowutil.isenabled(repo) or ui.configbool( | |
207 | b'remotefilelog', b'server' |
|
216 | b'remotefilelog', b'server' | |
208 | ): |
|
217 | ): | |
209 | if isinstance(proto, _sshv1server): |
|
218 | if isinstance(proto, _sshv1server): | |
210 | # legacy getfiles method which only works over ssh |
|
219 | # legacy getfiles method which only works over ssh | |
211 | caps.append(constants.NETWORK_CAP_LEGACY_SSH_GETFILES) |
|
220 | caps.append(constants.NETWORK_CAP_LEGACY_SSH_GETFILES) | |
212 | caps.append(b'x_rfl_getflogheads') |
|
221 | caps.append(b'x_rfl_getflogheads') | |
213 | caps.append(b'x_rfl_getfile') |
|
222 | caps.append(b'x_rfl_getfile') | |
214 | return caps |
|
223 | return caps | |
215 |
|
224 | |||
216 | extensions.wrapfunction(wireprotov1server, b'_capabilities', _capabilities) |
|
225 | extensions.wrapfunction(wireprotov1server, b'_capabilities', _capabilities) | |
217 |
|
226 | |||
218 | def _adjustlinkrev(orig, self, *args, **kwargs): |
|
227 | def _adjustlinkrev(orig, self, *args, **kwargs): | |
219 | # When generating file blobs, taking the real path is too slow on large |
|
228 | # When generating file blobs, taking the real path is too slow on large | |
220 | # repos, so force it to just return the linkrev directly. |
|
229 | # repos, so force it to just return the linkrev directly. | |
221 | repo = self._repo |
|
230 | repo = self._repo | |
222 | if util.safehasattr(repo, b'forcelinkrev') and repo.forcelinkrev: |
|
231 | if util.safehasattr(repo, b'forcelinkrev') and repo.forcelinkrev: | |
223 | return self._filelog.linkrev(self._filelog.rev(self._filenode)) |
|
232 | return self._filelog.linkrev(self._filelog.rev(self._filenode)) | |
224 | return orig(self, *args, **kwargs) |
|
233 | return orig(self, *args, **kwargs) | |
225 |
|
234 | |||
226 | extensions.wrapfunction( |
|
235 | extensions.wrapfunction( | |
227 | context.basefilectx, b'_adjustlinkrev', _adjustlinkrev |
|
236 | context.basefilectx, b'_adjustlinkrev', _adjustlinkrev | |
228 | ) |
|
237 | ) | |
229 |
|
238 | |||
230 | def _iscmd(orig, cmd): |
|
239 | def _iscmd(orig, cmd): | |
231 | if cmd == b'x_rfl_getfiles': |
|
240 | if cmd == b'x_rfl_getfiles': | |
232 | return False |
|
241 | return False | |
233 | return orig(cmd) |
|
242 | return orig(cmd) | |
234 |
|
243 | |||
235 | extensions.wrapfunction(wireprotoserver, b'iscmd', _iscmd) |
|
244 | extensions.wrapfunction(wireprotoserver, b'iscmd', _iscmd) | |
236 |
|
245 | |||
237 |
|
246 | |||
238 | def _loadfileblob(repo, cachepath, path, node): |
|
247 | def _loadfileblob(repo, cachepath, path, node): | |
239 | filecachepath = os.path.join(cachepath, path, hex(node)) |
|
248 | filecachepath = os.path.join(cachepath, path, hex(node)) | |
240 | if not os.path.exists(filecachepath) or os.path.getsize(filecachepath) == 0: |
|
249 | if not os.path.exists(filecachepath) or os.path.getsize(filecachepath) == 0: | |
241 | filectx = repo.filectx(path, fileid=node) |
|
250 | filectx = repo.filectx(path, fileid=node) | |
242 | if filectx.node() == repo.nullid: |
|
251 | if filectx.node() == repo.nullid: | |
243 | repo.changelog = changelog.changelog(repo.svfs) |
|
252 | repo.changelog = changelog.changelog(repo.svfs) | |
244 | filectx = repo.filectx(path, fileid=node) |
|
253 | filectx = repo.filectx(path, fileid=node) | |
245 |
|
254 | |||
246 | text = createfileblob(filectx) |
|
255 | text = createfileblob(filectx) | |
247 | # TODO configurable compression engines |
|
256 | # TODO configurable compression engines | |
248 | text = zlib.compress(text) |
|
257 | text = zlib.compress(text) | |
249 |
|
258 | |||
250 | # everything should be user & group read/writable |
|
259 | # everything should be user & group read/writable | |
251 | oldumask = os.umask(0o002) |
|
260 | oldumask = os.umask(0o002) | |
252 | try: |
|
261 | try: | |
253 | dirname = os.path.dirname(filecachepath) |
|
262 | dirname = os.path.dirname(filecachepath) | |
254 | if not os.path.exists(dirname): |
|
263 | if not os.path.exists(dirname): | |
255 | try: |
|
264 | try: | |
256 | os.makedirs(dirname) |
|
265 | os.makedirs(dirname) | |
257 | except FileExistsError: |
|
266 | except FileExistsError: | |
258 | pass |
|
267 | pass | |
259 |
|
268 | |||
260 | f = None |
|
269 | f = None | |
261 | try: |
|
270 | try: | |
262 | f = util.atomictempfile(filecachepath, b"wb") |
|
271 | f = util.atomictempfile(filecachepath, b"wb") | |
263 | f.write(text) |
|
272 | f.write(text) | |
264 | except (IOError, OSError): |
|
273 | except (IOError, OSError): | |
265 | # Don't abort if the user only has permission to read, |
|
274 | # Don't abort if the user only has permission to read, | |
266 | # and not write. |
|
275 | # and not write. | |
267 | pass |
|
276 | pass | |
268 | finally: |
|
277 | finally: | |
269 | if f: |
|
278 | if f: | |
270 | f.close() |
|
279 | f.close() | |
271 | finally: |
|
280 | finally: | |
272 | os.umask(oldumask) |
|
281 | os.umask(oldumask) | |
273 | else: |
|
282 | else: | |
274 | with open(filecachepath, b"rb") as f: |
|
283 | with open(filecachepath, b"rb") as f: | |
275 | text = f.read() |
|
284 | text = f.read() | |
276 | return text |
|
285 | return text | |
277 |
|
286 | |||
278 |
|
287 | |||
279 | def getflogheads(repo, proto, path): |
|
288 | def getflogheads(repo, proto, path): | |
280 | """A server api for requesting a filelog's heads""" |
|
289 | """A server api for requesting a filelog's heads""" | |
281 | flog = repo.file(path) |
|
290 | flog = repo.file(path) | |
282 | heads = flog.heads() |
|
291 | heads = flog.heads() | |
283 | return b'\n'.join((hex(head) for head in heads if head != repo.nullid)) |
|
292 | return b'\n'.join((hex(head) for head in heads if head != repo.nullid)) | |
284 |
|
293 | |||
285 |
|
294 | |||
286 | def getfile(repo, proto, file, node): |
|
295 | def getfile(repo, proto, file, node): | |
287 | """A server api for requesting a particular version of a file. Can be used |
|
296 | """A server api for requesting a particular version of a file. Can be used | |
288 | in batches to request many files at once. The return protocol is: |
|
297 | in batches to request many files at once. The return protocol is: | |
289 | <errorcode>\0<data/errormsg> where <errorcode> is 0 for success or |
|
298 | <errorcode>\0<data/errormsg> where <errorcode> is 0 for success or | |
290 | non-zero for an error. |
|
299 | non-zero for an error. | |
291 |
|
300 | |||
292 | data is a compressed blob with revlog flag and ancestors information. See |
|
301 | data is a compressed blob with revlog flag and ancestors information. See | |
293 | createfileblob for its content. |
|
302 | createfileblob for its content. | |
294 | """ |
|
303 | """ | |
295 | if shallowutil.isenabled(repo): |
|
304 | if shallowutil.isenabled(repo): | |
296 | return b'1\0' + _(b'cannot fetch remote files from shallow repo') |
|
305 | return b'1\0' + _(b'cannot fetch remote files from shallow repo') | |
297 | cachepath = repo.ui.config(b"remotefilelog", b"servercachepath") |
|
306 | cachepath = repo.ui.config(b"remotefilelog", b"servercachepath") | |
298 | if not cachepath: |
|
307 | if not cachepath: | |
299 | cachepath = os.path.join(repo.path, b"remotefilelogcache") |
|
308 | cachepath = os.path.join(repo.path, b"remotefilelogcache") | |
300 | node = bin(node.strip()) |
|
309 | node = bin(node.strip()) | |
301 | if node == repo.nullid: |
|
310 | if node == repo.nullid: | |
302 | return b'0\0' |
|
311 | return b'0\0' | |
303 | return b'0\0' + _loadfileblob(repo, cachepath, file, node) |
|
312 | return b'0\0' + _loadfileblob(repo, cachepath, file, node) | |
304 |
|
313 | |||
305 |
|
314 | |||
306 | def getfiles(repo, proto): |
|
315 | def getfiles(repo, proto): | |
307 | """A server api for requesting particular versions of particular files.""" |
|
316 | """A server api for requesting particular versions of particular files.""" | |
308 | if shallowutil.isenabled(repo): |
|
317 | if shallowutil.isenabled(repo): | |
309 | raise error.Abort(_(b'cannot fetch remote files from shallow repo')) |
|
318 | raise error.Abort(_(b'cannot fetch remote files from shallow repo')) | |
310 | if not isinstance(proto, _sshv1server): |
|
319 | if not isinstance(proto, _sshv1server): | |
311 | raise error.Abort(_(b'cannot fetch remote files over non-ssh protocol')) |
|
320 | raise error.Abort(_(b'cannot fetch remote files over non-ssh protocol')) | |
312 |
|
321 | |||
313 | def streamer(): |
|
322 | def streamer(): | |
314 | fin = proto._fin |
|
323 | fin = proto._fin | |
315 |
|
324 | |||
316 | cachepath = repo.ui.config(b"remotefilelog", b"servercachepath") |
|
325 | cachepath = repo.ui.config(b"remotefilelog", b"servercachepath") | |
317 | if not cachepath: |
|
326 | if not cachepath: | |
318 | cachepath = os.path.join(repo.path, b"remotefilelogcache") |
|
327 | cachepath = os.path.join(repo.path, b"remotefilelogcache") | |
319 |
|
328 | |||
320 | while True: |
|
329 | while True: | |
321 | request = fin.readline()[:-1] |
|
330 | request = fin.readline()[:-1] | |
322 | if not request: |
|
331 | if not request: | |
323 | break |
|
332 | break | |
324 |
|
333 | |||
325 | node = bin(request[:40]) |
|
334 | node = bin(request[:40]) | |
326 | if node == repo.nullid: |
|
335 | if node == repo.nullid: | |
327 | yield b'0\n' |
|
336 | yield b'0\n' | |
328 | continue |
|
337 | continue | |
329 |
|
338 | |||
330 | path = request[40:] |
|
339 | path = request[40:] | |
331 |
|
340 | |||
332 | text = _loadfileblob(repo, cachepath, path, node) |
|
341 | text = _loadfileblob(repo, cachepath, path, node) | |
333 |
|
342 | |||
334 | yield b'%d\n%s' % (len(text), text) |
|
343 | yield b'%d\n%s' % (len(text), text) | |
335 |
|
344 | |||
336 | # it would be better to only flush after processing a whole batch |
|
345 | # it would be better to only flush after processing a whole batch | |
337 | # but currently we don't know if there are more requests coming |
|
346 | # but currently we don't know if there are more requests coming | |
338 | proto._fout.flush() |
|
347 | proto._fout.flush() | |
339 |
|
348 | |||
340 | return wireprototypes.streamres(streamer()) |
|
349 | return wireprototypes.streamres(streamer()) | |
341 |
|
350 | |||
342 |
|
351 | |||
343 | def createfileblob(filectx): |
|
352 | def createfileblob(filectx): | |
344 | """ |
|
353 | """ | |
345 | format: |
|
354 | format: | |
346 | v0: |
|
355 | v0: | |
347 | str(len(rawtext)) + '\0' + rawtext + ancestortext |
|
356 | str(len(rawtext)) + '\0' + rawtext + ancestortext | |
348 | v1: |
|
357 | v1: | |
349 | 'v1' + '\n' + metalist + '\0' + rawtext + ancestortext |
|
358 | 'v1' + '\n' + metalist + '\0' + rawtext + ancestortext | |
350 | metalist := metalist + '\n' + meta | meta |
|
359 | metalist := metalist + '\n' + meta | meta | |
351 | meta := sizemeta | flagmeta |
|
360 | meta := sizemeta | flagmeta | |
352 | sizemeta := METAKEYSIZE + str(len(rawtext)) |
|
361 | sizemeta := METAKEYSIZE + str(len(rawtext)) | |
353 | flagmeta := METAKEYFLAG + str(flag) |
|
362 | flagmeta := METAKEYFLAG + str(flag) | |
354 |
|
363 | |||
355 | note: sizemeta must exist. METAKEYFLAG and METAKEYSIZE must have a |
|
364 | note: sizemeta must exist. METAKEYFLAG and METAKEYSIZE must have a | |
356 | length of 1. |
|
365 | length of 1. | |
357 | """ |
|
366 | """ | |
358 | flog = filectx.filelog() |
|
367 | flog = filectx.filelog() | |
359 | frev = filectx.filerev() |
|
368 | frev = filectx.filerev() | |
360 | revlogflags = flog._revlog.flags(frev) |
|
369 | revlogflags = flog._revlog.flags(frev) | |
361 | if revlogflags == 0: |
|
370 | if revlogflags == 0: | |
362 | # normal files |
|
371 | # normal files | |
363 | text = filectx.data() |
|
372 | text = filectx.data() | |
364 | else: |
|
373 | else: | |
365 | # lfs, read raw revision data |
|
374 | # lfs, read raw revision data | |
366 | text = flog.rawdata(frev) |
|
375 | text = flog.rawdata(frev) | |
367 |
|
376 | |||
368 | repo = filectx._repo |
|
377 | repo = filectx._repo | |
369 |
|
378 | |||
370 | ancestors = [filectx] |
|
379 | ancestors = [filectx] | |
371 |
|
380 | |||
372 | try: |
|
381 | try: | |
373 | repo.forcelinkrev = True |
|
382 | repo.forcelinkrev = True | |
374 | ancestors.extend([f for f in filectx.ancestors()]) |
|
383 | ancestors.extend([f for f in filectx.ancestors()]) | |
375 |
|
384 | |||
376 | ancestortext = b"" |
|
385 | ancestortext = b"" | |
377 | for ancestorctx in ancestors: |
|
386 | for ancestorctx in ancestors: | |
378 | parents = ancestorctx.parents() |
|
387 | parents = ancestorctx.parents() | |
379 | p1 = repo.nullid |
|
388 | p1 = repo.nullid | |
380 | p2 = repo.nullid |
|
389 | p2 = repo.nullid | |
381 | if len(parents) > 0: |
|
390 | if len(parents) > 0: | |
382 | p1 = parents[0].filenode() |
|
391 | p1 = parents[0].filenode() | |
383 | if len(parents) > 1: |
|
392 | if len(parents) > 1: | |
384 | p2 = parents[1].filenode() |
|
393 | p2 = parents[1].filenode() | |
385 |
|
394 | |||
386 | copyname = b"" |
|
395 | copyname = b"" | |
387 | rename = ancestorctx.renamed() |
|
396 | rename = ancestorctx.renamed() | |
388 | if rename: |
|
397 | if rename: | |
389 | copyname = rename[0] |
|
398 | copyname = rename[0] | |
390 | linknode = ancestorctx.node() |
|
399 | linknode = ancestorctx.node() | |
391 | ancestortext += b"%s%s%s%s%s\0" % ( |
|
400 | ancestortext += b"%s%s%s%s%s\0" % ( | |
392 | ancestorctx.filenode(), |
|
401 | ancestorctx.filenode(), | |
393 | p1, |
|
402 | p1, | |
394 | p2, |
|
403 | p2, | |
395 | linknode, |
|
404 | linknode, | |
396 | copyname, |
|
405 | copyname, | |
397 | ) |
|
406 | ) | |
398 | finally: |
|
407 | finally: | |
399 | repo.forcelinkrev = False |
|
408 | repo.forcelinkrev = False | |
400 |
|
409 | |||
401 | header = shallowutil.buildfileblobheader(len(text), revlogflags) |
|
410 | header = shallowutil.buildfileblobheader(len(text), revlogflags) | |
402 |
|
411 | |||
403 | return b"%s\0%s%s" % (header, text, ancestortext) |
|
412 | return b"%s\0%s%s" % (header, text, ancestortext) | |
404 |
|
413 | |||
405 |
|
414 | |||
406 | def gcserver(ui, repo): |
|
415 | def gcserver(ui, repo): | |
407 | if not repo.ui.configbool(b"remotefilelog", b"server"): |
|
416 | if not repo.ui.configbool(b"remotefilelog", b"server"): | |
408 | return |
|
417 | return | |
409 |
|
418 | |||
410 | neededfiles = set() |
|
419 | neededfiles = set() | |
411 | heads = repo.revs(b"heads(tip~25000:) - null") |
|
420 | heads = repo.revs(b"heads(tip~25000:) - null") | |
412 |
|
421 | |||
413 | cachepath = repo.vfs.join(b"remotefilelogcache") |
|
422 | cachepath = repo.vfs.join(b"remotefilelogcache") | |
414 | for head in heads: |
|
423 | for head in heads: | |
415 | mf = repo[head].manifest() |
|
424 | mf = repo[head].manifest() | |
416 | for filename, filenode in mf.items(): |
|
425 | for filename, filenode in mf.items(): | |
417 | filecachepath = os.path.join(cachepath, filename, hex(filenode)) |
|
426 | filecachepath = os.path.join(cachepath, filename, hex(filenode)) | |
418 | neededfiles.add(filecachepath) |
|
427 | neededfiles.add(filecachepath) | |
419 |
|
428 | |||
420 | # delete unneeded older files |
|
429 | # delete unneeded older files | |
421 | days = repo.ui.configint(b"remotefilelog", b"serverexpiration") |
|
430 | days = repo.ui.configint(b"remotefilelog", b"serverexpiration") | |
422 | expiration = time.time() - (days * 24 * 60 * 60) |
|
431 | expiration = time.time() - (days * 24 * 60 * 60) | |
423 |
|
432 | |||
424 | progress = ui.makeprogress(_(b"removing old server cache"), unit=b"files") |
|
433 | progress = ui.makeprogress(_(b"removing old server cache"), unit=b"files") | |
425 | progress.update(0) |
|
434 | progress.update(0) | |
426 | for root, dirs, files in os.walk(cachepath): |
|
435 | for root, dirs, files in os.walk(cachepath): | |
427 | for file in files: |
|
436 | for file in files: | |
428 | filepath = os.path.join(root, file) |
|
437 | filepath = os.path.join(root, file) | |
429 | progress.increment() |
|
438 | progress.increment() | |
430 | if filepath in neededfiles: |
|
439 | if filepath in neededfiles: | |
431 | continue |
|
440 | continue | |
432 |
|
441 | |||
433 | stat = os.stat(filepath) |
|
442 | stat = os.stat(filepath) | |
434 | if stat.st_mtime < expiration: |
|
443 | if stat.st_mtime < expiration: | |
435 | os.remove(filepath) |
|
444 | os.remove(filepath) | |
436 |
|
445 | |||
437 | progress.complete() |
|
446 | progress.complete() |
@@ -1,574 +1,576 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 |
|
9 | |||
10 | from .i18n import _ |
|
10 | from .i18n import _ | |
11 | from .node import ( |
|
11 | from .node import ( | |
12 | hex, |
|
12 | hex, | |
13 | short, |
|
13 | short, | |
14 | ) |
|
14 | ) | |
15 | from . import ( |
|
15 | from . import ( | |
16 | bundle2, |
|
16 | bundle2, | |
17 | changegroup, |
|
17 | changegroup, | |
18 | discovery, |
|
18 | discovery, | |
19 | error, |
|
19 | error, | |
20 | exchange, |
|
20 | exchange, | |
21 | obsolete, |
|
21 | obsolete, | |
22 | obsutil, |
|
22 | obsutil, | |
23 | pathutil, |
|
23 | pathutil, | |
24 | phases, |
|
24 | phases, | |
25 | requirements, |
|
25 | requirements, | |
26 | scmutil, |
|
26 | scmutil, | |
27 | transaction, |
|
27 | transaction, | |
28 | util, |
|
28 | util, | |
29 | ) |
|
29 | ) | |
30 | from .utils import ( |
|
30 | from .utils import ( | |
31 | hashutil, |
|
31 | hashutil, | |
32 | urlutil, |
|
32 | urlutil, | |
33 | ) |
|
33 | ) | |
34 |
|
34 | |||
35 |
|
35 | |||
36 | def backupbundle( |
|
36 | def backupbundle( | |
37 | repo, |
|
37 | repo, | |
38 | bases, |
|
38 | bases, | |
39 | heads, |
|
39 | heads, | |
40 | node, |
|
40 | node, | |
41 | suffix, |
|
41 | suffix, | |
42 | compress=True, |
|
42 | compress=True, | |
43 | obsolescence=True, |
|
43 | obsolescence=True, | |
44 | tmp_backup=False, |
|
44 | tmp_backup=False, | |
45 | ): |
|
45 | ): | |
46 | """create a bundle with the specified revisions as a backup""" |
|
46 | """create a bundle with the specified revisions as a backup""" | |
47 |
|
47 | |||
48 | backupdir = b"strip-backup" |
|
48 | backupdir = b"strip-backup" | |
49 | vfs = repo.vfs |
|
49 | vfs = repo.vfs | |
50 | if not vfs.isdir(backupdir): |
|
50 | if not vfs.isdir(backupdir): | |
51 | vfs.mkdir(backupdir) |
|
51 | vfs.mkdir(backupdir) | |
52 |
|
52 | |||
53 | # Include a hash of all the nodes in the filename for uniqueness |
|
53 | # Include a hash of all the nodes in the filename for uniqueness | |
54 | allcommits = repo.set(b'%ln::%ln', bases, heads) |
|
54 | allcommits = repo.set(b'%ln::%ln', bases, heads) | |
55 | allhashes = sorted(c.hex() for c in allcommits) |
|
55 | allhashes = sorted(c.hex() for c in allcommits) | |
56 | totalhash = hashutil.sha1(b''.join(allhashes)).digest() |
|
56 | totalhash = hashutil.sha1(b''.join(allhashes)).digest() | |
57 | name = b"%s/%s-%s-%s.hg" % ( |
|
57 | name = b"%s/%s-%s-%s.hg" % ( | |
58 | backupdir, |
|
58 | backupdir, | |
59 | short(node), |
|
59 | short(node), | |
60 | hex(totalhash[:4]), |
|
60 | hex(totalhash[:4]), | |
61 | suffix, |
|
61 | suffix, | |
62 | ) |
|
62 | ) | |
63 |
|
63 | |||
64 | cgversion = changegroup.localversion(repo) |
|
64 | cgversion = changegroup.localversion(repo) | |
65 | comp = None |
|
65 | comp = None | |
66 | if cgversion != b'01': |
|
66 | if cgversion != b'01': | |
67 | bundletype = b"HG20" |
|
67 | bundletype = b"HG20" | |
68 | if compress: |
|
68 | if compress: | |
69 | comp = b'BZ' |
|
69 | comp = b'BZ' | |
70 | elif compress: |
|
70 | elif compress: | |
71 | bundletype = b"HG10BZ" |
|
71 | bundletype = b"HG10BZ" | |
72 | else: |
|
72 | else: | |
73 | bundletype = b"HG10UN" |
|
73 | bundletype = b"HG10UN" | |
74 |
|
74 | |||
75 | outgoing = discovery.outgoing(repo, missingroots=bases, ancestorsof=heads) |
|
75 | outgoing = discovery.outgoing(repo, missingroots=bases, ancestorsof=heads) | |
76 | contentopts = { |
|
76 | contentopts = { | |
77 | b'cg.version': cgversion, |
|
77 | b'cg.version': cgversion, | |
78 | b'obsolescence': obsolescence, |
|
78 | b'obsolescence': obsolescence, | |
79 | b'phases': True, |
|
79 | b'phases': True, | |
80 | } |
|
80 | } | |
81 | return bundle2.writenewbundle( |
|
81 | return bundle2.writenewbundle( | |
82 | repo.ui, |
|
82 | repo.ui, | |
83 | repo, |
|
83 | repo, | |
84 | b'strip', |
|
84 | b'strip', | |
85 | name, |
|
85 | name, | |
86 | bundletype, |
|
86 | bundletype, | |
87 | outgoing, |
|
87 | outgoing, | |
88 | contentopts, |
|
88 | contentopts, | |
89 | vfs, |
|
89 | vfs, | |
90 | compression=comp, |
|
90 | compression=comp, | |
91 | allow_internal=tmp_backup, |
|
91 | allow_internal=tmp_backup, | |
92 | ) |
|
92 | ) | |
93 |
|
93 | |||
94 |
|
94 | |||
95 | def _collectfiles(repo, striprev): |
|
95 | def _collectfiles(repo, striprev): | |
96 | """find out the filelogs affected by the strip""" |
|
96 | """find out the filelogs affected by the strip""" | |
97 | files = set() |
|
97 | files = set() | |
98 |
|
98 | |||
99 | for x in range(striprev, len(repo)): |
|
99 | for x in range(striprev, len(repo)): | |
100 | files.update(repo[x].files()) |
|
100 | files.update(repo[x].files()) | |
101 |
|
101 | |||
102 | return sorted(files) |
|
102 | return sorted(files) | |
103 |
|
103 | |||
104 |
|
104 | |||
105 | def _collectrevlog(revlog, striprev): |
|
105 | def _collectrevlog(revlog, striprev): | |
106 | _, brokenset = revlog.getstrippoint(striprev) |
|
106 | _, brokenset = revlog.getstrippoint(striprev) | |
107 | return [revlog.linkrev(r) for r in brokenset] |
|
107 | return [revlog.linkrev(r) for r in brokenset] | |
108 |
|
108 | |||
109 |
|
109 | |||
110 | def _collectbrokencsets(repo, files, striprev): |
|
110 | def _collectbrokencsets(repo, files, striprev): | |
111 | """return the changesets which will be broken by the truncation""" |
|
111 | """return the changesets which will be broken by the truncation""" | |
112 | s = set() |
|
112 | s = set() | |
113 |
|
113 | |||
114 | for revlog in manifestrevlogs(repo): |
|
114 | for revlog in manifestrevlogs(repo): | |
115 | s.update(_collectrevlog(revlog, striprev)) |
|
115 | s.update(_collectrevlog(revlog, striprev)) | |
116 | for fname in files: |
|
116 | for fname in files: | |
117 | s.update(_collectrevlog(repo.file(fname), striprev)) |
|
117 | s.update(_collectrevlog(repo.file(fname), striprev)) | |
118 |
|
118 | |||
119 | return s |
|
119 | return s | |
120 |
|
120 | |||
121 |
|
121 | |||
122 | def strip(ui, repo, nodelist, backup=True, topic=b'backup'): |
|
122 | def strip(ui, repo, nodelist, backup=True, topic=b'backup'): | |
123 | # This function requires the caller to lock the repo, but it operates |
|
123 | # This function requires the caller to lock the repo, but it operates | |
124 | # within a transaction of its own, and thus requires there to be no current |
|
124 | # within a transaction of its own, and thus requires there to be no current | |
125 | # transaction when it is called. |
|
125 | # transaction when it is called. | |
126 | if repo.currenttransaction() is not None: |
|
126 | if repo.currenttransaction() is not None: | |
127 | raise error.ProgrammingError(b'cannot strip from inside a transaction') |
|
127 | raise error.ProgrammingError(b'cannot strip from inside a transaction') | |
128 |
|
128 | |||
129 | # Simple way to maintain backwards compatibility for this |
|
129 | # Simple way to maintain backwards compatibility for this | |
130 | # argument. |
|
130 | # argument. | |
131 | if backup in [b'none', b'strip']: |
|
131 | if backup in [b'none', b'strip']: | |
132 | backup = False |
|
132 | backup = False | |
133 |
|
133 | |||
134 | repo = repo.unfiltered() |
|
134 | repo = repo.unfiltered() | |
135 | repo.destroying() |
|
135 | repo.destroying() | |
136 | vfs = repo.vfs |
|
136 | vfs = repo.vfs | |
137 | # load bookmark before changelog to avoid side effect from outdated |
|
137 | # load bookmark before changelog to avoid side effect from outdated | |
138 | # changelog (see repo._refreshchangelog) |
|
138 | # changelog (see repo._refreshchangelog) | |
139 | repo._bookmarks |
|
139 | repo._bookmarks | |
140 | cl = repo.changelog |
|
140 | cl = repo.changelog | |
141 |
|
141 | |||
142 | # TODO handle undo of merge sets |
|
142 | # TODO handle undo of merge sets | |
143 | if isinstance(nodelist, bytes): |
|
143 | if isinstance(nodelist, bytes): | |
144 | nodelist = [nodelist] |
|
144 | nodelist = [nodelist] | |
145 | striplist = [cl.rev(node) for node in nodelist] |
|
145 | striplist = [cl.rev(node) for node in nodelist] | |
146 | striprev = min(striplist) |
|
146 | striprev = min(striplist) | |
147 |
|
147 | |||
148 | files = _collectfiles(repo, striprev) |
|
148 | files = _collectfiles(repo, striprev) | |
149 | saverevs = _collectbrokencsets(repo, files, striprev) |
|
149 | saverevs = _collectbrokencsets(repo, files, striprev) | |
150 |
|
150 | |||
151 | # Some revisions with rev > striprev may not be descendants of striprev. |
|
151 | # Some revisions with rev > striprev may not be descendants of striprev. | |
152 | # We have to find these revisions and put them in a bundle, so that |
|
152 | # We have to find these revisions and put them in a bundle, so that | |
153 | # we can restore them after the truncations. |
|
153 | # we can restore them after the truncations. | |
154 | # To create the bundle we use repo.changegroupsubset which requires |
|
154 | # To create the bundle we use repo.changegroupsubset which requires | |
155 | # the list of heads and bases of the set of interesting revisions. |
|
155 | # the list of heads and bases of the set of interesting revisions. | |
156 | # (head = revision in the set that has no descendant in the set; |
|
156 | # (head = revision in the set that has no descendant in the set; | |
157 | # base = revision in the set that has no ancestor in the set) |
|
157 | # base = revision in the set that has no ancestor in the set) | |
158 | tostrip = set(striplist) |
|
158 | tostrip = set(striplist) | |
159 | saveheads = set(saverevs) |
|
159 | saveheads = set(saverevs) | |
160 | for r in cl.revs(start=striprev + 1): |
|
160 | for r in cl.revs(start=striprev + 1): | |
161 | if any(p in tostrip for p in cl.parentrevs(r)): |
|
161 | if any(p in tostrip for p in cl.parentrevs(r)): | |
162 | tostrip.add(r) |
|
162 | tostrip.add(r) | |
163 |
|
163 | |||
164 | if r not in tostrip: |
|
164 | if r not in tostrip: | |
165 | saverevs.add(r) |
|
165 | saverevs.add(r) | |
166 | saveheads.difference_update(cl.parentrevs(r)) |
|
166 | saveheads.difference_update(cl.parentrevs(r)) | |
167 | saveheads.add(r) |
|
167 | saveheads.add(r) | |
168 | saveheads = [cl.node(r) for r in saveheads] |
|
168 | saveheads = [cl.node(r) for r in saveheads] | |
169 |
|
169 | |||
170 | # compute base nodes |
|
170 | # compute base nodes | |
171 | if saverevs: |
|
171 | if saverevs: | |
172 | descendants = set(cl.descendants(saverevs)) |
|
172 | descendants = set(cl.descendants(saverevs)) | |
173 | saverevs.difference_update(descendants) |
|
173 | saverevs.difference_update(descendants) | |
174 | savebases = [cl.node(r) for r in saverevs] |
|
174 | savebases = [cl.node(r) for r in saverevs] | |
175 | stripbases = [cl.node(r) for r in tostrip] |
|
175 | stripbases = [cl.node(r) for r in tostrip] | |
176 |
|
176 | |||
177 | stripobsidx = obsmarkers = () |
|
177 | stripobsidx = obsmarkers = () | |
178 | if repo.ui.configbool(b'devel', b'strip-obsmarkers'): |
|
178 | if repo.ui.configbool(b'devel', b'strip-obsmarkers'): | |
179 | obsmarkers = obsutil.exclusivemarkers(repo, stripbases) |
|
179 | obsmarkers = obsutil.exclusivemarkers(repo, stripbases) | |
180 | if obsmarkers: |
|
180 | if obsmarkers: | |
181 | stripobsidx = [ |
|
181 | stripobsidx = [ | |
182 | i for i, m in enumerate(repo.obsstore) if m in obsmarkers |
|
182 | i for i, m in enumerate(repo.obsstore) if m in obsmarkers | |
183 | ] |
|
183 | ] | |
184 |
|
184 | |||
185 | newbmtarget, updatebm = _bookmarkmovements(repo, tostrip) |
|
185 | newbmtarget, updatebm = _bookmarkmovements(repo, tostrip) | |
186 |
|
186 | |||
187 | backupfile = None |
|
187 | backupfile = None | |
188 | node = nodelist[-1] |
|
188 | node = nodelist[-1] | |
189 | if backup: |
|
189 | if backup: | |
190 | backupfile = _createstripbackup(repo, stripbases, node, topic) |
|
190 | backupfile = _createstripbackup(repo, stripbases, node, topic) | |
191 | # create a changegroup for all the branches we need to keep |
|
191 | # create a changegroup for all the branches we need to keep | |
192 | tmpbundlefile = None |
|
192 | tmpbundlefile = None | |
193 | if saveheads: |
|
193 | if saveheads: | |
194 | # do not compress temporary bundle if we remove it from disk later |
|
194 | # do not compress temporary bundle if we remove it from disk later | |
195 | # |
|
195 | # | |
196 | # We do not include obsolescence, it might re-introduce prune markers |
|
196 | # We do not include obsolescence, it might re-introduce prune markers | |
197 | # we are trying to strip. This is harmless since the stripped markers |
|
197 | # we are trying to strip. This is harmless since the stripped markers | |
198 | # are already backed up and we did not touched the markers for the |
|
198 | # are already backed up and we did not touched the markers for the | |
199 | # saved changesets. |
|
199 | # saved changesets. | |
200 | tmpbundlefile = backupbundle( |
|
200 | tmpbundlefile = backupbundle( | |
201 | repo, |
|
201 | repo, | |
202 | savebases, |
|
202 | savebases, | |
203 | saveheads, |
|
203 | saveheads, | |
204 | node, |
|
204 | node, | |
205 | b'temp', |
|
205 | b'temp', | |
206 | compress=False, |
|
206 | compress=False, | |
207 | obsolescence=False, |
|
207 | obsolescence=False, | |
208 | tmp_backup=True, |
|
208 | tmp_backup=True, | |
209 | ) |
|
209 | ) | |
210 |
|
210 | |||
211 | with ui.uninterruptible(): |
|
211 | with ui.uninterruptible(): | |
212 | try: |
|
212 | try: | |
213 | with repo.transaction(b"strip") as tr: |
|
213 | with repo.transaction(b"strip") as tr: | |
214 | # TODO this code violates the interface abstraction of the |
|
214 | # TODO this code violates the interface abstraction of the | |
215 | # transaction and makes assumptions that file storage is |
|
215 | # transaction and makes assumptions that file storage is | |
216 | # using append-only files. We'll need some kind of storage |
|
216 | # using append-only files. We'll need some kind of storage | |
217 | # API to handle stripping for us. |
|
217 | # API to handle stripping for us. | |
218 | oldfiles = set(tr._offsetmap.keys()) |
|
218 | oldfiles = set(tr._offsetmap.keys()) | |
219 | oldfiles.update(tr._newfiles) |
|
219 | oldfiles.update(tr._newfiles) | |
220 |
|
220 | |||
221 | tr.startgroup() |
|
221 | tr.startgroup() | |
222 | cl.strip(striprev, tr) |
|
222 | cl.strip(striprev, tr) | |
223 | stripmanifest(repo, striprev, tr, files) |
|
223 | stripmanifest(repo, striprev, tr, files) | |
224 |
|
224 | |||
225 | for fn in files: |
|
225 | for fn in files: | |
226 | repo.file(fn).strip(striprev, tr) |
|
226 | repo.file(fn).strip(striprev, tr) | |
227 | tr.endgroup() |
|
227 | tr.endgroup() | |
228 |
|
228 | |||
229 | entries = tr.readjournal() |
|
229 | entries = tr.readjournal() | |
230 |
|
230 | |||
231 | for file, troffset in entries: |
|
231 | for file, troffset in entries: | |
232 | if file in oldfiles: |
|
232 | if file in oldfiles: | |
233 | continue |
|
233 | continue | |
234 | with repo.svfs(file, b'a', checkambig=True) as fp: |
|
234 | with repo.svfs(file, b'a', checkambig=True) as fp: | |
235 | fp.truncate(troffset) |
|
235 | fp.truncate(troffset) | |
236 | if troffset == 0: |
|
236 | if troffset == 0: | |
237 | repo.store.markremoved(file) |
|
237 | repo.store.markremoved(file) | |
238 |
|
238 | |||
239 | deleteobsmarkers(repo.obsstore, stripobsidx) |
|
239 | deleteobsmarkers(repo.obsstore, stripobsidx) | |
240 | del repo.obsstore |
|
240 | del repo.obsstore | |
241 | repo.invalidatevolatilesets() |
|
241 | repo.invalidatevolatilesets() | |
242 | repo._phasecache.filterunknown(repo) |
|
242 | repo._phasecache.filterunknown(repo) | |
243 |
|
243 | |||
244 | if tmpbundlefile: |
|
244 | if tmpbundlefile: | |
245 | ui.note(_(b"adding branch\n")) |
|
245 | ui.note(_(b"adding branch\n")) | |
246 | f = vfs.open(tmpbundlefile, b"rb") |
|
246 | f = vfs.open(tmpbundlefile, b"rb") | |
247 | gen = exchange.readbundle(ui, f, tmpbundlefile, vfs) |
|
247 | gen = exchange.readbundle(ui, f, tmpbundlefile, vfs) | |
248 | # silence internal shuffling chatter |
|
248 | # silence internal shuffling chatter | |
249 | maybe_silent = ( |
|
249 | maybe_silent = ( | |
250 | repo.ui.silent() |
|
250 | repo.ui.silent() | |
251 | if not repo.ui.verbose |
|
251 | if not repo.ui.verbose | |
252 | else util.nullcontextmanager() |
|
252 | else util.nullcontextmanager() | |
253 | ) |
|
253 | ) | |
254 | with maybe_silent: |
|
254 | with maybe_silent: | |
255 | tmpbundleurl = b'bundle:' + vfs.join(tmpbundlefile) |
|
255 | tmpbundleurl = b'bundle:' + vfs.join(tmpbundlefile) | |
256 | txnname = b'strip' |
|
256 | txnname = b'strip' | |
257 | if not isinstance(gen, bundle2.unbundle20): |
|
257 | if not isinstance(gen, bundle2.unbundle20): | |
258 | txnname = b"strip\n%s" % urlutil.hidepassword( |
|
258 | txnname = b"strip\n%s" % urlutil.hidepassword( | |
259 | tmpbundleurl |
|
259 | tmpbundleurl | |
260 | ) |
|
260 | ) | |
261 | with repo.transaction(txnname) as tr: |
|
261 | with repo.transaction(txnname) as tr: | |
262 | bundle2.applybundle( |
|
262 | bundle2.applybundle( | |
263 | repo, gen, tr, source=b'strip', url=tmpbundleurl |
|
263 | repo, gen, tr, source=b'strip', url=tmpbundleurl | |
264 | ) |
|
264 | ) | |
265 | f.close() |
|
265 | f.close() | |
266 |
|
266 | |||
267 | with repo.transaction(b'repair') as tr: |
|
267 | with repo.transaction(b'repair') as tr: | |
268 | bmchanges = [(m, repo[newbmtarget].node()) for m in updatebm] |
|
268 | bmchanges = [(m, repo[newbmtarget].node()) for m in updatebm] | |
269 | repo._bookmarks.applychanges(repo, tr, bmchanges) |
|
269 | repo._bookmarks.applychanges(repo, tr, bmchanges) | |
270 |
|
270 | |||
271 | transaction.cleanup_undo_files(repo.ui.warn, repo.vfs_map) |
|
271 | transaction.cleanup_undo_files(repo.ui.warn, repo.vfs_map) | |
272 |
|
272 | |||
273 | except: # re-raises |
|
273 | except: # re-raises | |
274 | if backupfile: |
|
274 | if backupfile: | |
275 | ui.warn( |
|
275 | ui.warn( | |
276 | _(b"strip failed, backup bundle stored in '%s'\n") |
|
276 | _(b"strip failed, backup bundle stored in '%s'\n") | |
277 | % vfs.join(backupfile) |
|
277 | % vfs.join(backupfile) | |
278 | ) |
|
278 | ) | |
279 | if tmpbundlefile: |
|
279 | if tmpbundlefile: | |
280 | ui.warn( |
|
280 | ui.warn( | |
281 | _(b"strip failed, unrecovered changes stored in '%s'\n") |
|
281 | _(b"strip failed, unrecovered changes stored in '%s'\n") | |
282 | % vfs.join(tmpbundlefile) |
|
282 | % vfs.join(tmpbundlefile) | |
283 | ) |
|
283 | ) | |
284 | ui.warn( |
|
284 | ui.warn( | |
285 | _( |
|
285 | _( | |
286 | b"(fix the problem, then recover the changesets with " |
|
286 | b"(fix the problem, then recover the changesets with " | |
287 | b"\"hg unbundle '%s'\")\n" |
|
287 | b"\"hg unbundle '%s'\")\n" | |
288 | ) |
|
288 | ) | |
289 | % vfs.join(tmpbundlefile) |
|
289 | % vfs.join(tmpbundlefile) | |
290 | ) |
|
290 | ) | |
291 | raise |
|
291 | raise | |
292 | else: |
|
292 | else: | |
293 | if tmpbundlefile: |
|
293 | if tmpbundlefile: | |
294 | # Remove temporary bundle only if there were no exceptions |
|
294 | # Remove temporary bundle only if there were no exceptions | |
295 | vfs.unlink(tmpbundlefile) |
|
295 | vfs.unlink(tmpbundlefile) | |
296 |
|
296 | |||
297 | repo.destroyed() |
|
297 | repo.destroyed() | |
298 | # return the backup file path (or None if 'backup' was False) so |
|
298 | # return the backup file path (or None if 'backup' was False) so | |
299 | # extensions can use it |
|
299 | # extensions can use it | |
300 | return backupfile |
|
300 | return backupfile | |
301 |
|
301 | |||
302 |
|
302 | |||
303 | def softstrip(ui, repo, nodelist, backup=True, topic=b'backup'): |
|
303 | def softstrip(ui, repo, nodelist, backup=True, topic=b'backup'): | |
304 | """perform a "soft" strip using the archived phase""" |
|
304 | """perform a "soft" strip using the archived phase""" | |
305 | tostrip = [c.node() for c in repo.set(b'sort(%ln::)', nodelist)] |
|
305 | tostrip = [c.node() for c in repo.set(b'sort(%ln::)', nodelist)] | |
306 | if not tostrip: |
|
306 | if not tostrip: | |
307 | return None |
|
307 | return None | |
308 |
|
308 | |||
309 | backupfile = None |
|
309 | backupfile = None | |
310 | if backup: |
|
310 | if backup: | |
311 | node = tostrip[0] |
|
311 | node = tostrip[0] | |
312 | backupfile = _createstripbackup(repo, tostrip, node, topic) |
|
312 | backupfile = _createstripbackup(repo, tostrip, node, topic) | |
313 |
|
313 | |||
314 | newbmtarget, updatebm = _bookmarkmovements(repo, tostrip) |
|
314 | newbmtarget, updatebm = _bookmarkmovements(repo, tostrip) | |
315 | with repo.transaction(b'strip') as tr: |
|
315 | with repo.transaction(b'strip') as tr: | |
316 | phases.retractboundary(repo, tr, phases.archived, tostrip) |
|
316 | phases.retractboundary(repo, tr, phases.archived, tostrip) | |
317 | bmchanges = [(m, repo[newbmtarget].node()) for m in updatebm] |
|
317 | bmchanges = [(m, repo[newbmtarget].node()) for m in updatebm] | |
318 | repo._bookmarks.applychanges(repo, tr, bmchanges) |
|
318 | repo._bookmarks.applychanges(repo, tr, bmchanges) | |
319 | return backupfile |
|
319 | return backupfile | |
320 |
|
320 | |||
321 |
|
321 | |||
322 | def _bookmarkmovements(repo, tostrip): |
|
322 | def _bookmarkmovements(repo, tostrip): | |
323 | # compute necessary bookmark movement |
|
323 | # compute necessary bookmark movement | |
324 | bm = repo._bookmarks |
|
324 | bm = repo._bookmarks | |
325 | updatebm = [] |
|
325 | updatebm = [] | |
326 | for m in bm: |
|
326 | for m in bm: | |
327 | rev = repo[bm[m]].rev() |
|
327 | rev = repo[bm[m]].rev() | |
328 | if rev in tostrip: |
|
328 | if rev in tostrip: | |
329 | updatebm.append(m) |
|
329 | updatebm.append(m) | |
330 | newbmtarget = None |
|
330 | newbmtarget = None | |
331 | # If we need to move bookmarks, compute bookmark |
|
331 | # If we need to move bookmarks, compute bookmark | |
332 | # targets. Otherwise we can skip doing this logic. |
|
332 | # targets. Otherwise we can skip doing this logic. | |
333 | if updatebm: |
|
333 | if updatebm: | |
334 | # For a set s, max(parents(s) - s) is the same as max(heads(::s - s)), |
|
334 | # For a set s, max(parents(s) - s) is the same as max(heads(::s - s)), | |
335 | # but is much faster |
|
335 | # but is much faster | |
336 | newbmtarget = repo.revs(b'max(parents(%ld) - (%ld))', tostrip, tostrip) |
|
336 | newbmtarget = repo.revs(b'max(parents(%ld) - (%ld))', tostrip, tostrip) | |
337 | if newbmtarget: |
|
337 | if newbmtarget: | |
338 | newbmtarget = repo[newbmtarget.first()].node() |
|
338 | newbmtarget = repo[newbmtarget.first()].node() | |
339 | else: |
|
339 | else: | |
340 | newbmtarget = b'.' |
|
340 | newbmtarget = b'.' | |
341 | return newbmtarget, updatebm |
|
341 | return newbmtarget, updatebm | |
342 |
|
342 | |||
343 |
|
343 | |||
344 | def _createstripbackup(repo, stripbases, node, topic): |
|
344 | def _createstripbackup(repo, stripbases, node, topic): | |
345 | # backup the changeset we are about to strip |
|
345 | # backup the changeset we are about to strip | |
346 | vfs = repo.vfs |
|
346 | vfs = repo.vfs | |
347 | unfi = repo.unfiltered() |
|
347 | unfi = repo.unfiltered() | |
348 | to_node = unfi.changelog.node |
|
348 | to_node = unfi.changelog.node | |
349 | # internal changeset are internal implementation details that should not |
|
349 | # internal changeset are internal implementation details that should not | |
350 | # leave the repository and not be exposed to the users. In addition feature |
|
350 | # leave the repository and not be exposed to the users. In addition feature | |
351 | # using them requires to be resistant to strip. See test case for more |
|
351 | # using them requires to be resistant to strip. See test case for more | |
352 | # details. |
|
352 | # details. | |
353 | all_backup = unfi.revs( |
|
353 | all_backup = unfi.revs( | |
354 | b"(%ln)::(%ld) and not _internal()", |
|
354 | b"(%ln)::(%ld) and not _internal()", | |
355 | stripbases, |
|
355 | stripbases, | |
356 | unfi.changelog.headrevs(), |
|
356 | unfi.changelog.headrevs(), | |
357 | ) |
|
357 | ) | |
358 | if not all_backup: |
|
358 | if not all_backup: | |
359 | return None |
|
359 | return None | |
360 |
|
360 | |||
361 | def to_nodes(revs): |
|
361 | def to_nodes(revs): | |
362 | return [to_node(r) for r in revs] |
|
362 | return [to_node(r) for r in revs] | |
363 |
|
363 | |||
364 | bases = to_nodes(unfi.revs("roots(%ld)", all_backup)) |
|
364 | bases = to_nodes(unfi.revs("roots(%ld)", all_backup)) | |
365 | heads = to_nodes(unfi.revs("heads(%ld)", all_backup)) |
|
365 | heads = to_nodes(unfi.revs("heads(%ld)", all_backup)) | |
366 | backupfile = backupbundle(repo, bases, heads, node, topic) |
|
366 | backupfile = backupbundle(repo, bases, heads, node, topic) | |
367 | repo.ui.status(_(b"saved backup bundle to %s\n") % vfs.join(backupfile)) |
|
367 | repo.ui.status(_(b"saved backup bundle to %s\n") % vfs.join(backupfile)) | |
368 | repo.ui.log( |
|
368 | repo.ui.log( | |
369 | b"backupbundle", b"saved backup bundle to %s\n", vfs.join(backupfile) |
|
369 | b"backupbundle", b"saved backup bundle to %s\n", vfs.join(backupfile) | |
370 | ) |
|
370 | ) | |
371 | return backupfile |
|
371 | return backupfile | |
372 |
|
372 | |||
373 |
|
373 | |||
374 | def safestriproots(ui, repo, nodes): |
|
374 | def safestriproots(ui, repo, nodes): | |
375 | """return list of roots of nodes where descendants are covered by nodes""" |
|
375 | """return list of roots of nodes where descendants are covered by nodes""" | |
376 | torev = repo.unfiltered().changelog.rev |
|
376 | torev = repo.unfiltered().changelog.rev | |
377 | revs = {torev(n) for n in nodes} |
|
377 | revs = {torev(n) for n in nodes} | |
378 | # tostrip = wanted - unsafe = wanted - ancestors(orphaned) |
|
378 | # tostrip = wanted - unsafe = wanted - ancestors(orphaned) | |
379 | # orphaned = affected - wanted |
|
379 | # orphaned = affected - wanted | |
380 | # affected = descendants(roots(wanted)) |
|
380 | # affected = descendants(roots(wanted)) | |
381 | # wanted = revs |
|
381 | # wanted = revs | |
382 | revset = b'%ld - ( ::( (roots(%ld):: and not _phase(%s)) -%ld) )' |
|
382 | revset = b'%ld - ( ::( (roots(%ld):: and not _phase(%s)) -%ld) )' | |
383 | tostrip = set(repo.revs(revset, revs, revs, phases.internal, revs)) |
|
383 | tostrip = set(repo.revs(revset, revs, revs, phases.internal, revs)) | |
384 | notstrip = revs - tostrip |
|
384 | notstrip = revs - tostrip | |
385 | if notstrip: |
|
385 | if notstrip: | |
386 | nodestr = b', '.join(sorted(short(repo[n].node()) for n in notstrip)) |
|
386 | nodestr = b', '.join(sorted(short(repo[n].node()) for n in notstrip)) | |
387 | ui.warn( |
|
387 | ui.warn( | |
388 | _(b'warning: orphaned descendants detected, not stripping %s\n') |
|
388 | _(b'warning: orphaned descendants detected, not stripping %s\n') | |
389 | % nodestr |
|
389 | % nodestr | |
390 | ) |
|
390 | ) | |
391 | return [c.node() for c in repo.set(b'roots(%ld)', tostrip)] |
|
391 | return [c.node() for c in repo.set(b'roots(%ld)', tostrip)] | |
392 |
|
392 | |||
393 |
|
393 | |||
394 | class stripcallback: |
|
394 | class stripcallback: | |
395 | """used as a transaction postclose callback""" |
|
395 | """used as a transaction postclose callback""" | |
396 |
|
396 | |||
397 | def __init__(self, ui, repo, backup, topic): |
|
397 | def __init__(self, ui, repo, backup, topic): | |
398 | self.ui = ui |
|
398 | self.ui = ui | |
399 | self.repo = repo |
|
399 | self.repo = repo | |
400 | self.backup = backup |
|
400 | self.backup = backup | |
401 | self.topic = topic or b'backup' |
|
401 | self.topic = topic or b'backup' | |
402 | self.nodelist = [] |
|
402 | self.nodelist = [] | |
403 |
|
403 | |||
404 | def addnodes(self, nodes): |
|
404 | def addnodes(self, nodes): | |
405 | self.nodelist.extend(nodes) |
|
405 | self.nodelist.extend(nodes) | |
406 |
|
406 | |||
407 | def __call__(self, tr): |
|
407 | def __call__(self, tr): | |
408 | roots = safestriproots(self.ui, self.repo, self.nodelist) |
|
408 | roots = safestriproots(self.ui, self.repo, self.nodelist) | |
409 | if roots: |
|
409 | if roots: | |
410 | strip(self.ui, self.repo, roots, self.backup, self.topic) |
|
410 | strip(self.ui, self.repo, roots, self.backup, self.topic) | |
411 |
|
411 | |||
412 |
|
412 | |||
413 | def delayedstrip(ui, repo, nodelist, topic=None, backup=True): |
|
413 | def delayedstrip(ui, repo, nodelist, topic=None, backup=True): | |
414 | """like strip, but works inside transaction and won't strip irreverent revs |
|
414 | """like strip, but works inside transaction and won't strip irreverent revs | |
415 |
|
415 | |||
416 | nodelist must explicitly contain all descendants. Otherwise a warning will |
|
416 | nodelist must explicitly contain all descendants. Otherwise a warning will | |
417 | be printed that some nodes are not stripped. |
|
417 | be printed that some nodes are not stripped. | |
418 |
|
418 | |||
419 | Will do a backup if `backup` is True. The last non-None "topic" will be |
|
419 | Will do a backup if `backup` is True. The last non-None "topic" will be | |
420 | used as the backup topic name. The default backup topic name is "backup". |
|
420 | used as the backup topic name. The default backup topic name is "backup". | |
421 | """ |
|
421 | """ | |
422 | tr = repo.currenttransaction() |
|
422 | tr = repo.currenttransaction() | |
423 | if not tr: |
|
423 | if not tr: | |
424 | nodes = safestriproots(ui, repo, nodelist) |
|
424 | nodes = safestriproots(ui, repo, nodelist) | |
425 | return strip(ui, repo, nodes, backup=backup, topic=topic) |
|
425 | return strip(ui, repo, nodes, backup=backup, topic=topic) | |
426 | # transaction postclose callbacks are called in alphabet order. |
|
426 | # transaction postclose callbacks are called in alphabet order. | |
427 | # use '\xff' as prefix so we are likely to be called last. |
|
427 | # use '\xff' as prefix so we are likely to be called last. | |
428 | callback = tr.getpostclose(b'\xffstrip') |
|
428 | callback = tr.getpostclose(b'\xffstrip') | |
429 | if callback is None: |
|
429 | if callback is None: | |
430 | callback = stripcallback(ui, repo, backup=backup, topic=topic) |
|
430 | callback = stripcallback(ui, repo, backup=backup, topic=topic) | |
431 | tr.addpostclose(b'\xffstrip', callback) |
|
431 | tr.addpostclose(b'\xffstrip', callback) | |
432 | if topic: |
|
432 | if topic: | |
433 | callback.topic = topic |
|
433 | callback.topic = topic | |
434 | callback.addnodes(nodelist) |
|
434 | callback.addnodes(nodelist) | |
435 |
|
435 | |||
436 |
|
436 | |||
437 | def stripmanifest(repo, striprev, tr, files): |
|
437 | def stripmanifest(repo, striprev, tr, files): | |
438 | for revlog in manifestrevlogs(repo): |
|
438 | for revlog in manifestrevlogs(repo): | |
439 | revlog.strip(striprev, tr) |
|
439 | revlog.strip(striprev, tr) | |
440 |
|
440 | |||
441 |
|
441 | |||
442 | def manifestrevlogs(repo): |
|
442 | def manifestrevlogs(repo): | |
443 | yield repo.manifestlog.getstorage(b'') |
|
443 | yield repo.manifestlog.getstorage(b'') | |
444 | if scmutil.istreemanifest(repo): |
|
444 | if scmutil.istreemanifest(repo): | |
445 | # This logic is safe if treemanifest isn't enabled, but also |
|
445 | # This logic is safe if treemanifest isn't enabled, but also | |
446 | # pointless, so we skip it if treemanifest isn't enabled. |
|
446 | # pointless, so we skip it if treemanifest isn't enabled. | |
447 |
for |
|
447 | for entry in repo.store.datafiles(): | |
|
448 | unencoded = entry.unencoded_path | |||
|
449 | # XXX use the entry.revlog_type instead | |||
448 | if unencoded.startswith(b'meta/') and unencoded.endswith( |
|
450 | if unencoded.startswith(b'meta/') and unencoded.endswith( | |
449 | b'00manifest.i' |
|
451 | b'00manifest.i' | |
450 | ): |
|
452 | ): | |
451 | dir = unencoded[5:-12] |
|
453 | dir = unencoded[5:-12] | |
452 | yield repo.manifestlog.getstorage(dir) |
|
454 | yield repo.manifestlog.getstorage(dir) | |
453 |
|
455 | |||
454 |
|
456 | |||
455 | def rebuildfncache(ui, repo, only_data=False): |
|
457 | def rebuildfncache(ui, repo, only_data=False): | |
456 | """Rebuilds the fncache file from repo history. |
|
458 | """Rebuilds the fncache file from repo history. | |
457 |
|
459 | |||
458 | Missing entries will be added. Extra entries will be removed. |
|
460 | Missing entries will be added. Extra entries will be removed. | |
459 | """ |
|
461 | """ | |
460 | repo = repo.unfiltered() |
|
462 | repo = repo.unfiltered() | |
461 |
|
463 | |||
462 | if requirements.FNCACHE_REQUIREMENT not in repo.requirements: |
|
464 | if requirements.FNCACHE_REQUIREMENT not in repo.requirements: | |
463 | ui.warn( |
|
465 | ui.warn( | |
464 | _( |
|
466 | _( | |
465 | b'(not rebuilding fncache because repository does not ' |
|
467 | b'(not rebuilding fncache because repository does not ' | |
466 | b'support fncache)\n' |
|
468 | b'support fncache)\n' | |
467 | ) |
|
469 | ) | |
468 | ) |
|
470 | ) | |
469 | return |
|
471 | return | |
470 |
|
472 | |||
471 | with repo.lock(): |
|
473 | with repo.lock(): | |
472 | fnc = repo.store.fncache |
|
474 | fnc = repo.store.fncache | |
473 | fnc.ensureloaded(warn=ui.warn) |
|
475 | fnc.ensureloaded(warn=ui.warn) | |
474 |
|
476 | |||
475 | oldentries = set(fnc.entries) |
|
477 | oldentries = set(fnc.entries) | |
476 | newentries = set() |
|
478 | newentries = set() | |
477 | seenfiles = set() |
|
479 | seenfiles = set() | |
478 |
|
480 | |||
479 | if only_data: |
|
481 | if only_data: | |
480 | # Trust the listing of .i from the fncache, but not the .d. This is |
|
482 | # Trust the listing of .i from the fncache, but not the .d. This is | |
481 | # much faster, because we only need to stat every possible .d files, |
|
483 | # much faster, because we only need to stat every possible .d files, | |
482 | # instead of reading the full changelog |
|
484 | # instead of reading the full changelog | |
483 | for f in fnc: |
|
485 | for f in fnc: | |
484 | if f[:5] == b'data/' and f[-2:] == b'.i': |
|
486 | if f[:5] == b'data/' and f[-2:] == b'.i': | |
485 | seenfiles.add(f[5:-2]) |
|
487 | seenfiles.add(f[5:-2]) | |
486 | newentries.add(f) |
|
488 | newentries.add(f) | |
487 | dataf = f[:-2] + b'.d' |
|
489 | dataf = f[:-2] + b'.d' | |
488 | if repo.store._exists(dataf): |
|
490 | if repo.store._exists(dataf): | |
489 | newentries.add(dataf) |
|
491 | newentries.add(dataf) | |
490 | else: |
|
492 | else: | |
491 | progress = ui.makeprogress( |
|
493 | progress = ui.makeprogress( | |
492 | _(b'rebuilding'), unit=_(b'changesets'), total=len(repo) |
|
494 | _(b'rebuilding'), unit=_(b'changesets'), total=len(repo) | |
493 | ) |
|
495 | ) | |
494 | for rev in repo: |
|
496 | for rev in repo: | |
495 | progress.update(rev) |
|
497 | progress.update(rev) | |
496 |
|
498 | |||
497 | ctx = repo[rev] |
|
499 | ctx = repo[rev] | |
498 | for f in ctx.files(): |
|
500 | for f in ctx.files(): | |
499 | # This is to minimize I/O. |
|
501 | # This is to minimize I/O. | |
500 | if f in seenfiles: |
|
502 | if f in seenfiles: | |
501 | continue |
|
503 | continue | |
502 | seenfiles.add(f) |
|
504 | seenfiles.add(f) | |
503 |
|
505 | |||
504 | i = b'data/%s.i' % f |
|
506 | i = b'data/%s.i' % f | |
505 | d = b'data/%s.d' % f |
|
507 | d = b'data/%s.d' % f | |
506 |
|
508 | |||
507 | if repo.store._exists(i): |
|
509 | if repo.store._exists(i): | |
508 | newentries.add(i) |
|
510 | newentries.add(i) | |
509 | if repo.store._exists(d): |
|
511 | if repo.store._exists(d): | |
510 | newentries.add(d) |
|
512 | newentries.add(d) | |
511 |
|
513 | |||
512 | progress.complete() |
|
514 | progress.complete() | |
513 |
|
515 | |||
514 | if requirements.TREEMANIFEST_REQUIREMENT in repo.requirements: |
|
516 | if requirements.TREEMANIFEST_REQUIREMENT in repo.requirements: | |
515 | # This logic is safe if treemanifest isn't enabled, but also |
|
517 | # This logic is safe if treemanifest isn't enabled, but also | |
516 | # pointless, so we skip it if treemanifest isn't enabled. |
|
518 | # pointless, so we skip it if treemanifest isn't enabled. | |
517 | for dir in pathutil.dirs(seenfiles): |
|
519 | for dir in pathutil.dirs(seenfiles): | |
518 | i = b'meta/%s/00manifest.i' % dir |
|
520 | i = b'meta/%s/00manifest.i' % dir | |
519 | d = b'meta/%s/00manifest.d' % dir |
|
521 | d = b'meta/%s/00manifest.d' % dir | |
520 |
|
522 | |||
521 | if repo.store._exists(i): |
|
523 | if repo.store._exists(i): | |
522 | newentries.add(i) |
|
524 | newentries.add(i) | |
523 | if repo.store._exists(d): |
|
525 | if repo.store._exists(d): | |
524 | newentries.add(d) |
|
526 | newentries.add(d) | |
525 |
|
527 | |||
526 | addcount = len(newentries - oldentries) |
|
528 | addcount = len(newentries - oldentries) | |
527 | removecount = len(oldentries - newentries) |
|
529 | removecount = len(oldentries - newentries) | |
528 | for p in sorted(oldentries - newentries): |
|
530 | for p in sorted(oldentries - newentries): | |
529 | ui.write(_(b'removing %s\n') % p) |
|
531 | ui.write(_(b'removing %s\n') % p) | |
530 | for p in sorted(newentries - oldentries): |
|
532 | for p in sorted(newentries - oldentries): | |
531 | ui.write(_(b'adding %s\n') % p) |
|
533 | ui.write(_(b'adding %s\n') % p) | |
532 |
|
534 | |||
533 | if addcount or removecount: |
|
535 | if addcount or removecount: | |
534 | ui.write( |
|
536 | ui.write( | |
535 | _(b'%d items added, %d removed from fncache\n') |
|
537 | _(b'%d items added, %d removed from fncache\n') | |
536 | % (addcount, removecount) |
|
538 | % (addcount, removecount) | |
537 | ) |
|
539 | ) | |
538 | fnc.entries = newentries |
|
540 | fnc.entries = newentries | |
539 | fnc._dirty = True |
|
541 | fnc._dirty = True | |
540 |
|
542 | |||
541 | with repo.transaction(b'fncache') as tr: |
|
543 | with repo.transaction(b'fncache') as tr: | |
542 | fnc.write(tr) |
|
544 | fnc.write(tr) | |
543 | else: |
|
545 | else: | |
544 | ui.write(_(b'fncache already up to date\n')) |
|
546 | ui.write(_(b'fncache already up to date\n')) | |
545 |
|
547 | |||
546 |
|
548 | |||
547 | def deleteobsmarkers(obsstore, indices): |
|
549 | def deleteobsmarkers(obsstore, indices): | |
548 | """Delete some obsmarkers from obsstore and return how many were deleted |
|
550 | """Delete some obsmarkers from obsstore and return how many were deleted | |
549 |
|
551 | |||
550 | 'indices' is a list of ints which are the indices |
|
552 | 'indices' is a list of ints which are the indices | |
551 | of the markers to be deleted. |
|
553 | of the markers to be deleted. | |
552 |
|
554 | |||
553 | Every invocation of this function completely rewrites the obsstore file, |
|
555 | Every invocation of this function completely rewrites the obsstore file, | |
554 | skipping the markers we want to be removed. The new temporary file is |
|
556 | skipping the markers we want to be removed. The new temporary file is | |
555 | created, remaining markers are written there and on .close() this file |
|
557 | created, remaining markers are written there and on .close() this file | |
556 | gets atomically renamed to obsstore, thus guaranteeing consistency.""" |
|
558 | gets atomically renamed to obsstore, thus guaranteeing consistency.""" | |
557 | if not indices: |
|
559 | if not indices: | |
558 | # we don't want to rewrite the obsstore with the same content |
|
560 | # we don't want to rewrite the obsstore with the same content | |
559 | return |
|
561 | return | |
560 |
|
562 | |||
561 | left = [] |
|
563 | left = [] | |
562 | current = obsstore._all |
|
564 | current = obsstore._all | |
563 | n = 0 |
|
565 | n = 0 | |
564 | for i, m in enumerate(current): |
|
566 | for i, m in enumerate(current): | |
565 | if i in indices: |
|
567 | if i in indices: | |
566 | n += 1 |
|
568 | n += 1 | |
567 | continue |
|
569 | continue | |
568 | left.append(m) |
|
570 | left.append(m) | |
569 |
|
571 | |||
570 | newobsstorefile = obsstore.svfs(b'obsstore', b'w', atomictemp=True) |
|
572 | newobsstorefile = obsstore.svfs(b'obsstore', b'w', atomictemp=True) | |
571 | for bytes in obsolete.encodemarkers(left, True, obsstore._version): |
|
573 | for bytes in obsolete.encodemarkers(left, True, obsstore._version): | |
572 | newobsstorefile.write(bytes) |
|
574 | newobsstorefile.write(bytes) | |
573 | newobsstorefile.close() |
|
575 | newobsstorefile.close() | |
574 | return n |
|
576 | return n |
@@ -1,883 +1,888 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 |
|
592 | |||
593 | # raw text can be a `memoryview`, which doesn't implement `startswith` |
|
593 | # raw text can be a `memoryview`, which doesn't implement `startswith` | |
594 | has_meta = bytes(raw_text[:2]) == b'\x01\n' |
|
594 | has_meta = bytes(raw_text[:2]) == b'\x01\n' | |
595 | if metadata_cache is not None: |
|
595 | if metadata_cache is not None: | |
596 | metadata_cache[filerev] = has_meta |
|
596 | metadata_cache[filerev] = has_meta | |
597 | if has_meta: |
|
597 | if has_meta: | |
598 | (p1, p2) = parents_revs() |
|
598 | (p1, p2) = parents_revs() | |
599 | if p1 != nullrev and p2 == nullrev: |
|
599 | if p1 != nullrev and p2 == nullrev: | |
600 | return True |
|
600 | return True | |
601 | return False |
|
601 | return False | |
602 |
|
602 | |||
603 |
|
603 | |||
604 | def _is_revision_affected_fast(repo, fl, filerev, metadata_cache): |
|
604 | def _is_revision_affected_fast(repo, fl, filerev, metadata_cache): | |
605 | rl = fl._revlog |
|
605 | rl = fl._revlog | |
606 | is_censored = lambda: rl.iscensored(filerev) |
|
606 | is_censored = lambda: rl.iscensored(filerev) | |
607 | delta_base = lambda: rl.deltaparent(filerev) |
|
607 | delta_base = lambda: rl.deltaparent(filerev) | |
608 | delta = lambda: rl._chunk(filerev) |
|
608 | delta = lambda: rl._chunk(filerev) | |
609 | full_text = lambda: rl.rawdata(filerev) |
|
609 | full_text = lambda: rl.rawdata(filerev) | |
610 | parent_revs = lambda: rl.parentrevs(filerev) |
|
610 | parent_revs = lambda: rl.parentrevs(filerev) | |
611 | return _is_revision_affected_fast_inner( |
|
611 | return _is_revision_affected_fast_inner( | |
612 | is_censored, |
|
612 | is_censored, | |
613 | delta_base, |
|
613 | delta_base, | |
614 | delta, |
|
614 | delta, | |
615 | full_text, |
|
615 | full_text, | |
616 | parent_revs, |
|
616 | parent_revs, | |
617 | filerev, |
|
617 | filerev, | |
618 | metadata_cache, |
|
618 | metadata_cache, | |
619 | ) |
|
619 | ) | |
620 |
|
620 | |||
621 |
|
621 | |||
622 | def _is_revision_affected_fast_inner( |
|
622 | def _is_revision_affected_fast_inner( | |
623 | is_censored, |
|
623 | is_censored, | |
624 | delta_base, |
|
624 | delta_base, | |
625 | delta, |
|
625 | delta, | |
626 | full_text, |
|
626 | full_text, | |
627 | parent_revs, |
|
627 | parent_revs, | |
628 | filerev, |
|
628 | filerev, | |
629 | metadata_cache, |
|
629 | metadata_cache, | |
630 | ): |
|
630 | ): | |
631 | """Optimization fast-path for `_is_revision_affected`. |
|
631 | """Optimization fast-path for `_is_revision_affected`. | |
632 |
|
632 | |||
633 | `metadata_cache` is a dict of `{rev: has_metadata}` which allows any |
|
633 | `metadata_cache` is a dict of `{rev: has_metadata}` which allows any | |
634 | revision to check if its base has metadata, saving computation of the full |
|
634 | revision to check if its base has metadata, saving computation of the full | |
635 | text, instead looking at the current delta. |
|
635 | text, instead looking at the current delta. | |
636 |
|
636 | |||
637 | This optimization only works if the revisions are looked at in order.""" |
|
637 | This optimization only works if the revisions are looked at in order.""" | |
638 |
|
638 | |||
639 | if is_censored(): |
|
639 | if is_censored(): | |
640 | # Censored revisions don't contain metadata, so they cannot be affected |
|
640 | # Censored revisions don't contain metadata, so they cannot be affected | |
641 | metadata_cache[filerev] = False |
|
641 | metadata_cache[filerev] = False | |
642 | return False |
|
642 | return False | |
643 |
|
643 | |||
644 | p1, p2 = parent_revs() |
|
644 | p1, p2 = parent_revs() | |
645 | if p1 == nullrev or p2 != nullrev: |
|
645 | if p1 == nullrev or p2 != nullrev: | |
646 | return False |
|
646 | return False | |
647 |
|
647 | |||
648 | delta_parent = delta_base() |
|
648 | delta_parent = delta_base() | |
649 | parent_has_metadata = metadata_cache.get(delta_parent) |
|
649 | parent_has_metadata = metadata_cache.get(delta_parent) | |
650 | if parent_has_metadata is None: |
|
650 | if parent_has_metadata is None: | |
651 | return _is_revision_affected_inner( |
|
651 | return _is_revision_affected_inner( | |
652 | full_text, |
|
652 | full_text, | |
653 | parent_revs, |
|
653 | parent_revs, | |
654 | filerev, |
|
654 | filerev, | |
655 | metadata_cache, |
|
655 | metadata_cache, | |
656 | ) |
|
656 | ) | |
657 |
|
657 | |||
658 | chunk = delta() |
|
658 | chunk = delta() | |
659 | if not len(chunk): |
|
659 | if not len(chunk): | |
660 | # No diff for this revision |
|
660 | # No diff for this revision | |
661 | return parent_has_metadata |
|
661 | return parent_has_metadata | |
662 |
|
662 | |||
663 | header_length = 12 |
|
663 | header_length = 12 | |
664 | if len(chunk) < header_length: |
|
664 | if len(chunk) < header_length: | |
665 | raise error.Abort(_(b"patch cannot be decoded")) |
|
665 | raise error.Abort(_(b"patch cannot be decoded")) | |
666 |
|
666 | |||
667 | start, _end, _length = struct.unpack(b">lll", chunk[:header_length]) |
|
667 | start, _end, _length = struct.unpack(b">lll", chunk[:header_length]) | |
668 |
|
668 | |||
669 | if start < 2: # len(b'\x01\n') == 2 |
|
669 | if start < 2: # len(b'\x01\n') == 2 | |
670 | # This delta does *something* to the metadata marker (if any). |
|
670 | # This delta does *something* to the metadata marker (if any). | |
671 | # Check it the slow way |
|
671 | # Check it the slow way | |
672 | is_affected = _is_revision_affected_inner( |
|
672 | is_affected = _is_revision_affected_inner( | |
673 | full_text, |
|
673 | full_text, | |
674 | parent_revs, |
|
674 | parent_revs, | |
675 | filerev, |
|
675 | filerev, | |
676 | metadata_cache, |
|
676 | metadata_cache, | |
677 | ) |
|
677 | ) | |
678 | return is_affected |
|
678 | return is_affected | |
679 |
|
679 | |||
680 | # The diff did not remove or add the metadata header, it's then in the same |
|
680 | # The diff did not remove or add the metadata header, it's then in the same | |
681 | # situation as its parent |
|
681 | # situation as its parent | |
682 | metadata_cache[filerev] = parent_has_metadata |
|
682 | metadata_cache[filerev] = parent_has_metadata | |
683 | return parent_has_metadata |
|
683 | return parent_has_metadata | |
684 |
|
684 | |||
685 |
|
685 | |||
686 | def _from_report(ui, repo, context, from_report, dry_run): |
|
686 | def _from_report(ui, repo, context, from_report, dry_run): | |
687 | """ |
|
687 | """ | |
688 | Fix the revisions given in the `from_report` file, but still checks if the |
|
688 | Fix the revisions given in the `from_report` file, but still checks if the | |
689 | revisions are indeed affected to prevent an unfortunate cyclic situation |
|
689 | revisions are indeed affected to prevent an unfortunate cyclic situation | |
690 | where we'd swap well-ordered parents again. |
|
690 | where we'd swap well-ordered parents again. | |
691 |
|
691 | |||
692 | See the doc for `debug_fix_issue6528` for the format documentation. |
|
692 | See the doc for `debug_fix_issue6528` for the format documentation. | |
693 | """ |
|
693 | """ | |
694 | ui.write(_(b"loading report file '%s'\n") % from_report) |
|
694 | ui.write(_(b"loading report file '%s'\n") % from_report) | |
695 |
|
695 | |||
696 | with context(), open(from_report, mode='rb') as f: |
|
696 | with context(), open(from_report, mode='rb') as f: | |
697 | for line in f.read().split(b'\n'): |
|
697 | for line in f.read().split(b'\n'): | |
698 | if not line: |
|
698 | if not line: | |
699 | continue |
|
699 | continue | |
700 | filenodes, filename = line.split(b' ', 1) |
|
700 | filenodes, filename = line.split(b' ', 1) | |
701 | fl = _filelog_from_filename(repo, filename) |
|
701 | fl = _filelog_from_filename(repo, filename) | |
702 | to_fix = set( |
|
702 | to_fix = set( | |
703 | fl.rev(binascii.unhexlify(n)) for n in filenodes.split(b',') |
|
703 | fl.rev(binascii.unhexlify(n)) for n in filenodes.split(b',') | |
704 | ) |
|
704 | ) | |
705 | excluded = set() |
|
705 | excluded = set() | |
706 |
|
706 | |||
707 | for filerev in to_fix: |
|
707 | for filerev in to_fix: | |
708 | if _is_revision_affected(fl, filerev): |
|
708 | if _is_revision_affected(fl, filerev): | |
709 | msg = b"found affected revision %d for filelog '%s'\n" |
|
709 | msg = b"found affected revision %d for filelog '%s'\n" | |
710 | ui.warn(msg % (filerev, filename)) |
|
710 | ui.warn(msg % (filerev, filename)) | |
711 | else: |
|
711 | else: | |
712 | msg = _(b"revision %s of file '%s' is not affected\n") |
|
712 | msg = _(b"revision %s of file '%s' is not affected\n") | |
713 | msg %= (binascii.hexlify(fl.node(filerev)), filename) |
|
713 | msg %= (binascii.hexlify(fl.node(filerev)), filename) | |
714 | ui.warn(msg) |
|
714 | ui.warn(msg) | |
715 | excluded.add(filerev) |
|
715 | excluded.add(filerev) | |
716 |
|
716 | |||
717 | to_fix = to_fix - excluded |
|
717 | to_fix = to_fix - excluded | |
718 | if not to_fix: |
|
718 | if not to_fix: | |
719 | msg = _(b"no affected revisions were found for '%s'\n") |
|
719 | msg = _(b"no affected revisions were found for '%s'\n") | |
720 | ui.write(msg % filename) |
|
720 | ui.write(msg % filename) | |
721 | continue |
|
721 | continue | |
722 | if not dry_run: |
|
722 | if not dry_run: | |
723 | _reorder_filelog_parents(repo, fl, sorted(to_fix)) |
|
723 | _reorder_filelog_parents(repo, fl, sorted(to_fix)) | |
724 |
|
724 | |||
725 |
|
725 | |||
726 | def filter_delta_issue6528(revlog, deltas_iter): |
|
726 | def filter_delta_issue6528(revlog, deltas_iter): | |
727 | """filter incomind deltas to repaire issue 6528 on the fly""" |
|
727 | """filter incomind deltas to repaire issue 6528 on the fly""" | |
728 | metadata_cache = {} |
|
728 | metadata_cache = {} | |
729 |
|
729 | |||
730 | deltacomputer = deltas.deltacomputer(revlog) |
|
730 | deltacomputer = deltas.deltacomputer(revlog) | |
731 |
|
731 | |||
732 | for rev, d in enumerate(deltas_iter, len(revlog)): |
|
732 | for rev, d in enumerate(deltas_iter, len(revlog)): | |
733 | ( |
|
733 | ( | |
734 | node, |
|
734 | node, | |
735 | p1_node, |
|
735 | p1_node, | |
736 | p2_node, |
|
736 | p2_node, | |
737 | linknode, |
|
737 | linknode, | |
738 | deltabase, |
|
738 | deltabase, | |
739 | delta, |
|
739 | delta, | |
740 | flags, |
|
740 | flags, | |
741 | sidedata, |
|
741 | sidedata, | |
742 | ) = d |
|
742 | ) = d | |
743 |
|
743 | |||
744 | if not revlog.index.has_node(deltabase): |
|
744 | if not revlog.index.has_node(deltabase): | |
745 | raise error.LookupError( |
|
745 | raise error.LookupError( | |
746 | deltabase, revlog.radix, _(b'unknown parent') |
|
746 | deltabase, revlog.radix, _(b'unknown parent') | |
747 | ) |
|
747 | ) | |
748 | base_rev = revlog.rev(deltabase) |
|
748 | base_rev = revlog.rev(deltabase) | |
749 | if not revlog.index.has_node(p1_node): |
|
749 | if not revlog.index.has_node(p1_node): | |
750 | raise error.LookupError(p1_node, revlog.radix, _(b'unknown parent')) |
|
750 | raise error.LookupError(p1_node, revlog.radix, _(b'unknown parent')) | |
751 | p1_rev = revlog.rev(p1_node) |
|
751 | p1_rev = revlog.rev(p1_node) | |
752 | if not revlog.index.has_node(p2_node): |
|
752 | if not revlog.index.has_node(p2_node): | |
753 | raise error.LookupError(p2_node, revlog.radix, _(b'unknown parent')) |
|
753 | raise error.LookupError(p2_node, revlog.radix, _(b'unknown parent')) | |
754 | p2_rev = revlog.rev(p2_node) |
|
754 | p2_rev = revlog.rev(p2_node) | |
755 |
|
755 | |||
756 | is_censored = lambda: bool(flags & REVIDX_ISCENSORED) |
|
756 | is_censored = lambda: bool(flags & REVIDX_ISCENSORED) | |
757 | delta_base = lambda: revlog.rev(delta_base) |
|
757 | delta_base = lambda: revlog.rev(delta_base) | |
758 | delta_base = lambda: base_rev |
|
758 | delta_base = lambda: base_rev | |
759 | parent_revs = lambda: (p1_rev, p2_rev) |
|
759 | parent_revs = lambda: (p1_rev, p2_rev) | |
760 |
|
760 | |||
761 | def full_text(): |
|
761 | def full_text(): | |
762 | # note: being able to reuse the full text computation in the |
|
762 | # note: being able to reuse the full text computation in the | |
763 | # underlying addrevision would be useful however this is a bit too |
|
763 | # underlying addrevision would be useful however this is a bit too | |
764 | # intrusive the for the "quick" issue6528 we are writing before the |
|
764 | # intrusive the for the "quick" issue6528 we are writing before the | |
765 | # 5.8 release |
|
765 | # 5.8 release | |
766 | textlen = mdiff.patchedsize(revlog.size(base_rev), delta) |
|
766 | textlen = mdiff.patchedsize(revlog.size(base_rev), delta) | |
767 |
|
767 | |||
768 | revinfo = revlogutils.revisioninfo( |
|
768 | revinfo = revlogutils.revisioninfo( | |
769 | node, |
|
769 | node, | |
770 | p1_node, |
|
770 | p1_node, | |
771 | p2_node, |
|
771 | p2_node, | |
772 | [None], |
|
772 | [None], | |
773 | textlen, |
|
773 | textlen, | |
774 | (base_rev, delta), |
|
774 | (base_rev, delta), | |
775 | flags, |
|
775 | flags, | |
776 | ) |
|
776 | ) | |
777 | # cached by the global "writing" context |
|
777 | # cached by the global "writing" context | |
778 | assert revlog._writinghandles is not None |
|
778 | assert revlog._writinghandles is not None | |
779 | if revlog._inline: |
|
779 | if revlog._inline: | |
780 | fh = revlog._writinghandles[0] |
|
780 | fh = revlog._writinghandles[0] | |
781 | else: |
|
781 | else: | |
782 | fh = revlog._writinghandles[1] |
|
782 | fh = revlog._writinghandles[1] | |
783 | return deltacomputer.buildtext(revinfo, fh) |
|
783 | return deltacomputer.buildtext(revinfo, fh) | |
784 |
|
784 | |||
785 | is_affected = _is_revision_affected_fast_inner( |
|
785 | is_affected = _is_revision_affected_fast_inner( | |
786 | is_censored, |
|
786 | is_censored, | |
787 | delta_base, |
|
787 | delta_base, | |
788 | lambda: delta, |
|
788 | lambda: delta, | |
789 | full_text, |
|
789 | full_text, | |
790 | parent_revs, |
|
790 | parent_revs, | |
791 | rev, |
|
791 | rev, | |
792 | metadata_cache, |
|
792 | metadata_cache, | |
793 | ) |
|
793 | ) | |
794 | if is_affected: |
|
794 | if is_affected: | |
795 | d = ( |
|
795 | d = ( | |
796 | node, |
|
796 | node, | |
797 | p2_node, |
|
797 | p2_node, | |
798 | p1_node, |
|
798 | p1_node, | |
799 | linknode, |
|
799 | linknode, | |
800 | deltabase, |
|
800 | deltabase, | |
801 | delta, |
|
801 | delta, | |
802 | flags, |
|
802 | flags, | |
803 | sidedata, |
|
803 | sidedata, | |
804 | ) |
|
804 | ) | |
805 | yield d |
|
805 | yield d | |
806 |
|
806 | |||
807 |
|
807 | |||
808 | def repair_issue6528( |
|
808 | def repair_issue6528( | |
809 | ui, repo, dry_run=False, to_report=None, from_report=None, paranoid=False |
|
809 | ui, repo, dry_run=False, to_report=None, from_report=None, paranoid=False | |
810 | ): |
|
810 | ): | |
811 | from .. import store # avoid cycle |
|
811 | from .. import store # avoid cycle | |
812 |
|
812 | |||
813 | @contextlib.contextmanager |
|
813 | @contextlib.contextmanager | |
814 | def context(): |
|
814 | def context(): | |
815 | if dry_run or to_report: # No need for locking |
|
815 | if dry_run or to_report: # No need for locking | |
816 | yield |
|
816 | yield | |
817 | else: |
|
817 | else: | |
818 | with repo.wlock(), repo.lock(): |
|
818 | with repo.wlock(), repo.lock(): | |
819 | yield |
|
819 | yield | |
820 |
|
820 | |||
821 | if from_report: |
|
821 | if from_report: | |
822 | return _from_report(ui, repo, context, from_report, dry_run) |
|
822 | return _from_report(ui, repo, context, from_report, dry_run) | |
823 |
|
823 | |||
824 | report_entries = [] |
|
824 | report_entries = [] | |
825 |
|
825 | |||
826 | with context(): |
|
826 | with context(): | |
827 | files = list( |
|
827 | files = list( | |
828 |
|
|
828 | entry | |
829 |
for |
|
829 | for entry in repo.store.datafiles() | |
830 | if path.endswith(b'.i') and file_type & store.FILEFLAGS_FILELOG |
|
830 | if ( | |
|
831 | entry.unencoded_path.endswith(b'.i') | |||
|
832 | and entry.is_revlog | |||
|
833 | and entry.revlog_type == store.FILEFLAGS_FILELOG | |||
|
834 | ) | |||
831 | ) |
|
835 | ) | |
832 |
|
836 | |||
833 | progress = ui.makeprogress( |
|
837 | progress = ui.makeprogress( | |
834 | _(b"looking for affected revisions"), |
|
838 | _(b"looking for affected revisions"), | |
835 | unit=_(b"filelogs"), |
|
839 | unit=_(b"filelogs"), | |
836 | total=len(files), |
|
840 | total=len(files), | |
837 | ) |
|
841 | ) | |
838 | found_nothing = True |
|
842 | found_nothing = True | |
839 |
|
843 | |||
840 |
for |
|
844 | for entry in files: | |
|
845 | path = entry.unencoded_path | |||
841 | progress.increment() |
|
846 | progress.increment() | |
842 | filename = _get_filename_from_filelog_index(path) |
|
847 | filename = _get_filename_from_filelog_index(path) | |
843 | fl = _filelog_from_filename(repo, filename) |
|
848 | fl = _filelog_from_filename(repo, filename) | |
844 |
|
849 | |||
845 | # Set of filerevs (or hex filenodes if `to_report`) that need fixing |
|
850 | # Set of filerevs (or hex filenodes if `to_report`) that need fixing | |
846 | to_fix = set() |
|
851 | to_fix = set() | |
847 | metadata_cache = {} |
|
852 | metadata_cache = {} | |
848 | for filerev in fl.revs(): |
|
853 | for filerev in fl.revs(): | |
849 | affected = _is_revision_affected_fast( |
|
854 | affected = _is_revision_affected_fast( | |
850 | repo, fl, filerev, metadata_cache |
|
855 | repo, fl, filerev, metadata_cache | |
851 | ) |
|
856 | ) | |
852 | if paranoid: |
|
857 | if paranoid: | |
853 | slow = _is_revision_affected(fl, filerev) |
|
858 | slow = _is_revision_affected(fl, filerev) | |
854 | if slow != affected: |
|
859 | if slow != affected: | |
855 | msg = _(b"paranoid check failed for '%s' at node %s") |
|
860 | msg = _(b"paranoid check failed for '%s' at node %s") | |
856 | node = binascii.hexlify(fl.node(filerev)) |
|
861 | node = binascii.hexlify(fl.node(filerev)) | |
857 | raise error.Abort(msg % (filename, node)) |
|
862 | raise error.Abort(msg % (filename, node)) | |
858 | if affected: |
|
863 | if affected: | |
859 | msg = b"found affected revision %d for filelog '%s'\n" |
|
864 | msg = b"found affected revision %d for filelog '%s'\n" | |
860 | ui.warn(msg % (filerev, path)) |
|
865 | ui.warn(msg % (filerev, path)) | |
861 | found_nothing = False |
|
866 | found_nothing = False | |
862 | if not dry_run: |
|
867 | if not dry_run: | |
863 | if to_report: |
|
868 | if to_report: | |
864 | to_fix.add(binascii.hexlify(fl.node(filerev))) |
|
869 | to_fix.add(binascii.hexlify(fl.node(filerev))) | |
865 | else: |
|
870 | else: | |
866 | to_fix.add(filerev) |
|
871 | to_fix.add(filerev) | |
867 |
|
872 | |||
868 | if to_fix: |
|
873 | if to_fix: | |
869 | to_fix = sorted(to_fix) |
|
874 | to_fix = sorted(to_fix) | |
870 | if to_report: |
|
875 | if to_report: | |
871 | report_entries.append((filename, to_fix)) |
|
876 | report_entries.append((filename, to_fix)) | |
872 | else: |
|
877 | else: | |
873 | _reorder_filelog_parents(repo, fl, to_fix) |
|
878 | _reorder_filelog_parents(repo, fl, to_fix) | |
874 |
|
879 | |||
875 | if found_nothing: |
|
880 | if found_nothing: | |
876 | ui.write(_(b"no affected revisions were found\n")) |
|
881 | ui.write(_(b"no affected revisions were found\n")) | |
877 |
|
882 | |||
878 | if to_report and report_entries: |
|
883 | if to_report and report_entries: | |
879 | with open(to_report, mode="wb") as f: |
|
884 | with open(to_report, mode="wb") as f: | |
880 | for path, to_fix in report_entries: |
|
885 | for path, to_fix in report_entries: | |
881 | f.write(b"%s %s\n" % (b",".join(to_fix), path)) |
|
886 | f.write(b"%s %s\n" % (b",".join(to_fix), path)) | |
882 |
|
887 | |||
883 | progress.complete() |
|
888 | progress.complete() |
@@ -1,855 +1,901 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 |
|
8 | |||
9 | import functools |
|
9 | import functools | |
10 | import os |
|
10 | import os | |
11 | import re |
|
11 | import re | |
12 | import stat |
|
12 | import stat | |
|
13 | from typing import Generator | |||
13 |
|
14 | |||
14 | from .i18n import _ |
|
15 | from .i18n import _ | |
15 | from .pycompat import getattr |
|
16 | from .pycompat import getattr | |
|
17 | from .thirdparty import attr | |||
16 | from .node import hex |
|
18 | from .node import hex | |
17 | from . import ( |
|
19 | from . import ( | |
18 | changelog, |
|
20 | changelog, | |
19 | error, |
|
21 | error, | |
20 | manifest, |
|
22 | manifest, | |
21 | policy, |
|
23 | policy, | |
22 | pycompat, |
|
24 | pycompat, | |
23 | util, |
|
25 | util, | |
24 | vfs as vfsmod, |
|
26 | vfs as vfsmod, | |
25 | ) |
|
27 | ) | |
26 | from .utils import hashutil |
|
28 | from .utils import hashutil | |
27 |
|
29 | |||
28 | parsers = policy.importmod('parsers') |
|
30 | parsers = policy.importmod('parsers') | |
29 | # how much bytes should be read from fncache in one read |
|
31 | # how much bytes should be read from fncache in one read | |
30 | # It is done to prevent loading large fncache files into memory |
|
32 | # It is done to prevent loading large fncache files into memory | |
31 | fncache_chunksize = 10 ** 6 |
|
33 | fncache_chunksize = 10 ** 6 | |
32 |
|
34 | |||
33 |
|
35 | |||
34 | def _matchtrackedpath(path, matcher): |
|
36 | def _matchtrackedpath(path, matcher): | |
35 | """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 | |
36 | matched by matcher or not. |
|
38 | matched by matcher or not. | |
37 |
|
39 | |||
38 | If matcher is None, returns True""" |
|
40 | If matcher is None, returns True""" | |
39 |
|
41 | |||
40 | if matcher is None: |
|
42 | if matcher is None: | |
41 | return True |
|
43 | return True | |
42 | path = decodedir(path) |
|
44 | path = decodedir(path) | |
43 | if path.startswith(b'data/'): |
|
45 | if path.startswith(b'data/'): | |
44 | return matcher(path[len(b'data/') : -len(b'.i')]) |
|
46 | return matcher(path[len(b'data/') : -len(b'.i')]) | |
45 | elif path.startswith(b'meta/'): |
|
47 | elif path.startswith(b'meta/'): | |
46 | return matcher.visitdir(path[len(b'meta/') : -len(b'/00manifest.i')]) |
|
48 | return matcher.visitdir(path[len(b'meta/') : -len(b'/00manifest.i')]) | |
47 |
|
49 | |||
48 | raise error.ProgrammingError(b"cannot decode path %s" % path) |
|
50 | raise error.ProgrammingError(b"cannot decode path %s" % path) | |
49 |
|
51 | |||
50 |
|
52 | |||
51 | # 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 | |
52 | # foo.i or foo.d |
|
54 | # foo.i or foo.d | |
53 | def _encodedir(path): |
|
55 | def _encodedir(path): | |
54 | """ |
|
56 | """ | |
55 | >>> _encodedir(b'data/foo.i') |
|
57 | >>> _encodedir(b'data/foo.i') | |
56 | 'data/foo.i' |
|
58 | 'data/foo.i' | |
57 | >>> _encodedir(b'data/foo.i/bla.i') |
|
59 | >>> _encodedir(b'data/foo.i/bla.i') | |
58 | 'data/foo.i.hg/bla.i' |
|
60 | 'data/foo.i.hg/bla.i' | |
59 | >>> _encodedir(b'data/foo.i.hg/bla.i') |
|
61 | >>> _encodedir(b'data/foo.i.hg/bla.i') | |
60 | 'data/foo.i.hg.hg/bla.i' |
|
62 | 'data/foo.i.hg.hg/bla.i' | |
61 | >>> _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') | |
62 | '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' | |
63 | """ |
|
65 | """ | |
64 | return ( |
|
66 | return ( | |
65 | path.replace(b".hg/", b".hg.hg/") |
|
67 | path.replace(b".hg/", b".hg.hg/") | |
66 | .replace(b".i/", b".i.hg/") |
|
68 | .replace(b".i/", b".i.hg/") | |
67 | .replace(b".d/", b".d.hg/") |
|
69 | .replace(b".d/", b".d.hg/") | |
68 | ) |
|
70 | ) | |
69 |
|
71 | |||
70 |
|
72 | |||
71 | encodedir = getattr(parsers, 'encodedir', _encodedir) |
|
73 | encodedir = getattr(parsers, 'encodedir', _encodedir) | |
72 |
|
74 | |||
73 |
|
75 | |||
74 | def decodedir(path): |
|
76 | def decodedir(path): | |
75 | """ |
|
77 | """ | |
76 | >>> decodedir(b'data/foo.i') |
|
78 | >>> decodedir(b'data/foo.i') | |
77 | 'data/foo.i' |
|
79 | 'data/foo.i' | |
78 | >>> decodedir(b'data/foo.i.hg/bla.i') |
|
80 | >>> decodedir(b'data/foo.i.hg/bla.i') | |
79 | 'data/foo.i/bla.i' |
|
81 | 'data/foo.i/bla.i' | |
80 | >>> decodedir(b'data/foo.i.hg.hg/bla.i') |
|
82 | >>> decodedir(b'data/foo.i.hg.hg/bla.i') | |
81 | 'data/foo.i.hg/bla.i' |
|
83 | 'data/foo.i.hg/bla.i' | |
82 | """ |
|
84 | """ | |
83 | if b".hg/" not in path: |
|
85 | if b".hg/" not in path: | |
84 | return path |
|
86 | return path | |
85 | return ( |
|
87 | return ( | |
86 | path.replace(b".d.hg/", b".d/") |
|
88 | path.replace(b".d.hg/", b".d/") | |
87 | .replace(b".i.hg/", b".i/") |
|
89 | .replace(b".i.hg/", b".i/") | |
88 | .replace(b".hg.hg/", b".hg/") |
|
90 | .replace(b".hg.hg/", b".hg/") | |
89 | ) |
|
91 | ) | |
90 |
|
92 | |||
91 |
|
93 | |||
92 | def _reserved(): |
|
94 | def _reserved(): | |
93 | """characters that are problematic for filesystems |
|
95 | """characters that are problematic for filesystems | |
94 |
|
96 | |||
95 | * ascii escapes (0..31) |
|
97 | * ascii escapes (0..31) | |
96 | * ascii hi (126..255) |
|
98 | * ascii hi (126..255) | |
97 | * windows specials |
|
99 | * windows specials | |
98 |
|
100 | |||
99 | these characters will be escaped by encodefunctions |
|
101 | these characters will be escaped by encodefunctions | |
100 | """ |
|
102 | """ | |
101 | winreserved = [ord(x) for x in u'\\:*?"<>|'] |
|
103 | winreserved = [ord(x) for x in u'\\:*?"<>|'] | |
102 | for x in range(32): |
|
104 | for x in range(32): | |
103 | yield x |
|
105 | yield x | |
104 | for x in range(126, 256): |
|
106 | for x in range(126, 256): | |
105 | yield x |
|
107 | yield x | |
106 | for x in winreserved: |
|
108 | for x in winreserved: | |
107 | yield x |
|
109 | yield x | |
108 |
|
110 | |||
109 |
|
111 | |||
110 | def _buildencodefun(): |
|
112 | def _buildencodefun(): | |
111 | """ |
|
113 | """ | |
112 | >>> enc, dec = _buildencodefun() |
|
114 | >>> enc, dec = _buildencodefun() | |
113 |
|
115 | |||
114 | >>> enc(b'nothing/special.txt') |
|
116 | >>> enc(b'nothing/special.txt') | |
115 | 'nothing/special.txt' |
|
117 | 'nothing/special.txt' | |
116 | >>> dec(b'nothing/special.txt') |
|
118 | >>> dec(b'nothing/special.txt') | |
117 | 'nothing/special.txt' |
|
119 | 'nothing/special.txt' | |
118 |
|
120 | |||
119 | >>> enc(b'HELLO') |
|
121 | >>> enc(b'HELLO') | |
120 | '_h_e_l_l_o' |
|
122 | '_h_e_l_l_o' | |
121 | >>> dec(b'_h_e_l_l_o') |
|
123 | >>> dec(b'_h_e_l_l_o') | |
122 | 'HELLO' |
|
124 | 'HELLO' | |
123 |
|
125 | |||
124 | >>> enc(b'hello:world?') |
|
126 | >>> enc(b'hello:world?') | |
125 | 'hello~3aworld~3f' |
|
127 | 'hello~3aworld~3f' | |
126 | >>> dec(b'hello~3aworld~3f') |
|
128 | >>> dec(b'hello~3aworld~3f') | |
127 | 'hello:world?' |
|
129 | 'hello:world?' | |
128 |
|
130 | |||
129 | >>> enc(b'the\\x07quick\\xADshot') |
|
131 | >>> enc(b'the\\x07quick\\xADshot') | |
130 | 'the~07quick~adshot' |
|
132 | 'the~07quick~adshot' | |
131 | >>> dec(b'the~07quick~adshot') |
|
133 | >>> dec(b'the~07quick~adshot') | |
132 | 'the\\x07quick\\xadshot' |
|
134 | 'the\\x07quick\\xadshot' | |
133 | """ |
|
135 | """ | |
134 | e = b'_' |
|
136 | e = b'_' | |
135 | xchr = pycompat.bytechr |
|
137 | xchr = pycompat.bytechr | |
136 | asciistr = list(map(xchr, range(127))) |
|
138 | asciistr = list(map(xchr, range(127))) | |
137 | capitals = list(range(ord(b"A"), ord(b"Z") + 1)) |
|
139 | capitals = list(range(ord(b"A"), ord(b"Z") + 1)) | |
138 |
|
140 | |||
139 | cmap = {x: x for x in asciistr} |
|
141 | cmap = {x: x for x in asciistr} | |
140 | for x in _reserved(): |
|
142 | for x in _reserved(): | |
141 | cmap[xchr(x)] = b"~%02x" % x |
|
143 | cmap[xchr(x)] = b"~%02x" % x | |
142 | for x in capitals + [ord(e)]: |
|
144 | for x in capitals + [ord(e)]: | |
143 | cmap[xchr(x)] = e + xchr(x).lower() |
|
145 | cmap[xchr(x)] = e + xchr(x).lower() | |
144 |
|
146 | |||
145 | dmap = {} |
|
147 | dmap = {} | |
146 | for k, v in cmap.items(): |
|
148 | for k, v in cmap.items(): | |
147 | dmap[v] = k |
|
149 | dmap[v] = k | |
148 |
|
150 | |||
149 | def decode(s): |
|
151 | def decode(s): | |
150 | i = 0 |
|
152 | i = 0 | |
151 | while i < len(s): |
|
153 | while i < len(s): | |
152 | for l in range(1, 4): |
|
154 | for l in range(1, 4): | |
153 | try: |
|
155 | try: | |
154 | yield dmap[s[i : i + l]] |
|
156 | yield dmap[s[i : i + l]] | |
155 | i += l |
|
157 | i += l | |
156 | break |
|
158 | break | |
157 | except KeyError: |
|
159 | except KeyError: | |
158 | pass |
|
160 | pass | |
159 | else: |
|
161 | else: | |
160 | raise KeyError |
|
162 | raise KeyError | |
161 |
|
163 | |||
162 | return ( |
|
164 | return ( | |
163 | lambda s: b''.join([cmap[s[c : c + 1]] for c in range(len(s))]), |
|
165 | lambda s: b''.join([cmap[s[c : c + 1]] for c in range(len(s))]), | |
164 | lambda s: b''.join(list(decode(s))), |
|
166 | lambda s: b''.join(list(decode(s))), | |
165 | ) |
|
167 | ) | |
166 |
|
168 | |||
167 |
|
169 | |||
168 | _encodefname, _decodefname = _buildencodefun() |
|
170 | _encodefname, _decodefname = _buildencodefun() | |
169 |
|
171 | |||
170 |
|
172 | |||
171 | def encodefilename(s): |
|
173 | def encodefilename(s): | |
172 | """ |
|
174 | """ | |
173 | >>> encodefilename(b'foo.i/bar.d/bla.hg/hi:world?/HELLO') |
|
175 | >>> encodefilename(b'foo.i/bar.d/bla.hg/hi:world?/HELLO') | |
174 | 'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o' |
|
176 | 'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o' | |
175 | """ |
|
177 | """ | |
176 | return _encodefname(encodedir(s)) |
|
178 | return _encodefname(encodedir(s)) | |
177 |
|
179 | |||
178 |
|
180 | |||
179 | def decodefilename(s): |
|
181 | def decodefilename(s): | |
180 | """ |
|
182 | """ | |
181 | >>> decodefilename(b'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o') |
|
183 | >>> decodefilename(b'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o') | |
182 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' |
|
184 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' | |
183 | """ |
|
185 | """ | |
184 | return decodedir(_decodefname(s)) |
|
186 | return decodedir(_decodefname(s)) | |
185 |
|
187 | |||
186 |
|
188 | |||
187 | def _buildlowerencodefun(): |
|
189 | def _buildlowerencodefun(): | |
188 | """ |
|
190 | """ | |
189 | >>> f = _buildlowerencodefun() |
|
191 | >>> f = _buildlowerencodefun() | |
190 | >>> f(b'nothing/special.txt') |
|
192 | >>> f(b'nothing/special.txt') | |
191 | 'nothing/special.txt' |
|
193 | 'nothing/special.txt' | |
192 | >>> f(b'HELLO') |
|
194 | >>> f(b'HELLO') | |
193 | 'hello' |
|
195 | 'hello' | |
194 | >>> f(b'hello:world?') |
|
196 | >>> f(b'hello:world?') | |
195 | 'hello~3aworld~3f' |
|
197 | 'hello~3aworld~3f' | |
196 | >>> f(b'the\\x07quick\\xADshot') |
|
198 | >>> f(b'the\\x07quick\\xADshot') | |
197 | 'the~07quick~adshot' |
|
199 | 'the~07quick~adshot' | |
198 | """ |
|
200 | """ | |
199 | xchr = pycompat.bytechr |
|
201 | xchr = pycompat.bytechr | |
200 | cmap = {xchr(x): xchr(x) for x in range(127)} |
|
202 | cmap = {xchr(x): xchr(x) for x in range(127)} | |
201 | for x in _reserved(): |
|
203 | for x in _reserved(): | |
202 | cmap[xchr(x)] = b"~%02x" % x |
|
204 | cmap[xchr(x)] = b"~%02x" % x | |
203 | for x in range(ord(b"A"), ord(b"Z") + 1): |
|
205 | for x in range(ord(b"A"), ord(b"Z") + 1): | |
204 | cmap[xchr(x)] = xchr(x).lower() |
|
206 | cmap[xchr(x)] = xchr(x).lower() | |
205 |
|
207 | |||
206 | def lowerencode(s): |
|
208 | def lowerencode(s): | |
207 | return b"".join([cmap[c] for c in pycompat.iterbytestr(s)]) |
|
209 | return b"".join([cmap[c] for c in pycompat.iterbytestr(s)]) | |
208 |
|
210 | |||
209 | return lowerencode |
|
211 | return lowerencode | |
210 |
|
212 | |||
211 |
|
213 | |||
212 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() |
|
214 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() | |
213 |
|
215 | |||
214 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 |
|
216 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 | |
215 | _winres3 = (b'aux', b'con', b'prn', b'nul') # length 3 |
|
217 | _winres3 = (b'aux', b'con', b'prn', b'nul') # length 3 | |
216 | _winres4 = (b'com', b'lpt') # length 4 (with trailing 1..9) |
|
218 | _winres4 = (b'com', b'lpt') # length 4 (with trailing 1..9) | |
217 |
|
219 | |||
218 |
|
220 | |||
219 | def _auxencode(path, dotencode): |
|
221 | def _auxencode(path, dotencode): | |
220 | """ |
|
222 | """ | |
221 | Encodes filenames containing names reserved by Windows or which end in |
|
223 | Encodes filenames containing names reserved by Windows or which end in | |
222 | period or space. Does not touch other single reserved characters c. |
|
224 | period or space. Does not touch other single reserved characters c. | |
223 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. |
|
225 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. | |
224 | Additionally encodes space or period at the beginning, if dotencode is |
|
226 | Additionally encodes space or period at the beginning, if dotencode is | |
225 | True. Parameter path is assumed to be all lowercase. |
|
227 | True. Parameter path is assumed to be all lowercase. | |
226 | A segment only needs encoding if a reserved name appears as a |
|
228 | A segment only needs encoding if a reserved name appears as a | |
227 | basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux" |
|
229 | basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux" | |
228 | doesn't need encoding. |
|
230 | doesn't need encoding. | |
229 |
|
231 | |||
230 | >>> s = b'.foo/aux.txt/txt.aux/con/prn/nul/foo.' |
|
232 | >>> s = b'.foo/aux.txt/txt.aux/con/prn/nul/foo.' | |
231 | >>> _auxencode(s.split(b'/'), True) |
|
233 | >>> _auxencode(s.split(b'/'), True) | |
232 | ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e'] |
|
234 | ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e'] | |
233 | >>> s = b'.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' |
|
235 | >>> s = b'.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' | |
234 | >>> _auxencode(s.split(b'/'), False) |
|
236 | >>> _auxencode(s.split(b'/'), False) | |
235 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] |
|
237 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] | |
236 | >>> _auxencode([b'foo. '], True) |
|
238 | >>> _auxencode([b'foo. '], True) | |
237 | ['foo.~20'] |
|
239 | ['foo.~20'] | |
238 | >>> _auxencode([b' .foo'], True) |
|
240 | >>> _auxencode([b' .foo'], True) | |
239 | ['~20.foo'] |
|
241 | ['~20.foo'] | |
240 | """ |
|
242 | """ | |
241 | for i, n in enumerate(path): |
|
243 | for i, n in enumerate(path): | |
242 | if not n: |
|
244 | if not n: | |
243 | continue |
|
245 | continue | |
244 | if dotencode and n[0] in b'. ': |
|
246 | if dotencode and n[0] in b'. ': | |
245 | n = b"~%02x" % ord(n[0:1]) + n[1:] |
|
247 | n = b"~%02x" % ord(n[0:1]) + n[1:] | |
246 | path[i] = n |
|
248 | path[i] = n | |
247 | else: |
|
249 | else: | |
248 | l = n.find(b'.') |
|
250 | l = n.find(b'.') | |
249 | if l == -1: |
|
251 | if l == -1: | |
250 | l = len(n) |
|
252 | l = len(n) | |
251 | if (l == 3 and n[:3] in _winres3) or ( |
|
253 | if (l == 3 and n[:3] in _winres3) or ( | |
252 | l == 4 |
|
254 | l == 4 | |
253 | and n[3:4] <= b'9' |
|
255 | and n[3:4] <= b'9' | |
254 | and n[3:4] >= b'1' |
|
256 | and n[3:4] >= b'1' | |
255 | and n[:3] in _winres4 |
|
257 | and n[:3] in _winres4 | |
256 | ): |
|
258 | ): | |
257 | # encode third letter ('aux' -> 'au~78') |
|
259 | # encode third letter ('aux' -> 'au~78') | |
258 | ec = b"~%02x" % ord(n[2:3]) |
|
260 | ec = b"~%02x" % ord(n[2:3]) | |
259 | n = n[0:2] + ec + n[3:] |
|
261 | n = n[0:2] + ec + n[3:] | |
260 | path[i] = n |
|
262 | path[i] = n | |
261 | if n[-1] in b'. ': |
|
263 | if n[-1] in b'. ': | |
262 | # encode last period or space ('foo...' -> 'foo..~2e') |
|
264 | # encode last period or space ('foo...' -> 'foo..~2e') | |
263 | path[i] = n[:-1] + b"~%02x" % ord(n[-1:]) |
|
265 | path[i] = n[:-1] + b"~%02x" % ord(n[-1:]) | |
264 | return path |
|
266 | return path | |
265 |
|
267 | |||
266 |
|
268 | |||
267 | _maxstorepathlen = 120 |
|
269 | _maxstorepathlen = 120 | |
268 | _dirprefixlen = 8 |
|
270 | _dirprefixlen = 8 | |
269 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 |
|
271 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 | |
270 |
|
272 | |||
271 |
|
273 | |||
272 | def _hashencode(path, dotencode): |
|
274 | def _hashencode(path, dotencode): | |
273 | digest = hex(hashutil.sha1(path).digest()) |
|
275 | digest = hex(hashutil.sha1(path).digest()) | |
274 | le = lowerencode(path[5:]).split(b'/') # skips prefix 'data/' or 'meta/' |
|
276 | le = lowerencode(path[5:]).split(b'/') # skips prefix 'data/' or 'meta/' | |
275 | parts = _auxencode(le, dotencode) |
|
277 | parts = _auxencode(le, dotencode) | |
276 | basename = parts[-1] |
|
278 | basename = parts[-1] | |
277 | _root, ext = os.path.splitext(basename) |
|
279 | _root, ext = os.path.splitext(basename) | |
278 | sdirs = [] |
|
280 | sdirs = [] | |
279 | sdirslen = 0 |
|
281 | sdirslen = 0 | |
280 | for p in parts[:-1]: |
|
282 | for p in parts[:-1]: | |
281 | d = p[:_dirprefixlen] |
|
283 | d = p[:_dirprefixlen] | |
282 | if d[-1] in b'. ': |
|
284 | if d[-1] in b'. ': | |
283 | # Windows can't access dirs ending in period or space |
|
285 | # Windows can't access dirs ending in period or space | |
284 | d = d[:-1] + b'_' |
|
286 | d = d[:-1] + b'_' | |
285 | if sdirslen == 0: |
|
287 | if sdirslen == 0: | |
286 | t = len(d) |
|
288 | t = len(d) | |
287 | else: |
|
289 | else: | |
288 | t = sdirslen + 1 + len(d) |
|
290 | t = sdirslen + 1 + len(d) | |
289 | if t > _maxshortdirslen: |
|
291 | if t > _maxshortdirslen: | |
290 | break |
|
292 | break | |
291 | sdirs.append(d) |
|
293 | sdirs.append(d) | |
292 | sdirslen = t |
|
294 | sdirslen = t | |
293 | dirs = b'/'.join(sdirs) |
|
295 | dirs = b'/'.join(sdirs) | |
294 | if len(dirs) > 0: |
|
296 | if len(dirs) > 0: | |
295 | dirs += b'/' |
|
297 | dirs += b'/' | |
296 | res = b'dh/' + dirs + digest + ext |
|
298 | res = b'dh/' + dirs + digest + ext | |
297 | spaceleft = _maxstorepathlen - len(res) |
|
299 | spaceleft = _maxstorepathlen - len(res) | |
298 | if spaceleft > 0: |
|
300 | if spaceleft > 0: | |
299 | filler = basename[:spaceleft] |
|
301 | filler = basename[:spaceleft] | |
300 | res = b'dh/' + dirs + filler + digest + ext |
|
302 | res = b'dh/' + dirs + filler + digest + ext | |
301 | return res |
|
303 | return res | |
302 |
|
304 | |||
303 |
|
305 | |||
304 | def _hybridencode(path, dotencode): |
|
306 | def _hybridencode(path, dotencode): | |
305 | """encodes path with a length limit |
|
307 | """encodes path with a length limit | |
306 |
|
308 | |||
307 | Encodes all paths that begin with 'data/', according to the following. |
|
309 | Encodes all paths that begin with 'data/', according to the following. | |
308 |
|
310 | |||
309 | Default encoding (reversible): |
|
311 | Default encoding (reversible): | |
310 |
|
312 | |||
311 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal |
|
313 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal | |
312 | characters are encoded as '~xx', where xx is the two digit hex code |
|
314 | characters are encoded as '~xx', where xx is the two digit hex code | |
313 | of the character (see encodefilename). |
|
315 | of the character (see encodefilename). | |
314 | Relevant path components consisting of Windows reserved filenames are |
|
316 | Relevant path components consisting of Windows reserved filenames are | |
315 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). |
|
317 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). | |
316 |
|
318 | |||
317 | Hashed encoding (not reversible): |
|
319 | Hashed encoding (not reversible): | |
318 |
|
320 | |||
319 | If the default-encoded path is longer than _maxstorepathlen, a |
|
321 | If the default-encoded path is longer than _maxstorepathlen, a | |
320 | non-reversible hybrid hashing of the path is done instead. |
|
322 | non-reversible hybrid hashing of the path is done instead. | |
321 | This encoding uses up to _dirprefixlen characters of all directory |
|
323 | This encoding uses up to _dirprefixlen characters of all directory | |
322 | levels of the lowerencoded path, but not more levels than can fit into |
|
324 | levels of the lowerencoded path, but not more levels than can fit into | |
323 | _maxshortdirslen. |
|
325 | _maxshortdirslen. | |
324 | Then follows the filler followed by the sha digest of the full path. |
|
326 | Then follows the filler followed by the sha digest of the full path. | |
325 | The filler is the beginning of the basename of the lowerencoded path |
|
327 | The filler is the beginning of the basename of the lowerencoded path | |
326 | (the basename is everything after the last path separator). The filler |
|
328 | (the basename is everything after the last path separator). The filler | |
327 | is as long as possible, filling in characters from the basename until |
|
329 | is as long as possible, filling in characters from the basename until | |
328 | the encoded path has _maxstorepathlen characters (or all chars of the |
|
330 | the encoded path has _maxstorepathlen characters (or all chars of the | |
329 | basename have been taken). |
|
331 | basename have been taken). | |
330 | The extension (e.g. '.i' or '.d') is preserved. |
|
332 | The extension (e.g. '.i' or '.d') is preserved. | |
331 |
|
333 | |||
332 | The string 'data/' at the beginning is replaced with 'dh/', if the hashed |
|
334 | The string 'data/' at the beginning is replaced with 'dh/', if the hashed | |
333 | encoding was used. |
|
335 | encoding was used. | |
334 | """ |
|
336 | """ | |
335 | path = encodedir(path) |
|
337 | path = encodedir(path) | |
336 | ef = _encodefname(path).split(b'/') |
|
338 | ef = _encodefname(path).split(b'/') | |
337 | res = b'/'.join(_auxencode(ef, dotencode)) |
|
339 | res = b'/'.join(_auxencode(ef, dotencode)) | |
338 | if len(res) > _maxstorepathlen: |
|
340 | if len(res) > _maxstorepathlen: | |
339 | res = _hashencode(path, dotencode) |
|
341 | res = _hashencode(path, dotencode) | |
340 | return res |
|
342 | return res | |
341 |
|
343 | |||
342 |
|
344 | |||
343 | def _pathencode(path): |
|
345 | def _pathencode(path): | |
344 | de = encodedir(path) |
|
346 | de = encodedir(path) | |
345 | if len(path) > _maxstorepathlen: |
|
347 | if len(path) > _maxstorepathlen: | |
346 | return _hashencode(de, True) |
|
348 | return _hashencode(de, True) | |
347 | ef = _encodefname(de).split(b'/') |
|
349 | ef = _encodefname(de).split(b'/') | |
348 | res = b'/'.join(_auxencode(ef, True)) |
|
350 | res = b'/'.join(_auxencode(ef, True)) | |
349 | if len(res) > _maxstorepathlen: |
|
351 | if len(res) > _maxstorepathlen: | |
350 | return _hashencode(de, True) |
|
352 | return _hashencode(de, True) | |
351 | return res |
|
353 | return res | |
352 |
|
354 | |||
353 |
|
355 | |||
354 | _pathencode = getattr(parsers, 'pathencode', _pathencode) |
|
356 | _pathencode = getattr(parsers, 'pathencode', _pathencode) | |
355 |
|
357 | |||
356 |
|
358 | |||
357 | def _plainhybridencode(f): |
|
359 | def _plainhybridencode(f): | |
358 | return _hybridencode(f, False) |
|
360 | return _hybridencode(f, False) | |
359 |
|
361 | |||
360 |
|
362 | |||
361 | def _calcmode(vfs): |
|
363 | def _calcmode(vfs): | |
362 | try: |
|
364 | try: | |
363 | # files in .hg/ will be created using this mode |
|
365 | # files in .hg/ will be created using this mode | |
364 | mode = vfs.stat().st_mode |
|
366 | mode = vfs.stat().st_mode | |
365 | # avoid some useless chmods |
|
367 | # avoid some useless chmods | |
366 | if (0o777 & ~util.umask) == (0o777 & mode): |
|
368 | if (0o777 & ~util.umask) == (0o777 & mode): | |
367 | mode = None |
|
369 | mode = None | |
368 | except OSError: |
|
370 | except OSError: | |
369 | mode = None |
|
371 | mode = None | |
370 | return mode |
|
372 | return mode | |
371 |
|
373 | |||
372 |
|
374 | |||
373 | _data = [ |
|
375 | _data = [ | |
374 | b'bookmarks', |
|
376 | b'bookmarks', | |
375 | b'narrowspec', |
|
377 | b'narrowspec', | |
376 | b'data', |
|
378 | b'data', | |
377 | b'meta', |
|
379 | b'meta', | |
378 | b'00manifest.d', |
|
380 | b'00manifest.d', | |
379 | b'00manifest.i', |
|
381 | b'00manifest.i', | |
380 | b'00changelog.d', |
|
382 | b'00changelog.d', | |
381 | b'00changelog.i', |
|
383 | b'00changelog.i', | |
382 | b'phaseroots', |
|
384 | b'phaseroots', | |
383 | b'obsstore', |
|
385 | b'obsstore', | |
384 | b'requires', |
|
386 | b'requires', | |
385 | ] |
|
387 | ] | |
386 |
|
388 | |||
387 | REVLOG_FILES_MAIN_EXT = (b'.i',) |
|
389 | REVLOG_FILES_MAIN_EXT = (b'.i',) | |
388 | REVLOG_FILES_OTHER_EXT = ( |
|
390 | REVLOG_FILES_OTHER_EXT = ( | |
389 | b'.idx', |
|
391 | b'.idx', | |
390 | b'.d', |
|
392 | b'.d', | |
391 | b'.dat', |
|
393 | b'.dat', | |
392 | b'.n', |
|
394 | b'.n', | |
393 | b'.nd', |
|
395 | b'.nd', | |
394 | b'.sda', |
|
396 | b'.sda', | |
395 | ) |
|
397 | ) | |
396 | # files that are "volatile" and might change between listing and streaming |
|
398 | # files that are "volatile" and might change between listing and streaming | |
397 | # |
|
399 | # | |
398 | # note: the ".nd" file are nodemap data and won't "change" but they might be |
|
400 | # note: the ".nd" file are nodemap data and won't "change" but they might be | |
399 | # deleted. |
|
401 | # deleted. | |
400 | REVLOG_FILES_VOLATILE_EXT = (b'.n', b'.nd') |
|
402 | REVLOG_FILES_VOLATILE_EXT = (b'.n', b'.nd') | |
401 |
|
403 | |||
402 | # some exception to the above matching |
|
404 | # some exception to the above matching | |
403 | # |
|
405 | # | |
404 | # XXX This is currently not in use because of issue6542 |
|
406 | # XXX This is currently not in use because of issue6542 | |
405 | EXCLUDED = re.compile(br'.*undo\.[^/]+\.(nd?|i)$') |
|
407 | EXCLUDED = re.compile(br'.*undo\.[^/]+\.(nd?|i)$') | |
406 |
|
408 | |||
407 |
|
409 | |||
408 | def is_revlog(f, kind, st): |
|
410 | def is_revlog(f, kind, st): | |
409 | if kind != stat.S_IFREG: |
|
411 | if kind != stat.S_IFREG: | |
410 | return None |
|
412 | return None | |
411 | return revlog_type(f) |
|
413 | return revlog_type(f) | |
412 |
|
414 | |||
413 |
|
415 | |||
414 | def revlog_type(f): |
|
416 | def revlog_type(f): | |
415 | # XXX we need to filter `undo.` created by the transaction here, however |
|
417 | # XXX we need to filter `undo.` created by the transaction here, however | |
416 | # being naive about it also filter revlog for `undo.*` files, leading to |
|
418 | # being naive about it also filter revlog for `undo.*` files, leading to | |
417 | # issue6542. So we no longer use EXCLUDED. |
|
419 | # issue6542. So we no longer use EXCLUDED. | |
418 | if f.endswith(REVLOG_FILES_MAIN_EXT): |
|
420 | if f.endswith(REVLOG_FILES_MAIN_EXT): | |
419 | return FILEFLAGS_REVLOG_MAIN |
|
421 | return FILEFLAGS_REVLOG_MAIN | |
420 | elif f.endswith(REVLOG_FILES_OTHER_EXT): |
|
422 | elif f.endswith(REVLOG_FILES_OTHER_EXT): | |
421 | t = FILETYPE_FILELOG_OTHER |
|
423 | t = FILETYPE_FILELOG_OTHER | |
422 | if f.endswith(REVLOG_FILES_VOLATILE_EXT): |
|
424 | if f.endswith(REVLOG_FILES_VOLATILE_EXT): | |
423 | t |= FILEFLAGS_VOLATILE |
|
425 | t |= FILEFLAGS_VOLATILE | |
424 | return t |
|
426 | return t | |
425 | return None |
|
427 | return None | |
426 |
|
428 | |||
427 |
|
429 | |||
428 | # the file is part of changelog data |
|
430 | # the file is part of changelog data | |
429 | FILEFLAGS_CHANGELOG = 1 << 13 |
|
431 | FILEFLAGS_CHANGELOG = 1 << 13 | |
430 | # the file is part of manifest data |
|
432 | # the file is part of manifest data | |
431 | FILEFLAGS_MANIFESTLOG = 1 << 12 |
|
433 | FILEFLAGS_MANIFESTLOG = 1 << 12 | |
432 | # the file is part of filelog data |
|
434 | # the file is part of filelog data | |
433 | FILEFLAGS_FILELOG = 1 << 11 |
|
435 | FILEFLAGS_FILELOG = 1 << 11 | |
434 | # file that are not directly part of a revlog |
|
436 | # file that are not directly part of a revlog | |
435 | FILEFLAGS_OTHER = 1 << 10 |
|
437 | FILEFLAGS_OTHER = 1 << 10 | |
436 |
|
438 | |||
437 | # the main entry point for a revlog |
|
439 | # the main entry point for a revlog | |
438 | FILEFLAGS_REVLOG_MAIN = 1 << 1 |
|
440 | FILEFLAGS_REVLOG_MAIN = 1 << 1 | |
439 | # a secondary file for a revlog |
|
441 | # a secondary file for a revlog | |
440 | FILEFLAGS_REVLOG_OTHER = 1 << 0 |
|
442 | FILEFLAGS_REVLOG_OTHER = 1 << 0 | |
441 |
|
443 | |||
442 | # files that are "volatile" and might change between listing and streaming |
|
444 | # files that are "volatile" and might change between listing and streaming | |
443 | FILEFLAGS_VOLATILE = 1 << 20 |
|
445 | FILEFLAGS_VOLATILE = 1 << 20 | |
444 |
|
446 | |||
445 | FILETYPE_CHANGELOG_MAIN = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_MAIN |
|
447 | FILETYPE_CHANGELOG_MAIN = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_MAIN | |
446 | FILETYPE_CHANGELOG_OTHER = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_OTHER |
|
448 | FILETYPE_CHANGELOG_OTHER = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_OTHER | |
447 | FILETYPE_MANIFESTLOG_MAIN = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_MAIN |
|
449 | FILETYPE_MANIFESTLOG_MAIN = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_MAIN | |
448 | FILETYPE_MANIFESTLOG_OTHER = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_OTHER |
|
450 | FILETYPE_MANIFESTLOG_OTHER = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_OTHER | |
449 | FILETYPE_FILELOG_MAIN = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_MAIN |
|
451 | FILETYPE_FILELOG_MAIN = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_MAIN | |
450 | FILETYPE_FILELOG_OTHER = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_OTHER |
|
452 | FILETYPE_FILELOG_OTHER = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_OTHER | |
451 | FILETYPE_OTHER = FILEFLAGS_OTHER |
|
453 | FILETYPE_OTHER = FILEFLAGS_OTHER | |
452 |
|
454 | |||
453 |
|
455 | |||
|
456 | @attr.s(slots=True) | |||
|
457 | class StoreEntry: | |||
|
458 | """An entry in the store | |||
|
459 | ||||
|
460 | This is returned by `store.walk` and represent some data in the store.""" | |||
|
461 | ||||
|
462 | unencoded_path = attr.ib() | |||
|
463 | is_revlog = attr.ib(default=False) | |||
|
464 | revlog_type = attr.ib(default=None) | |||
|
465 | is_revlog_main = attr.ib(default=None) | |||
|
466 | is_volatile = attr.ib(default=False) | |||
|
467 | file_size = attr.ib(default=None) | |||
|
468 | ||||
|
469 | ||||
454 | class basicstore: |
|
470 | class basicstore: | |
455 | '''base class for local repository stores''' |
|
471 | '''base class for local repository stores''' | |
456 |
|
472 | |||
457 | def __init__(self, path, vfstype): |
|
473 | def __init__(self, path, vfstype): | |
458 | vfs = vfstype(path) |
|
474 | vfs = vfstype(path) | |
459 | self.path = vfs.base |
|
475 | self.path = vfs.base | |
460 | self.createmode = _calcmode(vfs) |
|
476 | self.createmode = _calcmode(vfs) | |
461 | vfs.createmode = self.createmode |
|
477 | vfs.createmode = self.createmode | |
462 | self.rawvfs = vfs |
|
478 | self.rawvfs = vfs | |
463 | self.vfs = vfsmod.filtervfs(vfs, encodedir) |
|
479 | self.vfs = vfsmod.filtervfs(vfs, encodedir) | |
464 | self.opener = self.vfs |
|
480 | self.opener = self.vfs | |
465 |
|
481 | |||
466 | def join(self, f): |
|
482 | def join(self, f): | |
467 | return self.path + b'/' + encodedir(f) |
|
483 | return self.path + b'/' + encodedir(f) | |
468 |
|
484 | |||
469 | def _walk(self, relpath, recurse): |
|
485 | def _walk(self, relpath, recurse): | |
470 | '''yields (revlog_type, unencoded, size)''' |
|
486 | '''yields (revlog_type, unencoded, size)''' | |
471 | path = self.path |
|
487 | path = self.path | |
472 | if relpath: |
|
488 | if relpath: | |
473 | path += b'/' + relpath |
|
489 | path += b'/' + relpath | |
474 | striplen = len(self.path) + 1 |
|
490 | striplen = len(self.path) + 1 | |
475 | l = [] |
|
491 | l = [] | |
476 | if self.rawvfs.isdir(path): |
|
492 | if self.rawvfs.isdir(path): | |
477 | visit = [path] |
|
493 | visit = [path] | |
478 | readdir = self.rawvfs.readdir |
|
494 | readdir = self.rawvfs.readdir | |
479 | while visit: |
|
495 | while visit: | |
480 | p = visit.pop() |
|
496 | p = visit.pop() | |
481 | for f, kind, st in readdir(p, stat=True): |
|
497 | for f, kind, st in readdir(p, stat=True): | |
482 | fp = p + b'/' + f |
|
498 | fp = p + b'/' + f | |
483 | rl_type = is_revlog(f, kind, st) |
|
499 | rl_type = is_revlog(f, kind, st) | |
484 | if rl_type is not None: |
|
500 | if rl_type is not None: | |
485 | n = util.pconvert(fp[striplen:]) |
|
501 | n = util.pconvert(fp[striplen:]) | |
486 | l.append((rl_type, decodedir(n), st.st_size)) |
|
502 | l.append((rl_type, decodedir(n), st.st_size)) | |
487 | elif kind == stat.S_IFDIR and recurse: |
|
503 | elif kind == stat.S_IFDIR and recurse: | |
488 | visit.append(fp) |
|
504 | visit.append(fp) | |
489 | l.sort() |
|
505 | l.sort() | |
490 | return l |
|
506 | return l | |
491 |
|
507 | |||
492 | def changelog(self, trypending, concurrencychecker=None): |
|
508 | def changelog(self, trypending, concurrencychecker=None): | |
493 | return changelog.changelog( |
|
509 | return changelog.changelog( | |
494 | self.vfs, |
|
510 | self.vfs, | |
495 | trypending=trypending, |
|
511 | trypending=trypending, | |
496 | concurrencychecker=concurrencychecker, |
|
512 | concurrencychecker=concurrencychecker, | |
497 | ) |
|
513 | ) | |
498 |
|
514 | |||
499 | def manifestlog(self, repo, storenarrowmatch): |
|
515 | def manifestlog(self, repo, storenarrowmatch): | |
500 | rootstore = manifest.manifestrevlog(repo.nodeconstants, self.vfs) |
|
516 | rootstore = manifest.manifestrevlog(repo.nodeconstants, self.vfs) | |
501 | return manifest.manifestlog(self.vfs, repo, rootstore, storenarrowmatch) |
|
517 | return manifest.manifestlog(self.vfs, repo, rootstore, storenarrowmatch) | |
502 |
|
518 | |||
503 | def datafiles(self, matcher=None, undecodable=None): |
|
519 | def datafiles( | |
|
520 | self, matcher=None, undecodable=None | |||
|
521 | ) -> Generator[StoreEntry, None, None]: | |||
504 | """Like walk, but excluding the changelog and root manifest. |
|
522 | """Like walk, but excluding the changelog and root manifest. | |
505 |
|
523 | |||
506 | When [undecodable] is None, revlogs names that can't be |
|
524 | When [undecodable] is None, revlogs names that can't be | |
507 | decoded cause an exception. When it is provided, it should |
|
525 | decoded cause an exception. When it is provided, it should | |
508 | be a list and the filenames that can't be decoded are added |
|
526 | be a list and the filenames that can't be decoded are added | |
509 | to it instead. This is very rarely needed.""" |
|
527 | to it instead. This is very rarely needed.""" | |
510 | files = self._walk(b'data', True) + self._walk(b'meta', True) |
|
528 | files = self._walk(b'data', True) + self._walk(b'meta', True) | |
511 | for (t, u, s) in files: |
|
529 | for (t, u, s) in files: | |
512 | if t is not None: |
|
530 | if t is not None: | |
513 | yield (FILEFLAGS_FILELOG | t, u, s) |
|
531 | yield StoreEntry( | |
|
532 | unencoded_path=u, | |||
|
533 | is_revlog=True, | |||
|
534 | revlog_type=FILEFLAGS_FILELOG, | |||
|
535 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), | |||
|
536 | is_volatile=bool(t & FILEFLAGS_VOLATILE), | |||
|
537 | file_size=s, | |||
|
538 | ) | |||
514 |
|
539 | |||
515 | def topfiles(self): |
|
540 | def topfiles(self) -> Generator[StoreEntry, None, None]: | |
516 | # yield manifest before changelog |
|
541 | # yield manifest before changelog | |
517 | files = reversed(self._walk(b'', False)) |
|
542 | files = reversed(self._walk(b'', False)) | |
518 | for (t, u, s) in files: |
|
543 | for (t, u, s) in files: | |
519 | if u.startswith(b'00changelog'): |
|
544 | if u.startswith(b'00changelog'): | |
520 |
|
|
545 | revlog_type = FILEFLAGS_CHANGELOG | |
521 | elif u.startswith(b'00manifest'): |
|
546 | elif u.startswith(b'00manifest'): | |
522 |
|
|
547 | revlog_type = FILEFLAGS_MANIFESTLOG | |
523 | else: |
|
548 | else: | |
524 | yield (FILETYPE_OTHER | t, u, s) |
|
549 | revlog_type = None | |
|
550 | yield StoreEntry( | |||
|
551 | unencoded_path=u, | |||
|
552 | is_revlog=revlog_type is not None, | |||
|
553 | revlog_type=revlog_type, | |||
|
554 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), | |||
|
555 | is_volatile=bool(t & FILEFLAGS_VOLATILE), | |||
|
556 | file_size=s, | |||
|
557 | ) | |||
525 |
|
558 | |||
526 | def walk(self, matcher=None): |
|
559 | def walk(self, matcher=None) -> Generator[StoreEntry, None, None]: | |
527 | """return files related to data storage (ie: revlogs) |
|
560 | """return files related to data storage (ie: revlogs) | |
528 |
|
561 | |||
529 | yields (file_type, unencoded, size) |
|
562 | yields (file_type, unencoded, size) | |
530 |
|
563 | |||
531 | if a matcher is passed, storage files of only those tracked paths |
|
564 | if a matcher is passed, storage files of only those tracked paths | |
532 | are passed with matches the matcher |
|
565 | are passed with matches the matcher | |
533 | """ |
|
566 | """ | |
534 | # yield data files first |
|
567 | # yield data files first | |
535 | for x in self.datafiles(matcher): |
|
568 | for x in self.datafiles(matcher): | |
536 | yield x |
|
569 | yield x | |
537 | for x in self.topfiles(): |
|
570 | for x in self.topfiles(): | |
538 | yield x |
|
571 | yield x | |
539 |
|
572 | |||
540 | def copylist(self): |
|
573 | def copylist(self): | |
541 | return _data |
|
574 | return _data | |
542 |
|
575 | |||
543 | def write(self, tr): |
|
576 | def write(self, tr): | |
544 | pass |
|
577 | pass | |
545 |
|
578 | |||
546 | def invalidatecaches(self): |
|
579 | def invalidatecaches(self): | |
547 | pass |
|
580 | pass | |
548 |
|
581 | |||
549 | def markremoved(self, fn): |
|
582 | def markremoved(self, fn): | |
550 | pass |
|
583 | pass | |
551 |
|
584 | |||
552 | def __contains__(self, path): |
|
585 | def __contains__(self, path): | |
553 | '''Checks if the store contains path''' |
|
586 | '''Checks if the store contains path''' | |
554 | path = b"/".join((b"data", path)) |
|
587 | path = b"/".join((b"data", path)) | |
555 | # file? |
|
588 | # file? | |
556 | if self.vfs.exists(path + b".i"): |
|
589 | if self.vfs.exists(path + b".i"): | |
557 | return True |
|
590 | return True | |
558 | # dir? |
|
591 | # dir? | |
559 | if not path.endswith(b"/"): |
|
592 | if not path.endswith(b"/"): | |
560 | path = path + b"/" |
|
593 | path = path + b"/" | |
561 | return self.vfs.exists(path) |
|
594 | return self.vfs.exists(path) | |
562 |
|
595 | |||
563 |
|
596 | |||
564 | class encodedstore(basicstore): |
|
597 | class encodedstore(basicstore): | |
565 | def __init__(self, path, vfstype): |
|
598 | def __init__(self, path, vfstype): | |
566 | vfs = vfstype(path + b'/store') |
|
599 | vfs = vfstype(path + b'/store') | |
567 | self.path = vfs.base |
|
600 | self.path = vfs.base | |
568 | self.createmode = _calcmode(vfs) |
|
601 | self.createmode = _calcmode(vfs) | |
569 | vfs.createmode = self.createmode |
|
602 | vfs.createmode = self.createmode | |
570 | self.rawvfs = vfs |
|
603 | self.rawvfs = vfs | |
571 | self.vfs = vfsmod.filtervfs(vfs, encodefilename) |
|
604 | self.vfs = vfsmod.filtervfs(vfs, encodefilename) | |
572 | self.opener = self.vfs |
|
605 | self.opener = self.vfs | |
573 |
|
606 | |||
574 | # note: topfiles would also need a decode phase. It is just that in |
|
607 | # note: topfiles would also need a decode phase. It is just that in | |
575 | # practice we do not have any file outside of `data/` that needs encoding. |
|
608 | # practice we do not have any file outside of `data/` that needs encoding. | |
576 | # However that might change so we should probably add a test and encoding |
|
609 | # However that might change so we should probably add a test and encoding | |
577 | # decoding for it too. see issue6548 |
|
610 | # decoding for it too. see issue6548 | |
578 |
|
611 | |||
579 | def datafiles(self, matcher=None, undecodable=None): |
|
612 | def datafiles( | |
580 | for t, f1, size in super(encodedstore, self).datafiles(): |
|
613 | self, matcher=None, undecodable=None | |
|
614 | ) -> Generator[StoreEntry, None, None]: | |||
|
615 | for entry in super(encodedstore, self).datafiles(): | |||
581 | try: |
|
616 | try: | |
|
617 | f1 = entry.unencoded_path | |||
582 | f2 = decodefilename(f1) |
|
618 | f2 = decodefilename(f1) | |
583 | except KeyError: |
|
619 | except KeyError: | |
584 | if undecodable is None: |
|
620 | if undecodable is None: | |
585 | msg = _(b'undecodable revlog name %s') % f1 |
|
621 | msg = _(b'undecodable revlog name %s') % f1 | |
586 | raise error.StorageError(msg) |
|
622 | raise error.StorageError(msg) | |
587 | else: |
|
623 | else: | |
588 | undecodable.append(f1) |
|
624 | undecodable.append(f1) | |
589 | continue |
|
625 | continue | |
590 | if not _matchtrackedpath(f2, matcher): |
|
626 | if not _matchtrackedpath(f2, matcher): | |
591 | continue |
|
627 | continue | |
592 | yield t, f2, size |
|
628 | entry.unencoded_path = f2 | |
|
629 | yield entry | |||
593 |
|
630 | |||
594 | def join(self, f): |
|
631 | def join(self, f): | |
595 | return self.path + b'/' + encodefilename(f) |
|
632 | return self.path + b'/' + encodefilename(f) | |
596 |
|
633 | |||
597 | def copylist(self): |
|
634 | def copylist(self): | |
598 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in _data] |
|
635 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in _data] | |
599 |
|
636 | |||
600 |
|
637 | |||
601 | class fncache: |
|
638 | class fncache: | |
602 | # the filename used to be partially encoded |
|
639 | # the filename used to be partially encoded | |
603 | # hence the encodedir/decodedir dance |
|
640 | # hence the encodedir/decodedir dance | |
604 | def __init__(self, vfs): |
|
641 | def __init__(self, vfs): | |
605 | self.vfs = vfs |
|
642 | self.vfs = vfs | |
606 | self._ignores = set() |
|
643 | self._ignores = set() | |
607 | self.entries = None |
|
644 | self.entries = None | |
608 | self._dirty = False |
|
645 | self._dirty = False | |
609 | # set of new additions to fncache |
|
646 | # set of new additions to fncache | |
610 | self.addls = set() |
|
647 | self.addls = set() | |
611 |
|
648 | |||
612 | def ensureloaded(self, warn=None): |
|
649 | def ensureloaded(self, warn=None): | |
613 | """read the fncache file if not already read. |
|
650 | """read the fncache file if not already read. | |
614 |
|
651 | |||
615 | If the file on disk is corrupted, raise. If warn is provided, |
|
652 | If the file on disk is corrupted, raise. If warn is provided, | |
616 | warn and keep going instead.""" |
|
653 | warn and keep going instead.""" | |
617 | if self.entries is None: |
|
654 | if self.entries is None: | |
618 | self._load(warn) |
|
655 | self._load(warn) | |
619 |
|
656 | |||
620 | def _load(self, warn=None): |
|
657 | def _load(self, warn=None): | |
621 | '''fill the entries from the fncache file''' |
|
658 | '''fill the entries from the fncache file''' | |
622 | self._dirty = False |
|
659 | self._dirty = False | |
623 | try: |
|
660 | try: | |
624 | fp = self.vfs(b'fncache', mode=b'rb') |
|
661 | fp = self.vfs(b'fncache', mode=b'rb') | |
625 | except IOError: |
|
662 | except IOError: | |
626 | # skip nonexistent file |
|
663 | # skip nonexistent file | |
627 | self.entries = set() |
|
664 | self.entries = set() | |
628 | return |
|
665 | return | |
629 |
|
666 | |||
630 | self.entries = set() |
|
667 | self.entries = set() | |
631 | chunk = b'' |
|
668 | chunk = b'' | |
632 | for c in iter(functools.partial(fp.read, fncache_chunksize), b''): |
|
669 | for c in iter(functools.partial(fp.read, fncache_chunksize), b''): | |
633 | chunk += c |
|
670 | chunk += c | |
634 | try: |
|
671 | try: | |
635 | p = chunk.rindex(b'\n') |
|
672 | p = chunk.rindex(b'\n') | |
636 | self.entries.update(decodedir(chunk[: p + 1]).splitlines()) |
|
673 | self.entries.update(decodedir(chunk[: p + 1]).splitlines()) | |
637 | chunk = chunk[p + 1 :] |
|
674 | chunk = chunk[p + 1 :] | |
638 | except ValueError: |
|
675 | except ValueError: | |
639 | # substring '\n' not found, maybe the entry is bigger than the |
|
676 | # substring '\n' not found, maybe the entry is bigger than the | |
640 | # chunksize, so let's keep iterating |
|
677 | # chunksize, so let's keep iterating | |
641 | pass |
|
678 | pass | |
642 |
|
679 | |||
643 | if chunk: |
|
680 | if chunk: | |
644 | msg = _(b"fncache does not ends with a newline") |
|
681 | msg = _(b"fncache does not ends with a newline") | |
645 | if warn: |
|
682 | if warn: | |
646 | warn(msg + b'\n') |
|
683 | warn(msg + b'\n') | |
647 | else: |
|
684 | else: | |
648 | raise error.Abort( |
|
685 | raise error.Abort( | |
649 | msg, |
|
686 | msg, | |
650 | hint=_( |
|
687 | hint=_( | |
651 | b"use 'hg debugrebuildfncache' to " |
|
688 | b"use 'hg debugrebuildfncache' to " | |
652 | b"rebuild the fncache" |
|
689 | b"rebuild the fncache" | |
653 | ), |
|
690 | ), | |
654 | ) |
|
691 | ) | |
655 | self._checkentries(fp, warn) |
|
692 | self._checkentries(fp, warn) | |
656 | fp.close() |
|
693 | fp.close() | |
657 |
|
694 | |||
658 | def _checkentries(self, fp, warn): |
|
695 | def _checkentries(self, fp, warn): | |
659 | """make sure there is no empty string in entries""" |
|
696 | """make sure there is no empty string in entries""" | |
660 | if b'' in self.entries: |
|
697 | if b'' in self.entries: | |
661 | fp.seek(0) |
|
698 | fp.seek(0) | |
662 | for n, line in enumerate(fp): |
|
699 | for n, line in enumerate(fp): | |
663 | if not line.rstrip(b'\n'): |
|
700 | if not line.rstrip(b'\n'): | |
664 | t = _(b'invalid entry in fncache, line %d') % (n + 1) |
|
701 | t = _(b'invalid entry in fncache, line %d') % (n + 1) | |
665 | if warn: |
|
702 | if warn: | |
666 | warn(t + b'\n') |
|
703 | warn(t + b'\n') | |
667 | else: |
|
704 | else: | |
668 | raise error.Abort(t) |
|
705 | raise error.Abort(t) | |
669 |
|
706 | |||
670 | def write(self, tr): |
|
707 | def write(self, tr): | |
671 | if self._dirty: |
|
708 | if self._dirty: | |
672 | assert self.entries is not None |
|
709 | assert self.entries is not None | |
673 | self.entries = self.entries | self.addls |
|
710 | self.entries = self.entries | self.addls | |
674 | self.addls = set() |
|
711 | self.addls = set() | |
675 | tr.addbackup(b'fncache') |
|
712 | tr.addbackup(b'fncache') | |
676 | fp = self.vfs(b'fncache', mode=b'wb', atomictemp=True) |
|
713 | fp = self.vfs(b'fncache', mode=b'wb', atomictemp=True) | |
677 | if self.entries: |
|
714 | if self.entries: | |
678 | fp.write(encodedir(b'\n'.join(self.entries) + b'\n')) |
|
715 | fp.write(encodedir(b'\n'.join(self.entries) + b'\n')) | |
679 | fp.close() |
|
716 | fp.close() | |
680 | self._dirty = False |
|
717 | self._dirty = False | |
681 | if self.addls: |
|
718 | if self.addls: | |
682 | # if we have just new entries, let's append them to the fncache |
|
719 | # if we have just new entries, let's append them to the fncache | |
683 | tr.addbackup(b'fncache') |
|
720 | tr.addbackup(b'fncache') | |
684 | fp = self.vfs(b'fncache', mode=b'ab', atomictemp=True) |
|
721 | fp = self.vfs(b'fncache', mode=b'ab', atomictemp=True) | |
685 | if self.addls: |
|
722 | if self.addls: | |
686 | fp.write(encodedir(b'\n'.join(self.addls) + b'\n')) |
|
723 | fp.write(encodedir(b'\n'.join(self.addls) + b'\n')) | |
687 | fp.close() |
|
724 | fp.close() | |
688 | self.entries = None |
|
725 | self.entries = None | |
689 | self.addls = set() |
|
726 | self.addls = set() | |
690 |
|
727 | |||
691 | def addignore(self, fn): |
|
728 | def addignore(self, fn): | |
692 | self._ignores.add(fn) |
|
729 | self._ignores.add(fn) | |
693 |
|
730 | |||
694 | def add(self, fn): |
|
731 | def add(self, fn): | |
695 | if fn in self._ignores: |
|
732 | if fn in self._ignores: | |
696 | return |
|
733 | return | |
697 | if self.entries is None: |
|
734 | if self.entries is None: | |
698 | self._load() |
|
735 | self._load() | |
699 | if fn not in self.entries: |
|
736 | if fn not in self.entries: | |
700 | self.addls.add(fn) |
|
737 | self.addls.add(fn) | |
701 |
|
738 | |||
702 | def remove(self, fn): |
|
739 | def remove(self, fn): | |
703 | if self.entries is None: |
|
740 | if self.entries is None: | |
704 | self._load() |
|
741 | self._load() | |
705 | if fn in self.addls: |
|
742 | if fn in self.addls: | |
706 | self.addls.remove(fn) |
|
743 | self.addls.remove(fn) | |
707 | return |
|
744 | return | |
708 | try: |
|
745 | try: | |
709 | self.entries.remove(fn) |
|
746 | self.entries.remove(fn) | |
710 | self._dirty = True |
|
747 | self._dirty = True | |
711 | except KeyError: |
|
748 | except KeyError: | |
712 | pass |
|
749 | pass | |
713 |
|
750 | |||
714 | def __contains__(self, fn): |
|
751 | def __contains__(self, fn): | |
715 | if fn in self.addls: |
|
752 | if fn in self.addls: | |
716 | return True |
|
753 | return True | |
717 | if self.entries is None: |
|
754 | if self.entries is None: | |
718 | self._load() |
|
755 | self._load() | |
719 | return fn in self.entries |
|
756 | return fn in self.entries | |
720 |
|
757 | |||
721 | def __iter__(self): |
|
758 | def __iter__(self): | |
722 | if self.entries is None: |
|
759 | if self.entries is None: | |
723 | self._load() |
|
760 | self._load() | |
724 | return iter(self.entries | self.addls) |
|
761 | return iter(self.entries | self.addls) | |
725 |
|
762 | |||
726 |
|
763 | |||
727 | class _fncachevfs(vfsmod.proxyvfs): |
|
764 | class _fncachevfs(vfsmod.proxyvfs): | |
728 | def __init__(self, vfs, fnc, encode): |
|
765 | def __init__(self, vfs, fnc, encode): | |
729 | vfsmod.proxyvfs.__init__(self, vfs) |
|
766 | vfsmod.proxyvfs.__init__(self, vfs) | |
730 | self.fncache = fnc |
|
767 | self.fncache = fnc | |
731 | self.encode = encode |
|
768 | self.encode = encode | |
732 |
|
769 | |||
733 | def __call__(self, path, mode=b'r', *args, **kw): |
|
770 | def __call__(self, path, mode=b'r', *args, **kw): | |
734 | encoded = self.encode(path) |
|
771 | encoded = self.encode(path) | |
735 | if ( |
|
772 | if ( | |
736 | mode not in (b'r', b'rb') |
|
773 | mode not in (b'r', b'rb') | |
737 | and (path.startswith(b'data/') or path.startswith(b'meta/')) |
|
774 | and (path.startswith(b'data/') or path.startswith(b'meta/')) | |
738 | and revlog_type(path) is not None |
|
775 | and revlog_type(path) is not None | |
739 | ): |
|
776 | ): | |
740 | # do not trigger a fncache load when adding a file that already is |
|
777 | # do not trigger a fncache load when adding a file that already is | |
741 | # known to exist. |
|
778 | # known to exist. | |
742 | notload = self.fncache.entries is None and self.vfs.exists(encoded) |
|
779 | notload = self.fncache.entries is None and self.vfs.exists(encoded) | |
743 | if notload and b'r+' in mode and not self.vfs.stat(encoded).st_size: |
|
780 | if notload and b'r+' in mode and not self.vfs.stat(encoded).st_size: | |
744 | # when appending to an existing file, if the file has size zero, |
|
781 | # when appending to an existing file, if the file has size zero, | |
745 | # it should be considered as missing. Such zero-size files are |
|
782 | # it should be considered as missing. Such zero-size files are | |
746 | # the result of truncation when a transaction is aborted. |
|
783 | # the result of truncation when a transaction is aborted. | |
747 | notload = False |
|
784 | notload = False | |
748 | if not notload: |
|
785 | if not notload: | |
749 | self.fncache.add(path) |
|
786 | self.fncache.add(path) | |
750 | return self.vfs(encoded, mode, *args, **kw) |
|
787 | return self.vfs(encoded, mode, *args, **kw) | |
751 |
|
788 | |||
752 | def join(self, path): |
|
789 | def join(self, path): | |
753 | if path: |
|
790 | if path: | |
754 | return self.vfs.join(self.encode(path)) |
|
791 | return self.vfs.join(self.encode(path)) | |
755 | else: |
|
792 | else: | |
756 | return self.vfs.join(path) |
|
793 | return self.vfs.join(path) | |
757 |
|
794 | |||
758 | def register_file(self, path): |
|
795 | def register_file(self, path): | |
759 | """generic hook point to lets fncache steer its stew""" |
|
796 | """generic hook point to lets fncache steer its stew""" | |
760 | if path.startswith(b'data/') or path.startswith(b'meta/'): |
|
797 | if path.startswith(b'data/') or path.startswith(b'meta/'): | |
761 | self.fncache.add(path) |
|
798 | self.fncache.add(path) | |
762 |
|
799 | |||
763 |
|
800 | |||
764 | class fncachestore(basicstore): |
|
801 | class fncachestore(basicstore): | |
765 | def __init__(self, path, vfstype, dotencode): |
|
802 | def __init__(self, path, vfstype, dotencode): | |
766 | if dotencode: |
|
803 | if dotencode: | |
767 | encode = _pathencode |
|
804 | encode = _pathencode | |
768 | else: |
|
805 | else: | |
769 | encode = _plainhybridencode |
|
806 | encode = _plainhybridencode | |
770 | self.encode = encode |
|
807 | self.encode = encode | |
771 | vfs = vfstype(path + b'/store') |
|
808 | vfs = vfstype(path + b'/store') | |
772 | self.path = vfs.base |
|
809 | self.path = vfs.base | |
773 | self.pathsep = self.path + b'/' |
|
810 | self.pathsep = self.path + b'/' | |
774 | self.createmode = _calcmode(vfs) |
|
811 | self.createmode = _calcmode(vfs) | |
775 | vfs.createmode = self.createmode |
|
812 | vfs.createmode = self.createmode | |
776 | self.rawvfs = vfs |
|
813 | self.rawvfs = vfs | |
777 | fnc = fncache(vfs) |
|
814 | fnc = fncache(vfs) | |
778 | self.fncache = fnc |
|
815 | self.fncache = fnc | |
779 | self.vfs = _fncachevfs(vfs, fnc, encode) |
|
816 | self.vfs = _fncachevfs(vfs, fnc, encode) | |
780 | self.opener = self.vfs |
|
817 | self.opener = self.vfs | |
781 |
|
818 | |||
782 | def join(self, f): |
|
819 | def join(self, f): | |
783 | return self.pathsep + self.encode(f) |
|
820 | return self.pathsep + self.encode(f) | |
784 |
|
821 | |||
785 | def getsize(self, path): |
|
822 | def getsize(self, path): | |
786 | return self.rawvfs.stat(path).st_size |
|
823 | return self.rawvfs.stat(path).st_size | |
787 |
|
824 | |||
788 | def datafiles(self, matcher=None, undecodable=None): |
|
825 | def datafiles( | |
|
826 | self, matcher=None, undecodable=None | |||
|
827 | ) -> Generator[StoreEntry, None, None]: | |||
789 | for f in sorted(self.fncache): |
|
828 | for f in sorted(self.fncache): | |
790 | if not _matchtrackedpath(f, matcher): |
|
829 | if not _matchtrackedpath(f, matcher): | |
791 | continue |
|
830 | continue | |
792 | ef = self.encode(f) |
|
831 | ef = self.encode(f) | |
793 | t = revlog_type(f) |
|
832 | t = revlog_type(f) | |
794 | if t is None: |
|
833 | if t is None: | |
795 | # Note: this should not be in the fncache then⦠|
|
834 | # Note: this should not be in the fncache then⦠| |
796 | # |
|
835 | # | |
797 | # However the fncache might contains such file added by |
|
836 | # However the fncache might contains such file added by | |
798 | # previous version of Mercurial. |
|
837 | # previous version of Mercurial. | |
799 | continue |
|
838 | continue | |
800 | t |= FILEFLAGS_FILELOG |
|
839 | t |= FILEFLAGS_FILELOG | |
801 | try: |
|
840 | try: | |
802 |
yield |
|
841 | yield StoreEntry( | |
|
842 | unencoded_path=f, | |||
|
843 | is_revlog=True, | |||
|
844 | revlog_type=FILEFLAGS_FILELOG, | |||
|
845 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), | |||
|
846 | is_volatile=bool(t & FILEFLAGS_VOLATILE), | |||
|
847 | file_size=self.getsize(ef), | |||
|
848 | ) | |||
803 | except FileNotFoundError: |
|
849 | except FileNotFoundError: | |
804 | pass |
|
850 | pass | |
805 |
|
851 | |||
806 | def copylist(self): |
|
852 | def copylist(self): | |
807 | d = ( |
|
853 | d = ( | |
808 | b'bookmarks', |
|
854 | b'bookmarks', | |
809 | b'narrowspec', |
|
855 | b'narrowspec', | |
810 | b'data', |
|
856 | b'data', | |
811 | b'meta', |
|
857 | b'meta', | |
812 | b'dh', |
|
858 | b'dh', | |
813 | b'fncache', |
|
859 | b'fncache', | |
814 | b'phaseroots', |
|
860 | b'phaseroots', | |
815 | b'obsstore', |
|
861 | b'obsstore', | |
816 | b'00manifest.d', |
|
862 | b'00manifest.d', | |
817 | b'00manifest.i', |
|
863 | b'00manifest.i', | |
818 | b'00changelog.d', |
|
864 | b'00changelog.d', | |
819 | b'00changelog.i', |
|
865 | b'00changelog.i', | |
820 | b'requires', |
|
866 | b'requires', | |
821 | ) |
|
867 | ) | |
822 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in d] |
|
868 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in d] | |
823 |
|
869 | |||
824 | def write(self, tr): |
|
870 | def write(self, tr): | |
825 | self.fncache.write(tr) |
|
871 | self.fncache.write(tr) | |
826 |
|
872 | |||
827 | def invalidatecaches(self): |
|
873 | def invalidatecaches(self): | |
828 | self.fncache.entries = None |
|
874 | self.fncache.entries = None | |
829 | self.fncache.addls = set() |
|
875 | self.fncache.addls = set() | |
830 |
|
876 | |||
831 | def markremoved(self, fn): |
|
877 | def markremoved(self, fn): | |
832 | self.fncache.remove(fn) |
|
878 | self.fncache.remove(fn) | |
833 |
|
879 | |||
834 | def _exists(self, f): |
|
880 | def _exists(self, f): | |
835 | ef = self.encode(f) |
|
881 | ef = self.encode(f) | |
836 | try: |
|
882 | try: | |
837 | self.getsize(ef) |
|
883 | self.getsize(ef) | |
838 | return True |
|
884 | return True | |
839 | except FileNotFoundError: |
|
885 | except FileNotFoundError: | |
840 | return False |
|
886 | return False | |
841 |
|
887 | |||
842 | def __contains__(self, path): |
|
888 | def __contains__(self, path): | |
843 | '''Checks if the store contains path''' |
|
889 | '''Checks if the store contains path''' | |
844 | path = b"/".join((b"data", path)) |
|
890 | path = b"/".join((b"data", path)) | |
845 | # check for files (exact match) |
|
891 | # check for files (exact match) | |
846 | e = path + b'.i' |
|
892 | e = path + b'.i' | |
847 | if e in self.fncache and self._exists(e): |
|
893 | if e in self.fncache and self._exists(e): | |
848 | return True |
|
894 | return True | |
849 | # now check for directories (prefix match) |
|
895 | # now check for directories (prefix match) | |
850 | if not path.endswith(b'/'): |
|
896 | if not path.endswith(b'/'): | |
851 | path += b'/' |
|
897 | path += b'/' | |
852 | for e in self.fncache: |
|
898 | for e in self.fncache: | |
853 | if e.startswith(path) and self._exists(e): |
|
899 | if e.startswith(path) and self._exists(e): | |
854 | return True |
|
900 | return True | |
855 | return False |
|
901 | return False |
@@ -1,935 +1,937 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 |
|
8 | |||
9 | import contextlib |
|
9 | import contextlib | |
10 | import os |
|
10 | import os | |
11 | import struct |
|
11 | import struct | |
12 |
|
12 | |||
13 | from .i18n import _ |
|
13 | from .i18n import _ | |
14 | from .pycompat import open |
|
14 | from .pycompat import open | |
15 | from .interfaces import repository |
|
15 | from .interfaces import repository | |
16 | from . import ( |
|
16 | from . import ( | |
17 | bookmarks, |
|
17 | bookmarks, | |
18 | cacheutil, |
|
18 | cacheutil, | |
19 | error, |
|
19 | error, | |
20 | narrowspec, |
|
20 | narrowspec, | |
21 | phases, |
|
21 | phases, | |
22 | pycompat, |
|
22 | pycompat, | |
23 | requirements as requirementsmod, |
|
23 | requirements as requirementsmod, | |
24 | scmutil, |
|
24 | scmutil, | |
25 | store, |
|
25 | store, | |
26 | transaction, |
|
26 | transaction, | |
27 | util, |
|
27 | util, | |
28 | ) |
|
28 | ) | |
29 | from .revlogutils import ( |
|
29 | from .revlogutils import ( | |
30 | nodemap, |
|
30 | nodemap, | |
31 | ) |
|
31 | ) | |
32 |
|
32 | |||
33 |
|
33 | |||
34 | def new_stream_clone_requirements(default_requirements, streamed_requirements): |
|
34 | def new_stream_clone_requirements(default_requirements, streamed_requirements): | |
35 | """determine the final set of requirement for a new stream clone |
|
35 | """determine the final set of requirement for a new stream clone | |
36 |
|
36 | |||
37 | this method combine the "default" requirements that a new repository would |
|
37 | this method combine the "default" requirements that a new repository would | |
38 | use with the constaint we get from the stream clone content. We keep local |
|
38 | use with the constaint we get from the stream clone content. We keep local | |
39 | configuration choice when possible. |
|
39 | configuration choice when possible. | |
40 | """ |
|
40 | """ | |
41 | requirements = set(default_requirements) |
|
41 | requirements = set(default_requirements) | |
42 | requirements -= requirementsmod.STREAM_FIXED_REQUIREMENTS |
|
42 | requirements -= requirementsmod.STREAM_FIXED_REQUIREMENTS | |
43 | requirements.update(streamed_requirements) |
|
43 | requirements.update(streamed_requirements) | |
44 | return requirements |
|
44 | return requirements | |
45 |
|
45 | |||
46 |
|
46 | |||
47 | def streamed_requirements(repo): |
|
47 | def streamed_requirements(repo): | |
48 | """the set of requirement the new clone will have to support |
|
48 | """the set of requirement the new clone will have to support | |
49 |
|
49 | |||
50 | This is used for advertising the stream options and to generate the actual |
|
50 | This is used for advertising the stream options and to generate the actual | |
51 | stream content.""" |
|
51 | stream content.""" | |
52 | requiredformats = ( |
|
52 | requiredformats = ( | |
53 | repo.requirements & requirementsmod.STREAM_FIXED_REQUIREMENTS |
|
53 | repo.requirements & requirementsmod.STREAM_FIXED_REQUIREMENTS | |
54 | ) |
|
54 | ) | |
55 | return requiredformats |
|
55 | return requiredformats | |
56 |
|
56 | |||
57 |
|
57 | |||
58 | def canperformstreamclone(pullop, bundle2=False): |
|
58 | def canperformstreamclone(pullop, bundle2=False): | |
59 | """Whether it is possible to perform a streaming clone as part of pull. |
|
59 | """Whether it is possible to perform a streaming clone as part of pull. | |
60 |
|
60 | |||
61 | ``bundle2`` will cause the function to consider stream clone through |
|
61 | ``bundle2`` will cause the function to consider stream clone through | |
62 | bundle2 and only through bundle2. |
|
62 | bundle2 and only through bundle2. | |
63 |
|
63 | |||
64 | Returns a tuple of (supported, requirements). ``supported`` is True if |
|
64 | Returns a tuple of (supported, requirements). ``supported`` is True if | |
65 | streaming clone is supported and False otherwise. ``requirements`` is |
|
65 | streaming clone is supported and False otherwise. ``requirements`` is | |
66 | a set of repo requirements from the remote, or ``None`` if stream clone |
|
66 | a set of repo requirements from the remote, or ``None`` if stream clone | |
67 | isn't supported. |
|
67 | isn't supported. | |
68 | """ |
|
68 | """ | |
69 | repo = pullop.repo |
|
69 | repo = pullop.repo | |
70 | remote = pullop.remote |
|
70 | remote = pullop.remote | |
71 |
|
71 | |||
72 | bundle2supported = False |
|
72 | bundle2supported = False | |
73 | if pullop.canusebundle2: |
|
73 | if pullop.canusebundle2: | |
74 | if b'v2' in pullop.remotebundle2caps.get(b'stream', []): |
|
74 | if b'v2' in pullop.remotebundle2caps.get(b'stream', []): | |
75 | bundle2supported = True |
|
75 | bundle2supported = True | |
76 | # else |
|
76 | # else | |
77 | # Server doesn't support bundle2 stream clone or doesn't support |
|
77 | # Server doesn't support bundle2 stream clone or doesn't support | |
78 | # the versions we support. Fall back and possibly allow legacy. |
|
78 | # the versions we support. Fall back and possibly allow legacy. | |
79 |
|
79 | |||
80 | # Ensures legacy code path uses available bundle2. |
|
80 | # Ensures legacy code path uses available bundle2. | |
81 | if bundle2supported and not bundle2: |
|
81 | if bundle2supported and not bundle2: | |
82 | return False, None |
|
82 | return False, None | |
83 | # Ensures bundle2 doesn't try to do a stream clone if it isn't supported. |
|
83 | # Ensures bundle2 doesn't try to do a stream clone if it isn't supported. | |
84 | elif bundle2 and not bundle2supported: |
|
84 | elif bundle2 and not bundle2supported: | |
85 | return False, None |
|
85 | return False, None | |
86 |
|
86 | |||
87 | # Streaming clone only works on empty repositories. |
|
87 | # Streaming clone only works on empty repositories. | |
88 | if len(repo): |
|
88 | if len(repo): | |
89 | return False, None |
|
89 | return False, None | |
90 |
|
90 | |||
91 | # Streaming clone only works if all data is being requested. |
|
91 | # Streaming clone only works if all data is being requested. | |
92 | if pullop.heads: |
|
92 | if pullop.heads: | |
93 | return False, None |
|
93 | return False, None | |
94 |
|
94 | |||
95 | streamrequested = pullop.streamclonerequested |
|
95 | streamrequested = pullop.streamclonerequested | |
96 |
|
96 | |||
97 | # If we don't have a preference, let the server decide for us. This |
|
97 | # If we don't have a preference, let the server decide for us. This | |
98 | # likely only comes into play in LANs. |
|
98 | # likely only comes into play in LANs. | |
99 | if streamrequested is None: |
|
99 | if streamrequested is None: | |
100 | # The server can advertise whether to prefer streaming clone. |
|
100 | # The server can advertise whether to prefer streaming clone. | |
101 | streamrequested = remote.capable(b'stream-preferred') |
|
101 | streamrequested = remote.capable(b'stream-preferred') | |
102 |
|
102 | |||
103 | if not streamrequested: |
|
103 | if not streamrequested: | |
104 | return False, None |
|
104 | return False, None | |
105 |
|
105 | |||
106 | # In order for stream clone to work, the client has to support all the |
|
106 | # In order for stream clone to work, the client has to support all the | |
107 | # requirements advertised by the server. |
|
107 | # requirements advertised by the server. | |
108 | # |
|
108 | # | |
109 | # The server advertises its requirements via the "stream" and "streamreqs" |
|
109 | # The server advertises its requirements via the "stream" and "streamreqs" | |
110 | # capability. "stream" (a value-less capability) is advertised if and only |
|
110 | # capability. "stream" (a value-less capability) is advertised if and only | |
111 | # if the only requirement is "revlogv1." Else, the "streamreqs" capability |
|
111 | # if the only requirement is "revlogv1." Else, the "streamreqs" capability | |
112 | # is advertised and contains a comma-delimited list of requirements. |
|
112 | # is advertised and contains a comma-delimited list of requirements. | |
113 | requirements = set() |
|
113 | requirements = set() | |
114 | if remote.capable(b'stream'): |
|
114 | if remote.capable(b'stream'): | |
115 | requirements.add(requirementsmod.REVLOGV1_REQUIREMENT) |
|
115 | requirements.add(requirementsmod.REVLOGV1_REQUIREMENT) | |
116 | else: |
|
116 | else: | |
117 | streamreqs = remote.capable(b'streamreqs') |
|
117 | streamreqs = remote.capable(b'streamreqs') | |
118 | # This is weird and shouldn't happen with modern servers. |
|
118 | # This is weird and shouldn't happen with modern servers. | |
119 | if not streamreqs: |
|
119 | if not streamreqs: | |
120 | pullop.repo.ui.warn( |
|
120 | pullop.repo.ui.warn( | |
121 | _( |
|
121 | _( | |
122 | b'warning: stream clone requested but server has them ' |
|
122 | b'warning: stream clone requested but server has them ' | |
123 | b'disabled\n' |
|
123 | b'disabled\n' | |
124 | ) |
|
124 | ) | |
125 | ) |
|
125 | ) | |
126 | return False, None |
|
126 | return False, None | |
127 |
|
127 | |||
128 | streamreqs = set(streamreqs.split(b',')) |
|
128 | streamreqs = set(streamreqs.split(b',')) | |
129 | # Server requires something we don't support. Bail. |
|
129 | # Server requires something we don't support. Bail. | |
130 | missingreqs = streamreqs - repo.supported |
|
130 | missingreqs = streamreqs - repo.supported | |
131 | if missingreqs: |
|
131 | if missingreqs: | |
132 | pullop.repo.ui.warn( |
|
132 | pullop.repo.ui.warn( | |
133 | _( |
|
133 | _( | |
134 | b'warning: stream clone requested but client is missing ' |
|
134 | b'warning: stream clone requested but client is missing ' | |
135 | b'requirements: %s\n' |
|
135 | b'requirements: %s\n' | |
136 | ) |
|
136 | ) | |
137 | % b', '.join(sorted(missingreqs)) |
|
137 | % b', '.join(sorted(missingreqs)) | |
138 | ) |
|
138 | ) | |
139 | pullop.repo.ui.warn( |
|
139 | pullop.repo.ui.warn( | |
140 | _( |
|
140 | _( | |
141 | b'(see https://www.mercurial-scm.org/wiki/MissingRequirement ' |
|
141 | b'(see https://www.mercurial-scm.org/wiki/MissingRequirement ' | |
142 | b'for more information)\n' |
|
142 | b'for more information)\n' | |
143 | ) |
|
143 | ) | |
144 | ) |
|
144 | ) | |
145 | return False, None |
|
145 | return False, None | |
146 | requirements = streamreqs |
|
146 | requirements = streamreqs | |
147 |
|
147 | |||
148 | return True, requirements |
|
148 | return True, requirements | |
149 |
|
149 | |||
150 |
|
150 | |||
151 | def maybeperformlegacystreamclone(pullop): |
|
151 | def maybeperformlegacystreamclone(pullop): | |
152 | """Possibly perform a legacy stream clone operation. |
|
152 | """Possibly perform a legacy stream clone operation. | |
153 |
|
153 | |||
154 | Legacy stream clones are performed as part of pull but before all other |
|
154 | Legacy stream clones are performed as part of pull but before all other | |
155 | operations. |
|
155 | operations. | |
156 |
|
156 | |||
157 | A legacy stream clone will not be performed if a bundle2 stream clone is |
|
157 | A legacy stream clone will not be performed if a bundle2 stream clone is | |
158 | supported. |
|
158 | supported. | |
159 | """ |
|
159 | """ | |
160 | from . import localrepo |
|
160 | from . import localrepo | |
161 |
|
161 | |||
162 | supported, requirements = canperformstreamclone(pullop) |
|
162 | supported, requirements = canperformstreamclone(pullop) | |
163 |
|
163 | |||
164 | if not supported: |
|
164 | if not supported: | |
165 | return |
|
165 | return | |
166 |
|
166 | |||
167 | repo = pullop.repo |
|
167 | repo = pullop.repo | |
168 | remote = pullop.remote |
|
168 | remote = pullop.remote | |
169 |
|
169 | |||
170 | # Save remote branchmap. We will use it later to speed up branchcache |
|
170 | # Save remote branchmap. We will use it later to speed up branchcache | |
171 | # creation. |
|
171 | # creation. | |
172 | rbranchmap = None |
|
172 | rbranchmap = None | |
173 | if remote.capable(b'branchmap'): |
|
173 | if remote.capable(b'branchmap'): | |
174 | with remote.commandexecutor() as e: |
|
174 | with remote.commandexecutor() as e: | |
175 | rbranchmap = e.callcommand(b'branchmap', {}).result() |
|
175 | rbranchmap = e.callcommand(b'branchmap', {}).result() | |
176 |
|
176 | |||
177 | repo.ui.status(_(b'streaming all changes\n')) |
|
177 | repo.ui.status(_(b'streaming all changes\n')) | |
178 |
|
178 | |||
179 | with remote.commandexecutor() as e: |
|
179 | with remote.commandexecutor() as e: | |
180 | fp = e.callcommand(b'stream_out', {}).result() |
|
180 | fp = e.callcommand(b'stream_out', {}).result() | |
181 |
|
181 | |||
182 | # TODO strictly speaking, this code should all be inside the context |
|
182 | # TODO strictly speaking, this code should all be inside the context | |
183 | # manager because the context manager is supposed to ensure all wire state |
|
183 | # manager because the context manager is supposed to ensure all wire state | |
184 | # is flushed when exiting. But the legacy peers don't do this, so it |
|
184 | # is flushed when exiting. But the legacy peers don't do this, so it | |
185 | # doesn't matter. |
|
185 | # doesn't matter. | |
186 | l = fp.readline() |
|
186 | l = fp.readline() | |
187 | try: |
|
187 | try: | |
188 | resp = int(l) |
|
188 | resp = int(l) | |
189 | except ValueError: |
|
189 | except ValueError: | |
190 | raise error.ResponseError( |
|
190 | raise error.ResponseError( | |
191 | _(b'unexpected response from remote server:'), l |
|
191 | _(b'unexpected response from remote server:'), l | |
192 | ) |
|
192 | ) | |
193 | if resp == 1: |
|
193 | if resp == 1: | |
194 | raise error.Abort(_(b'operation forbidden by server')) |
|
194 | raise error.Abort(_(b'operation forbidden by server')) | |
195 | elif resp == 2: |
|
195 | elif resp == 2: | |
196 | raise error.Abort(_(b'locking the remote repository failed')) |
|
196 | raise error.Abort(_(b'locking the remote repository failed')) | |
197 | elif resp != 0: |
|
197 | elif resp != 0: | |
198 | raise error.Abort(_(b'the server sent an unknown error code')) |
|
198 | raise error.Abort(_(b'the server sent an unknown error code')) | |
199 |
|
199 | |||
200 | l = fp.readline() |
|
200 | l = fp.readline() | |
201 | try: |
|
201 | try: | |
202 | filecount, bytecount = map(int, l.split(b' ', 1)) |
|
202 | filecount, bytecount = map(int, l.split(b' ', 1)) | |
203 | except (ValueError, TypeError): |
|
203 | except (ValueError, TypeError): | |
204 | raise error.ResponseError( |
|
204 | raise error.ResponseError( | |
205 | _(b'unexpected response from remote server:'), l |
|
205 | _(b'unexpected response from remote server:'), l | |
206 | ) |
|
206 | ) | |
207 |
|
207 | |||
208 | with repo.lock(): |
|
208 | with repo.lock(): | |
209 | consumev1(repo, fp, filecount, bytecount) |
|
209 | consumev1(repo, fp, filecount, bytecount) | |
210 | repo.requirements = new_stream_clone_requirements( |
|
210 | repo.requirements = new_stream_clone_requirements( | |
211 | repo.requirements, |
|
211 | repo.requirements, | |
212 | requirements, |
|
212 | requirements, | |
213 | ) |
|
213 | ) | |
214 | repo.svfs.options = localrepo.resolvestorevfsoptions( |
|
214 | repo.svfs.options = localrepo.resolvestorevfsoptions( | |
215 | repo.ui, repo.requirements, repo.features |
|
215 | repo.ui, repo.requirements, repo.features | |
216 | ) |
|
216 | ) | |
217 | scmutil.writereporequirements(repo) |
|
217 | scmutil.writereporequirements(repo) | |
218 | nodemap.post_stream_cleanup(repo) |
|
218 | nodemap.post_stream_cleanup(repo) | |
219 |
|
219 | |||
220 | if rbranchmap: |
|
220 | if rbranchmap: | |
221 | repo._branchcaches.replace(repo, rbranchmap) |
|
221 | repo._branchcaches.replace(repo, rbranchmap) | |
222 |
|
222 | |||
223 | repo.invalidate() |
|
223 | repo.invalidate() | |
224 |
|
224 | |||
225 |
|
225 | |||
226 | def allowservergeneration(repo): |
|
226 | def allowservergeneration(repo): | |
227 | """Whether streaming clones are allowed from the server.""" |
|
227 | """Whether streaming clones are allowed from the server.""" | |
228 | if repository.REPO_FEATURE_STREAM_CLONE not in repo.features: |
|
228 | if repository.REPO_FEATURE_STREAM_CLONE not in repo.features: | |
229 | return False |
|
229 | return False | |
230 |
|
230 | |||
231 | if not repo.ui.configbool(b'server', b'uncompressed', untrusted=True): |
|
231 | if not repo.ui.configbool(b'server', b'uncompressed', untrusted=True): | |
232 | return False |
|
232 | return False | |
233 |
|
233 | |||
234 | # The way stream clone works makes it impossible to hide secret changesets. |
|
234 | # The way stream clone works makes it impossible to hide secret changesets. | |
235 | # So don't allow this by default. |
|
235 | # So don't allow this by default. | |
236 | secret = phases.hassecret(repo) |
|
236 | secret = phases.hassecret(repo) | |
237 | if secret: |
|
237 | if secret: | |
238 | return repo.ui.configbool(b'server', b'uncompressedallowsecret') |
|
238 | return repo.ui.configbool(b'server', b'uncompressedallowsecret') | |
239 |
|
239 | |||
240 | return True |
|
240 | return True | |
241 |
|
241 | |||
242 |
|
242 | |||
243 | # This is it's own function so extensions can override it. |
|
243 | # This is it's own function so extensions can override it. | |
244 | def _walkstreamfiles(repo, matcher=None): |
|
244 | def _walkstreamfiles(repo, matcher=None): | |
245 | return repo.store.walk(matcher) |
|
245 | return repo.store.walk(matcher) | |
246 |
|
246 | |||
247 |
|
247 | |||
248 | def generatev1(repo): |
|
248 | def generatev1(repo): | |
249 | """Emit content for version 1 of a streaming clone. |
|
249 | """Emit content for version 1 of a streaming clone. | |
250 |
|
250 | |||
251 | This returns a 3-tuple of (file count, byte size, data iterator). |
|
251 | This returns a 3-tuple of (file count, byte size, data iterator). | |
252 |
|
252 | |||
253 | The data iterator consists of N entries for each file being transferred. |
|
253 | The data iterator consists of N entries for each file being transferred. | |
254 | Each file entry starts as a line with the file name and integer size |
|
254 | Each file entry starts as a line with the file name and integer size | |
255 | delimited by a null byte. |
|
255 | delimited by a null byte. | |
256 |
|
256 | |||
257 | The raw file data follows. Following the raw file data is the next file |
|
257 | The raw file data follows. Following the raw file data is the next file | |
258 | entry, or EOF. |
|
258 | entry, or EOF. | |
259 |
|
259 | |||
260 | When used on the wire protocol, an additional line indicating protocol |
|
260 | When used on the wire protocol, an additional line indicating protocol | |
261 | success will be prepended to the stream. This function is not responsible |
|
261 | success will be prepended to the stream. This function is not responsible | |
262 | for adding it. |
|
262 | for adding it. | |
263 |
|
263 | |||
264 | This function will obtain a repository lock to ensure a consistent view of |
|
264 | This function will obtain a repository lock to ensure a consistent view of | |
265 | the store is captured. It therefore may raise LockError. |
|
265 | the store is captured. It therefore may raise LockError. | |
266 | """ |
|
266 | """ | |
267 | entries = [] |
|
267 | entries = [] | |
268 | total_bytes = 0 |
|
268 | total_bytes = 0 | |
269 | # Get consistent snapshot of repo, lock during scan. |
|
269 | # Get consistent snapshot of repo, lock during scan. | |
270 | with repo.lock(): |
|
270 | with repo.lock(): | |
271 | repo.ui.debug(b'scanning\n') |
|
271 | repo.ui.debug(b'scanning\n') | |
272 |
for |
|
272 | for entry in _walkstreamfiles(repo): | |
273 | if size: |
|
273 | if entry.file_size: | |
274 |
entries.append(( |
|
274 | entries.append((entry.unencoded_path, entry.file_size)) | |
275 | total_bytes += size |
|
275 | total_bytes += entry.file_size | |
276 | _test_sync_point_walk_1(repo) |
|
276 | _test_sync_point_walk_1(repo) | |
277 | _test_sync_point_walk_2(repo) |
|
277 | _test_sync_point_walk_2(repo) | |
278 |
|
278 | |||
279 | repo.ui.debug( |
|
279 | repo.ui.debug( | |
280 | b'%d files, %d bytes to transfer\n' % (len(entries), total_bytes) |
|
280 | b'%d files, %d bytes to transfer\n' % (len(entries), total_bytes) | |
281 | ) |
|
281 | ) | |
282 |
|
282 | |||
283 | svfs = repo.svfs |
|
283 | svfs = repo.svfs | |
284 | debugflag = repo.ui.debugflag |
|
284 | debugflag = repo.ui.debugflag | |
285 |
|
285 | |||
286 | def emitrevlogdata(): |
|
286 | def emitrevlogdata(): | |
287 | for name, size in entries: |
|
287 | for name, size in entries: | |
288 | if debugflag: |
|
288 | if debugflag: | |
289 | repo.ui.debug(b'sending %s (%d bytes)\n' % (name, size)) |
|
289 | repo.ui.debug(b'sending %s (%d bytes)\n' % (name, size)) | |
290 | # partially encode name over the wire for backwards compat |
|
290 | # partially encode name over the wire for backwards compat | |
291 | yield b'%s\0%d\n' % (store.encodedir(name), size) |
|
291 | yield b'%s\0%d\n' % (store.encodedir(name), size) | |
292 | # auditing at this stage is both pointless (paths are already |
|
292 | # auditing at this stage is both pointless (paths are already | |
293 | # trusted by the local repo) and expensive |
|
293 | # trusted by the local repo) and expensive | |
294 | with svfs(name, b'rb', auditpath=False) as fp: |
|
294 | with svfs(name, b'rb', auditpath=False) as fp: | |
295 | if size <= 65536: |
|
295 | if size <= 65536: | |
296 | yield fp.read(size) |
|
296 | yield fp.read(size) | |
297 | else: |
|
297 | else: | |
298 | for chunk in util.filechunkiter(fp, limit=size): |
|
298 | for chunk in util.filechunkiter(fp, limit=size): | |
299 | yield chunk |
|
299 | yield chunk | |
300 |
|
300 | |||
301 | return len(entries), total_bytes, emitrevlogdata() |
|
301 | return len(entries), total_bytes, emitrevlogdata() | |
302 |
|
302 | |||
303 |
|
303 | |||
304 | def generatev1wireproto(repo): |
|
304 | def generatev1wireproto(repo): | |
305 | """Emit content for version 1 of streaming clone suitable for the wire. |
|
305 | """Emit content for version 1 of streaming clone suitable for the wire. | |
306 |
|
306 | |||
307 | This is the data output from ``generatev1()`` with 2 header lines. The |
|
307 | This is the data output from ``generatev1()`` with 2 header lines. The | |
308 | first line indicates overall success. The 2nd contains the file count and |
|
308 | first line indicates overall success. The 2nd contains the file count and | |
309 | byte size of payload. |
|
309 | byte size of payload. | |
310 |
|
310 | |||
311 | The success line contains "0" for success, "1" for stream generation not |
|
311 | The success line contains "0" for success, "1" for stream generation not | |
312 | allowed, and "2" for error locking the repository (possibly indicating |
|
312 | allowed, and "2" for error locking the repository (possibly indicating | |
313 | a permissions error for the server process). |
|
313 | a permissions error for the server process). | |
314 | """ |
|
314 | """ | |
315 | if not allowservergeneration(repo): |
|
315 | if not allowservergeneration(repo): | |
316 | yield b'1\n' |
|
316 | yield b'1\n' | |
317 | return |
|
317 | return | |
318 |
|
318 | |||
319 | try: |
|
319 | try: | |
320 | filecount, bytecount, it = generatev1(repo) |
|
320 | filecount, bytecount, it = generatev1(repo) | |
321 | except error.LockError: |
|
321 | except error.LockError: | |
322 | yield b'2\n' |
|
322 | yield b'2\n' | |
323 | return |
|
323 | return | |
324 |
|
324 | |||
325 | # Indicates successful response. |
|
325 | # Indicates successful response. | |
326 | yield b'0\n' |
|
326 | yield b'0\n' | |
327 | yield b'%d %d\n' % (filecount, bytecount) |
|
327 | yield b'%d %d\n' % (filecount, bytecount) | |
328 | for chunk in it: |
|
328 | for chunk in it: | |
329 | yield chunk |
|
329 | yield chunk | |
330 |
|
330 | |||
331 |
|
331 | |||
332 | def generatebundlev1(repo, compression=b'UN'): |
|
332 | def generatebundlev1(repo, compression=b'UN'): | |
333 | """Emit content for version 1 of a stream clone bundle. |
|
333 | """Emit content for version 1 of a stream clone bundle. | |
334 |
|
334 | |||
335 | The first 4 bytes of the output ("HGS1") denote this as stream clone |
|
335 | The first 4 bytes of the output ("HGS1") denote this as stream clone | |
336 | bundle version 1. |
|
336 | bundle version 1. | |
337 |
|
337 | |||
338 | The next 2 bytes indicate the compression type. Only "UN" is currently |
|
338 | The next 2 bytes indicate the compression type. Only "UN" is currently | |
339 | supported. |
|
339 | supported. | |
340 |
|
340 | |||
341 | The next 16 bytes are two 64-bit big endian unsigned integers indicating |
|
341 | The next 16 bytes are two 64-bit big endian unsigned integers indicating | |
342 | file count and byte count, respectively. |
|
342 | file count and byte count, respectively. | |
343 |
|
343 | |||
344 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length |
|
344 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length | |
345 | of the requirements string, including a trailing \0. The following N bytes |
|
345 | of the requirements string, including a trailing \0. The following N bytes | |
346 | are the requirements string, which is ASCII containing a comma-delimited |
|
346 | are the requirements string, which is ASCII containing a comma-delimited | |
347 | list of repo requirements that are needed to support the data. |
|
347 | list of repo requirements that are needed to support the data. | |
348 |
|
348 | |||
349 | The remaining content is the output of ``generatev1()`` (which may be |
|
349 | The remaining content is the output of ``generatev1()`` (which may be | |
350 | compressed in the future). |
|
350 | compressed in the future). | |
351 |
|
351 | |||
352 | Returns a tuple of (requirements, data generator). |
|
352 | Returns a tuple of (requirements, data generator). | |
353 | """ |
|
353 | """ | |
354 | if compression != b'UN': |
|
354 | if compression != b'UN': | |
355 | raise ValueError(b'we do not support the compression argument yet') |
|
355 | raise ValueError(b'we do not support the compression argument yet') | |
356 |
|
356 | |||
357 | requirements = streamed_requirements(repo) |
|
357 | requirements = streamed_requirements(repo) | |
358 | requires = b','.join(sorted(requirements)) |
|
358 | requires = b','.join(sorted(requirements)) | |
359 |
|
359 | |||
360 | def gen(): |
|
360 | def gen(): | |
361 | yield b'HGS1' |
|
361 | yield b'HGS1' | |
362 | yield compression |
|
362 | yield compression | |
363 |
|
363 | |||
364 | filecount, bytecount, it = generatev1(repo) |
|
364 | filecount, bytecount, it = generatev1(repo) | |
365 | repo.ui.status( |
|
365 | repo.ui.status( | |
366 | _(b'writing %d bytes for %d files\n') % (bytecount, filecount) |
|
366 | _(b'writing %d bytes for %d files\n') % (bytecount, filecount) | |
367 | ) |
|
367 | ) | |
368 |
|
368 | |||
369 | yield struct.pack(b'>QQ', filecount, bytecount) |
|
369 | yield struct.pack(b'>QQ', filecount, bytecount) | |
370 | yield struct.pack(b'>H', len(requires) + 1) |
|
370 | yield struct.pack(b'>H', len(requires) + 1) | |
371 | yield requires + b'\0' |
|
371 | yield requires + b'\0' | |
372 |
|
372 | |||
373 | # This is where we'll add compression in the future. |
|
373 | # This is where we'll add compression in the future. | |
374 | assert compression == b'UN' |
|
374 | assert compression == b'UN' | |
375 |
|
375 | |||
376 | progress = repo.ui.makeprogress( |
|
376 | progress = repo.ui.makeprogress( | |
377 | _(b'bundle'), total=bytecount, unit=_(b'bytes') |
|
377 | _(b'bundle'), total=bytecount, unit=_(b'bytes') | |
378 | ) |
|
378 | ) | |
379 | progress.update(0) |
|
379 | progress.update(0) | |
380 |
|
380 | |||
381 | for chunk in it: |
|
381 | for chunk in it: | |
382 | progress.increment(step=len(chunk)) |
|
382 | progress.increment(step=len(chunk)) | |
383 | yield chunk |
|
383 | yield chunk | |
384 |
|
384 | |||
385 | progress.complete() |
|
385 | progress.complete() | |
386 |
|
386 | |||
387 | return requirements, gen() |
|
387 | return requirements, gen() | |
388 |
|
388 | |||
389 |
|
389 | |||
390 | def consumev1(repo, fp, filecount, bytecount): |
|
390 | def consumev1(repo, fp, filecount, bytecount): | |
391 | """Apply the contents from version 1 of a streaming clone file handle. |
|
391 | """Apply the contents from version 1 of a streaming clone file handle. | |
392 |
|
392 | |||
393 | This takes the output from "stream_out" and applies it to the specified |
|
393 | This takes the output from "stream_out" and applies it to the specified | |
394 | repository. |
|
394 | repository. | |
395 |
|
395 | |||
396 | Like "stream_out," the status line added by the wire protocol is not |
|
396 | Like "stream_out," the status line added by the wire protocol is not | |
397 | handled by this function. |
|
397 | handled by this function. | |
398 | """ |
|
398 | """ | |
399 | with repo.lock(): |
|
399 | with repo.lock(): | |
400 | repo.ui.status( |
|
400 | repo.ui.status( | |
401 | _(b'%d files to transfer, %s of data\n') |
|
401 | _(b'%d files to transfer, %s of data\n') | |
402 | % (filecount, util.bytecount(bytecount)) |
|
402 | % (filecount, util.bytecount(bytecount)) | |
403 | ) |
|
403 | ) | |
404 | progress = repo.ui.makeprogress( |
|
404 | progress = repo.ui.makeprogress( | |
405 | _(b'clone'), total=bytecount, unit=_(b'bytes') |
|
405 | _(b'clone'), total=bytecount, unit=_(b'bytes') | |
406 | ) |
|
406 | ) | |
407 | progress.update(0) |
|
407 | progress.update(0) | |
408 | start = util.timer() |
|
408 | start = util.timer() | |
409 |
|
409 | |||
410 | # TODO: get rid of (potential) inconsistency |
|
410 | # TODO: get rid of (potential) inconsistency | |
411 | # |
|
411 | # | |
412 | # If transaction is started and any @filecache property is |
|
412 | # If transaction is started and any @filecache property is | |
413 | # changed at this point, it causes inconsistency between |
|
413 | # changed at this point, it causes inconsistency between | |
414 | # in-memory cached property and streamclone-ed file on the |
|
414 | # in-memory cached property and streamclone-ed file on the | |
415 | # disk. Nested transaction prevents transaction scope "clone" |
|
415 | # disk. Nested transaction prevents transaction scope "clone" | |
416 | # below from writing in-memory changes out at the end of it, |
|
416 | # below from writing in-memory changes out at the end of it, | |
417 | # even though in-memory changes are discarded at the end of it |
|
417 | # even though in-memory changes are discarded at the end of it | |
418 | # regardless of transaction nesting. |
|
418 | # regardless of transaction nesting. | |
419 | # |
|
419 | # | |
420 | # But transaction nesting can't be simply prohibited, because |
|
420 | # But transaction nesting can't be simply prohibited, because | |
421 | # nesting occurs also in ordinary case (e.g. enabling |
|
421 | # nesting occurs also in ordinary case (e.g. enabling | |
422 | # clonebundles). |
|
422 | # clonebundles). | |
423 |
|
423 | |||
424 | with repo.transaction(b'clone'): |
|
424 | with repo.transaction(b'clone'): | |
425 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): |
|
425 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): | |
426 | for i in range(filecount): |
|
426 | for i in range(filecount): | |
427 | # XXX doesn't support '\n' or '\r' in filenames |
|
427 | # XXX doesn't support '\n' or '\r' in filenames | |
428 | l = fp.readline() |
|
428 | l = fp.readline() | |
429 | try: |
|
429 | try: | |
430 | name, size = l.split(b'\0', 1) |
|
430 | name, size = l.split(b'\0', 1) | |
431 | size = int(size) |
|
431 | size = int(size) | |
432 | except (ValueError, TypeError): |
|
432 | except (ValueError, TypeError): | |
433 | raise error.ResponseError( |
|
433 | raise error.ResponseError( | |
434 | _(b'unexpected response from remote server:'), l |
|
434 | _(b'unexpected response from remote server:'), l | |
435 | ) |
|
435 | ) | |
436 | if repo.ui.debugflag: |
|
436 | if repo.ui.debugflag: | |
437 | repo.ui.debug( |
|
437 | repo.ui.debug( | |
438 | b'adding %s (%s)\n' % (name, util.bytecount(size)) |
|
438 | b'adding %s (%s)\n' % (name, util.bytecount(size)) | |
439 | ) |
|
439 | ) | |
440 | # for backwards compat, name was partially encoded |
|
440 | # for backwards compat, name was partially encoded | |
441 | path = store.decodedir(name) |
|
441 | path = store.decodedir(name) | |
442 | with repo.svfs(path, b'w', backgroundclose=True) as ofp: |
|
442 | with repo.svfs(path, b'w', backgroundclose=True) as ofp: | |
443 | for chunk in util.filechunkiter(fp, limit=size): |
|
443 | for chunk in util.filechunkiter(fp, limit=size): | |
444 | progress.increment(step=len(chunk)) |
|
444 | progress.increment(step=len(chunk)) | |
445 | ofp.write(chunk) |
|
445 | ofp.write(chunk) | |
446 |
|
446 | |||
447 | # force @filecache properties to be reloaded from |
|
447 | # force @filecache properties to be reloaded from | |
448 | # streamclone-ed file at next access |
|
448 | # streamclone-ed file at next access | |
449 | repo.invalidate(clearfilecache=True) |
|
449 | repo.invalidate(clearfilecache=True) | |
450 |
|
450 | |||
451 | elapsed = util.timer() - start |
|
451 | elapsed = util.timer() - start | |
452 | if elapsed <= 0: |
|
452 | if elapsed <= 0: | |
453 | elapsed = 0.001 |
|
453 | elapsed = 0.001 | |
454 | progress.complete() |
|
454 | progress.complete() | |
455 | repo.ui.status( |
|
455 | repo.ui.status( | |
456 | _(b'transferred %s in %.1f seconds (%s/sec)\n') |
|
456 | _(b'transferred %s in %.1f seconds (%s/sec)\n') | |
457 | % ( |
|
457 | % ( | |
458 | util.bytecount(bytecount), |
|
458 | util.bytecount(bytecount), | |
459 | elapsed, |
|
459 | elapsed, | |
460 | util.bytecount(bytecount / elapsed), |
|
460 | util.bytecount(bytecount / elapsed), | |
461 | ) |
|
461 | ) | |
462 | ) |
|
462 | ) | |
463 |
|
463 | |||
464 |
|
464 | |||
465 | def readbundle1header(fp): |
|
465 | def readbundle1header(fp): | |
466 | compression = fp.read(2) |
|
466 | compression = fp.read(2) | |
467 | if compression != b'UN': |
|
467 | if compression != b'UN': | |
468 | raise error.Abort( |
|
468 | raise error.Abort( | |
469 | _( |
|
469 | _( | |
470 | b'only uncompressed stream clone bundles are ' |
|
470 | b'only uncompressed stream clone bundles are ' | |
471 | b'supported; got %s' |
|
471 | b'supported; got %s' | |
472 | ) |
|
472 | ) | |
473 | % compression |
|
473 | % compression | |
474 | ) |
|
474 | ) | |
475 |
|
475 | |||
476 | filecount, bytecount = struct.unpack(b'>QQ', fp.read(16)) |
|
476 | filecount, bytecount = struct.unpack(b'>QQ', fp.read(16)) | |
477 | requireslen = struct.unpack(b'>H', fp.read(2))[0] |
|
477 | requireslen = struct.unpack(b'>H', fp.read(2))[0] | |
478 | requires = fp.read(requireslen) |
|
478 | requires = fp.read(requireslen) | |
479 |
|
479 | |||
480 | if not requires.endswith(b'\0'): |
|
480 | if not requires.endswith(b'\0'): | |
481 | raise error.Abort( |
|
481 | raise error.Abort( | |
482 | _( |
|
482 | _( | |
483 | b'malformed stream clone bundle: ' |
|
483 | b'malformed stream clone bundle: ' | |
484 | b'requirements not properly encoded' |
|
484 | b'requirements not properly encoded' | |
485 | ) |
|
485 | ) | |
486 | ) |
|
486 | ) | |
487 |
|
487 | |||
488 | requirements = set(requires.rstrip(b'\0').split(b',')) |
|
488 | requirements = set(requires.rstrip(b'\0').split(b',')) | |
489 |
|
489 | |||
490 | return filecount, bytecount, requirements |
|
490 | return filecount, bytecount, requirements | |
491 |
|
491 | |||
492 |
|
492 | |||
493 | def applybundlev1(repo, fp): |
|
493 | def applybundlev1(repo, fp): | |
494 | """Apply the content from a stream clone bundle version 1. |
|
494 | """Apply the content from a stream clone bundle version 1. | |
495 |
|
495 | |||
496 | We assume the 4 byte header has been read and validated and the file handle |
|
496 | We assume the 4 byte header has been read and validated and the file handle | |
497 | is at the 2 byte compression identifier. |
|
497 | is at the 2 byte compression identifier. | |
498 | """ |
|
498 | """ | |
499 | if len(repo): |
|
499 | if len(repo): | |
500 | raise error.Abort( |
|
500 | raise error.Abort( | |
501 | _(b'cannot apply stream clone bundle on non-empty repo') |
|
501 | _(b'cannot apply stream clone bundle on non-empty repo') | |
502 | ) |
|
502 | ) | |
503 |
|
503 | |||
504 | filecount, bytecount, requirements = readbundle1header(fp) |
|
504 | filecount, bytecount, requirements = readbundle1header(fp) | |
505 | missingreqs = requirements - repo.supported |
|
505 | missingreqs = requirements - repo.supported | |
506 | if missingreqs: |
|
506 | if missingreqs: | |
507 | raise error.Abort( |
|
507 | raise error.Abort( | |
508 | _(b'unable to apply stream clone: unsupported format: %s') |
|
508 | _(b'unable to apply stream clone: unsupported format: %s') | |
509 | % b', '.join(sorted(missingreqs)) |
|
509 | % b', '.join(sorted(missingreqs)) | |
510 | ) |
|
510 | ) | |
511 |
|
511 | |||
512 | consumev1(repo, fp, filecount, bytecount) |
|
512 | consumev1(repo, fp, filecount, bytecount) | |
513 | nodemap.post_stream_cleanup(repo) |
|
513 | nodemap.post_stream_cleanup(repo) | |
514 |
|
514 | |||
515 |
|
515 | |||
516 | class streamcloneapplier: |
|
516 | class streamcloneapplier: | |
517 | """Class to manage applying streaming clone bundles. |
|
517 | """Class to manage applying streaming clone bundles. | |
518 |
|
518 | |||
519 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle |
|
519 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle | |
520 | readers to perform bundle type-specific functionality. |
|
520 | readers to perform bundle type-specific functionality. | |
521 | """ |
|
521 | """ | |
522 |
|
522 | |||
523 | def __init__(self, fh): |
|
523 | def __init__(self, fh): | |
524 | self._fh = fh |
|
524 | self._fh = fh | |
525 |
|
525 | |||
526 | def apply(self, repo): |
|
526 | def apply(self, repo): | |
527 | return applybundlev1(repo, self._fh) |
|
527 | return applybundlev1(repo, self._fh) | |
528 |
|
528 | |||
529 |
|
529 | |||
530 | # type of file to stream |
|
530 | # type of file to stream | |
531 | _fileappend = 0 # append only file |
|
531 | _fileappend = 0 # append only file | |
532 | _filefull = 1 # full snapshot file |
|
532 | _filefull = 1 # full snapshot file | |
533 |
|
533 | |||
534 | # Source of the file |
|
534 | # Source of the file | |
535 | _srcstore = b's' # store (svfs) |
|
535 | _srcstore = b's' # store (svfs) | |
536 | _srccache = b'c' # cache (cache) |
|
536 | _srccache = b'c' # cache (cache) | |
537 |
|
537 | |||
538 | # This is it's own function so extensions can override it. |
|
538 | # This is it's own function so extensions can override it. | |
539 | def _walkstreamfullstorefiles(repo): |
|
539 | def _walkstreamfullstorefiles(repo): | |
540 | """list snapshot file from the store""" |
|
540 | """list snapshot file from the store""" | |
541 | fnames = [] |
|
541 | fnames = [] | |
542 | if not repo.publishing(): |
|
542 | if not repo.publishing(): | |
543 | fnames.append(b'phaseroots') |
|
543 | fnames.append(b'phaseroots') | |
544 | return fnames |
|
544 | return fnames | |
545 |
|
545 | |||
546 |
|
546 | |||
547 | def _filterfull(entry, copy, vfsmap): |
|
547 | def _filterfull(entry, copy, vfsmap): | |
548 | """actually copy the snapshot files""" |
|
548 | """actually copy the snapshot files""" | |
549 | src, name, ftype, data = entry |
|
549 | src, name, ftype, data = entry | |
550 | if ftype != _filefull: |
|
550 | if ftype != _filefull: | |
551 | return entry |
|
551 | return entry | |
552 | return (src, name, ftype, copy(vfsmap[src].join(name))) |
|
552 | return (src, name, ftype, copy(vfsmap[src].join(name))) | |
553 |
|
553 | |||
554 |
|
554 | |||
555 | @contextlib.contextmanager |
|
555 | @contextlib.contextmanager | |
556 | def maketempcopies(): |
|
556 | def maketempcopies(): | |
557 | """return a function to temporary copy file""" |
|
557 | """return a function to temporary copy file""" | |
558 |
|
558 | |||
559 | files = [] |
|
559 | files = [] | |
560 | dst_dir = pycompat.mkdtemp(prefix=b'hg-clone-') |
|
560 | dst_dir = pycompat.mkdtemp(prefix=b'hg-clone-') | |
561 | try: |
|
561 | try: | |
562 |
|
562 | |||
563 | def copy(src): |
|
563 | def copy(src): | |
564 | fd, dst = pycompat.mkstemp( |
|
564 | fd, dst = pycompat.mkstemp( | |
565 | prefix=os.path.basename(src), dir=dst_dir |
|
565 | prefix=os.path.basename(src), dir=dst_dir | |
566 | ) |
|
566 | ) | |
567 | os.close(fd) |
|
567 | os.close(fd) | |
568 | files.append(dst) |
|
568 | files.append(dst) | |
569 | util.copyfiles(src, dst, hardlink=True) |
|
569 | util.copyfiles(src, dst, hardlink=True) | |
570 | return dst |
|
570 | return dst | |
571 |
|
571 | |||
572 | yield copy |
|
572 | yield copy | |
573 | finally: |
|
573 | finally: | |
574 | for tmp in files: |
|
574 | for tmp in files: | |
575 | util.tryunlink(tmp) |
|
575 | util.tryunlink(tmp) | |
576 | util.tryrmdir(dst_dir) |
|
576 | util.tryrmdir(dst_dir) | |
577 |
|
577 | |||
578 |
|
578 | |||
579 | def _makemap(repo): |
|
579 | def _makemap(repo): | |
580 | """make a (src -> vfs) map for the repo""" |
|
580 | """make a (src -> vfs) map for the repo""" | |
581 | vfsmap = { |
|
581 | vfsmap = { | |
582 | _srcstore: repo.svfs, |
|
582 | _srcstore: repo.svfs, | |
583 | _srccache: repo.cachevfs, |
|
583 | _srccache: repo.cachevfs, | |
584 | } |
|
584 | } | |
585 | # we keep repo.vfs out of the on purpose, ther are too many danger there |
|
585 | # we keep repo.vfs out of the on purpose, ther are too many danger there | |
586 | # (eg: .hg/hgrc) |
|
586 | # (eg: .hg/hgrc) | |
587 | assert repo.vfs not in vfsmap.values() |
|
587 | assert repo.vfs not in vfsmap.values() | |
588 |
|
588 | |||
589 | return vfsmap |
|
589 | return vfsmap | |
590 |
|
590 | |||
591 |
|
591 | |||
592 | def _emit2(repo, entries, totalfilesize): |
|
592 | def _emit2(repo, entries, totalfilesize): | |
593 | """actually emit the stream bundle""" |
|
593 | """actually emit the stream bundle""" | |
594 | vfsmap = _makemap(repo) |
|
594 | vfsmap = _makemap(repo) | |
595 | # we keep repo.vfs out of the on purpose, ther are too many danger there |
|
595 | # we keep repo.vfs out of the on purpose, ther are too many danger there | |
596 | # (eg: .hg/hgrc), |
|
596 | # (eg: .hg/hgrc), | |
597 | # |
|
597 | # | |
598 | # this assert is duplicated (from _makemap) as author might think this is |
|
598 | # this assert is duplicated (from _makemap) as author might think this is | |
599 | # fine, while this is really not fine. |
|
599 | # fine, while this is really not fine. | |
600 | if repo.vfs in vfsmap.values(): |
|
600 | if repo.vfs in vfsmap.values(): | |
601 | raise error.ProgrammingError( |
|
601 | raise error.ProgrammingError( | |
602 | b'repo.vfs must not be added to vfsmap for security reasons' |
|
602 | b'repo.vfs must not be added to vfsmap for security reasons' | |
603 | ) |
|
603 | ) | |
604 |
|
604 | |||
605 | progress = repo.ui.makeprogress( |
|
605 | progress = repo.ui.makeprogress( | |
606 | _(b'bundle'), total=totalfilesize, unit=_(b'bytes') |
|
606 | _(b'bundle'), total=totalfilesize, unit=_(b'bytes') | |
607 | ) |
|
607 | ) | |
608 | progress.update(0) |
|
608 | progress.update(0) | |
609 | with maketempcopies() as copy, progress: |
|
609 | with maketempcopies() as copy, progress: | |
610 | # copy is delayed until we are in the try |
|
610 | # copy is delayed until we are in the try | |
611 | entries = [_filterfull(e, copy, vfsmap) for e in entries] |
|
611 | entries = [_filterfull(e, copy, vfsmap) for e in entries] | |
612 | yield None # this release the lock on the repository |
|
612 | yield None # this release the lock on the repository | |
613 | totalbytecount = 0 |
|
613 | totalbytecount = 0 | |
614 |
|
614 | |||
615 | for src, name, ftype, data in entries: |
|
615 | for src, name, ftype, data in entries: | |
616 | vfs = vfsmap[src] |
|
616 | vfs = vfsmap[src] | |
617 | yield src |
|
617 | yield src | |
618 | yield util.uvarintencode(len(name)) |
|
618 | yield util.uvarintencode(len(name)) | |
619 | if ftype == _fileappend: |
|
619 | if ftype == _fileappend: | |
620 | fp = vfs(name) |
|
620 | fp = vfs(name) | |
621 | size = data |
|
621 | size = data | |
622 | elif ftype == _filefull: |
|
622 | elif ftype == _filefull: | |
623 | fp = open(data, b'rb') |
|
623 | fp = open(data, b'rb') | |
624 | size = util.fstat(fp).st_size |
|
624 | size = util.fstat(fp).st_size | |
625 | bytecount = 0 |
|
625 | bytecount = 0 | |
626 | try: |
|
626 | try: | |
627 | yield util.uvarintencode(size) |
|
627 | yield util.uvarintencode(size) | |
628 | yield name |
|
628 | yield name | |
629 | if size <= 65536: |
|
629 | if size <= 65536: | |
630 | chunks = (fp.read(size),) |
|
630 | chunks = (fp.read(size),) | |
631 | else: |
|
631 | else: | |
632 | chunks = util.filechunkiter(fp, limit=size) |
|
632 | chunks = util.filechunkiter(fp, limit=size) | |
633 | for chunk in chunks: |
|
633 | for chunk in chunks: | |
634 | bytecount += len(chunk) |
|
634 | bytecount += len(chunk) | |
635 | totalbytecount += len(chunk) |
|
635 | totalbytecount += len(chunk) | |
636 | progress.update(totalbytecount) |
|
636 | progress.update(totalbytecount) | |
637 | yield chunk |
|
637 | yield chunk | |
638 | if bytecount != size: |
|
638 | if bytecount != size: | |
639 | # Would most likely be caused by a race due to `hg strip` or |
|
639 | # Would most likely be caused by a race due to `hg strip` or | |
640 | # a revlog split |
|
640 | # a revlog split | |
641 | raise error.Abort( |
|
641 | raise error.Abort( | |
642 | _( |
|
642 | _( | |
643 | b'clone could only read %d bytes from %s, but ' |
|
643 | b'clone could only read %d bytes from %s, but ' | |
644 | b'expected %d bytes' |
|
644 | b'expected %d bytes' | |
645 | ) |
|
645 | ) | |
646 | % (bytecount, name, size) |
|
646 | % (bytecount, name, size) | |
647 | ) |
|
647 | ) | |
648 | finally: |
|
648 | finally: | |
649 | fp.close() |
|
649 | fp.close() | |
650 |
|
650 | |||
651 |
|
651 | |||
652 | def _test_sync_point_walk_1(repo): |
|
652 | def _test_sync_point_walk_1(repo): | |
653 | """a function for synchronisation during tests""" |
|
653 | """a function for synchronisation during tests""" | |
654 |
|
654 | |||
655 |
|
655 | |||
656 | def _test_sync_point_walk_2(repo): |
|
656 | def _test_sync_point_walk_2(repo): | |
657 | """a function for synchronisation during tests""" |
|
657 | """a function for synchronisation during tests""" | |
658 |
|
658 | |||
659 |
|
659 | |||
660 | def _v2_walk(repo, includes, excludes, includeobsmarkers): |
|
660 | def _v2_walk(repo, includes, excludes, includeobsmarkers): | |
661 | """emit a seris of files information useful to clone a repo |
|
661 | """emit a seris of files information useful to clone a repo | |
662 |
|
662 | |||
663 | return (entries, totalfilesize) |
|
663 | return (entries, totalfilesize) | |
664 |
|
664 | |||
665 | entries is a list of tuple (vfs-key, file-path, file-type, size) |
|
665 | entries is a list of tuple (vfs-key, file-path, file-type, size) | |
666 |
|
666 | |||
667 | - `vfs-key`: is a key to the right vfs to write the file (see _makemap) |
|
667 | - `vfs-key`: is a key to the right vfs to write the file (see _makemap) | |
668 | - `name`: file path of the file to copy (to be feed to the vfss) |
|
668 | - `name`: file path of the file to copy (to be feed to the vfss) | |
669 | - `file-type`: do this file need to be copied with the source lock ? |
|
669 | - `file-type`: do this file need to be copied with the source lock ? | |
670 | - `size`: the size of the file (or None) |
|
670 | - `size`: the size of the file (or None) | |
671 | """ |
|
671 | """ | |
672 | assert repo._currentlock(repo._lockref) is not None |
|
672 | assert repo._currentlock(repo._lockref) is not None | |
673 | entries = [] |
|
673 | entries = [] | |
674 | totalfilesize = 0 |
|
674 | totalfilesize = 0 | |
675 |
|
675 | |||
676 | matcher = None |
|
676 | matcher = None | |
677 | if includes or excludes: |
|
677 | if includes or excludes: | |
678 | matcher = narrowspec.match(repo.root, includes, excludes) |
|
678 | matcher = narrowspec.match(repo.root, includes, excludes) | |
679 |
|
679 | |||
680 |
for |
|
680 | for entry in _walkstreamfiles(repo, matcher): | |
681 | if size: |
|
681 | if entry.file_size: | |
682 | ft = _fileappend |
|
682 | ft = _fileappend | |
683 | if rl_type & store.FILEFLAGS_VOLATILE: |
|
683 | if entry.is_volatile: | |
684 | ft = _filefull |
|
684 | ft = _filefull | |
685 |
entries.append( |
|
685 | entries.append( | |
686 | totalfilesize += size |
|
686 | (_srcstore, entry.unencoded_path, ft, entry.file_size) | |
|
687 | ) | |||
|
688 | totalfilesize += entry.file_size | |||
687 | for name in _walkstreamfullstorefiles(repo): |
|
689 | for name in _walkstreamfullstorefiles(repo): | |
688 | if repo.svfs.exists(name): |
|
690 | if repo.svfs.exists(name): | |
689 | totalfilesize += repo.svfs.lstat(name).st_size |
|
691 | totalfilesize += repo.svfs.lstat(name).st_size | |
690 | entries.append((_srcstore, name, _filefull, None)) |
|
692 | entries.append((_srcstore, name, _filefull, None)) | |
691 | if includeobsmarkers and repo.svfs.exists(b'obsstore'): |
|
693 | if includeobsmarkers and repo.svfs.exists(b'obsstore'): | |
692 | totalfilesize += repo.svfs.lstat(b'obsstore').st_size |
|
694 | totalfilesize += repo.svfs.lstat(b'obsstore').st_size | |
693 | entries.append((_srcstore, b'obsstore', _filefull, None)) |
|
695 | entries.append((_srcstore, b'obsstore', _filefull, None)) | |
694 | for name in cacheutil.cachetocopy(repo): |
|
696 | for name in cacheutil.cachetocopy(repo): | |
695 | if repo.cachevfs.exists(name): |
|
697 | if repo.cachevfs.exists(name): | |
696 | totalfilesize += repo.cachevfs.lstat(name).st_size |
|
698 | totalfilesize += repo.cachevfs.lstat(name).st_size | |
697 | entries.append((_srccache, name, _filefull, None)) |
|
699 | entries.append((_srccache, name, _filefull, None)) | |
698 | return entries, totalfilesize |
|
700 | return entries, totalfilesize | |
699 |
|
701 | |||
700 |
|
702 | |||
701 | def generatev2(repo, includes, excludes, includeobsmarkers): |
|
703 | def generatev2(repo, includes, excludes, includeobsmarkers): | |
702 | """Emit content for version 2 of a streaming clone. |
|
704 | """Emit content for version 2 of a streaming clone. | |
703 |
|
705 | |||
704 | the data stream consists the following entries: |
|
706 | the data stream consists the following entries: | |
705 | 1) A char representing the file destination (eg: store or cache) |
|
707 | 1) A char representing the file destination (eg: store or cache) | |
706 | 2) A varint containing the length of the filename |
|
708 | 2) A varint containing the length of the filename | |
707 | 3) A varint containing the length of file data |
|
709 | 3) A varint containing the length of file data | |
708 | 4) N bytes containing the filename (the internal, store-agnostic form) |
|
710 | 4) N bytes containing the filename (the internal, store-agnostic form) | |
709 | 5) N bytes containing the file data |
|
711 | 5) N bytes containing the file data | |
710 |
|
712 | |||
711 | Returns a 3-tuple of (file count, file size, data iterator). |
|
713 | Returns a 3-tuple of (file count, file size, data iterator). | |
712 | """ |
|
714 | """ | |
713 |
|
715 | |||
714 | with repo.lock(): |
|
716 | with repo.lock(): | |
715 |
|
717 | |||
716 | repo.ui.debug(b'scanning\n') |
|
718 | repo.ui.debug(b'scanning\n') | |
717 |
|
719 | |||
718 | entries, totalfilesize = _v2_walk( |
|
720 | entries, totalfilesize = _v2_walk( | |
719 | repo, |
|
721 | repo, | |
720 | includes=includes, |
|
722 | includes=includes, | |
721 | excludes=excludes, |
|
723 | excludes=excludes, | |
722 | includeobsmarkers=includeobsmarkers, |
|
724 | includeobsmarkers=includeobsmarkers, | |
723 | ) |
|
725 | ) | |
724 |
|
726 | |||
725 | chunks = _emit2(repo, entries, totalfilesize) |
|
727 | chunks = _emit2(repo, entries, totalfilesize) | |
726 | first = next(chunks) |
|
728 | first = next(chunks) | |
727 | assert first is None |
|
729 | assert first is None | |
728 | _test_sync_point_walk_1(repo) |
|
730 | _test_sync_point_walk_1(repo) | |
729 | _test_sync_point_walk_2(repo) |
|
731 | _test_sync_point_walk_2(repo) | |
730 |
|
732 | |||
731 | return len(entries), totalfilesize, chunks |
|
733 | return len(entries), totalfilesize, chunks | |
732 |
|
734 | |||
733 |
|
735 | |||
734 | @contextlib.contextmanager |
|
736 | @contextlib.contextmanager | |
735 | def nested(*ctxs): |
|
737 | def nested(*ctxs): | |
736 | this = ctxs[0] |
|
738 | this = ctxs[0] | |
737 | rest = ctxs[1:] |
|
739 | rest = ctxs[1:] | |
738 | with this: |
|
740 | with this: | |
739 | if rest: |
|
741 | if rest: | |
740 | with nested(*rest): |
|
742 | with nested(*rest): | |
741 | yield |
|
743 | yield | |
742 | else: |
|
744 | else: | |
743 | yield |
|
745 | yield | |
744 |
|
746 | |||
745 |
|
747 | |||
746 | def consumev2(repo, fp, filecount, filesize): |
|
748 | def consumev2(repo, fp, filecount, filesize): | |
747 | """Apply the contents from a version 2 streaming clone. |
|
749 | """Apply the contents from a version 2 streaming clone. | |
748 |
|
750 | |||
749 | Data is read from an object that only needs to provide a ``read(size)`` |
|
751 | Data is read from an object that only needs to provide a ``read(size)`` | |
750 | method. |
|
752 | method. | |
751 | """ |
|
753 | """ | |
752 | with repo.lock(): |
|
754 | with repo.lock(): | |
753 | repo.ui.status( |
|
755 | repo.ui.status( | |
754 | _(b'%d files to transfer, %s of data\n') |
|
756 | _(b'%d files to transfer, %s of data\n') | |
755 | % (filecount, util.bytecount(filesize)) |
|
757 | % (filecount, util.bytecount(filesize)) | |
756 | ) |
|
758 | ) | |
757 |
|
759 | |||
758 | start = util.timer() |
|
760 | start = util.timer() | |
759 | progress = repo.ui.makeprogress( |
|
761 | progress = repo.ui.makeprogress( | |
760 | _(b'clone'), total=filesize, unit=_(b'bytes') |
|
762 | _(b'clone'), total=filesize, unit=_(b'bytes') | |
761 | ) |
|
763 | ) | |
762 | progress.update(0) |
|
764 | progress.update(0) | |
763 |
|
765 | |||
764 | vfsmap = _makemap(repo) |
|
766 | vfsmap = _makemap(repo) | |
765 | # we keep repo.vfs out of the on purpose, ther are too many danger |
|
767 | # we keep repo.vfs out of the on purpose, ther are too many danger | |
766 | # there (eg: .hg/hgrc), |
|
768 | # there (eg: .hg/hgrc), | |
767 | # |
|
769 | # | |
768 | # this assert is duplicated (from _makemap) as author might think this |
|
770 | # this assert is duplicated (from _makemap) as author might think this | |
769 | # is fine, while this is really not fine. |
|
771 | # is fine, while this is really not fine. | |
770 | if repo.vfs in vfsmap.values(): |
|
772 | if repo.vfs in vfsmap.values(): | |
771 | raise error.ProgrammingError( |
|
773 | raise error.ProgrammingError( | |
772 | b'repo.vfs must not be added to vfsmap for security reasons' |
|
774 | b'repo.vfs must not be added to vfsmap for security reasons' | |
773 | ) |
|
775 | ) | |
774 |
|
776 | |||
775 | with repo.transaction(b'clone'): |
|
777 | with repo.transaction(b'clone'): | |
776 | ctxs = (vfs.backgroundclosing(repo.ui) for vfs in vfsmap.values()) |
|
778 | ctxs = (vfs.backgroundclosing(repo.ui) for vfs in vfsmap.values()) | |
777 | with nested(*ctxs): |
|
779 | with nested(*ctxs): | |
778 | for i in range(filecount): |
|
780 | for i in range(filecount): | |
779 | src = util.readexactly(fp, 1) |
|
781 | src = util.readexactly(fp, 1) | |
780 | vfs = vfsmap[src] |
|
782 | vfs = vfsmap[src] | |
781 | namelen = util.uvarintdecodestream(fp) |
|
783 | namelen = util.uvarintdecodestream(fp) | |
782 | datalen = util.uvarintdecodestream(fp) |
|
784 | datalen = util.uvarintdecodestream(fp) | |
783 |
|
785 | |||
784 | name = util.readexactly(fp, namelen) |
|
786 | name = util.readexactly(fp, namelen) | |
785 |
|
787 | |||
786 | if repo.ui.debugflag: |
|
788 | if repo.ui.debugflag: | |
787 | repo.ui.debug( |
|
789 | repo.ui.debug( | |
788 | b'adding [%s] %s (%s)\n' |
|
790 | b'adding [%s] %s (%s)\n' | |
789 | % (src, name, util.bytecount(datalen)) |
|
791 | % (src, name, util.bytecount(datalen)) | |
790 | ) |
|
792 | ) | |
791 |
|
793 | |||
792 | with vfs(name, b'w') as ofp: |
|
794 | with vfs(name, b'w') as ofp: | |
793 | for chunk in util.filechunkiter(fp, limit=datalen): |
|
795 | for chunk in util.filechunkiter(fp, limit=datalen): | |
794 | progress.increment(step=len(chunk)) |
|
796 | progress.increment(step=len(chunk)) | |
795 | ofp.write(chunk) |
|
797 | ofp.write(chunk) | |
796 |
|
798 | |||
797 | # force @filecache properties to be reloaded from |
|
799 | # force @filecache properties to be reloaded from | |
798 | # streamclone-ed file at next access |
|
800 | # streamclone-ed file at next access | |
799 | repo.invalidate(clearfilecache=True) |
|
801 | repo.invalidate(clearfilecache=True) | |
800 |
|
802 | |||
801 | elapsed = util.timer() - start |
|
803 | elapsed = util.timer() - start | |
802 | if elapsed <= 0: |
|
804 | if elapsed <= 0: | |
803 | elapsed = 0.001 |
|
805 | elapsed = 0.001 | |
804 | repo.ui.status( |
|
806 | repo.ui.status( | |
805 | _(b'transferred %s in %.1f seconds (%s/sec)\n') |
|
807 | _(b'transferred %s in %.1f seconds (%s/sec)\n') | |
806 | % ( |
|
808 | % ( | |
807 | util.bytecount(progress.pos), |
|
809 | util.bytecount(progress.pos), | |
808 | elapsed, |
|
810 | elapsed, | |
809 | util.bytecount(progress.pos / elapsed), |
|
811 | util.bytecount(progress.pos / elapsed), | |
810 | ) |
|
812 | ) | |
811 | ) |
|
813 | ) | |
812 | progress.complete() |
|
814 | progress.complete() | |
813 |
|
815 | |||
814 |
|
816 | |||
815 | def applybundlev2(repo, fp, filecount, filesize, requirements): |
|
817 | def applybundlev2(repo, fp, filecount, filesize, requirements): | |
816 | from . import localrepo |
|
818 | from . import localrepo | |
817 |
|
819 | |||
818 | missingreqs = [r for r in requirements if r not in repo.supported] |
|
820 | missingreqs = [r for r in requirements if r not in repo.supported] | |
819 | if missingreqs: |
|
821 | if missingreqs: | |
820 | raise error.Abort( |
|
822 | raise error.Abort( | |
821 | _(b'unable to apply stream clone: unsupported format: %s') |
|
823 | _(b'unable to apply stream clone: unsupported format: %s') | |
822 | % b', '.join(sorted(missingreqs)) |
|
824 | % b', '.join(sorted(missingreqs)) | |
823 | ) |
|
825 | ) | |
824 |
|
826 | |||
825 | consumev2(repo, fp, filecount, filesize) |
|
827 | consumev2(repo, fp, filecount, filesize) | |
826 |
|
828 | |||
827 | repo.requirements = new_stream_clone_requirements( |
|
829 | repo.requirements = new_stream_clone_requirements( | |
828 | repo.requirements, |
|
830 | repo.requirements, | |
829 | requirements, |
|
831 | requirements, | |
830 | ) |
|
832 | ) | |
831 | repo.svfs.options = localrepo.resolvestorevfsoptions( |
|
833 | repo.svfs.options = localrepo.resolvestorevfsoptions( | |
832 | repo.ui, repo.requirements, repo.features |
|
834 | repo.ui, repo.requirements, repo.features | |
833 | ) |
|
835 | ) | |
834 | scmutil.writereporequirements(repo) |
|
836 | scmutil.writereporequirements(repo) | |
835 | nodemap.post_stream_cleanup(repo) |
|
837 | nodemap.post_stream_cleanup(repo) | |
836 |
|
838 | |||
837 |
|
839 | |||
838 | def _copy_files(src_vfs_map, dst_vfs_map, entries, progress): |
|
840 | def _copy_files(src_vfs_map, dst_vfs_map, entries, progress): | |
839 | hardlink = [True] |
|
841 | hardlink = [True] | |
840 |
|
842 | |||
841 | def copy_used(): |
|
843 | def copy_used(): | |
842 | hardlink[0] = False |
|
844 | hardlink[0] = False | |
843 | progress.topic = _(b'copying') |
|
845 | progress.topic = _(b'copying') | |
844 |
|
846 | |||
845 | for k, path, size in entries: |
|
847 | for k, path, size in entries: | |
846 | src_vfs = src_vfs_map[k] |
|
848 | src_vfs = src_vfs_map[k] | |
847 | dst_vfs = dst_vfs_map[k] |
|
849 | dst_vfs = dst_vfs_map[k] | |
848 | src_path = src_vfs.join(path) |
|
850 | src_path = src_vfs.join(path) | |
849 | dst_path = dst_vfs.join(path) |
|
851 | dst_path = dst_vfs.join(path) | |
850 | # We cannot use dirname and makedirs of dst_vfs here because the store |
|
852 | # We cannot use dirname and makedirs of dst_vfs here because the store | |
851 | # encoding confuses them. See issue 6581 for details. |
|
853 | # encoding confuses them. See issue 6581 for details. | |
852 | dirname = os.path.dirname(dst_path) |
|
854 | dirname = os.path.dirname(dst_path) | |
853 | if not os.path.exists(dirname): |
|
855 | if not os.path.exists(dirname): | |
854 | util.makedirs(dirname) |
|
856 | util.makedirs(dirname) | |
855 | dst_vfs.register_file(path) |
|
857 | dst_vfs.register_file(path) | |
856 | # XXX we could use the #nb_bytes argument. |
|
858 | # XXX we could use the #nb_bytes argument. | |
857 | util.copyfile( |
|
859 | util.copyfile( | |
858 | src_path, |
|
860 | src_path, | |
859 | dst_path, |
|
861 | dst_path, | |
860 | hardlink=hardlink[0], |
|
862 | hardlink=hardlink[0], | |
861 | no_hardlink_cb=copy_used, |
|
863 | no_hardlink_cb=copy_used, | |
862 | check_fs_hardlink=False, |
|
864 | check_fs_hardlink=False, | |
863 | ) |
|
865 | ) | |
864 | progress.increment() |
|
866 | progress.increment() | |
865 | return hardlink[0] |
|
867 | return hardlink[0] | |
866 |
|
868 | |||
867 |
|
869 | |||
868 | def local_copy(src_repo, dest_repo): |
|
870 | def local_copy(src_repo, dest_repo): | |
869 | """copy all content from one local repository to another |
|
871 | """copy all content from one local repository to another | |
870 |
|
872 | |||
871 | This is useful for local clone""" |
|
873 | This is useful for local clone""" | |
872 | src_store_requirements = { |
|
874 | src_store_requirements = { | |
873 | r |
|
875 | r | |
874 | for r in src_repo.requirements |
|
876 | for r in src_repo.requirements | |
875 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS |
|
877 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS | |
876 | } |
|
878 | } | |
877 | dest_store_requirements = { |
|
879 | dest_store_requirements = { | |
878 | r |
|
880 | r | |
879 | for r in dest_repo.requirements |
|
881 | for r in dest_repo.requirements | |
880 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS |
|
882 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS | |
881 | } |
|
883 | } | |
882 | assert src_store_requirements == dest_store_requirements |
|
884 | assert src_store_requirements == dest_store_requirements | |
883 |
|
885 | |||
884 | with dest_repo.lock(): |
|
886 | with dest_repo.lock(): | |
885 | with src_repo.lock(): |
|
887 | with src_repo.lock(): | |
886 |
|
888 | |||
887 | # bookmark is not integrated to the streaming as it might use the |
|
889 | # bookmark is not integrated to the streaming as it might use the | |
888 | # `repo.vfs` and they are too many sentitive data accessible |
|
890 | # `repo.vfs` and they are too many sentitive data accessible | |
889 | # through `repo.vfs` to expose it to streaming clone. |
|
891 | # through `repo.vfs` to expose it to streaming clone. | |
890 | src_book_vfs = bookmarks.bookmarksvfs(src_repo) |
|
892 | src_book_vfs = bookmarks.bookmarksvfs(src_repo) | |
891 | srcbookmarks = src_book_vfs.join(b'bookmarks') |
|
893 | srcbookmarks = src_book_vfs.join(b'bookmarks') | |
892 | bm_count = 0 |
|
894 | bm_count = 0 | |
893 | if os.path.exists(srcbookmarks): |
|
895 | if os.path.exists(srcbookmarks): | |
894 | bm_count = 1 |
|
896 | bm_count = 1 | |
895 |
|
897 | |||
896 | entries, totalfilesize = _v2_walk( |
|
898 | entries, totalfilesize = _v2_walk( | |
897 | src_repo, |
|
899 | src_repo, | |
898 | includes=None, |
|
900 | includes=None, | |
899 | excludes=None, |
|
901 | excludes=None, | |
900 | includeobsmarkers=True, |
|
902 | includeobsmarkers=True, | |
901 | ) |
|
903 | ) | |
902 | src_vfs_map = _makemap(src_repo) |
|
904 | src_vfs_map = _makemap(src_repo) | |
903 | dest_vfs_map = _makemap(dest_repo) |
|
905 | dest_vfs_map = _makemap(dest_repo) | |
904 | progress = src_repo.ui.makeprogress( |
|
906 | progress = src_repo.ui.makeprogress( | |
905 | topic=_(b'linking'), |
|
907 | topic=_(b'linking'), | |
906 | total=len(entries) + bm_count, |
|
908 | total=len(entries) + bm_count, | |
907 | unit=_(b'files'), |
|
909 | unit=_(b'files'), | |
908 | ) |
|
910 | ) | |
909 | # copy files |
|
911 | # copy files | |
910 | # |
|
912 | # | |
911 | # We could copy the full file while the source repository is locked |
|
913 | # We could copy the full file while the source repository is locked | |
912 | # and the other one without the lock. However, in the linking case, |
|
914 | # and the other one without the lock. However, in the linking case, | |
913 | # this would also requires checks that nobody is appending any data |
|
915 | # this would also requires checks that nobody is appending any data | |
914 | # to the files while we do the clone, so this is not done yet. We |
|
916 | # to the files while we do the clone, so this is not done yet. We | |
915 | # could do this blindly when copying files. |
|
917 | # could do this blindly when copying files. | |
916 | files = ((k, path, size) for k, path, ftype, size in entries) |
|
918 | files = ((k, path, size) for k, path, ftype, size in entries) | |
917 | hardlink = _copy_files(src_vfs_map, dest_vfs_map, files, progress) |
|
919 | hardlink = _copy_files(src_vfs_map, dest_vfs_map, files, progress) | |
918 |
|
920 | |||
919 | # copy bookmarks over |
|
921 | # copy bookmarks over | |
920 | if bm_count: |
|
922 | if bm_count: | |
921 | dst_book_vfs = bookmarks.bookmarksvfs(dest_repo) |
|
923 | dst_book_vfs = bookmarks.bookmarksvfs(dest_repo) | |
922 | dstbookmarks = dst_book_vfs.join(b'bookmarks') |
|
924 | dstbookmarks = dst_book_vfs.join(b'bookmarks') | |
923 | util.copyfile(srcbookmarks, dstbookmarks) |
|
925 | util.copyfile(srcbookmarks, dstbookmarks) | |
924 | progress.complete() |
|
926 | progress.complete() | |
925 | if hardlink: |
|
927 | if hardlink: | |
926 | msg = b'linked %d files\n' |
|
928 | msg = b'linked %d files\n' | |
927 | else: |
|
929 | else: | |
928 | msg = b'copied %d files\n' |
|
930 | msg = b'copied %d files\n' | |
929 | src_repo.ui.debug(msg % (len(entries) + bm_count)) |
|
931 | src_repo.ui.debug(msg % (len(entries) + bm_count)) | |
930 |
|
932 | |||
931 | with dest_repo.transaction(b"localclone") as tr: |
|
933 | with dest_repo.transaction(b"localclone") as tr: | |
932 | dest_repo.store.write(tr) |
|
934 | dest_repo.store.write(tr) | |
933 |
|
935 | |||
934 | # clean up transaction file as they do not make sense |
|
936 | # clean up transaction file as they do not make sense | |
935 | transaction.cleanup_undo_files(dest_repo.ui.warn, dest_repo.vfs_map) |
|
937 | transaction.cleanup_undo_files(dest_repo.ui.warn, dest_repo.vfs_map) |
@@ -1,689 +1,690 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 |
|
8 | |||
9 | import stat |
|
9 | import stat | |
10 |
|
10 | |||
11 | from ..i18n import _ |
|
11 | from ..i18n import _ | |
12 | from ..pycompat import getattr |
|
12 | from ..pycompat import getattr | |
13 | from .. import ( |
|
13 | from .. import ( | |
14 | changelog, |
|
14 | changelog, | |
15 | error, |
|
15 | error, | |
16 | filelog, |
|
16 | filelog, | |
17 | manifest, |
|
17 | manifest, | |
18 | metadata, |
|
18 | metadata, | |
19 | pycompat, |
|
19 | pycompat, | |
20 | requirements, |
|
20 | requirements, | |
21 | scmutil, |
|
21 | scmutil, | |
22 | store, |
|
22 | store, | |
23 | util, |
|
23 | util, | |
24 | vfs as vfsmod, |
|
24 | vfs as vfsmod, | |
25 | ) |
|
25 | ) | |
26 | from ..revlogutils import ( |
|
26 | from ..revlogutils import ( | |
27 | constants as revlogconst, |
|
27 | constants as revlogconst, | |
28 | flagutil, |
|
28 | flagutil, | |
29 | nodemap, |
|
29 | nodemap, | |
30 | sidedata as sidedatamod, |
|
30 | sidedata as sidedatamod, | |
31 | ) |
|
31 | ) | |
32 | from . import actions as upgrade_actions |
|
32 | from . import actions as upgrade_actions | |
33 |
|
33 | |||
34 |
|
34 | |||
35 | def get_sidedata_helpers(srcrepo, dstrepo): |
|
35 | def get_sidedata_helpers(srcrepo, dstrepo): | |
36 | use_w = srcrepo.ui.configbool(b'experimental', b'worker.repository-upgrade') |
|
36 | use_w = srcrepo.ui.configbool(b'experimental', b'worker.repository-upgrade') | |
37 | sequential = pycompat.iswindows or not use_w |
|
37 | sequential = pycompat.iswindows or not use_w | |
38 | if not sequential: |
|
38 | if not sequential: | |
39 | srcrepo.register_sidedata_computer( |
|
39 | srcrepo.register_sidedata_computer( | |
40 | revlogconst.KIND_CHANGELOG, |
|
40 | revlogconst.KIND_CHANGELOG, | |
41 | sidedatamod.SD_FILES, |
|
41 | sidedatamod.SD_FILES, | |
42 | (sidedatamod.SD_FILES,), |
|
42 | (sidedatamod.SD_FILES,), | |
43 | metadata._get_worker_sidedata_adder(srcrepo, dstrepo), |
|
43 | metadata._get_worker_sidedata_adder(srcrepo, dstrepo), | |
44 | flagutil.REVIDX_HASCOPIESINFO, |
|
44 | flagutil.REVIDX_HASCOPIESINFO, | |
45 | replace=True, |
|
45 | replace=True, | |
46 | ) |
|
46 | ) | |
47 | return sidedatamod.get_sidedata_helpers(srcrepo, dstrepo._wanted_sidedata) |
|
47 | return sidedatamod.get_sidedata_helpers(srcrepo, dstrepo._wanted_sidedata) | |
48 |
|
48 | |||
49 |
|
49 | |||
50 | def _revlogfrompath(repo, rl_type, path): |
|
50 | def _revlogfrompath(repo, rl_type, path): | |
51 | """Obtain a revlog from a repo path. |
|
51 | """Obtain a revlog from a repo path. | |
52 |
|
52 | |||
53 | An instance of the appropriate class is returned. |
|
53 | An instance of the appropriate class is returned. | |
54 | """ |
|
54 | """ | |
55 | if rl_type & store.FILEFLAGS_CHANGELOG: |
|
55 | if rl_type & store.FILEFLAGS_CHANGELOG: | |
56 | return changelog.changelog(repo.svfs) |
|
56 | return changelog.changelog(repo.svfs) | |
57 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: |
|
57 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: | |
58 | mandir = b'' |
|
58 | mandir = b'' | |
59 | if b'/' in path: |
|
59 | if b'/' in path: | |
60 | mandir = path.rsplit(b'/', 1)[0] |
|
60 | mandir = path.rsplit(b'/', 1)[0] | |
61 | return manifest.manifestrevlog( |
|
61 | return manifest.manifestrevlog( | |
62 | repo.nodeconstants, repo.svfs, tree=mandir |
|
62 | repo.nodeconstants, repo.svfs, tree=mandir | |
63 | ) |
|
63 | ) | |
64 | else: |
|
64 | else: | |
65 | # drop the extension and the `data/` prefix |
|
65 | # drop the extension and the `data/` prefix | |
66 | path_part = path.rsplit(b'.', 1)[0].split(b'/', 1) |
|
66 | path_part = path.rsplit(b'.', 1)[0].split(b'/', 1) | |
67 | if len(path_part) < 2: |
|
67 | if len(path_part) < 2: | |
68 | msg = _(b'cannot recognize revlog from filename: %s') |
|
68 | msg = _(b'cannot recognize revlog from filename: %s') | |
69 | msg %= path |
|
69 | msg %= path | |
70 | raise error.Abort(msg) |
|
70 | raise error.Abort(msg) | |
71 | path = path_part[1] |
|
71 | path = path_part[1] | |
72 | return filelog.filelog(repo.svfs, path) |
|
72 | return filelog.filelog(repo.svfs, path) | |
73 |
|
73 | |||
74 |
|
74 | |||
75 | def _copyrevlog(tr, destrepo, oldrl, rl_type, unencodedname): |
|
75 | def _copyrevlog(tr, destrepo, oldrl, rl_type, unencodedname): | |
76 | """copy all relevant files for `oldrl` into `destrepo` store |
|
76 | """copy all relevant files for `oldrl` into `destrepo` store | |
77 |
|
77 | |||
78 | Files are copied "as is" without any transformation. The copy is performed |
|
78 | Files are copied "as is" without any transformation. The copy is performed | |
79 | without extra checks. Callers are responsible for making sure the copied |
|
79 | without extra checks. Callers are responsible for making sure the copied | |
80 | content is compatible with format of the destination repository. |
|
80 | content is compatible with format of the destination repository. | |
81 | """ |
|
81 | """ | |
82 | oldrl = getattr(oldrl, '_revlog', oldrl) |
|
82 | oldrl = getattr(oldrl, '_revlog', oldrl) | |
83 | newrl = _revlogfrompath(destrepo, rl_type, unencodedname) |
|
83 | newrl = _revlogfrompath(destrepo, rl_type, unencodedname) | |
84 | newrl = getattr(newrl, '_revlog', newrl) |
|
84 | newrl = getattr(newrl, '_revlog', newrl) | |
85 |
|
85 | |||
86 | oldvfs = oldrl.opener |
|
86 | oldvfs = oldrl.opener | |
87 | newvfs = newrl.opener |
|
87 | newvfs = newrl.opener | |
88 | oldindex = oldvfs.join(oldrl._indexfile) |
|
88 | oldindex = oldvfs.join(oldrl._indexfile) | |
89 | newindex = newvfs.join(newrl._indexfile) |
|
89 | newindex = newvfs.join(newrl._indexfile) | |
90 | olddata = oldvfs.join(oldrl._datafile) |
|
90 | olddata = oldvfs.join(oldrl._datafile) | |
91 | newdata = newvfs.join(newrl._datafile) |
|
91 | newdata = newvfs.join(newrl._datafile) | |
92 |
|
92 | |||
93 | with newvfs(newrl._indexfile, b'w'): |
|
93 | with newvfs(newrl._indexfile, b'w'): | |
94 | pass # create all the directories |
|
94 | pass # create all the directories | |
95 |
|
95 | |||
96 | util.copyfile(oldindex, newindex) |
|
96 | util.copyfile(oldindex, newindex) | |
97 | copydata = oldrl.opener.exists(oldrl._datafile) |
|
97 | copydata = oldrl.opener.exists(oldrl._datafile) | |
98 | if copydata: |
|
98 | if copydata: | |
99 | util.copyfile(olddata, newdata) |
|
99 | util.copyfile(olddata, newdata) | |
100 |
|
100 | |||
101 | if rl_type & store.FILEFLAGS_FILELOG: |
|
101 | if rl_type & store.FILEFLAGS_FILELOG: | |
102 | destrepo.svfs.fncache.add(unencodedname) |
|
102 | destrepo.svfs.fncache.add(unencodedname) | |
103 | if copydata: |
|
103 | if copydata: | |
104 | destrepo.svfs.fncache.add(unencodedname[:-2] + b'.d') |
|
104 | destrepo.svfs.fncache.add(unencodedname[:-2] + b'.d') | |
105 |
|
105 | |||
106 |
|
106 | |||
107 | UPGRADE_CHANGELOG = b"changelog" |
|
107 | UPGRADE_CHANGELOG = b"changelog" | |
108 | UPGRADE_MANIFEST = b"manifest" |
|
108 | UPGRADE_MANIFEST = b"manifest" | |
109 | UPGRADE_FILELOGS = b"all-filelogs" |
|
109 | UPGRADE_FILELOGS = b"all-filelogs" | |
110 |
|
110 | |||
111 | UPGRADE_ALL_REVLOGS = frozenset( |
|
111 | UPGRADE_ALL_REVLOGS = frozenset( | |
112 | [UPGRADE_CHANGELOG, UPGRADE_MANIFEST, UPGRADE_FILELOGS] |
|
112 | [UPGRADE_CHANGELOG, UPGRADE_MANIFEST, UPGRADE_FILELOGS] | |
113 | ) |
|
113 | ) | |
114 |
|
114 | |||
115 |
|
115 | |||
116 | def matchrevlog(revlogfilter, rl_type): |
|
116 | def matchrevlog(revlogfilter, rl_type): | |
117 | """check if a revlog is selected for cloning. |
|
117 | """check if a revlog is selected for cloning. | |
118 |
|
118 | |||
119 | In other words, are there any updates which need to be done on revlog |
|
119 | In other words, are there any updates which need to be done on revlog | |
120 | or it can be blindly copied. |
|
120 | or it can be blindly copied. | |
121 |
|
121 | |||
122 | The store entry is checked against the passed filter""" |
|
122 | The store entry is checked against the passed filter""" | |
123 | if rl_type & store.FILEFLAGS_CHANGELOG: |
|
123 | if rl_type & store.FILEFLAGS_CHANGELOG: | |
124 | return UPGRADE_CHANGELOG in revlogfilter |
|
124 | return UPGRADE_CHANGELOG in revlogfilter | |
125 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: |
|
125 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: | |
126 | return UPGRADE_MANIFEST in revlogfilter |
|
126 | return UPGRADE_MANIFEST in revlogfilter | |
127 | assert rl_type & store.FILEFLAGS_FILELOG |
|
127 | assert rl_type & store.FILEFLAGS_FILELOG | |
128 | return UPGRADE_FILELOGS in revlogfilter |
|
128 | return UPGRADE_FILELOGS in revlogfilter | |
129 |
|
129 | |||
130 |
|
130 | |||
131 | def _perform_clone( |
|
131 | def _perform_clone( | |
132 | ui, |
|
132 | ui, | |
133 | dstrepo, |
|
133 | dstrepo, | |
134 | tr, |
|
134 | tr, | |
135 | old_revlog, |
|
135 | old_revlog, | |
136 | rl_type, |
|
136 | rl_type, | |
137 | unencoded, |
|
137 | unencoded, | |
138 | upgrade_op, |
|
138 | upgrade_op, | |
139 | sidedata_helpers, |
|
139 | sidedata_helpers, | |
140 | oncopiedrevision, |
|
140 | oncopiedrevision, | |
141 | ): |
|
141 | ): | |
142 | """returns the new revlog object created""" |
|
142 | """returns the new revlog object created""" | |
143 | newrl = None |
|
143 | newrl = None | |
144 | if matchrevlog(upgrade_op.revlogs_to_process, rl_type): |
|
144 | if matchrevlog(upgrade_op.revlogs_to_process, rl_type): | |
145 | ui.note( |
|
145 | ui.note( | |
146 | _(b'cloning %d revisions from %s\n') % (len(old_revlog), unencoded) |
|
146 | _(b'cloning %d revisions from %s\n') % (len(old_revlog), unencoded) | |
147 | ) |
|
147 | ) | |
148 | newrl = _revlogfrompath(dstrepo, rl_type, unencoded) |
|
148 | newrl = _revlogfrompath(dstrepo, rl_type, unencoded) | |
149 | old_revlog.clone( |
|
149 | old_revlog.clone( | |
150 | tr, |
|
150 | tr, | |
151 | newrl, |
|
151 | newrl, | |
152 | addrevisioncb=oncopiedrevision, |
|
152 | addrevisioncb=oncopiedrevision, | |
153 | deltareuse=upgrade_op.delta_reuse_mode, |
|
153 | deltareuse=upgrade_op.delta_reuse_mode, | |
154 | forcedeltabothparents=upgrade_op.force_re_delta_both_parents, |
|
154 | forcedeltabothparents=upgrade_op.force_re_delta_both_parents, | |
155 | sidedata_helpers=sidedata_helpers, |
|
155 | sidedata_helpers=sidedata_helpers, | |
156 | ) |
|
156 | ) | |
157 | else: |
|
157 | else: | |
158 | msg = _(b'blindly copying %s containing %i revisions\n') |
|
158 | msg = _(b'blindly copying %s containing %i revisions\n') | |
159 | ui.note(msg % (unencoded, len(old_revlog))) |
|
159 | ui.note(msg % (unencoded, len(old_revlog))) | |
160 | _copyrevlog(tr, dstrepo, old_revlog, rl_type, unencoded) |
|
160 | _copyrevlog(tr, dstrepo, old_revlog, rl_type, unencoded) | |
161 |
|
161 | |||
162 | newrl = _revlogfrompath(dstrepo, rl_type, unencoded) |
|
162 | newrl = _revlogfrompath(dstrepo, rl_type, unencoded) | |
163 | return newrl |
|
163 | return newrl | |
164 |
|
164 | |||
165 |
|
165 | |||
166 | def _clonerevlogs( |
|
166 | def _clonerevlogs( | |
167 | ui, |
|
167 | ui, | |
168 | srcrepo, |
|
168 | srcrepo, | |
169 | dstrepo, |
|
169 | dstrepo, | |
170 | tr, |
|
170 | tr, | |
171 | upgrade_op, |
|
171 | upgrade_op, | |
172 | ): |
|
172 | ): | |
173 | """Copy revlogs between 2 repos.""" |
|
173 | """Copy revlogs between 2 repos.""" | |
174 | revcount = 0 |
|
174 | revcount = 0 | |
175 | srcsize = 0 |
|
175 | srcsize = 0 | |
176 | srcrawsize = 0 |
|
176 | srcrawsize = 0 | |
177 | dstsize = 0 |
|
177 | dstsize = 0 | |
178 | fcount = 0 |
|
178 | fcount = 0 | |
179 | frevcount = 0 |
|
179 | frevcount = 0 | |
180 | fsrcsize = 0 |
|
180 | fsrcsize = 0 | |
181 | frawsize = 0 |
|
181 | frawsize = 0 | |
182 | fdstsize = 0 |
|
182 | fdstsize = 0 | |
183 | mcount = 0 |
|
183 | mcount = 0 | |
184 | mrevcount = 0 |
|
184 | mrevcount = 0 | |
185 | msrcsize = 0 |
|
185 | msrcsize = 0 | |
186 | mrawsize = 0 |
|
186 | mrawsize = 0 | |
187 | mdstsize = 0 |
|
187 | mdstsize = 0 | |
188 | crevcount = 0 |
|
188 | crevcount = 0 | |
189 | csrcsize = 0 |
|
189 | csrcsize = 0 | |
190 | crawsize = 0 |
|
190 | crawsize = 0 | |
191 | cdstsize = 0 |
|
191 | cdstsize = 0 | |
192 |
|
192 | |||
193 | alldatafiles = list(srcrepo.store.walk()) |
|
193 | alldatafiles = list(srcrepo.store.walk()) | |
194 | # mapping of data files which needs to be cloned |
|
194 | # mapping of data files which needs to be cloned | |
195 | # key is unencoded filename |
|
195 | # key is unencoded filename | |
196 | # value is revlog_object_from_srcrepo |
|
196 | # value is revlog_object_from_srcrepo | |
197 | manifests = {} |
|
197 | manifests = {} | |
198 | changelogs = {} |
|
198 | changelogs = {} | |
199 | filelogs = {} |
|
199 | filelogs = {} | |
200 |
|
200 | |||
201 | # Perform a pass to collect metadata. This validates we can open all |
|
201 | # Perform a pass to collect metadata. This validates we can open all | |
202 | # source files and allows a unified progress bar to be displayed. |
|
202 | # source files and allows a unified progress bar to be displayed. | |
203 |
for r |
|
203 | for entry in alldatafiles: | |
204 | if not rl_type & store.FILEFLAGS_REVLOG_MAIN: |
|
204 | if not (entry.is_revlog and entry.is_revlog_main): | |
205 | continue |
|
205 | continue | |
|
206 | unencoded = entry.unencoded_path | |||
206 |
|
207 | |||
207 | # the store.walk function will wrongly pickup transaction backup and |
|
208 | # the store.walk function will wrongly pickup transaction backup and | |
208 | # 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. | |
209 | # (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 | |
210 | # hack together) |
|
211 | # hack together) | |
211 | skip_undo = ( |
|
212 | skip_undo = ( | |
212 | b'undo.backup.00changelog.i', |
|
213 | b'undo.backup.00changelog.i', | |
213 | b'undo.backup.00manifest.i', |
|
214 | b'undo.backup.00manifest.i', | |
214 | ) |
|
215 | ) | |
215 | if unencoded in skip_undo: |
|
216 | if unencoded in skip_undo: | |
216 | continue |
|
217 | continue | |
217 |
|
218 | |||
218 | rl = _revlogfrompath(srcrepo, rl_type, unencoded) |
|
219 | rl = _revlogfrompath(srcrepo, entry.revlog_type, unencoded) | |
219 |
|
220 | |||
220 | info = rl.storageinfo( |
|
221 | info = rl.storageinfo( | |
221 | exclusivefiles=True, |
|
222 | exclusivefiles=True, | |
222 | revisionscount=True, |
|
223 | revisionscount=True, | |
223 | trackedsize=True, |
|
224 | trackedsize=True, | |
224 | storedsize=True, |
|
225 | storedsize=True, | |
225 | ) |
|
226 | ) | |
226 |
|
227 | |||
227 | revcount += info[b'revisionscount'] or 0 |
|
228 | revcount += info[b'revisionscount'] or 0 | |
228 | datasize = info[b'storedsize'] or 0 |
|
229 | datasize = info[b'storedsize'] or 0 | |
229 | rawsize = info[b'trackedsize'] or 0 |
|
230 | rawsize = info[b'trackedsize'] or 0 | |
230 |
|
231 | |||
231 | srcsize += datasize |
|
232 | srcsize += datasize | |
232 | srcrawsize += rawsize |
|
233 | srcrawsize += rawsize | |
233 |
|
234 | |||
234 | # This is for the separate progress bars. |
|
235 | # This is for the separate progress bars. | |
235 | if rl_type & store.FILEFLAGS_CHANGELOG: |
|
236 | if entry.revlog_type & store.FILEFLAGS_CHANGELOG: | |
236 | changelogs[unencoded] = rl_type |
|
237 | changelogs[unencoded] = entry.revlog_type | |
237 | crevcount += len(rl) |
|
238 | crevcount += len(rl) | |
238 | csrcsize += datasize |
|
239 | csrcsize += datasize | |
239 | crawsize += rawsize |
|
240 | crawsize += rawsize | |
240 | elif rl_type & store.FILEFLAGS_MANIFESTLOG: |
|
241 | elif entry.revlog_type & store.FILEFLAGS_MANIFESTLOG: | |
241 | manifests[unencoded] = rl_type |
|
242 | manifests[unencoded] = entry.revlog_type | |
242 | mcount += 1 |
|
243 | mcount += 1 | |
243 | mrevcount += len(rl) |
|
244 | mrevcount += len(rl) | |
244 | msrcsize += datasize |
|
245 | msrcsize += datasize | |
245 | mrawsize += rawsize |
|
246 | mrawsize += rawsize | |
246 | elif rl_type & store.FILEFLAGS_FILELOG: |
|
247 | elif entry.revlog_type & store.FILEFLAGS_FILELOG: | |
247 | filelogs[unencoded] = rl_type |
|
248 | filelogs[unencoded] = entry.revlog_type | |
248 | fcount += 1 |
|
249 | fcount += 1 | |
249 | frevcount += len(rl) |
|
250 | frevcount += len(rl) | |
250 | fsrcsize += datasize |
|
251 | fsrcsize += datasize | |
251 | frawsize += rawsize |
|
252 | frawsize += rawsize | |
252 | else: |
|
253 | else: | |
253 | error.ProgrammingError(b'unknown revlog type') |
|
254 | error.ProgrammingError(b'unknown revlog type') | |
254 |
|
255 | |||
255 | if not revcount: |
|
256 | if not revcount: | |
256 | return |
|
257 | return | |
257 |
|
258 | |||
258 | ui.status( |
|
259 | ui.status( | |
259 | _( |
|
260 | _( | |
260 | b'migrating %d total revisions (%d in filelogs, %d in manifests, ' |
|
261 | b'migrating %d total revisions (%d in filelogs, %d in manifests, ' | |
261 | b'%d in changelog)\n' |
|
262 | b'%d in changelog)\n' | |
262 | ) |
|
263 | ) | |
263 | % (revcount, frevcount, mrevcount, crevcount) |
|
264 | % (revcount, frevcount, mrevcount, crevcount) | |
264 | ) |
|
265 | ) | |
265 | ui.status( |
|
266 | ui.status( | |
266 | _(b'migrating %s in store; %s tracked data\n') |
|
267 | _(b'migrating %s in store; %s tracked data\n') | |
267 | % ((util.bytecount(srcsize), util.bytecount(srcrawsize))) |
|
268 | % ((util.bytecount(srcsize), util.bytecount(srcrawsize))) | |
268 | ) |
|
269 | ) | |
269 |
|
270 | |||
270 | # Used to keep track of progress. |
|
271 | # Used to keep track of progress. | |
271 | progress = None |
|
272 | progress = None | |
272 |
|
273 | |||
273 | def oncopiedrevision(rl, rev, node): |
|
274 | def oncopiedrevision(rl, rev, node): | |
274 | progress.increment() |
|
275 | progress.increment() | |
275 |
|
276 | |||
276 | sidedata_helpers = get_sidedata_helpers(srcrepo, dstrepo) |
|
277 | sidedata_helpers = get_sidedata_helpers(srcrepo, dstrepo) | |
277 |
|
278 | |||
278 | # Migrating filelogs |
|
279 | # Migrating filelogs | |
279 | ui.status( |
|
280 | ui.status( | |
280 | _( |
|
281 | _( | |
281 | b'migrating %d filelogs containing %d revisions ' |
|
282 | b'migrating %d filelogs containing %d revisions ' | |
282 | b'(%s in store; %s tracked data)\n' |
|
283 | b'(%s in store; %s tracked data)\n' | |
283 | ) |
|
284 | ) | |
284 | % ( |
|
285 | % ( | |
285 | fcount, |
|
286 | fcount, | |
286 | frevcount, |
|
287 | frevcount, | |
287 | util.bytecount(fsrcsize), |
|
288 | util.bytecount(fsrcsize), | |
288 | util.bytecount(frawsize), |
|
289 | util.bytecount(frawsize), | |
289 | ) |
|
290 | ) | |
290 | ) |
|
291 | ) | |
291 | progress = srcrepo.ui.makeprogress(_(b'file revisions'), total=frevcount) |
|
292 | progress = srcrepo.ui.makeprogress(_(b'file revisions'), total=frevcount) | |
292 | for unencoded, rl_type in sorted(filelogs.items()): |
|
293 | for unencoded, rl_type in sorted(filelogs.items()): | |
293 | oldrl = _revlogfrompath(srcrepo, rl_type, unencoded) |
|
294 | oldrl = _revlogfrompath(srcrepo, rl_type, unencoded) | |
294 |
|
295 | |||
295 | newrl = _perform_clone( |
|
296 | newrl = _perform_clone( | |
296 | ui, |
|
297 | ui, | |
297 | dstrepo, |
|
298 | dstrepo, | |
298 | tr, |
|
299 | tr, | |
299 | oldrl, |
|
300 | oldrl, | |
300 | rl_type, |
|
301 | rl_type, | |
301 | unencoded, |
|
302 | unencoded, | |
302 | upgrade_op, |
|
303 | upgrade_op, | |
303 | sidedata_helpers, |
|
304 | sidedata_helpers, | |
304 | oncopiedrevision, |
|
305 | oncopiedrevision, | |
305 | ) |
|
306 | ) | |
306 | info = newrl.storageinfo(storedsize=True) |
|
307 | info = newrl.storageinfo(storedsize=True) | |
307 | fdstsize += info[b'storedsize'] or 0 |
|
308 | fdstsize += info[b'storedsize'] or 0 | |
308 | ui.status( |
|
309 | ui.status( | |
309 | _( |
|
310 | _( | |
310 | b'finished migrating %d filelog revisions across %d ' |
|
311 | b'finished migrating %d filelog revisions across %d ' | |
311 | b'filelogs; change in size: %s\n' |
|
312 | b'filelogs; change in size: %s\n' | |
312 | ) |
|
313 | ) | |
313 | % (frevcount, fcount, util.bytecount(fdstsize - fsrcsize)) |
|
314 | % (frevcount, fcount, util.bytecount(fdstsize - fsrcsize)) | |
314 | ) |
|
315 | ) | |
315 |
|
316 | |||
316 | # Migrating manifests |
|
317 | # Migrating manifests | |
317 | ui.status( |
|
318 | ui.status( | |
318 | _( |
|
319 | _( | |
319 | b'migrating %d manifests containing %d revisions ' |
|
320 | b'migrating %d manifests containing %d revisions ' | |
320 | b'(%s in store; %s tracked data)\n' |
|
321 | b'(%s in store; %s tracked data)\n' | |
321 | ) |
|
322 | ) | |
322 | % ( |
|
323 | % ( | |
323 | mcount, |
|
324 | mcount, | |
324 | mrevcount, |
|
325 | mrevcount, | |
325 | util.bytecount(msrcsize), |
|
326 | util.bytecount(msrcsize), | |
326 | util.bytecount(mrawsize), |
|
327 | util.bytecount(mrawsize), | |
327 | ) |
|
328 | ) | |
328 | ) |
|
329 | ) | |
329 | if progress: |
|
330 | if progress: | |
330 | progress.complete() |
|
331 | progress.complete() | |
331 | progress = srcrepo.ui.makeprogress( |
|
332 | progress = srcrepo.ui.makeprogress( | |
332 | _(b'manifest revisions'), total=mrevcount |
|
333 | _(b'manifest revisions'), total=mrevcount | |
333 | ) |
|
334 | ) | |
334 | for unencoded, rl_type in sorted(manifests.items()): |
|
335 | for unencoded, rl_type in sorted(manifests.items()): | |
335 | oldrl = _revlogfrompath(srcrepo, rl_type, unencoded) |
|
336 | oldrl = _revlogfrompath(srcrepo, rl_type, unencoded) | |
336 | newrl = _perform_clone( |
|
337 | newrl = _perform_clone( | |
337 | ui, |
|
338 | ui, | |
338 | dstrepo, |
|
339 | dstrepo, | |
339 | tr, |
|
340 | tr, | |
340 | oldrl, |
|
341 | oldrl, | |
341 | rl_type, |
|
342 | rl_type, | |
342 | unencoded, |
|
343 | unencoded, | |
343 | upgrade_op, |
|
344 | upgrade_op, | |
344 | sidedata_helpers, |
|
345 | sidedata_helpers, | |
345 | oncopiedrevision, |
|
346 | oncopiedrevision, | |
346 | ) |
|
347 | ) | |
347 | info = newrl.storageinfo(storedsize=True) |
|
348 | info = newrl.storageinfo(storedsize=True) | |
348 | mdstsize += info[b'storedsize'] or 0 |
|
349 | mdstsize += info[b'storedsize'] or 0 | |
349 | ui.status( |
|
350 | ui.status( | |
350 | _( |
|
351 | _( | |
351 | b'finished migrating %d manifest revisions across %d ' |
|
352 | b'finished migrating %d manifest revisions across %d ' | |
352 | b'manifests; change in size: %s\n' |
|
353 | b'manifests; change in size: %s\n' | |
353 | ) |
|
354 | ) | |
354 | % (mrevcount, mcount, util.bytecount(mdstsize - msrcsize)) |
|
355 | % (mrevcount, mcount, util.bytecount(mdstsize - msrcsize)) | |
355 | ) |
|
356 | ) | |
356 |
|
357 | |||
357 | # Migrating changelog |
|
358 | # Migrating changelog | |
358 | ui.status( |
|
359 | ui.status( | |
359 | _( |
|
360 | _( | |
360 | b'migrating changelog containing %d revisions ' |
|
361 | b'migrating changelog containing %d revisions ' | |
361 | b'(%s in store; %s tracked data)\n' |
|
362 | b'(%s in store; %s tracked data)\n' | |
362 | ) |
|
363 | ) | |
363 | % ( |
|
364 | % ( | |
364 | crevcount, |
|
365 | crevcount, | |
365 | util.bytecount(csrcsize), |
|
366 | util.bytecount(csrcsize), | |
366 | util.bytecount(crawsize), |
|
367 | util.bytecount(crawsize), | |
367 | ) |
|
368 | ) | |
368 | ) |
|
369 | ) | |
369 | if progress: |
|
370 | if progress: | |
370 | progress.complete() |
|
371 | progress.complete() | |
371 | progress = srcrepo.ui.makeprogress( |
|
372 | progress = srcrepo.ui.makeprogress( | |
372 | _(b'changelog revisions'), total=crevcount |
|
373 | _(b'changelog revisions'), total=crevcount | |
373 | ) |
|
374 | ) | |
374 | for unencoded, rl_type in sorted(changelogs.items()): |
|
375 | for unencoded, rl_type in sorted(changelogs.items()): | |
375 | oldrl = _revlogfrompath(srcrepo, rl_type, unencoded) |
|
376 | oldrl = _revlogfrompath(srcrepo, rl_type, unencoded) | |
376 | newrl = _perform_clone( |
|
377 | newrl = _perform_clone( | |
377 | ui, |
|
378 | ui, | |
378 | dstrepo, |
|
379 | dstrepo, | |
379 | tr, |
|
380 | tr, | |
380 | oldrl, |
|
381 | oldrl, | |
381 | rl_type, |
|
382 | rl_type, | |
382 | unencoded, |
|
383 | unencoded, | |
383 | upgrade_op, |
|
384 | upgrade_op, | |
384 | sidedata_helpers, |
|
385 | sidedata_helpers, | |
385 | oncopiedrevision, |
|
386 | oncopiedrevision, | |
386 | ) |
|
387 | ) | |
387 | info = newrl.storageinfo(storedsize=True) |
|
388 | info = newrl.storageinfo(storedsize=True) | |
388 | cdstsize += info[b'storedsize'] or 0 |
|
389 | cdstsize += info[b'storedsize'] or 0 | |
389 | progress.complete() |
|
390 | progress.complete() | |
390 | ui.status( |
|
391 | ui.status( | |
391 | _( |
|
392 | _( | |
392 | b'finished migrating %d changelog revisions; change in size: ' |
|
393 | b'finished migrating %d changelog revisions; change in size: ' | |
393 | b'%s\n' |
|
394 | b'%s\n' | |
394 | ) |
|
395 | ) | |
395 | % (crevcount, util.bytecount(cdstsize - csrcsize)) |
|
396 | % (crevcount, util.bytecount(cdstsize - csrcsize)) | |
396 | ) |
|
397 | ) | |
397 |
|
398 | |||
398 | dstsize = fdstsize + mdstsize + cdstsize |
|
399 | dstsize = fdstsize + mdstsize + cdstsize | |
399 | ui.status( |
|
400 | ui.status( | |
400 | _( |
|
401 | _( | |
401 | b'finished migrating %d total revisions; total change in store ' |
|
402 | b'finished migrating %d total revisions; total change in store ' | |
402 | b'size: %s\n' |
|
403 | b'size: %s\n' | |
403 | ) |
|
404 | ) | |
404 | % (revcount, util.bytecount(dstsize - srcsize)) |
|
405 | % (revcount, util.bytecount(dstsize - srcsize)) | |
405 | ) |
|
406 | ) | |
406 |
|
407 | |||
407 |
|
408 | |||
408 | def _files_to_copy_post_revlog_clone(srcrepo): |
|
409 | def _files_to_copy_post_revlog_clone(srcrepo): | |
409 | """yields files which should be copied to destination after revlogs |
|
410 | """yields files which should be copied to destination after revlogs | |
410 | are cloned""" |
|
411 | are cloned""" | |
411 | for path, kind, st in sorted(srcrepo.store.vfs.readdir(b'', stat=True)): |
|
412 | for path, kind, st in sorted(srcrepo.store.vfs.readdir(b'', stat=True)): | |
412 | # don't copy revlogs as they are already cloned |
|
413 | # don't copy revlogs as they are already cloned | |
413 | if store.revlog_type(path) is not None: |
|
414 | if store.revlog_type(path) is not None: | |
414 | continue |
|
415 | continue | |
415 | # Skip transaction related files. |
|
416 | # Skip transaction related files. | |
416 | if path.startswith(b'undo'): |
|
417 | if path.startswith(b'undo'): | |
417 | continue |
|
418 | continue | |
418 | # Only copy regular files. |
|
419 | # Only copy regular files. | |
419 | if kind != stat.S_IFREG: |
|
420 | if kind != stat.S_IFREG: | |
420 | continue |
|
421 | continue | |
421 | # Skip other skipped files. |
|
422 | # Skip other skipped files. | |
422 | if path in (b'lock', b'fncache'): |
|
423 | if path in (b'lock', b'fncache'): | |
423 | continue |
|
424 | continue | |
424 | # TODO: should we skip cache too? |
|
425 | # TODO: should we skip cache too? | |
425 |
|
426 | |||
426 | yield path |
|
427 | yield path | |
427 |
|
428 | |||
428 |
|
429 | |||
429 | def _replacestores(currentrepo, upgradedrepo, backupvfs, upgrade_op): |
|
430 | def _replacestores(currentrepo, upgradedrepo, backupvfs, upgrade_op): | |
430 | """Replace the stores after current repository is upgraded |
|
431 | """Replace the stores after current repository is upgraded | |
431 |
|
432 | |||
432 | Creates a backup of current repository store at backup path |
|
433 | Creates a backup of current repository store at backup path | |
433 | Replaces upgraded store files in current repo from upgraded one |
|
434 | Replaces upgraded store files in current repo from upgraded one | |
434 |
|
435 | |||
435 | Arguments: |
|
436 | Arguments: | |
436 | currentrepo: repo object of current repository |
|
437 | currentrepo: repo object of current repository | |
437 | upgradedrepo: repo object of the upgraded data |
|
438 | upgradedrepo: repo object of the upgraded data | |
438 | backupvfs: vfs object for the backup path |
|
439 | backupvfs: vfs object for the backup path | |
439 | upgrade_op: upgrade operation object |
|
440 | upgrade_op: upgrade operation object | |
440 | to be used to decide what all is upgraded |
|
441 | to be used to decide what all is upgraded | |
441 | """ |
|
442 | """ | |
442 | # TODO: don't blindly rename everything in store |
|
443 | # TODO: don't blindly rename everything in store | |
443 | # There can be upgrades where store is not touched at all |
|
444 | # There can be upgrades where store is not touched at all | |
444 | if upgrade_op.backup_store: |
|
445 | if upgrade_op.backup_store: | |
445 | util.rename(currentrepo.spath, backupvfs.join(b'store')) |
|
446 | util.rename(currentrepo.spath, backupvfs.join(b'store')) | |
446 | else: |
|
447 | else: | |
447 | currentrepo.vfs.rmtree(b'store', forcibly=True) |
|
448 | currentrepo.vfs.rmtree(b'store', forcibly=True) | |
448 | util.rename(upgradedrepo.spath, currentrepo.spath) |
|
449 | util.rename(upgradedrepo.spath, currentrepo.spath) | |
449 |
|
450 | |||
450 |
|
451 | |||
451 | def finishdatamigration(ui, srcrepo, dstrepo, requirements): |
|
452 | def finishdatamigration(ui, srcrepo, dstrepo, requirements): | |
452 | """Hook point for extensions to perform additional actions during upgrade. |
|
453 | """Hook point for extensions to perform additional actions during upgrade. | |
453 |
|
454 | |||
454 | This function is called after revlogs and store files have been copied but |
|
455 | This function is called after revlogs and store files have been copied but | |
455 | before the new store is swapped into the original location. |
|
456 | before the new store is swapped into the original location. | |
456 | """ |
|
457 | """ | |
457 |
|
458 | |||
458 |
|
459 | |||
459 | def upgrade(ui, srcrepo, dstrepo, upgrade_op): |
|
460 | def upgrade(ui, srcrepo, dstrepo, upgrade_op): | |
460 | """Do the low-level work of upgrading a repository. |
|
461 | """Do the low-level work of upgrading a repository. | |
461 |
|
462 | |||
462 | The upgrade is effectively performed as a copy between a source |
|
463 | The upgrade is effectively performed as a copy between a source | |
463 | repository and a temporary destination repository. |
|
464 | repository and a temporary destination repository. | |
464 |
|
465 | |||
465 | The source repository is unmodified for as long as possible so the |
|
466 | The source repository is unmodified for as long as possible so the | |
466 | upgrade can abort at any time without causing loss of service for |
|
467 | upgrade can abort at any time without causing loss of service for | |
467 | readers and without corrupting the source repository. |
|
468 | readers and without corrupting the source repository. | |
468 | """ |
|
469 | """ | |
469 | assert srcrepo.currentwlock() |
|
470 | assert srcrepo.currentwlock() | |
470 | assert dstrepo.currentwlock() |
|
471 | assert dstrepo.currentwlock() | |
471 | backuppath = None |
|
472 | backuppath = None | |
472 | backupvfs = None |
|
473 | backupvfs = None | |
473 |
|
474 | |||
474 | ui.status( |
|
475 | ui.status( | |
475 | _( |
|
476 | _( | |
476 | b'(it is safe to interrupt this process any time before ' |
|
477 | b'(it is safe to interrupt this process any time before ' | |
477 | b'data migration completes)\n' |
|
478 | b'data migration completes)\n' | |
478 | ) |
|
479 | ) | |
479 | ) |
|
480 | ) | |
480 |
|
481 | |||
481 | if upgrade_actions.dirstatev2 in upgrade_op.upgrade_actions: |
|
482 | if upgrade_actions.dirstatev2 in upgrade_op.upgrade_actions: | |
482 | ui.status(_(b'upgrading to dirstate-v2 from v1\n')) |
|
483 | ui.status(_(b'upgrading to dirstate-v2 from v1\n')) | |
483 | upgrade_dirstate(ui, srcrepo, upgrade_op, b'v1', b'v2') |
|
484 | upgrade_dirstate(ui, srcrepo, upgrade_op, b'v1', b'v2') | |
484 | upgrade_op.upgrade_actions.remove(upgrade_actions.dirstatev2) |
|
485 | upgrade_op.upgrade_actions.remove(upgrade_actions.dirstatev2) | |
485 |
|
486 | |||
486 | if upgrade_actions.dirstatev2 in upgrade_op.removed_actions: |
|
487 | if upgrade_actions.dirstatev2 in upgrade_op.removed_actions: | |
487 | ui.status(_(b'downgrading from dirstate-v2 to v1\n')) |
|
488 | ui.status(_(b'downgrading from dirstate-v2 to v1\n')) | |
488 | upgrade_dirstate(ui, srcrepo, upgrade_op, b'v2', b'v1') |
|
489 | upgrade_dirstate(ui, srcrepo, upgrade_op, b'v2', b'v1') | |
489 | upgrade_op.removed_actions.remove(upgrade_actions.dirstatev2) |
|
490 | upgrade_op.removed_actions.remove(upgrade_actions.dirstatev2) | |
490 |
|
491 | |||
491 | if upgrade_actions.dirstatetrackedkey in upgrade_op.upgrade_actions: |
|
492 | if upgrade_actions.dirstatetrackedkey in upgrade_op.upgrade_actions: | |
492 | ui.status(_(b'create dirstate-tracked-hint file\n')) |
|
493 | ui.status(_(b'create dirstate-tracked-hint file\n')) | |
493 | upgrade_tracked_hint(ui, srcrepo, upgrade_op, add=True) |
|
494 | upgrade_tracked_hint(ui, srcrepo, upgrade_op, add=True) | |
494 | upgrade_op.upgrade_actions.remove(upgrade_actions.dirstatetrackedkey) |
|
495 | upgrade_op.upgrade_actions.remove(upgrade_actions.dirstatetrackedkey) | |
495 | elif upgrade_actions.dirstatetrackedkey in upgrade_op.removed_actions: |
|
496 | elif upgrade_actions.dirstatetrackedkey in upgrade_op.removed_actions: | |
496 | ui.status(_(b'remove dirstate-tracked-hint file\n')) |
|
497 | ui.status(_(b'remove dirstate-tracked-hint file\n')) | |
497 | upgrade_tracked_hint(ui, srcrepo, upgrade_op, add=False) |
|
498 | upgrade_tracked_hint(ui, srcrepo, upgrade_op, add=False) | |
498 | upgrade_op.removed_actions.remove(upgrade_actions.dirstatetrackedkey) |
|
499 | upgrade_op.removed_actions.remove(upgrade_actions.dirstatetrackedkey) | |
499 |
|
500 | |||
500 | if not (upgrade_op.upgrade_actions or upgrade_op.removed_actions): |
|
501 | if not (upgrade_op.upgrade_actions or upgrade_op.removed_actions): | |
501 | return |
|
502 | return | |
502 |
|
503 | |||
503 | if upgrade_op.requirements_only: |
|
504 | if upgrade_op.requirements_only: | |
504 | ui.status(_(b'upgrading repository requirements\n')) |
|
505 | ui.status(_(b'upgrading repository requirements\n')) | |
505 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
506 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
506 | # if there is only one action and that is persistent nodemap upgrade |
|
507 | # if there is only one action and that is persistent nodemap upgrade | |
507 | # directly write the nodemap file and update requirements instead of going |
|
508 | # directly write the nodemap file and update requirements instead of going | |
508 | # through the whole cloning process |
|
509 | # through the whole cloning process | |
509 | elif ( |
|
510 | elif ( | |
510 | len(upgrade_op.upgrade_actions) == 1 |
|
511 | len(upgrade_op.upgrade_actions) == 1 | |
511 | and b'persistent-nodemap' in upgrade_op.upgrade_actions_names |
|
512 | and b'persistent-nodemap' in upgrade_op.upgrade_actions_names | |
512 | and not upgrade_op.removed_actions |
|
513 | and not upgrade_op.removed_actions | |
513 | ): |
|
514 | ): | |
514 | ui.status( |
|
515 | ui.status( | |
515 | _(b'upgrading repository to use persistent nodemap feature\n') |
|
516 | _(b'upgrading repository to use persistent nodemap feature\n') | |
516 | ) |
|
517 | ) | |
517 | with srcrepo.transaction(b'upgrade') as tr: |
|
518 | with srcrepo.transaction(b'upgrade') as tr: | |
518 | unfi = srcrepo.unfiltered() |
|
519 | unfi = srcrepo.unfiltered() | |
519 | cl = unfi.changelog |
|
520 | cl = unfi.changelog | |
520 | nodemap.persist_nodemap(tr, cl, force=True) |
|
521 | nodemap.persist_nodemap(tr, cl, force=True) | |
521 | # we want to directly operate on the underlying revlog to force |
|
522 | # we want to directly operate on the underlying revlog to force | |
522 | # create a nodemap file. This is fine since this is upgrade code |
|
523 | # create a nodemap file. This is fine since this is upgrade code | |
523 | # and it heavily relies on repository being revlog based |
|
524 | # and it heavily relies on repository being revlog based | |
524 | # hence accessing private attributes can be justified |
|
525 | # hence accessing private attributes can be justified | |
525 | nodemap.persist_nodemap( |
|
526 | nodemap.persist_nodemap( | |
526 | tr, unfi.manifestlog._rootstore._revlog, force=True |
|
527 | tr, unfi.manifestlog._rootstore._revlog, force=True | |
527 | ) |
|
528 | ) | |
528 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
529 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
529 | elif ( |
|
530 | elif ( | |
530 | len(upgrade_op.removed_actions) == 1 |
|
531 | len(upgrade_op.removed_actions) == 1 | |
531 | and [ |
|
532 | and [ | |
532 | x |
|
533 | x | |
533 | for x in upgrade_op.removed_actions |
|
534 | for x in upgrade_op.removed_actions | |
534 | if x.name == b'persistent-nodemap' |
|
535 | if x.name == b'persistent-nodemap' | |
535 | ] |
|
536 | ] | |
536 | and not upgrade_op.upgrade_actions |
|
537 | and not upgrade_op.upgrade_actions | |
537 | ): |
|
538 | ): | |
538 | ui.status( |
|
539 | ui.status( | |
539 | _(b'downgrading repository to not use persistent nodemap feature\n') |
|
540 | _(b'downgrading repository to not use persistent nodemap feature\n') | |
540 | ) |
|
541 | ) | |
541 | with srcrepo.transaction(b'upgrade') as tr: |
|
542 | with srcrepo.transaction(b'upgrade') as tr: | |
542 | unfi = srcrepo.unfiltered() |
|
543 | unfi = srcrepo.unfiltered() | |
543 | cl = unfi.changelog |
|
544 | cl = unfi.changelog | |
544 | nodemap.delete_nodemap(tr, srcrepo, cl) |
|
545 | nodemap.delete_nodemap(tr, srcrepo, cl) | |
545 | # check comment 20 lines above for accessing private attributes |
|
546 | # check comment 20 lines above for accessing private attributes | |
546 | nodemap.delete_nodemap( |
|
547 | nodemap.delete_nodemap( | |
547 | tr, srcrepo, unfi.manifestlog._rootstore._revlog |
|
548 | tr, srcrepo, unfi.manifestlog._rootstore._revlog | |
548 | ) |
|
549 | ) | |
549 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
550 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
550 | else: |
|
551 | else: | |
551 | with dstrepo.transaction(b'upgrade') as tr: |
|
552 | with dstrepo.transaction(b'upgrade') as tr: | |
552 | _clonerevlogs( |
|
553 | _clonerevlogs( | |
553 | ui, |
|
554 | ui, | |
554 | srcrepo, |
|
555 | srcrepo, | |
555 | dstrepo, |
|
556 | dstrepo, | |
556 | tr, |
|
557 | tr, | |
557 | upgrade_op, |
|
558 | upgrade_op, | |
558 | ) |
|
559 | ) | |
559 |
|
560 | |||
560 | # Now copy other files in the store directory. |
|
561 | # Now copy other files in the store directory. | |
561 | for p in _files_to_copy_post_revlog_clone(srcrepo): |
|
562 | for p in _files_to_copy_post_revlog_clone(srcrepo): | |
562 | srcrepo.ui.status(_(b'copying %s\n') % p) |
|
563 | srcrepo.ui.status(_(b'copying %s\n') % p) | |
563 | src = srcrepo.store.rawvfs.join(p) |
|
564 | src = srcrepo.store.rawvfs.join(p) | |
564 | dst = dstrepo.store.rawvfs.join(p) |
|
565 | dst = dstrepo.store.rawvfs.join(p) | |
565 | util.copyfile(src, dst, copystat=True) |
|
566 | util.copyfile(src, dst, copystat=True) | |
566 |
|
567 | |||
567 | finishdatamigration(ui, srcrepo, dstrepo, requirements) |
|
568 | finishdatamigration(ui, srcrepo, dstrepo, requirements) | |
568 |
|
569 | |||
569 | ui.status(_(b'data fully upgraded in a temporary repository\n')) |
|
570 | ui.status(_(b'data fully upgraded in a temporary repository\n')) | |
570 |
|
571 | |||
571 | if upgrade_op.backup_store: |
|
572 | if upgrade_op.backup_store: | |
572 | backuppath = pycompat.mkdtemp( |
|
573 | backuppath = pycompat.mkdtemp( | |
573 | prefix=b'upgradebackup.', dir=srcrepo.path |
|
574 | prefix=b'upgradebackup.', dir=srcrepo.path | |
574 | ) |
|
575 | ) | |
575 | backupvfs = vfsmod.vfs(backuppath) |
|
576 | backupvfs = vfsmod.vfs(backuppath) | |
576 |
|
577 | |||
577 | # Make a backup of requires file first, as it is the first to be modified. |
|
578 | # Make a backup of requires file first, as it is the first to be modified. | |
578 | util.copyfile( |
|
579 | util.copyfile( | |
579 | srcrepo.vfs.join(b'requires'), backupvfs.join(b'requires') |
|
580 | srcrepo.vfs.join(b'requires'), backupvfs.join(b'requires') | |
580 | ) |
|
581 | ) | |
581 |
|
582 | |||
582 | # We install an arbitrary requirement that clients must not support |
|
583 | # We install an arbitrary requirement that clients must not support | |
583 | # as a mechanism to lock out new clients during the data swap. This is |
|
584 | # as a mechanism to lock out new clients during the data swap. This is | |
584 | # better than allowing a client to continue while the repository is in |
|
585 | # better than allowing a client to continue while the repository is in | |
585 | # an inconsistent state. |
|
586 | # an inconsistent state. | |
586 | ui.status( |
|
587 | ui.status( | |
587 | _( |
|
588 | _( | |
588 | b'marking source repository as being upgraded; clients will be ' |
|
589 | b'marking source repository as being upgraded; clients will be ' | |
589 | b'unable to read from repository\n' |
|
590 | b'unable to read from repository\n' | |
590 | ) |
|
591 | ) | |
591 | ) |
|
592 | ) | |
592 | scmutil.writereporequirements( |
|
593 | scmutil.writereporequirements( | |
593 | srcrepo, srcrepo.requirements | {b'upgradeinprogress'} |
|
594 | srcrepo, srcrepo.requirements | {b'upgradeinprogress'} | |
594 | ) |
|
595 | ) | |
595 |
|
596 | |||
596 | ui.status(_(b'starting in-place swap of repository data\n')) |
|
597 | ui.status(_(b'starting in-place swap of repository data\n')) | |
597 | if upgrade_op.backup_store: |
|
598 | if upgrade_op.backup_store: | |
598 | ui.status( |
|
599 | ui.status( | |
599 | _(b'replaced files will be backed up at %s\n') % backuppath |
|
600 | _(b'replaced files will be backed up at %s\n') % backuppath | |
600 | ) |
|
601 | ) | |
601 |
|
602 | |||
602 | # Now swap in the new store directory. Doing it as a rename should make |
|
603 | # Now swap in the new store directory. Doing it as a rename should make | |
603 | # the operation nearly instantaneous and atomic (at least in well-behaved |
|
604 | # the operation nearly instantaneous and atomic (at least in well-behaved | |
604 | # environments). |
|
605 | # environments). | |
605 | ui.status(_(b'replacing store...\n')) |
|
606 | ui.status(_(b'replacing store...\n')) | |
606 | tstart = util.timer() |
|
607 | tstart = util.timer() | |
607 | _replacestores(srcrepo, dstrepo, backupvfs, upgrade_op) |
|
608 | _replacestores(srcrepo, dstrepo, backupvfs, upgrade_op) | |
608 | elapsed = util.timer() - tstart |
|
609 | elapsed = util.timer() - tstart | |
609 | ui.status( |
|
610 | ui.status( | |
610 | _( |
|
611 | _( | |
611 | b'store replacement complete; repository was inconsistent for ' |
|
612 | b'store replacement complete; repository was inconsistent for ' | |
612 | b'%0.1fs\n' |
|
613 | b'%0.1fs\n' | |
613 | ) |
|
614 | ) | |
614 | % elapsed |
|
615 | % elapsed | |
615 | ) |
|
616 | ) | |
616 |
|
617 | |||
617 | # We first write the requirements file. Any new requirements will lock |
|
618 | # We first write the requirements file. Any new requirements will lock | |
618 | # out legacy clients. |
|
619 | # out legacy clients. | |
619 | ui.status( |
|
620 | ui.status( | |
620 | _( |
|
621 | _( | |
621 | b'finalizing requirements file and making repository readable ' |
|
622 | b'finalizing requirements file and making repository readable ' | |
622 | b'again\n' |
|
623 | b'again\n' | |
623 | ) |
|
624 | ) | |
624 | ) |
|
625 | ) | |
625 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
626 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
626 |
|
627 | |||
627 | if upgrade_op.backup_store: |
|
628 | if upgrade_op.backup_store: | |
628 | # The lock file from the old store won't be removed because nothing has a |
|
629 | # The lock file from the old store won't be removed because nothing has a | |
629 | # reference to its new location. So clean it up manually. Alternatively, we |
|
630 | # reference to its new location. So clean it up manually. Alternatively, we | |
630 | # could update srcrepo.svfs and other variables to point to the new |
|
631 | # could update srcrepo.svfs and other variables to point to the new | |
631 | # location. This is simpler. |
|
632 | # location. This is simpler. | |
632 | assert backupvfs is not None # help pytype |
|
633 | assert backupvfs is not None # help pytype | |
633 | backupvfs.unlink(b'store/lock') |
|
634 | backupvfs.unlink(b'store/lock') | |
634 |
|
635 | |||
635 | return backuppath |
|
636 | return backuppath | |
636 |
|
637 | |||
637 |
|
638 | |||
638 | def upgrade_dirstate(ui, srcrepo, upgrade_op, old, new): |
|
639 | def upgrade_dirstate(ui, srcrepo, upgrade_op, old, new): | |
639 | if upgrade_op.backup_store: |
|
640 | if upgrade_op.backup_store: | |
640 | backuppath = pycompat.mkdtemp( |
|
641 | backuppath = pycompat.mkdtemp( | |
641 | prefix=b'upgradebackup.', dir=srcrepo.path |
|
642 | prefix=b'upgradebackup.', dir=srcrepo.path | |
642 | ) |
|
643 | ) | |
643 | ui.status(_(b'replaced files will be backed up at %s\n') % backuppath) |
|
644 | ui.status(_(b'replaced files will be backed up at %s\n') % backuppath) | |
644 | backupvfs = vfsmod.vfs(backuppath) |
|
645 | backupvfs = vfsmod.vfs(backuppath) | |
645 | util.copyfile( |
|
646 | util.copyfile( | |
646 | srcrepo.vfs.join(b'requires'), backupvfs.join(b'requires') |
|
647 | srcrepo.vfs.join(b'requires'), backupvfs.join(b'requires') | |
647 | ) |
|
648 | ) | |
648 | try: |
|
649 | try: | |
649 | util.copyfile( |
|
650 | util.copyfile( | |
650 | srcrepo.vfs.join(b'dirstate'), backupvfs.join(b'dirstate') |
|
651 | srcrepo.vfs.join(b'dirstate'), backupvfs.join(b'dirstate') | |
651 | ) |
|
652 | ) | |
652 | except FileNotFoundError: |
|
653 | except FileNotFoundError: | |
653 | # The dirstate does not exist on an empty repo or a repo with no |
|
654 | # The dirstate does not exist on an empty repo or a repo with no | |
654 | # revision checked out |
|
655 | # revision checked out | |
655 | pass |
|
656 | pass | |
656 |
|
657 | |||
657 | assert srcrepo.dirstate._use_dirstate_v2 == (old == b'v2') |
|
658 | assert srcrepo.dirstate._use_dirstate_v2 == (old == b'v2') | |
658 | use_v2 = new == b'v2' |
|
659 | use_v2 = new == b'v2' | |
659 | if use_v2: |
|
660 | if use_v2: | |
660 | # Write the requirements *before* upgrading |
|
661 | # Write the requirements *before* upgrading | |
661 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
662 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
662 |
|
663 | |||
663 | srcrepo.dirstate._map.preload() |
|
664 | srcrepo.dirstate._map.preload() | |
664 | srcrepo.dirstate._use_dirstate_v2 = use_v2 |
|
665 | srcrepo.dirstate._use_dirstate_v2 = use_v2 | |
665 | srcrepo.dirstate._map._use_dirstate_v2 = use_v2 |
|
666 | srcrepo.dirstate._map._use_dirstate_v2 = use_v2 | |
666 | srcrepo.dirstate._dirty = True |
|
667 | srcrepo.dirstate._dirty = True | |
667 | try: |
|
668 | try: | |
668 | srcrepo.vfs.unlink(b'dirstate') |
|
669 | srcrepo.vfs.unlink(b'dirstate') | |
669 | except FileNotFoundError: |
|
670 | except FileNotFoundError: | |
670 | # The dirstate does not exist on an empty repo or a repo with no |
|
671 | # The dirstate does not exist on an empty repo or a repo with no | |
671 | # revision checked out |
|
672 | # revision checked out | |
672 | pass |
|
673 | pass | |
673 |
|
674 | |||
674 | srcrepo.dirstate.write(None) |
|
675 | srcrepo.dirstate.write(None) | |
675 | if not use_v2: |
|
676 | if not use_v2: | |
676 | # Remove the v2 requirement *after* downgrading |
|
677 | # Remove the v2 requirement *after* downgrading | |
677 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
678 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) | |
678 |
|
679 | |||
679 |
|
680 | |||
680 | def upgrade_tracked_hint(ui, srcrepo, upgrade_op, add): |
|
681 | def upgrade_tracked_hint(ui, srcrepo, upgrade_op, add): | |
681 | if add: |
|
682 | if add: | |
682 | srcrepo.dirstate._use_tracked_hint = True |
|
683 | srcrepo.dirstate._use_tracked_hint = True | |
683 | srcrepo.dirstate._dirty = True |
|
684 | srcrepo.dirstate._dirty = True | |
684 | srcrepo.dirstate._dirty_tracked_set = True |
|
685 | srcrepo.dirstate._dirty_tracked_set = True | |
685 | srcrepo.dirstate.write(None) |
|
686 | srcrepo.dirstate.write(None) | |
686 | if not add: |
|
687 | if not add: | |
687 | srcrepo.dirstate.delete_tracked_hint() |
|
688 | srcrepo.dirstate.delete_tracked_hint() | |
688 |
|
689 | |||
689 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
|
690 | scmutil.writereporequirements(srcrepo, upgrade_op.new_requirements) |
@@ -1,621 +1,625 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 |
|
8 | |||
9 | import os |
|
9 | import os | |
10 |
|
10 | |||
11 | from .i18n import _ |
|
11 | from .i18n import _ | |
12 | from .node import short |
|
12 | from .node import short | |
13 | from .utils import stringutil |
|
13 | from .utils import stringutil | |
14 |
|
14 | |||
15 | from . import ( |
|
15 | from . import ( | |
16 | error, |
|
16 | error, | |
17 | pycompat, |
|
17 | pycompat, | |
18 | requirements, |
|
18 | requirements, | |
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: |
|
58 | class verifier: | |
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 | if self.errors: |
|
214 | if self.errors: | |
215 | ui.warn(_(b"not checking dirstate because of previous errors\n")) |
|
215 | ui.warn(_(b"not checking dirstate because of previous errors\n")) | |
216 | dirstate_errors = 0 |
|
216 | dirstate_errors = 0 | |
217 | else: |
|
217 | else: | |
218 | dirstate_errors = self._verify_dirstate() |
|
218 | dirstate_errors = self._verify_dirstate() | |
219 |
|
219 | |||
220 | # final report |
|
220 | # final report | |
221 | ui.status( |
|
221 | ui.status( | |
222 | _(b"checked %d changesets with %d changes to %d files\n") |
|
222 | _(b"checked %d changesets with %d changes to %d files\n") | |
223 | % (len(repo.changelog), filerevisions, totalfiles) |
|
223 | % (len(repo.changelog), filerevisions, totalfiles) | |
224 | ) |
|
224 | ) | |
225 | if self.warnings: |
|
225 | if self.warnings: | |
226 | ui.warn(_(b"%d warnings encountered!\n") % self.warnings) |
|
226 | ui.warn(_(b"%d warnings encountered!\n") % self.warnings) | |
227 | if self.fncachewarned: |
|
227 | if self.fncachewarned: | |
228 | ui.warn(HINT_FNCACHE) |
|
228 | ui.warn(HINT_FNCACHE) | |
229 | if self.errors: |
|
229 | if self.errors: | |
230 | ui.warn(_(b"%d integrity errors encountered!\n") % self.errors) |
|
230 | ui.warn(_(b"%d integrity errors encountered!\n") % self.errors) | |
231 | if self.badrevs: |
|
231 | if self.badrevs: | |
232 | msg = _(b"(first damaged changeset appears to be %d)\n") |
|
232 | msg = _(b"(first damaged changeset appears to be %d)\n") | |
233 | msg %= min(self.badrevs) |
|
233 | msg %= min(self.badrevs) | |
234 | ui.warn(msg) |
|
234 | ui.warn(msg) | |
235 | if dirstate_errors: |
|
235 | if dirstate_errors: | |
236 | ui.warn( |
|
236 | ui.warn( | |
237 | _(b"dirstate inconsistent with current parent's manifest\n") |
|
237 | _(b"dirstate inconsistent with current parent's manifest\n") | |
238 | ) |
|
238 | ) | |
239 | ui.warn(_(b"%d dirstate errors\n") % dirstate_errors) |
|
239 | ui.warn(_(b"%d dirstate errors\n") % dirstate_errors) | |
240 | return 1 |
|
240 | return 1 | |
241 | return 0 |
|
241 | return 0 | |
242 |
|
242 | |||
243 | def _verifychangelog(self): |
|
243 | def _verifychangelog(self): | |
244 | """verify the changelog of a repository |
|
244 | """verify the changelog of a repository | |
245 |
|
245 | |||
246 | The following checks are performed: |
|
246 | The following checks are performed: | |
247 | - all of `_checkrevlog` checks, |
|
247 | - all of `_checkrevlog` checks, | |
248 | - all of `_checkentry` checks (for each revisions), |
|
248 | - all of `_checkentry` checks (for each revisions), | |
249 | - each revision can be read. |
|
249 | - each revision can be read. | |
250 |
|
250 | |||
251 | The function returns some of the data observed in the changesets as a |
|
251 | The function returns some of the data observed in the changesets as a | |
252 | (mflinkrevs, filelinkrevs) tuples: |
|
252 | (mflinkrevs, filelinkrevs) tuples: | |
253 | - mflinkrevs: is a { manifest-node -> [changelog-rev] } mapping |
|
253 | - mflinkrevs: is a { manifest-node -> [changelog-rev] } mapping | |
254 | - filelinkrevs: is a { file-path -> [changelog-rev] } mapping |
|
254 | - filelinkrevs: is a { file-path -> [changelog-rev] } mapping | |
255 |
|
255 | |||
256 | If a matcher was specified, filelinkrevs will only contains matched |
|
256 | If a matcher was specified, filelinkrevs will only contains matched | |
257 | files. |
|
257 | files. | |
258 | """ |
|
258 | """ | |
259 | ui = self.ui |
|
259 | ui = self.ui | |
260 | repo = self.repo |
|
260 | repo = self.repo | |
261 | match = self.match |
|
261 | match = self.match | |
262 | cl = repo.changelog |
|
262 | cl = repo.changelog | |
263 |
|
263 | |||
264 | ui.status(_(b"checking changesets\n")) |
|
264 | ui.status(_(b"checking changesets\n")) | |
265 | mflinkrevs = {} |
|
265 | mflinkrevs = {} | |
266 | filelinkrevs = {} |
|
266 | filelinkrevs = {} | |
267 | seen = {} |
|
267 | seen = {} | |
268 | self._checkrevlog(cl, b"changelog", 0) |
|
268 | self._checkrevlog(cl, b"changelog", 0) | |
269 | progress = ui.makeprogress( |
|
269 | progress = ui.makeprogress( | |
270 | _(b'checking'), unit=_(b'changesets'), total=len(repo) |
|
270 | _(b'checking'), unit=_(b'changesets'), total=len(repo) | |
271 | ) |
|
271 | ) | |
272 | for i in repo: |
|
272 | for i in repo: | |
273 | progress.update(i) |
|
273 | progress.update(i) | |
274 | n = cl.node(i) |
|
274 | n = cl.node(i) | |
275 | self._checkentry(cl, i, n, seen, [i], b"changelog") |
|
275 | self._checkentry(cl, i, n, seen, [i], b"changelog") | |
276 |
|
276 | |||
277 | try: |
|
277 | try: | |
278 | changes = cl.read(n) |
|
278 | changes = cl.read(n) | |
279 | if changes[0] != self.repo.nullid: |
|
279 | if changes[0] != self.repo.nullid: | |
280 | mflinkrevs.setdefault(changes[0], []).append(i) |
|
280 | mflinkrevs.setdefault(changes[0], []).append(i) | |
281 | self.refersmf = True |
|
281 | self.refersmf = True | |
282 | for f in changes[3]: |
|
282 | for f in changes[3]: | |
283 | if match(f): |
|
283 | if match(f): | |
284 | filelinkrevs.setdefault(_normpath(f), []).append(i) |
|
284 | filelinkrevs.setdefault(_normpath(f), []).append(i) | |
285 | except Exception as inst: |
|
285 | except Exception as inst: | |
286 | self.refersmf = True |
|
286 | self.refersmf = True | |
287 | self._exc(i, _(b"unpacking changeset %s") % short(n), inst) |
|
287 | self._exc(i, _(b"unpacking changeset %s") % short(n), inst) | |
288 | progress.complete() |
|
288 | progress.complete() | |
289 | return mflinkrevs, filelinkrevs |
|
289 | return mflinkrevs, filelinkrevs | |
290 |
|
290 | |||
291 | def _verifymanifest( |
|
291 | def _verifymanifest( | |
292 | self, mflinkrevs, dir=b"", storefiles=None, subdirprogress=None |
|
292 | self, mflinkrevs, dir=b"", storefiles=None, subdirprogress=None | |
293 | ): |
|
293 | ): | |
294 | """verify the manifestlog content |
|
294 | """verify the manifestlog content | |
295 |
|
295 | |||
296 | Inputs: |
|
296 | Inputs: | |
297 | - mflinkrevs: a {manifest-node -> [changelog-revisions]} mapping |
|
297 | - mflinkrevs: a {manifest-node -> [changelog-revisions]} mapping | |
298 | - dir: a subdirectory to check (for tree manifest repo) |
|
298 | - dir: a subdirectory to check (for tree manifest repo) | |
299 | - storefiles: set of currently "orphan" files. |
|
299 | - storefiles: set of currently "orphan" files. | |
300 | - subdirprogress: a progress object |
|
300 | - subdirprogress: a progress object | |
301 |
|
301 | |||
302 | This function checks: |
|
302 | This function checks: | |
303 | * all of `_checkrevlog` checks (for all manifest related revlogs) |
|
303 | * all of `_checkrevlog` checks (for all manifest related revlogs) | |
304 | * all of `_checkentry` checks (for all manifest related revisions) |
|
304 | * all of `_checkentry` checks (for all manifest related revisions) | |
305 | * nodes for subdirectory exists in the sub-directory manifest |
|
305 | * nodes for subdirectory exists in the sub-directory manifest | |
306 | * each manifest entries have a file path |
|
306 | * each manifest entries have a file path | |
307 | * each manifest node refered in mflinkrevs exist in the manifest log |
|
307 | * each manifest node refered in mflinkrevs exist in the manifest log | |
308 |
|
308 | |||
309 | If tree manifest is in use and a matchers is specified, only the |
|
309 | If tree manifest is in use and a matchers is specified, only the | |
310 | sub-directories matching it will be verified. |
|
310 | sub-directories matching it will be verified. | |
311 |
|
311 | |||
312 | return a two level mapping: |
|
312 | return a two level mapping: | |
313 | {"path" -> { filenode -> changelog-revision}} |
|
313 | {"path" -> { filenode -> changelog-revision}} | |
314 |
|
314 | |||
315 | This mapping primarily contains entries for every files in the |
|
315 | This mapping primarily contains entries for every files in the | |
316 | repository. In addition, when tree-manifest is used, it also contains |
|
316 | repository. In addition, when tree-manifest is used, it also contains | |
317 | sub-directory entries. |
|
317 | sub-directory entries. | |
318 |
|
318 | |||
319 | If a matcher is provided, only matching paths will be included. |
|
319 | If a matcher is provided, only matching paths will be included. | |
320 | """ |
|
320 | """ | |
321 | repo = self.repo |
|
321 | repo = self.repo | |
322 | ui = self.ui |
|
322 | ui = self.ui | |
323 | match = self.match |
|
323 | match = self.match | |
324 | mfl = self.repo.manifestlog |
|
324 | mfl = self.repo.manifestlog | |
325 | mf = mfl.getstorage(dir) |
|
325 | mf = mfl.getstorage(dir) | |
326 |
|
326 | |||
327 | if not dir: |
|
327 | if not dir: | |
328 | self.ui.status(_(b"checking manifests\n")) |
|
328 | self.ui.status(_(b"checking manifests\n")) | |
329 |
|
329 | |||
330 | filenodes = {} |
|
330 | filenodes = {} | |
331 | subdirnodes = {} |
|
331 | subdirnodes = {} | |
332 | seen = {} |
|
332 | seen = {} | |
333 | label = b"manifest" |
|
333 | label = b"manifest" | |
334 | if dir: |
|
334 | if dir: | |
335 | label = dir |
|
335 | label = dir | |
336 | revlogfiles = mf.files() |
|
336 | revlogfiles = mf.files() | |
337 | storefiles.difference_update(revlogfiles) |
|
337 | storefiles.difference_update(revlogfiles) | |
338 | if subdirprogress: # should be true since we're in a subdirectory |
|
338 | if subdirprogress: # should be true since we're in a subdirectory | |
339 | subdirprogress.increment() |
|
339 | subdirprogress.increment() | |
340 | if self.refersmf: |
|
340 | if self.refersmf: | |
341 | # Do not check manifest if there are only changelog entries with |
|
341 | # Do not check manifest if there are only changelog entries with | |
342 | # null manifests. |
|
342 | # null manifests. | |
343 | self._checkrevlog(mf._revlog, label, 0) |
|
343 | self._checkrevlog(mf._revlog, label, 0) | |
344 | progress = ui.makeprogress( |
|
344 | progress = ui.makeprogress( | |
345 | _(b'checking'), unit=_(b'manifests'), total=len(mf) |
|
345 | _(b'checking'), unit=_(b'manifests'), total=len(mf) | |
346 | ) |
|
346 | ) | |
347 | for i in mf: |
|
347 | for i in mf: | |
348 | if not dir: |
|
348 | if not dir: | |
349 | progress.update(i) |
|
349 | progress.update(i) | |
350 | n = mf.node(i) |
|
350 | n = mf.node(i) | |
351 | lr = self._checkentry(mf, i, n, seen, mflinkrevs.get(n, []), label) |
|
351 | lr = self._checkentry(mf, i, n, seen, mflinkrevs.get(n, []), label) | |
352 | if n in mflinkrevs: |
|
352 | if n in mflinkrevs: | |
353 | del mflinkrevs[n] |
|
353 | del mflinkrevs[n] | |
354 | elif dir: |
|
354 | elif dir: | |
355 | msg = _(b"%s not in parent-directory manifest") % short(n) |
|
355 | msg = _(b"%s not in parent-directory manifest") % short(n) | |
356 | self._err(lr, msg, label) |
|
356 | self._err(lr, msg, label) | |
357 | else: |
|
357 | else: | |
358 | self._err(lr, _(b"%s not in changesets") % short(n), label) |
|
358 | self._err(lr, _(b"%s not in changesets") % short(n), label) | |
359 |
|
359 | |||
360 | try: |
|
360 | try: | |
361 | mfdelta = mfl.get(dir, n).readdelta(shallow=True) |
|
361 | mfdelta = mfl.get(dir, n).readdelta(shallow=True) | |
362 | for f, fn, fl in mfdelta.iterentries(): |
|
362 | for f, fn, fl in mfdelta.iterentries(): | |
363 | if not f: |
|
363 | if not f: | |
364 | self._err(lr, _(b"entry without name in manifest")) |
|
364 | self._err(lr, _(b"entry without name in manifest")) | |
365 | elif f == b"/dev/null": # ignore this in very old repos |
|
365 | elif f == b"/dev/null": # ignore this in very old repos | |
366 | continue |
|
366 | continue | |
367 | fullpath = dir + _normpath(f) |
|
367 | fullpath = dir + _normpath(f) | |
368 | if fl == b't': |
|
368 | if fl == b't': | |
369 | if not match.visitdir(fullpath): |
|
369 | if not match.visitdir(fullpath): | |
370 | continue |
|
370 | continue | |
371 | sdn = subdirnodes.setdefault(fullpath + b'/', {}) |
|
371 | sdn = subdirnodes.setdefault(fullpath + b'/', {}) | |
372 | sdn.setdefault(fn, []).append(lr) |
|
372 | sdn.setdefault(fn, []).append(lr) | |
373 | else: |
|
373 | else: | |
374 | if not match(fullpath): |
|
374 | if not match(fullpath): | |
375 | continue |
|
375 | continue | |
376 | filenodes.setdefault(fullpath, {}).setdefault(fn, lr) |
|
376 | filenodes.setdefault(fullpath, {}).setdefault(fn, lr) | |
377 | except Exception as inst: |
|
377 | except Exception as inst: | |
378 | self._exc(lr, _(b"reading delta %s") % short(n), inst, label) |
|
378 | self._exc(lr, _(b"reading delta %s") % short(n), inst, label) | |
379 | if self._level >= VERIFY_FULL: |
|
379 | if self._level >= VERIFY_FULL: | |
380 | try: |
|
380 | try: | |
381 | # Various issues can affect manifest. So we read each full |
|
381 | # Various issues can affect manifest. So we read each full | |
382 | # text from storage. This triggers the checks from the core |
|
382 | # text from storage. This triggers the checks from the core | |
383 | # code (eg: hash verification, filename are ordered, etc.) |
|
383 | # code (eg: hash verification, filename are ordered, etc.) | |
384 | mfdelta = mfl.get(dir, n).read() |
|
384 | mfdelta = mfl.get(dir, n).read() | |
385 | except Exception as inst: |
|
385 | except Exception as inst: | |
386 | msg = _(b"reading full manifest %s") % short(n) |
|
386 | msg = _(b"reading full manifest %s") % short(n) | |
387 | self._exc(lr, msg, inst, label) |
|
387 | self._exc(lr, msg, inst, label) | |
388 |
|
388 | |||
389 | if not dir: |
|
389 | if not dir: | |
390 | progress.complete() |
|
390 | progress.complete() | |
391 |
|
391 | |||
392 | if self.havemf: |
|
392 | if self.havemf: | |
393 | # since we delete entry in `mflinkrevs` during iteration, any |
|
393 | # since we delete entry in `mflinkrevs` during iteration, any | |
394 | # remaining entries are "missing". We need to issue errors for them. |
|
394 | # remaining entries are "missing". We need to issue errors for them. | |
395 | changesetpairs = [(c, m) for m in mflinkrevs for c in mflinkrevs[m]] |
|
395 | changesetpairs = [(c, m) for m in mflinkrevs for c in mflinkrevs[m]] | |
396 | for c, m in sorted(changesetpairs): |
|
396 | for c, m in sorted(changesetpairs): | |
397 | if dir: |
|
397 | if dir: | |
398 | self._err(c, WARN_PARENT_DIR_UNKNOWN_REV % short(m), label) |
|
398 | self._err(c, WARN_PARENT_DIR_UNKNOWN_REV % short(m), label) | |
399 | else: |
|
399 | else: | |
400 | msg = _(b"changeset refers to unknown revision %s") |
|
400 | msg = _(b"changeset refers to unknown revision %s") | |
401 | msg %= short(m) |
|
401 | msg %= short(m) | |
402 | self._err(c, msg, label) |
|
402 | self._err(c, msg, label) | |
403 |
|
403 | |||
404 | if not dir and subdirnodes: |
|
404 | if not dir and subdirnodes: | |
405 | self.ui.status(_(b"checking directory manifests\n")) |
|
405 | self.ui.status(_(b"checking directory manifests\n")) | |
406 | storefiles = set() |
|
406 | storefiles = set() | |
407 | subdirs = set() |
|
407 | subdirs = set() | |
408 | revlogv1 = self.revlogv1 |
|
408 | revlogv1 = self.revlogv1 | |
409 | undecodable = [] |
|
409 | undecodable = [] | |
410 |
for |
|
410 | for entry in repo.store.datafiles(undecodable=undecodable): | |
|
411 | f = entry.unencoded_path | |||
|
412 | size = entry.file_size | |||
411 | if (size > 0 or not revlogv1) and f.startswith(b'meta/'): |
|
413 | if (size > 0 or not revlogv1) and f.startswith(b'meta/'): | |
412 | storefiles.add(_normpath(f)) |
|
414 | storefiles.add(_normpath(f)) | |
413 | subdirs.add(os.path.dirname(f)) |
|
415 | subdirs.add(os.path.dirname(f)) | |
414 | for f in undecodable: |
|
416 | for f in undecodable: | |
415 | self._err(None, _(b"cannot decode filename '%s'") % f) |
|
417 | self._err(None, _(b"cannot decode filename '%s'") % f) | |
416 | subdirprogress = ui.makeprogress( |
|
418 | subdirprogress = ui.makeprogress( | |
417 | _(b'checking'), unit=_(b'manifests'), total=len(subdirs) |
|
419 | _(b'checking'), unit=_(b'manifests'), total=len(subdirs) | |
418 | ) |
|
420 | ) | |
419 |
|
421 | |||
420 | for subdir, linkrevs in subdirnodes.items(): |
|
422 | for subdir, linkrevs in subdirnodes.items(): | |
421 | subdirfilenodes = self._verifymanifest( |
|
423 | subdirfilenodes = self._verifymanifest( | |
422 | linkrevs, subdir, storefiles, subdirprogress |
|
424 | linkrevs, subdir, storefiles, subdirprogress | |
423 | ) |
|
425 | ) | |
424 | for f, onefilenodes in subdirfilenodes.items(): |
|
426 | for f, onefilenodes in subdirfilenodes.items(): | |
425 | filenodes.setdefault(f, {}).update(onefilenodes) |
|
427 | filenodes.setdefault(f, {}).update(onefilenodes) | |
426 |
|
428 | |||
427 | if not dir and subdirnodes: |
|
429 | if not dir and subdirnodes: | |
428 | assert subdirprogress is not None # help pytype |
|
430 | assert subdirprogress is not None # help pytype | |
429 | subdirprogress.complete() |
|
431 | subdirprogress.complete() | |
430 | if self.warnorphanstorefiles: |
|
432 | if self.warnorphanstorefiles: | |
431 | for f in sorted(storefiles): |
|
433 | for f in sorted(storefiles): | |
432 | self._warn(_(b"warning: orphan data file '%s'") % f) |
|
434 | self._warn(_(b"warning: orphan data file '%s'") % f) | |
433 |
|
435 | |||
434 | return filenodes |
|
436 | return filenodes | |
435 |
|
437 | |||
436 | def _crosscheckfiles(self, filelinkrevs, filenodes): |
|
438 | def _crosscheckfiles(self, filelinkrevs, filenodes): | |
437 | repo = self.repo |
|
439 | repo = self.repo | |
438 | ui = self.ui |
|
440 | ui = self.ui | |
439 | ui.status(_(b"crosschecking files in changesets and manifests\n")) |
|
441 | ui.status(_(b"crosschecking files in changesets and manifests\n")) | |
440 |
|
442 | |||
441 | total = len(filelinkrevs) + len(filenodes) |
|
443 | total = len(filelinkrevs) + len(filenodes) | |
442 | progress = ui.makeprogress( |
|
444 | progress = ui.makeprogress( | |
443 | _(b'crosschecking'), unit=_(b'files'), total=total |
|
445 | _(b'crosschecking'), unit=_(b'files'), total=total | |
444 | ) |
|
446 | ) | |
445 | if self.havemf: |
|
447 | if self.havemf: | |
446 | for f in sorted(filelinkrevs): |
|
448 | for f in sorted(filelinkrevs): | |
447 | progress.increment() |
|
449 | progress.increment() | |
448 | if f not in filenodes: |
|
450 | if f not in filenodes: | |
449 | lr = filelinkrevs[f][0] |
|
451 | lr = filelinkrevs[f][0] | |
450 | self._err(lr, _(b"in changeset but not in manifest"), f) |
|
452 | self._err(lr, _(b"in changeset but not in manifest"), f) | |
451 |
|
453 | |||
452 | if self.havecl: |
|
454 | if self.havecl: | |
453 | for f in sorted(filenodes): |
|
455 | for f in sorted(filenodes): | |
454 | progress.increment() |
|
456 | progress.increment() | |
455 | if f not in filelinkrevs: |
|
457 | if f not in filelinkrevs: | |
456 | try: |
|
458 | try: | |
457 | fl = repo.file(f) |
|
459 | fl = repo.file(f) | |
458 | lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]]) |
|
460 | lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]]) | |
459 | except Exception: |
|
461 | except Exception: | |
460 | lr = None |
|
462 | lr = None | |
461 | self._err(lr, _(b"in manifest but not in changeset"), f) |
|
463 | self._err(lr, _(b"in manifest but not in changeset"), f) | |
462 |
|
464 | |||
463 | progress.complete() |
|
465 | progress.complete() | |
464 |
|
466 | |||
465 | def _verifyfiles(self, filenodes, filelinkrevs): |
|
467 | def _verifyfiles(self, filenodes, filelinkrevs): | |
466 | repo = self.repo |
|
468 | repo = self.repo | |
467 | ui = self.ui |
|
469 | ui = self.ui | |
468 | lrugetctx = self.lrugetctx |
|
470 | lrugetctx = self.lrugetctx | |
469 | revlogv1 = self.revlogv1 |
|
471 | revlogv1 = self.revlogv1 | |
470 | havemf = self.havemf |
|
472 | havemf = self.havemf | |
471 | ui.status(_(b"checking files\n")) |
|
473 | ui.status(_(b"checking files\n")) | |
472 |
|
474 | |||
473 | storefiles = set() |
|
475 | storefiles = set() | |
474 | undecodable = [] |
|
476 | undecodable = [] | |
475 |
for |
|
477 | for entry in repo.store.datafiles(undecodable=undecodable): | |
|
478 | size = entry.file_size | |||
|
479 | f = entry.unencoded_path | |||
476 | if (size > 0 or not revlogv1) and f.startswith(b'data/'): |
|
480 | if (size > 0 or not revlogv1) and f.startswith(b'data/'): | |
477 | storefiles.add(_normpath(f)) |
|
481 | storefiles.add(_normpath(f)) | |
478 | for f in undecodable: |
|
482 | for f in undecodable: | |
479 | self._err(None, _(b"cannot decode filename '%s'") % f) |
|
483 | self._err(None, _(b"cannot decode filename '%s'") % f) | |
480 |
|
484 | |||
481 | state = { |
|
485 | state = { | |
482 | # TODO this assumes revlog storage for changelog. |
|
486 | # TODO this assumes revlog storage for changelog. | |
483 | b'expectedversion': self.repo.changelog._format_version, |
|
487 | b'expectedversion': self.repo.changelog._format_version, | |
484 | b'skipflags': self.skipflags, |
|
488 | b'skipflags': self.skipflags, | |
485 | # experimental config: censor.policy |
|
489 | # experimental config: censor.policy | |
486 | b'erroroncensored': ui.config(b'censor', b'policy') == b'abort', |
|
490 | b'erroroncensored': ui.config(b'censor', b'policy') == b'abort', | |
487 | } |
|
491 | } | |
488 |
|
492 | |||
489 | files = sorted(set(filenodes) | set(filelinkrevs)) |
|
493 | files = sorted(set(filenodes) | set(filelinkrevs)) | |
490 | revisions = 0 |
|
494 | revisions = 0 | |
491 | progress = ui.makeprogress( |
|
495 | progress = ui.makeprogress( | |
492 | _(b'checking'), unit=_(b'files'), total=len(files) |
|
496 | _(b'checking'), unit=_(b'files'), total=len(files) | |
493 | ) |
|
497 | ) | |
494 | for i, f in enumerate(files): |
|
498 | for i, f in enumerate(files): | |
495 | progress.update(i, item=f) |
|
499 | progress.update(i, item=f) | |
496 | try: |
|
500 | try: | |
497 | linkrevs = filelinkrevs[f] |
|
501 | linkrevs = filelinkrevs[f] | |
498 | except KeyError: |
|
502 | except KeyError: | |
499 | # in manifest but not in changelog |
|
503 | # in manifest but not in changelog | |
500 | linkrevs = [] |
|
504 | linkrevs = [] | |
501 |
|
505 | |||
502 | if linkrevs: |
|
506 | if linkrevs: | |
503 | lr = linkrevs[0] |
|
507 | lr = linkrevs[0] | |
504 | else: |
|
508 | else: | |
505 | lr = None |
|
509 | lr = None | |
506 |
|
510 | |||
507 | try: |
|
511 | try: | |
508 | fl = repo.file(f) |
|
512 | fl = repo.file(f) | |
509 | except error.StorageError as e: |
|
513 | except error.StorageError as e: | |
510 | self._err(lr, _(b"broken revlog! (%s)") % e, f) |
|
514 | self._err(lr, _(b"broken revlog! (%s)") % e, f) | |
511 | continue |
|
515 | continue | |
512 |
|
516 | |||
513 | for ff in fl.files(): |
|
517 | for ff in fl.files(): | |
514 | try: |
|
518 | try: | |
515 | storefiles.remove(ff) |
|
519 | storefiles.remove(ff) | |
516 | except KeyError: |
|
520 | except KeyError: | |
517 | if self.warnorphanstorefiles: |
|
521 | if self.warnorphanstorefiles: | |
518 | msg = _(b" warning: revlog '%s' not in fncache!") |
|
522 | msg = _(b" warning: revlog '%s' not in fncache!") | |
519 | self._warn(msg % ff) |
|
523 | self._warn(msg % ff) | |
520 | self.fncachewarned = True |
|
524 | self.fncachewarned = True | |
521 |
|
525 | |||
522 | if not len(fl) and (self.havecl or self.havemf): |
|
526 | if not len(fl) and (self.havecl or self.havemf): | |
523 | self._err(lr, _(b"empty or missing %s") % f) |
|
527 | self._err(lr, _(b"empty or missing %s") % f) | |
524 | else: |
|
528 | else: | |
525 | # Guard against implementations not setting this. |
|
529 | # Guard against implementations not setting this. | |
526 | state[b'skipread'] = set() |
|
530 | state[b'skipread'] = set() | |
527 | state[b'safe_renamed'] = set() |
|
531 | state[b'safe_renamed'] = set() | |
528 |
|
532 | |||
529 | for problem in fl.verifyintegrity(state): |
|
533 | for problem in fl.verifyintegrity(state): | |
530 | if problem.node is not None: |
|
534 | if problem.node is not None: | |
531 | linkrev = fl.linkrev(fl.rev(problem.node)) |
|
535 | linkrev = fl.linkrev(fl.rev(problem.node)) | |
532 | else: |
|
536 | else: | |
533 | linkrev = None |
|
537 | linkrev = None | |
534 |
|
538 | |||
535 | if problem.warning: |
|
539 | if problem.warning: | |
536 | self._warn(problem.warning) |
|
540 | self._warn(problem.warning) | |
537 | elif problem.error: |
|
541 | elif problem.error: | |
538 | linkrev_msg = linkrev if linkrev is not None else lr |
|
542 | linkrev_msg = linkrev if linkrev is not None else lr | |
539 | self._err(linkrev_msg, problem.error, f) |
|
543 | self._err(linkrev_msg, problem.error, f) | |
540 | else: |
|
544 | else: | |
541 | raise error.ProgrammingError( |
|
545 | raise error.ProgrammingError( | |
542 | b'problem instance does not set warning or error ' |
|
546 | b'problem instance does not set warning or error ' | |
543 | b'attribute: %s' % problem.msg |
|
547 | b'attribute: %s' % problem.msg | |
544 | ) |
|
548 | ) | |
545 |
|
549 | |||
546 | seen = {} |
|
550 | seen = {} | |
547 | for i in fl: |
|
551 | for i in fl: | |
548 | revisions += 1 |
|
552 | revisions += 1 | |
549 | n = fl.node(i) |
|
553 | n = fl.node(i) | |
550 | lr = self._checkentry(fl, i, n, seen, linkrevs, f) |
|
554 | lr = self._checkentry(fl, i, n, seen, linkrevs, f) | |
551 | if f in filenodes: |
|
555 | if f in filenodes: | |
552 | if havemf and n not in filenodes[f]: |
|
556 | if havemf and n not in filenodes[f]: | |
553 | self._err(lr, _(b"%s not in manifests") % (short(n)), f) |
|
557 | self._err(lr, _(b"%s not in manifests") % (short(n)), f) | |
554 | else: |
|
558 | else: | |
555 | del filenodes[f][n] |
|
559 | del filenodes[f][n] | |
556 |
|
560 | |||
557 | if n in state[b'skipread'] and n not in state[b'safe_renamed']: |
|
561 | if n in state[b'skipread'] and n not in state[b'safe_renamed']: | |
558 | continue |
|
562 | continue | |
559 |
|
563 | |||
560 | # check renames |
|
564 | # check renames | |
561 | try: |
|
565 | try: | |
562 | # This requires resolving fulltext (at least on revlogs, |
|
566 | # This requires resolving fulltext (at least on revlogs, | |
563 | # though not with LFS revisions). We may want |
|
567 | # though not with LFS revisions). We may want | |
564 | # ``verifyintegrity()`` to pass a set of nodes with |
|
568 | # ``verifyintegrity()`` to pass a set of nodes with | |
565 | # rename metadata as an optimization. |
|
569 | # rename metadata as an optimization. | |
566 | rp = fl.renamed(n) |
|
570 | rp = fl.renamed(n) | |
567 | if rp: |
|
571 | if rp: | |
568 | if lr is not None and ui.verbose: |
|
572 | if lr is not None and ui.verbose: | |
569 | ctx = lrugetctx(lr) |
|
573 | ctx = lrugetctx(lr) | |
570 | if not any(rp[0] in pctx for pctx in ctx.parents()): |
|
574 | if not any(rp[0] in pctx for pctx in ctx.parents()): | |
571 | self._warn(WARN_UNKNOWN_COPY_SOURCE % (f, ctx)) |
|
575 | self._warn(WARN_UNKNOWN_COPY_SOURCE % (f, ctx)) | |
572 | fl2 = repo.file(rp[0]) |
|
576 | fl2 = repo.file(rp[0]) | |
573 | if not len(fl2): |
|
577 | if not len(fl2): | |
574 | m = _(b"empty or missing copy source revlog %s:%s") |
|
578 | m = _(b"empty or missing copy source revlog %s:%s") | |
575 | self._err(lr, m % (rp[0], short(rp[1])), f) |
|
579 | self._err(lr, m % (rp[0], short(rp[1])), f) | |
576 | elif rp[1] == self.repo.nullid: |
|
580 | elif rp[1] == self.repo.nullid: | |
577 | msg = WARN_NULLID_COPY_SOURCE |
|
581 | msg = WARN_NULLID_COPY_SOURCE | |
578 | msg %= (f, lr, rp[0], short(rp[1])) |
|
582 | msg %= (f, lr, rp[0], short(rp[1])) | |
579 | ui.note(msg) |
|
583 | ui.note(msg) | |
580 | else: |
|
584 | else: | |
581 | fl2.rev(rp[1]) |
|
585 | fl2.rev(rp[1]) | |
582 | except Exception as inst: |
|
586 | except Exception as inst: | |
583 | self._exc( |
|
587 | self._exc( | |
584 | lr, _(b"checking rename of %s") % short(n), inst, f |
|
588 | lr, _(b"checking rename of %s") % short(n), inst, f | |
585 | ) |
|
589 | ) | |
586 |
|
590 | |||
587 | # cross-check |
|
591 | # cross-check | |
588 | if f in filenodes: |
|
592 | if f in filenodes: | |
589 | fns = [(v, k) for k, v in filenodes[f].items()] |
|
593 | fns = [(v, k) for k, v in filenodes[f].items()] | |
590 | for lr, node in sorted(fns): |
|
594 | for lr, node in sorted(fns): | |
591 | msg = _(b"manifest refers to unknown revision %s") |
|
595 | msg = _(b"manifest refers to unknown revision %s") | |
592 | self._err(lr, msg % short(node), f) |
|
596 | self._err(lr, msg % short(node), f) | |
593 | progress.complete() |
|
597 | progress.complete() | |
594 |
|
598 | |||
595 | if self.warnorphanstorefiles: |
|
599 | if self.warnorphanstorefiles: | |
596 | for f in sorted(storefiles): |
|
600 | for f in sorted(storefiles): | |
597 | self._warn(_(b"warning: orphan data file '%s'") % f) |
|
601 | self._warn(_(b"warning: orphan data file '%s'") % f) | |
598 |
|
602 | |||
599 | return len(files), revisions |
|
603 | return len(files), revisions | |
600 |
|
604 | |||
601 | def _verify_dirstate(self): |
|
605 | def _verify_dirstate(self): | |
602 | """Check that the dirstate is consistent with the parent's manifest""" |
|
606 | """Check that the dirstate is consistent with the parent's manifest""" | |
603 | repo = self.repo |
|
607 | repo = self.repo | |
604 | ui = self.ui |
|
608 | ui = self.ui | |
605 | ui.status(_(b"checking dirstate\n")) |
|
609 | ui.status(_(b"checking dirstate\n")) | |
606 |
|
610 | |||
607 | parent1, parent2 = repo.dirstate.parents() |
|
611 | parent1, parent2 = repo.dirstate.parents() | |
608 | m1 = repo[parent1].manifest() |
|
612 | m1 = repo[parent1].manifest() | |
609 | m2 = repo[parent2].manifest() |
|
613 | m2 = repo[parent2].manifest() | |
610 | dirstate_errors = 0 |
|
614 | dirstate_errors = 0 | |
611 |
|
615 | |||
612 | is_narrow = requirements.NARROW_REQUIREMENT in repo.requirements |
|
616 | is_narrow = requirements.NARROW_REQUIREMENT in repo.requirements | |
613 | narrow_matcher = repo.narrowmatch() if is_narrow else None |
|
617 | narrow_matcher = repo.narrowmatch() if is_narrow else None | |
614 |
|
618 | |||
615 | for err in repo.dirstate.verify(m1, m2, parent1, narrow_matcher): |
|
619 | for err in repo.dirstate.verify(m1, m2, parent1, narrow_matcher): | |
616 | ui.error(err) |
|
620 | ui.error(err) | |
617 | dirstate_errors += 1 |
|
621 | dirstate_errors += 1 | |
618 |
|
622 | |||
619 | if dirstate_errors: |
|
623 | if dirstate_errors: | |
620 | self.errors += dirstate_errors |
|
624 | self.errors += dirstate_errors | |
621 | return dirstate_errors |
|
625 | return dirstate_errors |
General Comments 0
You need to be logged in to leave comments.
Login now