Show More
@@ -1,963 +1,967 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 | from typing import Generator | |
14 |
|
14 | |||
15 | from .i18n import _ |
|
15 | from .i18n import _ | |
16 | from .pycompat import getattr |
|
16 | from .pycompat import getattr | |
17 | from .thirdparty import attr |
|
17 | from .thirdparty import attr | |
18 | from .node import hex |
|
18 | from .node import hex | |
19 | from . import ( |
|
19 | from . import ( | |
20 | changelog, |
|
20 | changelog, | |
21 | error, |
|
21 | error, | |
22 | manifest, |
|
22 | manifest, | |
23 | policy, |
|
23 | policy, | |
24 | pycompat, |
|
24 | pycompat, | |
25 | util, |
|
25 | util, | |
26 | vfs as vfsmod, |
|
26 | vfs as vfsmod, | |
27 | ) |
|
27 | ) | |
28 | from .utils import hashutil |
|
28 | from .utils import hashutil | |
29 |
|
29 | |||
30 | parsers = policy.importmod('parsers') |
|
30 | parsers = policy.importmod('parsers') | |
31 | # how much bytes should be read from fncache in one read |
|
31 | # how much bytes should be read from fncache in one read | |
32 | # It is done to prevent loading large fncache files into memory |
|
32 | # It is done to prevent loading large fncache files into memory | |
33 | fncache_chunksize = 10 ** 6 |
|
33 | fncache_chunksize = 10 ** 6 | |
34 |
|
34 | |||
35 |
|
35 | |||
36 | def _matchtrackedpath(path, matcher): |
|
36 | def _matchtrackedpath(path, matcher): | |
37 | """parses a fncache entry and returns whether the entry is tracking a path |
|
37 | """parses a fncache entry and returns whether the entry is tracking a path | |
38 | matched by matcher or not. |
|
38 | matched by matcher or not. | |
39 |
|
39 | |||
40 | If matcher is None, returns True""" |
|
40 | If matcher is None, returns True""" | |
41 |
|
41 | |||
42 | if matcher is None: |
|
42 | if matcher is None: | |
43 | return True |
|
43 | return True | |
44 | path = decodedir(path) |
|
44 | path = decodedir(path) | |
45 | if path.startswith(b'data/'): |
|
45 | if path.startswith(b'data/'): | |
46 | return matcher(path[len(b'data/') : -len(b'.i')]) |
|
46 | return matcher(path[len(b'data/') : -len(b'.i')]) | |
47 | elif path.startswith(b'meta/'): |
|
47 | elif path.startswith(b'meta/'): | |
48 | return matcher.visitdir(path[len(b'meta/') : -len(b'/00manifest.i')]) |
|
48 | return matcher.visitdir(path[len(b'meta/') : -len(b'/00manifest.i')]) | |
49 |
|
49 | |||
50 | raise error.ProgrammingError(b"cannot decode path %s" % path) |
|
50 | raise error.ProgrammingError(b"cannot decode path %s" % path) | |
51 |
|
51 | |||
52 |
|
52 | |||
53 | # This avoids a collision between a file named foo and a dir named |
|
53 | # This avoids a collision between a file named foo and a dir named | |
54 | # foo.i or foo.d |
|
54 | # foo.i or foo.d | |
55 | def _encodedir(path): |
|
55 | def _encodedir(path): | |
56 | """ |
|
56 | """ | |
57 | >>> _encodedir(b'data/foo.i') |
|
57 | >>> _encodedir(b'data/foo.i') | |
58 | 'data/foo.i' |
|
58 | 'data/foo.i' | |
59 | >>> _encodedir(b'data/foo.i/bla.i') |
|
59 | >>> _encodedir(b'data/foo.i/bla.i') | |
60 | 'data/foo.i.hg/bla.i' |
|
60 | 'data/foo.i.hg/bla.i' | |
61 | >>> _encodedir(b'data/foo.i.hg/bla.i') |
|
61 | >>> _encodedir(b'data/foo.i.hg/bla.i') | |
62 | 'data/foo.i.hg.hg/bla.i' |
|
62 | 'data/foo.i.hg.hg/bla.i' | |
63 | >>> _encodedir(b'data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n') |
|
63 | >>> _encodedir(b'data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n') | |
64 | 'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n' |
|
64 | 'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n' | |
65 | """ |
|
65 | """ | |
66 | return ( |
|
66 | return ( | |
67 | path.replace(b".hg/", b".hg.hg/") |
|
67 | path.replace(b".hg/", b".hg.hg/") | |
68 | .replace(b".i/", b".i.hg/") |
|
68 | .replace(b".i/", b".i.hg/") | |
69 | .replace(b".d/", b".d.hg/") |
|
69 | .replace(b".d/", b".d.hg/") | |
70 | ) |
|
70 | ) | |
71 |
|
71 | |||
72 |
|
72 | |||
73 | encodedir = getattr(parsers, 'encodedir', _encodedir) |
|
73 | encodedir = getattr(parsers, 'encodedir', _encodedir) | |
74 |
|
74 | |||
75 |
|
75 | |||
76 | def decodedir(path): |
|
76 | def decodedir(path): | |
77 | """ |
|
77 | """ | |
78 | >>> decodedir(b'data/foo.i') |
|
78 | >>> decodedir(b'data/foo.i') | |
79 | 'data/foo.i' |
|
79 | 'data/foo.i' | |
80 | >>> decodedir(b'data/foo.i.hg/bla.i') |
|
80 | >>> decodedir(b'data/foo.i.hg/bla.i') | |
81 | 'data/foo.i/bla.i' |
|
81 | 'data/foo.i/bla.i' | |
82 | >>> decodedir(b'data/foo.i.hg.hg/bla.i') |
|
82 | >>> decodedir(b'data/foo.i.hg.hg/bla.i') | |
83 | 'data/foo.i.hg/bla.i' |
|
83 | 'data/foo.i.hg/bla.i' | |
84 | """ |
|
84 | """ | |
85 | if b".hg/" not in path: |
|
85 | if b".hg/" not in path: | |
86 | return path |
|
86 | return path | |
87 | return ( |
|
87 | return ( | |
88 | path.replace(b".d.hg/", b".d/") |
|
88 | path.replace(b".d.hg/", b".d/") | |
89 | .replace(b".i.hg/", b".i/") |
|
89 | .replace(b".i.hg/", b".i/") | |
90 | .replace(b".hg.hg/", b".hg/") |
|
90 | .replace(b".hg.hg/", b".hg/") | |
91 | ) |
|
91 | ) | |
92 |
|
92 | |||
93 |
|
93 | |||
94 | def _reserved(): |
|
94 | def _reserved(): | |
95 | """characters that are problematic for filesystems |
|
95 | """characters that are problematic for filesystems | |
96 |
|
96 | |||
97 | * ascii escapes (0..31) |
|
97 | * ascii escapes (0..31) | |
98 | * ascii hi (126..255) |
|
98 | * ascii hi (126..255) | |
99 | * windows specials |
|
99 | * windows specials | |
100 |
|
100 | |||
101 | these characters will be escaped by encodefunctions |
|
101 | these characters will be escaped by encodefunctions | |
102 | """ |
|
102 | """ | |
103 | winreserved = [ord(x) for x in u'\\:*?"<>|'] |
|
103 | winreserved = [ord(x) for x in u'\\:*?"<>|'] | |
104 | for x in range(32): |
|
104 | for x in range(32): | |
105 | yield x |
|
105 | yield x | |
106 | for x in range(126, 256): |
|
106 | for x in range(126, 256): | |
107 | yield x |
|
107 | yield x | |
108 | for x in winreserved: |
|
108 | for x in winreserved: | |
109 | yield x |
|
109 | yield x | |
110 |
|
110 | |||
111 |
|
111 | |||
112 | def _buildencodefun(): |
|
112 | def _buildencodefun(): | |
113 | """ |
|
113 | """ | |
114 | >>> enc, dec = _buildencodefun() |
|
114 | >>> enc, dec = _buildencodefun() | |
115 |
|
115 | |||
116 | >>> enc(b'nothing/special.txt') |
|
116 | >>> enc(b'nothing/special.txt') | |
117 | 'nothing/special.txt' |
|
117 | 'nothing/special.txt' | |
118 | >>> dec(b'nothing/special.txt') |
|
118 | >>> dec(b'nothing/special.txt') | |
119 | 'nothing/special.txt' |
|
119 | 'nothing/special.txt' | |
120 |
|
120 | |||
121 | >>> enc(b'HELLO') |
|
121 | >>> enc(b'HELLO') | |
122 | '_h_e_l_l_o' |
|
122 | '_h_e_l_l_o' | |
123 | >>> dec(b'_h_e_l_l_o') |
|
123 | >>> dec(b'_h_e_l_l_o') | |
124 | 'HELLO' |
|
124 | 'HELLO' | |
125 |
|
125 | |||
126 | >>> enc(b'hello:world?') |
|
126 | >>> enc(b'hello:world?') | |
127 | 'hello~3aworld~3f' |
|
127 | 'hello~3aworld~3f' | |
128 | >>> dec(b'hello~3aworld~3f') |
|
128 | >>> dec(b'hello~3aworld~3f') | |
129 | 'hello:world?' |
|
129 | 'hello:world?' | |
130 |
|
130 | |||
131 | >>> enc(b'the\\x07quick\\xADshot') |
|
131 | >>> enc(b'the\\x07quick\\xADshot') | |
132 | 'the~07quick~adshot' |
|
132 | 'the~07quick~adshot' | |
133 | >>> dec(b'the~07quick~adshot') |
|
133 | >>> dec(b'the~07quick~adshot') | |
134 | 'the\\x07quick\\xadshot' |
|
134 | 'the\\x07quick\\xadshot' | |
135 | """ |
|
135 | """ | |
136 | e = b'_' |
|
136 | e = b'_' | |
137 | xchr = pycompat.bytechr |
|
137 | xchr = pycompat.bytechr | |
138 | asciistr = list(map(xchr, range(127))) |
|
138 | asciistr = list(map(xchr, range(127))) | |
139 | capitals = list(range(ord(b"A"), ord(b"Z") + 1)) |
|
139 | capitals = list(range(ord(b"A"), ord(b"Z") + 1)) | |
140 |
|
140 | |||
141 | cmap = {x: x for x in asciistr} |
|
141 | cmap = {x: x for x in asciistr} | |
142 | for x in _reserved(): |
|
142 | for x in _reserved(): | |
143 | cmap[xchr(x)] = b"~%02x" % x |
|
143 | cmap[xchr(x)] = b"~%02x" % x | |
144 | for x in capitals + [ord(e)]: |
|
144 | for x in capitals + [ord(e)]: | |
145 | cmap[xchr(x)] = e + xchr(x).lower() |
|
145 | cmap[xchr(x)] = e + xchr(x).lower() | |
146 |
|
146 | |||
147 | dmap = {} |
|
147 | dmap = {} | |
148 | for k, v in cmap.items(): |
|
148 | for k, v in cmap.items(): | |
149 | dmap[v] = k |
|
149 | dmap[v] = k | |
150 |
|
150 | |||
151 | def decode(s): |
|
151 | def decode(s): | |
152 | i = 0 |
|
152 | i = 0 | |
153 | while i < len(s): |
|
153 | while i < len(s): | |
154 | for l in range(1, 4): |
|
154 | for l in range(1, 4): | |
155 | try: |
|
155 | try: | |
156 | yield dmap[s[i : i + l]] |
|
156 | yield dmap[s[i : i + l]] | |
157 | i += l |
|
157 | i += l | |
158 | break |
|
158 | break | |
159 | except KeyError: |
|
159 | except KeyError: | |
160 | pass |
|
160 | pass | |
161 | else: |
|
161 | else: | |
162 | raise KeyError |
|
162 | raise KeyError | |
163 |
|
163 | |||
164 | return ( |
|
164 | return ( | |
165 | lambda s: b''.join([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))]), | |
166 | lambda s: b''.join(list(decode(s))), |
|
166 | lambda s: b''.join(list(decode(s))), | |
167 | ) |
|
167 | ) | |
168 |
|
168 | |||
169 |
|
169 | |||
170 | _encodefname, _decodefname = _buildencodefun() |
|
170 | _encodefname, _decodefname = _buildencodefun() | |
171 |
|
171 | |||
172 |
|
172 | |||
173 | def encodefilename(s): |
|
173 | def encodefilename(s): | |
174 | """ |
|
174 | """ | |
175 | >>> encodefilename(b'foo.i/bar.d/bla.hg/hi:world?/HELLO') |
|
175 | >>> encodefilename(b'foo.i/bar.d/bla.hg/hi:world?/HELLO') | |
176 | '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' | |
177 | """ |
|
177 | """ | |
178 | return _encodefname(encodedir(s)) |
|
178 | return _encodefname(encodedir(s)) | |
179 |
|
179 | |||
180 |
|
180 | |||
181 | def decodefilename(s): |
|
181 | def decodefilename(s): | |
182 | """ |
|
182 | """ | |
183 | >>> 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') | |
184 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' |
|
184 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' | |
185 | """ |
|
185 | """ | |
186 | return decodedir(_decodefname(s)) |
|
186 | return decodedir(_decodefname(s)) | |
187 |
|
187 | |||
188 |
|
188 | |||
189 | def _buildlowerencodefun(): |
|
189 | def _buildlowerencodefun(): | |
190 | """ |
|
190 | """ | |
191 | >>> f = _buildlowerencodefun() |
|
191 | >>> f = _buildlowerencodefun() | |
192 | >>> f(b'nothing/special.txt') |
|
192 | >>> f(b'nothing/special.txt') | |
193 | 'nothing/special.txt' |
|
193 | 'nothing/special.txt' | |
194 | >>> f(b'HELLO') |
|
194 | >>> f(b'HELLO') | |
195 | 'hello' |
|
195 | 'hello' | |
196 | >>> f(b'hello:world?') |
|
196 | >>> f(b'hello:world?') | |
197 | 'hello~3aworld~3f' |
|
197 | 'hello~3aworld~3f' | |
198 | >>> f(b'the\\x07quick\\xADshot') |
|
198 | >>> f(b'the\\x07quick\\xADshot') | |
199 | 'the~07quick~adshot' |
|
199 | 'the~07quick~adshot' | |
200 | """ |
|
200 | """ | |
201 | xchr = pycompat.bytechr |
|
201 | xchr = pycompat.bytechr | |
202 | cmap = {xchr(x): xchr(x) for x in range(127)} |
|
202 | cmap = {xchr(x): xchr(x) for x in range(127)} | |
203 | for x in _reserved(): |
|
203 | for x in _reserved(): | |
204 | cmap[xchr(x)] = b"~%02x" % x |
|
204 | cmap[xchr(x)] = b"~%02x" % x | |
205 | for x in range(ord(b"A"), ord(b"Z") + 1): |
|
205 | for x in range(ord(b"A"), ord(b"Z") + 1): | |
206 | cmap[xchr(x)] = xchr(x).lower() |
|
206 | cmap[xchr(x)] = xchr(x).lower() | |
207 |
|
207 | |||
208 | def lowerencode(s): |
|
208 | def lowerencode(s): | |
209 | return b"".join([cmap[c] for c in pycompat.iterbytestr(s)]) |
|
209 | return b"".join([cmap[c] for c in pycompat.iterbytestr(s)]) | |
210 |
|
210 | |||
211 | return lowerencode |
|
211 | return lowerencode | |
212 |
|
212 | |||
213 |
|
213 | |||
214 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() |
|
214 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() | |
215 |
|
215 | |||
216 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 |
|
216 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 | |
217 | _winres3 = (b'aux', b'con', b'prn', b'nul') # length 3 |
|
217 | _winres3 = (b'aux', b'con', b'prn', b'nul') # length 3 | |
218 | _winres4 = (b'com', b'lpt') # length 4 (with trailing 1..9) |
|
218 | _winres4 = (b'com', b'lpt') # length 4 (with trailing 1..9) | |
219 |
|
219 | |||
220 |
|
220 | |||
221 | def _auxencode(path, dotencode): |
|
221 | def _auxencode(path, dotencode): | |
222 | """ |
|
222 | """ | |
223 | Encodes filenames containing names reserved by Windows or which end in |
|
223 | Encodes filenames containing names reserved by Windows or which end in | |
224 | period or space. Does not touch other single reserved characters c. |
|
224 | period or space. Does not touch other single reserved characters c. | |
225 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. |
|
225 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. | |
226 | Additionally encodes space or period at the beginning, if dotencode is |
|
226 | Additionally encodes space or period at the beginning, if dotencode is | |
227 | True. Parameter path is assumed to be all lowercase. |
|
227 | True. Parameter path is assumed to be all lowercase. | |
228 | 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 | |
229 | 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" | |
230 | doesn't need encoding. |
|
230 | doesn't need encoding. | |
231 |
|
231 | |||
232 | >>> s = b'.foo/aux.txt/txt.aux/con/prn/nul/foo.' |
|
232 | >>> s = b'.foo/aux.txt/txt.aux/con/prn/nul/foo.' | |
233 | >>> _auxencode(s.split(b'/'), True) |
|
233 | >>> _auxencode(s.split(b'/'), True) | |
234 | ['~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'] | |
235 | >>> s = b'.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' |
|
235 | >>> s = b'.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' | |
236 | >>> _auxencode(s.split(b'/'), False) |
|
236 | >>> _auxencode(s.split(b'/'), False) | |
237 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] |
|
237 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] | |
238 | >>> _auxencode([b'foo. '], True) |
|
238 | >>> _auxencode([b'foo. '], True) | |
239 | ['foo.~20'] |
|
239 | ['foo.~20'] | |
240 | >>> _auxencode([b' .foo'], True) |
|
240 | >>> _auxencode([b' .foo'], True) | |
241 | ['~20.foo'] |
|
241 | ['~20.foo'] | |
242 | """ |
|
242 | """ | |
243 | for i, n in enumerate(path): |
|
243 | for i, n in enumerate(path): | |
244 | if not n: |
|
244 | if not n: | |
245 | continue |
|
245 | continue | |
246 | if dotencode and n[0] in b'. ': |
|
246 | if dotencode and n[0] in b'. ': | |
247 | n = b"~%02x" % ord(n[0:1]) + n[1:] |
|
247 | n = b"~%02x" % ord(n[0:1]) + n[1:] | |
248 | path[i] = n |
|
248 | path[i] = n | |
249 | else: |
|
249 | else: | |
250 | l = n.find(b'.') |
|
250 | l = n.find(b'.') | |
251 | if l == -1: |
|
251 | if l == -1: | |
252 | l = len(n) |
|
252 | l = len(n) | |
253 | if (l == 3 and n[:3] in _winres3) or ( |
|
253 | if (l == 3 and n[:3] in _winres3) or ( | |
254 | l == 4 |
|
254 | l == 4 | |
255 | and n[3:4] <= b'9' |
|
255 | and n[3:4] <= b'9' | |
256 | and n[3:4] >= b'1' |
|
256 | and n[3:4] >= b'1' | |
257 | and n[:3] in _winres4 |
|
257 | and n[:3] in _winres4 | |
258 | ): |
|
258 | ): | |
259 | # encode third letter ('aux' -> 'au~78') |
|
259 | # encode third letter ('aux' -> 'au~78') | |
260 | ec = b"~%02x" % ord(n[2:3]) |
|
260 | ec = b"~%02x" % ord(n[2:3]) | |
261 | n = n[0:2] + ec + n[3:] |
|
261 | n = n[0:2] + ec + n[3:] | |
262 | path[i] = n |
|
262 | path[i] = n | |
263 | if n[-1] in b'. ': |
|
263 | if n[-1] in b'. ': | |
264 | # encode last period or space ('foo...' -> 'foo..~2e') |
|
264 | # encode last period or space ('foo...' -> 'foo..~2e') | |
265 | path[i] = n[:-1] + b"~%02x" % ord(n[-1:]) |
|
265 | path[i] = n[:-1] + b"~%02x" % ord(n[-1:]) | |
266 | return path |
|
266 | return path | |
267 |
|
267 | |||
268 |
|
268 | |||
269 | _maxstorepathlen = 120 |
|
269 | _maxstorepathlen = 120 | |
270 | _dirprefixlen = 8 |
|
270 | _dirprefixlen = 8 | |
271 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 |
|
271 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 | |
272 |
|
272 | |||
273 |
|
273 | |||
274 | def _hashencode(path, dotencode): |
|
274 | def _hashencode(path, dotencode): | |
275 | digest = hex(hashutil.sha1(path).digest()) |
|
275 | digest = hex(hashutil.sha1(path).digest()) | |
276 | le = lowerencode(path[5:]).split(b'/') # skips prefix 'data/' or 'meta/' |
|
276 | le = lowerencode(path[5:]).split(b'/') # skips prefix 'data/' or 'meta/' | |
277 | parts = _auxencode(le, dotencode) |
|
277 | parts = _auxencode(le, dotencode) | |
278 | basename = parts[-1] |
|
278 | basename = parts[-1] | |
279 | _root, ext = os.path.splitext(basename) |
|
279 | _root, ext = os.path.splitext(basename) | |
280 | sdirs = [] |
|
280 | sdirs = [] | |
281 | sdirslen = 0 |
|
281 | sdirslen = 0 | |
282 | for p in parts[:-1]: |
|
282 | for p in parts[:-1]: | |
283 | d = p[:_dirprefixlen] |
|
283 | d = p[:_dirprefixlen] | |
284 | if d[-1] in b'. ': |
|
284 | if d[-1] in b'. ': | |
285 | # Windows can't access dirs ending in period or space |
|
285 | # Windows can't access dirs ending in period or space | |
286 | d = d[:-1] + b'_' |
|
286 | d = d[:-1] + b'_' | |
287 | if sdirslen == 0: |
|
287 | if sdirslen == 0: | |
288 | t = len(d) |
|
288 | t = len(d) | |
289 | else: |
|
289 | else: | |
290 | t = sdirslen + 1 + len(d) |
|
290 | t = sdirslen + 1 + len(d) | |
291 | if t > _maxshortdirslen: |
|
291 | if t > _maxshortdirslen: | |
292 | break |
|
292 | break | |
293 | sdirs.append(d) |
|
293 | sdirs.append(d) | |
294 | sdirslen = t |
|
294 | sdirslen = t | |
295 | dirs = b'/'.join(sdirs) |
|
295 | dirs = b'/'.join(sdirs) | |
296 | if len(dirs) > 0: |
|
296 | if len(dirs) > 0: | |
297 | dirs += b'/' |
|
297 | dirs += b'/' | |
298 | res = b'dh/' + dirs + digest + ext |
|
298 | res = b'dh/' + dirs + digest + ext | |
299 | spaceleft = _maxstorepathlen - len(res) |
|
299 | spaceleft = _maxstorepathlen - len(res) | |
300 | if spaceleft > 0: |
|
300 | if spaceleft > 0: | |
301 | filler = basename[:spaceleft] |
|
301 | filler = basename[:spaceleft] | |
302 | res = b'dh/' + dirs + filler + digest + ext |
|
302 | res = b'dh/' + dirs + filler + digest + ext | |
303 | return res |
|
303 | return res | |
304 |
|
304 | |||
305 |
|
305 | |||
306 | def _hybridencode(path, dotencode): |
|
306 | def _hybridencode(path, dotencode): | |
307 | """encodes path with a length limit |
|
307 | """encodes path with a length limit | |
308 |
|
308 | |||
309 | Encodes all paths that begin with 'data/', according to the following. |
|
309 | Encodes all paths that begin with 'data/', according to the following. | |
310 |
|
310 | |||
311 | Default encoding (reversible): |
|
311 | Default encoding (reversible): | |
312 |
|
312 | |||
313 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal |
|
313 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal | |
314 | 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 | |
315 | of the character (see encodefilename). |
|
315 | of the character (see encodefilename). | |
316 | Relevant path components consisting of Windows reserved filenames are |
|
316 | Relevant path components consisting of Windows reserved filenames are | |
317 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). |
|
317 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). | |
318 |
|
318 | |||
319 | Hashed encoding (not reversible): |
|
319 | Hashed encoding (not reversible): | |
320 |
|
320 | |||
321 | If the default-encoded path is longer than _maxstorepathlen, a |
|
321 | If the default-encoded path is longer than _maxstorepathlen, a | |
322 | non-reversible hybrid hashing of the path is done instead. |
|
322 | non-reversible hybrid hashing of the path is done instead. | |
323 | This encoding uses up to _dirprefixlen characters of all directory |
|
323 | This encoding uses up to _dirprefixlen characters of all directory | |
324 | 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 | |
325 | _maxshortdirslen. |
|
325 | _maxshortdirslen. | |
326 | 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. | |
327 | 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 | |
328 | (the basename is everything after the last path separator). The filler |
|
328 | (the basename is everything after the last path separator). The filler | |
329 | 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 | |
330 | the encoded path has _maxstorepathlen characters (or all chars of the |
|
330 | the encoded path has _maxstorepathlen characters (or all chars of the | |
331 | basename have been taken). |
|
331 | basename have been taken). | |
332 | The extension (e.g. '.i' or '.d') is preserved. |
|
332 | The extension (e.g. '.i' or '.d') is preserved. | |
333 |
|
333 | |||
334 | 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 | |
335 | encoding was used. |
|
335 | encoding was used. | |
336 | """ |
|
336 | """ | |
337 | path = encodedir(path) |
|
337 | path = encodedir(path) | |
338 | ef = _encodefname(path).split(b'/') |
|
338 | ef = _encodefname(path).split(b'/') | |
339 | res = b'/'.join(_auxencode(ef, dotencode)) |
|
339 | res = b'/'.join(_auxencode(ef, dotencode)) | |
340 | if len(res) > _maxstorepathlen: |
|
340 | if len(res) > _maxstorepathlen: | |
341 | res = _hashencode(path, dotencode) |
|
341 | res = _hashencode(path, dotencode) | |
342 | return res |
|
342 | return res | |
343 |
|
343 | |||
344 |
|
344 | |||
345 | def _pathencode(path): |
|
345 | def _pathencode(path): | |
346 | de = encodedir(path) |
|
346 | de = encodedir(path) | |
347 | if len(path) > _maxstorepathlen: |
|
347 | if len(path) > _maxstorepathlen: | |
348 | return _hashencode(de, True) |
|
348 | return _hashencode(de, True) | |
349 | ef = _encodefname(de).split(b'/') |
|
349 | ef = _encodefname(de).split(b'/') | |
350 | res = b'/'.join(_auxencode(ef, True)) |
|
350 | res = b'/'.join(_auxencode(ef, True)) | |
351 | if len(res) > _maxstorepathlen: |
|
351 | if len(res) > _maxstorepathlen: | |
352 | return _hashencode(de, True) |
|
352 | return _hashencode(de, True) | |
353 | return res |
|
353 | return res | |
354 |
|
354 | |||
355 |
|
355 | |||
356 | _pathencode = getattr(parsers, 'pathencode', _pathencode) |
|
356 | _pathencode = getattr(parsers, 'pathencode', _pathencode) | |
357 |
|
357 | |||
358 |
|
358 | |||
359 | def _plainhybridencode(f): |
|
359 | def _plainhybridencode(f): | |
360 | return _hybridencode(f, False) |
|
360 | return _hybridencode(f, False) | |
361 |
|
361 | |||
362 |
|
362 | |||
363 | def _calcmode(vfs): |
|
363 | def _calcmode(vfs): | |
364 | try: |
|
364 | try: | |
365 | # files in .hg/ will be created using this mode |
|
365 | # files in .hg/ will be created using this mode | |
366 | mode = vfs.stat().st_mode |
|
366 | mode = vfs.stat().st_mode | |
367 | # avoid some useless chmods |
|
367 | # avoid some useless chmods | |
368 | if (0o777 & ~util.umask) == (0o777 & mode): |
|
368 | if (0o777 & ~util.umask) == (0o777 & mode): | |
369 | mode = None |
|
369 | mode = None | |
370 | except OSError: |
|
370 | except OSError: | |
371 | mode = None |
|
371 | mode = None | |
372 | return mode |
|
372 | return mode | |
373 |
|
373 | |||
374 |
|
374 | |||
375 | _data = [ |
|
375 | _data = [ | |
376 | b'bookmarks', |
|
376 | b'bookmarks', | |
377 | b'narrowspec', |
|
377 | b'narrowspec', | |
378 | b'data', |
|
378 | b'data', | |
379 | b'meta', |
|
379 | b'meta', | |
380 | b'00manifest.d', |
|
380 | b'00manifest.d', | |
381 | b'00manifest.i', |
|
381 | b'00manifest.i', | |
382 | b'00changelog.d', |
|
382 | b'00changelog.d', | |
383 | b'00changelog.i', |
|
383 | b'00changelog.i', | |
384 | b'phaseroots', |
|
384 | b'phaseroots', | |
385 | b'obsstore', |
|
385 | b'obsstore', | |
386 | b'requires', |
|
386 | b'requires', | |
387 | ] |
|
387 | ] | |
388 |
|
388 | |||
389 | REVLOG_FILES_MAIN_EXT = (b'.i',) |
|
389 | REVLOG_FILES_MAIN_EXT = (b'.i',) | |
390 | REVLOG_FILES_OTHER_EXT = ( |
|
390 | REVLOG_FILES_OTHER_EXT = ( | |
391 | b'.idx', |
|
391 | b'.idx', | |
392 | b'.d', |
|
392 | b'.d', | |
393 | b'.dat', |
|
393 | b'.dat', | |
394 | b'.n', |
|
394 | b'.n', | |
395 | b'.nd', |
|
395 | b'.nd', | |
396 | b'.sda', |
|
396 | b'.sda', | |
397 | ) |
|
397 | ) | |
398 | # files that are "volatile" and might change between listing and streaming |
|
398 | # files that are "volatile" and might change between listing and streaming | |
399 | # |
|
399 | # | |
400 | # 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 | |
401 | # deleted. |
|
401 | # deleted. | |
402 | REVLOG_FILES_VOLATILE_EXT = (b'.n', b'.nd') |
|
402 | REVLOG_FILES_VOLATILE_EXT = (b'.n', b'.nd') | |
403 |
|
403 | |||
404 | # some exception to the above matching |
|
404 | # some exception to the above matching | |
405 | # |
|
405 | # | |
406 | # XXX This is currently not in use because of issue6542 |
|
406 | # XXX This is currently not in use because of issue6542 | |
407 | EXCLUDED = re.compile(br'.*undo\.[^/]+\.(nd?|i)$') |
|
407 | EXCLUDED = re.compile(br'.*undo\.[^/]+\.(nd?|i)$') | |
408 |
|
408 | |||
409 |
|
409 | |||
410 | def is_revlog(f, kind, st): |
|
410 | def is_revlog(f, kind, st): | |
411 | if kind != stat.S_IFREG: |
|
411 | if kind != stat.S_IFREG: | |
412 | return None |
|
412 | return None | |
413 | return revlog_type(f) |
|
413 | return revlog_type(f) | |
414 |
|
414 | |||
415 |
|
415 | |||
416 | def revlog_type(f): |
|
416 | def revlog_type(f): | |
417 | # 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 | |
418 | # 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 | |
419 | # issue6542. So we no longer use EXCLUDED. |
|
419 | # issue6542. So we no longer use EXCLUDED. | |
420 | if f.endswith(REVLOG_FILES_MAIN_EXT): |
|
420 | if f.endswith(REVLOG_FILES_MAIN_EXT): | |
421 | return FILEFLAGS_REVLOG_MAIN |
|
421 | return FILEFLAGS_REVLOG_MAIN | |
422 | elif f.endswith(REVLOG_FILES_OTHER_EXT): |
|
422 | elif f.endswith(REVLOG_FILES_OTHER_EXT): | |
423 | t = FILETYPE_FILELOG_OTHER |
|
423 | t = FILETYPE_FILELOG_OTHER | |
424 | if f.endswith(REVLOG_FILES_VOLATILE_EXT): |
|
424 | if f.endswith(REVLOG_FILES_VOLATILE_EXT): | |
425 | t |= FILEFLAGS_VOLATILE |
|
425 | t |= FILEFLAGS_VOLATILE | |
426 | return t |
|
426 | return t | |
427 | return None |
|
427 | return None | |
428 |
|
428 | |||
429 |
|
429 | |||
430 | # the file is part of changelog data |
|
430 | # the file is part of changelog data | |
431 | FILEFLAGS_CHANGELOG = 1 << 13 |
|
431 | FILEFLAGS_CHANGELOG = 1 << 13 | |
432 | # the file is part of manifest data |
|
432 | # the file is part of manifest data | |
433 | FILEFLAGS_MANIFESTLOG = 1 << 12 |
|
433 | FILEFLAGS_MANIFESTLOG = 1 << 12 | |
434 | # the file is part of filelog data |
|
434 | # the file is part of filelog data | |
435 | FILEFLAGS_FILELOG = 1 << 11 |
|
435 | FILEFLAGS_FILELOG = 1 << 11 | |
436 | # file that are not directly part of a revlog |
|
436 | # file that are not directly part of a revlog | |
437 | FILEFLAGS_OTHER = 1 << 10 |
|
437 | FILEFLAGS_OTHER = 1 << 10 | |
438 |
|
438 | |||
439 | # the main entry point for a revlog |
|
439 | # the main entry point for a revlog | |
440 | FILEFLAGS_REVLOG_MAIN = 1 << 1 |
|
440 | FILEFLAGS_REVLOG_MAIN = 1 << 1 | |
441 | # a secondary file for a revlog |
|
441 | # a secondary file for a revlog | |
442 | FILEFLAGS_REVLOG_OTHER = 1 << 0 |
|
442 | FILEFLAGS_REVLOG_OTHER = 1 << 0 | |
443 |
|
443 | |||
444 | # files that are "volatile" and might change between listing and streaming |
|
444 | # files that are "volatile" and might change between listing and streaming | |
445 | FILEFLAGS_VOLATILE = 1 << 20 |
|
445 | FILEFLAGS_VOLATILE = 1 << 20 | |
446 |
|
446 | |||
447 | FILETYPE_CHANGELOG_MAIN = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_MAIN |
|
447 | FILETYPE_CHANGELOG_MAIN = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_MAIN | |
448 | FILETYPE_CHANGELOG_OTHER = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_OTHER |
|
448 | FILETYPE_CHANGELOG_OTHER = FILEFLAGS_CHANGELOG | FILEFLAGS_REVLOG_OTHER | |
449 | FILETYPE_MANIFESTLOG_MAIN = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_MAIN |
|
449 | FILETYPE_MANIFESTLOG_MAIN = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_MAIN | |
450 | FILETYPE_MANIFESTLOG_OTHER = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_OTHER |
|
450 | FILETYPE_MANIFESTLOG_OTHER = FILEFLAGS_MANIFESTLOG | FILEFLAGS_REVLOG_OTHER | |
451 | FILETYPE_FILELOG_MAIN = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_MAIN |
|
451 | FILETYPE_FILELOG_MAIN = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_MAIN | |
452 | FILETYPE_FILELOG_OTHER = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_OTHER |
|
452 | FILETYPE_FILELOG_OTHER = FILEFLAGS_FILELOG | FILEFLAGS_REVLOG_OTHER | |
453 | FILETYPE_OTHER = FILEFLAGS_OTHER |
|
453 | FILETYPE_OTHER = FILEFLAGS_OTHER | |
454 |
|
454 | |||
455 |
|
455 | |||
456 | @attr.s(slots=True, init=False) |
|
456 | @attr.s(slots=True, init=False) | |
457 | class BaseStoreEntry: |
|
457 | class BaseStoreEntry: | |
458 | """An entry in the store |
|
458 | """An entry in the store | |
459 |
|
459 | |||
460 | This is returned by `store.walk` and represent some data in the store.""" |
|
460 | This is returned by `store.walk` and represent some data in the store.""" | |
461 |
|
461 | |||
462 | unencoded_path = attr.ib() |
|
462 | unencoded_path = attr.ib() | |
463 | _is_volatile = attr.ib(default=False) |
|
463 | _is_volatile = attr.ib(default=False) | |
464 | _file_size = attr.ib(default=None) |
|
464 | _file_size = attr.ib(default=None) | |
465 |
|
465 | |||
466 | def __init__( |
|
466 | def __init__( | |
467 | self, |
|
467 | self, | |
468 | unencoded_path, |
|
468 | unencoded_path, | |
469 | is_volatile=False, |
|
469 | is_volatile=False, | |
470 | file_size=None, |
|
470 | file_size=None, | |
471 | ): |
|
471 | ): | |
472 | self.unencoded_path = unencoded_path |
|
472 | self.unencoded_path = unencoded_path | |
473 | self._is_volatile = is_volatile |
|
473 | self._is_volatile = is_volatile | |
474 | self._file_size = file_size |
|
474 | self._file_size = file_size | |
475 |
|
475 | |||
476 | def files(self): |
|
476 | def files(self): | |
477 | return [ |
|
477 | return [ | |
478 | StoreFile( |
|
478 | StoreFile( | |
479 | unencoded_path=self.unencoded_path, |
|
479 | unencoded_path=self.unencoded_path, | |
480 | file_size=self._file_size, |
|
480 | file_size=self._file_size, | |
481 | is_volatile=self._is_volatile, |
|
481 | is_volatile=self._is_volatile, | |
482 | ) |
|
482 | ) | |
483 | ] |
|
483 | ] | |
484 |
|
484 | |||
485 |
|
485 | |||
486 | @attr.s(slots=True, init=False) |
|
486 | @attr.s(slots=True, init=False) | |
487 | class SimpleStoreEntry(BaseStoreEntry): |
|
487 | class SimpleStoreEntry(BaseStoreEntry): | |
488 | """A generic entry in the store""" |
|
488 | """A generic entry in the store""" | |
489 |
|
489 | |||
490 | is_revlog = False |
|
490 | is_revlog = False | |
491 |
|
491 | |||
492 |
|
492 | |||
493 | @attr.s(slots=True, init=False) |
|
493 | @attr.s(slots=True, init=False) | |
494 | class RevlogStoreEntry(BaseStoreEntry): |
|
494 | class RevlogStoreEntry(BaseStoreEntry): | |
495 | """A revlog entry in the store""" |
|
495 | """A revlog entry in the store""" | |
496 |
|
496 | |||
497 | is_revlog = True |
|
497 | is_revlog = True | |
498 | revlog_type = attr.ib(default=None) |
|
498 | revlog_type = attr.ib(default=None) | |
499 | is_revlog_main = attr.ib(default=None) |
|
499 | is_revlog_main = attr.ib(default=None) | |
500 |
|
500 | |||
501 | def __init__( |
|
501 | def __init__( | |
502 | self, |
|
502 | self, | |
503 | unencoded_path, |
|
503 | unencoded_path, | |
504 | revlog_type, |
|
504 | revlog_type, | |
505 | is_revlog_main=False, |
|
505 | is_revlog_main=False, | |
506 | is_volatile=False, |
|
506 | is_volatile=False, | |
507 | file_size=None, |
|
507 | file_size=None, | |
508 | ): |
|
508 | ): | |
509 | super().__init__( |
|
509 | super().__init__( | |
510 | unencoded_path=unencoded_path, |
|
510 | unencoded_path=unencoded_path, | |
511 | is_volatile=is_volatile, |
|
511 | is_volatile=is_volatile, | |
512 | file_size=file_size, |
|
512 | file_size=file_size, | |
513 | ) |
|
513 | ) | |
514 | self.revlog_type = revlog_type |
|
514 | self.revlog_type = revlog_type | |
515 | self.is_revlog_main = is_revlog_main |
|
515 | self.is_revlog_main = is_revlog_main | |
516 |
|
516 | |||
517 |
|
517 | |||
518 | @attr.s(slots=True) |
|
518 | @attr.s(slots=True) | |
519 | class StoreFile: |
|
519 | class StoreFile: | |
520 | """a file matching an entry""" |
|
520 | """a file matching an entry""" | |
521 |
|
521 | |||
522 | unencoded_path = attr.ib() |
|
522 | unencoded_path = attr.ib() | |
523 | file_size = attr.ib() |
|
523 | _file_size = attr.ib(default=False) | |
524 | is_volatile = attr.ib(default=False) |
|
524 | is_volatile = attr.ib(default=False) | |
525 |
|
525 | |||
|
526 | def file_size(self, vfs): | |||
|
527 | if self._file_size is not None: | |||
|
528 | return self._file_size | |||
|
529 | try: | |||
|
530 | return vfs.stat(self.unencoded_path).st_size | |||
|
531 | except FileNotFoundError: | |||
|
532 | return 0 | |||
|
533 | ||||
526 |
|
534 | |||
527 | class basicstore: |
|
535 | class basicstore: | |
528 | '''base class for local repository stores''' |
|
536 | '''base class for local repository stores''' | |
529 |
|
537 | |||
530 | def __init__(self, path, vfstype): |
|
538 | def __init__(self, path, vfstype): | |
531 | vfs = vfstype(path) |
|
539 | vfs = vfstype(path) | |
532 | self.path = vfs.base |
|
540 | self.path = vfs.base | |
533 | self.createmode = _calcmode(vfs) |
|
541 | self.createmode = _calcmode(vfs) | |
534 | vfs.createmode = self.createmode |
|
542 | vfs.createmode = self.createmode | |
535 | self.rawvfs = vfs |
|
543 | self.rawvfs = vfs | |
536 | self.vfs = vfsmod.filtervfs(vfs, encodedir) |
|
544 | self.vfs = vfsmod.filtervfs(vfs, encodedir) | |
537 | self.opener = self.vfs |
|
545 | self.opener = self.vfs | |
538 |
|
546 | |||
539 | def join(self, f): |
|
547 | def join(self, f): | |
540 | return self.path + b'/' + encodedir(f) |
|
548 | return self.path + b'/' + encodedir(f) | |
541 |
|
549 | |||
542 | def _walk(self, relpath, recurse): |
|
550 | def _walk(self, relpath, recurse): | |
543 | '''yields (revlog_type, unencoded, size)''' |
|
551 | '''yields (revlog_type, unencoded, size)''' | |
544 | path = self.path |
|
552 | path = self.path | |
545 | if relpath: |
|
553 | if relpath: | |
546 | path += b'/' + relpath |
|
554 | path += b'/' + relpath | |
547 | striplen = len(self.path) + 1 |
|
555 | striplen = len(self.path) + 1 | |
548 | l = [] |
|
556 | l = [] | |
549 | if self.rawvfs.isdir(path): |
|
557 | if self.rawvfs.isdir(path): | |
550 | visit = [path] |
|
558 | visit = [path] | |
551 | readdir = self.rawvfs.readdir |
|
559 | readdir = self.rawvfs.readdir | |
552 | while visit: |
|
560 | while visit: | |
553 | p = visit.pop() |
|
561 | p = visit.pop() | |
554 | for f, kind, st in readdir(p, stat=True): |
|
562 | for f, kind, st in readdir(p, stat=True): | |
555 | fp = p + b'/' + f |
|
563 | fp = p + b'/' + f | |
556 | rl_type = is_revlog(f, kind, st) |
|
564 | rl_type = is_revlog(f, kind, st) | |
557 | if rl_type is not None: |
|
565 | if rl_type is not None: | |
558 | n = util.pconvert(fp[striplen:]) |
|
566 | n = util.pconvert(fp[striplen:]) | |
559 | l.append((rl_type, decodedir(n), st.st_size)) |
|
567 | l.append((rl_type, decodedir(n), st.st_size)) | |
560 | elif kind == stat.S_IFDIR and recurse: |
|
568 | elif kind == stat.S_IFDIR and recurse: | |
561 | visit.append(fp) |
|
569 | visit.append(fp) | |
562 | l.sort() |
|
570 | l.sort() | |
563 | return l |
|
571 | return l | |
564 |
|
572 | |||
565 | def changelog(self, trypending, concurrencychecker=None): |
|
573 | def changelog(self, trypending, concurrencychecker=None): | |
566 | return changelog.changelog( |
|
574 | return changelog.changelog( | |
567 | self.vfs, |
|
575 | self.vfs, | |
568 | trypending=trypending, |
|
576 | trypending=trypending, | |
569 | concurrencychecker=concurrencychecker, |
|
577 | concurrencychecker=concurrencychecker, | |
570 | ) |
|
578 | ) | |
571 |
|
579 | |||
572 | def manifestlog(self, repo, storenarrowmatch): |
|
580 | def manifestlog(self, repo, storenarrowmatch): | |
573 | rootstore = manifest.manifestrevlog(repo.nodeconstants, self.vfs) |
|
581 | rootstore = manifest.manifestrevlog(repo.nodeconstants, self.vfs) | |
574 | return manifest.manifestlog(self.vfs, repo, rootstore, storenarrowmatch) |
|
582 | return manifest.manifestlog(self.vfs, repo, rootstore, storenarrowmatch) | |
575 |
|
583 | |||
576 | def datafiles( |
|
584 | def datafiles( | |
577 | self, matcher=None, undecodable=None |
|
585 | self, matcher=None, undecodable=None | |
578 | ) -> Generator[BaseStoreEntry, None, None]: |
|
586 | ) -> Generator[BaseStoreEntry, None, None]: | |
579 | """Like walk, but excluding the changelog and root manifest. |
|
587 | """Like walk, but excluding the changelog and root manifest. | |
580 |
|
588 | |||
581 | When [undecodable] is None, revlogs names that can't be |
|
589 | When [undecodable] is None, revlogs names that can't be | |
582 | decoded cause an exception. When it is provided, it should |
|
590 | decoded cause an exception. When it is provided, it should | |
583 | be a list and the filenames that can't be decoded are added |
|
591 | be a list and the filenames that can't be decoded are added | |
584 | to it instead. This is very rarely needed.""" |
|
592 | to it instead. This is very rarely needed.""" | |
585 | files = self._walk(b'data', True) + self._walk(b'meta', True) |
|
593 | files = self._walk(b'data', True) + self._walk(b'meta', True) | |
586 | for (t, u, s) in files: |
|
594 | for (t, u, s) in files: | |
587 | if t is not None: |
|
595 | if t is not None: | |
588 | yield RevlogStoreEntry( |
|
596 | yield RevlogStoreEntry( | |
589 | unencoded_path=u, |
|
597 | unencoded_path=u, | |
590 | revlog_type=FILEFLAGS_FILELOG, |
|
598 | revlog_type=FILEFLAGS_FILELOG, | |
591 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), |
|
599 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), | |
592 | is_volatile=bool(t & FILEFLAGS_VOLATILE), |
|
600 | is_volatile=bool(t & FILEFLAGS_VOLATILE), | |
593 | file_size=s, |
|
601 | file_size=s, | |
594 | ) |
|
602 | ) | |
595 |
|
603 | |||
596 | def topfiles(self) -> Generator[BaseStoreEntry, None, None]: |
|
604 | def topfiles(self) -> Generator[BaseStoreEntry, None, None]: | |
597 | # yield manifest before changelog |
|
605 | # yield manifest before changelog | |
598 | files = reversed(self._walk(b'', False)) |
|
606 | files = reversed(self._walk(b'', False)) | |
599 | for (t, u, s) in files: |
|
607 | for (t, u, s) in files: | |
600 | if u.startswith(b'00changelog'): |
|
608 | if u.startswith(b'00changelog'): | |
601 | yield RevlogStoreEntry( |
|
609 | yield RevlogStoreEntry( | |
602 | unencoded_path=u, |
|
610 | unencoded_path=u, | |
603 | revlog_type=FILEFLAGS_CHANGELOG, |
|
611 | revlog_type=FILEFLAGS_CHANGELOG, | |
604 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), |
|
612 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), | |
605 | is_volatile=bool(t & FILEFLAGS_VOLATILE), |
|
613 | is_volatile=bool(t & FILEFLAGS_VOLATILE), | |
606 | file_size=s, |
|
614 | file_size=s, | |
607 | ) |
|
615 | ) | |
608 | elif u.startswith(b'00manifest'): |
|
616 | elif u.startswith(b'00manifest'): | |
609 | yield RevlogStoreEntry( |
|
617 | yield RevlogStoreEntry( | |
610 | unencoded_path=u, |
|
618 | unencoded_path=u, | |
611 | revlog_type=FILEFLAGS_MANIFESTLOG, |
|
619 | revlog_type=FILEFLAGS_MANIFESTLOG, | |
612 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), |
|
620 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), | |
613 | is_volatile=bool(t & FILEFLAGS_VOLATILE), |
|
621 | is_volatile=bool(t & FILEFLAGS_VOLATILE), | |
614 | file_size=s, |
|
622 | file_size=s, | |
615 | ) |
|
623 | ) | |
616 | else: |
|
624 | else: | |
617 | yield SimpleStoreEntry( |
|
625 | yield SimpleStoreEntry( | |
618 | unencoded_path=u, |
|
626 | unencoded_path=u, | |
619 | is_volatile=bool(t & FILEFLAGS_VOLATILE), |
|
627 | is_volatile=bool(t & FILEFLAGS_VOLATILE), | |
620 | file_size=s, |
|
628 | file_size=s, | |
621 | ) |
|
629 | ) | |
622 |
|
630 | |||
623 | def walk(self, matcher=None) -> Generator[BaseStoreEntry, None, None]: |
|
631 | def walk(self, matcher=None) -> Generator[BaseStoreEntry, None, None]: | |
624 | """return files related to data storage (ie: revlogs) |
|
632 | """return files related to data storage (ie: revlogs) | |
625 |
|
633 | |||
626 | yields (file_type, unencoded, size) |
|
634 | yields (file_type, unencoded, size) | |
627 |
|
635 | |||
628 | if a matcher is passed, storage files of only those tracked paths |
|
636 | if a matcher is passed, storage files of only those tracked paths | |
629 | are passed with matches the matcher |
|
637 | are passed with matches the matcher | |
630 | """ |
|
638 | """ | |
631 | # yield data files first |
|
639 | # yield data files first | |
632 | for x in self.datafiles(matcher): |
|
640 | for x in self.datafiles(matcher): | |
633 | yield x |
|
641 | yield x | |
634 | for x in self.topfiles(): |
|
642 | for x in self.topfiles(): | |
635 | yield x |
|
643 | yield x | |
636 |
|
644 | |||
637 | def copylist(self): |
|
645 | def copylist(self): | |
638 | return _data |
|
646 | return _data | |
639 |
|
647 | |||
640 | def write(self, tr): |
|
648 | def write(self, tr): | |
641 | pass |
|
649 | pass | |
642 |
|
650 | |||
643 | def invalidatecaches(self): |
|
651 | def invalidatecaches(self): | |
644 | pass |
|
652 | pass | |
645 |
|
653 | |||
646 | def markremoved(self, fn): |
|
654 | def markremoved(self, fn): | |
647 | pass |
|
655 | pass | |
648 |
|
656 | |||
649 | def __contains__(self, path): |
|
657 | def __contains__(self, path): | |
650 | '''Checks if the store contains path''' |
|
658 | '''Checks if the store contains path''' | |
651 | path = b"/".join((b"data", path)) |
|
659 | path = b"/".join((b"data", path)) | |
652 | # file? |
|
660 | # file? | |
653 | if self.vfs.exists(path + b".i"): |
|
661 | if self.vfs.exists(path + b".i"): | |
654 | return True |
|
662 | return True | |
655 | # dir? |
|
663 | # dir? | |
656 | if not path.endswith(b"/"): |
|
664 | if not path.endswith(b"/"): | |
657 | path = path + b"/" |
|
665 | path = path + b"/" | |
658 | return self.vfs.exists(path) |
|
666 | return self.vfs.exists(path) | |
659 |
|
667 | |||
660 |
|
668 | |||
661 | class encodedstore(basicstore): |
|
669 | class encodedstore(basicstore): | |
662 | def __init__(self, path, vfstype): |
|
670 | def __init__(self, path, vfstype): | |
663 | vfs = vfstype(path + b'/store') |
|
671 | vfs = vfstype(path + b'/store') | |
664 | self.path = vfs.base |
|
672 | self.path = vfs.base | |
665 | self.createmode = _calcmode(vfs) |
|
673 | self.createmode = _calcmode(vfs) | |
666 | vfs.createmode = self.createmode |
|
674 | vfs.createmode = self.createmode | |
667 | self.rawvfs = vfs |
|
675 | self.rawvfs = vfs | |
668 | self.vfs = vfsmod.filtervfs(vfs, encodefilename) |
|
676 | self.vfs = vfsmod.filtervfs(vfs, encodefilename) | |
669 | self.opener = self.vfs |
|
677 | self.opener = self.vfs | |
670 |
|
678 | |||
671 | # note: topfiles would also need a decode phase. It is just that in |
|
679 | # note: topfiles would also need a decode phase. It is just that in | |
672 | # practice we do not have any file outside of `data/` that needs encoding. |
|
680 | # practice we do not have any file outside of `data/` that needs encoding. | |
673 | # However that might change so we should probably add a test and encoding |
|
681 | # However that might change so we should probably add a test and encoding | |
674 | # decoding for it too. see issue6548 |
|
682 | # decoding for it too. see issue6548 | |
675 |
|
683 | |||
676 | def datafiles( |
|
684 | def datafiles( | |
677 | self, matcher=None, undecodable=None |
|
685 | self, matcher=None, undecodable=None | |
678 | ) -> Generator[BaseStoreEntry, None, None]: |
|
686 | ) -> Generator[BaseStoreEntry, None, None]: | |
679 | for entry in super(encodedstore, self).datafiles(): |
|
687 | for entry in super(encodedstore, self).datafiles(): | |
680 | try: |
|
688 | try: | |
681 | f1 = entry.unencoded_path |
|
689 | f1 = entry.unencoded_path | |
682 | f2 = decodefilename(f1) |
|
690 | f2 = decodefilename(f1) | |
683 | except KeyError: |
|
691 | except KeyError: | |
684 | if undecodable is None: |
|
692 | if undecodable is None: | |
685 | msg = _(b'undecodable revlog name %s') % f1 |
|
693 | msg = _(b'undecodable revlog name %s') % f1 | |
686 | raise error.StorageError(msg) |
|
694 | raise error.StorageError(msg) | |
687 | else: |
|
695 | else: | |
688 | undecodable.append(f1) |
|
696 | undecodable.append(f1) | |
689 | continue |
|
697 | continue | |
690 | if not _matchtrackedpath(f2, matcher): |
|
698 | if not _matchtrackedpath(f2, matcher): | |
691 | continue |
|
699 | continue | |
692 | entry.unencoded_path = f2 |
|
700 | entry.unencoded_path = f2 | |
693 | yield entry |
|
701 | yield entry | |
694 |
|
702 | |||
695 | def join(self, f): |
|
703 | def join(self, f): | |
696 | return self.path + b'/' + encodefilename(f) |
|
704 | return self.path + b'/' + encodefilename(f) | |
697 |
|
705 | |||
698 | def copylist(self): |
|
706 | def copylist(self): | |
699 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in _data] |
|
707 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in _data] | |
700 |
|
708 | |||
701 |
|
709 | |||
702 | class fncache: |
|
710 | class fncache: | |
703 | # the filename used to be partially encoded |
|
711 | # the filename used to be partially encoded | |
704 | # hence the encodedir/decodedir dance |
|
712 | # hence the encodedir/decodedir dance | |
705 | def __init__(self, vfs): |
|
713 | def __init__(self, vfs): | |
706 | self.vfs = vfs |
|
714 | self.vfs = vfs | |
707 | self._ignores = set() |
|
715 | self._ignores = set() | |
708 | self.entries = None |
|
716 | self.entries = None | |
709 | self._dirty = False |
|
717 | self._dirty = False | |
710 | # set of new additions to fncache |
|
718 | # set of new additions to fncache | |
711 | self.addls = set() |
|
719 | self.addls = set() | |
712 |
|
720 | |||
713 | def ensureloaded(self, warn=None): |
|
721 | def ensureloaded(self, warn=None): | |
714 | """read the fncache file if not already read. |
|
722 | """read the fncache file if not already read. | |
715 |
|
723 | |||
716 | If the file on disk is corrupted, raise. If warn is provided, |
|
724 | If the file on disk is corrupted, raise. If warn is provided, | |
717 | warn and keep going instead.""" |
|
725 | warn and keep going instead.""" | |
718 | if self.entries is None: |
|
726 | if self.entries is None: | |
719 | self._load(warn) |
|
727 | self._load(warn) | |
720 |
|
728 | |||
721 | def _load(self, warn=None): |
|
729 | def _load(self, warn=None): | |
722 | '''fill the entries from the fncache file''' |
|
730 | '''fill the entries from the fncache file''' | |
723 | self._dirty = False |
|
731 | self._dirty = False | |
724 | try: |
|
732 | try: | |
725 | fp = self.vfs(b'fncache', mode=b'rb') |
|
733 | fp = self.vfs(b'fncache', mode=b'rb') | |
726 | except IOError: |
|
734 | except IOError: | |
727 | # skip nonexistent file |
|
735 | # skip nonexistent file | |
728 | self.entries = set() |
|
736 | self.entries = set() | |
729 | return |
|
737 | return | |
730 |
|
738 | |||
731 | self.entries = set() |
|
739 | self.entries = set() | |
732 | chunk = b'' |
|
740 | chunk = b'' | |
733 | for c in iter(functools.partial(fp.read, fncache_chunksize), b''): |
|
741 | for c in iter(functools.partial(fp.read, fncache_chunksize), b''): | |
734 | chunk += c |
|
742 | chunk += c | |
735 | try: |
|
743 | try: | |
736 | p = chunk.rindex(b'\n') |
|
744 | p = chunk.rindex(b'\n') | |
737 | self.entries.update(decodedir(chunk[: p + 1]).splitlines()) |
|
745 | self.entries.update(decodedir(chunk[: p + 1]).splitlines()) | |
738 | chunk = chunk[p + 1 :] |
|
746 | chunk = chunk[p + 1 :] | |
739 | except ValueError: |
|
747 | except ValueError: | |
740 | # substring '\n' not found, maybe the entry is bigger than the |
|
748 | # substring '\n' not found, maybe the entry is bigger than the | |
741 | # chunksize, so let's keep iterating |
|
749 | # chunksize, so let's keep iterating | |
742 | pass |
|
750 | pass | |
743 |
|
751 | |||
744 | if chunk: |
|
752 | if chunk: | |
745 | msg = _(b"fncache does not ends with a newline") |
|
753 | msg = _(b"fncache does not ends with a newline") | |
746 | if warn: |
|
754 | if warn: | |
747 | warn(msg + b'\n') |
|
755 | warn(msg + b'\n') | |
748 | else: |
|
756 | else: | |
749 | raise error.Abort( |
|
757 | raise error.Abort( | |
750 | msg, |
|
758 | msg, | |
751 | hint=_( |
|
759 | hint=_( | |
752 | b"use 'hg debugrebuildfncache' to " |
|
760 | b"use 'hg debugrebuildfncache' to " | |
753 | b"rebuild the fncache" |
|
761 | b"rebuild the fncache" | |
754 | ), |
|
762 | ), | |
755 | ) |
|
763 | ) | |
756 | self._checkentries(fp, warn) |
|
764 | self._checkentries(fp, warn) | |
757 | fp.close() |
|
765 | fp.close() | |
758 |
|
766 | |||
759 | def _checkentries(self, fp, warn): |
|
767 | def _checkentries(self, fp, warn): | |
760 | """make sure there is no empty string in entries""" |
|
768 | """make sure there is no empty string in entries""" | |
761 | if b'' in self.entries: |
|
769 | if b'' in self.entries: | |
762 | fp.seek(0) |
|
770 | fp.seek(0) | |
763 | for n, line in enumerate(fp): |
|
771 | for n, line in enumerate(fp): | |
764 | if not line.rstrip(b'\n'): |
|
772 | if not line.rstrip(b'\n'): | |
765 | t = _(b'invalid entry in fncache, line %d') % (n + 1) |
|
773 | t = _(b'invalid entry in fncache, line %d') % (n + 1) | |
766 | if warn: |
|
774 | if warn: | |
767 | warn(t + b'\n') |
|
775 | warn(t + b'\n') | |
768 | else: |
|
776 | else: | |
769 | raise error.Abort(t) |
|
777 | raise error.Abort(t) | |
770 |
|
778 | |||
771 | def write(self, tr): |
|
779 | def write(self, tr): | |
772 | if self._dirty: |
|
780 | if self._dirty: | |
773 | assert self.entries is not None |
|
781 | assert self.entries is not None | |
774 | self.entries = self.entries | self.addls |
|
782 | self.entries = self.entries | self.addls | |
775 | self.addls = set() |
|
783 | self.addls = set() | |
776 | tr.addbackup(b'fncache') |
|
784 | tr.addbackup(b'fncache') | |
777 | fp = self.vfs(b'fncache', mode=b'wb', atomictemp=True) |
|
785 | fp = self.vfs(b'fncache', mode=b'wb', atomictemp=True) | |
778 | if self.entries: |
|
786 | if self.entries: | |
779 | fp.write(encodedir(b'\n'.join(self.entries) + b'\n')) |
|
787 | fp.write(encodedir(b'\n'.join(self.entries) + b'\n')) | |
780 | fp.close() |
|
788 | fp.close() | |
781 | self._dirty = False |
|
789 | self._dirty = False | |
782 | if self.addls: |
|
790 | if self.addls: | |
783 | # if we have just new entries, let's append them to the fncache |
|
791 | # if we have just new entries, let's append them to the fncache | |
784 | tr.addbackup(b'fncache') |
|
792 | tr.addbackup(b'fncache') | |
785 | fp = self.vfs(b'fncache', mode=b'ab', atomictemp=True) |
|
793 | fp = self.vfs(b'fncache', mode=b'ab', atomictemp=True) | |
786 | if self.addls: |
|
794 | if self.addls: | |
787 | fp.write(encodedir(b'\n'.join(self.addls) + b'\n')) |
|
795 | fp.write(encodedir(b'\n'.join(self.addls) + b'\n')) | |
788 | fp.close() |
|
796 | fp.close() | |
789 | self.entries = None |
|
797 | self.entries = None | |
790 | self.addls = set() |
|
798 | self.addls = set() | |
791 |
|
799 | |||
792 | def addignore(self, fn): |
|
800 | def addignore(self, fn): | |
793 | self._ignores.add(fn) |
|
801 | self._ignores.add(fn) | |
794 |
|
802 | |||
795 | def add(self, fn): |
|
803 | def add(self, fn): | |
796 | if fn in self._ignores: |
|
804 | if fn in self._ignores: | |
797 | return |
|
805 | return | |
798 | if self.entries is None: |
|
806 | if self.entries is None: | |
799 | self._load() |
|
807 | self._load() | |
800 | if fn not in self.entries: |
|
808 | if fn not in self.entries: | |
801 | self.addls.add(fn) |
|
809 | self.addls.add(fn) | |
802 |
|
810 | |||
803 | def remove(self, fn): |
|
811 | def remove(self, fn): | |
804 | if self.entries is None: |
|
812 | if self.entries is None: | |
805 | self._load() |
|
813 | self._load() | |
806 | if fn in self.addls: |
|
814 | if fn in self.addls: | |
807 | self.addls.remove(fn) |
|
815 | self.addls.remove(fn) | |
808 | return |
|
816 | return | |
809 | try: |
|
817 | try: | |
810 | self.entries.remove(fn) |
|
818 | self.entries.remove(fn) | |
811 | self._dirty = True |
|
819 | self._dirty = True | |
812 | except KeyError: |
|
820 | except KeyError: | |
813 | pass |
|
821 | pass | |
814 |
|
822 | |||
815 | def __contains__(self, fn): |
|
823 | def __contains__(self, fn): | |
816 | if fn in self.addls: |
|
824 | if fn in self.addls: | |
817 | return True |
|
825 | return True | |
818 | if self.entries is None: |
|
826 | if self.entries is None: | |
819 | self._load() |
|
827 | self._load() | |
820 | return fn in self.entries |
|
828 | return fn in self.entries | |
821 |
|
829 | |||
822 | def __iter__(self): |
|
830 | def __iter__(self): | |
823 | if self.entries is None: |
|
831 | if self.entries is None: | |
824 | self._load() |
|
832 | self._load() | |
825 | return iter(self.entries | self.addls) |
|
833 | return iter(self.entries | self.addls) | |
826 |
|
834 | |||
827 |
|
835 | |||
828 | class _fncachevfs(vfsmod.proxyvfs): |
|
836 | class _fncachevfs(vfsmod.proxyvfs): | |
829 | def __init__(self, vfs, fnc, encode): |
|
837 | def __init__(self, vfs, fnc, encode): | |
830 | vfsmod.proxyvfs.__init__(self, vfs) |
|
838 | vfsmod.proxyvfs.__init__(self, vfs) | |
831 | self.fncache = fnc |
|
839 | self.fncache = fnc | |
832 | self.encode = encode |
|
840 | self.encode = encode | |
833 |
|
841 | |||
834 | def __call__(self, path, mode=b'r', *args, **kw): |
|
842 | def __call__(self, path, mode=b'r', *args, **kw): | |
835 | encoded = self.encode(path) |
|
843 | encoded = self.encode(path) | |
836 | if ( |
|
844 | if ( | |
837 | mode not in (b'r', b'rb') |
|
845 | mode not in (b'r', b'rb') | |
838 | and (path.startswith(b'data/') or path.startswith(b'meta/')) |
|
846 | and (path.startswith(b'data/') or path.startswith(b'meta/')) | |
839 | and revlog_type(path) is not None |
|
847 | and revlog_type(path) is not None | |
840 | ): |
|
848 | ): | |
841 | # do not trigger a fncache load when adding a file that already is |
|
849 | # do not trigger a fncache load when adding a file that already is | |
842 | # known to exist. |
|
850 | # known to exist. | |
843 | notload = self.fncache.entries is None and self.vfs.exists(encoded) |
|
851 | notload = self.fncache.entries is None and self.vfs.exists(encoded) | |
844 | if notload and b'r+' in mode and not self.vfs.stat(encoded).st_size: |
|
852 | if notload and b'r+' in mode and not self.vfs.stat(encoded).st_size: | |
845 | # when appending to an existing file, if the file has size zero, |
|
853 | # when appending to an existing file, if the file has size zero, | |
846 | # it should be considered as missing. Such zero-size files are |
|
854 | # it should be considered as missing. Such zero-size files are | |
847 | # the result of truncation when a transaction is aborted. |
|
855 | # the result of truncation when a transaction is aborted. | |
848 | notload = False |
|
856 | notload = False | |
849 | if not notload: |
|
857 | if not notload: | |
850 | self.fncache.add(path) |
|
858 | self.fncache.add(path) | |
851 | return self.vfs(encoded, mode, *args, **kw) |
|
859 | return self.vfs(encoded, mode, *args, **kw) | |
852 |
|
860 | |||
853 | def join(self, path): |
|
861 | def join(self, path): | |
854 | if path: |
|
862 | if path: | |
855 | return self.vfs.join(self.encode(path)) |
|
863 | return self.vfs.join(self.encode(path)) | |
856 | else: |
|
864 | else: | |
857 | return self.vfs.join(path) |
|
865 | return self.vfs.join(path) | |
858 |
|
866 | |||
859 | def register_file(self, path): |
|
867 | def register_file(self, path): | |
860 | """generic hook point to lets fncache steer its stew""" |
|
868 | """generic hook point to lets fncache steer its stew""" | |
861 | if path.startswith(b'data/') or path.startswith(b'meta/'): |
|
869 | if path.startswith(b'data/') or path.startswith(b'meta/'): | |
862 | self.fncache.add(path) |
|
870 | self.fncache.add(path) | |
863 |
|
871 | |||
864 |
|
872 | |||
865 | class fncachestore(basicstore): |
|
873 | class fncachestore(basicstore): | |
866 | def __init__(self, path, vfstype, dotencode): |
|
874 | def __init__(self, path, vfstype, dotencode): | |
867 | if dotencode: |
|
875 | if dotencode: | |
868 | encode = _pathencode |
|
876 | encode = _pathencode | |
869 | else: |
|
877 | else: | |
870 | encode = _plainhybridencode |
|
878 | encode = _plainhybridencode | |
871 | self.encode = encode |
|
879 | self.encode = encode | |
872 | vfs = vfstype(path + b'/store') |
|
880 | vfs = vfstype(path + b'/store') | |
873 | self.path = vfs.base |
|
881 | self.path = vfs.base | |
874 | self.pathsep = self.path + b'/' |
|
882 | self.pathsep = self.path + b'/' | |
875 | self.createmode = _calcmode(vfs) |
|
883 | self.createmode = _calcmode(vfs) | |
876 | vfs.createmode = self.createmode |
|
884 | vfs.createmode = self.createmode | |
877 | self.rawvfs = vfs |
|
885 | self.rawvfs = vfs | |
878 | fnc = fncache(vfs) |
|
886 | fnc = fncache(vfs) | |
879 | self.fncache = fnc |
|
887 | self.fncache = fnc | |
880 | self.vfs = _fncachevfs(vfs, fnc, encode) |
|
888 | self.vfs = _fncachevfs(vfs, fnc, encode) | |
881 | self.opener = self.vfs |
|
889 | self.opener = self.vfs | |
882 |
|
890 | |||
883 | def join(self, f): |
|
891 | def join(self, f): | |
884 | return self.pathsep + self.encode(f) |
|
892 | return self.pathsep + self.encode(f) | |
885 |
|
893 | |||
886 | def getsize(self, path): |
|
894 | def getsize(self, path): | |
887 | return self.rawvfs.stat(path).st_size |
|
895 | return self.rawvfs.stat(path).st_size | |
888 |
|
896 | |||
889 | def datafiles( |
|
897 | def datafiles( | |
890 | self, matcher=None, undecodable=None |
|
898 | self, matcher=None, undecodable=None | |
891 | ) -> Generator[BaseStoreEntry, None, None]: |
|
899 | ) -> Generator[BaseStoreEntry, None, None]: | |
892 | for f in sorted(self.fncache): |
|
900 | for f in sorted(self.fncache): | |
893 | if not _matchtrackedpath(f, matcher): |
|
901 | if not _matchtrackedpath(f, matcher): | |
894 | continue |
|
902 | continue | |
895 | ef = self.encode(f) |
|
903 | ef = self.encode(f) | |
896 | t = revlog_type(f) |
|
904 | t = revlog_type(f) | |
897 | if t is None: |
|
905 | if t is None: | |
898 | # Note: this should not be in the fncache then⦠|
|
906 | # Note: this should not be in the fncache then⦠| |
899 | # |
|
907 | # | |
900 | # However the fncache might contains such file added by |
|
908 | # However the fncache might contains such file added by | |
901 | # previous version of Mercurial. |
|
909 | # previous version of Mercurial. | |
902 | continue |
|
910 | continue | |
903 | try: |
|
911 | yield RevlogStoreEntry( | |
904 | yield RevlogStoreEntry( |
|
912 | unencoded_path=f, | |
905 | unencoded_path=f, |
|
913 | revlog_type=FILEFLAGS_FILELOG, | |
906 |
|
|
914 | is_revlog_main=bool(t & FILEFLAGS_REVLOG_MAIN), | |
907 |
|
|
915 | is_volatile=bool(t & FILEFLAGS_VOLATILE), | |
908 | is_volatile=bool(t & FILEFLAGS_VOLATILE), |
|
916 | ) | |
909 | file_size=self.getsize(ef), |
|
|||
910 | ) |
|
|||
911 | except FileNotFoundError: |
|
|||
912 | pass |
|
|||
913 |
|
917 | |||
914 | def copylist(self): |
|
918 | def copylist(self): | |
915 | d = ( |
|
919 | d = ( | |
916 | b'bookmarks', |
|
920 | b'bookmarks', | |
917 | b'narrowspec', |
|
921 | b'narrowspec', | |
918 | b'data', |
|
922 | b'data', | |
919 | b'meta', |
|
923 | b'meta', | |
920 | b'dh', |
|
924 | b'dh', | |
921 | b'fncache', |
|
925 | b'fncache', | |
922 | b'phaseroots', |
|
926 | b'phaseroots', | |
923 | b'obsstore', |
|
927 | b'obsstore', | |
924 | b'00manifest.d', |
|
928 | b'00manifest.d', | |
925 | b'00manifest.i', |
|
929 | b'00manifest.i', | |
926 | b'00changelog.d', |
|
930 | b'00changelog.d', | |
927 | b'00changelog.i', |
|
931 | b'00changelog.i', | |
928 | b'requires', |
|
932 | b'requires', | |
929 | ) |
|
933 | ) | |
930 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in d] |
|
934 | return [b'requires', b'00changelog.i'] + [b'store/' + f for f in d] | |
931 |
|
935 | |||
932 | def write(self, tr): |
|
936 | def write(self, tr): | |
933 | self.fncache.write(tr) |
|
937 | self.fncache.write(tr) | |
934 |
|
938 | |||
935 | def invalidatecaches(self): |
|
939 | def invalidatecaches(self): | |
936 | self.fncache.entries = None |
|
940 | self.fncache.entries = None | |
937 | self.fncache.addls = set() |
|
941 | self.fncache.addls = set() | |
938 |
|
942 | |||
939 | def markremoved(self, fn): |
|
943 | def markremoved(self, fn): | |
940 | self.fncache.remove(fn) |
|
944 | self.fncache.remove(fn) | |
941 |
|
945 | |||
942 | def _exists(self, f): |
|
946 | def _exists(self, f): | |
943 | ef = self.encode(f) |
|
947 | ef = self.encode(f) | |
944 | try: |
|
948 | try: | |
945 | self.getsize(ef) |
|
949 | self.getsize(ef) | |
946 | return True |
|
950 | return True | |
947 | except FileNotFoundError: |
|
951 | except FileNotFoundError: | |
948 | return False |
|
952 | return False | |
949 |
|
953 | |||
950 | def __contains__(self, path): |
|
954 | def __contains__(self, path): | |
951 | '''Checks if the store contains path''' |
|
955 | '''Checks if the store contains path''' | |
952 | path = b"/".join((b"data", path)) |
|
956 | path = b"/".join((b"data", path)) | |
953 | # check for files (exact match) |
|
957 | # check for files (exact match) | |
954 | e = path + b'.i' |
|
958 | e = path + b'.i' | |
955 | if e in self.fncache and self._exists(e): |
|
959 | if e in self.fncache and self._exists(e): | |
956 | return True |
|
960 | return True | |
957 | # now check for directories (prefix match) |
|
961 | # now check for directories (prefix match) | |
958 | if not path.endswith(b'/'): |
|
962 | if not path.endswith(b'/'): | |
959 | path += b'/' |
|
963 | path += b'/' | |
960 | for e in self.fncache: |
|
964 | for e in self.fncache: | |
961 | if e.startswith(path) and self._exists(e): |
|
965 | if e.startswith(path) and self._exists(e): | |
962 | return True |
|
966 | return True | |
963 | return False |
|
967 | return False |
@@ -1,937 +1,939 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 entry in _walkstreamfiles(repo): |
|
272 | for entry in _walkstreamfiles(repo): | |
273 | for f in entry.files(): |
|
273 | for f in entry.files(): | |
274 |
|
|
274 | file_size = f.file_size(repo.store.vfs) | |
275 | entries.append((f.unencoded_path, f.file_size)) |
|
275 | if file_size: | |
276 |
|
|
276 | entries.append((f.unencoded_path, file_size)) | |
|
277 | total_bytes += file_size | |||
277 | _test_sync_point_walk_1(repo) |
|
278 | _test_sync_point_walk_1(repo) | |
278 | _test_sync_point_walk_2(repo) |
|
279 | _test_sync_point_walk_2(repo) | |
279 |
|
280 | |||
280 | repo.ui.debug( |
|
281 | repo.ui.debug( | |
281 | b'%d files, %d bytes to transfer\n' % (len(entries), total_bytes) |
|
282 | b'%d files, %d bytes to transfer\n' % (len(entries), total_bytes) | |
282 | ) |
|
283 | ) | |
283 |
|
284 | |||
284 | svfs = repo.svfs |
|
285 | svfs = repo.svfs | |
285 | debugflag = repo.ui.debugflag |
|
286 | debugflag = repo.ui.debugflag | |
286 |
|
287 | |||
287 | def emitrevlogdata(): |
|
288 | def emitrevlogdata(): | |
288 | for name, size in entries: |
|
289 | for name, size in entries: | |
289 | if debugflag: |
|
290 | if debugflag: | |
290 | repo.ui.debug(b'sending %s (%d bytes)\n' % (name, size)) |
|
291 | repo.ui.debug(b'sending %s (%d bytes)\n' % (name, size)) | |
291 | # partially encode name over the wire for backwards compat |
|
292 | # partially encode name over the wire for backwards compat | |
292 | yield b'%s\0%d\n' % (store.encodedir(name), size) |
|
293 | yield b'%s\0%d\n' % (store.encodedir(name), size) | |
293 | # auditing at this stage is both pointless (paths are already |
|
294 | # auditing at this stage is both pointless (paths are already | |
294 | # trusted by the local repo) and expensive |
|
295 | # trusted by the local repo) and expensive | |
295 | with svfs(name, b'rb', auditpath=False) as fp: |
|
296 | with svfs(name, b'rb', auditpath=False) as fp: | |
296 | if size <= 65536: |
|
297 | if size <= 65536: | |
297 | yield fp.read(size) |
|
298 | yield fp.read(size) | |
298 | else: |
|
299 | else: | |
299 | for chunk in util.filechunkiter(fp, limit=size): |
|
300 | for chunk in util.filechunkiter(fp, limit=size): | |
300 | yield chunk |
|
301 | yield chunk | |
301 |
|
302 | |||
302 | return len(entries), total_bytes, emitrevlogdata() |
|
303 | return len(entries), total_bytes, emitrevlogdata() | |
303 |
|
304 | |||
304 |
|
305 | |||
305 | def generatev1wireproto(repo): |
|
306 | def generatev1wireproto(repo): | |
306 | """Emit content for version 1 of streaming clone suitable for the wire. |
|
307 | """Emit content for version 1 of streaming clone suitable for the wire. | |
307 |
|
308 | |||
308 | This is the data output from ``generatev1()`` with 2 header lines. The |
|
309 | This is the data output from ``generatev1()`` with 2 header lines. The | |
309 | first line indicates overall success. The 2nd contains the file count and |
|
310 | first line indicates overall success. The 2nd contains the file count and | |
310 | byte size of payload. |
|
311 | byte size of payload. | |
311 |
|
312 | |||
312 | The success line contains "0" for success, "1" for stream generation not |
|
313 | The success line contains "0" for success, "1" for stream generation not | |
313 | allowed, and "2" for error locking the repository (possibly indicating |
|
314 | allowed, and "2" for error locking the repository (possibly indicating | |
314 | a permissions error for the server process). |
|
315 | a permissions error for the server process). | |
315 | """ |
|
316 | """ | |
316 | if not allowservergeneration(repo): |
|
317 | if not allowservergeneration(repo): | |
317 | yield b'1\n' |
|
318 | yield b'1\n' | |
318 | return |
|
319 | return | |
319 |
|
320 | |||
320 | try: |
|
321 | try: | |
321 | filecount, bytecount, it = generatev1(repo) |
|
322 | filecount, bytecount, it = generatev1(repo) | |
322 | except error.LockError: |
|
323 | except error.LockError: | |
323 | yield b'2\n' |
|
324 | yield b'2\n' | |
324 | return |
|
325 | return | |
325 |
|
326 | |||
326 | # Indicates successful response. |
|
327 | # Indicates successful response. | |
327 | yield b'0\n' |
|
328 | yield b'0\n' | |
328 | yield b'%d %d\n' % (filecount, bytecount) |
|
329 | yield b'%d %d\n' % (filecount, bytecount) | |
329 | for chunk in it: |
|
330 | for chunk in it: | |
330 | yield chunk |
|
331 | yield chunk | |
331 |
|
332 | |||
332 |
|
333 | |||
333 | def generatebundlev1(repo, compression=b'UN'): |
|
334 | def generatebundlev1(repo, compression=b'UN'): | |
334 | """Emit content for version 1 of a stream clone bundle. |
|
335 | """Emit content for version 1 of a stream clone bundle. | |
335 |
|
336 | |||
336 | The first 4 bytes of the output ("HGS1") denote this as stream clone |
|
337 | The first 4 bytes of the output ("HGS1") denote this as stream clone | |
337 | bundle version 1. |
|
338 | bundle version 1. | |
338 |
|
339 | |||
339 | The next 2 bytes indicate the compression type. Only "UN" is currently |
|
340 | The next 2 bytes indicate the compression type. Only "UN" is currently | |
340 | supported. |
|
341 | supported. | |
341 |
|
342 | |||
342 | The next 16 bytes are two 64-bit big endian unsigned integers indicating |
|
343 | The next 16 bytes are two 64-bit big endian unsigned integers indicating | |
343 | file count and byte count, respectively. |
|
344 | file count and byte count, respectively. | |
344 |
|
345 | |||
345 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length |
|
346 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length | |
346 | of the requirements string, including a trailing \0. The following N bytes |
|
347 | of the requirements string, including a trailing \0. The following N bytes | |
347 | are the requirements string, which is ASCII containing a comma-delimited |
|
348 | are the requirements string, which is ASCII containing a comma-delimited | |
348 | list of repo requirements that are needed to support the data. |
|
349 | list of repo requirements that are needed to support the data. | |
349 |
|
350 | |||
350 | The remaining content is the output of ``generatev1()`` (which may be |
|
351 | The remaining content is the output of ``generatev1()`` (which may be | |
351 | compressed in the future). |
|
352 | compressed in the future). | |
352 |
|
353 | |||
353 | Returns a tuple of (requirements, data generator). |
|
354 | Returns a tuple of (requirements, data generator). | |
354 | """ |
|
355 | """ | |
355 | if compression != b'UN': |
|
356 | if compression != b'UN': | |
356 | raise ValueError(b'we do not support the compression argument yet') |
|
357 | raise ValueError(b'we do not support the compression argument yet') | |
357 |
|
358 | |||
358 | requirements = streamed_requirements(repo) |
|
359 | requirements = streamed_requirements(repo) | |
359 | requires = b','.join(sorted(requirements)) |
|
360 | requires = b','.join(sorted(requirements)) | |
360 |
|
361 | |||
361 | def gen(): |
|
362 | def gen(): | |
362 | yield b'HGS1' |
|
363 | yield b'HGS1' | |
363 | yield compression |
|
364 | yield compression | |
364 |
|
365 | |||
365 | filecount, bytecount, it = generatev1(repo) |
|
366 | filecount, bytecount, it = generatev1(repo) | |
366 | repo.ui.status( |
|
367 | repo.ui.status( | |
367 | _(b'writing %d bytes for %d files\n') % (bytecount, filecount) |
|
368 | _(b'writing %d bytes for %d files\n') % (bytecount, filecount) | |
368 | ) |
|
369 | ) | |
369 |
|
370 | |||
370 | yield struct.pack(b'>QQ', filecount, bytecount) |
|
371 | yield struct.pack(b'>QQ', filecount, bytecount) | |
371 | yield struct.pack(b'>H', len(requires) + 1) |
|
372 | yield struct.pack(b'>H', len(requires) + 1) | |
372 | yield requires + b'\0' |
|
373 | yield requires + b'\0' | |
373 |
|
374 | |||
374 | # This is where we'll add compression in the future. |
|
375 | # This is where we'll add compression in the future. | |
375 | assert compression == b'UN' |
|
376 | assert compression == b'UN' | |
376 |
|
377 | |||
377 | progress = repo.ui.makeprogress( |
|
378 | progress = repo.ui.makeprogress( | |
378 | _(b'bundle'), total=bytecount, unit=_(b'bytes') |
|
379 | _(b'bundle'), total=bytecount, unit=_(b'bytes') | |
379 | ) |
|
380 | ) | |
380 | progress.update(0) |
|
381 | progress.update(0) | |
381 |
|
382 | |||
382 | for chunk in it: |
|
383 | for chunk in it: | |
383 | progress.increment(step=len(chunk)) |
|
384 | progress.increment(step=len(chunk)) | |
384 | yield chunk |
|
385 | yield chunk | |
385 |
|
386 | |||
386 | progress.complete() |
|
387 | progress.complete() | |
387 |
|
388 | |||
388 | return requirements, gen() |
|
389 | return requirements, gen() | |
389 |
|
390 | |||
390 |
|
391 | |||
391 | def consumev1(repo, fp, filecount, bytecount): |
|
392 | def consumev1(repo, fp, filecount, bytecount): | |
392 | """Apply the contents from version 1 of a streaming clone file handle. |
|
393 | """Apply the contents from version 1 of a streaming clone file handle. | |
393 |
|
394 | |||
394 | This takes the output from "stream_out" and applies it to the specified |
|
395 | This takes the output from "stream_out" and applies it to the specified | |
395 | repository. |
|
396 | repository. | |
396 |
|
397 | |||
397 | Like "stream_out," the status line added by the wire protocol is not |
|
398 | Like "stream_out," the status line added by the wire protocol is not | |
398 | handled by this function. |
|
399 | handled by this function. | |
399 | """ |
|
400 | """ | |
400 | with repo.lock(): |
|
401 | with repo.lock(): | |
401 | repo.ui.status( |
|
402 | repo.ui.status( | |
402 | _(b'%d files to transfer, %s of data\n') |
|
403 | _(b'%d files to transfer, %s of data\n') | |
403 | % (filecount, util.bytecount(bytecount)) |
|
404 | % (filecount, util.bytecount(bytecount)) | |
404 | ) |
|
405 | ) | |
405 | progress = repo.ui.makeprogress( |
|
406 | progress = repo.ui.makeprogress( | |
406 | _(b'clone'), total=bytecount, unit=_(b'bytes') |
|
407 | _(b'clone'), total=bytecount, unit=_(b'bytes') | |
407 | ) |
|
408 | ) | |
408 | progress.update(0) |
|
409 | progress.update(0) | |
409 | start = util.timer() |
|
410 | start = util.timer() | |
410 |
|
411 | |||
411 | # TODO: get rid of (potential) inconsistency |
|
412 | # TODO: get rid of (potential) inconsistency | |
412 | # |
|
413 | # | |
413 | # If transaction is started and any @filecache property is |
|
414 | # If transaction is started and any @filecache property is | |
414 | # changed at this point, it causes inconsistency between |
|
415 | # changed at this point, it causes inconsistency between | |
415 | # in-memory cached property and streamclone-ed file on the |
|
416 | # in-memory cached property and streamclone-ed file on the | |
416 | # disk. Nested transaction prevents transaction scope "clone" |
|
417 | # disk. Nested transaction prevents transaction scope "clone" | |
417 | # below from writing in-memory changes out at the end of it, |
|
418 | # below from writing in-memory changes out at the end of it, | |
418 | # even though in-memory changes are discarded at the end of it |
|
419 | # even though in-memory changes are discarded at the end of it | |
419 | # regardless of transaction nesting. |
|
420 | # regardless of transaction nesting. | |
420 | # |
|
421 | # | |
421 | # But transaction nesting can't be simply prohibited, because |
|
422 | # But transaction nesting can't be simply prohibited, because | |
422 | # nesting occurs also in ordinary case (e.g. enabling |
|
423 | # nesting occurs also in ordinary case (e.g. enabling | |
423 | # clonebundles). |
|
424 | # clonebundles). | |
424 |
|
425 | |||
425 | with repo.transaction(b'clone'): |
|
426 | with repo.transaction(b'clone'): | |
426 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): |
|
427 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): | |
427 | for i in range(filecount): |
|
428 | for i in range(filecount): | |
428 | # XXX doesn't support '\n' or '\r' in filenames |
|
429 | # XXX doesn't support '\n' or '\r' in filenames | |
429 | l = fp.readline() |
|
430 | l = fp.readline() | |
430 | try: |
|
431 | try: | |
431 | name, size = l.split(b'\0', 1) |
|
432 | name, size = l.split(b'\0', 1) | |
432 | size = int(size) |
|
433 | size = int(size) | |
433 | except (ValueError, TypeError): |
|
434 | except (ValueError, TypeError): | |
434 | raise error.ResponseError( |
|
435 | raise error.ResponseError( | |
435 | _(b'unexpected response from remote server:'), l |
|
436 | _(b'unexpected response from remote server:'), l | |
436 | ) |
|
437 | ) | |
437 | if repo.ui.debugflag: |
|
438 | if repo.ui.debugflag: | |
438 | repo.ui.debug( |
|
439 | repo.ui.debug( | |
439 | b'adding %s (%s)\n' % (name, util.bytecount(size)) |
|
440 | b'adding %s (%s)\n' % (name, util.bytecount(size)) | |
440 | ) |
|
441 | ) | |
441 | # for backwards compat, name was partially encoded |
|
442 | # for backwards compat, name was partially encoded | |
442 | path = store.decodedir(name) |
|
443 | path = store.decodedir(name) | |
443 | with repo.svfs(path, b'w', backgroundclose=True) as ofp: |
|
444 | with repo.svfs(path, b'w', backgroundclose=True) as ofp: | |
444 | for chunk in util.filechunkiter(fp, limit=size): |
|
445 | for chunk in util.filechunkiter(fp, limit=size): | |
445 | progress.increment(step=len(chunk)) |
|
446 | progress.increment(step=len(chunk)) | |
446 | ofp.write(chunk) |
|
447 | ofp.write(chunk) | |
447 |
|
448 | |||
448 | # force @filecache properties to be reloaded from |
|
449 | # force @filecache properties to be reloaded from | |
449 | # streamclone-ed file at next access |
|
450 | # streamclone-ed file at next access | |
450 | repo.invalidate(clearfilecache=True) |
|
451 | repo.invalidate(clearfilecache=True) | |
451 |
|
452 | |||
452 | elapsed = util.timer() - start |
|
453 | elapsed = util.timer() - start | |
453 | if elapsed <= 0: |
|
454 | if elapsed <= 0: | |
454 | elapsed = 0.001 |
|
455 | elapsed = 0.001 | |
455 | progress.complete() |
|
456 | progress.complete() | |
456 | repo.ui.status( |
|
457 | repo.ui.status( | |
457 | _(b'transferred %s in %.1f seconds (%s/sec)\n') |
|
458 | _(b'transferred %s in %.1f seconds (%s/sec)\n') | |
458 | % ( |
|
459 | % ( | |
459 | util.bytecount(bytecount), |
|
460 | util.bytecount(bytecount), | |
460 | elapsed, |
|
461 | elapsed, | |
461 | util.bytecount(bytecount / elapsed), |
|
462 | util.bytecount(bytecount / elapsed), | |
462 | ) |
|
463 | ) | |
463 | ) |
|
464 | ) | |
464 |
|
465 | |||
465 |
|
466 | |||
466 | def readbundle1header(fp): |
|
467 | def readbundle1header(fp): | |
467 | compression = fp.read(2) |
|
468 | compression = fp.read(2) | |
468 | if compression != b'UN': |
|
469 | if compression != b'UN': | |
469 | raise error.Abort( |
|
470 | raise error.Abort( | |
470 | _( |
|
471 | _( | |
471 | b'only uncompressed stream clone bundles are ' |
|
472 | b'only uncompressed stream clone bundles are ' | |
472 | b'supported; got %s' |
|
473 | b'supported; got %s' | |
473 | ) |
|
474 | ) | |
474 | % compression |
|
475 | % compression | |
475 | ) |
|
476 | ) | |
476 |
|
477 | |||
477 | filecount, bytecount = struct.unpack(b'>QQ', fp.read(16)) |
|
478 | filecount, bytecount = struct.unpack(b'>QQ', fp.read(16)) | |
478 | requireslen = struct.unpack(b'>H', fp.read(2))[0] |
|
479 | requireslen = struct.unpack(b'>H', fp.read(2))[0] | |
479 | requires = fp.read(requireslen) |
|
480 | requires = fp.read(requireslen) | |
480 |
|
481 | |||
481 | if not requires.endswith(b'\0'): |
|
482 | if not requires.endswith(b'\0'): | |
482 | raise error.Abort( |
|
483 | raise error.Abort( | |
483 | _( |
|
484 | _( | |
484 | b'malformed stream clone bundle: ' |
|
485 | b'malformed stream clone bundle: ' | |
485 | b'requirements not properly encoded' |
|
486 | b'requirements not properly encoded' | |
486 | ) |
|
487 | ) | |
487 | ) |
|
488 | ) | |
488 |
|
489 | |||
489 | requirements = set(requires.rstrip(b'\0').split(b',')) |
|
490 | requirements = set(requires.rstrip(b'\0').split(b',')) | |
490 |
|
491 | |||
491 | return filecount, bytecount, requirements |
|
492 | return filecount, bytecount, requirements | |
492 |
|
493 | |||
493 |
|
494 | |||
494 | def applybundlev1(repo, fp): |
|
495 | def applybundlev1(repo, fp): | |
495 | """Apply the content from a stream clone bundle version 1. |
|
496 | """Apply the content from a stream clone bundle version 1. | |
496 |
|
497 | |||
497 | We assume the 4 byte header has been read and validated and the file handle |
|
498 | We assume the 4 byte header has been read and validated and the file handle | |
498 | is at the 2 byte compression identifier. |
|
499 | is at the 2 byte compression identifier. | |
499 | """ |
|
500 | """ | |
500 | if len(repo): |
|
501 | if len(repo): | |
501 | raise error.Abort( |
|
502 | raise error.Abort( | |
502 | _(b'cannot apply stream clone bundle on non-empty repo') |
|
503 | _(b'cannot apply stream clone bundle on non-empty repo') | |
503 | ) |
|
504 | ) | |
504 |
|
505 | |||
505 | filecount, bytecount, requirements = readbundle1header(fp) |
|
506 | filecount, bytecount, requirements = readbundle1header(fp) | |
506 | missingreqs = requirements - repo.supported |
|
507 | missingreqs = requirements - repo.supported | |
507 | if missingreqs: |
|
508 | if missingreqs: | |
508 | raise error.Abort( |
|
509 | raise error.Abort( | |
509 | _(b'unable to apply stream clone: unsupported format: %s') |
|
510 | _(b'unable to apply stream clone: unsupported format: %s') | |
510 | % b', '.join(sorted(missingreqs)) |
|
511 | % b', '.join(sorted(missingreqs)) | |
511 | ) |
|
512 | ) | |
512 |
|
513 | |||
513 | consumev1(repo, fp, filecount, bytecount) |
|
514 | consumev1(repo, fp, filecount, bytecount) | |
514 | nodemap.post_stream_cleanup(repo) |
|
515 | nodemap.post_stream_cleanup(repo) | |
515 |
|
516 | |||
516 |
|
517 | |||
517 | class streamcloneapplier: |
|
518 | class streamcloneapplier: | |
518 | """Class to manage applying streaming clone bundles. |
|
519 | """Class to manage applying streaming clone bundles. | |
519 |
|
520 | |||
520 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle |
|
521 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle | |
521 | readers to perform bundle type-specific functionality. |
|
522 | readers to perform bundle type-specific functionality. | |
522 | """ |
|
523 | """ | |
523 |
|
524 | |||
524 | def __init__(self, fh): |
|
525 | def __init__(self, fh): | |
525 | self._fh = fh |
|
526 | self._fh = fh | |
526 |
|
527 | |||
527 | def apply(self, repo): |
|
528 | def apply(self, repo): | |
528 | return applybundlev1(repo, self._fh) |
|
529 | return applybundlev1(repo, self._fh) | |
529 |
|
530 | |||
530 |
|
531 | |||
531 | # type of file to stream |
|
532 | # type of file to stream | |
532 | _fileappend = 0 # append only file |
|
533 | _fileappend = 0 # append only file | |
533 | _filefull = 1 # full snapshot file |
|
534 | _filefull = 1 # full snapshot file | |
534 |
|
535 | |||
535 | # Source of the file |
|
536 | # Source of the file | |
536 | _srcstore = b's' # store (svfs) |
|
537 | _srcstore = b's' # store (svfs) | |
537 | _srccache = b'c' # cache (cache) |
|
538 | _srccache = b'c' # cache (cache) | |
538 |
|
539 | |||
539 | # This is it's own function so extensions can override it. |
|
540 | # This is it's own function so extensions can override it. | |
540 | def _walkstreamfullstorefiles(repo): |
|
541 | def _walkstreamfullstorefiles(repo): | |
541 | """list snapshot file from the store""" |
|
542 | """list snapshot file from the store""" | |
542 | fnames = [] |
|
543 | fnames = [] | |
543 | if not repo.publishing(): |
|
544 | if not repo.publishing(): | |
544 | fnames.append(b'phaseroots') |
|
545 | fnames.append(b'phaseroots') | |
545 | return fnames |
|
546 | return fnames | |
546 |
|
547 | |||
547 |
|
548 | |||
548 | def _filterfull(entry, copy, vfsmap): |
|
549 | def _filterfull(entry, copy, vfsmap): | |
549 | """actually copy the snapshot files""" |
|
550 | """actually copy the snapshot files""" | |
550 | src, name, ftype, data = entry |
|
551 | src, name, ftype, data = entry | |
551 | if ftype != _filefull: |
|
552 | if ftype != _filefull: | |
552 | return entry |
|
553 | return entry | |
553 | return (src, name, ftype, copy(vfsmap[src].join(name))) |
|
554 | return (src, name, ftype, copy(vfsmap[src].join(name))) | |
554 |
|
555 | |||
555 |
|
556 | |||
556 | @contextlib.contextmanager |
|
557 | @contextlib.contextmanager | |
557 | def maketempcopies(): |
|
558 | def maketempcopies(): | |
558 | """return a function to temporary copy file""" |
|
559 | """return a function to temporary copy file""" | |
559 |
|
560 | |||
560 | files = [] |
|
561 | files = [] | |
561 | dst_dir = pycompat.mkdtemp(prefix=b'hg-clone-') |
|
562 | dst_dir = pycompat.mkdtemp(prefix=b'hg-clone-') | |
562 | try: |
|
563 | try: | |
563 |
|
564 | |||
564 | def copy(src): |
|
565 | def copy(src): | |
565 | fd, dst = pycompat.mkstemp( |
|
566 | fd, dst = pycompat.mkstemp( | |
566 | prefix=os.path.basename(src), dir=dst_dir |
|
567 | prefix=os.path.basename(src), dir=dst_dir | |
567 | ) |
|
568 | ) | |
568 | os.close(fd) |
|
569 | os.close(fd) | |
569 | files.append(dst) |
|
570 | files.append(dst) | |
570 | util.copyfiles(src, dst, hardlink=True) |
|
571 | util.copyfiles(src, dst, hardlink=True) | |
571 | return dst |
|
572 | return dst | |
572 |
|
573 | |||
573 | yield copy |
|
574 | yield copy | |
574 | finally: |
|
575 | finally: | |
575 | for tmp in files: |
|
576 | for tmp in files: | |
576 | util.tryunlink(tmp) |
|
577 | util.tryunlink(tmp) | |
577 | util.tryrmdir(dst_dir) |
|
578 | util.tryrmdir(dst_dir) | |
578 |
|
579 | |||
579 |
|
580 | |||
580 | def _makemap(repo): |
|
581 | def _makemap(repo): | |
581 | """make a (src -> vfs) map for the repo""" |
|
582 | """make a (src -> vfs) map for the repo""" | |
582 | vfsmap = { |
|
583 | vfsmap = { | |
583 | _srcstore: repo.svfs, |
|
584 | _srcstore: repo.svfs, | |
584 | _srccache: repo.cachevfs, |
|
585 | _srccache: repo.cachevfs, | |
585 | } |
|
586 | } | |
586 | # we keep repo.vfs out of the on purpose, ther are too many danger there |
|
587 | # we keep repo.vfs out of the on purpose, ther are too many danger there | |
587 | # (eg: .hg/hgrc) |
|
588 | # (eg: .hg/hgrc) | |
588 | assert repo.vfs not in vfsmap.values() |
|
589 | assert repo.vfs not in vfsmap.values() | |
589 |
|
590 | |||
590 | return vfsmap |
|
591 | return vfsmap | |
591 |
|
592 | |||
592 |
|
593 | |||
593 | def _emit2(repo, entries, totalfilesize): |
|
594 | def _emit2(repo, entries, totalfilesize): | |
594 | """actually emit the stream bundle""" |
|
595 | """actually emit the stream bundle""" | |
595 | vfsmap = _makemap(repo) |
|
596 | vfsmap = _makemap(repo) | |
596 | # we keep repo.vfs out of the on purpose, ther are too many danger there |
|
597 | # we keep repo.vfs out of the on purpose, ther are too many danger there | |
597 | # (eg: .hg/hgrc), |
|
598 | # (eg: .hg/hgrc), | |
598 | # |
|
599 | # | |
599 | # this assert is duplicated (from _makemap) as author might think this is |
|
600 | # this assert is duplicated (from _makemap) as author might think this is | |
600 | # fine, while this is really not fine. |
|
601 | # fine, while this is really not fine. | |
601 | if repo.vfs in vfsmap.values(): |
|
602 | if repo.vfs in vfsmap.values(): | |
602 | raise error.ProgrammingError( |
|
603 | raise error.ProgrammingError( | |
603 | b'repo.vfs must not be added to vfsmap for security reasons' |
|
604 | b'repo.vfs must not be added to vfsmap for security reasons' | |
604 | ) |
|
605 | ) | |
605 |
|
606 | |||
606 | progress = repo.ui.makeprogress( |
|
607 | progress = repo.ui.makeprogress( | |
607 | _(b'bundle'), total=totalfilesize, unit=_(b'bytes') |
|
608 | _(b'bundle'), total=totalfilesize, unit=_(b'bytes') | |
608 | ) |
|
609 | ) | |
609 | progress.update(0) |
|
610 | progress.update(0) | |
610 | with maketempcopies() as copy, progress: |
|
611 | with maketempcopies() as copy, progress: | |
611 | # copy is delayed until we are in the try |
|
612 | # copy is delayed until we are in the try | |
612 | entries = [_filterfull(e, copy, vfsmap) for e in entries] |
|
613 | entries = [_filterfull(e, copy, vfsmap) for e in entries] | |
613 | yield None # this release the lock on the repository |
|
614 | yield None # this release the lock on the repository | |
614 | totalbytecount = 0 |
|
615 | totalbytecount = 0 | |
615 |
|
616 | |||
616 | for src, name, ftype, data in entries: |
|
617 | for src, name, ftype, data in entries: | |
617 | vfs = vfsmap[src] |
|
618 | vfs = vfsmap[src] | |
618 | yield src |
|
619 | yield src | |
619 | yield util.uvarintencode(len(name)) |
|
620 | yield util.uvarintencode(len(name)) | |
620 | if ftype == _fileappend: |
|
621 | if ftype == _fileappend: | |
621 | fp = vfs(name) |
|
622 | fp = vfs(name) | |
622 | size = data |
|
623 | size = data | |
623 | elif ftype == _filefull: |
|
624 | elif ftype == _filefull: | |
624 | fp = open(data, b'rb') |
|
625 | fp = open(data, b'rb') | |
625 | size = util.fstat(fp).st_size |
|
626 | size = util.fstat(fp).st_size | |
626 | bytecount = 0 |
|
627 | bytecount = 0 | |
627 | try: |
|
628 | try: | |
628 | yield util.uvarintencode(size) |
|
629 | yield util.uvarintencode(size) | |
629 | yield name |
|
630 | yield name | |
630 | if size <= 65536: |
|
631 | if size <= 65536: | |
631 | chunks = (fp.read(size),) |
|
632 | chunks = (fp.read(size),) | |
632 | else: |
|
633 | else: | |
633 | chunks = util.filechunkiter(fp, limit=size) |
|
634 | chunks = util.filechunkiter(fp, limit=size) | |
634 | for chunk in chunks: |
|
635 | for chunk in chunks: | |
635 | bytecount += len(chunk) |
|
636 | bytecount += len(chunk) | |
636 | totalbytecount += len(chunk) |
|
637 | totalbytecount += len(chunk) | |
637 | progress.update(totalbytecount) |
|
638 | progress.update(totalbytecount) | |
638 | yield chunk |
|
639 | yield chunk | |
639 | if bytecount != size: |
|
640 | if bytecount != size: | |
640 | # Would most likely be caused by a race due to `hg strip` or |
|
641 | # Would most likely be caused by a race due to `hg strip` or | |
641 | # a revlog split |
|
642 | # a revlog split | |
642 | raise error.Abort( |
|
643 | raise error.Abort( | |
643 | _( |
|
644 | _( | |
644 | b'clone could only read %d bytes from %s, but ' |
|
645 | b'clone could only read %d bytes from %s, but ' | |
645 | b'expected %d bytes' |
|
646 | b'expected %d bytes' | |
646 | ) |
|
647 | ) | |
647 | % (bytecount, name, size) |
|
648 | % (bytecount, name, size) | |
648 | ) |
|
649 | ) | |
649 | finally: |
|
650 | finally: | |
650 | fp.close() |
|
651 | fp.close() | |
651 |
|
652 | |||
652 |
|
653 | |||
653 | def _test_sync_point_walk_1(repo): |
|
654 | def _test_sync_point_walk_1(repo): | |
654 | """a function for synchronisation during tests""" |
|
655 | """a function for synchronisation during tests""" | |
655 |
|
656 | |||
656 |
|
657 | |||
657 | def _test_sync_point_walk_2(repo): |
|
658 | def _test_sync_point_walk_2(repo): | |
658 | """a function for synchronisation during tests""" |
|
659 | """a function for synchronisation during tests""" | |
659 |
|
660 | |||
660 |
|
661 | |||
661 | def _v2_walk(repo, includes, excludes, includeobsmarkers): |
|
662 | def _v2_walk(repo, includes, excludes, includeobsmarkers): | |
662 | """emit a seris of files information useful to clone a repo |
|
663 | """emit a seris of files information useful to clone a repo | |
663 |
|
664 | |||
664 | return (entries, totalfilesize) |
|
665 | return (entries, totalfilesize) | |
665 |
|
666 | |||
666 | entries is a list of tuple (vfs-key, file-path, file-type, size) |
|
667 | entries is a list of tuple (vfs-key, file-path, file-type, size) | |
667 |
|
668 | |||
668 | - `vfs-key`: is a key to the right vfs to write the file (see _makemap) |
|
669 | - `vfs-key`: is a key to the right vfs to write the file (see _makemap) | |
669 | - `name`: file path of the file to copy (to be feed to the vfss) |
|
670 | - `name`: file path of the file to copy (to be feed to the vfss) | |
670 | - `file-type`: do this file need to be copied with the source lock ? |
|
671 | - `file-type`: do this file need to be copied with the source lock ? | |
671 | - `size`: the size of the file (or None) |
|
672 | - `size`: the size of the file (or None) | |
672 | """ |
|
673 | """ | |
673 | assert repo._currentlock(repo._lockref) is not None |
|
674 | assert repo._currentlock(repo._lockref) is not None | |
674 | entries = [] |
|
675 | entries = [] | |
675 | totalfilesize = 0 |
|
676 | totalfilesize = 0 | |
676 |
|
677 | |||
677 | matcher = None |
|
678 | matcher = None | |
678 | if includes or excludes: |
|
679 | if includes or excludes: | |
679 | matcher = narrowspec.match(repo.root, includes, excludes) |
|
680 | matcher = narrowspec.match(repo.root, includes, excludes) | |
680 |
|
681 | |||
681 | for entry in _walkstreamfiles(repo, matcher): |
|
682 | for entry in _walkstreamfiles(repo, matcher): | |
682 | for f in entry.files(): |
|
683 | for f in entry.files(): | |
683 |
|
|
684 | file_size = f.file_size(repo.store.vfs) | |
|
685 | if file_size: | |||
684 | ft = _fileappend |
|
686 | ft = _fileappend | |
685 | if f.is_volatile: |
|
687 | if f.is_volatile: | |
686 | ft = _filefull |
|
688 | ft = _filefull | |
687 |
entries.append((_srcstore, f.unencoded_path, ft, |
|
689 | entries.append((_srcstore, f.unencoded_path, ft, file_size)) | |
688 |
totalfilesize += |
|
690 | totalfilesize += file_size | |
689 | for name in _walkstreamfullstorefiles(repo): |
|
691 | for name in _walkstreamfullstorefiles(repo): | |
690 | if repo.svfs.exists(name): |
|
692 | if repo.svfs.exists(name): | |
691 | totalfilesize += repo.svfs.lstat(name).st_size |
|
693 | totalfilesize += repo.svfs.lstat(name).st_size | |
692 | entries.append((_srcstore, name, _filefull, None)) |
|
694 | entries.append((_srcstore, name, _filefull, None)) | |
693 | if includeobsmarkers and repo.svfs.exists(b'obsstore'): |
|
695 | if includeobsmarkers and repo.svfs.exists(b'obsstore'): | |
694 | totalfilesize += repo.svfs.lstat(b'obsstore').st_size |
|
696 | totalfilesize += repo.svfs.lstat(b'obsstore').st_size | |
695 | entries.append((_srcstore, b'obsstore', _filefull, None)) |
|
697 | entries.append((_srcstore, b'obsstore', _filefull, None)) | |
696 | for name in cacheutil.cachetocopy(repo): |
|
698 | for name in cacheutil.cachetocopy(repo): | |
697 | if repo.cachevfs.exists(name): |
|
699 | if repo.cachevfs.exists(name): | |
698 | totalfilesize += repo.cachevfs.lstat(name).st_size |
|
700 | totalfilesize += repo.cachevfs.lstat(name).st_size | |
699 | entries.append((_srccache, name, _filefull, None)) |
|
701 | entries.append((_srccache, name, _filefull, None)) | |
700 | return entries, totalfilesize |
|
702 | return entries, totalfilesize | |
701 |
|
703 | |||
702 |
|
704 | |||
703 | def generatev2(repo, includes, excludes, includeobsmarkers): |
|
705 | def generatev2(repo, includes, excludes, includeobsmarkers): | |
704 | """Emit content for version 2 of a streaming clone. |
|
706 | """Emit content for version 2 of a streaming clone. | |
705 |
|
707 | |||
706 | the data stream consists the following entries: |
|
708 | the data stream consists the following entries: | |
707 | 1) A char representing the file destination (eg: store or cache) |
|
709 | 1) A char representing the file destination (eg: store or cache) | |
708 | 2) A varint containing the length of the filename |
|
710 | 2) A varint containing the length of the filename | |
709 | 3) A varint containing the length of file data |
|
711 | 3) A varint containing the length of file data | |
710 | 4) N bytes containing the filename (the internal, store-agnostic form) |
|
712 | 4) N bytes containing the filename (the internal, store-agnostic form) | |
711 | 5) N bytes containing the file data |
|
713 | 5) N bytes containing the file data | |
712 |
|
714 | |||
713 | Returns a 3-tuple of (file count, file size, data iterator). |
|
715 | Returns a 3-tuple of (file count, file size, data iterator). | |
714 | """ |
|
716 | """ | |
715 |
|
717 | |||
716 | with repo.lock(): |
|
718 | with repo.lock(): | |
717 |
|
719 | |||
718 | repo.ui.debug(b'scanning\n') |
|
720 | repo.ui.debug(b'scanning\n') | |
719 |
|
721 | |||
720 | entries, totalfilesize = _v2_walk( |
|
722 | entries, totalfilesize = _v2_walk( | |
721 | repo, |
|
723 | repo, | |
722 | includes=includes, |
|
724 | includes=includes, | |
723 | excludes=excludes, |
|
725 | excludes=excludes, | |
724 | includeobsmarkers=includeobsmarkers, |
|
726 | includeobsmarkers=includeobsmarkers, | |
725 | ) |
|
727 | ) | |
726 |
|
728 | |||
727 | chunks = _emit2(repo, entries, totalfilesize) |
|
729 | chunks = _emit2(repo, entries, totalfilesize) | |
728 | first = next(chunks) |
|
730 | first = next(chunks) | |
729 | assert first is None |
|
731 | assert first is None | |
730 | _test_sync_point_walk_1(repo) |
|
732 | _test_sync_point_walk_1(repo) | |
731 | _test_sync_point_walk_2(repo) |
|
733 | _test_sync_point_walk_2(repo) | |
732 |
|
734 | |||
733 | return len(entries), totalfilesize, chunks |
|
735 | return len(entries), totalfilesize, chunks | |
734 |
|
736 | |||
735 |
|
737 | |||
736 | @contextlib.contextmanager |
|
738 | @contextlib.contextmanager | |
737 | def nested(*ctxs): |
|
739 | def nested(*ctxs): | |
738 | this = ctxs[0] |
|
740 | this = ctxs[0] | |
739 | rest = ctxs[1:] |
|
741 | rest = ctxs[1:] | |
740 | with this: |
|
742 | with this: | |
741 | if rest: |
|
743 | if rest: | |
742 | with nested(*rest): |
|
744 | with nested(*rest): | |
743 | yield |
|
745 | yield | |
744 | else: |
|
746 | else: | |
745 | yield |
|
747 | yield | |
746 |
|
748 | |||
747 |
|
749 | |||
748 | def consumev2(repo, fp, filecount, filesize): |
|
750 | def consumev2(repo, fp, filecount, filesize): | |
749 | """Apply the contents from a version 2 streaming clone. |
|
751 | """Apply the contents from a version 2 streaming clone. | |
750 |
|
752 | |||
751 | Data is read from an object that only needs to provide a ``read(size)`` |
|
753 | Data is read from an object that only needs to provide a ``read(size)`` | |
752 | method. |
|
754 | method. | |
753 | """ |
|
755 | """ | |
754 | with repo.lock(): |
|
756 | with repo.lock(): | |
755 | repo.ui.status( |
|
757 | repo.ui.status( | |
756 | _(b'%d files to transfer, %s of data\n') |
|
758 | _(b'%d files to transfer, %s of data\n') | |
757 | % (filecount, util.bytecount(filesize)) |
|
759 | % (filecount, util.bytecount(filesize)) | |
758 | ) |
|
760 | ) | |
759 |
|
761 | |||
760 | start = util.timer() |
|
762 | start = util.timer() | |
761 | progress = repo.ui.makeprogress( |
|
763 | progress = repo.ui.makeprogress( | |
762 | _(b'clone'), total=filesize, unit=_(b'bytes') |
|
764 | _(b'clone'), total=filesize, unit=_(b'bytes') | |
763 | ) |
|
765 | ) | |
764 | progress.update(0) |
|
766 | progress.update(0) | |
765 |
|
767 | |||
766 | vfsmap = _makemap(repo) |
|
768 | vfsmap = _makemap(repo) | |
767 | # we keep repo.vfs out of the on purpose, ther are too many danger |
|
769 | # we keep repo.vfs out of the on purpose, ther are too many danger | |
768 | # there (eg: .hg/hgrc), |
|
770 | # there (eg: .hg/hgrc), | |
769 | # |
|
771 | # | |
770 | # this assert is duplicated (from _makemap) as author might think this |
|
772 | # this assert is duplicated (from _makemap) as author might think this | |
771 | # is fine, while this is really not fine. |
|
773 | # is fine, while this is really not fine. | |
772 | if repo.vfs in vfsmap.values(): |
|
774 | if repo.vfs in vfsmap.values(): | |
773 | raise error.ProgrammingError( |
|
775 | raise error.ProgrammingError( | |
774 | b'repo.vfs must not be added to vfsmap for security reasons' |
|
776 | b'repo.vfs must not be added to vfsmap for security reasons' | |
775 | ) |
|
777 | ) | |
776 |
|
778 | |||
777 | with repo.transaction(b'clone'): |
|
779 | with repo.transaction(b'clone'): | |
778 | ctxs = (vfs.backgroundclosing(repo.ui) for vfs in vfsmap.values()) |
|
780 | ctxs = (vfs.backgroundclosing(repo.ui) for vfs in vfsmap.values()) | |
779 | with nested(*ctxs): |
|
781 | with nested(*ctxs): | |
780 | for i in range(filecount): |
|
782 | for i in range(filecount): | |
781 | src = util.readexactly(fp, 1) |
|
783 | src = util.readexactly(fp, 1) | |
782 | vfs = vfsmap[src] |
|
784 | vfs = vfsmap[src] | |
783 | namelen = util.uvarintdecodestream(fp) |
|
785 | namelen = util.uvarintdecodestream(fp) | |
784 | datalen = util.uvarintdecodestream(fp) |
|
786 | datalen = util.uvarintdecodestream(fp) | |
785 |
|
787 | |||
786 | name = util.readexactly(fp, namelen) |
|
788 | name = util.readexactly(fp, namelen) | |
787 |
|
789 | |||
788 | if repo.ui.debugflag: |
|
790 | if repo.ui.debugflag: | |
789 | repo.ui.debug( |
|
791 | repo.ui.debug( | |
790 | b'adding [%s] %s (%s)\n' |
|
792 | b'adding [%s] %s (%s)\n' | |
791 | % (src, name, util.bytecount(datalen)) |
|
793 | % (src, name, util.bytecount(datalen)) | |
792 | ) |
|
794 | ) | |
793 |
|
795 | |||
794 | with vfs(name, b'w') as ofp: |
|
796 | with vfs(name, b'w') as ofp: | |
795 | for chunk in util.filechunkiter(fp, limit=datalen): |
|
797 | for chunk in util.filechunkiter(fp, limit=datalen): | |
796 | progress.increment(step=len(chunk)) |
|
798 | progress.increment(step=len(chunk)) | |
797 | ofp.write(chunk) |
|
799 | ofp.write(chunk) | |
798 |
|
800 | |||
799 | # force @filecache properties to be reloaded from |
|
801 | # force @filecache properties to be reloaded from | |
800 | # streamclone-ed file at next access |
|
802 | # streamclone-ed file at next access | |
801 | repo.invalidate(clearfilecache=True) |
|
803 | repo.invalidate(clearfilecache=True) | |
802 |
|
804 | |||
803 | elapsed = util.timer() - start |
|
805 | elapsed = util.timer() - start | |
804 | if elapsed <= 0: |
|
806 | if elapsed <= 0: | |
805 | elapsed = 0.001 |
|
807 | elapsed = 0.001 | |
806 | repo.ui.status( |
|
808 | repo.ui.status( | |
807 | _(b'transferred %s in %.1f seconds (%s/sec)\n') |
|
809 | _(b'transferred %s in %.1f seconds (%s/sec)\n') | |
808 | % ( |
|
810 | % ( | |
809 | util.bytecount(progress.pos), |
|
811 | util.bytecount(progress.pos), | |
810 | elapsed, |
|
812 | elapsed, | |
811 | util.bytecount(progress.pos / elapsed), |
|
813 | util.bytecount(progress.pos / elapsed), | |
812 | ) |
|
814 | ) | |
813 | ) |
|
815 | ) | |
814 | progress.complete() |
|
816 | progress.complete() | |
815 |
|
817 | |||
816 |
|
818 | |||
817 | def applybundlev2(repo, fp, filecount, filesize, requirements): |
|
819 | def applybundlev2(repo, fp, filecount, filesize, requirements): | |
818 | from . import localrepo |
|
820 | from . import localrepo | |
819 |
|
821 | |||
820 | missingreqs = [r for r in requirements if r not in repo.supported] |
|
822 | missingreqs = [r for r in requirements if r not in repo.supported] | |
821 | if missingreqs: |
|
823 | if missingreqs: | |
822 | raise error.Abort( |
|
824 | raise error.Abort( | |
823 | _(b'unable to apply stream clone: unsupported format: %s') |
|
825 | _(b'unable to apply stream clone: unsupported format: %s') | |
824 | % b', '.join(sorted(missingreqs)) |
|
826 | % b', '.join(sorted(missingreqs)) | |
825 | ) |
|
827 | ) | |
826 |
|
828 | |||
827 | consumev2(repo, fp, filecount, filesize) |
|
829 | consumev2(repo, fp, filecount, filesize) | |
828 |
|
830 | |||
829 | repo.requirements = new_stream_clone_requirements( |
|
831 | repo.requirements = new_stream_clone_requirements( | |
830 | repo.requirements, |
|
832 | repo.requirements, | |
831 | requirements, |
|
833 | requirements, | |
832 | ) |
|
834 | ) | |
833 | repo.svfs.options = localrepo.resolvestorevfsoptions( |
|
835 | repo.svfs.options = localrepo.resolvestorevfsoptions( | |
834 | repo.ui, repo.requirements, repo.features |
|
836 | repo.ui, repo.requirements, repo.features | |
835 | ) |
|
837 | ) | |
836 | scmutil.writereporequirements(repo) |
|
838 | scmutil.writereporequirements(repo) | |
837 | nodemap.post_stream_cleanup(repo) |
|
839 | nodemap.post_stream_cleanup(repo) | |
838 |
|
840 | |||
839 |
|
841 | |||
840 | def _copy_files(src_vfs_map, dst_vfs_map, entries, progress): |
|
842 | def _copy_files(src_vfs_map, dst_vfs_map, entries, progress): | |
841 | hardlink = [True] |
|
843 | hardlink = [True] | |
842 |
|
844 | |||
843 | def copy_used(): |
|
845 | def copy_used(): | |
844 | hardlink[0] = False |
|
846 | hardlink[0] = False | |
845 | progress.topic = _(b'copying') |
|
847 | progress.topic = _(b'copying') | |
846 |
|
848 | |||
847 | for k, path, size in entries: |
|
849 | for k, path, size in entries: | |
848 | src_vfs = src_vfs_map[k] |
|
850 | src_vfs = src_vfs_map[k] | |
849 | dst_vfs = dst_vfs_map[k] |
|
851 | dst_vfs = dst_vfs_map[k] | |
850 | src_path = src_vfs.join(path) |
|
852 | src_path = src_vfs.join(path) | |
851 | dst_path = dst_vfs.join(path) |
|
853 | dst_path = dst_vfs.join(path) | |
852 | # We cannot use dirname and makedirs of dst_vfs here because the store |
|
854 | # We cannot use dirname and makedirs of dst_vfs here because the store | |
853 | # encoding confuses them. See issue 6581 for details. |
|
855 | # encoding confuses them. See issue 6581 for details. | |
854 | dirname = os.path.dirname(dst_path) |
|
856 | dirname = os.path.dirname(dst_path) | |
855 | if not os.path.exists(dirname): |
|
857 | if not os.path.exists(dirname): | |
856 | util.makedirs(dirname) |
|
858 | util.makedirs(dirname) | |
857 | dst_vfs.register_file(path) |
|
859 | dst_vfs.register_file(path) | |
858 | # XXX we could use the #nb_bytes argument. |
|
860 | # XXX we could use the #nb_bytes argument. | |
859 | util.copyfile( |
|
861 | util.copyfile( | |
860 | src_path, |
|
862 | src_path, | |
861 | dst_path, |
|
863 | dst_path, | |
862 | hardlink=hardlink[0], |
|
864 | hardlink=hardlink[0], | |
863 | no_hardlink_cb=copy_used, |
|
865 | no_hardlink_cb=copy_used, | |
864 | check_fs_hardlink=False, |
|
866 | check_fs_hardlink=False, | |
865 | ) |
|
867 | ) | |
866 | progress.increment() |
|
868 | progress.increment() | |
867 | return hardlink[0] |
|
869 | return hardlink[0] | |
868 |
|
870 | |||
869 |
|
871 | |||
870 | def local_copy(src_repo, dest_repo): |
|
872 | def local_copy(src_repo, dest_repo): | |
871 | """copy all content from one local repository to another |
|
873 | """copy all content from one local repository to another | |
872 |
|
874 | |||
873 | This is useful for local clone""" |
|
875 | This is useful for local clone""" | |
874 | src_store_requirements = { |
|
876 | src_store_requirements = { | |
875 | r |
|
877 | r | |
876 | for r in src_repo.requirements |
|
878 | for r in src_repo.requirements | |
877 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS |
|
879 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS | |
878 | } |
|
880 | } | |
879 | dest_store_requirements = { |
|
881 | dest_store_requirements = { | |
880 | r |
|
882 | r | |
881 | for r in dest_repo.requirements |
|
883 | for r in dest_repo.requirements | |
882 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS |
|
884 | if r not in requirementsmod.WORKING_DIR_REQUIREMENTS | |
883 | } |
|
885 | } | |
884 | assert src_store_requirements == dest_store_requirements |
|
886 | assert src_store_requirements == dest_store_requirements | |
885 |
|
887 | |||
886 | with dest_repo.lock(): |
|
888 | with dest_repo.lock(): | |
887 | with src_repo.lock(): |
|
889 | with src_repo.lock(): | |
888 |
|
890 | |||
889 | # bookmark is not integrated to the streaming as it might use the |
|
891 | # bookmark is not integrated to the streaming as it might use the | |
890 | # `repo.vfs` and they are too many sentitive data accessible |
|
892 | # `repo.vfs` and they are too many sentitive data accessible | |
891 | # through `repo.vfs` to expose it to streaming clone. |
|
893 | # through `repo.vfs` to expose it to streaming clone. | |
892 | src_book_vfs = bookmarks.bookmarksvfs(src_repo) |
|
894 | src_book_vfs = bookmarks.bookmarksvfs(src_repo) | |
893 | srcbookmarks = src_book_vfs.join(b'bookmarks') |
|
895 | srcbookmarks = src_book_vfs.join(b'bookmarks') | |
894 | bm_count = 0 |
|
896 | bm_count = 0 | |
895 | if os.path.exists(srcbookmarks): |
|
897 | if os.path.exists(srcbookmarks): | |
896 | bm_count = 1 |
|
898 | bm_count = 1 | |
897 |
|
899 | |||
898 | entries, totalfilesize = _v2_walk( |
|
900 | entries, totalfilesize = _v2_walk( | |
899 | src_repo, |
|
901 | src_repo, | |
900 | includes=None, |
|
902 | includes=None, | |
901 | excludes=None, |
|
903 | excludes=None, | |
902 | includeobsmarkers=True, |
|
904 | includeobsmarkers=True, | |
903 | ) |
|
905 | ) | |
904 | src_vfs_map = _makemap(src_repo) |
|
906 | src_vfs_map = _makemap(src_repo) | |
905 | dest_vfs_map = _makemap(dest_repo) |
|
907 | dest_vfs_map = _makemap(dest_repo) | |
906 | progress = src_repo.ui.makeprogress( |
|
908 | progress = src_repo.ui.makeprogress( | |
907 | topic=_(b'linking'), |
|
909 | topic=_(b'linking'), | |
908 | total=len(entries) + bm_count, |
|
910 | total=len(entries) + bm_count, | |
909 | unit=_(b'files'), |
|
911 | unit=_(b'files'), | |
910 | ) |
|
912 | ) | |
911 | # copy files |
|
913 | # copy files | |
912 | # |
|
914 | # | |
913 | # We could copy the full file while the source repository is locked |
|
915 | # We could copy the full file while the source repository is locked | |
914 | # and the other one without the lock. However, in the linking case, |
|
916 | # and the other one without the lock. However, in the linking case, | |
915 | # this would also requires checks that nobody is appending any data |
|
917 | # this would also requires checks that nobody is appending any data | |
916 | # to the files while we do the clone, so this is not done yet. We |
|
918 | # to the files while we do the clone, so this is not done yet. We | |
917 | # could do this blindly when copying files. |
|
919 | # could do this blindly when copying files. | |
918 | files = ((k, path, size) for k, path, ftype, size in entries) |
|
920 | files = ((k, path, size) for k, path, ftype, size in entries) | |
919 | hardlink = _copy_files(src_vfs_map, dest_vfs_map, files, progress) |
|
921 | hardlink = _copy_files(src_vfs_map, dest_vfs_map, files, progress) | |
920 |
|
922 | |||
921 | # copy bookmarks over |
|
923 | # copy bookmarks over | |
922 | if bm_count: |
|
924 | if bm_count: | |
923 | dst_book_vfs = bookmarks.bookmarksvfs(dest_repo) |
|
925 | dst_book_vfs = bookmarks.bookmarksvfs(dest_repo) | |
924 | dstbookmarks = dst_book_vfs.join(b'bookmarks') |
|
926 | dstbookmarks = dst_book_vfs.join(b'bookmarks') | |
925 | util.copyfile(srcbookmarks, dstbookmarks) |
|
927 | util.copyfile(srcbookmarks, dstbookmarks) | |
926 | progress.complete() |
|
928 | progress.complete() | |
927 | if hardlink: |
|
929 | if hardlink: | |
928 | msg = b'linked %d files\n' |
|
930 | msg = b'linked %d files\n' | |
929 | else: |
|
931 | else: | |
930 | msg = b'copied %d files\n' |
|
932 | msg = b'copied %d files\n' | |
931 | src_repo.ui.debug(msg % (len(entries) + bm_count)) |
|
933 | src_repo.ui.debug(msg % (len(entries) + bm_count)) | |
932 |
|
934 | |||
933 | with dest_repo.transaction(b"localclone") as tr: |
|
935 | with dest_repo.transaction(b"localclone") as tr: | |
934 | dest_repo.store.write(tr) |
|
936 | dest_repo.store.write(tr) | |
935 |
|
937 | |||
936 | # clean up transaction file as they do not make sense |
|
938 | # clean up transaction file as they do not make sense | |
937 | transaction.cleanup_undo_files(dest_repo.ui.warn, dest_repo.vfs_map) |
|
939 | transaction.cleanup_undo_files(dest_repo.ui.warn, dest_repo.vfs_map) |
@@ -1,627 +1,627 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 entry in repo.store.datafiles(undecodable=undecodable): |
|
410 | for entry in repo.store.datafiles(undecodable=undecodable): | |
411 | for file_ in entry.files(): |
|
411 | for file_ in entry.files(): | |
412 | f = file_.unencoded_path |
|
412 | f = file_.unencoded_path | |
413 | size = file_.file_size |
|
413 | size = file_.file_size(repo.store.vfs) | |
414 | if (size > 0 or not revlogv1) and f.startswith(b'meta/'): |
|
414 | if (size > 0 or not revlogv1) and f.startswith(b'meta/'): | |
415 | storefiles.add(_normpath(f)) |
|
415 | storefiles.add(_normpath(f)) | |
416 | subdirs.add(os.path.dirname(f)) |
|
416 | subdirs.add(os.path.dirname(f)) | |
417 | for f in undecodable: |
|
417 | for f in undecodable: | |
418 | self._err(None, _(b"cannot decode filename '%s'") % f) |
|
418 | self._err(None, _(b"cannot decode filename '%s'") % f) | |
419 | subdirprogress = ui.makeprogress( |
|
419 | subdirprogress = ui.makeprogress( | |
420 | _(b'checking'), unit=_(b'manifests'), total=len(subdirs) |
|
420 | _(b'checking'), unit=_(b'manifests'), total=len(subdirs) | |
421 | ) |
|
421 | ) | |
422 |
|
422 | |||
423 | for subdir, linkrevs in subdirnodes.items(): |
|
423 | for subdir, linkrevs in subdirnodes.items(): | |
424 | subdirfilenodes = self._verifymanifest( |
|
424 | subdirfilenodes = self._verifymanifest( | |
425 | linkrevs, subdir, storefiles, subdirprogress |
|
425 | linkrevs, subdir, storefiles, subdirprogress | |
426 | ) |
|
426 | ) | |
427 | for f, onefilenodes in subdirfilenodes.items(): |
|
427 | for f, onefilenodes in subdirfilenodes.items(): | |
428 | filenodes.setdefault(f, {}).update(onefilenodes) |
|
428 | filenodes.setdefault(f, {}).update(onefilenodes) | |
429 |
|
429 | |||
430 | if not dir and subdirnodes: |
|
430 | if not dir and subdirnodes: | |
431 | assert subdirprogress is not None # help pytype |
|
431 | assert subdirprogress is not None # help pytype | |
432 | subdirprogress.complete() |
|
432 | subdirprogress.complete() | |
433 | if self.warnorphanstorefiles: |
|
433 | if self.warnorphanstorefiles: | |
434 | for f in sorted(storefiles): |
|
434 | for f in sorted(storefiles): | |
435 | self._warn(_(b"warning: orphan data file '%s'") % f) |
|
435 | self._warn(_(b"warning: orphan data file '%s'") % f) | |
436 |
|
436 | |||
437 | return filenodes |
|
437 | return filenodes | |
438 |
|
438 | |||
439 | def _crosscheckfiles(self, filelinkrevs, filenodes): |
|
439 | def _crosscheckfiles(self, filelinkrevs, filenodes): | |
440 | repo = self.repo |
|
440 | repo = self.repo | |
441 | ui = self.ui |
|
441 | ui = self.ui | |
442 | ui.status(_(b"crosschecking files in changesets and manifests\n")) |
|
442 | ui.status(_(b"crosschecking files in changesets and manifests\n")) | |
443 |
|
443 | |||
444 | total = len(filelinkrevs) + len(filenodes) |
|
444 | total = len(filelinkrevs) + len(filenodes) | |
445 | progress = ui.makeprogress( |
|
445 | progress = ui.makeprogress( | |
446 | _(b'crosschecking'), unit=_(b'files'), total=total |
|
446 | _(b'crosschecking'), unit=_(b'files'), total=total | |
447 | ) |
|
447 | ) | |
448 | if self.havemf: |
|
448 | if self.havemf: | |
449 | for f in sorted(filelinkrevs): |
|
449 | for f in sorted(filelinkrevs): | |
450 | progress.increment() |
|
450 | progress.increment() | |
451 | if f not in filenodes: |
|
451 | if f not in filenodes: | |
452 | lr = filelinkrevs[f][0] |
|
452 | lr = filelinkrevs[f][0] | |
453 | self._err(lr, _(b"in changeset but not in manifest"), f) |
|
453 | self._err(lr, _(b"in changeset but not in manifest"), f) | |
454 |
|
454 | |||
455 | if self.havecl: |
|
455 | if self.havecl: | |
456 | for f in sorted(filenodes): |
|
456 | for f in sorted(filenodes): | |
457 | progress.increment() |
|
457 | progress.increment() | |
458 | if f not in filelinkrevs: |
|
458 | if f not in filelinkrevs: | |
459 | try: |
|
459 | try: | |
460 | fl = repo.file(f) |
|
460 | fl = repo.file(f) | |
461 | lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]]) |
|
461 | lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]]) | |
462 | except Exception: |
|
462 | except Exception: | |
463 | lr = None |
|
463 | lr = None | |
464 | self._err(lr, _(b"in manifest but not in changeset"), f) |
|
464 | self._err(lr, _(b"in manifest but not in changeset"), f) | |
465 |
|
465 | |||
466 | progress.complete() |
|
466 | progress.complete() | |
467 |
|
467 | |||
468 | def _verifyfiles(self, filenodes, filelinkrevs): |
|
468 | def _verifyfiles(self, filenodes, filelinkrevs): | |
469 | repo = self.repo |
|
469 | repo = self.repo | |
470 | ui = self.ui |
|
470 | ui = self.ui | |
471 | lrugetctx = self.lrugetctx |
|
471 | lrugetctx = self.lrugetctx | |
472 | revlogv1 = self.revlogv1 |
|
472 | revlogv1 = self.revlogv1 | |
473 | havemf = self.havemf |
|
473 | havemf = self.havemf | |
474 | ui.status(_(b"checking files\n")) |
|
474 | ui.status(_(b"checking files\n")) | |
475 |
|
475 | |||
476 | storefiles = set() |
|
476 | storefiles = set() | |
477 | undecodable = [] |
|
477 | undecodable = [] | |
478 | for entry in repo.store.datafiles(undecodable=undecodable): |
|
478 | for entry in repo.store.datafiles(undecodable=undecodable): | |
479 | for file_ in entry.files(): |
|
479 | for file_ in entry.files(): | |
480 | size = file_.file_size |
|
480 | size = file_.file_size(repo.store.vfs) | |
481 | f = file_.unencoded_path |
|
481 | f = file_.unencoded_path | |
482 | if (size > 0 or not revlogv1) and f.startswith(b'data/'): |
|
482 | if (size > 0 or not revlogv1) and f.startswith(b'data/'): | |
483 | storefiles.add(_normpath(f)) |
|
483 | storefiles.add(_normpath(f)) | |
484 | for f in undecodable: |
|
484 | for f in undecodable: | |
485 | self._err(None, _(b"cannot decode filename '%s'") % f) |
|
485 | self._err(None, _(b"cannot decode filename '%s'") % f) | |
486 |
|
486 | |||
487 | state = { |
|
487 | state = { | |
488 | # TODO this assumes revlog storage for changelog. |
|
488 | # TODO this assumes revlog storage for changelog. | |
489 | b'expectedversion': self.repo.changelog._format_version, |
|
489 | b'expectedversion': self.repo.changelog._format_version, | |
490 | b'skipflags': self.skipflags, |
|
490 | b'skipflags': self.skipflags, | |
491 | # experimental config: censor.policy |
|
491 | # experimental config: censor.policy | |
492 | b'erroroncensored': ui.config(b'censor', b'policy') == b'abort', |
|
492 | b'erroroncensored': ui.config(b'censor', b'policy') == b'abort', | |
493 | } |
|
493 | } | |
494 |
|
494 | |||
495 | files = sorted(set(filenodes) | set(filelinkrevs)) |
|
495 | files = sorted(set(filenodes) | set(filelinkrevs)) | |
496 | revisions = 0 |
|
496 | revisions = 0 | |
497 | progress = ui.makeprogress( |
|
497 | progress = ui.makeprogress( | |
498 | _(b'checking'), unit=_(b'files'), total=len(files) |
|
498 | _(b'checking'), unit=_(b'files'), total=len(files) | |
499 | ) |
|
499 | ) | |
500 | for i, f in enumerate(files): |
|
500 | for i, f in enumerate(files): | |
501 | progress.update(i, item=f) |
|
501 | progress.update(i, item=f) | |
502 | try: |
|
502 | try: | |
503 | linkrevs = filelinkrevs[f] |
|
503 | linkrevs = filelinkrevs[f] | |
504 | except KeyError: |
|
504 | except KeyError: | |
505 | # in manifest but not in changelog |
|
505 | # in manifest but not in changelog | |
506 | linkrevs = [] |
|
506 | linkrevs = [] | |
507 |
|
507 | |||
508 | if linkrevs: |
|
508 | if linkrevs: | |
509 | lr = linkrevs[0] |
|
509 | lr = linkrevs[0] | |
510 | else: |
|
510 | else: | |
511 | lr = None |
|
511 | lr = None | |
512 |
|
512 | |||
513 | try: |
|
513 | try: | |
514 | fl = repo.file(f) |
|
514 | fl = repo.file(f) | |
515 | except error.StorageError as e: |
|
515 | except error.StorageError as e: | |
516 | self._err(lr, _(b"broken revlog! (%s)") % e, f) |
|
516 | self._err(lr, _(b"broken revlog! (%s)") % e, f) | |
517 | continue |
|
517 | continue | |
518 |
|
518 | |||
519 | for ff in fl.files(): |
|
519 | for ff in fl.files(): | |
520 | try: |
|
520 | try: | |
521 | storefiles.remove(ff) |
|
521 | storefiles.remove(ff) | |
522 | except KeyError: |
|
522 | except KeyError: | |
523 | if self.warnorphanstorefiles: |
|
523 | if self.warnorphanstorefiles: | |
524 | msg = _(b" warning: revlog '%s' not in fncache!") |
|
524 | msg = _(b" warning: revlog '%s' not in fncache!") | |
525 | self._warn(msg % ff) |
|
525 | self._warn(msg % ff) | |
526 | self.fncachewarned = True |
|
526 | self.fncachewarned = True | |
527 |
|
527 | |||
528 | if not len(fl) and (self.havecl or self.havemf): |
|
528 | if not len(fl) and (self.havecl or self.havemf): | |
529 | self._err(lr, _(b"empty or missing %s") % f) |
|
529 | self._err(lr, _(b"empty or missing %s") % f) | |
530 | else: |
|
530 | else: | |
531 | # Guard against implementations not setting this. |
|
531 | # Guard against implementations not setting this. | |
532 | state[b'skipread'] = set() |
|
532 | state[b'skipread'] = set() | |
533 | state[b'safe_renamed'] = set() |
|
533 | state[b'safe_renamed'] = set() | |
534 |
|
534 | |||
535 | for problem in fl.verifyintegrity(state): |
|
535 | for problem in fl.verifyintegrity(state): | |
536 | if problem.node is not None: |
|
536 | if problem.node is not None: | |
537 | linkrev = fl.linkrev(fl.rev(problem.node)) |
|
537 | linkrev = fl.linkrev(fl.rev(problem.node)) | |
538 | else: |
|
538 | else: | |
539 | linkrev = None |
|
539 | linkrev = None | |
540 |
|
540 | |||
541 | if problem.warning: |
|
541 | if problem.warning: | |
542 | self._warn(problem.warning) |
|
542 | self._warn(problem.warning) | |
543 | elif problem.error: |
|
543 | elif problem.error: | |
544 | linkrev_msg = linkrev if linkrev is not None else lr |
|
544 | linkrev_msg = linkrev if linkrev is not None else lr | |
545 | self._err(linkrev_msg, problem.error, f) |
|
545 | self._err(linkrev_msg, problem.error, f) | |
546 | else: |
|
546 | else: | |
547 | raise error.ProgrammingError( |
|
547 | raise error.ProgrammingError( | |
548 | b'problem instance does not set warning or error ' |
|
548 | b'problem instance does not set warning or error ' | |
549 | b'attribute: %s' % problem.msg |
|
549 | b'attribute: %s' % problem.msg | |
550 | ) |
|
550 | ) | |
551 |
|
551 | |||
552 | seen = {} |
|
552 | seen = {} | |
553 | for i in fl: |
|
553 | for i in fl: | |
554 | revisions += 1 |
|
554 | revisions += 1 | |
555 | n = fl.node(i) |
|
555 | n = fl.node(i) | |
556 | lr = self._checkentry(fl, i, n, seen, linkrevs, f) |
|
556 | lr = self._checkentry(fl, i, n, seen, linkrevs, f) | |
557 | if f in filenodes: |
|
557 | if f in filenodes: | |
558 | if havemf and n not in filenodes[f]: |
|
558 | if havemf and n not in filenodes[f]: | |
559 | self._err(lr, _(b"%s not in manifests") % (short(n)), f) |
|
559 | self._err(lr, _(b"%s not in manifests") % (short(n)), f) | |
560 | else: |
|
560 | else: | |
561 | del filenodes[f][n] |
|
561 | del filenodes[f][n] | |
562 |
|
562 | |||
563 | if n in state[b'skipread'] and n not in state[b'safe_renamed']: |
|
563 | if n in state[b'skipread'] and n not in state[b'safe_renamed']: | |
564 | continue |
|
564 | continue | |
565 |
|
565 | |||
566 | # check renames |
|
566 | # check renames | |
567 | try: |
|
567 | try: | |
568 | # This requires resolving fulltext (at least on revlogs, |
|
568 | # This requires resolving fulltext (at least on revlogs, | |
569 | # though not with LFS revisions). We may want |
|
569 | # though not with LFS revisions). We may want | |
570 | # ``verifyintegrity()`` to pass a set of nodes with |
|
570 | # ``verifyintegrity()`` to pass a set of nodes with | |
571 | # rename metadata as an optimization. |
|
571 | # rename metadata as an optimization. | |
572 | rp = fl.renamed(n) |
|
572 | rp = fl.renamed(n) | |
573 | if rp: |
|
573 | if rp: | |
574 | if lr is not None and ui.verbose: |
|
574 | if lr is not None and ui.verbose: | |
575 | ctx = lrugetctx(lr) |
|
575 | ctx = lrugetctx(lr) | |
576 | if not any(rp[0] in pctx for pctx in ctx.parents()): |
|
576 | if not any(rp[0] in pctx for pctx in ctx.parents()): | |
577 | self._warn(WARN_UNKNOWN_COPY_SOURCE % (f, ctx)) |
|
577 | self._warn(WARN_UNKNOWN_COPY_SOURCE % (f, ctx)) | |
578 | fl2 = repo.file(rp[0]) |
|
578 | fl2 = repo.file(rp[0]) | |
579 | if not len(fl2): |
|
579 | if not len(fl2): | |
580 | m = _(b"empty or missing copy source revlog %s:%s") |
|
580 | m = _(b"empty or missing copy source revlog %s:%s") | |
581 | self._err(lr, m % (rp[0], short(rp[1])), f) |
|
581 | self._err(lr, m % (rp[0], short(rp[1])), f) | |
582 | elif rp[1] == self.repo.nullid: |
|
582 | elif rp[1] == self.repo.nullid: | |
583 | msg = WARN_NULLID_COPY_SOURCE |
|
583 | msg = WARN_NULLID_COPY_SOURCE | |
584 | msg %= (f, lr, rp[0], short(rp[1])) |
|
584 | msg %= (f, lr, rp[0], short(rp[1])) | |
585 | ui.note(msg) |
|
585 | ui.note(msg) | |
586 | else: |
|
586 | else: | |
587 | fl2.rev(rp[1]) |
|
587 | fl2.rev(rp[1]) | |
588 | except Exception as inst: |
|
588 | except Exception as inst: | |
589 | self._exc( |
|
589 | self._exc( | |
590 | lr, _(b"checking rename of %s") % short(n), inst, f |
|
590 | lr, _(b"checking rename of %s") % short(n), inst, f | |
591 | ) |
|
591 | ) | |
592 |
|
592 | |||
593 | # cross-check |
|
593 | # cross-check | |
594 | if f in filenodes: |
|
594 | if f in filenodes: | |
595 | fns = [(v, k) for k, v in filenodes[f].items()] |
|
595 | fns = [(v, k) for k, v in filenodes[f].items()] | |
596 | for lr, node in sorted(fns): |
|
596 | for lr, node in sorted(fns): | |
597 | msg = _(b"manifest refers to unknown revision %s") |
|
597 | msg = _(b"manifest refers to unknown revision %s") | |
598 | self._err(lr, msg % short(node), f) |
|
598 | self._err(lr, msg % short(node), f) | |
599 | progress.complete() |
|
599 | progress.complete() | |
600 |
|
600 | |||
601 | if self.warnorphanstorefiles: |
|
601 | if self.warnorphanstorefiles: | |
602 | for f in sorted(storefiles): |
|
602 | for f in sorted(storefiles): | |
603 | self._warn(_(b"warning: orphan data file '%s'") % f) |
|
603 | self._warn(_(b"warning: orphan data file '%s'") % f) | |
604 |
|
604 | |||
605 | return len(files), revisions |
|
605 | return len(files), revisions | |
606 |
|
606 | |||
607 | def _verify_dirstate(self): |
|
607 | def _verify_dirstate(self): | |
608 | """Check that the dirstate is consistent with the parent's manifest""" |
|
608 | """Check that the dirstate is consistent with the parent's manifest""" | |
609 | repo = self.repo |
|
609 | repo = self.repo | |
610 | ui = self.ui |
|
610 | ui = self.ui | |
611 | ui.status(_(b"checking dirstate\n")) |
|
611 | ui.status(_(b"checking dirstate\n")) | |
612 |
|
612 | |||
613 | parent1, parent2 = repo.dirstate.parents() |
|
613 | parent1, parent2 = repo.dirstate.parents() | |
614 | m1 = repo[parent1].manifest() |
|
614 | m1 = repo[parent1].manifest() | |
615 | m2 = repo[parent2].manifest() |
|
615 | m2 = repo[parent2].manifest() | |
616 | dirstate_errors = 0 |
|
616 | dirstate_errors = 0 | |
617 |
|
617 | |||
618 | is_narrow = requirements.NARROW_REQUIREMENT in repo.requirements |
|
618 | is_narrow = requirements.NARROW_REQUIREMENT in repo.requirements | |
619 | narrow_matcher = repo.narrowmatch() if is_narrow else None |
|
619 | narrow_matcher = repo.narrowmatch() if is_narrow else None | |
620 |
|
620 | |||
621 | for err in repo.dirstate.verify(m1, m2, parent1, narrow_matcher): |
|
621 | for err in repo.dirstate.verify(m1, m2, parent1, narrow_matcher): | |
622 | ui.error(err) |
|
622 | ui.error(err) | |
623 | dirstate_errors += 1 |
|
623 | dirstate_errors += 1 | |
624 |
|
624 | |||
625 | if dirstate_errors: |
|
625 | if dirstate_errors: | |
626 | self.errors += dirstate_errors |
|
626 | self.errors += dirstate_errors | |
627 | return dirstate_errors |
|
627 | return dirstate_errors |
General Comments 0
You need to be logged in to leave comments.
Login now