Show More
@@ -1,577 +1,577 b'' | |||||
1 | # store.py - repository store handling for Mercurial |
|
1 | # store.py - repository store handling for Mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2008 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2008 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import errno |
|
10 | import errno | |
11 | import hashlib |
|
11 | import hashlib | |
12 | import os |
|
12 | import os | |
13 | import stat |
|
13 | import stat | |
14 |
|
14 | |||
15 | from .i18n import _ |
|
15 | from .i18n import _ | |
16 | from . import ( |
|
16 | from . import ( | |
17 | error, |
|
17 | error, | |
18 | parsers, |
|
18 | parsers, | |
19 | pycompat, |
|
19 | pycompat, | |
20 | scmutil, |
|
|||
21 | util, |
|
20 | util, | |
|
21 | vfs as vfsmod, | |||
22 | ) |
|
22 | ) | |
23 |
|
23 | |||
24 | # This avoids a collision between a file named foo and a dir named |
|
24 | # This avoids a collision between a file named foo and a dir named | |
25 | # foo.i or foo.d |
|
25 | # foo.i or foo.d | |
26 | def _encodedir(path): |
|
26 | def _encodedir(path): | |
27 | ''' |
|
27 | ''' | |
28 | >>> _encodedir('data/foo.i') |
|
28 | >>> _encodedir('data/foo.i') | |
29 | 'data/foo.i' |
|
29 | 'data/foo.i' | |
30 | >>> _encodedir('data/foo.i/bla.i') |
|
30 | >>> _encodedir('data/foo.i/bla.i') | |
31 | 'data/foo.i.hg/bla.i' |
|
31 | 'data/foo.i.hg/bla.i' | |
32 | >>> _encodedir('data/foo.i.hg/bla.i') |
|
32 | >>> _encodedir('data/foo.i.hg/bla.i') | |
33 | 'data/foo.i.hg.hg/bla.i' |
|
33 | 'data/foo.i.hg.hg/bla.i' | |
34 | >>> _encodedir('data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n') |
|
34 | >>> _encodedir('data/foo.i\\ndata/foo.i/bla.i\\ndata/foo.i.hg/bla.i\\n') | |
35 | 'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n' |
|
35 | 'data/foo.i\\ndata/foo.i.hg/bla.i\\ndata/foo.i.hg.hg/bla.i\\n' | |
36 | ''' |
|
36 | ''' | |
37 | return (path |
|
37 | return (path | |
38 | .replace(".hg/", ".hg.hg/") |
|
38 | .replace(".hg/", ".hg.hg/") | |
39 | .replace(".i/", ".i.hg/") |
|
39 | .replace(".i/", ".i.hg/") | |
40 | .replace(".d/", ".d.hg/")) |
|
40 | .replace(".d/", ".d.hg/")) | |
41 |
|
41 | |||
42 | encodedir = getattr(parsers, 'encodedir', _encodedir) |
|
42 | encodedir = getattr(parsers, 'encodedir', _encodedir) | |
43 |
|
43 | |||
44 | def decodedir(path): |
|
44 | def decodedir(path): | |
45 | ''' |
|
45 | ''' | |
46 | >>> decodedir('data/foo.i') |
|
46 | >>> decodedir('data/foo.i') | |
47 | 'data/foo.i' |
|
47 | 'data/foo.i' | |
48 | >>> decodedir('data/foo.i.hg/bla.i') |
|
48 | >>> decodedir('data/foo.i.hg/bla.i') | |
49 | 'data/foo.i/bla.i' |
|
49 | 'data/foo.i/bla.i' | |
50 | >>> decodedir('data/foo.i.hg.hg/bla.i') |
|
50 | >>> decodedir('data/foo.i.hg.hg/bla.i') | |
51 | 'data/foo.i.hg/bla.i' |
|
51 | 'data/foo.i.hg/bla.i' | |
52 | ''' |
|
52 | ''' | |
53 | if ".hg/" not in path: |
|
53 | if ".hg/" not in path: | |
54 | return path |
|
54 | return path | |
55 | return (path |
|
55 | return (path | |
56 | .replace(".d.hg/", ".d/") |
|
56 | .replace(".d.hg/", ".d/") | |
57 | .replace(".i.hg/", ".i/") |
|
57 | .replace(".i.hg/", ".i/") | |
58 | .replace(".hg.hg/", ".hg/")) |
|
58 | .replace(".hg.hg/", ".hg/")) | |
59 |
|
59 | |||
60 | def _reserved(): |
|
60 | def _reserved(): | |
61 | ''' characters that are problematic for filesystems |
|
61 | ''' characters that are problematic for filesystems | |
62 |
|
62 | |||
63 | * ascii escapes (0..31) |
|
63 | * ascii escapes (0..31) | |
64 | * ascii hi (126..255) |
|
64 | * ascii hi (126..255) | |
65 | * windows specials |
|
65 | * windows specials | |
66 |
|
66 | |||
67 | these characters will be escaped by encodefunctions |
|
67 | these characters will be escaped by encodefunctions | |
68 | ''' |
|
68 | ''' | |
69 | winreserved = [ord(x) for x in u'\\:*?"<>|'] |
|
69 | winreserved = [ord(x) for x in u'\\:*?"<>|'] | |
70 | for x in range(32): |
|
70 | for x in range(32): | |
71 | yield x |
|
71 | yield x | |
72 | for x in range(126, 256): |
|
72 | for x in range(126, 256): | |
73 | yield x |
|
73 | yield x | |
74 | for x in winreserved: |
|
74 | for x in winreserved: | |
75 | yield x |
|
75 | yield x | |
76 |
|
76 | |||
77 | def _buildencodefun(): |
|
77 | def _buildencodefun(): | |
78 | ''' |
|
78 | ''' | |
79 | >>> enc, dec = _buildencodefun() |
|
79 | >>> enc, dec = _buildencodefun() | |
80 |
|
80 | |||
81 | >>> enc('nothing/special.txt') |
|
81 | >>> enc('nothing/special.txt') | |
82 | 'nothing/special.txt' |
|
82 | 'nothing/special.txt' | |
83 | >>> dec('nothing/special.txt') |
|
83 | >>> dec('nothing/special.txt') | |
84 | 'nothing/special.txt' |
|
84 | 'nothing/special.txt' | |
85 |
|
85 | |||
86 | >>> enc('HELLO') |
|
86 | >>> enc('HELLO') | |
87 | '_h_e_l_l_o' |
|
87 | '_h_e_l_l_o' | |
88 | >>> dec('_h_e_l_l_o') |
|
88 | >>> dec('_h_e_l_l_o') | |
89 | 'HELLO' |
|
89 | 'HELLO' | |
90 |
|
90 | |||
91 | >>> enc('hello:world?') |
|
91 | >>> enc('hello:world?') | |
92 | 'hello~3aworld~3f' |
|
92 | 'hello~3aworld~3f' | |
93 | >>> dec('hello~3aworld~3f') |
|
93 | >>> dec('hello~3aworld~3f') | |
94 | 'hello:world?' |
|
94 | 'hello:world?' | |
95 |
|
95 | |||
96 | >>> enc('the\x07quick\xADshot') |
|
96 | >>> enc('the\x07quick\xADshot') | |
97 | 'the~07quick~adshot' |
|
97 | 'the~07quick~adshot' | |
98 | >>> dec('the~07quick~adshot') |
|
98 | >>> dec('the~07quick~adshot') | |
99 | 'the\\x07quick\\xadshot' |
|
99 | 'the\\x07quick\\xadshot' | |
100 | ''' |
|
100 | ''' | |
101 | e = '_' |
|
101 | e = '_' | |
102 | if pycompat.ispy3: |
|
102 | if pycompat.ispy3: | |
103 | xchr = lambda x: bytes([x]) |
|
103 | xchr = lambda x: bytes([x]) | |
104 | asciistr = [bytes([a]) for a in range(127)] |
|
104 | asciistr = [bytes([a]) for a in range(127)] | |
105 | else: |
|
105 | else: | |
106 | xchr = chr |
|
106 | xchr = chr | |
107 | asciistr = map(chr, xrange(127)) |
|
107 | asciistr = map(chr, xrange(127)) | |
108 | capitals = list(range(ord("A"), ord("Z") + 1)) |
|
108 | capitals = list(range(ord("A"), ord("Z") + 1)) | |
109 |
|
109 | |||
110 | cmap = dict((x, x) for x in asciistr) |
|
110 | cmap = dict((x, x) for x in asciistr) | |
111 | for x in _reserved(): |
|
111 | for x in _reserved(): | |
112 | cmap[xchr(x)] = "~%02x" % x |
|
112 | cmap[xchr(x)] = "~%02x" % x | |
113 | for x in capitals + [ord(e)]: |
|
113 | for x in capitals + [ord(e)]: | |
114 | cmap[xchr(x)] = e + xchr(x).lower() |
|
114 | cmap[xchr(x)] = e + xchr(x).lower() | |
115 |
|
115 | |||
116 | dmap = {} |
|
116 | dmap = {} | |
117 | for k, v in cmap.iteritems(): |
|
117 | for k, v in cmap.iteritems(): | |
118 | dmap[v] = k |
|
118 | dmap[v] = k | |
119 | def decode(s): |
|
119 | def decode(s): | |
120 | i = 0 |
|
120 | i = 0 | |
121 | while i < len(s): |
|
121 | while i < len(s): | |
122 | for l in xrange(1, 4): |
|
122 | for l in xrange(1, 4): | |
123 | try: |
|
123 | try: | |
124 | yield dmap[s[i:i + l]] |
|
124 | yield dmap[s[i:i + l]] | |
125 | i += l |
|
125 | i += l | |
126 | break |
|
126 | break | |
127 | except KeyError: |
|
127 | except KeyError: | |
128 | pass |
|
128 | pass | |
129 | else: |
|
129 | else: | |
130 | raise KeyError |
|
130 | raise KeyError | |
131 | return (lambda s: ''.join([cmap[s[c:c + 1]] for c in xrange(len(s))]), |
|
131 | return (lambda s: ''.join([cmap[s[c:c + 1]] for c in xrange(len(s))]), | |
132 | lambda s: ''.join(list(decode(s)))) |
|
132 | lambda s: ''.join(list(decode(s)))) | |
133 |
|
133 | |||
134 | _encodefname, _decodefname = _buildencodefun() |
|
134 | _encodefname, _decodefname = _buildencodefun() | |
135 |
|
135 | |||
136 | def encodefilename(s): |
|
136 | def encodefilename(s): | |
137 | ''' |
|
137 | ''' | |
138 | >>> encodefilename('foo.i/bar.d/bla.hg/hi:world?/HELLO') |
|
138 | >>> encodefilename('foo.i/bar.d/bla.hg/hi:world?/HELLO') | |
139 | 'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o' |
|
139 | 'foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o' | |
140 | ''' |
|
140 | ''' | |
141 | return _encodefname(encodedir(s)) |
|
141 | return _encodefname(encodedir(s)) | |
142 |
|
142 | |||
143 | def decodefilename(s): |
|
143 | def decodefilename(s): | |
144 | ''' |
|
144 | ''' | |
145 | >>> decodefilename('foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o') |
|
145 | >>> decodefilename('foo.i.hg/bar.d.hg/bla.hg.hg/hi~3aworld~3f/_h_e_l_l_o') | |
146 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' |
|
146 | 'foo.i/bar.d/bla.hg/hi:world?/HELLO' | |
147 | ''' |
|
147 | ''' | |
148 | return decodedir(_decodefname(s)) |
|
148 | return decodedir(_decodefname(s)) | |
149 |
|
149 | |||
150 | def _buildlowerencodefun(): |
|
150 | def _buildlowerencodefun(): | |
151 | ''' |
|
151 | ''' | |
152 | >>> f = _buildlowerencodefun() |
|
152 | >>> f = _buildlowerencodefun() | |
153 | >>> f('nothing/special.txt') |
|
153 | >>> f('nothing/special.txt') | |
154 | 'nothing/special.txt' |
|
154 | 'nothing/special.txt' | |
155 | >>> f('HELLO') |
|
155 | >>> f('HELLO') | |
156 | 'hello' |
|
156 | 'hello' | |
157 | >>> f('hello:world?') |
|
157 | >>> f('hello:world?') | |
158 | 'hello~3aworld~3f' |
|
158 | 'hello~3aworld~3f' | |
159 | >>> f('the\x07quick\xADshot') |
|
159 | >>> f('the\x07quick\xADshot') | |
160 | 'the~07quick~adshot' |
|
160 | 'the~07quick~adshot' | |
161 | ''' |
|
161 | ''' | |
162 | cmap = dict([(chr(x), chr(x)) for x in xrange(127)]) |
|
162 | cmap = dict([(chr(x), chr(x)) for x in xrange(127)]) | |
163 | for x in _reserved(): |
|
163 | for x in _reserved(): | |
164 | cmap[chr(x)] = "~%02x" % x |
|
164 | cmap[chr(x)] = "~%02x" % x | |
165 | for x in range(ord("A"), ord("Z") + 1): |
|
165 | for x in range(ord("A"), ord("Z") + 1): | |
166 | cmap[chr(x)] = chr(x).lower() |
|
166 | cmap[chr(x)] = chr(x).lower() | |
167 | return lambda s: "".join([cmap[c] for c in s]) |
|
167 | return lambda s: "".join([cmap[c] for c in s]) | |
168 |
|
168 | |||
169 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() |
|
169 | lowerencode = getattr(parsers, 'lowerencode', None) or _buildlowerencodefun() | |
170 |
|
170 | |||
171 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 |
|
171 | # Windows reserved names: con, prn, aux, nul, com1..com9, lpt1..lpt9 | |
172 | _winres3 = ('aux', 'con', 'prn', 'nul') # length 3 |
|
172 | _winres3 = ('aux', 'con', 'prn', 'nul') # length 3 | |
173 | _winres4 = ('com', 'lpt') # length 4 (with trailing 1..9) |
|
173 | _winres4 = ('com', 'lpt') # length 4 (with trailing 1..9) | |
174 | def _auxencode(path, dotencode): |
|
174 | def _auxencode(path, dotencode): | |
175 | ''' |
|
175 | ''' | |
176 | Encodes filenames containing names reserved by Windows or which end in |
|
176 | Encodes filenames containing names reserved by Windows or which end in | |
177 | period or space. Does not touch other single reserved characters c. |
|
177 | period or space. Does not touch other single reserved characters c. | |
178 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. |
|
178 | Specifically, c in '\\:*?"<>|' or ord(c) <= 31 are *not* encoded here. | |
179 | Additionally encodes space or period at the beginning, if dotencode is |
|
179 | Additionally encodes space or period at the beginning, if dotencode is | |
180 | True. Parameter path is assumed to be all lowercase. |
|
180 | True. Parameter path is assumed to be all lowercase. | |
181 | A segment only needs encoding if a reserved name appears as a |
|
181 | A segment only needs encoding if a reserved name appears as a | |
182 | basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux" |
|
182 | basename (e.g. "aux", "aux.foo"). A directory or file named "foo.aux" | |
183 | doesn't need encoding. |
|
183 | doesn't need encoding. | |
184 |
|
184 | |||
185 | >>> s = '.foo/aux.txt/txt.aux/con/prn/nul/foo.' |
|
185 | >>> s = '.foo/aux.txt/txt.aux/con/prn/nul/foo.' | |
186 | >>> _auxencode(s.split('/'), True) |
|
186 | >>> _auxencode(s.split('/'), True) | |
187 | ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e'] |
|
187 | ['~2efoo', 'au~78.txt', 'txt.aux', 'co~6e', 'pr~6e', 'nu~6c', 'foo~2e'] | |
188 | >>> s = '.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' |
|
188 | >>> s = '.com1com2/lpt9.lpt4.lpt1/conprn/com0/lpt0/foo.' | |
189 | >>> _auxencode(s.split('/'), False) |
|
189 | >>> _auxencode(s.split('/'), False) | |
190 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] |
|
190 | ['.com1com2', 'lp~749.lpt4.lpt1', 'conprn', 'com0', 'lpt0', 'foo~2e'] | |
191 | >>> _auxencode(['foo. '], True) |
|
191 | >>> _auxencode(['foo. '], True) | |
192 | ['foo.~20'] |
|
192 | ['foo.~20'] | |
193 | >>> _auxencode([' .foo'], True) |
|
193 | >>> _auxencode([' .foo'], True) | |
194 | ['~20.foo'] |
|
194 | ['~20.foo'] | |
195 | ''' |
|
195 | ''' | |
196 | for i, n in enumerate(path): |
|
196 | for i, n in enumerate(path): | |
197 | if not n: |
|
197 | if not n: | |
198 | continue |
|
198 | continue | |
199 | if dotencode and n[0] in '. ': |
|
199 | if dotencode and n[0] in '. ': | |
200 | n = "~%02x" % ord(n[0]) + n[1:] |
|
200 | n = "~%02x" % ord(n[0]) + n[1:] | |
201 | path[i] = n |
|
201 | path[i] = n | |
202 | else: |
|
202 | else: | |
203 | l = n.find('.') |
|
203 | l = n.find('.') | |
204 | if l == -1: |
|
204 | if l == -1: | |
205 | l = len(n) |
|
205 | l = len(n) | |
206 | if ((l == 3 and n[:3] in _winres3) or |
|
206 | if ((l == 3 and n[:3] in _winres3) or | |
207 | (l == 4 and n[3] <= '9' and n[3] >= '1' |
|
207 | (l == 4 and n[3] <= '9' and n[3] >= '1' | |
208 | and n[:3] in _winres4)): |
|
208 | and n[:3] in _winres4)): | |
209 | # encode third letter ('aux' -> 'au~78') |
|
209 | # encode third letter ('aux' -> 'au~78') | |
210 | ec = "~%02x" % ord(n[2]) |
|
210 | ec = "~%02x" % ord(n[2]) | |
211 | n = n[0:2] + ec + n[3:] |
|
211 | n = n[0:2] + ec + n[3:] | |
212 | path[i] = n |
|
212 | path[i] = n | |
213 | if n[-1] in '. ': |
|
213 | if n[-1] in '. ': | |
214 | # encode last period or space ('foo...' -> 'foo..~2e') |
|
214 | # encode last period or space ('foo...' -> 'foo..~2e') | |
215 | path[i] = n[:-1] + "~%02x" % ord(n[-1]) |
|
215 | path[i] = n[:-1] + "~%02x" % ord(n[-1]) | |
216 | return path |
|
216 | return path | |
217 |
|
217 | |||
218 | _maxstorepathlen = 120 |
|
218 | _maxstorepathlen = 120 | |
219 | _dirprefixlen = 8 |
|
219 | _dirprefixlen = 8 | |
220 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 |
|
220 | _maxshortdirslen = 8 * (_dirprefixlen + 1) - 4 | |
221 |
|
221 | |||
222 | def _hashencode(path, dotencode): |
|
222 | def _hashencode(path, dotencode): | |
223 | digest = hashlib.sha1(path).hexdigest() |
|
223 | digest = hashlib.sha1(path).hexdigest() | |
224 | le = lowerencode(path[5:]).split('/') # skips prefix 'data/' or 'meta/' |
|
224 | le = lowerencode(path[5:]).split('/') # skips prefix 'data/' or 'meta/' | |
225 | parts = _auxencode(le, dotencode) |
|
225 | parts = _auxencode(le, dotencode) | |
226 | basename = parts[-1] |
|
226 | basename = parts[-1] | |
227 | _root, ext = os.path.splitext(basename) |
|
227 | _root, ext = os.path.splitext(basename) | |
228 | sdirs = [] |
|
228 | sdirs = [] | |
229 | sdirslen = 0 |
|
229 | sdirslen = 0 | |
230 | for p in parts[:-1]: |
|
230 | for p in parts[:-1]: | |
231 | d = p[:_dirprefixlen] |
|
231 | d = p[:_dirprefixlen] | |
232 | if d[-1] in '. ': |
|
232 | if d[-1] in '. ': | |
233 | # Windows can't access dirs ending in period or space |
|
233 | # Windows can't access dirs ending in period or space | |
234 | d = d[:-1] + '_' |
|
234 | d = d[:-1] + '_' | |
235 | if sdirslen == 0: |
|
235 | if sdirslen == 0: | |
236 | t = len(d) |
|
236 | t = len(d) | |
237 | else: |
|
237 | else: | |
238 | t = sdirslen + 1 + len(d) |
|
238 | t = sdirslen + 1 + len(d) | |
239 | if t > _maxshortdirslen: |
|
239 | if t > _maxshortdirslen: | |
240 | break |
|
240 | break | |
241 | sdirs.append(d) |
|
241 | sdirs.append(d) | |
242 | sdirslen = t |
|
242 | sdirslen = t | |
243 | dirs = '/'.join(sdirs) |
|
243 | dirs = '/'.join(sdirs) | |
244 | if len(dirs) > 0: |
|
244 | if len(dirs) > 0: | |
245 | dirs += '/' |
|
245 | dirs += '/' | |
246 | res = 'dh/' + dirs + digest + ext |
|
246 | res = 'dh/' + dirs + digest + ext | |
247 | spaceleft = _maxstorepathlen - len(res) |
|
247 | spaceleft = _maxstorepathlen - len(res) | |
248 | if spaceleft > 0: |
|
248 | if spaceleft > 0: | |
249 | filler = basename[:spaceleft] |
|
249 | filler = basename[:spaceleft] | |
250 | res = 'dh/' + dirs + filler + digest + ext |
|
250 | res = 'dh/' + dirs + filler + digest + ext | |
251 | return res |
|
251 | return res | |
252 |
|
252 | |||
253 | def _hybridencode(path, dotencode): |
|
253 | def _hybridencode(path, dotencode): | |
254 | '''encodes path with a length limit |
|
254 | '''encodes path with a length limit | |
255 |
|
255 | |||
256 | Encodes all paths that begin with 'data/', according to the following. |
|
256 | Encodes all paths that begin with 'data/', according to the following. | |
257 |
|
257 | |||
258 | Default encoding (reversible): |
|
258 | Default encoding (reversible): | |
259 |
|
259 | |||
260 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal |
|
260 | Encodes all uppercase letters 'X' as '_x'. All reserved or illegal | |
261 | characters are encoded as '~xx', where xx is the two digit hex code |
|
261 | characters are encoded as '~xx', where xx is the two digit hex code | |
262 | of the character (see encodefilename). |
|
262 | of the character (see encodefilename). | |
263 | Relevant path components consisting of Windows reserved filenames are |
|
263 | Relevant path components consisting of Windows reserved filenames are | |
264 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). |
|
264 | masked by encoding the third character ('aux' -> 'au~78', see _auxencode). | |
265 |
|
265 | |||
266 | Hashed encoding (not reversible): |
|
266 | Hashed encoding (not reversible): | |
267 |
|
267 | |||
268 | If the default-encoded path is longer than _maxstorepathlen, a |
|
268 | If the default-encoded path is longer than _maxstorepathlen, a | |
269 | non-reversible hybrid hashing of the path is done instead. |
|
269 | non-reversible hybrid hashing of the path is done instead. | |
270 | This encoding uses up to _dirprefixlen characters of all directory |
|
270 | This encoding uses up to _dirprefixlen characters of all directory | |
271 | levels of the lowerencoded path, but not more levels than can fit into |
|
271 | levels of the lowerencoded path, but not more levels than can fit into | |
272 | _maxshortdirslen. |
|
272 | _maxshortdirslen. | |
273 | Then follows the filler followed by the sha digest of the full path. |
|
273 | Then follows the filler followed by the sha digest of the full path. | |
274 | The filler is the beginning of the basename of the lowerencoded path |
|
274 | The filler is the beginning of the basename of the lowerencoded path | |
275 | (the basename is everything after the last path separator). The filler |
|
275 | (the basename is everything after the last path separator). The filler | |
276 | is as long as possible, filling in characters from the basename until |
|
276 | is as long as possible, filling in characters from the basename until | |
277 | the encoded path has _maxstorepathlen characters (or all chars of the |
|
277 | the encoded path has _maxstorepathlen characters (or all chars of the | |
278 | basename have been taken). |
|
278 | basename have been taken). | |
279 | The extension (e.g. '.i' or '.d') is preserved. |
|
279 | The extension (e.g. '.i' or '.d') is preserved. | |
280 |
|
280 | |||
281 | The string 'data/' at the beginning is replaced with 'dh/', if the hashed |
|
281 | The string 'data/' at the beginning is replaced with 'dh/', if the hashed | |
282 | encoding was used. |
|
282 | encoding was used. | |
283 | ''' |
|
283 | ''' | |
284 | path = encodedir(path) |
|
284 | path = encodedir(path) | |
285 | ef = _encodefname(path).split('/') |
|
285 | ef = _encodefname(path).split('/') | |
286 | res = '/'.join(_auxencode(ef, dotencode)) |
|
286 | res = '/'.join(_auxencode(ef, dotencode)) | |
287 | if len(res) > _maxstorepathlen: |
|
287 | if len(res) > _maxstorepathlen: | |
288 | res = _hashencode(path, dotencode) |
|
288 | res = _hashencode(path, dotencode) | |
289 | return res |
|
289 | return res | |
290 |
|
290 | |||
291 | def _pathencode(path): |
|
291 | def _pathencode(path): | |
292 | de = encodedir(path) |
|
292 | de = encodedir(path) | |
293 | if len(path) > _maxstorepathlen: |
|
293 | if len(path) > _maxstorepathlen: | |
294 | return _hashencode(de, True) |
|
294 | return _hashencode(de, True) | |
295 | ef = _encodefname(de).split('/') |
|
295 | ef = _encodefname(de).split('/') | |
296 | res = '/'.join(_auxencode(ef, True)) |
|
296 | res = '/'.join(_auxencode(ef, True)) | |
297 | if len(res) > _maxstorepathlen: |
|
297 | if len(res) > _maxstorepathlen: | |
298 | return _hashencode(de, True) |
|
298 | return _hashencode(de, True) | |
299 | return res |
|
299 | return res | |
300 |
|
300 | |||
301 | _pathencode = getattr(parsers, 'pathencode', _pathencode) |
|
301 | _pathencode = getattr(parsers, 'pathencode', _pathencode) | |
302 |
|
302 | |||
303 | def _plainhybridencode(f): |
|
303 | def _plainhybridencode(f): | |
304 | return _hybridencode(f, False) |
|
304 | return _hybridencode(f, False) | |
305 |
|
305 | |||
306 | def _calcmode(vfs): |
|
306 | def _calcmode(vfs): | |
307 | try: |
|
307 | try: | |
308 | # files in .hg/ will be created using this mode |
|
308 | # files in .hg/ will be created using this mode | |
309 | mode = vfs.stat().st_mode |
|
309 | mode = vfs.stat().st_mode | |
310 | # avoid some useless chmods |
|
310 | # avoid some useless chmods | |
311 | if (0o777 & ~util.umask) == (0o777 & mode): |
|
311 | if (0o777 & ~util.umask) == (0o777 & mode): | |
312 | mode = None |
|
312 | mode = None | |
313 | except OSError: |
|
313 | except OSError: | |
314 | mode = None |
|
314 | mode = None | |
315 | return mode |
|
315 | return mode | |
316 |
|
316 | |||
317 | _data = ('data meta 00manifest.d 00manifest.i 00changelog.d 00changelog.i' |
|
317 | _data = ('data meta 00manifest.d 00manifest.i 00changelog.d 00changelog.i' | |
318 | ' phaseroots obsstore') |
|
318 | ' phaseroots obsstore') | |
319 |
|
319 | |||
320 | class basicstore(object): |
|
320 | class basicstore(object): | |
321 | '''base class for local repository stores''' |
|
321 | '''base class for local repository stores''' | |
322 | def __init__(self, path, vfstype): |
|
322 | def __init__(self, path, vfstype): | |
323 | vfs = vfstype(path) |
|
323 | vfs = vfstype(path) | |
324 | self.path = vfs.base |
|
324 | self.path = vfs.base | |
325 | self.createmode = _calcmode(vfs) |
|
325 | self.createmode = _calcmode(vfs) | |
326 | vfs.createmode = self.createmode |
|
326 | vfs.createmode = self.createmode | |
327 | self.rawvfs = vfs |
|
327 | self.rawvfs = vfs | |
328 |
self.vfs = |
|
328 | self.vfs = vfsmod.filtervfs(vfs, encodedir) | |
329 | self.opener = self.vfs |
|
329 | self.opener = self.vfs | |
330 |
|
330 | |||
331 | def join(self, f): |
|
331 | def join(self, f): | |
332 | return self.path + '/' + encodedir(f) |
|
332 | return self.path + '/' + encodedir(f) | |
333 |
|
333 | |||
334 | def _walk(self, relpath, recurse): |
|
334 | def _walk(self, relpath, recurse): | |
335 | '''yields (unencoded, encoded, size)''' |
|
335 | '''yields (unencoded, encoded, size)''' | |
336 | path = self.path |
|
336 | path = self.path | |
337 | if relpath: |
|
337 | if relpath: | |
338 | path += '/' + relpath |
|
338 | path += '/' + relpath | |
339 | striplen = len(self.path) + 1 |
|
339 | striplen = len(self.path) + 1 | |
340 | l = [] |
|
340 | l = [] | |
341 | if self.rawvfs.isdir(path): |
|
341 | if self.rawvfs.isdir(path): | |
342 | visit = [path] |
|
342 | visit = [path] | |
343 | readdir = self.rawvfs.readdir |
|
343 | readdir = self.rawvfs.readdir | |
344 | while visit: |
|
344 | while visit: | |
345 | p = visit.pop() |
|
345 | p = visit.pop() | |
346 | for f, kind, st in readdir(p, stat=True): |
|
346 | for f, kind, st in readdir(p, stat=True): | |
347 | fp = p + '/' + f |
|
347 | fp = p + '/' + f | |
348 | if kind == stat.S_IFREG and f[-2:] in ('.d', '.i'): |
|
348 | if kind == stat.S_IFREG and f[-2:] in ('.d', '.i'): | |
349 | n = util.pconvert(fp[striplen:]) |
|
349 | n = util.pconvert(fp[striplen:]) | |
350 | l.append((decodedir(n), n, st.st_size)) |
|
350 | l.append((decodedir(n), n, st.st_size)) | |
351 | elif kind == stat.S_IFDIR and recurse: |
|
351 | elif kind == stat.S_IFDIR and recurse: | |
352 | visit.append(fp) |
|
352 | visit.append(fp) | |
353 | l.sort() |
|
353 | l.sort() | |
354 | return l |
|
354 | return l | |
355 |
|
355 | |||
356 | def datafiles(self): |
|
356 | def datafiles(self): | |
357 | return self._walk('data', True) + self._walk('meta', True) |
|
357 | return self._walk('data', True) + self._walk('meta', True) | |
358 |
|
358 | |||
359 | def topfiles(self): |
|
359 | def topfiles(self): | |
360 | # yield manifest before changelog |
|
360 | # yield manifest before changelog | |
361 | return reversed(self._walk('', False)) |
|
361 | return reversed(self._walk('', False)) | |
362 |
|
362 | |||
363 | def walk(self): |
|
363 | def walk(self): | |
364 | '''yields (unencoded, encoded, size)''' |
|
364 | '''yields (unencoded, encoded, size)''' | |
365 | # yield data files first |
|
365 | # yield data files first | |
366 | for x in self.datafiles(): |
|
366 | for x in self.datafiles(): | |
367 | yield x |
|
367 | yield x | |
368 | for x in self.topfiles(): |
|
368 | for x in self.topfiles(): | |
369 | yield x |
|
369 | yield x | |
370 |
|
370 | |||
371 | def copylist(self): |
|
371 | def copylist(self): | |
372 | return ['requires'] + _data.split() |
|
372 | return ['requires'] + _data.split() | |
373 |
|
373 | |||
374 | def write(self, tr): |
|
374 | def write(self, tr): | |
375 | pass |
|
375 | pass | |
376 |
|
376 | |||
377 | def invalidatecaches(self): |
|
377 | def invalidatecaches(self): | |
378 | pass |
|
378 | pass | |
379 |
|
379 | |||
380 | def markremoved(self, fn): |
|
380 | def markremoved(self, fn): | |
381 | pass |
|
381 | pass | |
382 |
|
382 | |||
383 | def __contains__(self, path): |
|
383 | def __contains__(self, path): | |
384 | '''Checks if the store contains path''' |
|
384 | '''Checks if the store contains path''' | |
385 | path = "/".join(("data", path)) |
|
385 | path = "/".join(("data", path)) | |
386 | # file? |
|
386 | # file? | |
387 | if self.vfs.exists(path + ".i"): |
|
387 | if self.vfs.exists(path + ".i"): | |
388 | return True |
|
388 | return True | |
389 | # dir? |
|
389 | # dir? | |
390 | if not path.endswith("/"): |
|
390 | if not path.endswith("/"): | |
391 | path = path + "/" |
|
391 | path = path + "/" | |
392 | return self.vfs.exists(path) |
|
392 | return self.vfs.exists(path) | |
393 |
|
393 | |||
394 | class encodedstore(basicstore): |
|
394 | class encodedstore(basicstore): | |
395 | def __init__(self, path, vfstype): |
|
395 | def __init__(self, path, vfstype): | |
396 | vfs = vfstype(path + '/store') |
|
396 | vfs = vfstype(path + '/store') | |
397 | self.path = vfs.base |
|
397 | self.path = vfs.base | |
398 | self.createmode = _calcmode(vfs) |
|
398 | self.createmode = _calcmode(vfs) | |
399 | vfs.createmode = self.createmode |
|
399 | vfs.createmode = self.createmode | |
400 | self.rawvfs = vfs |
|
400 | self.rawvfs = vfs | |
401 |
self.vfs = |
|
401 | self.vfs = vfsmod.filtervfs(vfs, encodefilename) | |
402 | self.opener = self.vfs |
|
402 | self.opener = self.vfs | |
403 |
|
403 | |||
404 | def datafiles(self): |
|
404 | def datafiles(self): | |
405 | for a, b, size in super(encodedstore, self).datafiles(): |
|
405 | for a, b, size in super(encodedstore, self).datafiles(): | |
406 | try: |
|
406 | try: | |
407 | a = decodefilename(a) |
|
407 | a = decodefilename(a) | |
408 | except KeyError: |
|
408 | except KeyError: | |
409 | a = None |
|
409 | a = None | |
410 | yield a, b, size |
|
410 | yield a, b, size | |
411 |
|
411 | |||
412 | def join(self, f): |
|
412 | def join(self, f): | |
413 | return self.path + '/' + encodefilename(f) |
|
413 | return self.path + '/' + encodefilename(f) | |
414 |
|
414 | |||
415 | def copylist(self): |
|
415 | def copylist(self): | |
416 | return (['requires', '00changelog.i'] + |
|
416 | return (['requires', '00changelog.i'] + | |
417 | ['store/' + f for f in _data.split()]) |
|
417 | ['store/' + f for f in _data.split()]) | |
418 |
|
418 | |||
419 | class fncache(object): |
|
419 | class fncache(object): | |
420 | # the filename used to be partially encoded |
|
420 | # the filename used to be partially encoded | |
421 | # hence the encodedir/decodedir dance |
|
421 | # hence the encodedir/decodedir dance | |
422 | def __init__(self, vfs): |
|
422 | def __init__(self, vfs): | |
423 | self.vfs = vfs |
|
423 | self.vfs = vfs | |
424 | self.entries = None |
|
424 | self.entries = None | |
425 | self._dirty = False |
|
425 | self._dirty = False | |
426 |
|
426 | |||
427 | def _load(self): |
|
427 | def _load(self): | |
428 | '''fill the entries from the fncache file''' |
|
428 | '''fill the entries from the fncache file''' | |
429 | self._dirty = False |
|
429 | self._dirty = False | |
430 | try: |
|
430 | try: | |
431 | fp = self.vfs('fncache', mode='rb') |
|
431 | fp = self.vfs('fncache', mode='rb') | |
432 | except IOError: |
|
432 | except IOError: | |
433 | # skip nonexistent file |
|
433 | # skip nonexistent file | |
434 | self.entries = set() |
|
434 | self.entries = set() | |
435 | return |
|
435 | return | |
436 | self.entries = set(decodedir(fp.read()).splitlines()) |
|
436 | self.entries = set(decodedir(fp.read()).splitlines()) | |
437 | if '' in self.entries: |
|
437 | if '' in self.entries: | |
438 | fp.seek(0) |
|
438 | fp.seek(0) | |
439 | for n, line in enumerate(util.iterfile(fp)): |
|
439 | for n, line in enumerate(util.iterfile(fp)): | |
440 | if not line.rstrip('\n'): |
|
440 | if not line.rstrip('\n'): | |
441 | t = _('invalid entry in fncache, line %d') % (n + 1) |
|
441 | t = _('invalid entry in fncache, line %d') % (n + 1) | |
442 | raise error.Abort(t) |
|
442 | raise error.Abort(t) | |
443 | fp.close() |
|
443 | fp.close() | |
444 |
|
444 | |||
445 | def write(self, tr): |
|
445 | def write(self, tr): | |
446 | if self._dirty: |
|
446 | if self._dirty: | |
447 | tr.addbackup('fncache') |
|
447 | tr.addbackup('fncache') | |
448 | fp = self.vfs('fncache', mode='wb', atomictemp=True) |
|
448 | fp = self.vfs('fncache', mode='wb', atomictemp=True) | |
449 | if self.entries: |
|
449 | if self.entries: | |
450 | fp.write(encodedir('\n'.join(self.entries) + '\n')) |
|
450 | fp.write(encodedir('\n'.join(self.entries) + '\n')) | |
451 | fp.close() |
|
451 | fp.close() | |
452 | self._dirty = False |
|
452 | self._dirty = False | |
453 |
|
453 | |||
454 | def add(self, fn): |
|
454 | def add(self, fn): | |
455 | if self.entries is None: |
|
455 | if self.entries is None: | |
456 | self._load() |
|
456 | self._load() | |
457 | if fn not in self.entries: |
|
457 | if fn not in self.entries: | |
458 | self._dirty = True |
|
458 | self._dirty = True | |
459 | self.entries.add(fn) |
|
459 | self.entries.add(fn) | |
460 |
|
460 | |||
461 | def remove(self, fn): |
|
461 | def remove(self, fn): | |
462 | if self.entries is None: |
|
462 | if self.entries is None: | |
463 | self._load() |
|
463 | self._load() | |
464 | try: |
|
464 | try: | |
465 | self.entries.remove(fn) |
|
465 | self.entries.remove(fn) | |
466 | self._dirty = True |
|
466 | self._dirty = True | |
467 | except KeyError: |
|
467 | except KeyError: | |
468 | pass |
|
468 | pass | |
469 |
|
469 | |||
470 | def __contains__(self, fn): |
|
470 | def __contains__(self, fn): | |
471 | if self.entries is None: |
|
471 | if self.entries is None: | |
472 | self._load() |
|
472 | self._load() | |
473 | return fn in self.entries |
|
473 | return fn in self.entries | |
474 |
|
474 | |||
475 | def __iter__(self): |
|
475 | def __iter__(self): | |
476 | if self.entries is None: |
|
476 | if self.entries is None: | |
477 | self._load() |
|
477 | self._load() | |
478 | return iter(self.entries) |
|
478 | return iter(self.entries) | |
479 |
|
479 | |||
480 |
class _fncachevfs( |
|
480 | class _fncachevfs(vfsmod.abstractvfs, vfsmod.auditvfs): | |
481 | def __init__(self, vfs, fnc, encode): |
|
481 | def __init__(self, vfs, fnc, encode): | |
482 |
|
|
482 | vfsmod.auditvfs.__init__(self, vfs) | |
483 | self.fncache = fnc |
|
483 | self.fncache = fnc | |
484 | self.encode = encode |
|
484 | self.encode = encode | |
485 |
|
485 | |||
486 | def __call__(self, path, mode='r', *args, **kw): |
|
486 | def __call__(self, path, mode='r', *args, **kw): | |
487 | if mode not in ('r', 'rb') and (path.startswith('data/') or |
|
487 | if mode not in ('r', 'rb') and (path.startswith('data/') or | |
488 | path.startswith('meta/')): |
|
488 | path.startswith('meta/')): | |
489 | self.fncache.add(path) |
|
489 | self.fncache.add(path) | |
490 | return self.vfs(self.encode(path), mode, *args, **kw) |
|
490 | return self.vfs(self.encode(path), mode, *args, **kw) | |
491 |
|
491 | |||
492 | def join(self, path): |
|
492 | def join(self, path): | |
493 | if path: |
|
493 | if path: | |
494 | return self.vfs.join(self.encode(path)) |
|
494 | return self.vfs.join(self.encode(path)) | |
495 | else: |
|
495 | else: | |
496 | return self.vfs.join(path) |
|
496 | return self.vfs.join(path) | |
497 |
|
497 | |||
498 | class fncachestore(basicstore): |
|
498 | class fncachestore(basicstore): | |
499 | def __init__(self, path, vfstype, dotencode): |
|
499 | def __init__(self, path, vfstype, dotencode): | |
500 | if dotencode: |
|
500 | if dotencode: | |
501 | encode = _pathencode |
|
501 | encode = _pathencode | |
502 | else: |
|
502 | else: | |
503 | encode = _plainhybridencode |
|
503 | encode = _plainhybridencode | |
504 | self.encode = encode |
|
504 | self.encode = encode | |
505 | vfs = vfstype(path + '/store') |
|
505 | vfs = vfstype(path + '/store') | |
506 | self.path = vfs.base |
|
506 | self.path = vfs.base | |
507 | self.pathsep = self.path + '/' |
|
507 | self.pathsep = self.path + '/' | |
508 | self.createmode = _calcmode(vfs) |
|
508 | self.createmode = _calcmode(vfs) | |
509 | vfs.createmode = self.createmode |
|
509 | vfs.createmode = self.createmode | |
510 | self.rawvfs = vfs |
|
510 | self.rawvfs = vfs | |
511 | fnc = fncache(vfs) |
|
511 | fnc = fncache(vfs) | |
512 | self.fncache = fnc |
|
512 | self.fncache = fnc | |
513 | self.vfs = _fncachevfs(vfs, fnc, encode) |
|
513 | self.vfs = _fncachevfs(vfs, fnc, encode) | |
514 | self.opener = self.vfs |
|
514 | self.opener = self.vfs | |
515 |
|
515 | |||
516 | def join(self, f): |
|
516 | def join(self, f): | |
517 | return self.pathsep + self.encode(f) |
|
517 | return self.pathsep + self.encode(f) | |
518 |
|
518 | |||
519 | def getsize(self, path): |
|
519 | def getsize(self, path): | |
520 | return self.rawvfs.stat(path).st_size |
|
520 | return self.rawvfs.stat(path).st_size | |
521 |
|
521 | |||
522 | def datafiles(self): |
|
522 | def datafiles(self): | |
523 | for f in sorted(self.fncache): |
|
523 | for f in sorted(self.fncache): | |
524 | ef = self.encode(f) |
|
524 | ef = self.encode(f) | |
525 | try: |
|
525 | try: | |
526 | yield f, ef, self.getsize(ef) |
|
526 | yield f, ef, self.getsize(ef) | |
527 | except OSError as err: |
|
527 | except OSError as err: | |
528 | if err.errno != errno.ENOENT: |
|
528 | if err.errno != errno.ENOENT: | |
529 | raise |
|
529 | raise | |
530 |
|
530 | |||
531 | def copylist(self): |
|
531 | def copylist(self): | |
532 | d = ('data meta dh fncache phaseroots obsstore' |
|
532 | d = ('data meta dh fncache phaseroots obsstore' | |
533 | ' 00manifest.d 00manifest.i 00changelog.d 00changelog.i') |
|
533 | ' 00manifest.d 00manifest.i 00changelog.d 00changelog.i') | |
534 | return (['requires', '00changelog.i'] + |
|
534 | return (['requires', '00changelog.i'] + | |
535 | ['store/' + f for f in d.split()]) |
|
535 | ['store/' + f for f in d.split()]) | |
536 |
|
536 | |||
537 | def write(self, tr): |
|
537 | def write(self, tr): | |
538 | self.fncache.write(tr) |
|
538 | self.fncache.write(tr) | |
539 |
|
539 | |||
540 | def invalidatecaches(self): |
|
540 | def invalidatecaches(self): | |
541 | self.fncache.entries = None |
|
541 | self.fncache.entries = None | |
542 |
|
542 | |||
543 | def markremoved(self, fn): |
|
543 | def markremoved(self, fn): | |
544 | self.fncache.remove(fn) |
|
544 | self.fncache.remove(fn) | |
545 |
|
545 | |||
546 | def _exists(self, f): |
|
546 | def _exists(self, f): | |
547 | ef = self.encode(f) |
|
547 | ef = self.encode(f) | |
548 | try: |
|
548 | try: | |
549 | self.getsize(ef) |
|
549 | self.getsize(ef) | |
550 | return True |
|
550 | return True | |
551 | except OSError as err: |
|
551 | except OSError as err: | |
552 | if err.errno != errno.ENOENT: |
|
552 | if err.errno != errno.ENOENT: | |
553 | raise |
|
553 | raise | |
554 | # nonexistent entry |
|
554 | # nonexistent entry | |
555 | return False |
|
555 | return False | |
556 |
|
556 | |||
557 | def __contains__(self, path): |
|
557 | def __contains__(self, path): | |
558 | '''Checks if the store contains path''' |
|
558 | '''Checks if the store contains path''' | |
559 | path = "/".join(("data", path)) |
|
559 | path = "/".join(("data", path)) | |
560 | # check for files (exact match) |
|
560 | # check for files (exact match) | |
561 | e = path + '.i' |
|
561 | e = path + '.i' | |
562 | if e in self.fncache and self._exists(e): |
|
562 | if e in self.fncache and self._exists(e): | |
563 | return True |
|
563 | return True | |
564 | # now check for directories (prefix match) |
|
564 | # now check for directories (prefix match) | |
565 | if not path.endswith('/'): |
|
565 | if not path.endswith('/'): | |
566 | path += '/' |
|
566 | path += '/' | |
567 | for e in self.fncache: |
|
567 | for e in self.fncache: | |
568 | if e.startswith(path) and self._exists(e): |
|
568 | if e.startswith(path) and self._exists(e): | |
569 | return True |
|
569 | return True | |
570 | return False |
|
570 | return False | |
571 |
|
571 | |||
572 | def store(requirements, path, vfstype): |
|
572 | def store(requirements, path, vfstype): | |
573 | if 'store' in requirements: |
|
573 | if 'store' in requirements: | |
574 | if 'fncache' in requirements: |
|
574 | if 'fncache' in requirements: | |
575 | return fncachestore(path, vfstype, 'dotencode' in requirements) |
|
575 | return fncachestore(path, vfstype, 'dotencode' in requirements) | |
576 | return encodedstore(path, vfstype) |
|
576 | return encodedstore(path, vfstype) | |
577 | return basicstore(path, vfstype) |
|
577 | return basicstore(path, vfstype) |
General Comments 0
You need to be logged in to leave comments.
Login now