Show More
@@ -1,605 +1,585 b'' | |||||
1 | # changelog.py - changelog class for mercurial |
|
1 | # changelog.py - changelog class for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2005-2007 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 | from .i18n import _ |
|
10 | from .i18n import _ | |
11 | from .node import ( |
|
11 | from .node import ( | |
12 | bin, |
|
12 | bin, | |
13 | hex, |
|
13 | hex, | |
14 | nullid, |
|
14 | nullid, | |
15 | ) |
|
15 | ) | |
16 | from .thirdparty import attr |
|
16 | from .thirdparty import attr | |
17 |
|
17 | |||
18 | from . import ( |
|
18 | from . import ( | |
19 | encoding, |
|
19 | encoding, | |
20 | error, |
|
20 | error, | |
21 | metadata, |
|
21 | metadata, | |
22 | pycompat, |
|
22 | pycompat, | |
23 | revlog, |
|
23 | revlog, | |
24 | ) |
|
24 | ) | |
25 | from .utils import ( |
|
25 | from .utils import ( | |
26 | dateutil, |
|
26 | dateutil, | |
27 | stringutil, |
|
27 | stringutil, | |
28 | ) |
|
28 | ) | |
29 |
|
29 | |||
30 | from .revlogutils import sidedata as sidedatamod |
|
30 | from .revlogutils import sidedata as sidedatamod | |
31 |
|
31 | |||
32 | _defaultextra = {b'branch': b'default'} |
|
32 | _defaultextra = {b'branch': b'default'} | |
33 |
|
33 | |||
34 |
|
34 | |||
35 | def _string_escape(text): |
|
35 | def _string_escape(text): | |
36 | """ |
|
36 | """ | |
37 | >>> from .pycompat import bytechr as chr |
|
37 | >>> from .pycompat import bytechr as chr | |
38 | >>> d = {b'nl': chr(10), b'bs': chr(92), b'cr': chr(13), b'nul': chr(0)} |
|
38 | >>> d = {b'nl': chr(10), b'bs': chr(92), b'cr': chr(13), b'nul': chr(0)} | |
39 | >>> s = b"ab%(nl)scd%(bs)s%(bs)sn%(nul)s12ab%(cr)scd%(bs)s%(nl)s" % d |
|
39 | >>> s = b"ab%(nl)scd%(bs)s%(bs)sn%(nul)s12ab%(cr)scd%(bs)s%(nl)s" % d | |
40 | >>> s |
|
40 | >>> s | |
41 | 'ab\\ncd\\\\\\\\n\\x0012ab\\rcd\\\\\\n' |
|
41 | 'ab\\ncd\\\\\\\\n\\x0012ab\\rcd\\\\\\n' | |
42 | >>> res = _string_escape(s) |
|
42 | >>> res = _string_escape(s) | |
43 | >>> s == _string_unescape(res) |
|
43 | >>> s == _string_unescape(res) | |
44 | True |
|
44 | True | |
45 | """ |
|
45 | """ | |
46 | # subset of the string_escape codec |
|
46 | # subset of the string_escape codec | |
47 | text = ( |
|
47 | text = ( | |
48 | text.replace(b'\\', b'\\\\') |
|
48 | text.replace(b'\\', b'\\\\') | |
49 | .replace(b'\n', b'\\n') |
|
49 | .replace(b'\n', b'\\n') | |
50 | .replace(b'\r', b'\\r') |
|
50 | .replace(b'\r', b'\\r') | |
51 | ) |
|
51 | ) | |
52 | return text.replace(b'\0', b'\\0') |
|
52 | return text.replace(b'\0', b'\\0') | |
53 |
|
53 | |||
54 |
|
54 | |||
55 | def _string_unescape(text): |
|
55 | def _string_unescape(text): | |
56 | if b'\\0' in text: |
|
56 | if b'\\0' in text: | |
57 | # fix up \0 without getting into trouble with \\0 |
|
57 | # fix up \0 without getting into trouble with \\0 | |
58 | text = text.replace(b'\\\\', b'\\\\\n') |
|
58 | text = text.replace(b'\\\\', b'\\\\\n') | |
59 | text = text.replace(b'\\0', b'\0') |
|
59 | text = text.replace(b'\\0', b'\0') | |
60 | text = text.replace(b'\n', b'') |
|
60 | text = text.replace(b'\n', b'') | |
61 | return stringutil.unescapestr(text) |
|
61 | return stringutil.unescapestr(text) | |
62 |
|
62 | |||
63 |
|
63 | |||
64 | def decodeextra(text): |
|
64 | def decodeextra(text): | |
65 | """ |
|
65 | """ | |
66 | >>> from .pycompat import bytechr as chr |
|
66 | >>> from .pycompat import bytechr as chr | |
67 | >>> sorted(decodeextra(encodeextra({b'foo': b'bar', b'baz': chr(0) + b'2'}) |
|
67 | >>> sorted(decodeextra(encodeextra({b'foo': b'bar', b'baz': chr(0) + b'2'}) | |
68 | ... ).items()) |
|
68 | ... ).items()) | |
69 | [('baz', '\\x002'), ('branch', 'default'), ('foo', 'bar')] |
|
69 | [('baz', '\\x002'), ('branch', 'default'), ('foo', 'bar')] | |
70 | >>> sorted(decodeextra(encodeextra({b'foo': b'bar', |
|
70 | >>> sorted(decodeextra(encodeextra({b'foo': b'bar', | |
71 | ... b'baz': chr(92) + chr(0) + b'2'}) |
|
71 | ... b'baz': chr(92) + chr(0) + b'2'}) | |
72 | ... ).items()) |
|
72 | ... ).items()) | |
73 | [('baz', '\\\\\\x002'), ('branch', 'default'), ('foo', 'bar')] |
|
73 | [('baz', '\\\\\\x002'), ('branch', 'default'), ('foo', 'bar')] | |
74 | """ |
|
74 | """ | |
75 | extra = _defaultextra.copy() |
|
75 | extra = _defaultextra.copy() | |
76 | for l in text.split(b'\0'): |
|
76 | for l in text.split(b'\0'): | |
77 | if l: |
|
77 | if l: | |
78 | k, v = _string_unescape(l).split(b':', 1) |
|
78 | k, v = _string_unescape(l).split(b':', 1) | |
79 | extra[k] = v |
|
79 | extra[k] = v | |
80 | return extra |
|
80 | return extra | |
81 |
|
81 | |||
82 |
|
82 | |||
83 | def encodeextra(d): |
|
83 | def encodeextra(d): | |
84 | # keys must be sorted to produce a deterministic changelog entry |
|
84 | # keys must be sorted to produce a deterministic changelog entry | |
85 | items = [_string_escape(b'%s:%s' % (k, d[k])) for k in sorted(d)] |
|
85 | items = [_string_escape(b'%s:%s' % (k, d[k])) for k in sorted(d)] | |
86 | return b"\0".join(items) |
|
86 | return b"\0".join(items) | |
87 |
|
87 | |||
88 |
|
88 | |||
89 | def stripdesc(desc): |
|
89 | def stripdesc(desc): | |
90 | """strip trailing whitespace and leading and trailing empty lines""" |
|
90 | """strip trailing whitespace and leading and trailing empty lines""" | |
91 | return b'\n'.join([l.rstrip() for l in desc.splitlines()]).strip(b'\n') |
|
91 | return b'\n'.join([l.rstrip() for l in desc.splitlines()]).strip(b'\n') | |
92 |
|
92 | |||
93 |
|
93 | |||
94 | class appender(object): |
|
94 | class appender(object): | |
95 | '''the changelog index must be updated last on disk, so we use this class |
|
95 | '''the changelog index must be updated last on disk, so we use this class | |
96 | to delay writes to it''' |
|
96 | to delay writes to it''' | |
97 |
|
97 | |||
98 | def __init__(self, vfs, name, mode, buf): |
|
98 | def __init__(self, vfs, name, mode, buf): | |
99 | self.data = buf |
|
99 | self.data = buf | |
100 | fp = vfs(name, mode) |
|
100 | fp = vfs(name, mode) | |
101 | self.fp = fp |
|
101 | self.fp = fp | |
102 | self.offset = fp.tell() |
|
102 | self.offset = fp.tell() | |
103 | self.size = vfs.fstat(fp).st_size |
|
103 | self.size = vfs.fstat(fp).st_size | |
104 | self._end = self.size |
|
104 | self._end = self.size | |
105 |
|
105 | |||
106 | def end(self): |
|
106 | def end(self): | |
107 | return self._end |
|
107 | return self._end | |
108 |
|
108 | |||
109 | def tell(self): |
|
109 | def tell(self): | |
110 | return self.offset |
|
110 | return self.offset | |
111 |
|
111 | |||
112 | def flush(self): |
|
112 | def flush(self): | |
113 | pass |
|
113 | pass | |
114 |
|
114 | |||
115 | @property |
|
115 | @property | |
116 | def closed(self): |
|
116 | def closed(self): | |
117 | return self.fp.closed |
|
117 | return self.fp.closed | |
118 |
|
118 | |||
119 | def close(self): |
|
119 | def close(self): | |
120 | self.fp.close() |
|
120 | self.fp.close() | |
121 |
|
121 | |||
122 | def seek(self, offset, whence=0): |
|
122 | def seek(self, offset, whence=0): | |
123 | '''virtual file offset spans real file and data''' |
|
123 | '''virtual file offset spans real file and data''' | |
124 | if whence == 0: |
|
124 | if whence == 0: | |
125 | self.offset = offset |
|
125 | self.offset = offset | |
126 | elif whence == 1: |
|
126 | elif whence == 1: | |
127 | self.offset += offset |
|
127 | self.offset += offset | |
128 | elif whence == 2: |
|
128 | elif whence == 2: | |
129 | self.offset = self.end() + offset |
|
129 | self.offset = self.end() + offset | |
130 | if self.offset < self.size: |
|
130 | if self.offset < self.size: | |
131 | self.fp.seek(self.offset) |
|
131 | self.fp.seek(self.offset) | |
132 |
|
132 | |||
133 | def read(self, count=-1): |
|
133 | def read(self, count=-1): | |
134 | '''only trick here is reads that span real file and data''' |
|
134 | '''only trick here is reads that span real file and data''' | |
135 | ret = b"" |
|
135 | ret = b"" | |
136 | if self.offset < self.size: |
|
136 | if self.offset < self.size: | |
137 | s = self.fp.read(count) |
|
137 | s = self.fp.read(count) | |
138 | ret = s |
|
138 | ret = s | |
139 | self.offset += len(s) |
|
139 | self.offset += len(s) | |
140 | if count > 0: |
|
140 | if count > 0: | |
141 | count -= len(s) |
|
141 | count -= len(s) | |
142 | if count != 0: |
|
142 | if count != 0: | |
143 | doff = self.offset - self.size |
|
143 | doff = self.offset - self.size | |
144 | self.data.insert(0, b"".join(self.data)) |
|
144 | self.data.insert(0, b"".join(self.data)) | |
145 | del self.data[1:] |
|
145 | del self.data[1:] | |
146 | s = self.data[0][doff : doff + count] |
|
146 | s = self.data[0][doff : doff + count] | |
147 | self.offset += len(s) |
|
147 | self.offset += len(s) | |
148 | ret += s |
|
148 | ret += s | |
149 | return ret |
|
149 | return ret | |
150 |
|
150 | |||
151 | def write(self, s): |
|
151 | def write(self, s): | |
152 | self.data.append(bytes(s)) |
|
152 | self.data.append(bytes(s)) | |
153 | self.offset += len(s) |
|
153 | self.offset += len(s) | |
154 | self._end += len(s) |
|
154 | self._end += len(s) | |
155 |
|
155 | |||
156 | def __enter__(self): |
|
156 | def __enter__(self): | |
157 | self.fp.__enter__() |
|
157 | self.fp.__enter__() | |
158 | return self |
|
158 | return self | |
159 |
|
159 | |||
160 | def __exit__(self, *args): |
|
160 | def __exit__(self, *args): | |
161 | return self.fp.__exit__(*args) |
|
161 | return self.fp.__exit__(*args) | |
162 |
|
162 | |||
163 |
|
163 | |||
164 | class _divertopener(object): |
|
164 | class _divertopener(object): | |
165 | def __init__(self, opener, target): |
|
165 | def __init__(self, opener, target): | |
166 | self._opener = opener |
|
166 | self._opener = opener | |
167 | self._target = target |
|
167 | self._target = target | |
168 |
|
168 | |||
169 | def __call__(self, name, mode=b'r', checkambig=False, **kwargs): |
|
169 | def __call__(self, name, mode=b'r', checkambig=False, **kwargs): | |
170 | if name != self._target: |
|
170 | if name != self._target: | |
171 | return self._opener(name, mode, **kwargs) |
|
171 | return self._opener(name, mode, **kwargs) | |
172 | return self._opener(name + b".a", mode, **kwargs) |
|
172 | return self._opener(name + b".a", mode, **kwargs) | |
173 |
|
173 | |||
174 | def __getattr__(self, attr): |
|
174 | def __getattr__(self, attr): | |
175 | return getattr(self._opener, attr) |
|
175 | return getattr(self._opener, attr) | |
176 |
|
176 | |||
177 |
|
177 | |||
178 | def _delayopener(opener, target, buf): |
|
178 | def _delayopener(opener, target, buf): | |
179 | """build an opener that stores chunks in 'buf' instead of 'target'""" |
|
179 | """build an opener that stores chunks in 'buf' instead of 'target'""" | |
180 |
|
180 | |||
181 | def _delay(name, mode=b'r', checkambig=False, **kwargs): |
|
181 | def _delay(name, mode=b'r', checkambig=False, **kwargs): | |
182 | if name != target: |
|
182 | if name != target: | |
183 | return opener(name, mode, **kwargs) |
|
183 | return opener(name, mode, **kwargs) | |
184 | assert not kwargs |
|
184 | assert not kwargs | |
185 | return appender(opener, name, mode, buf) |
|
185 | return appender(opener, name, mode, buf) | |
186 |
|
186 | |||
187 | return _delay |
|
187 | return _delay | |
188 |
|
188 | |||
189 |
|
189 | |||
190 | @attr.s |
|
190 | @attr.s | |
191 | class _changelogrevision(object): |
|
191 | class _changelogrevision(object): | |
192 | # Extensions might modify _defaultextra, so let the constructor below pass |
|
192 | # Extensions might modify _defaultextra, so let the constructor below pass | |
193 | # it in |
|
193 | # it in | |
194 | extra = attr.ib() |
|
194 | extra = attr.ib() | |
195 | manifest = attr.ib(default=nullid) |
|
195 | manifest = attr.ib(default=nullid) | |
196 | user = attr.ib(default=b'') |
|
196 | user = attr.ib(default=b'') | |
197 | date = attr.ib(default=(0, 0)) |
|
197 | date = attr.ib(default=(0, 0)) | |
198 | files = attr.ib(default=attr.Factory(list)) |
|
198 | files = attr.ib(default=attr.Factory(list)) | |
199 | filesadded = attr.ib(default=None) |
|
199 | filesadded = attr.ib(default=None) | |
200 | filesremoved = attr.ib(default=None) |
|
200 | filesremoved = attr.ib(default=None) | |
201 | p1copies = attr.ib(default=None) |
|
201 | p1copies = attr.ib(default=None) | |
202 | p2copies = attr.ib(default=None) |
|
202 | p2copies = attr.ib(default=None) | |
203 | description = attr.ib(default=b'') |
|
203 | description = attr.ib(default=b'') | |
204 |
|
204 | |||
205 |
|
205 | |||
206 | class changelogrevision(object): |
|
206 | class changelogrevision(object): | |
207 | """Holds results of a parsed changelog revision. |
|
207 | """Holds results of a parsed changelog revision. | |
208 |
|
208 | |||
209 | Changelog revisions consist of multiple pieces of data, including |
|
209 | Changelog revisions consist of multiple pieces of data, including | |
210 | the manifest node, user, and date. This object exposes a view into |
|
210 | the manifest node, user, and date. This object exposes a view into | |
211 | the parsed object. |
|
211 | the parsed object. | |
212 | """ |
|
212 | """ | |
213 |
|
213 | |||
214 | __slots__ = ( |
|
214 | __slots__ = ( | |
215 | '_offsets', |
|
215 | '_offsets', | |
216 | '_text', |
|
216 | '_text', | |
217 | '_sidedata', |
|
217 | '_sidedata', | |
218 | '_cpsd', |
|
218 | '_cpsd', | |
219 | ) |
|
219 | ) | |
220 |
|
220 | |||
221 | def __new__(cls, text, sidedata, cpsd): |
|
221 | def __new__(cls, text, sidedata, cpsd): | |
222 | if not text: |
|
222 | if not text: | |
223 | return _changelogrevision(extra=_defaultextra) |
|
223 | return _changelogrevision(extra=_defaultextra) | |
224 |
|
224 | |||
225 | self = super(changelogrevision, cls).__new__(cls) |
|
225 | self = super(changelogrevision, cls).__new__(cls) | |
226 | # We could return here and implement the following as an __init__. |
|
226 | # We could return here and implement the following as an __init__. | |
227 | # But doing it here is equivalent and saves an extra function call. |
|
227 | # But doing it here is equivalent and saves an extra function call. | |
228 |
|
228 | |||
229 | # format used: |
|
229 | # format used: | |
230 | # nodeid\n : manifest node in ascii |
|
230 | # nodeid\n : manifest node in ascii | |
231 | # user\n : user, no \n or \r allowed |
|
231 | # user\n : user, no \n or \r allowed | |
232 | # time tz extra\n : date (time is int or float, timezone is int) |
|
232 | # time tz extra\n : date (time is int or float, timezone is int) | |
233 | # : extra is metadata, encoded and separated by '\0' |
|
233 | # : extra is metadata, encoded and separated by '\0' | |
234 | # : older versions ignore it |
|
234 | # : older versions ignore it | |
235 | # files\n\n : files modified by the cset, no \n or \r allowed |
|
235 | # files\n\n : files modified by the cset, no \n or \r allowed | |
236 | # (.*) : comment (free text, ideally utf-8) |
|
236 | # (.*) : comment (free text, ideally utf-8) | |
237 | # |
|
237 | # | |
238 | # changelog v0 doesn't use extra |
|
238 | # changelog v0 doesn't use extra | |
239 |
|
239 | |||
240 | nl1 = text.index(b'\n') |
|
240 | nl1 = text.index(b'\n') | |
241 | nl2 = text.index(b'\n', nl1 + 1) |
|
241 | nl2 = text.index(b'\n', nl1 + 1) | |
242 | nl3 = text.index(b'\n', nl2 + 1) |
|
242 | nl3 = text.index(b'\n', nl2 + 1) | |
243 |
|
243 | |||
244 | # The list of files may be empty. Which means nl3 is the first of the |
|
244 | # The list of files may be empty. Which means nl3 is the first of the | |
245 | # double newline that precedes the description. |
|
245 | # double newline that precedes the description. | |
246 | if text[nl3 + 1 : nl3 + 2] == b'\n': |
|
246 | if text[nl3 + 1 : nl3 + 2] == b'\n': | |
247 | doublenl = nl3 |
|
247 | doublenl = nl3 | |
248 | else: |
|
248 | else: | |
249 | doublenl = text.index(b'\n\n', nl3 + 1) |
|
249 | doublenl = text.index(b'\n\n', nl3 + 1) | |
250 |
|
250 | |||
251 | self._offsets = (nl1, nl2, nl3, doublenl) |
|
251 | self._offsets = (nl1, nl2, nl3, doublenl) | |
252 | self._text = text |
|
252 | self._text = text | |
253 | self._sidedata = sidedata |
|
253 | self._sidedata = sidedata | |
254 | self._cpsd = cpsd |
|
254 | self._cpsd = cpsd | |
255 |
|
255 | |||
256 | return self |
|
256 | return self | |
257 |
|
257 | |||
258 | @property |
|
258 | @property | |
259 | def manifest(self): |
|
259 | def manifest(self): | |
260 | return bin(self._text[0 : self._offsets[0]]) |
|
260 | return bin(self._text[0 : self._offsets[0]]) | |
261 |
|
261 | |||
262 | @property |
|
262 | @property | |
263 | def user(self): |
|
263 | def user(self): | |
264 | off = self._offsets |
|
264 | off = self._offsets | |
265 | return encoding.tolocal(self._text[off[0] + 1 : off[1]]) |
|
265 | return encoding.tolocal(self._text[off[0] + 1 : off[1]]) | |
266 |
|
266 | |||
267 | @property |
|
267 | @property | |
268 | def _rawdate(self): |
|
268 | def _rawdate(self): | |
269 | off = self._offsets |
|
269 | off = self._offsets | |
270 | dateextra = self._text[off[1] + 1 : off[2]] |
|
270 | dateextra = self._text[off[1] + 1 : off[2]] | |
271 | return dateextra.split(b' ', 2)[0:2] |
|
271 | return dateextra.split(b' ', 2)[0:2] | |
272 |
|
272 | |||
273 | @property |
|
273 | @property | |
274 | def _rawextra(self): |
|
274 | def _rawextra(self): | |
275 | off = self._offsets |
|
275 | off = self._offsets | |
276 | dateextra = self._text[off[1] + 1 : off[2]] |
|
276 | dateextra = self._text[off[1] + 1 : off[2]] | |
277 | fields = dateextra.split(b' ', 2) |
|
277 | fields = dateextra.split(b' ', 2) | |
278 | if len(fields) != 3: |
|
278 | if len(fields) != 3: | |
279 | return None |
|
279 | return None | |
280 |
|
280 | |||
281 | return fields[2] |
|
281 | return fields[2] | |
282 |
|
282 | |||
283 | @property |
|
283 | @property | |
284 | def date(self): |
|
284 | def date(self): | |
285 | raw = self._rawdate |
|
285 | raw = self._rawdate | |
286 | time = float(raw[0]) |
|
286 | time = float(raw[0]) | |
287 | # Various tools did silly things with the timezone. |
|
287 | # Various tools did silly things with the timezone. | |
288 | try: |
|
288 | try: | |
289 | timezone = int(raw[1]) |
|
289 | timezone = int(raw[1]) | |
290 | except ValueError: |
|
290 | except ValueError: | |
291 | timezone = 0 |
|
291 | timezone = 0 | |
292 |
|
292 | |||
293 | return time, timezone |
|
293 | return time, timezone | |
294 |
|
294 | |||
295 | @property |
|
295 | @property | |
296 | def extra(self): |
|
296 | def extra(self): | |
297 | raw = self._rawextra |
|
297 | raw = self._rawextra | |
298 | if raw is None: |
|
298 | if raw is None: | |
299 | return _defaultextra |
|
299 | return _defaultextra | |
300 |
|
300 | |||
301 | return decodeextra(raw) |
|
301 | return decodeextra(raw) | |
302 |
|
302 | |||
303 | @property |
|
303 | @property | |
304 | def files(self): |
|
304 | def files(self): | |
305 | off = self._offsets |
|
305 | off = self._offsets | |
306 | if off[2] == off[3]: |
|
306 | if off[2] == off[3]: | |
307 | return [] |
|
307 | return [] | |
308 |
|
308 | |||
309 | return self._text[off[2] + 1 : off[3]].split(b'\n') |
|
309 | return self._text[off[2] + 1 : off[3]].split(b'\n') | |
310 |
|
310 | |||
311 | @property |
|
311 | @property | |
312 | def filesadded(self): |
|
312 | def filesadded(self): | |
313 | if self._cpsd: |
|
313 | if self._cpsd: | |
314 | rawindices = self._sidedata.get(sidedatamod.SD_FILESADDED) |
|
314 | rawindices = self._sidedata.get(sidedatamod.SD_FILESADDED) | |
315 | if not rawindices: |
|
315 | if not rawindices: | |
316 | return [] |
|
316 | return [] | |
317 | else: |
|
317 | else: | |
318 | rawindices = self.extra.get(b'filesadded') |
|
318 | rawindices = self.extra.get(b'filesadded') | |
319 | if rawindices is None: |
|
319 | if rawindices is None: | |
320 | return None |
|
320 | return None | |
321 | return metadata.decodefileindices(self.files, rawindices) |
|
321 | return metadata.decodefileindices(self.files, rawindices) | |
322 |
|
322 | |||
323 | @property |
|
323 | @property | |
324 | def filesremoved(self): |
|
324 | def filesremoved(self): | |
325 | if self._cpsd: |
|
325 | if self._cpsd: | |
326 | rawindices = self._sidedata.get(sidedatamod.SD_FILESREMOVED) |
|
326 | rawindices = self._sidedata.get(sidedatamod.SD_FILESREMOVED) | |
327 | if not rawindices: |
|
327 | if not rawindices: | |
328 | return [] |
|
328 | return [] | |
329 | else: |
|
329 | else: | |
330 | rawindices = self.extra.get(b'filesremoved') |
|
330 | rawindices = self.extra.get(b'filesremoved') | |
331 | if rawindices is None: |
|
331 | if rawindices is None: | |
332 | return None |
|
332 | return None | |
333 | return metadata.decodefileindices(self.files, rawindices) |
|
333 | return metadata.decodefileindices(self.files, rawindices) | |
334 |
|
334 | |||
335 | @property |
|
335 | @property | |
336 | def p1copies(self): |
|
336 | def p1copies(self): | |
337 | if self._cpsd: |
|
337 | if self._cpsd: | |
338 | rawcopies = self._sidedata.get(sidedatamod.SD_P1COPIES) |
|
338 | rawcopies = self._sidedata.get(sidedatamod.SD_P1COPIES) | |
339 | if not rawcopies: |
|
339 | if not rawcopies: | |
340 | return {} |
|
340 | return {} | |
341 | else: |
|
341 | else: | |
342 | rawcopies = self.extra.get(b'p1copies') |
|
342 | rawcopies = self.extra.get(b'p1copies') | |
343 | if rawcopies is None: |
|
343 | if rawcopies is None: | |
344 | return None |
|
344 | return None | |
345 | return metadata.decodecopies(self.files, rawcopies) |
|
345 | return metadata.decodecopies(self.files, rawcopies) | |
346 |
|
346 | |||
347 | @property |
|
347 | @property | |
348 | def p2copies(self): |
|
348 | def p2copies(self): | |
349 | if self._cpsd: |
|
349 | if self._cpsd: | |
350 | rawcopies = self._sidedata.get(sidedatamod.SD_P2COPIES) |
|
350 | rawcopies = self._sidedata.get(sidedatamod.SD_P2COPIES) | |
351 | if not rawcopies: |
|
351 | if not rawcopies: | |
352 | return {} |
|
352 | return {} | |
353 | else: |
|
353 | else: | |
354 | rawcopies = self.extra.get(b'p2copies') |
|
354 | rawcopies = self.extra.get(b'p2copies') | |
355 | if rawcopies is None: |
|
355 | if rawcopies is None: | |
356 | return None |
|
356 | return None | |
357 | return metadata.decodecopies(self.files, rawcopies) |
|
357 | return metadata.decodecopies(self.files, rawcopies) | |
358 |
|
358 | |||
359 | @property |
|
359 | @property | |
360 | def description(self): |
|
360 | def description(self): | |
361 | return encoding.tolocal(self._text[self._offsets[3] + 2 :]) |
|
361 | return encoding.tolocal(self._text[self._offsets[3] + 2 :]) | |
362 |
|
362 | |||
363 |
|
363 | |||
364 | class changelog(revlog.revlog): |
|
364 | class changelog(revlog.revlog): | |
365 | def __init__(self, opener, trypending=False): |
|
365 | def __init__(self, opener, trypending=False): | |
366 | """Load a changelog revlog using an opener. |
|
366 | """Load a changelog revlog using an opener. | |
367 |
|
367 | |||
368 | If ``trypending`` is true, we attempt to load the index from a |
|
368 | If ``trypending`` is true, we attempt to load the index from a | |
369 | ``00changelog.i.a`` file instead of the default ``00changelog.i``. |
|
369 | ``00changelog.i.a`` file instead of the default ``00changelog.i``. | |
370 | The ``00changelog.i.a`` file contains index (and possibly inline |
|
370 | The ``00changelog.i.a`` file contains index (and possibly inline | |
371 | revision) data for a transaction that hasn't been finalized yet. |
|
371 | revision) data for a transaction that hasn't been finalized yet. | |
372 | It exists in a separate file to facilitate readers (such as |
|
372 | It exists in a separate file to facilitate readers (such as | |
373 | hooks processes) accessing data before a transaction is finalized. |
|
373 | hooks processes) accessing data before a transaction is finalized. | |
374 | """ |
|
374 | """ | |
375 | if trypending and opener.exists(b'00changelog.i.a'): |
|
375 | if trypending and opener.exists(b'00changelog.i.a'): | |
376 | indexfile = b'00changelog.i.a' |
|
376 | indexfile = b'00changelog.i.a' | |
377 | else: |
|
377 | else: | |
378 | indexfile = b'00changelog.i' |
|
378 | indexfile = b'00changelog.i' | |
379 |
|
379 | |||
380 | datafile = b'00changelog.d' |
|
380 | datafile = b'00changelog.d' | |
381 | revlog.revlog.__init__( |
|
381 | revlog.revlog.__init__( | |
382 | self, |
|
382 | self, | |
383 | opener, |
|
383 | opener, | |
384 | indexfile, |
|
384 | indexfile, | |
385 | datafile=datafile, |
|
385 | datafile=datafile, | |
386 | checkambig=True, |
|
386 | checkambig=True, | |
387 | mmaplargeindex=True, |
|
387 | mmaplargeindex=True, | |
388 | persistentnodemap=opener.options.get(b'persistent-nodemap', False), |
|
388 | persistentnodemap=opener.options.get(b'persistent-nodemap', False), | |
389 | ) |
|
389 | ) | |
390 |
|
390 | |||
391 | if self._initempty and (self.version & 0xFFFF == revlog.REVLOGV1): |
|
391 | if self._initempty and (self.version & 0xFFFF == revlog.REVLOGV1): | |
392 | # changelogs don't benefit from generaldelta. |
|
392 | # changelogs don't benefit from generaldelta. | |
393 |
|
393 | |||
394 | self.version &= ~revlog.FLAG_GENERALDELTA |
|
394 | self.version &= ~revlog.FLAG_GENERALDELTA | |
395 | self._generaldelta = False |
|
395 | self._generaldelta = False | |
396 |
|
396 | |||
397 | # Delta chains for changelogs tend to be very small because entries |
|
397 | # Delta chains for changelogs tend to be very small because entries | |
398 | # tend to be small and don't delta well with each. So disable delta |
|
398 | # tend to be small and don't delta well with each. So disable delta | |
399 | # chains. |
|
399 | # chains. | |
400 | self._storedeltachains = False |
|
400 | self._storedeltachains = False | |
401 |
|
401 | |||
402 | self._realopener = opener |
|
402 | self._realopener = opener | |
403 | self._delayed = False |
|
403 | self._delayed = False | |
404 | self._delaybuf = None |
|
404 | self._delaybuf = None | |
405 | self._divert = False |
|
405 | self._divert = False | |
406 | self.filteredrevs = frozenset() |
|
406 | self.filteredrevs = frozenset() | |
407 | self._copiesstorage = opener.options.get(b'copies-storage') |
|
407 | self._copiesstorage = opener.options.get(b'copies-storage') | |
408 |
|
408 | |||
409 | def delayupdate(self, tr): |
|
409 | def delayupdate(self, tr): | |
410 | """delay visibility of index updates to other readers""" |
|
410 | """delay visibility of index updates to other readers""" | |
411 |
|
411 | |||
412 | if not self._delayed: |
|
412 | if not self._delayed: | |
413 | if len(self) == 0: |
|
413 | if len(self) == 0: | |
414 | self._divert = True |
|
414 | self._divert = True | |
415 | if self._realopener.exists(self.indexfile + b'.a'): |
|
415 | if self._realopener.exists(self.indexfile + b'.a'): | |
416 | self._realopener.unlink(self.indexfile + b'.a') |
|
416 | self._realopener.unlink(self.indexfile + b'.a') | |
417 | self.opener = _divertopener(self._realopener, self.indexfile) |
|
417 | self.opener = _divertopener(self._realopener, self.indexfile) | |
418 | else: |
|
418 | else: | |
419 | self._delaybuf = [] |
|
419 | self._delaybuf = [] | |
420 | self.opener = _delayopener( |
|
420 | self.opener = _delayopener( | |
421 | self._realopener, self.indexfile, self._delaybuf |
|
421 | self._realopener, self.indexfile, self._delaybuf | |
422 | ) |
|
422 | ) | |
423 | self._delayed = True |
|
423 | self._delayed = True | |
424 | tr.addpending(b'cl-%i' % id(self), self._writepending) |
|
424 | tr.addpending(b'cl-%i' % id(self), self._writepending) | |
425 | tr.addfinalize(b'cl-%i' % id(self), self._finalize) |
|
425 | tr.addfinalize(b'cl-%i' % id(self), self._finalize) | |
426 |
|
426 | |||
427 | def _finalize(self, tr): |
|
427 | def _finalize(self, tr): | |
428 | """finalize index updates""" |
|
428 | """finalize index updates""" | |
429 | self._delayed = False |
|
429 | self._delayed = False | |
430 | self.opener = self._realopener |
|
430 | self.opener = self._realopener | |
431 | # move redirected index data back into place |
|
431 | # move redirected index data back into place | |
432 | if self._divert: |
|
432 | if self._divert: | |
433 | assert not self._delaybuf |
|
433 | assert not self._delaybuf | |
434 | tmpname = self.indexfile + b".a" |
|
434 | tmpname = self.indexfile + b".a" | |
435 | nfile = self.opener.open(tmpname) |
|
435 | nfile = self.opener.open(tmpname) | |
436 | nfile.close() |
|
436 | nfile.close() | |
437 | self.opener.rename(tmpname, self.indexfile, checkambig=True) |
|
437 | self.opener.rename(tmpname, self.indexfile, checkambig=True) | |
438 | elif self._delaybuf: |
|
438 | elif self._delaybuf: | |
439 | fp = self.opener(self.indexfile, b'a', checkambig=True) |
|
439 | fp = self.opener(self.indexfile, b'a', checkambig=True) | |
440 | fp.write(b"".join(self._delaybuf)) |
|
440 | fp.write(b"".join(self._delaybuf)) | |
441 | fp.close() |
|
441 | fp.close() | |
442 | self._delaybuf = None |
|
442 | self._delaybuf = None | |
443 | self._divert = False |
|
443 | self._divert = False | |
444 | # split when we're done |
|
444 | # split when we're done | |
445 | self._enforceinlinesize(tr) |
|
445 | self._enforceinlinesize(tr) | |
446 |
|
446 | |||
447 | def _writepending(self, tr): |
|
447 | def _writepending(self, tr): | |
448 | """create a file containing the unfinalized state for |
|
448 | """create a file containing the unfinalized state for | |
449 | pretxnchangegroup""" |
|
449 | pretxnchangegroup""" | |
450 | if self._delaybuf: |
|
450 | if self._delaybuf: | |
451 | # make a temporary copy of the index |
|
451 | # make a temporary copy of the index | |
452 | fp1 = self._realopener(self.indexfile) |
|
452 | fp1 = self._realopener(self.indexfile) | |
453 | pendingfilename = self.indexfile + b".a" |
|
453 | pendingfilename = self.indexfile + b".a" | |
454 | # register as a temp file to ensure cleanup on failure |
|
454 | # register as a temp file to ensure cleanup on failure | |
455 | tr.registertmp(pendingfilename) |
|
455 | tr.registertmp(pendingfilename) | |
456 | # write existing data |
|
456 | # write existing data | |
457 | fp2 = self._realopener(pendingfilename, b"w") |
|
457 | fp2 = self._realopener(pendingfilename, b"w") | |
458 | fp2.write(fp1.read()) |
|
458 | fp2.write(fp1.read()) | |
459 | # add pending data |
|
459 | # add pending data | |
460 | fp2.write(b"".join(self._delaybuf)) |
|
460 | fp2.write(b"".join(self._delaybuf)) | |
461 | fp2.close() |
|
461 | fp2.close() | |
462 | # switch modes so finalize can simply rename |
|
462 | # switch modes so finalize can simply rename | |
463 | self._delaybuf = None |
|
463 | self._delaybuf = None | |
464 | self._divert = True |
|
464 | self._divert = True | |
465 | self.opener = _divertopener(self._realopener, self.indexfile) |
|
465 | self.opener = _divertopener(self._realopener, self.indexfile) | |
466 |
|
466 | |||
467 | if self._divert: |
|
467 | if self._divert: | |
468 | return True |
|
468 | return True | |
469 |
|
469 | |||
470 | return False |
|
470 | return False | |
471 |
|
471 | |||
472 | def _enforceinlinesize(self, tr, fp=None): |
|
472 | def _enforceinlinesize(self, tr, fp=None): | |
473 | if not self._delayed: |
|
473 | if not self._delayed: | |
474 | revlog.revlog._enforceinlinesize(self, tr, fp) |
|
474 | revlog.revlog._enforceinlinesize(self, tr, fp) | |
475 |
|
475 | |||
476 | def read(self, node): |
|
476 | def read(self, node): | |
477 | """Obtain data from a parsed changelog revision. |
|
477 | """Obtain data from a parsed changelog revision. | |
478 |
|
478 | |||
479 | Returns a 6-tuple of: |
|
479 | Returns a 6-tuple of: | |
480 |
|
480 | |||
481 | - manifest node in binary |
|
481 | - manifest node in binary | |
482 | - author/user as a localstr |
|
482 | - author/user as a localstr | |
483 | - date as a 2-tuple of (time, timezone) |
|
483 | - date as a 2-tuple of (time, timezone) | |
484 | - list of files |
|
484 | - list of files | |
485 | - commit message as a localstr |
|
485 | - commit message as a localstr | |
486 | - dict of extra metadata |
|
486 | - dict of extra metadata | |
487 |
|
487 | |||
488 | Unless you need to access all fields, consider calling |
|
488 | Unless you need to access all fields, consider calling | |
489 | ``changelogrevision`` instead, as it is faster for partial object |
|
489 | ``changelogrevision`` instead, as it is faster for partial object | |
490 | access. |
|
490 | access. | |
491 | """ |
|
491 | """ | |
492 | d, s = self._revisiondata(node) |
|
492 | d, s = self._revisiondata(node) | |
493 | c = changelogrevision( |
|
493 | c = changelogrevision( | |
494 | d, s, self._copiesstorage == b'changeset-sidedata' |
|
494 | d, s, self._copiesstorage == b'changeset-sidedata' | |
495 | ) |
|
495 | ) | |
496 | return (c.manifest, c.user, c.date, c.files, c.description, c.extra) |
|
496 | return (c.manifest, c.user, c.date, c.files, c.description, c.extra) | |
497 |
|
497 | |||
498 | def changelogrevision(self, nodeorrev): |
|
498 | def changelogrevision(self, nodeorrev): | |
499 | """Obtain a ``changelogrevision`` for a node or revision.""" |
|
499 | """Obtain a ``changelogrevision`` for a node or revision.""" | |
500 | text, sidedata = self._revisiondata(nodeorrev) |
|
500 | text, sidedata = self._revisiondata(nodeorrev) | |
501 | return changelogrevision( |
|
501 | return changelogrevision( | |
502 | text, sidedata, self._copiesstorage == b'changeset-sidedata' |
|
502 | text, sidedata, self._copiesstorage == b'changeset-sidedata' | |
503 | ) |
|
503 | ) | |
504 |
|
504 | |||
505 | def readfiles(self, node): |
|
505 | def readfiles(self, node): | |
506 | """ |
|
506 | """ | |
507 | short version of read that only returns the files modified by the cset |
|
507 | short version of read that only returns the files modified by the cset | |
508 | """ |
|
508 | """ | |
509 | text = self.revision(node) |
|
509 | text = self.revision(node) | |
510 | if not text: |
|
510 | if not text: | |
511 | return [] |
|
511 | return [] | |
512 | last = text.index(b"\n\n") |
|
512 | last = text.index(b"\n\n") | |
513 | l = text[:last].split(b'\n') |
|
513 | l = text[:last].split(b'\n') | |
514 | return l[3:] |
|
514 | return l[3:] | |
515 |
|
515 | |||
516 | def add( |
|
516 | def add( | |
517 | self, |
|
517 | self, | |
518 | manifest, |
|
518 | manifest, | |
519 | files, |
|
519 | files, | |
520 | desc, |
|
520 | desc, | |
521 | transaction, |
|
521 | transaction, | |
522 | p1, |
|
522 | p1, | |
523 | p2, |
|
523 | p2, | |
524 | user, |
|
524 | user, | |
525 | date=None, |
|
525 | date=None, | |
526 | extra=None, |
|
526 | extra=None, | |
527 | ): |
|
527 | ): | |
528 | # Convert to UTF-8 encoded bytestrings as the very first |
|
528 | # Convert to UTF-8 encoded bytestrings as the very first | |
529 | # thing: calling any method on a localstr object will turn it |
|
529 | # thing: calling any method on a localstr object will turn it | |
530 | # into a str object and the cached UTF-8 string is thus lost. |
|
530 | # into a str object and the cached UTF-8 string is thus lost. | |
531 | user, desc = encoding.fromlocal(user), encoding.fromlocal(desc) |
|
531 | user, desc = encoding.fromlocal(user), encoding.fromlocal(desc) | |
532 |
|
532 | |||
533 | user = user.strip() |
|
533 | user = user.strip() | |
534 | # An empty username or a username with a "\n" will make the |
|
534 | # An empty username or a username with a "\n" will make the | |
535 | # revision text contain two "\n\n" sequences -> corrupt |
|
535 | # revision text contain two "\n\n" sequences -> corrupt | |
536 | # repository since read cannot unpack the revision. |
|
536 | # repository since read cannot unpack the revision. | |
537 | if not user: |
|
537 | if not user: | |
538 | raise error.StorageError(_(b"empty username")) |
|
538 | raise error.StorageError(_(b"empty username")) | |
539 | if b"\n" in user: |
|
539 | if b"\n" in user: | |
540 | raise error.StorageError( |
|
540 | raise error.StorageError( | |
541 | _(b"username %r contains a newline") % pycompat.bytestr(user) |
|
541 | _(b"username %r contains a newline") % pycompat.bytestr(user) | |
542 | ) |
|
542 | ) | |
543 |
|
543 | |||
544 | desc = stripdesc(desc) |
|
544 | desc = stripdesc(desc) | |
545 |
|
545 | |||
546 | if date: |
|
546 | if date: | |
547 | parseddate = b"%d %d" % dateutil.parsedate(date) |
|
547 | parseddate = b"%d %d" % dateutil.parsedate(date) | |
548 | else: |
|
548 | else: | |
549 | parseddate = b"%d %d" % dateutil.makedate() |
|
549 | parseddate = b"%d %d" % dateutil.makedate() | |
550 | if extra: |
|
550 | if extra: | |
551 | branch = extra.get(b"branch") |
|
551 | branch = extra.get(b"branch") | |
552 | if branch in (b"default", b""): |
|
552 | if branch in (b"default", b""): | |
553 | del extra[b"branch"] |
|
553 | del extra[b"branch"] | |
554 | elif branch in (b".", b"null", b"tip"): |
|
554 | elif branch in (b".", b"null", b"tip"): | |
555 | raise error.StorageError( |
|
555 | raise error.StorageError( | |
556 | _(b'the name \'%s\' is reserved') % branch |
|
556 | _(b'the name \'%s\' is reserved') % branch | |
557 | ) |
|
557 | ) | |
558 | sortedfiles = sorted(files.touched) |
|
558 | sortedfiles = sorted(files.touched) | |
559 | sidedata = None |
|
559 | sidedata = None | |
560 | if self._copiesstorage == b'changeset-sidedata': |
|
560 | if self._copiesstorage == b'changeset-sidedata': | |
561 | sidedata = {} |
|
561 | sidedata = metadata.encode_copies_sidedata(files) | |
562 | p1copies = files.copied_from_p1 |
|
|||
563 | if p1copies: |
|
|||
564 | p1copies = metadata.encodecopies(sortedfiles, p1copies) |
|
|||
565 | sidedata[sidedatamod.SD_P1COPIES] = p1copies |
|
|||
566 | p2copies = files.copied_from_p2 |
|
|||
567 | if p2copies: |
|
|||
568 | p2copies = metadata.encodecopies(sortedfiles, p2copies) |
|
|||
569 | sidedata[sidedatamod.SD_P2COPIES] = p2copies |
|
|||
570 | filesadded = files.added |
|
|||
571 | if filesadded: |
|
|||
572 | filesadded = metadata.encodefileindices(sortedfiles, filesadded) |
|
|||
573 | sidedata[sidedatamod.SD_FILESADDED] = filesadded |
|
|||
574 | filesremoved = files.removed |
|
|||
575 | if filesremoved: |
|
|||
576 | filesremoved = metadata.encodefileindices( |
|
|||
577 | sortedfiles, filesremoved |
|
|||
578 | ) |
|
|||
579 | sidedata[sidedatamod.SD_FILESREMOVED] = filesremoved |
|
|||
580 | if not sidedata: |
|
|||
581 | sidedata = None |
|
|||
582 |
|
562 | |||
583 | if extra: |
|
563 | if extra: | |
584 | extra = encodeextra(extra) |
|
564 | extra = encodeextra(extra) | |
585 | parseddate = b"%s %s" % (parseddate, extra) |
|
565 | parseddate = b"%s %s" % (parseddate, extra) | |
586 | l = [hex(manifest), user, parseddate] + sortedfiles + [b"", desc] |
|
566 | l = [hex(manifest), user, parseddate] + sortedfiles + [b"", desc] | |
587 | text = b"\n".join(l) |
|
567 | text = b"\n".join(l) | |
588 | return self.addrevision( |
|
568 | return self.addrevision( | |
589 | text, transaction, len(self), p1, p2, sidedata=sidedata |
|
569 | text, transaction, len(self), p1, p2, sidedata=sidedata | |
590 | ) |
|
570 | ) | |
591 |
|
571 | |||
592 | def branchinfo(self, rev): |
|
572 | def branchinfo(self, rev): | |
593 | """return the branch name and open/close state of a revision |
|
573 | """return the branch name and open/close state of a revision | |
594 |
|
574 | |||
595 | This function exists because creating a changectx object |
|
575 | This function exists because creating a changectx object | |
596 | just to access this is costly.""" |
|
576 | just to access this is costly.""" | |
597 | extra = self.read(rev)[5] |
|
577 | extra = self.read(rev)[5] | |
598 | return encoding.tolocal(extra.get(b"branch")), b'close' in extra |
|
578 | return encoding.tolocal(extra.get(b"branch")), b'close' in extra | |
599 |
|
579 | |||
600 | def _nodeduplicatecallback(self, transaction, node): |
|
580 | def _nodeduplicatecallback(self, transaction, node): | |
601 | # keep track of revisions that got "re-added", eg: unbunde of know rev. |
|
581 | # keep track of revisions that got "re-added", eg: unbunde of know rev. | |
602 | # |
|
582 | # | |
603 | # We track them in a list to preserve their order from the source bundle |
|
583 | # We track them in a list to preserve their order from the source bundle | |
604 | duplicates = transaction.changes.setdefault(b'revduplicates', []) |
|
584 | duplicates = transaction.changes.setdefault(b'revduplicates', []) | |
605 | duplicates.append(self.rev(node)) |
|
585 | duplicates.append(self.rev(node)) |
@@ -1,400 +1,424 b'' | |||||
1 | # metadata.py -- code related to various metadata computation and access. |
|
1 | # metadata.py -- code related to various metadata computation and access. | |
2 | # |
|
2 | # | |
3 | # Copyright 2019 Google, Inc <martinvonz@google.com> |
|
3 | # Copyright 2019 Google, Inc <martinvonz@google.com> | |
4 | # Copyright 2020 Pierre-Yves David <pierre-yves.david@octobus.net> |
|
4 | # Copyright 2020 Pierre-Yves David <pierre-yves.david@octobus.net> | |
5 | # |
|
5 | # | |
6 | # This software may be used and distributed according to the terms of the |
|
6 | # This software may be used and distributed according to the terms of the | |
7 | # GNU General Public License version 2 or any later version. |
|
7 | # GNU General Public License version 2 or any later version. | |
8 | from __future__ import absolute_import, print_function |
|
8 | from __future__ import absolute_import, print_function | |
9 |
|
9 | |||
10 | import multiprocessing |
|
10 | import multiprocessing | |
11 |
|
11 | |||
12 | from . import ( |
|
12 | from . import ( | |
13 | error, |
|
13 | error, | |
14 | node, |
|
14 | node, | |
15 | pycompat, |
|
15 | pycompat, | |
16 | util, |
|
16 | util, | |
17 | ) |
|
17 | ) | |
18 |
|
18 | |||
19 | from .revlogutils import ( |
|
19 | from .revlogutils import ( | |
20 | flagutil as sidedataflag, |
|
20 | flagutil as sidedataflag, | |
21 | sidedata as sidedatamod, |
|
21 | sidedata as sidedatamod, | |
22 | ) |
|
22 | ) | |
23 |
|
23 | |||
24 |
|
24 | |||
25 | class ChangingFiles(object): |
|
25 | class ChangingFiles(object): | |
26 | """A class recording the changes made to a file by a revision |
|
26 | """A class recording the changes made to a file by a revision | |
27 | """ |
|
27 | """ | |
28 |
|
28 | |||
29 | def __init__( |
|
29 | def __init__( | |
30 | self, touched=(), added=(), removed=(), p1_copies=(), p2_copies=(), |
|
30 | self, touched=(), added=(), removed=(), p1_copies=(), p2_copies=(), | |
31 | ): |
|
31 | ): | |
32 | self._added = set(added) |
|
32 | self._added = set(added) | |
33 | self._removed = set(removed) |
|
33 | self._removed = set(removed) | |
34 | self._touched = set(touched) |
|
34 | self._touched = set(touched) | |
35 | self._touched.update(self._added) |
|
35 | self._touched.update(self._added) | |
36 | self._touched.update(self._removed) |
|
36 | self._touched.update(self._removed) | |
37 | self._p1_copies = dict(p1_copies) |
|
37 | self._p1_copies = dict(p1_copies) | |
38 | self._p2_copies = dict(p2_copies) |
|
38 | self._p2_copies = dict(p2_copies) | |
39 |
|
39 | |||
40 | @property |
|
40 | @property | |
41 | def added(self): |
|
41 | def added(self): | |
42 | return frozenset(self._added) |
|
42 | return frozenset(self._added) | |
43 |
|
43 | |||
44 | def mark_added(self, filename): |
|
44 | def mark_added(self, filename): | |
45 | self._added.add(filename) |
|
45 | self._added.add(filename) | |
46 | self._touched.add(filename) |
|
46 | self._touched.add(filename) | |
47 |
|
47 | |||
48 | def update_added(self, filenames): |
|
48 | def update_added(self, filenames): | |
49 | for f in filenames: |
|
49 | for f in filenames: | |
50 | self.mark_added(f) |
|
50 | self.mark_added(f) | |
51 |
|
51 | |||
52 | @property |
|
52 | @property | |
53 | def removed(self): |
|
53 | def removed(self): | |
54 | return frozenset(self._removed) |
|
54 | return frozenset(self._removed) | |
55 |
|
55 | |||
56 | def mark_removed(self, filename): |
|
56 | def mark_removed(self, filename): | |
57 | self._removed.add(filename) |
|
57 | self._removed.add(filename) | |
58 | self._touched.add(filename) |
|
58 | self._touched.add(filename) | |
59 |
|
59 | |||
60 | def update_removed(self, filenames): |
|
60 | def update_removed(self, filenames): | |
61 | for f in filenames: |
|
61 | for f in filenames: | |
62 | self.mark_removed(f) |
|
62 | self.mark_removed(f) | |
63 |
|
63 | |||
64 | @property |
|
64 | @property | |
65 | def touched(self): |
|
65 | def touched(self): | |
66 | return frozenset(self._touched) |
|
66 | return frozenset(self._touched) | |
67 |
|
67 | |||
68 | def mark_touched(self, filename): |
|
68 | def mark_touched(self, filename): | |
69 | self._touched.add(filename) |
|
69 | self._touched.add(filename) | |
70 |
|
70 | |||
71 | def update_touched(self, filenames): |
|
71 | def update_touched(self, filenames): | |
72 | for f in filenames: |
|
72 | for f in filenames: | |
73 | self.mark_touched(f) |
|
73 | self.mark_touched(f) | |
74 |
|
74 | |||
75 | @property |
|
75 | @property | |
76 | def copied_from_p1(self): |
|
76 | def copied_from_p1(self): | |
77 | return self._p1_copies.copy() |
|
77 | return self._p1_copies.copy() | |
78 |
|
78 | |||
79 | def mark_copied_from_p1(self, source, dest): |
|
79 | def mark_copied_from_p1(self, source, dest): | |
80 | self._p1_copies[dest] = source |
|
80 | self._p1_copies[dest] = source | |
81 |
|
81 | |||
82 | def update_copies_from_p1(self, copies): |
|
82 | def update_copies_from_p1(self, copies): | |
83 | for dest, source in copies.items(): |
|
83 | for dest, source in copies.items(): | |
84 | self.mark_copied_from_p1(source, dest) |
|
84 | self.mark_copied_from_p1(source, dest) | |
85 |
|
85 | |||
86 | @property |
|
86 | @property | |
87 | def copied_from_p2(self): |
|
87 | def copied_from_p2(self): | |
88 | return self._p2_copies.copy() |
|
88 | return self._p2_copies.copy() | |
89 |
|
89 | |||
90 | def mark_copied_from_p2(self, source, dest): |
|
90 | def mark_copied_from_p2(self, source, dest): | |
91 | self._p2_copies[dest] = source |
|
91 | self._p2_copies[dest] = source | |
92 |
|
92 | |||
93 | def update_copies_from_p2(self, copies): |
|
93 | def update_copies_from_p2(self, copies): | |
94 | for dest, source in copies.items(): |
|
94 | for dest, source in copies.items(): | |
95 | self.mark_copied_from_p2(source, dest) |
|
95 | self.mark_copied_from_p2(source, dest) | |
96 |
|
96 | |||
97 |
|
97 | |||
98 | def computechangesetfilesadded(ctx): |
|
98 | def computechangesetfilesadded(ctx): | |
99 | """return the list of files added in a changeset |
|
99 | """return the list of files added in a changeset | |
100 | """ |
|
100 | """ | |
101 | added = [] |
|
101 | added = [] | |
102 | for f in ctx.files(): |
|
102 | for f in ctx.files(): | |
103 | if not any(f in p for p in ctx.parents()): |
|
103 | if not any(f in p for p in ctx.parents()): | |
104 | added.append(f) |
|
104 | added.append(f) | |
105 | return added |
|
105 | return added | |
106 |
|
106 | |||
107 |
|
107 | |||
108 | def get_removal_filter(ctx, x=None): |
|
108 | def get_removal_filter(ctx, x=None): | |
109 | """return a function to detect files "wrongly" detected as `removed` |
|
109 | """return a function to detect files "wrongly" detected as `removed` | |
110 |
|
110 | |||
111 | When a file is removed relative to p1 in a merge, this |
|
111 | When a file is removed relative to p1 in a merge, this | |
112 | function determines whether the absence is due to a |
|
112 | function determines whether the absence is due to a | |
113 | deletion from a parent, or whether the merge commit |
|
113 | deletion from a parent, or whether the merge commit | |
114 | itself deletes the file. We decide this by doing a |
|
114 | itself deletes the file. We decide this by doing a | |
115 | simplified three way merge of the manifest entry for |
|
115 | simplified three way merge of the manifest entry for | |
116 | the file. There are two ways we decide the merge |
|
116 | the file. There are two ways we decide the merge | |
117 | itself didn't delete a file: |
|
117 | itself didn't delete a file: | |
118 | - neither parent (nor the merge) contain the file |
|
118 | - neither parent (nor the merge) contain the file | |
119 | - exactly one parent contains the file, and that |
|
119 | - exactly one parent contains the file, and that | |
120 | parent has the same filelog entry as the merge |
|
120 | parent has the same filelog entry as the merge | |
121 | ancestor (or all of them if there two). In other |
|
121 | ancestor (or all of them if there two). In other | |
122 | words, that parent left the file unchanged while the |
|
122 | words, that parent left the file unchanged while the | |
123 | other one deleted it. |
|
123 | other one deleted it. | |
124 | One way to think about this is that deleting a file is |
|
124 | One way to think about this is that deleting a file is | |
125 | similar to emptying it, so the list of changed files |
|
125 | similar to emptying it, so the list of changed files | |
126 | should be similar either way. The computation |
|
126 | should be similar either way. The computation | |
127 | described above is not done directly in _filecommit |
|
127 | described above is not done directly in _filecommit | |
128 | when creating the list of changed files, however |
|
128 | when creating the list of changed files, however | |
129 | it does something very similar by comparing filelog |
|
129 | it does something very similar by comparing filelog | |
130 | nodes. |
|
130 | nodes. | |
131 | """ |
|
131 | """ | |
132 |
|
132 | |||
133 | if x is not None: |
|
133 | if x is not None: | |
134 | p1, p2, m1, m2 = x |
|
134 | p1, p2, m1, m2 = x | |
135 | else: |
|
135 | else: | |
136 | p1 = ctx.p1() |
|
136 | p1 = ctx.p1() | |
137 | p2 = ctx.p2() |
|
137 | p2 = ctx.p2() | |
138 | m1 = p1.manifest() |
|
138 | m1 = p1.manifest() | |
139 | m2 = p2.manifest() |
|
139 | m2 = p2.manifest() | |
140 |
|
140 | |||
141 | @util.cachefunc |
|
141 | @util.cachefunc | |
142 | def mas(): |
|
142 | def mas(): | |
143 | p1n = p1.node() |
|
143 | p1n = p1.node() | |
144 | p2n = p2.node() |
|
144 | p2n = p2.node() | |
145 | cahs = ctx.repo().changelog.commonancestorsheads(p1n, p2n) |
|
145 | cahs = ctx.repo().changelog.commonancestorsheads(p1n, p2n) | |
146 | if not cahs: |
|
146 | if not cahs: | |
147 | cahs = [node.nullrev] |
|
147 | cahs = [node.nullrev] | |
148 | return [ctx.repo()[r].manifest() for r in cahs] |
|
148 | return [ctx.repo()[r].manifest() for r in cahs] | |
149 |
|
149 | |||
150 | def deletionfromparent(f): |
|
150 | def deletionfromparent(f): | |
151 | if f in m1: |
|
151 | if f in m1: | |
152 | return f not in m2 and all( |
|
152 | return f not in m2 and all( | |
153 | f in ma and ma.find(f) == m1.find(f) for ma in mas() |
|
153 | f in ma and ma.find(f) == m1.find(f) for ma in mas() | |
154 | ) |
|
154 | ) | |
155 | elif f in m2: |
|
155 | elif f in m2: | |
156 | return all(f in ma and ma.find(f) == m2.find(f) for ma in mas()) |
|
156 | return all(f in ma and ma.find(f) == m2.find(f) for ma in mas()) | |
157 | else: |
|
157 | else: | |
158 | return True |
|
158 | return True | |
159 |
|
159 | |||
160 | return deletionfromparent |
|
160 | return deletionfromparent | |
161 |
|
161 | |||
162 |
|
162 | |||
163 | def computechangesetfilesremoved(ctx): |
|
163 | def computechangesetfilesremoved(ctx): | |
164 | """return the list of files removed in a changeset |
|
164 | """return the list of files removed in a changeset | |
165 | """ |
|
165 | """ | |
166 | removed = [] |
|
166 | removed = [] | |
167 | for f in ctx.files(): |
|
167 | for f in ctx.files(): | |
168 | if f not in ctx: |
|
168 | if f not in ctx: | |
169 | removed.append(f) |
|
169 | removed.append(f) | |
170 | if removed: |
|
170 | if removed: | |
171 | rf = get_removal_filter(ctx) |
|
171 | rf = get_removal_filter(ctx) | |
172 | removed = [r for r in removed if not rf(r)] |
|
172 | removed = [r for r in removed if not rf(r)] | |
173 | return removed |
|
173 | return removed | |
174 |
|
174 | |||
175 |
|
175 | |||
176 | def computechangesetcopies(ctx): |
|
176 | def computechangesetcopies(ctx): | |
177 | """return the copies data for a changeset |
|
177 | """return the copies data for a changeset | |
178 |
|
178 | |||
179 | The copies data are returned as a pair of dictionnary (p1copies, p2copies). |
|
179 | The copies data are returned as a pair of dictionnary (p1copies, p2copies). | |
180 |
|
180 | |||
181 | Each dictionnary are in the form: `{newname: oldname}` |
|
181 | Each dictionnary are in the form: `{newname: oldname}` | |
182 | """ |
|
182 | """ | |
183 | p1copies = {} |
|
183 | p1copies = {} | |
184 | p2copies = {} |
|
184 | p2copies = {} | |
185 | p1 = ctx.p1() |
|
185 | p1 = ctx.p1() | |
186 | p2 = ctx.p2() |
|
186 | p2 = ctx.p2() | |
187 | narrowmatch = ctx._repo.narrowmatch() |
|
187 | narrowmatch = ctx._repo.narrowmatch() | |
188 | for dst in ctx.files(): |
|
188 | for dst in ctx.files(): | |
189 | if not narrowmatch(dst) or dst not in ctx: |
|
189 | if not narrowmatch(dst) or dst not in ctx: | |
190 | continue |
|
190 | continue | |
191 | copied = ctx[dst].renamed() |
|
191 | copied = ctx[dst].renamed() | |
192 | if not copied: |
|
192 | if not copied: | |
193 | continue |
|
193 | continue | |
194 | src, srcnode = copied |
|
194 | src, srcnode = copied | |
195 | if src in p1 and p1[src].filenode() == srcnode: |
|
195 | if src in p1 and p1[src].filenode() == srcnode: | |
196 | p1copies[dst] = src |
|
196 | p1copies[dst] = src | |
197 | elif src in p2 and p2[src].filenode() == srcnode: |
|
197 | elif src in p2 and p2[src].filenode() == srcnode: | |
198 | p2copies[dst] = src |
|
198 | p2copies[dst] = src | |
199 | return p1copies, p2copies |
|
199 | return p1copies, p2copies | |
200 |
|
200 | |||
201 |
|
201 | |||
202 | def encodecopies(files, copies): |
|
202 | def encodecopies(files, copies): | |
203 | items = [] |
|
203 | items = [] | |
204 | for i, dst in enumerate(files): |
|
204 | for i, dst in enumerate(files): | |
205 | if dst in copies: |
|
205 | if dst in copies: | |
206 | items.append(b'%d\0%s' % (i, copies[dst])) |
|
206 | items.append(b'%d\0%s' % (i, copies[dst])) | |
207 | if len(items) != len(copies): |
|
207 | if len(items) != len(copies): | |
208 | raise error.ProgrammingError( |
|
208 | raise error.ProgrammingError( | |
209 | b'some copy targets missing from file list' |
|
209 | b'some copy targets missing from file list' | |
210 | ) |
|
210 | ) | |
211 | return b"\n".join(items) |
|
211 | return b"\n".join(items) | |
212 |
|
212 | |||
213 |
|
213 | |||
214 | def decodecopies(files, data): |
|
214 | def decodecopies(files, data): | |
215 | try: |
|
215 | try: | |
216 | copies = {} |
|
216 | copies = {} | |
217 | if not data: |
|
217 | if not data: | |
218 | return copies |
|
218 | return copies | |
219 | for l in data.split(b'\n'): |
|
219 | for l in data.split(b'\n'): | |
220 | strindex, src = l.split(b'\0') |
|
220 | strindex, src = l.split(b'\0') | |
221 | i = int(strindex) |
|
221 | i = int(strindex) | |
222 | dst = files[i] |
|
222 | dst = files[i] | |
223 | copies[dst] = src |
|
223 | copies[dst] = src | |
224 | return copies |
|
224 | return copies | |
225 | except (ValueError, IndexError): |
|
225 | except (ValueError, IndexError): | |
226 | # Perhaps someone had chosen the same key name (e.g. "p1copies") and |
|
226 | # Perhaps someone had chosen the same key name (e.g. "p1copies") and | |
227 | # used different syntax for the value. |
|
227 | # used different syntax for the value. | |
228 | return None |
|
228 | return None | |
229 |
|
229 | |||
230 |
|
230 | |||
231 | def encodefileindices(files, subset): |
|
231 | def encodefileindices(files, subset): | |
232 | subset = set(subset) |
|
232 | subset = set(subset) | |
233 | indices = [] |
|
233 | indices = [] | |
234 | for i, f in enumerate(files): |
|
234 | for i, f in enumerate(files): | |
235 | if f in subset: |
|
235 | if f in subset: | |
236 | indices.append(b'%d' % i) |
|
236 | indices.append(b'%d' % i) | |
237 | return b'\n'.join(indices) |
|
237 | return b'\n'.join(indices) | |
238 |
|
238 | |||
239 |
|
239 | |||
240 | def decodefileindices(files, data): |
|
240 | def decodefileindices(files, data): | |
241 | try: |
|
241 | try: | |
242 | subset = [] |
|
242 | subset = [] | |
243 | if not data: |
|
243 | if not data: | |
244 | return subset |
|
244 | return subset | |
245 | for strindex in data.split(b'\n'): |
|
245 | for strindex in data.split(b'\n'): | |
246 | i = int(strindex) |
|
246 | i = int(strindex) | |
247 | if i < 0 or i >= len(files): |
|
247 | if i < 0 or i >= len(files): | |
248 | return None |
|
248 | return None | |
249 | subset.append(files[i]) |
|
249 | subset.append(files[i]) | |
250 | return subset |
|
250 | return subset | |
251 | except (ValueError, IndexError): |
|
251 | except (ValueError, IndexError): | |
252 | # Perhaps someone had chosen the same key name (e.g. "added") and |
|
252 | # Perhaps someone had chosen the same key name (e.g. "added") and | |
253 | # used different syntax for the value. |
|
253 | # used different syntax for the value. | |
254 | return None |
|
254 | return None | |
255 |
|
255 | |||
256 |
|
256 | |||
|
257 | def encode_copies_sidedata(files): | |||
|
258 | sortedfiles = sorted(files.touched) | |||
|
259 | sidedata = {} | |||
|
260 | p1copies = files.copied_from_p1 | |||
|
261 | if p1copies: | |||
|
262 | p1copies = encodecopies(sortedfiles, p1copies) | |||
|
263 | sidedata[sidedatamod.SD_P1COPIES] = p1copies | |||
|
264 | p2copies = files.copied_from_p2 | |||
|
265 | if p2copies: | |||
|
266 | p2copies = encodecopies(sortedfiles, p2copies) | |||
|
267 | sidedata[sidedatamod.SD_P2COPIES] = p2copies | |||
|
268 | filesadded = files.added | |||
|
269 | if filesadded: | |||
|
270 | filesadded = encodefileindices(sortedfiles, filesadded) | |||
|
271 | sidedata[sidedatamod.SD_FILESADDED] = filesadded | |||
|
272 | filesremoved = files.removed | |||
|
273 | if filesremoved: | |||
|
274 | filesremoved = encodefileindices(sortedfiles, filesremoved) | |||
|
275 | sidedata[sidedatamod.SD_FILESREMOVED] = filesremoved | |||
|
276 | if not sidedata: | |||
|
277 | sidedata = None | |||
|
278 | return sidedata | |||
|
279 | ||||
|
280 | ||||
257 | def _getsidedata(srcrepo, rev): |
|
281 | def _getsidedata(srcrepo, rev): | |
258 | ctx = srcrepo[rev] |
|
282 | ctx = srcrepo[rev] | |
259 | filescopies = computechangesetcopies(ctx) |
|
283 | filescopies = computechangesetcopies(ctx) | |
260 | filesadded = computechangesetfilesadded(ctx) |
|
284 | filesadded = computechangesetfilesadded(ctx) | |
261 | filesremoved = computechangesetfilesremoved(ctx) |
|
285 | filesremoved = computechangesetfilesremoved(ctx) | |
262 | sidedata = {} |
|
286 | sidedata = {} | |
263 | if any([filescopies, filesadded, filesremoved]): |
|
287 | if any([filescopies, filesadded, filesremoved]): | |
264 | sortedfiles = sorted(ctx.files()) |
|
288 | sortedfiles = sorted(ctx.files()) | |
265 | p1copies, p2copies = filescopies |
|
289 | p1copies, p2copies = filescopies | |
266 | p1copies = encodecopies(sortedfiles, p1copies) |
|
290 | p1copies = encodecopies(sortedfiles, p1copies) | |
267 | p2copies = encodecopies(sortedfiles, p2copies) |
|
291 | p2copies = encodecopies(sortedfiles, p2copies) | |
268 | filesadded = encodefileindices(sortedfiles, filesadded) |
|
292 | filesadded = encodefileindices(sortedfiles, filesadded) | |
269 | filesremoved = encodefileindices(sortedfiles, filesremoved) |
|
293 | filesremoved = encodefileindices(sortedfiles, filesremoved) | |
270 | if p1copies: |
|
294 | if p1copies: | |
271 | sidedata[sidedatamod.SD_P1COPIES] = p1copies |
|
295 | sidedata[sidedatamod.SD_P1COPIES] = p1copies | |
272 | if p2copies: |
|
296 | if p2copies: | |
273 | sidedata[sidedatamod.SD_P2COPIES] = p2copies |
|
297 | sidedata[sidedatamod.SD_P2COPIES] = p2copies | |
274 | if filesadded: |
|
298 | if filesadded: | |
275 | sidedata[sidedatamod.SD_FILESADDED] = filesadded |
|
299 | sidedata[sidedatamod.SD_FILESADDED] = filesadded | |
276 | if filesremoved: |
|
300 | if filesremoved: | |
277 | sidedata[sidedatamod.SD_FILESREMOVED] = filesremoved |
|
301 | sidedata[sidedatamod.SD_FILESREMOVED] = filesremoved | |
278 | return sidedata |
|
302 | return sidedata | |
279 |
|
303 | |||
280 |
|
304 | |||
281 | def getsidedataadder(srcrepo, destrepo): |
|
305 | def getsidedataadder(srcrepo, destrepo): | |
282 | use_w = srcrepo.ui.configbool(b'experimental', b'worker.repository-upgrade') |
|
306 | use_w = srcrepo.ui.configbool(b'experimental', b'worker.repository-upgrade') | |
283 | if pycompat.iswindows or not use_w: |
|
307 | if pycompat.iswindows or not use_w: | |
284 | return _get_simple_sidedata_adder(srcrepo, destrepo) |
|
308 | return _get_simple_sidedata_adder(srcrepo, destrepo) | |
285 | else: |
|
309 | else: | |
286 | return _get_worker_sidedata_adder(srcrepo, destrepo) |
|
310 | return _get_worker_sidedata_adder(srcrepo, destrepo) | |
287 |
|
311 | |||
288 |
|
312 | |||
289 | def _sidedata_worker(srcrepo, revs_queue, sidedata_queue, tokens): |
|
313 | def _sidedata_worker(srcrepo, revs_queue, sidedata_queue, tokens): | |
290 | """The function used by worker precomputing sidedata |
|
314 | """The function used by worker precomputing sidedata | |
291 |
|
315 | |||
292 | It read an input queue containing revision numbers |
|
316 | It read an input queue containing revision numbers | |
293 | It write in an output queue containing (rev, <sidedata-map>) |
|
317 | It write in an output queue containing (rev, <sidedata-map>) | |
294 |
|
318 | |||
295 | The `None` input value is used as a stop signal. |
|
319 | The `None` input value is used as a stop signal. | |
296 |
|
320 | |||
297 | The `tokens` semaphore is user to avoid having too many unprocessed |
|
321 | The `tokens` semaphore is user to avoid having too many unprocessed | |
298 | entries. The workers needs to acquire one token before fetching a task. |
|
322 | entries. The workers needs to acquire one token before fetching a task. | |
299 | They will be released by the consumer of the produced data. |
|
323 | They will be released by the consumer of the produced data. | |
300 | """ |
|
324 | """ | |
301 | tokens.acquire() |
|
325 | tokens.acquire() | |
302 | rev = revs_queue.get() |
|
326 | rev = revs_queue.get() | |
303 | while rev is not None: |
|
327 | while rev is not None: | |
304 | data = _getsidedata(srcrepo, rev) |
|
328 | data = _getsidedata(srcrepo, rev) | |
305 | sidedata_queue.put((rev, data)) |
|
329 | sidedata_queue.put((rev, data)) | |
306 | tokens.acquire() |
|
330 | tokens.acquire() | |
307 | rev = revs_queue.get() |
|
331 | rev = revs_queue.get() | |
308 | # processing of `None` is completed, release the token. |
|
332 | # processing of `None` is completed, release the token. | |
309 | tokens.release() |
|
333 | tokens.release() | |
310 |
|
334 | |||
311 |
|
335 | |||
312 | BUFF_PER_WORKER = 50 |
|
336 | BUFF_PER_WORKER = 50 | |
313 |
|
337 | |||
314 |
|
338 | |||
315 | def _get_worker_sidedata_adder(srcrepo, destrepo): |
|
339 | def _get_worker_sidedata_adder(srcrepo, destrepo): | |
316 | """The parallel version of the sidedata computation |
|
340 | """The parallel version of the sidedata computation | |
317 |
|
341 | |||
318 | This code spawn a pool of worker that precompute a buffer of sidedata |
|
342 | This code spawn a pool of worker that precompute a buffer of sidedata | |
319 | before we actually need them""" |
|
343 | before we actually need them""" | |
320 | # avoid circular import copies -> scmutil -> worker -> copies |
|
344 | # avoid circular import copies -> scmutil -> worker -> copies | |
321 | from . import worker |
|
345 | from . import worker | |
322 |
|
346 | |||
323 | nbworkers = worker._numworkers(srcrepo.ui) |
|
347 | nbworkers = worker._numworkers(srcrepo.ui) | |
324 |
|
348 | |||
325 | tokens = multiprocessing.BoundedSemaphore(nbworkers * BUFF_PER_WORKER) |
|
349 | tokens = multiprocessing.BoundedSemaphore(nbworkers * BUFF_PER_WORKER) | |
326 | revsq = multiprocessing.Queue() |
|
350 | revsq = multiprocessing.Queue() | |
327 | sidedataq = multiprocessing.Queue() |
|
351 | sidedataq = multiprocessing.Queue() | |
328 |
|
352 | |||
329 | assert srcrepo.filtername is None |
|
353 | assert srcrepo.filtername is None | |
330 | # queue all tasks beforehand, revision numbers are small and it make |
|
354 | # queue all tasks beforehand, revision numbers are small and it make | |
331 | # synchronisation simpler |
|
355 | # synchronisation simpler | |
332 | # |
|
356 | # | |
333 | # Since the computation for each node can be quite expensive, the overhead |
|
357 | # Since the computation for each node can be quite expensive, the overhead | |
334 | # of using a single queue is not revelant. In practice, most computation |
|
358 | # of using a single queue is not revelant. In practice, most computation | |
335 | # are fast but some are very expensive and dominate all the other smaller |
|
359 | # are fast but some are very expensive and dominate all the other smaller | |
336 | # cost. |
|
360 | # cost. | |
337 | for r in srcrepo.changelog.revs(): |
|
361 | for r in srcrepo.changelog.revs(): | |
338 | revsq.put(r) |
|
362 | revsq.put(r) | |
339 | # queue the "no more tasks" markers |
|
363 | # queue the "no more tasks" markers | |
340 | for i in range(nbworkers): |
|
364 | for i in range(nbworkers): | |
341 | revsq.put(None) |
|
365 | revsq.put(None) | |
342 |
|
366 | |||
343 | allworkers = [] |
|
367 | allworkers = [] | |
344 | for i in range(nbworkers): |
|
368 | for i in range(nbworkers): | |
345 | args = (srcrepo, revsq, sidedataq, tokens) |
|
369 | args = (srcrepo, revsq, sidedataq, tokens) | |
346 | w = multiprocessing.Process(target=_sidedata_worker, args=args) |
|
370 | w = multiprocessing.Process(target=_sidedata_worker, args=args) | |
347 | allworkers.append(w) |
|
371 | allworkers.append(w) | |
348 | w.start() |
|
372 | w.start() | |
349 |
|
373 | |||
350 | # dictionnary to store results for revision higher than we one we are |
|
374 | # dictionnary to store results for revision higher than we one we are | |
351 | # looking for. For example, if we need the sidedatamap for 42, and 43 is |
|
375 | # looking for. For example, if we need the sidedatamap for 42, and 43 is | |
352 | # received, when shelve 43 for later use. |
|
376 | # received, when shelve 43 for later use. | |
353 | staging = {} |
|
377 | staging = {} | |
354 |
|
378 | |||
355 | def sidedata_companion(revlog, rev): |
|
379 | def sidedata_companion(revlog, rev): | |
356 | sidedata = {} |
|
380 | sidedata = {} | |
357 | if util.safehasattr(revlog, b'filteredrevs'): # this is a changelog |
|
381 | if util.safehasattr(revlog, b'filteredrevs'): # this is a changelog | |
358 | # Is the data previously shelved ? |
|
382 | # Is the data previously shelved ? | |
359 | sidedata = staging.pop(rev, None) |
|
383 | sidedata = staging.pop(rev, None) | |
360 | if sidedata is None: |
|
384 | if sidedata is None: | |
361 | # look at the queued result until we find the one we are lookig |
|
385 | # look at the queued result until we find the one we are lookig | |
362 | # for (shelve the other ones) |
|
386 | # for (shelve the other ones) | |
363 | r, sidedata = sidedataq.get() |
|
387 | r, sidedata = sidedataq.get() | |
364 | while r != rev: |
|
388 | while r != rev: | |
365 | staging[r] = sidedata |
|
389 | staging[r] = sidedata | |
366 | r, sidedata = sidedataq.get() |
|
390 | r, sidedata = sidedataq.get() | |
367 | tokens.release() |
|
391 | tokens.release() | |
368 | return False, (), sidedata |
|
392 | return False, (), sidedata | |
369 |
|
393 | |||
370 | return sidedata_companion |
|
394 | return sidedata_companion | |
371 |
|
395 | |||
372 |
|
396 | |||
373 | def _get_simple_sidedata_adder(srcrepo, destrepo): |
|
397 | def _get_simple_sidedata_adder(srcrepo, destrepo): | |
374 | """The simple version of the sidedata computation |
|
398 | """The simple version of the sidedata computation | |
375 |
|
399 | |||
376 | It just compute it in the same thread on request""" |
|
400 | It just compute it in the same thread on request""" | |
377 |
|
401 | |||
378 | def sidedatacompanion(revlog, rev): |
|
402 | def sidedatacompanion(revlog, rev): | |
379 | sidedata = {} |
|
403 | sidedata = {} | |
380 | if util.safehasattr(revlog, 'filteredrevs'): # this is a changelog |
|
404 | if util.safehasattr(revlog, 'filteredrevs'): # this is a changelog | |
381 | sidedata = _getsidedata(srcrepo, rev) |
|
405 | sidedata = _getsidedata(srcrepo, rev) | |
382 | return False, (), sidedata |
|
406 | return False, (), sidedata | |
383 |
|
407 | |||
384 | return sidedatacompanion |
|
408 | return sidedatacompanion | |
385 |
|
409 | |||
386 |
|
410 | |||
387 | def getsidedataremover(srcrepo, destrepo): |
|
411 | def getsidedataremover(srcrepo, destrepo): | |
388 | def sidedatacompanion(revlog, rev): |
|
412 | def sidedatacompanion(revlog, rev): | |
389 | f = () |
|
413 | f = () | |
390 | if util.safehasattr(revlog, 'filteredrevs'): # this is a changelog |
|
414 | if util.safehasattr(revlog, 'filteredrevs'): # this is a changelog | |
391 | if revlog.flags(rev) & sidedataflag.REVIDX_SIDEDATA: |
|
415 | if revlog.flags(rev) & sidedataflag.REVIDX_SIDEDATA: | |
392 | f = ( |
|
416 | f = ( | |
393 | sidedatamod.SD_P1COPIES, |
|
417 | sidedatamod.SD_P1COPIES, | |
394 | sidedatamod.SD_P2COPIES, |
|
418 | sidedatamod.SD_P2COPIES, | |
395 | sidedatamod.SD_FILESADDED, |
|
419 | sidedatamod.SD_FILESADDED, | |
396 | sidedatamod.SD_FILESREMOVED, |
|
420 | sidedatamod.SD_FILESREMOVED, | |
397 | ) |
|
421 | ) | |
398 | return False, f, {} |
|
422 | return False, f, {} | |
399 |
|
423 | |||
400 | return sidedatacompanion |
|
424 | return sidedatacompanion |
General Comments 0
You need to be logged in to leave comments.
Login now