Show More
@@ -1,644 +1,575 b'' | |||||
1 | # changelog.py - changelog class for mercurial |
|
1 | # changelog.py - changelog class for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> |
|
3 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 |
|
8 | |||
9 | from .i18n import _ |
|
9 | from .i18n import _ | |
10 | from .node import ( |
|
10 | from .node import ( | |
11 | bin, |
|
11 | bin, | |
12 | hex, |
|
12 | hex, | |
13 | ) |
|
13 | ) | |
14 | from .thirdparty import attr |
|
14 | from .thirdparty import attr | |
15 |
|
15 | |||
16 | from . import ( |
|
16 | from . import ( | |
17 | encoding, |
|
17 | encoding, | |
18 | error, |
|
18 | error, | |
19 | metadata, |
|
19 | metadata, | |
20 | pycompat, |
|
20 | pycompat, | |
21 | revlog, |
|
21 | revlog, | |
22 | ) |
|
22 | ) | |
23 | from .utils import ( |
|
23 | from .utils import ( | |
24 | dateutil, |
|
24 | dateutil, | |
25 | stringutil, |
|
25 | stringutil, | |
26 | ) |
|
26 | ) | |
27 | from .revlogutils import ( |
|
27 | from .revlogutils import ( | |
28 | constants as revlog_constants, |
|
28 | constants as revlog_constants, | |
29 | flagutil, |
|
29 | flagutil, | |
|
30 | randomaccessfile, | |||
30 | ) |
|
31 | ) | |
31 |
|
32 | |||
32 | _defaultextra = {b'branch': b'default'} |
|
33 | _defaultextra = {b'branch': b'default'} | |
33 |
|
34 | |||
34 |
|
35 | |||
35 | def _string_escape(text): |
|
36 | def _string_escape(text): | |
36 | """ |
|
37 | """ | |
37 | >>> from .pycompat import bytechr as chr |
|
38 | >>> from .pycompat import bytechr as chr | |
38 | >>> d = {b'nl': chr(10), b'bs': chr(92), b'cr': chr(13), b'nul': chr(0)} |
|
39 | >>> 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 |
|
40 | >>> s = b"ab%(nl)scd%(bs)s%(bs)sn%(nul)s12ab%(cr)scd%(bs)s%(nl)s" % d | |
40 | >>> s |
|
41 | >>> s | |
41 | 'ab\\ncd\\\\\\\\n\\x0012ab\\rcd\\\\\\n' |
|
42 | 'ab\\ncd\\\\\\\\n\\x0012ab\\rcd\\\\\\n' | |
42 | >>> res = _string_escape(s) |
|
43 | >>> res = _string_escape(s) | |
43 | >>> s == _string_unescape(res) |
|
44 | >>> s == _string_unescape(res) | |
44 | True |
|
45 | True | |
45 | """ |
|
46 | """ | |
46 | # subset of the string_escape codec |
|
47 | # subset of the string_escape codec | |
47 | text = ( |
|
48 | text = ( | |
48 | text.replace(b'\\', b'\\\\') |
|
49 | text.replace(b'\\', b'\\\\') | |
49 | .replace(b'\n', b'\\n') |
|
50 | .replace(b'\n', b'\\n') | |
50 | .replace(b'\r', b'\\r') |
|
51 | .replace(b'\r', b'\\r') | |
51 | ) |
|
52 | ) | |
52 | return text.replace(b'\0', b'\\0') |
|
53 | return text.replace(b'\0', b'\\0') | |
53 |
|
54 | |||
54 |
|
55 | |||
55 | def _string_unescape(text): |
|
56 | def _string_unescape(text): | |
56 | if b'\\0' in text: |
|
57 | if b'\\0' in text: | |
57 | # fix up \0 without getting into trouble with \\0 |
|
58 | # fix up \0 without getting into trouble with \\0 | |
58 | text = text.replace(b'\\\\', b'\\\\\n') |
|
59 | text = text.replace(b'\\\\', b'\\\\\n') | |
59 | text = text.replace(b'\\0', b'\0') |
|
60 | text = text.replace(b'\\0', b'\0') | |
60 | text = text.replace(b'\n', b'') |
|
61 | text = text.replace(b'\n', b'') | |
61 | return stringutil.unescapestr(text) |
|
62 | return stringutil.unescapestr(text) | |
62 |
|
63 | |||
63 |
|
64 | |||
64 | def decodeextra(text): |
|
65 | def decodeextra(text): | |
65 | """ |
|
66 | """ | |
66 | >>> from .pycompat import bytechr as chr |
|
67 | >>> from .pycompat import bytechr as chr | |
67 | >>> sorted(decodeextra(encodeextra({b'foo': b'bar', b'baz': chr(0) + b'2'}) |
|
68 | >>> sorted(decodeextra(encodeextra({b'foo': b'bar', b'baz': chr(0) + b'2'}) | |
68 | ... ).items()) |
|
69 | ... ).items()) | |
69 | [('baz', '\\x002'), ('branch', 'default'), ('foo', 'bar')] |
|
70 | [('baz', '\\x002'), ('branch', 'default'), ('foo', 'bar')] | |
70 | >>> sorted(decodeextra(encodeextra({b'foo': b'bar', |
|
71 | >>> sorted(decodeextra(encodeextra({b'foo': b'bar', | |
71 | ... b'baz': chr(92) + chr(0) + b'2'}) |
|
72 | ... b'baz': chr(92) + chr(0) + b'2'}) | |
72 | ... ).items()) |
|
73 | ... ).items()) | |
73 | [('baz', '\\\\\\x002'), ('branch', 'default'), ('foo', 'bar')] |
|
74 | [('baz', '\\\\\\x002'), ('branch', 'default'), ('foo', 'bar')] | |
74 | """ |
|
75 | """ | |
75 | extra = _defaultextra.copy() |
|
76 | extra = _defaultextra.copy() | |
76 | for l in text.split(b'\0'): |
|
77 | for l in text.split(b'\0'): | |
77 | if l: |
|
78 | if l: | |
78 | k, v = _string_unescape(l).split(b':', 1) |
|
79 | k, v = _string_unescape(l).split(b':', 1) | |
79 | extra[k] = v |
|
80 | extra[k] = v | |
80 | return extra |
|
81 | return extra | |
81 |
|
82 | |||
82 |
|
83 | |||
83 | def encodeextra(d): |
|
84 | def encodeextra(d): | |
84 | # keys must be sorted to produce a deterministic changelog entry |
|
85 | # 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)] |
|
86 | items = [_string_escape(b'%s:%s' % (k, d[k])) for k in sorted(d)] | |
86 | return b"\0".join(items) |
|
87 | return b"\0".join(items) | |
87 |
|
88 | |||
88 |
|
89 | |||
89 | def stripdesc(desc): |
|
90 | def stripdesc(desc): | |
90 | """strip trailing whitespace and leading and trailing empty lines""" |
|
91 | """strip trailing whitespace and leading and trailing empty lines""" | |
91 | return b'\n'.join([l.rstrip() for l in desc.splitlines()]).strip(b'\n') |
|
92 | return b'\n'.join([l.rstrip() for l in desc.splitlines()]).strip(b'\n') | |
92 |
|
93 | |||
93 |
|
94 | |||
94 | class appender: |
|
|||
95 | """the changelog index must be updated last on disk, so we use this class |
|
|||
96 | to delay writes to it""" |
|
|||
97 |
|
||||
98 | def __init__(self, vfs, name, mode, buf): |
|
|||
99 | self.data = buf |
|
|||
100 | fp = vfs(name, mode) |
|
|||
101 | self.fp = fp |
|
|||
102 | self.offset = fp.tell() |
|
|||
103 | self.size = vfs.fstat(fp).st_size |
|
|||
104 | self._end = self.size |
|
|||
105 |
|
||||
106 | def end(self): |
|
|||
107 | return self._end |
|
|||
108 |
|
||||
109 | def tell(self): |
|
|||
110 | return self.offset |
|
|||
111 |
|
||||
112 | def flush(self): |
|
|||
113 | pass |
|
|||
114 |
|
||||
115 | @property |
|
|||
116 | def closed(self): |
|
|||
117 | return self.fp.closed |
|
|||
118 |
|
||||
119 | def close(self): |
|
|||
120 | self.fp.close() |
|
|||
121 |
|
||||
122 | def seek(self, offset, whence=0): |
|
|||
123 | '''virtual file offset spans real file and data''' |
|
|||
124 | if whence == 0: |
|
|||
125 | self.offset = offset |
|
|||
126 | elif whence == 1: |
|
|||
127 | self.offset += offset |
|
|||
128 | elif whence == 2: |
|
|||
129 | self.offset = self.end() + offset |
|
|||
130 | if self.offset < self.size: |
|
|||
131 | self.fp.seek(self.offset) |
|
|||
132 |
|
||||
133 | def read(self, count=-1): |
|
|||
134 | '''only trick here is reads that span real file and data''' |
|
|||
135 | ret = b"" |
|
|||
136 | if self.offset < self.size: |
|
|||
137 | s = self.fp.read(count) |
|
|||
138 | ret = s |
|
|||
139 | self.offset += len(s) |
|
|||
140 | if count > 0: |
|
|||
141 | count -= len(s) |
|
|||
142 | if count != 0: |
|
|||
143 | doff = self.offset - self.size |
|
|||
144 | self.data.insert(0, b"".join(self.data)) |
|
|||
145 | del self.data[1:] |
|
|||
146 | s = self.data[0][doff : doff + count] |
|
|||
147 | self.offset += len(s) |
|
|||
148 | ret += s |
|
|||
149 | return ret |
|
|||
150 |
|
||||
151 | def write(self, s): |
|
|||
152 | self.data.append(bytes(s)) |
|
|||
153 | self.offset += len(s) |
|
|||
154 | self._end += len(s) |
|
|||
155 |
|
||||
156 | def __enter__(self): |
|
|||
157 | self.fp.__enter__() |
|
|||
158 | return self |
|
|||
159 |
|
||||
160 | def __exit__(self, *args): |
|
|||
161 | return self.fp.__exit__(*args) |
|
|||
162 |
|
||||
163 |
|
||||
164 | class _divertopener: |
|
95 | class _divertopener: | |
165 | def __init__(self, opener, target): |
|
96 | def __init__(self, opener, target): | |
166 | self._opener = opener |
|
97 | self._opener = opener | |
167 | self._target = target |
|
98 | self._target = target | |
168 |
|
99 | |||
169 | def __call__(self, name, mode=b'r', checkambig=False, **kwargs): |
|
100 | def __call__(self, name, mode=b'r', checkambig=False, **kwargs): | |
170 | if name != self._target: |
|
101 | if name != self._target: | |
171 | return self._opener(name, mode, **kwargs) |
|
102 | return self._opener(name, mode, **kwargs) | |
172 | return self._opener(name + b".a", mode, **kwargs) |
|
103 | return self._opener(name + b".a", mode, **kwargs) | |
173 |
|
104 | |||
174 | def __getattr__(self, attr): |
|
105 | def __getattr__(self, attr): | |
175 | return getattr(self._opener, attr) |
|
106 | return getattr(self._opener, attr) | |
176 |
|
107 | |||
177 |
|
108 | |||
178 | class _delayopener: |
|
109 | class _delayopener: | |
179 | """build an opener that stores chunks in 'buf' instead of 'target'""" |
|
110 | """build an opener that stores chunks in 'buf' instead of 'target'""" | |
180 |
|
111 | |||
181 | def __init__(self, opener, target, buf): |
|
112 | def __init__(self, opener, target, buf): | |
182 | self._opener = opener |
|
113 | self._opener = opener | |
183 | self._target = target |
|
114 | self._target = target | |
184 | self._buf = buf |
|
115 | self._buf = buf | |
185 |
|
116 | |||
186 | def __call__(self, name, mode=b'r', checkambig=False, **kwargs): |
|
117 | def __call__(self, name, mode=b'r', checkambig=False, **kwargs): | |
187 | if name != self._target: |
|
118 | if name != self._target: | |
188 | return self._opener(name, mode, **kwargs) |
|
119 | return self._opener(name, mode, **kwargs) | |
189 | assert not kwargs |
|
120 | assert not kwargs | |
190 | return appender(self._opener, name, mode, self._buf) |
|
121 | return randomaccessfile.appender(self._opener, name, mode, self._buf) | |
191 |
|
122 | |||
192 | def __getattr__(self, attr): |
|
123 | def __getattr__(self, attr): | |
193 | return getattr(self._opener, attr) |
|
124 | return getattr(self._opener, attr) | |
194 |
|
125 | |||
195 |
|
126 | |||
196 | @attr.s |
|
127 | @attr.s | |
197 | class _changelogrevision: |
|
128 | class _changelogrevision: | |
198 | # Extensions might modify _defaultextra, so let the constructor below pass |
|
129 | # Extensions might modify _defaultextra, so let the constructor below pass | |
199 | # it in |
|
130 | # it in | |
200 | extra = attr.ib() |
|
131 | extra = attr.ib() | |
201 | manifest = attr.ib() |
|
132 | manifest = attr.ib() | |
202 | user = attr.ib(default=b'') |
|
133 | user = attr.ib(default=b'') | |
203 | date = attr.ib(default=(0, 0)) |
|
134 | date = attr.ib(default=(0, 0)) | |
204 | files = attr.ib(default=attr.Factory(list)) |
|
135 | files = attr.ib(default=attr.Factory(list)) | |
205 | filesadded = attr.ib(default=None) |
|
136 | filesadded = attr.ib(default=None) | |
206 | filesremoved = attr.ib(default=None) |
|
137 | filesremoved = attr.ib(default=None) | |
207 | p1copies = attr.ib(default=None) |
|
138 | p1copies = attr.ib(default=None) | |
208 | p2copies = attr.ib(default=None) |
|
139 | p2copies = attr.ib(default=None) | |
209 | description = attr.ib(default=b'') |
|
140 | description = attr.ib(default=b'') | |
210 | branchinfo = attr.ib(default=(_defaultextra[b'branch'], False)) |
|
141 | branchinfo = attr.ib(default=(_defaultextra[b'branch'], False)) | |
211 |
|
142 | |||
212 |
|
143 | |||
213 | class changelogrevision: |
|
144 | class changelogrevision: | |
214 | """Holds results of a parsed changelog revision. |
|
145 | """Holds results of a parsed changelog revision. | |
215 |
|
146 | |||
216 | Changelog revisions consist of multiple pieces of data, including |
|
147 | Changelog revisions consist of multiple pieces of data, including | |
217 | the manifest node, user, and date. This object exposes a view into |
|
148 | the manifest node, user, and date. This object exposes a view into | |
218 | the parsed object. |
|
149 | the parsed object. | |
219 | """ |
|
150 | """ | |
220 |
|
151 | |||
221 | __slots__ = ( |
|
152 | __slots__ = ( | |
222 | '_offsets', |
|
153 | '_offsets', | |
223 | '_text', |
|
154 | '_text', | |
224 | '_sidedata', |
|
155 | '_sidedata', | |
225 | '_cpsd', |
|
156 | '_cpsd', | |
226 | '_changes', |
|
157 | '_changes', | |
227 | ) |
|
158 | ) | |
228 |
|
159 | |||
229 | def __new__(cls, cl, text, sidedata, cpsd): |
|
160 | def __new__(cls, cl, text, sidedata, cpsd): | |
230 | if not text: |
|
161 | if not text: | |
231 | return _changelogrevision(extra=_defaultextra, manifest=cl.nullid) |
|
162 | return _changelogrevision(extra=_defaultextra, manifest=cl.nullid) | |
232 |
|
163 | |||
233 | self = super(changelogrevision, cls).__new__(cls) |
|
164 | self = super(changelogrevision, cls).__new__(cls) | |
234 | # We could return here and implement the following as an __init__. |
|
165 | # We could return here and implement the following as an __init__. | |
235 | # But doing it here is equivalent and saves an extra function call. |
|
166 | # But doing it here is equivalent and saves an extra function call. | |
236 |
|
167 | |||
237 | # format used: |
|
168 | # format used: | |
238 | # nodeid\n : manifest node in ascii |
|
169 | # nodeid\n : manifest node in ascii | |
239 | # user\n : user, no \n or \r allowed |
|
170 | # user\n : user, no \n or \r allowed | |
240 | # time tz extra\n : date (time is int or float, timezone is int) |
|
171 | # time tz extra\n : date (time is int or float, timezone is int) | |
241 | # : extra is metadata, encoded and separated by '\0' |
|
172 | # : extra is metadata, encoded and separated by '\0' | |
242 | # : older versions ignore it |
|
173 | # : older versions ignore it | |
243 | # files\n\n : files modified by the cset, no \n or \r allowed |
|
174 | # files\n\n : files modified by the cset, no \n or \r allowed | |
244 | # (.*) : comment (free text, ideally utf-8) |
|
175 | # (.*) : comment (free text, ideally utf-8) | |
245 | # |
|
176 | # | |
246 | # changelog v0 doesn't use extra |
|
177 | # changelog v0 doesn't use extra | |
247 |
|
178 | |||
248 | nl1 = text.index(b'\n') |
|
179 | nl1 = text.index(b'\n') | |
249 | nl2 = text.index(b'\n', nl1 + 1) |
|
180 | nl2 = text.index(b'\n', nl1 + 1) | |
250 | nl3 = text.index(b'\n', nl2 + 1) |
|
181 | nl3 = text.index(b'\n', nl2 + 1) | |
251 |
|
182 | |||
252 | # The list of files may be empty. Which means nl3 is the first of the |
|
183 | # The list of files may be empty. Which means nl3 is the first of the | |
253 | # double newline that precedes the description. |
|
184 | # double newline that precedes the description. | |
254 | if text[nl3 + 1 : nl3 + 2] == b'\n': |
|
185 | if text[nl3 + 1 : nl3 + 2] == b'\n': | |
255 | doublenl = nl3 |
|
186 | doublenl = nl3 | |
256 | else: |
|
187 | else: | |
257 | doublenl = text.index(b'\n\n', nl3 + 1) |
|
188 | doublenl = text.index(b'\n\n', nl3 + 1) | |
258 |
|
189 | |||
259 | self._offsets = (nl1, nl2, nl3, doublenl) |
|
190 | self._offsets = (nl1, nl2, nl3, doublenl) | |
260 | self._text = text |
|
191 | self._text = text | |
261 | self._sidedata = sidedata |
|
192 | self._sidedata = sidedata | |
262 | self._cpsd = cpsd |
|
193 | self._cpsd = cpsd | |
263 | self._changes = None |
|
194 | self._changes = None | |
264 |
|
195 | |||
265 | return self |
|
196 | return self | |
266 |
|
197 | |||
267 | @property |
|
198 | @property | |
268 | def manifest(self): |
|
199 | def manifest(self): | |
269 | return bin(self._text[0 : self._offsets[0]]) |
|
200 | return bin(self._text[0 : self._offsets[0]]) | |
270 |
|
201 | |||
271 | @property |
|
202 | @property | |
272 | def user(self): |
|
203 | def user(self): | |
273 | off = self._offsets |
|
204 | off = self._offsets | |
274 | return encoding.tolocal(self._text[off[0] + 1 : off[1]]) |
|
205 | return encoding.tolocal(self._text[off[0] + 1 : off[1]]) | |
275 |
|
206 | |||
276 | @property |
|
207 | @property | |
277 | def _rawdate(self): |
|
208 | def _rawdate(self): | |
278 | off = self._offsets |
|
209 | off = self._offsets | |
279 | dateextra = self._text[off[1] + 1 : off[2]] |
|
210 | dateextra = self._text[off[1] + 1 : off[2]] | |
280 | return dateextra.split(b' ', 2)[0:2] |
|
211 | return dateextra.split(b' ', 2)[0:2] | |
281 |
|
212 | |||
282 | @property |
|
213 | @property | |
283 | def _rawextra(self): |
|
214 | def _rawextra(self): | |
284 | off = self._offsets |
|
215 | off = self._offsets | |
285 | dateextra = self._text[off[1] + 1 : off[2]] |
|
216 | dateextra = self._text[off[1] + 1 : off[2]] | |
286 | fields = dateextra.split(b' ', 2) |
|
217 | fields = dateextra.split(b' ', 2) | |
287 | if len(fields) != 3: |
|
218 | if len(fields) != 3: | |
288 | return None |
|
219 | return None | |
289 |
|
220 | |||
290 | return fields[2] |
|
221 | return fields[2] | |
291 |
|
222 | |||
292 | @property |
|
223 | @property | |
293 | def date(self): |
|
224 | def date(self): | |
294 | raw = self._rawdate |
|
225 | raw = self._rawdate | |
295 | time = float(raw[0]) |
|
226 | time = float(raw[0]) | |
296 | # Various tools did silly things with the timezone. |
|
227 | # Various tools did silly things with the timezone. | |
297 | try: |
|
228 | try: | |
298 | timezone = int(raw[1]) |
|
229 | timezone = int(raw[1]) | |
299 | except ValueError: |
|
230 | except ValueError: | |
300 | timezone = 0 |
|
231 | timezone = 0 | |
301 |
|
232 | |||
302 | return time, timezone |
|
233 | return time, timezone | |
303 |
|
234 | |||
304 | @property |
|
235 | @property | |
305 | def extra(self): |
|
236 | def extra(self): | |
306 | raw = self._rawextra |
|
237 | raw = self._rawextra | |
307 | if raw is None: |
|
238 | if raw is None: | |
308 | return _defaultextra |
|
239 | return _defaultextra | |
309 |
|
240 | |||
310 | return decodeextra(raw) |
|
241 | return decodeextra(raw) | |
311 |
|
242 | |||
312 | @property |
|
243 | @property | |
313 | def changes(self): |
|
244 | def changes(self): | |
314 | if self._changes is not None: |
|
245 | if self._changes is not None: | |
315 | return self._changes |
|
246 | return self._changes | |
316 | if self._cpsd: |
|
247 | if self._cpsd: | |
317 | changes = metadata.decode_files_sidedata(self._sidedata) |
|
248 | changes = metadata.decode_files_sidedata(self._sidedata) | |
318 | else: |
|
249 | else: | |
319 | changes = metadata.ChangingFiles( |
|
250 | changes = metadata.ChangingFiles( | |
320 | touched=self.files or (), |
|
251 | touched=self.files or (), | |
321 | added=self.filesadded or (), |
|
252 | added=self.filesadded or (), | |
322 | removed=self.filesremoved or (), |
|
253 | removed=self.filesremoved or (), | |
323 | p1_copies=self.p1copies or {}, |
|
254 | p1_copies=self.p1copies or {}, | |
324 | p2_copies=self.p2copies or {}, |
|
255 | p2_copies=self.p2copies or {}, | |
325 | ) |
|
256 | ) | |
326 | self._changes = changes |
|
257 | self._changes = changes | |
327 | return changes |
|
258 | return changes | |
328 |
|
259 | |||
329 | @property |
|
260 | @property | |
330 | def files(self): |
|
261 | def files(self): | |
331 | if self._cpsd: |
|
262 | if self._cpsd: | |
332 | return sorted(self.changes.touched) |
|
263 | return sorted(self.changes.touched) | |
333 | off = self._offsets |
|
264 | off = self._offsets | |
334 | if off[2] == off[3]: |
|
265 | if off[2] == off[3]: | |
335 | return [] |
|
266 | return [] | |
336 |
|
267 | |||
337 | return self._text[off[2] + 1 : off[3]].split(b'\n') |
|
268 | return self._text[off[2] + 1 : off[3]].split(b'\n') | |
338 |
|
269 | |||
339 | @property |
|
270 | @property | |
340 | def filesadded(self): |
|
271 | def filesadded(self): | |
341 | if self._cpsd: |
|
272 | if self._cpsd: | |
342 | return self.changes.added |
|
273 | return self.changes.added | |
343 | else: |
|
274 | else: | |
344 | rawindices = self.extra.get(b'filesadded') |
|
275 | rawindices = self.extra.get(b'filesadded') | |
345 | if rawindices is None: |
|
276 | if rawindices is None: | |
346 | return None |
|
277 | return None | |
347 | return metadata.decodefileindices(self.files, rawindices) |
|
278 | return metadata.decodefileindices(self.files, rawindices) | |
348 |
|
279 | |||
349 | @property |
|
280 | @property | |
350 | def filesremoved(self): |
|
281 | def filesremoved(self): | |
351 | if self._cpsd: |
|
282 | if self._cpsd: | |
352 | return self.changes.removed |
|
283 | return self.changes.removed | |
353 | else: |
|
284 | else: | |
354 | rawindices = self.extra.get(b'filesremoved') |
|
285 | rawindices = self.extra.get(b'filesremoved') | |
355 | if rawindices is None: |
|
286 | if rawindices is None: | |
356 | return None |
|
287 | return None | |
357 | return metadata.decodefileindices(self.files, rawindices) |
|
288 | return metadata.decodefileindices(self.files, rawindices) | |
358 |
|
289 | |||
359 | @property |
|
290 | @property | |
360 | def p1copies(self): |
|
291 | def p1copies(self): | |
361 | if self._cpsd: |
|
292 | if self._cpsd: | |
362 | return self.changes.copied_from_p1 |
|
293 | return self.changes.copied_from_p1 | |
363 | else: |
|
294 | else: | |
364 | rawcopies = self.extra.get(b'p1copies') |
|
295 | rawcopies = self.extra.get(b'p1copies') | |
365 | if rawcopies is None: |
|
296 | if rawcopies is None: | |
366 | return None |
|
297 | return None | |
367 | return metadata.decodecopies(self.files, rawcopies) |
|
298 | return metadata.decodecopies(self.files, rawcopies) | |
368 |
|
299 | |||
369 | @property |
|
300 | @property | |
370 | def p2copies(self): |
|
301 | def p2copies(self): | |
371 | if self._cpsd: |
|
302 | if self._cpsd: | |
372 | return self.changes.copied_from_p2 |
|
303 | return self.changes.copied_from_p2 | |
373 | else: |
|
304 | else: | |
374 | rawcopies = self.extra.get(b'p2copies') |
|
305 | rawcopies = self.extra.get(b'p2copies') | |
375 | if rawcopies is None: |
|
306 | if rawcopies is None: | |
376 | return None |
|
307 | return None | |
377 | return metadata.decodecopies(self.files, rawcopies) |
|
308 | return metadata.decodecopies(self.files, rawcopies) | |
378 |
|
309 | |||
379 | @property |
|
310 | @property | |
380 | def description(self): |
|
311 | def description(self): | |
381 | return encoding.tolocal(self._text[self._offsets[3] + 2 :]) |
|
312 | return encoding.tolocal(self._text[self._offsets[3] + 2 :]) | |
382 |
|
313 | |||
383 | @property |
|
314 | @property | |
384 | def branchinfo(self): |
|
315 | def branchinfo(self): | |
385 | extra = self.extra |
|
316 | extra = self.extra | |
386 | return encoding.tolocal(extra.get(b"branch")), b'close' in extra |
|
317 | return encoding.tolocal(extra.get(b"branch")), b'close' in extra | |
387 |
|
318 | |||
388 |
|
319 | |||
389 | class changelog(revlog.revlog): |
|
320 | class changelog(revlog.revlog): | |
390 | def __init__(self, opener, trypending=False, concurrencychecker=None): |
|
321 | def __init__(self, opener, trypending=False, concurrencychecker=None): | |
391 | """Load a changelog revlog using an opener. |
|
322 | """Load a changelog revlog using an opener. | |
392 |
|
323 | |||
393 | If ``trypending`` is true, we attempt to load the index from a |
|
324 | If ``trypending`` is true, we attempt to load the index from a | |
394 | ``00changelog.i.a`` file instead of the default ``00changelog.i``. |
|
325 | ``00changelog.i.a`` file instead of the default ``00changelog.i``. | |
395 | The ``00changelog.i.a`` file contains index (and possibly inline |
|
326 | The ``00changelog.i.a`` file contains index (and possibly inline | |
396 | revision) data for a transaction that hasn't been finalized yet. |
|
327 | revision) data for a transaction that hasn't been finalized yet. | |
397 | It exists in a separate file to facilitate readers (such as |
|
328 | It exists in a separate file to facilitate readers (such as | |
398 | hooks processes) accessing data before a transaction is finalized. |
|
329 | hooks processes) accessing data before a transaction is finalized. | |
399 |
|
330 | |||
400 | ``concurrencychecker`` will be passed to the revlog init function, see |
|
331 | ``concurrencychecker`` will be passed to the revlog init function, see | |
401 | the documentation there. |
|
332 | the documentation there. | |
402 | """ |
|
333 | """ | |
403 | revlog.revlog.__init__( |
|
334 | revlog.revlog.__init__( | |
404 | self, |
|
335 | self, | |
405 | opener, |
|
336 | opener, | |
406 | target=(revlog_constants.KIND_CHANGELOG, None), |
|
337 | target=(revlog_constants.KIND_CHANGELOG, None), | |
407 | radix=b'00changelog', |
|
338 | radix=b'00changelog', | |
408 | checkambig=True, |
|
339 | checkambig=True, | |
409 | mmaplargeindex=True, |
|
340 | mmaplargeindex=True, | |
410 | persistentnodemap=opener.options.get(b'persistent-nodemap', False), |
|
341 | persistentnodemap=opener.options.get(b'persistent-nodemap', False), | |
411 | concurrencychecker=concurrencychecker, |
|
342 | concurrencychecker=concurrencychecker, | |
412 | trypending=trypending, |
|
343 | trypending=trypending, | |
413 | ) |
|
344 | ) | |
414 |
|
345 | |||
415 | if self._initempty and (self._format_version == revlog.REVLOGV1): |
|
346 | if self._initempty and (self._format_version == revlog.REVLOGV1): | |
416 | # changelogs don't benefit from generaldelta. |
|
347 | # changelogs don't benefit from generaldelta. | |
417 |
|
348 | |||
418 | self._format_flags &= ~revlog.FLAG_GENERALDELTA |
|
349 | self._format_flags &= ~revlog.FLAG_GENERALDELTA | |
419 | self.delta_config.general_delta = False |
|
350 | self.delta_config.general_delta = False | |
420 |
|
351 | |||
421 | # Delta chains for changelogs tend to be very small because entries |
|
352 | # Delta chains for changelogs tend to be very small because entries | |
422 | # tend to be small and don't delta well with each. So disable delta |
|
353 | # tend to be small and don't delta well with each. So disable delta | |
423 | # chains. |
|
354 | # chains. | |
424 | self._storedeltachains = False |
|
355 | self._storedeltachains = False | |
425 |
|
356 | |||
426 | self._realopener = opener |
|
357 | self._realopener = opener | |
427 | self._delayed = False |
|
358 | self._delayed = False | |
428 | self._delaybuf = None |
|
359 | self._delaybuf = None | |
429 | self._divert = False |
|
360 | self._divert = False | |
430 | self._filteredrevs = frozenset() |
|
361 | self._filteredrevs = frozenset() | |
431 | self._filteredrevs_hashcache = {} |
|
362 | self._filteredrevs_hashcache = {} | |
432 | self._copiesstorage = opener.options.get(b'copies-storage') |
|
363 | self._copiesstorage = opener.options.get(b'copies-storage') | |
433 |
|
364 | |||
434 | @property |
|
365 | @property | |
435 | def filteredrevs(self): |
|
366 | def filteredrevs(self): | |
436 | return self._filteredrevs |
|
367 | return self._filteredrevs | |
437 |
|
368 | |||
438 | @filteredrevs.setter |
|
369 | @filteredrevs.setter | |
439 | def filteredrevs(self, val): |
|
370 | def filteredrevs(self, val): | |
440 | # Ensure all updates go through this function |
|
371 | # Ensure all updates go through this function | |
441 | assert isinstance(val, frozenset) |
|
372 | assert isinstance(val, frozenset) | |
442 | self._filteredrevs = val |
|
373 | self._filteredrevs = val | |
443 | self._filteredrevs_hashcache = {} |
|
374 | self._filteredrevs_hashcache = {} | |
444 |
|
375 | |||
445 | def _write_docket(self, tr): |
|
376 | def _write_docket(self, tr): | |
446 | if not self._delayed: |
|
377 | if not self._delayed: | |
447 | super(changelog, self)._write_docket(tr) |
|
378 | super(changelog, self)._write_docket(tr) | |
448 |
|
379 | |||
449 | def delayupdate(self, tr): |
|
380 | def delayupdate(self, tr): | |
450 | """delay visibility of index updates to other readers""" |
|
381 | """delay visibility of index updates to other readers""" | |
451 | assert not self._inner.is_open |
|
382 | assert not self._inner.is_open | |
452 | if self._docket is None and not self._delayed: |
|
383 | if self._docket is None and not self._delayed: | |
453 | if len(self) == 0: |
|
384 | if len(self) == 0: | |
454 | self._divert = True |
|
385 | self._divert = True | |
455 | if self._realopener.exists(self._indexfile + b'.a'): |
|
386 | if self._realopener.exists(self._indexfile + b'.a'): | |
456 | self._realopener.unlink(self._indexfile + b'.a') |
|
387 | self._realopener.unlink(self._indexfile + b'.a') | |
457 | self.opener = _divertopener(self._realopener, self._indexfile) |
|
388 | self.opener = _divertopener(self._realopener, self._indexfile) | |
458 | else: |
|
389 | else: | |
459 | self._delaybuf = [] |
|
390 | self._delaybuf = [] | |
460 | self.opener = _delayopener( |
|
391 | self.opener = _delayopener( | |
461 | self._realopener, self._indexfile, self._delaybuf |
|
392 | self._realopener, self._indexfile, self._delaybuf | |
462 | ) |
|
393 | ) | |
463 | self._inner.opener = self.opener |
|
394 | self._inner.opener = self.opener | |
464 | self._inner._segmentfile.opener = self.opener |
|
395 | self._inner._segmentfile.opener = self.opener | |
465 | self._inner._segmentfile_sidedata.opener = self.opener |
|
396 | self._inner._segmentfile_sidedata.opener = self.opener | |
466 | self._delayed = True |
|
397 | self._delayed = True | |
467 | tr.addpending(b'cl-%i' % id(self), self._writepending) |
|
398 | tr.addpending(b'cl-%i' % id(self), self._writepending) | |
468 | tr.addfinalize(b'cl-%i' % id(self), self._finalize) |
|
399 | tr.addfinalize(b'cl-%i' % id(self), self._finalize) | |
469 |
|
400 | |||
470 | def _finalize(self, tr): |
|
401 | def _finalize(self, tr): | |
471 | """finalize index updates""" |
|
402 | """finalize index updates""" | |
472 | assert not self._inner.is_open |
|
403 | assert not self._inner.is_open | |
473 | self._delayed = False |
|
404 | self._delayed = False | |
474 | self.opener = self._realopener |
|
405 | self.opener = self._realopener | |
475 | self._inner.opener = self.opener |
|
406 | self._inner.opener = self.opener | |
476 | self._inner._segmentfile.opener = self.opener |
|
407 | self._inner._segmentfile.opener = self.opener | |
477 | self._inner._segmentfile_sidedata.opener = self.opener |
|
408 | self._inner._segmentfile_sidedata.opener = self.opener | |
478 | # move redirected index data back into place |
|
409 | # move redirected index data back into place | |
479 | if self._docket is not None: |
|
410 | if self._docket is not None: | |
480 | self._write_docket(tr) |
|
411 | self._write_docket(tr) | |
481 | elif self._divert: |
|
412 | elif self._divert: | |
482 | assert not self._delaybuf |
|
413 | assert not self._delaybuf | |
483 | tmpname = self._indexfile + b".a" |
|
414 | tmpname = self._indexfile + b".a" | |
484 | nfile = self.opener.open(tmpname) |
|
415 | nfile = self.opener.open(tmpname) | |
485 | nfile.close() |
|
416 | nfile.close() | |
486 | self.opener.rename(tmpname, self._indexfile, checkambig=True) |
|
417 | self.opener.rename(tmpname, self._indexfile, checkambig=True) | |
487 | elif self._delaybuf: |
|
418 | elif self._delaybuf: | |
488 | fp = self.opener(self._indexfile, b'a', checkambig=True) |
|
419 | fp = self.opener(self._indexfile, b'a', checkambig=True) | |
489 | fp.write(b"".join(self._delaybuf)) |
|
420 | fp.write(b"".join(self._delaybuf)) | |
490 | fp.close() |
|
421 | fp.close() | |
491 | self._delaybuf = None |
|
422 | self._delaybuf = None | |
492 | self._divert = False |
|
423 | self._divert = False | |
493 | # split when we're done |
|
424 | # split when we're done | |
494 | self._enforceinlinesize(tr, side_write=False) |
|
425 | self._enforceinlinesize(tr, side_write=False) | |
495 |
|
426 | |||
496 | def _writepending(self, tr): |
|
427 | def _writepending(self, tr): | |
497 | """create a file containing the unfinalized state for |
|
428 | """create a file containing the unfinalized state for | |
498 | pretxnchangegroup""" |
|
429 | pretxnchangegroup""" | |
499 | assert not self._inner.is_open |
|
430 | assert not self._inner.is_open | |
500 | if self._docket: |
|
431 | if self._docket: | |
501 | return self._docket.write(tr, pending=True) |
|
432 | return self._docket.write(tr, pending=True) | |
502 | if self._delaybuf: |
|
433 | if self._delaybuf: | |
503 | # make a temporary copy of the index |
|
434 | # make a temporary copy of the index | |
504 | fp1 = self._realopener(self._indexfile) |
|
435 | fp1 = self._realopener(self._indexfile) | |
505 | pendingfilename = self._indexfile + b".a" |
|
436 | pendingfilename = self._indexfile + b".a" | |
506 | # register as a temp file to ensure cleanup on failure |
|
437 | # register as a temp file to ensure cleanup on failure | |
507 | tr.registertmp(pendingfilename) |
|
438 | tr.registertmp(pendingfilename) | |
508 | # write existing data |
|
439 | # write existing data | |
509 | fp2 = self._realopener(pendingfilename, b"w") |
|
440 | fp2 = self._realopener(pendingfilename, b"w") | |
510 | fp2.write(fp1.read()) |
|
441 | fp2.write(fp1.read()) | |
511 | # add pending data |
|
442 | # add pending data | |
512 | fp2.write(b"".join(self._delaybuf)) |
|
443 | fp2.write(b"".join(self._delaybuf)) | |
513 | fp2.close() |
|
444 | fp2.close() | |
514 | # switch modes so finalize can simply rename |
|
445 | # switch modes so finalize can simply rename | |
515 | self._delaybuf = None |
|
446 | self._delaybuf = None | |
516 | self._divert = True |
|
447 | self._divert = True | |
517 | self.opener = _divertopener(self._realopener, self._indexfile) |
|
448 | self.opener = _divertopener(self._realopener, self._indexfile) | |
518 | self._inner.opener = self.opener |
|
449 | self._inner.opener = self.opener | |
519 | self._inner._segmentfile.opener = self.opener |
|
450 | self._inner._segmentfile.opener = self.opener | |
520 | self._inner._segmentfile_sidedata.opener = self.opener |
|
451 | self._inner._segmentfile_sidedata.opener = self.opener | |
521 |
|
452 | |||
522 | if self._divert: |
|
453 | if self._divert: | |
523 | return True |
|
454 | return True | |
524 |
|
455 | |||
525 | return False |
|
456 | return False | |
526 |
|
457 | |||
527 | def _enforceinlinesize(self, tr, side_write=True): |
|
458 | def _enforceinlinesize(self, tr, side_write=True): | |
528 | if not self._delayed: |
|
459 | if not self._delayed: | |
529 | revlog.revlog._enforceinlinesize(self, tr, side_write=side_write) |
|
460 | revlog.revlog._enforceinlinesize(self, tr, side_write=side_write) | |
530 |
|
461 | |||
531 | def read(self, nodeorrev): |
|
462 | def read(self, nodeorrev): | |
532 | """Obtain data from a parsed changelog revision. |
|
463 | """Obtain data from a parsed changelog revision. | |
533 |
|
464 | |||
534 | Returns a 6-tuple of: |
|
465 | Returns a 6-tuple of: | |
535 |
|
466 | |||
536 | - manifest node in binary |
|
467 | - manifest node in binary | |
537 | - author/user as a localstr |
|
468 | - author/user as a localstr | |
538 | - date as a 2-tuple of (time, timezone) |
|
469 | - date as a 2-tuple of (time, timezone) | |
539 | - list of files |
|
470 | - list of files | |
540 | - commit message as a localstr |
|
471 | - commit message as a localstr | |
541 | - dict of extra metadata |
|
472 | - dict of extra metadata | |
542 |
|
473 | |||
543 | Unless you need to access all fields, consider calling |
|
474 | Unless you need to access all fields, consider calling | |
544 | ``changelogrevision`` instead, as it is faster for partial object |
|
475 | ``changelogrevision`` instead, as it is faster for partial object | |
545 | access. |
|
476 | access. | |
546 | """ |
|
477 | """ | |
547 | d = self._revisiondata(nodeorrev) |
|
478 | d = self._revisiondata(nodeorrev) | |
548 | sidedata = self.sidedata(nodeorrev) |
|
479 | sidedata = self.sidedata(nodeorrev) | |
549 | copy_sd = self._copiesstorage == b'changeset-sidedata' |
|
480 | copy_sd = self._copiesstorage == b'changeset-sidedata' | |
550 | c = changelogrevision(self, d, sidedata, copy_sd) |
|
481 | c = changelogrevision(self, d, sidedata, copy_sd) | |
551 | return (c.manifest, c.user, c.date, c.files, c.description, c.extra) |
|
482 | return (c.manifest, c.user, c.date, c.files, c.description, c.extra) | |
552 |
|
483 | |||
553 | def changelogrevision(self, nodeorrev): |
|
484 | def changelogrevision(self, nodeorrev): | |
554 | """Obtain a ``changelogrevision`` for a node or revision.""" |
|
485 | """Obtain a ``changelogrevision`` for a node or revision.""" | |
555 | text = self._revisiondata(nodeorrev) |
|
486 | text = self._revisiondata(nodeorrev) | |
556 | sidedata = self.sidedata(nodeorrev) |
|
487 | sidedata = self.sidedata(nodeorrev) | |
557 | return changelogrevision( |
|
488 | return changelogrevision( | |
558 | self, text, sidedata, self._copiesstorage == b'changeset-sidedata' |
|
489 | self, text, sidedata, self._copiesstorage == b'changeset-sidedata' | |
559 | ) |
|
490 | ) | |
560 |
|
491 | |||
561 | def readfiles(self, nodeorrev): |
|
492 | def readfiles(self, nodeorrev): | |
562 | """ |
|
493 | """ | |
563 | short version of read that only returns the files modified by the cset |
|
494 | short version of read that only returns the files modified by the cset | |
564 | """ |
|
495 | """ | |
565 | text = self.revision(nodeorrev) |
|
496 | text = self.revision(nodeorrev) | |
566 | if not text: |
|
497 | if not text: | |
567 | return [] |
|
498 | return [] | |
568 | last = text.index(b"\n\n") |
|
499 | last = text.index(b"\n\n") | |
569 | l = text[:last].split(b'\n') |
|
500 | l = text[:last].split(b'\n') | |
570 | return l[3:] |
|
501 | return l[3:] | |
571 |
|
502 | |||
572 | def add( |
|
503 | def add( | |
573 | self, |
|
504 | self, | |
574 | manifest, |
|
505 | manifest, | |
575 | files, |
|
506 | files, | |
576 | desc, |
|
507 | desc, | |
577 | transaction, |
|
508 | transaction, | |
578 | p1, |
|
509 | p1, | |
579 | p2, |
|
510 | p2, | |
580 | user, |
|
511 | user, | |
581 | date=None, |
|
512 | date=None, | |
582 | extra=None, |
|
513 | extra=None, | |
583 | ): |
|
514 | ): | |
584 | # Convert to UTF-8 encoded bytestrings as the very first |
|
515 | # Convert to UTF-8 encoded bytestrings as the very first | |
585 | # thing: calling any method on a localstr object will turn it |
|
516 | # thing: calling any method on a localstr object will turn it | |
586 | # into a str object and the cached UTF-8 string is thus lost. |
|
517 | # into a str object and the cached UTF-8 string is thus lost. | |
587 | user, desc = encoding.fromlocal(user), encoding.fromlocal(desc) |
|
518 | user, desc = encoding.fromlocal(user), encoding.fromlocal(desc) | |
588 |
|
519 | |||
589 | user = user.strip() |
|
520 | user = user.strip() | |
590 | # An empty username or a username with a "\n" will make the |
|
521 | # An empty username or a username with a "\n" will make the | |
591 | # revision text contain two "\n\n" sequences -> corrupt |
|
522 | # revision text contain two "\n\n" sequences -> corrupt | |
592 | # repository since read cannot unpack the revision. |
|
523 | # repository since read cannot unpack the revision. | |
593 | if not user: |
|
524 | if not user: | |
594 | raise error.StorageError(_(b"empty username")) |
|
525 | raise error.StorageError(_(b"empty username")) | |
595 | if b"\n" in user: |
|
526 | if b"\n" in user: | |
596 | raise error.StorageError( |
|
527 | raise error.StorageError( | |
597 | _(b"username %r contains a newline") % pycompat.bytestr(user) |
|
528 | _(b"username %r contains a newline") % pycompat.bytestr(user) | |
598 | ) |
|
529 | ) | |
599 |
|
530 | |||
600 | desc = stripdesc(desc) |
|
531 | desc = stripdesc(desc) | |
601 |
|
532 | |||
602 | if date: |
|
533 | if date: | |
603 | parseddate = b"%d %d" % dateutil.parsedate(date) |
|
534 | parseddate = b"%d %d" % dateutil.parsedate(date) | |
604 | else: |
|
535 | else: | |
605 | parseddate = b"%d %d" % dateutil.makedate() |
|
536 | parseddate = b"%d %d" % dateutil.makedate() | |
606 | if extra: |
|
537 | if extra: | |
607 | branch = extra.get(b"branch") |
|
538 | branch = extra.get(b"branch") | |
608 | if branch in (b"default", b""): |
|
539 | if branch in (b"default", b""): | |
609 | del extra[b"branch"] |
|
540 | del extra[b"branch"] | |
610 | elif branch in (b".", b"null", b"tip"): |
|
541 | elif branch in (b".", b"null", b"tip"): | |
611 | raise error.StorageError( |
|
542 | raise error.StorageError( | |
612 | _(b'the name \'%s\' is reserved') % branch |
|
543 | _(b'the name \'%s\' is reserved') % branch | |
613 | ) |
|
544 | ) | |
614 | sortedfiles = sorted(files.touched) |
|
545 | sortedfiles = sorted(files.touched) | |
615 | flags = 0 |
|
546 | flags = 0 | |
616 | sidedata = None |
|
547 | sidedata = None | |
617 | if self._copiesstorage == b'changeset-sidedata': |
|
548 | if self._copiesstorage == b'changeset-sidedata': | |
618 | if files.has_copies_info: |
|
549 | if files.has_copies_info: | |
619 | flags |= flagutil.REVIDX_HASCOPIESINFO |
|
550 | flags |= flagutil.REVIDX_HASCOPIESINFO | |
620 | sidedata = metadata.encode_files_sidedata(files) |
|
551 | sidedata = metadata.encode_files_sidedata(files) | |
621 |
|
552 | |||
622 | if extra: |
|
553 | if extra: | |
623 | extra = encodeextra(extra) |
|
554 | extra = encodeextra(extra) | |
624 | parseddate = b"%s %s" % (parseddate, extra) |
|
555 | parseddate = b"%s %s" % (parseddate, extra) | |
625 | l = [hex(manifest), user, parseddate] + sortedfiles + [b"", desc] |
|
556 | l = [hex(manifest), user, parseddate] + sortedfiles + [b"", desc] | |
626 | text = b"\n".join(l) |
|
557 | text = b"\n".join(l) | |
627 | rev = self.addrevision( |
|
558 | rev = self.addrevision( | |
628 | text, transaction, len(self), p1, p2, sidedata=sidedata, flags=flags |
|
559 | text, transaction, len(self), p1, p2, sidedata=sidedata, flags=flags | |
629 | ) |
|
560 | ) | |
630 | return self.node(rev) |
|
561 | return self.node(rev) | |
631 |
|
562 | |||
632 | def branchinfo(self, rev): |
|
563 | def branchinfo(self, rev): | |
633 | """return the branch name and open/close state of a revision |
|
564 | """return the branch name and open/close state of a revision | |
634 |
|
565 | |||
635 | This function exists because creating a changectx object |
|
566 | This function exists because creating a changectx object | |
636 | just to access this is costly.""" |
|
567 | just to access this is costly.""" | |
637 | return self.changelogrevision(rev).branchinfo |
|
568 | return self.changelogrevision(rev).branchinfo | |
638 |
|
569 | |||
639 | def _nodeduplicatecallback(self, transaction, rev): |
|
570 | def _nodeduplicatecallback(self, transaction, rev): | |
640 | # keep track of revisions that got "re-added", eg: unbunde of know rev. |
|
571 | # keep track of revisions that got "re-added", eg: unbunde of know rev. | |
641 | # |
|
572 | # | |
642 | # We track them in a list to preserve their order from the source bundle |
|
573 | # We track them in a list to preserve their order from the source bundle | |
643 | duplicates = transaction.changes.setdefault(b'revduplicates', []) |
|
574 | duplicates = transaction.changes.setdefault(b'revduplicates', []) | |
644 | duplicates.append(rev) |
|
575 | duplicates.append(rev) |
@@ -1,164 +1,234 b'' | |||||
1 | # Copyright Mercurial Contributors |
|
1 | # Copyright Mercurial Contributors | |
2 | # |
|
2 | # | |
3 | # This software may be used and distributed according to the terms of the |
|
3 | # This software may be used and distributed according to the terms of the | |
4 | # GNU General Public License version 2 or any later version. |
|
4 | # GNU General Public License version 2 or any later version. | |
5 |
|
5 | |||
6 | import contextlib |
|
6 | import contextlib | |
7 |
|
7 | |||
8 | from ..i18n import _ |
|
8 | from ..i18n import _ | |
9 | from .. import ( |
|
9 | from .. import ( | |
10 | error, |
|
10 | error, | |
11 | util, |
|
11 | util, | |
12 | ) |
|
12 | ) | |
13 |
|
13 | |||
14 |
|
14 | |||
15 | _MAX_CACHED_CHUNK_SIZE = 1048576 # 1 MiB |
|
15 | _MAX_CACHED_CHUNK_SIZE = 1048576 # 1 MiB | |
16 |
|
16 | |||
17 | PARTIAL_READ_MSG = _( |
|
17 | PARTIAL_READ_MSG = _( | |
18 | b'partial read of revlog %s; expected %d bytes from offset %d, got %d' |
|
18 | b'partial read of revlog %s; expected %d bytes from offset %d, got %d' | |
19 | ) |
|
19 | ) | |
20 |
|
20 | |||
21 |
|
21 | |||
22 | def _is_power_of_two(n): |
|
22 | def _is_power_of_two(n): | |
23 | return (n & (n - 1) == 0) and n != 0 |
|
23 | return (n & (n - 1) == 0) and n != 0 | |
24 |
|
24 | |||
25 |
|
25 | |||
|
26 | class appender: | |||
|
27 | """the changelog index must be updated last on disk, so we use this class | |||
|
28 | to delay writes to it""" | |||
|
29 | ||||
|
30 | def __init__(self, vfs, name, mode, buf): | |||
|
31 | self.data = buf | |||
|
32 | fp = vfs(name, mode) | |||
|
33 | self.fp = fp | |||
|
34 | self.offset = fp.tell() | |||
|
35 | self.size = vfs.fstat(fp).st_size | |||
|
36 | self._end = self.size | |||
|
37 | ||||
|
38 | def end(self): | |||
|
39 | return self._end | |||
|
40 | ||||
|
41 | def tell(self): | |||
|
42 | return self.offset | |||
|
43 | ||||
|
44 | def flush(self): | |||
|
45 | pass | |||
|
46 | ||||
|
47 | @property | |||
|
48 | def closed(self): | |||
|
49 | return self.fp.closed | |||
|
50 | ||||
|
51 | def close(self): | |||
|
52 | self.fp.close() | |||
|
53 | ||||
|
54 | def seek(self, offset, whence=0): | |||
|
55 | '''virtual file offset spans real file and data''' | |||
|
56 | if whence == 0: | |||
|
57 | self.offset = offset | |||
|
58 | elif whence == 1: | |||
|
59 | self.offset += offset | |||
|
60 | elif whence == 2: | |||
|
61 | self.offset = self.end() + offset | |||
|
62 | if self.offset < self.size: | |||
|
63 | self.fp.seek(self.offset) | |||
|
64 | ||||
|
65 | def read(self, count=-1): | |||
|
66 | '''only trick here is reads that span real file and data''' | |||
|
67 | ret = b"" | |||
|
68 | if self.offset < self.size: | |||
|
69 | s = self.fp.read(count) | |||
|
70 | ret = s | |||
|
71 | self.offset += len(s) | |||
|
72 | if count > 0: | |||
|
73 | count -= len(s) | |||
|
74 | if count != 0: | |||
|
75 | doff = self.offset - self.size | |||
|
76 | self.data.insert(0, b"".join(self.data)) | |||
|
77 | del self.data[1:] | |||
|
78 | s = self.data[0][doff : doff + count] | |||
|
79 | self.offset += len(s) | |||
|
80 | ret += s | |||
|
81 | return ret | |||
|
82 | ||||
|
83 | def write(self, s): | |||
|
84 | self.data.append(bytes(s)) | |||
|
85 | self.offset += len(s) | |||
|
86 | self._end += len(s) | |||
|
87 | ||||
|
88 | def __enter__(self): | |||
|
89 | self.fp.__enter__() | |||
|
90 | return self | |||
|
91 | ||||
|
92 | def __exit__(self, *args): | |||
|
93 | return self.fp.__exit__(*args) | |||
|
94 | ||||
|
95 | ||||
26 | class randomaccessfile: |
|
96 | class randomaccessfile: | |
27 | """Accessing arbitrary chuncks of data within a file, with some caching""" |
|
97 | """Accessing arbitrary chuncks of data within a file, with some caching""" | |
28 |
|
98 | |||
29 | def __init__( |
|
99 | def __init__( | |
30 | self, |
|
100 | self, | |
31 | opener, |
|
101 | opener, | |
32 | filename, |
|
102 | filename, | |
33 | default_cached_chunk_size, |
|
103 | default_cached_chunk_size, | |
34 | initial_cache=None, |
|
104 | initial_cache=None, | |
35 | ): |
|
105 | ): | |
36 | # Required by bitwise manipulation below |
|
106 | # Required by bitwise manipulation below | |
37 | assert _is_power_of_two(default_cached_chunk_size) |
|
107 | assert _is_power_of_two(default_cached_chunk_size) | |
38 |
|
108 | |||
39 | self.opener = opener |
|
109 | self.opener = opener | |
40 | self.filename = filename |
|
110 | self.filename = filename | |
41 | self.default_cached_chunk_size = default_cached_chunk_size |
|
111 | self.default_cached_chunk_size = default_cached_chunk_size | |
42 | self.writing_handle = None # This is set from revlog.py |
|
112 | self.writing_handle = None # This is set from revlog.py | |
43 | self.reading_handle = None |
|
113 | self.reading_handle = None | |
44 | self._cached_chunk = b'' |
|
114 | self._cached_chunk = b'' | |
45 | self._cached_chunk_position = 0 # Offset from the start of the file |
|
115 | self._cached_chunk_position = 0 # Offset from the start of the file | |
46 | if initial_cache: |
|
116 | if initial_cache: | |
47 | self._cached_chunk_position, self._cached_chunk = initial_cache |
|
117 | self._cached_chunk_position, self._cached_chunk = initial_cache | |
48 |
|
118 | |||
49 | def clear_cache(self): |
|
119 | def clear_cache(self): | |
50 | self._cached_chunk = b'' |
|
120 | self._cached_chunk = b'' | |
51 | self._cached_chunk_position = 0 |
|
121 | self._cached_chunk_position = 0 | |
52 |
|
122 | |||
53 | @property |
|
123 | @property | |
54 | def is_open(self): |
|
124 | def is_open(self): | |
55 | """True if any file handle is being held |
|
125 | """True if any file handle is being held | |
56 |
|
126 | |||
57 | Used for assert and debug in the python code""" |
|
127 | Used for assert and debug in the python code""" | |
58 | return ( |
|
128 | return ( | |
59 | self.reading_handle is not None or self.writing_handle is not None |
|
129 | self.reading_handle is not None or self.writing_handle is not None | |
60 | ) |
|
130 | ) | |
61 |
|
131 | |||
62 | def _open(self, mode=b'r'): |
|
132 | def _open(self, mode=b'r'): | |
63 | """Return a file object""" |
|
133 | """Return a file object""" | |
64 | return self.opener(self.filename, mode=mode) |
|
134 | return self.opener(self.filename, mode=mode) | |
65 |
|
135 | |||
66 | @contextlib.contextmanager |
|
136 | @contextlib.contextmanager | |
67 | def _read_handle(self): |
|
137 | def _read_handle(self): | |
68 | """File object suitable for reading data""" |
|
138 | """File object suitable for reading data""" | |
69 | # Use a file handle being actively used for writes, if available. |
|
139 | # Use a file handle being actively used for writes, if available. | |
70 | # There is some danger to doing this because reads will seek the |
|
140 | # There is some danger to doing this because reads will seek the | |
71 | # file. However, revlog._writeentry performs a SEEK_END before all |
|
141 | # file. However, revlog._writeentry performs a SEEK_END before all | |
72 | # writes, so we should be safe. |
|
142 | # writes, so we should be safe. | |
73 | if self.writing_handle: |
|
143 | if self.writing_handle: | |
74 | yield self.writing_handle |
|
144 | yield self.writing_handle | |
75 |
|
145 | |||
76 | elif self.reading_handle: |
|
146 | elif self.reading_handle: | |
77 | yield self.reading_handle |
|
147 | yield self.reading_handle | |
78 |
|
148 | |||
79 | # Otherwise open a new file handle. |
|
149 | # Otherwise open a new file handle. | |
80 | else: |
|
150 | else: | |
81 | with self._open() as fp: |
|
151 | with self._open() as fp: | |
82 | yield fp |
|
152 | yield fp | |
83 |
|
153 | |||
84 | @contextlib.contextmanager |
|
154 | @contextlib.contextmanager | |
85 | def reading(self): |
|
155 | def reading(self): | |
86 | """Context manager that keeps the file open for reading""" |
|
156 | """Context manager that keeps the file open for reading""" | |
87 | if ( |
|
157 | if ( | |
88 | self.reading_handle is None |
|
158 | self.reading_handle is None | |
89 | and self.writing_handle is None |
|
159 | and self.writing_handle is None | |
90 | and self.filename is not None |
|
160 | and self.filename is not None | |
91 | ): |
|
161 | ): | |
92 | with self._open() as fp: |
|
162 | with self._open() as fp: | |
93 | self.reading_handle = fp |
|
163 | self.reading_handle = fp | |
94 | try: |
|
164 | try: | |
95 | yield |
|
165 | yield | |
96 | finally: |
|
166 | finally: | |
97 | self.reading_handle = None |
|
167 | self.reading_handle = None | |
98 | else: |
|
168 | else: | |
99 | yield |
|
169 | yield | |
100 |
|
170 | |||
101 | def read_chunk(self, offset, length): |
|
171 | def read_chunk(self, offset, length): | |
102 | """Read a chunk of bytes from the file. |
|
172 | """Read a chunk of bytes from the file. | |
103 |
|
173 | |||
104 | Accepts an absolute offset, length to read, and an optional existing |
|
174 | Accepts an absolute offset, length to read, and an optional existing | |
105 | file handle to read from. |
|
175 | file handle to read from. | |
106 |
|
176 | |||
107 | If an existing file handle is passed, it will be seeked and the |
|
177 | If an existing file handle is passed, it will be seeked and the | |
108 | original seek position will NOT be restored. |
|
178 | original seek position will NOT be restored. | |
109 |
|
179 | |||
110 | Returns a str or buffer of raw byte data. |
|
180 | Returns a str or buffer of raw byte data. | |
111 |
|
181 | |||
112 | Raises if the requested number of bytes could not be read. |
|
182 | Raises if the requested number of bytes could not be read. | |
113 | """ |
|
183 | """ | |
114 | end = offset + length |
|
184 | end = offset + length | |
115 | cache_start = self._cached_chunk_position |
|
185 | cache_start = self._cached_chunk_position | |
116 | cache_end = cache_start + len(self._cached_chunk) |
|
186 | cache_end = cache_start + len(self._cached_chunk) | |
117 | # Is the requested chunk within the cache? |
|
187 | # Is the requested chunk within the cache? | |
118 | if cache_start <= offset and end <= cache_end: |
|
188 | if cache_start <= offset and end <= cache_end: | |
119 | if cache_start == offset and end == cache_end: |
|
189 | if cache_start == offset and end == cache_end: | |
120 | return self._cached_chunk # avoid a copy |
|
190 | return self._cached_chunk # avoid a copy | |
121 | relative_start = offset - cache_start |
|
191 | relative_start = offset - cache_start | |
122 | return util.buffer(self._cached_chunk, relative_start, length) |
|
192 | return util.buffer(self._cached_chunk, relative_start, length) | |
123 |
|
193 | |||
124 | return self._read_and_update_cache(offset, length) |
|
194 | return self._read_and_update_cache(offset, length) | |
125 |
|
195 | |||
126 | def _read_and_update_cache(self, offset, length): |
|
196 | def _read_and_update_cache(self, offset, length): | |
127 | # Cache data both forward and backward around the requested |
|
197 | # Cache data both forward and backward around the requested | |
128 | # data, in a fixed size window. This helps speed up operations |
|
198 | # data, in a fixed size window. This helps speed up operations | |
129 | # involving reading the revlog backwards. |
|
199 | # involving reading the revlog backwards. | |
130 | real_offset = offset & ~(self.default_cached_chunk_size - 1) |
|
200 | real_offset = offset & ~(self.default_cached_chunk_size - 1) | |
131 | real_length = ( |
|
201 | real_length = ( | |
132 | (offset + length + self.default_cached_chunk_size) |
|
202 | (offset + length + self.default_cached_chunk_size) | |
133 | & ~(self.default_cached_chunk_size - 1) |
|
203 | & ~(self.default_cached_chunk_size - 1) | |
134 | ) - real_offset |
|
204 | ) - real_offset | |
135 | with self._read_handle() as file_obj: |
|
205 | with self._read_handle() as file_obj: | |
136 | file_obj.seek(real_offset) |
|
206 | file_obj.seek(real_offset) | |
137 | data = file_obj.read(real_length) |
|
207 | data = file_obj.read(real_length) | |
138 |
|
208 | |||
139 | self._add_cached_chunk(real_offset, data) |
|
209 | self._add_cached_chunk(real_offset, data) | |
140 |
|
210 | |||
141 | relative_offset = offset - real_offset |
|
211 | relative_offset = offset - real_offset | |
142 | got = len(data) - relative_offset |
|
212 | got = len(data) - relative_offset | |
143 | if got < length: |
|
213 | if got < length: | |
144 | message = PARTIAL_READ_MSG % (self.filename, length, offset, got) |
|
214 | message = PARTIAL_READ_MSG % (self.filename, length, offset, got) | |
145 | raise error.RevlogError(message) |
|
215 | raise error.RevlogError(message) | |
146 |
|
216 | |||
147 | if offset != real_offset or real_length != length: |
|
217 | if offset != real_offset or real_length != length: | |
148 | return util.buffer(data, relative_offset, length) |
|
218 | return util.buffer(data, relative_offset, length) | |
149 | return data |
|
219 | return data | |
150 |
|
220 | |||
151 | def _add_cached_chunk(self, offset, data): |
|
221 | def _add_cached_chunk(self, offset, data): | |
152 | """Add to or replace the cached data chunk. |
|
222 | """Add to or replace the cached data chunk. | |
153 |
|
223 | |||
154 | Accepts an absolute offset and the data that is at that location. |
|
224 | Accepts an absolute offset and the data that is at that location. | |
155 | """ |
|
225 | """ | |
156 | if ( |
|
226 | if ( | |
157 | self._cached_chunk_position + len(self._cached_chunk) == offset |
|
227 | self._cached_chunk_position + len(self._cached_chunk) == offset | |
158 | and len(self._cached_chunk) + len(data) < _MAX_CACHED_CHUNK_SIZE |
|
228 | and len(self._cached_chunk) + len(data) < _MAX_CACHED_CHUNK_SIZE | |
159 | ): |
|
229 | ): | |
160 | # add to existing cache |
|
230 | # add to existing cache | |
161 | self._cached_chunk += data |
|
231 | self._cached_chunk += data | |
162 | else: |
|
232 | else: | |
163 | self._cached_chunk = data |
|
233 | self._cached_chunk = data | |
164 | self._cached_chunk_position = offset |
|
234 | self._cached_chunk_position = offset |
General Comments 0
You need to be logged in to leave comments.
Login now