Mercurial > hg > mercurial-crew-with-dirclash
annotate mercurial/revlog.py @ 1450:199bb2b4ed4a
Add Benoit to CONTRIBUTORS
author | Matt Mackall <mpm@selenic.com> |
---|---|
date | Tue, 25 Oct 2005 16:31:38 -0700 |
parents | bc3e66edb04c |
children | 0847c45ffee6 |
rev | line source |
---|---|
1083 | 1 """ |
2 revlog.py - storage back-end for mercurial | |
3 | |
4 This provides efficient delta storage with O(1) retrieve and append | |
5 and O(changes) merge between branches | |
6 | |
7 Copyright 2005 Matt Mackall <mpm@selenic.com> | |
8 | |
9 This software may be used and distributed according to the terms | |
10 of the GNU General Public License, incorporated herein by reference. | |
11 """ | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
12 |
1089 | 13 from node import * |
1400
cf9a1233738a
i18n first part: make '_' available for files who need it
Benoit Boissinot <benoit.boissinot@ens-lyon.org
parents:
1393
diff
changeset
|
14 from i18n import gettext as _ |
1322
b3d44e9b3092
Make revlog constructor more discerning in its treatment of errors.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1232
diff
changeset
|
15 from demandload import demandload |
1325
57220daf40e9
Move urllib error handling from revlog into statichttprepo, where it belongs.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1322
diff
changeset
|
16 demandload(globals(), "binascii errno heapq mdiff sha struct zlib") |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
17 |
1091
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
18 def hash(text, p1, p2): |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
19 """generate a hash from the given text and its parent hashes |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
20 |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
21 This hash combines both the current file contents and its history |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
22 in a manner that makes it easy to distinguish nodes with the same |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
23 content in the revision graph. |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
24 """ |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
25 l = [p1, p2] |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
26 l.sort() |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
27 s = sha.new(l[0]) |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
28 s.update(l[1]) |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
29 s.update(text) |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
30 return s.digest() |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
31 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
32 def compress(text): |
1083 | 33 """ generate a possibly-compressed representation of text """ |
112 | 34 if not text: return text |
35 if len(text) < 44: | |
36 if text[0] == '\0': return text | |
37 return 'u' + text | |
38 bin = zlib.compress(text) | |
39 if len(bin) > len(text): | |
40 if text[0] == '\0': return text | |
41 return 'u' + text | |
42 return bin | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
43 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
44 def decompress(bin): |
1083 | 45 """ decompress the given input """ |
112 | 46 if not bin: return bin |
47 t = bin[0] | |
48 if t == '\0': return bin | |
49 if t == 'x': return zlib.decompress(bin) | |
50 if t == 'u': return bin[1:] | |
1402
9d2c2e6b32b5
i18n part2: use '_' for all strings who are part of the user interface
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents:
1400
diff
changeset
|
51 raise RevlogError(_("unknown compression type %s") % t) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
52 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
53 indexformat = ">4l20s20s20s" |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
54 |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
55 class lazyparser: |
1083 | 56 """ |
57 this class avoids the need to parse the entirety of large indices | |
58 | |
59 By default we parse and load 1000 entries at a time. | |
60 | |
61 If no position is specified, we load the whole index, and replace | |
62 the lazy objects in revlog with the underlying objects for | |
63 efficiency in cases where we look at most of the nodes. | |
64 """ | |
323 | 65 def __init__(self, data, revlog): |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
66 self.data = data |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
67 self.s = struct.calcsize(indexformat) |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
68 self.l = len(data)/self.s |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
69 self.index = [None] * self.l |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
70 self.map = {nullid: -1} |
323 | 71 self.all = 0 |
72 self.revlog = revlog | |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
73 |
323 | 74 def load(self, pos=None): |
75 if self.all: return | |
76 if pos is not None: | |
77 block = pos / 1000 | |
78 i = block * 1000 | |
79 end = min(self.l, i + 1000) | |
80 else: | |
81 self.all = 1 | |
82 i = 0 | |
83 end = self.l | |
84 self.revlog.index = self.index | |
85 self.revlog.nodemap = self.map | |
515 | 86 |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
87 while i < end: |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
88 d = self.data[i * self.s: (i + 1) * self.s] |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
89 e = struct.unpack(indexformat, d) |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
90 self.index[i] = e |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
91 self.map[e[6]] = i |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
92 i += 1 |
515 | 93 |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
94 class lazyindex: |
1083 | 95 """a lazy version of the index array""" |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
96 def __init__(self, parser): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
97 self.p = parser |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
98 def __len__(self): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
99 return len(self.p.index) |
115 | 100 def load(self, pos): |
1403
bc3e66edb04c
lazyindex fix, make load handle negative indexes properly.
Eric Hopper <hopper@omnifarious.org>
parents:
1402
diff
changeset
|
101 if pos < 0: |
bc3e66edb04c
lazyindex fix, make load handle negative indexes properly.
Eric Hopper <hopper@omnifarious.org>
parents:
1402
diff
changeset
|
102 pos += len(self.p.index) |
115 | 103 self.p.load(pos) |
104 return self.p.index[pos] | |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
105 def __getitem__(self, pos): |
115 | 106 return self.p.index[pos] or self.load(pos) |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
107 def append(self, e): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
108 self.p.index.append(e) |
515 | 109 |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
110 class lazymap: |
1083 | 111 """a lazy version of the node map""" |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
112 def __init__(self, parser): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
113 self.p = parser |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
114 def load(self, key): |
323 | 115 if self.p.all: return |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
116 n = self.p.data.find(key) |
1214 | 117 if n < 0: |
118 raise KeyError(key) | |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
119 pos = n / self.p.s |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
120 self.p.load(pos) |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
121 def __contains__(self, key): |
323 | 122 self.p.load() |
123 return key in self.p.map | |
97 | 124 def __iter__(self): |
469 | 125 yield nullid |
97 | 126 for i in xrange(self.p.l): |
127 try: | |
128 yield self.p.index[i][6] | |
129 except: | |
130 self.p.load(i) | |
131 yield self.p.index[i][6] | |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
132 def __getitem__(self, key): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
133 try: |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
134 return self.p.map[key] |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
135 except KeyError: |
86
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
136 try: |
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
137 self.load(key) |
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
138 return self.p.map[key] |
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
139 except KeyError: |
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
140 raise KeyError("node " + hex(key)) |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
141 def __setitem__(self, key, val): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
142 self.p.map[key] = val |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
143 |
1073
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
144 class RevlogError(Exception): pass |
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
145 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
146 class revlog: |
1083 | 147 """ |
148 the underlying revision storage object | |
149 | |
150 A revlog consists of two parts, an index and the revision data. | |
151 | |
152 The index is a file with a fixed record size containing | |
153 information on each revision, includings its nodeid (hash), the | |
154 nodeids of its parents, the position and offset of its data within | |
155 the data file, and the revision it's based on. Finally, each entry | |
156 contains a linkrev entry that can serve as a pointer to external | |
157 data. | |
158 | |
159 The revision data itself is a linear collection of data chunks. | |
160 Each chunk represents a revision and is usually represented as a | |
161 delta against the previous chunk. To bound lookup time, runs of | |
162 deltas are limited to about 2 times the length of the original | |
163 version data. This makes retrieval of a version proportional to | |
164 its size, or O(1) relative to the number of revisions. | |
165 | |
166 Both pieces of the revlog are written to in an append-only | |
167 fashion, which means we never need to rewrite a file to insert or | |
168 remove data, and can use some simple techniques to avoid the need | |
169 for locking while reading. | |
170 """ | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
171 def __init__(self, opener, indexfile, datafile): |
1083 | 172 """ |
173 create a revlog object | |
174 | |
175 opener is a function that abstracts the file opening operation | |
176 and can be used to implement COW semantics or the like. | |
177 """ | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
178 self.indexfile = indexfile |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
179 self.datafile = datafile |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
180 self.opener = opener |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
181 self.cache = None |
116
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
182 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
183 try: |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
184 i = self.opener(self.indexfile).read() |
1322
b3d44e9b3092
Make revlog constructor more discerning in its treatment of errors.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1232
diff
changeset
|
185 except IOError, inst: |
b3d44e9b3092
Make revlog constructor more discerning in its treatment of errors.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1232
diff
changeset
|
186 if inst.errno != errno.ENOENT: |
b3d44e9b3092
Make revlog constructor more discerning in its treatment of errors.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1232
diff
changeset
|
187 raise |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
188 i = "" |
116
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
189 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
190 if len(i) > 10000: |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
191 # big index, let's parse it on demand |
323 | 192 parser = lazyparser(i, self) |
116
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
193 self.index = lazyindex(parser) |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
194 self.nodemap = lazymap(parser) |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
195 else: |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
196 s = struct.calcsize(indexformat) |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
197 l = len(i) / s |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
198 self.index = [None] * l |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
199 m = [None] * l |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
200 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
201 n = 0 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
202 for f in xrange(0, len(i), s): |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
203 # offset, size, base, linkrev, p1, p2, nodeid |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
204 e = struct.unpack(indexformat, i[f:f + s]) |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
205 m[n] = (e[6], n) |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
206 self.index[n] = e |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
207 n += 1 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
208 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
209 self.nodemap = dict(m) |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
210 self.nodemap[nullid] = -1 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
211 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
212 def tip(self): return self.node(len(self.index) - 1) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
213 def count(self): return len(self.index) |
26 | 214 def node(self, rev): return (rev < 0) and nullid or self.index[rev][6] |
1201
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
215 def rev(self, node): |
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
216 try: |
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
217 return self.nodemap[node] |
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
218 except KeyError: |
1402
9d2c2e6b32b5
i18n part2: use '_' for all strings who are part of the user interface
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents:
1400
diff
changeset
|
219 raise RevlogError(_('%s: no node %s') % (self.indexfile, hex(node))) |
1201
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
220 def linkrev(self, node): return self.index[self.rev(node)][3] |
2 | 221 def parents(self, node): |
222 if node == nullid: return (nullid, nullid) | |
1201
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
223 return self.index[self.rev(node)][4:6] |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
224 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
225 def start(self, rev): return self.index[rev][0] |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
226 def length(self, rev): return self.index[rev][1] |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
227 def end(self, rev): return self.start(rev) + self.length(rev) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
228 def base(self, rev): return self.index[rev][2] |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
229 |
1074
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
230 def reachable(self, rev, stop=None): |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
231 reachable = {} |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
232 visit = [rev] |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
233 reachable[rev] = 1 |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
234 if stop: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
235 stopn = self.rev(stop) |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
236 else: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
237 stopn = 0 |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
238 while visit: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
239 n = visit.pop(0) |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
240 if n == stop: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
241 continue |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
242 if n == nullid: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
243 continue |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
244 for p in self.parents(n): |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
245 if self.rev(p) < stopn: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
246 continue |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
247 if p not in reachable: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
248 reachable[p] = 1 |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
249 visit.append(p) |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
250 return reachable |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
251 |
902 | 252 def heads(self, stop=None): |
1083 | 253 """return the list of all nodes that have no children""" |
221 | 254 p = {} |
255 h = [] | |
902 | 256 stoprev = 0 |
257 if stop and stop in self.nodemap: | |
258 stoprev = self.rev(stop) | |
1083 | 259 |
243 | 260 for r in range(self.count() - 1, -1, -1): |
221 | 261 n = self.node(r) |
262 if n not in p: | |
263 h.append(n) | |
902 | 264 if n == stop: |
265 break | |
266 if r < stoprev: | |
267 break | |
221 | 268 for pn in self.parents(n): |
269 p[pn] = 1 | |
270 return h | |
370 | 271 |
272 def children(self, node): | |
1083 | 273 """find the children of a given node""" |
370 | 274 c = [] |
275 p = self.rev(node) | |
276 for r in range(p + 1, self.count()): | |
277 n = self.node(r) | |
278 for pn in self.parents(n): | |
854
473c030d34a6
Fixed revlog.children.
Tristan Wibberley <tristan@wibberley.org>
parents:
655
diff
changeset
|
279 if pn == node: |
473c030d34a6
Fixed revlog.children.
Tristan Wibberley <tristan@wibberley.org>
parents:
655
diff
changeset
|
280 c.append(n) |
370 | 281 continue |
282 elif pn == nullid: | |
283 continue | |
284 return c | |
515 | 285 |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
286 def lookup(self, id): |
1083 | 287 """locate a node based on revision number or subset of hex nodeid""" |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
288 try: |
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
289 rev = int(id) |
469 | 290 if str(rev) != id: raise ValueError |
291 if rev < 0: rev = self.count() + rev | |
476
0a338d506268
Really _call_ method revlog.count in revlog.lookup()
Thomas Arendsen Hein <thomas@intevation.de>
parents:
469
diff
changeset
|
292 if rev < 0 or rev >= self.count(): raise ValueError |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
293 return self.node(rev) |
469 | 294 except (ValueError, OverflowError): |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
295 c = [] |
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
296 for n in self.nodemap: |
469 | 297 if hex(n).startswith(id): |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
298 c.append(n) |
1402
9d2c2e6b32b5
i18n part2: use '_' for all strings who are part of the user interface
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents:
1400
diff
changeset
|
299 if len(c) > 1: raise RevlogError(_("Ambiguous identifier")) |
9d2c2e6b32b5
i18n part2: use '_' for all strings who are part of the user interface
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents:
1400
diff
changeset
|
300 if len(c) < 1: raise RevlogError(_("No match found")) |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
301 return c[0] |
515 | 302 |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
303 return None |
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
304 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
305 def diff(self, a, b): |
1083 | 306 """return a delta between two revisions""" |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
307 return mdiff.textdiff(a, b) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
308 |
73 | 309 def patches(self, t, pl): |
1083 | 310 """apply a list of patches to a string""" |
73 | 311 return mdiff.patches(t, pl) |
312 | |
119
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
313 def delta(self, node): |
1083 | 314 """return or calculate a delta between a node and its predecessor""" |
119
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
315 r = self.rev(node) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
316 b = self.base(r) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
317 if r == b: |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
318 return self.diff(self.revision(self.node(r - 1)), |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
319 self.revision(node)) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
320 else: |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
321 f = self.opener(self.datafile) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
322 f.seek(self.start(r)) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
323 data = f.read(self.length(r)) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
324 return decompress(data) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
325 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
326 def revision(self, node): |
1083 | 327 """return an uncompressed revision of a given""" |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
328 if node == nullid: return "" |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
329 if self.cache and self.cache[0] == node: return self.cache[2] |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
330 |
1083 | 331 # look up what we need to read |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
332 text = None |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
333 rev = self.rev(node) |
117 | 334 start, length, base, link, p1, p2, node = self.index[rev] |
335 end = start + length | |
336 if base != rev: start = self.start(base) | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
337 |
1083 | 338 # do we have useful data cached? |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
339 if self.cache and self.cache[1] >= base and self.cache[1] < rev: |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
340 base = self.cache[1] |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
341 start = self.start(base + 1) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
342 text = self.cache[2] |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
343 last = 0 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
344 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
345 f = self.opener(self.datafile) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
346 f.seek(start) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
347 data = f.read(end - start) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
348 |
651 | 349 if text is None: |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
350 last = self.length(base) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
351 text = decompress(data[:last]) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
352 |
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
353 bins = [] |
64 | 354 for r in xrange(base + 1, rev + 1): |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
355 s = self.length(r) |
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
356 bins.append(decompress(data[last:last + s])) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
357 last = last + s |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
358 |
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
359 text = mdiff.patches(text, bins) |
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
360 |
26 | 361 if node != hash(text, p1, p2): |
1402
9d2c2e6b32b5
i18n part2: use '_' for all strings who are part of the user interface
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents:
1400
diff
changeset
|
362 raise RevlogError(_("integrity check failed on %s:%d") |
98 | 363 % (self.datafile, rev)) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
364 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
365 self.cache = (node, rev, text) |
515 | 366 return text |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
367 |
644 | 368 def addrevision(self, text, transaction, link, p1=None, p2=None, d=None): |
1083 | 369 """add a revision to the log |
370 | |
371 text - the revision data to add | |
372 transaction - the transaction object used for rollback | |
373 link - the linkrev data to add | |
374 p1, p2 - the parent nodeids of the revision | |
375 d - an optional precomputed delta | |
376 """ | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
377 if text is None: text = "" |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
378 if p1 is None: p1 = self.tip() |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
379 if p2 is None: p2 = nullid |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
380 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
381 node = hash(text, p1, p2) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
382 |
301 | 383 if node in self.nodemap: |
384 return node | |
385 | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
386 n = self.count() |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
387 t = n - 1 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
388 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
389 if n: |
64 | 390 base = self.base(t) |
391 start = self.start(base) | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
392 end = self.end(t) |
644 | 393 if not d: |
394 prev = self.revision(self.tip()) | |
395 d = self.diff(prev, text) | |
98 | 396 data = compress(d) |
64 | 397 dist = end - start + len(data) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
398 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
399 # full versions are inserted when the needed deltas |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
400 # become comparable to the uncompressed text |
64 | 401 if not n or dist > len(text) * 2: |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
402 data = compress(text) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
403 base = n |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
404 else: |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
405 base = self.base(t) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
406 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
407 offset = 0 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
408 if t >= 0: |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
409 offset = self.end(t) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
410 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
411 e = (offset, len(data), base, link, p1, p2, node) |
515 | 412 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
413 self.index.append(e) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
414 self.nodemap[node] = n |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
415 entry = struct.pack(indexformat, *e) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
416 |
26 | 417 transaction.add(self.datafile, e[0]) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
418 self.opener(self.datafile, "a").write(data) |
41 | 419 transaction.add(self.indexfile, n * len(entry)) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
420 self.opener(self.indexfile, "a").write(entry) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
421 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
422 self.cache = (node, n, text) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
423 return node |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
424 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
425 def ancestor(self, a, b): |
1083 | 426 """calculate the least common ancestor of nodes a and b""" |
147 | 427 # calculate the distance of every node from root |
428 dist = {nullid: 0} | |
429 for i in xrange(self.count()): | |
430 n = self.node(i) | |
431 p1, p2 = self.parents(n) | |
432 dist[n] = max(dist[p1], dist[p2]) + 1 | |
515 | 433 |
147 | 434 # traverse ancestors in order of decreasing distance from root |
435 def ancestors(node): | |
436 # we store negative distances because heap returns smallest member | |
437 h = [(-dist[node], node)] | |
438 seen = {} | |
439 earliest = self.count() | |
440 while h: | |
441 d, n = heapq.heappop(h) | |
442 if n not in seen: | |
443 seen[n] = 1 | |
381 | 444 r = self.rev(n) |
1351
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
445 yield (-d, n) |
147 | 446 for p in self.parents(n): |
447 heapq.heappush(h, (-dist[p], p)) | |
45
f2b2d5daec30
Fix recursion depth trouble with ancestor algorithm
mpm@selenic.com
parents:
41
diff
changeset
|
448 |
1351
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
449 def generations(node): |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
450 sg, s = None, {} |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
451 for g,n in ancestors(node): |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
452 if g != sg: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
453 if sg: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
454 yield sg, s |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
455 sg, s = g, {n:1} |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
456 else: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
457 s[n] = 1 |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
458 yield sg, s |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
459 |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
460 x = generations(a) |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
461 y = generations(b) |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
462 gx = x.next() |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
463 gy = y.next() |
45
f2b2d5daec30
Fix recursion depth trouble with ancestor algorithm
mpm@selenic.com
parents:
41
diff
changeset
|
464 |
147 | 465 # increment each ancestor list until it is closer to root than |
466 # the other, or they match | |
467 while 1: | |
1351
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
468 #print "ancestor gen %s %s" % (gx[0], gy[0]) |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
469 if gx[0] == gy[0]: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
470 # find the intersection |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
471 i = [ n for n in gx[1] if n in gy[1] ] |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
472 if i: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
473 return i[0] |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
474 else: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
475 #print "next" |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
476 gy = y.next() |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
477 gx = x.next() |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
478 elif gx[0] < gy[0]: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
479 #print "next y" |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
480 gy = y.next() |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
481 else: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
482 #print "next x" |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
483 gx = x.next() |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
484 |
46 | 485 def group(self, linkmap): |
1083 | 486 """calculate a delta group |
46 | 487 |
1083 | 488 Given a list of changeset revs, return a set of deltas and |
489 metadata corresponding to nodes. the first delta is | |
490 parent(nodes[0]) -> nodes[0] the receiver is guaranteed to | |
491 have this parent as it has all history before these | |
492 changesets. parent is parent[0] | |
493 """ | |
46 | 494 revs = [] |
495 needed = {} | |
496 | |
497 # find file nodes/revs that match changeset revs | |
498 for i in xrange(0, self.count()): | |
499 if self.index[i][3] in linkmap: | |
500 revs.append(i) | |
501 needed[i] = 1 | |
502 | |
503 # if we don't have any revisions touched by these changesets, bail | |
192 | 504 if not revs: |
505 yield struct.pack(">l", 0) | |
506 return | |
46 | 507 |
508 # add the parent of the first rev | |
509 p = self.parents(self.node(revs[0]))[0] | |
510 revs.insert(0, self.rev(p)) | |
511 | |
512 # for each delta that isn't contiguous in the log, we need to | |
513 # reconstruct the base, reconstruct the result, and then | |
514 # calculate the delta. We also need to do this where we've | |
515 # stored a full version and not a delta | |
516 for i in xrange(0, len(revs) - 1): | |
517 a, b = revs[i], revs[i + 1] | |
518 if a + 1 != b or self.base(b) == b: | |
519 for j in xrange(self.base(a), a + 1): | |
520 needed[j] = 1 | |
521 for j in xrange(self.base(b), b + 1): | |
522 needed[j] = 1 | |
523 | |
524 # calculate spans to retrieve from datafile | |
525 needed = needed.keys() | |
526 needed.sort() | |
527 spans = [] | |
192 | 528 oo = -1 |
529 ol = 0 | |
46 | 530 for n in needed: |
531 if n < 0: continue | |
532 o = self.start(n) | |
533 l = self.length(n) | |
192 | 534 if oo + ol == o: # can we merge with the previous? |
535 nl = spans[-1][2] | |
536 nl.append((n, l)) | |
537 ol += l | |
538 spans[-1] = (oo, ol, nl) | |
46 | 539 else: |
192 | 540 oo = o |
541 ol = l | |
542 spans.append((oo, ol, [(n, l)])) | |
46 | 543 |
544 # read spans in, divide up chunks | |
545 chunks = {} | |
192 | 546 for span in spans: |
46 | 547 # we reopen the file for each span to make http happy for now |
548 f = self.opener(self.datafile) | |
549 f.seek(span[0]) | |
550 data = f.read(span[1]) | |
551 | |
552 # divide up the span | |
553 pos = 0 | |
554 for r, l in span[2]: | |
192 | 555 chunks[r] = decompress(data[pos: pos + l]) |
46 | 556 pos += l |
557 | |
558 # helper to reconstruct intermediate versions | |
559 def construct(text, base, rev): | |
192 | 560 bins = [chunks[r] for r in xrange(base + 1, rev + 1)] |
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
561 return mdiff.patches(text, bins) |
46 | 562 |
563 # build deltas | |
564 deltas = [] | |
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
565 for d in xrange(0, len(revs) - 1): |
46 | 566 a, b = revs[d], revs[d + 1] |
567 n = self.node(b) | |
192 | 568 |
569 # do we need to construct a new delta? | |
46 | 570 if a + 1 != b or self.base(b) == b: |
571 if a >= 0: | |
572 base = self.base(a) | |
192 | 573 ta = chunks[self.base(a)] |
46 | 574 ta = construct(ta, base, a) |
575 else: | |
576 ta = "" | |
515 | 577 |
46 | 578 base = self.base(b) |
579 if a > base: | |
580 base = a | |
581 tb = ta | |
582 else: | |
192 | 583 tb = chunks[self.base(b)] |
46 | 584 tb = construct(tb, base, b) |
585 d = self.diff(ta, tb) | |
586 else: | |
192 | 587 d = chunks[b] |
46 | 588 |
589 p = self.parents(n) | |
590 meta = n + p[0] + p[1] + linkmap[self.linkrev(n)] | |
591 l = struct.pack(">l", len(meta) + len(d) + 4) | |
192 | 592 yield l |
593 yield meta | |
594 yield d | |
46 | 595 |
192 | 596 yield struct.pack(">l", 0) |
597 | |
1062 | 598 def addgroup(self, revs, linkmapper, transaction, unique=0): |
1083 | 599 """ |
600 add a delta group | |
46 | 601 |
1083 | 602 given a set of deltas, add them to the revision log. the |
603 first delta is against its parent, which should be in our | |
604 log, the rest are against the previous delta. | |
605 """ | |
606 | |
607 #track the base of the current delta log | |
46 | 608 r = self.count() |
609 t = r - 1 | |
192 | 610 node = nullid |
515 | 611 |
655 | 612 base = prev = -1 |
653
94cdd02792b5
Fix corruption resulting from skipping parts of a revision group
Matt Mackall <mpm@selenic.com>
parents:
651
diff
changeset
|
613 start = end = measure = 0 |
46 | 614 if r: |
615 start = self.start(self.base(t)) | |
616 end = self.end(t) | |
617 measure = self.length(self.base(t)) | |
618 base = self.base(t) | |
619 prev = self.tip() | |
620 | |
621 transaction.add(self.datafile, end) | |
622 transaction.add(self.indexfile, r * struct.calcsize(indexformat)) | |
623 dfh = self.opener(self.datafile, "a") | |
624 ifh = self.opener(self.indexfile, "a") | |
625 | |
626 # loop through our set of deltas | |
192 | 627 chain = None |
628 for chunk in revs: | |
629 node, p1, p2, cs = struct.unpack("20s20s20s20s", chunk[:80]) | |
94 | 630 link = linkmapper(cs) |
77 | 631 if node in self.nodemap: |
224
ccbcc4d76f81
fix bad assumption about uniqueness of file versions
mpm@selenic.com
parents:
221
diff
changeset
|
632 # this can happen if two branches make the same change |
1218
cde6818e082a
Add preliminary support for the bundle and unbundle commands
mpm@selenic.com
parents:
1214
diff
changeset
|
633 # if unique: |
1402
9d2c2e6b32b5
i18n part2: use '_' for all strings who are part of the user interface
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents:
1400
diff
changeset
|
634 # raise RevlogError(_("already have %s") % hex(node[:4])) |
653
94cdd02792b5
Fix corruption resulting from skipping parts of a revision group
Matt Mackall <mpm@selenic.com>
parents:
651
diff
changeset
|
635 chain = node |
224
ccbcc4d76f81
fix bad assumption about uniqueness of file versions
mpm@selenic.com
parents:
221
diff
changeset
|
636 continue |
192 | 637 delta = chunk[80:] |
638 | |
639 if not chain: | |
640 # retrieve the parent revision of the delta chain | |
641 chain = p1 | |
642 if not chain in self.nodemap: | |
1402
9d2c2e6b32b5
i18n part2: use '_' for all strings who are part of the user interface
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents:
1400
diff
changeset
|
643 raise RevlogError(_("unknown base %s") % short(chain[:4])) |
46 | 644 |
645 # full versions are inserted when the needed deltas become | |
646 # comparable to the uncompressed text or when the previous | |
647 # version is not the one we have a delta against. We use | |
648 # the size of the previous full rev as a proxy for the | |
649 # current size. | |
650 | |
651 if chain == prev: | |
652 cdelta = compress(delta) | |
653 | |
654 if chain != prev or (end - start + len(cdelta)) > measure * 2: | |
655 # flush our writes here so we can read it in revision | |
656 dfh.flush() | |
657 ifh.flush() | |
65
d40cc5aacc31
Fix up a bunch of bugs in the new merge code
mpm@selenic.com
parents:
64
diff
changeset
|
658 text = self.revision(chain) |
73 | 659 text = self.patches(text, [delta]) |
46 | 660 chk = self.addrevision(text, transaction, link, p1, p2) |
661 if chk != node: | |
1402
9d2c2e6b32b5
i18n part2: use '_' for all strings who are part of the user interface
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents:
1400
diff
changeset
|
662 raise RevlogError(_("consistency error adding group")) |
46 | 663 measure = len(text) |
664 else: | |
665 e = (end, len(cdelta), self.base(t), link, p1, p2, node) | |
666 self.index.append(e) | |
667 self.nodemap[node] = r | |
668 dfh.write(cdelta) | |
669 ifh.write(struct.pack(indexformat, *e)) | |
670 | |
65
d40cc5aacc31
Fix up a bunch of bugs in the new merge code
mpm@selenic.com
parents:
64
diff
changeset
|
671 t, r, chain, prev = r, r + 1, node, node |
46 | 672 start = self.start(self.base(t)) |
673 end = self.end(t) | |
674 | |
675 dfh.close() | |
676 ifh.close() | |
677 return node |