Mercurial > public > mercurial-scm > hg-stable
annotate mercurial/revlog.py @ 1351:0e2be889ccd7
Repair ancestor logic, fix up test cases
author | Matt Mackall <mpm@selenic.com> |
---|---|
date | Mon, 26 Sep 2005 16:52:47 -0700 |
parents | 57220daf40e9 |
children | 67779d34cb52 518da3c3b6ce |
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 * |
1322
b3d44e9b3092
Make revlog constructor more discerning in its treatment of errors.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1232
diff
changeset
|
14 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
|
15 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
|
16 |
1091
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
17 def hash(text, p1, p2): |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
18 """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
|
19 |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
20 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
|
21 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
|
22 content in the revision graph. |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
23 """ |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
24 l = [p1, p2] |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
25 l.sort() |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
26 s = sha.new(l[0]) |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
27 s.update(l[1]) |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
28 s.update(text) |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
29 return s.digest() |
d62130f99a73
Move hash function back to revlog from node
mpm@selenic.com
parents:
1089
diff
changeset
|
30 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
31 def compress(text): |
1083 | 32 """ generate a possibly-compressed representation of text """ |
112 | 33 if not text: return text |
34 if len(text) < 44: | |
35 if text[0] == '\0': return text | |
36 return 'u' + text | |
37 bin = zlib.compress(text) | |
38 if len(bin) > len(text): | |
39 if text[0] == '\0': return text | |
40 return 'u' + text | |
41 return bin | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
42 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
43 def decompress(bin): |
1083 | 44 """ decompress the given input """ |
112 | 45 if not bin: return bin |
46 t = bin[0] | |
47 if t == '\0': return bin | |
48 if t == 'x': return zlib.decompress(bin) | |
49 if t == 'u': return bin[1:] | |
1073
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
50 raise RevlogError("unknown compression type %s" % t) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
51 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
52 indexformat = ">4l20s20s20s" |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
53 |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
54 class lazyparser: |
1083 | 55 """ |
56 this class avoids the need to parse the entirety of large indices | |
57 | |
58 By default we parse and load 1000 entries at a time. | |
59 | |
60 If no position is specified, we load the whole index, and replace | |
61 the lazy objects in revlog with the underlying objects for | |
62 efficiency in cases where we look at most of the nodes. | |
63 """ | |
323 | 64 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
|
65 self.data = data |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
66 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
|
67 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
|
68 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
|
69 self.map = {nullid: -1} |
323 | 70 self.all = 0 |
71 self.revlog = revlog | |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
72 |
323 | 73 def load(self, pos=None): |
74 if self.all: return | |
75 if pos is not None: | |
76 block = pos / 1000 | |
77 i = block * 1000 | |
78 end = min(self.l, i + 1000) | |
79 else: | |
80 self.all = 1 | |
81 i = 0 | |
82 end = self.l | |
83 self.revlog.index = self.index | |
84 self.revlog.nodemap = self.map | |
515 | 85 |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
86 while i < end: |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
87 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
|
88 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
|
89 self.index[i] = e |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
90 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
|
91 i += 1 |
515 | 92 |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
93 class lazyindex: |
1083 | 94 """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
|
95 def __init__(self, parser): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
96 self.p = parser |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
97 def __len__(self): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
98 return len(self.p.index) |
115 | 99 def load(self, pos): |
100 self.p.load(pos) | |
101 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
|
102 def __getitem__(self, pos): |
115 | 103 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
|
104 def append(self, e): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
105 self.p.index.append(e) |
515 | 106 |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
107 class lazymap: |
1083 | 108 """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
|
109 def __init__(self, parser): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
110 self.p = parser |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
111 def load(self, key): |
323 | 112 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
|
113 n = self.p.data.find(key) |
1214 | 114 if n < 0: |
115 raise KeyError(key) | |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
116 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
|
117 self.p.load(pos) |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
118 def __contains__(self, key): |
323 | 119 self.p.load() |
120 return key in self.p.map | |
97 | 121 def __iter__(self): |
469 | 122 yield nullid |
97 | 123 for i in xrange(self.p.l): |
124 try: | |
125 yield self.p.index[i][6] | |
126 except: | |
127 self.p.load(i) | |
128 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
|
129 def __getitem__(self, key): |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
130 try: |
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
131 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
|
132 except KeyError: |
86
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
133 try: |
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
134 self.load(key) |
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
135 return self.p.map[key] |
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
136 except KeyError: |
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
137 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
|
138 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
|
139 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
|
140 |
1073
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
141 class RevlogError(Exception): pass |
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
142 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
143 class revlog: |
1083 | 144 """ |
145 the underlying revision storage object | |
146 | |
147 A revlog consists of two parts, an index and the revision data. | |
148 | |
149 The index is a file with a fixed record size containing | |
150 information on each revision, includings its nodeid (hash), the | |
151 nodeids of its parents, the position and offset of its data within | |
152 the data file, and the revision it's based on. Finally, each entry | |
153 contains a linkrev entry that can serve as a pointer to external | |
154 data. | |
155 | |
156 The revision data itself is a linear collection of data chunks. | |
157 Each chunk represents a revision and is usually represented as a | |
158 delta against the previous chunk. To bound lookup time, runs of | |
159 deltas are limited to about 2 times the length of the original | |
160 version data. This makes retrieval of a version proportional to | |
161 its size, or O(1) relative to the number of revisions. | |
162 | |
163 Both pieces of the revlog are written to in an append-only | |
164 fashion, which means we never need to rewrite a file to insert or | |
165 remove data, and can use some simple techniques to avoid the need | |
166 for locking while reading. | |
167 """ | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
168 def __init__(self, opener, indexfile, datafile): |
1083 | 169 """ |
170 create a revlog object | |
171 | |
172 opener is a function that abstracts the file opening operation | |
173 and can be used to implement COW semantics or the like. | |
174 """ | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
175 self.indexfile = indexfile |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
176 self.datafile = datafile |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
177 self.opener = opener |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
178 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
|
179 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
180 try: |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
181 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
|
182 except IOError, inst: |
b3d44e9b3092
Make revlog constructor more discerning in its treatment of errors.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1232
diff
changeset
|
183 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
|
184 raise |
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
185 i = "" |
116
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
186 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
187 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
|
188 # big index, let's parse it on demand |
323 | 189 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
|
190 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
|
191 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
|
192 else: |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
193 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
|
194 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
|
195 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
|
196 m = [None] * l |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
197 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
198 n = 0 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
199 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
|
200 # 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
|
201 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
|
202 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
|
203 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
|
204 n += 1 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
205 |
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
206 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
|
207 self.nodemap[nullid] = -1 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
208 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
209 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
|
210 def count(self): return len(self.index) |
26 | 211 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
|
212 def rev(self, node): |
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
213 try: |
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
214 return self.nodemap[node] |
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
215 except KeyError: |
1214 | 216 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
|
217 def linkrev(self, node): return self.index[self.rev(node)][3] |
2 | 218 def parents(self, node): |
219 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
|
220 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
|
221 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
222 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
|
223 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
|
224 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
|
225 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
|
226 |
1074
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
227 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
|
228 reachable = {} |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
229 visit = [rev] |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
230 reachable[rev] = 1 |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
231 if stop: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
232 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
|
233 else: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
234 stopn = 0 |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
235 while visit: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
236 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
|
237 if n == stop: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
238 continue |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
239 if n == nullid: |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
240 continue |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
241 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
|
242 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
|
243 continue |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
244 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
|
245 reachable[p] = 1 |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
246 visit.append(p) |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
247 return reachable |
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
248 |
902 | 249 def heads(self, stop=None): |
1083 | 250 """return the list of all nodes that have no children""" |
221 | 251 p = {} |
252 h = [] | |
902 | 253 stoprev = 0 |
254 if stop and stop in self.nodemap: | |
255 stoprev = self.rev(stop) | |
1083 | 256 |
243 | 257 for r in range(self.count() - 1, -1, -1): |
221 | 258 n = self.node(r) |
259 if n not in p: | |
260 h.append(n) | |
902 | 261 if n == stop: |
262 break | |
263 if r < stoprev: | |
264 break | |
221 | 265 for pn in self.parents(n): |
266 p[pn] = 1 | |
267 return h | |
370 | 268 |
269 def children(self, node): | |
1083 | 270 """find the children of a given node""" |
370 | 271 c = [] |
272 p = self.rev(node) | |
273 for r in range(p + 1, self.count()): | |
274 n = self.node(r) | |
275 for pn in self.parents(n): | |
854
473c030d34a6
Fixed revlog.children.
Tristan Wibberley <tristan@wibberley.org>
parents:
655
diff
changeset
|
276 if pn == node: |
473c030d34a6
Fixed revlog.children.
Tristan Wibberley <tristan@wibberley.org>
parents:
655
diff
changeset
|
277 c.append(n) |
370 | 278 continue |
279 elif pn == nullid: | |
280 continue | |
281 return c | |
515 | 282 |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
283 def lookup(self, id): |
1083 | 284 """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
|
285 try: |
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
286 rev = int(id) |
469 | 287 if str(rev) != id: raise ValueError |
288 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
|
289 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
|
290 return self.node(rev) |
469 | 291 except (ValueError, OverflowError): |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
292 c = [] |
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
293 for n in self.nodemap: |
469 | 294 if hex(n).startswith(id): |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
295 c.append(n) |
1232
eb3cc5e2eb89
Revert some exception type changes in revlog
mpm@selenic.com
parents:
1218
diff
changeset
|
296 if len(c) > 1: raise KeyError("Ambiguous identifier") |
eb3cc5e2eb89
Revert some exception type changes in revlog
mpm@selenic.com
parents:
1218
diff
changeset
|
297 if len(c) < 1: raise KeyError("No match found") |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
298 return c[0] |
515 | 299 |
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
300 return None |
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
301 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
302 def diff(self, a, b): |
1083 | 303 """return a delta between two revisions""" |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
304 return mdiff.textdiff(a, b) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
305 |
73 | 306 def patches(self, t, pl): |
1083 | 307 """apply a list of patches to a string""" |
73 | 308 return mdiff.patches(t, pl) |
309 | |
119
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
310 def delta(self, node): |
1083 | 311 """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
|
312 r = self.rev(node) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
313 b = self.base(r) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
314 if r == b: |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
315 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
|
316 self.revision(node)) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
317 else: |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
318 f = self.opener(self.datafile) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
319 f.seek(self.start(r)) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
320 data = f.read(self.length(r)) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
321 return decompress(data) |
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
322 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
323 def revision(self, node): |
1083 | 324 """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
|
325 if node == nullid: return "" |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
326 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
|
327 |
1083 | 328 # look up what we need to read |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
329 text = None |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
330 rev = self.rev(node) |
117 | 331 start, length, base, link, p1, p2, node = self.index[rev] |
332 end = start + length | |
333 if base != rev: start = self.start(base) | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
334 |
1083 | 335 # do we have useful data cached? |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
336 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
|
337 base = self.cache[1] |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
338 start = self.start(base + 1) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
339 text = self.cache[2] |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
340 last = 0 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
341 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
342 f = self.opener(self.datafile) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
343 f.seek(start) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
344 data = f.read(end - start) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
345 |
651 | 346 if text is None: |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
347 last = self.length(base) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
348 text = decompress(data[:last]) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
349 |
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
350 bins = [] |
64 | 351 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
|
352 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
|
353 bins.append(decompress(data[last:last + s])) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
354 last = last + s |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
355 |
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
356 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
|
357 |
26 | 358 if node != hash(text, p1, p2): |
1214 | 359 raise RevlogError("integrity check failed on %s:%d" |
98 | 360 % (self.datafile, rev)) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
361 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
362 self.cache = (node, rev, text) |
515 | 363 return text |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
364 |
644 | 365 def addrevision(self, text, transaction, link, p1=None, p2=None, d=None): |
1083 | 366 """add a revision to the log |
367 | |
368 text - the revision data to add | |
369 transaction - the transaction object used for rollback | |
370 link - the linkrev data to add | |
371 p1, p2 - the parent nodeids of the revision | |
372 d - an optional precomputed delta | |
373 """ | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
374 if text is None: text = "" |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
375 if p1 is None: p1 = self.tip() |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
376 if p2 is None: p2 = nullid |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
377 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
378 node = hash(text, p1, p2) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
379 |
301 | 380 if node in self.nodemap: |
381 return node | |
382 | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
383 n = self.count() |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
384 t = n - 1 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
385 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
386 if n: |
64 | 387 base = self.base(t) |
388 start = self.start(base) | |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
389 end = self.end(t) |
644 | 390 if not d: |
391 prev = self.revision(self.tip()) | |
392 d = self.diff(prev, text) | |
98 | 393 data = compress(d) |
64 | 394 dist = end - start + len(data) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
395 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
396 # full versions are inserted when the needed deltas |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
397 # become comparable to the uncompressed text |
64 | 398 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
|
399 data = compress(text) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
400 base = n |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
401 else: |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
402 base = self.base(t) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
403 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
404 offset = 0 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
405 if t >= 0: |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
406 offset = self.end(t) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
407 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
408 e = (offset, len(data), base, link, p1, p2, node) |
515 | 409 |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
410 self.index.append(e) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
411 self.nodemap[node] = n |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
412 entry = struct.pack(indexformat, *e) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
413 |
26 | 414 transaction.add(self.datafile, e[0]) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
415 self.opener(self.datafile, "a").write(data) |
41 | 416 transaction.add(self.indexfile, n * len(entry)) |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
417 self.opener(self.indexfile, "a").write(entry) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
418 |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
419 self.cache = (node, n, text) |
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
420 return node |
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 def ancestor(self, a, b): |
1083 | 423 """calculate the least common ancestor of nodes a and b""" |
147 | 424 # calculate the distance of every node from root |
425 dist = {nullid: 0} | |
426 for i in xrange(self.count()): | |
427 n = self.node(i) | |
428 p1, p2 = self.parents(n) | |
429 dist[n] = max(dist[p1], dist[p2]) + 1 | |
515 | 430 |
147 | 431 # traverse ancestors in order of decreasing distance from root |
432 def ancestors(node): | |
433 # we store negative distances because heap returns smallest member | |
434 h = [(-dist[node], node)] | |
435 seen = {} | |
436 earliest = self.count() | |
437 while h: | |
438 d, n = heapq.heappop(h) | |
439 if n not in seen: | |
440 seen[n] = 1 | |
381 | 441 r = self.rev(n) |
1351
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
442 yield (-d, n) |
147 | 443 for p in self.parents(n): |
444 heapq.heappush(h, (-dist[p], p)) | |
45
f2b2d5daec30
Fix recursion depth trouble with ancestor algorithm
mpm@selenic.com
parents:
41
diff
changeset
|
445 |
1351
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
446 def generations(node): |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
447 sg, s = None, {} |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
448 for g,n in ancestors(node): |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
449 if g != sg: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
450 if sg: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
451 yield sg, s |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
452 sg, s = g, {n:1} |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
453 else: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
454 s[n] = 1 |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
455 yield sg, s |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
456 |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
457 x = generations(a) |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
458 y = generations(b) |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
459 gx = x.next() |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
460 gy = y.next() |
45
f2b2d5daec30
Fix recursion depth trouble with ancestor algorithm
mpm@selenic.com
parents:
41
diff
changeset
|
461 |
147 | 462 # increment each ancestor list until it is closer to root than |
463 # the other, or they match | |
464 while 1: | |
1351
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
465 #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
|
466 if gx[0] == gy[0]: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
467 # find the intersection |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
468 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
|
469 if i: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
470 return i[0] |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
471 else: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
472 #print "next" |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
473 gy = y.next() |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
474 gx = x.next() |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
475 elif gx[0] < gy[0]: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
476 #print "next y" |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
477 gy = y.next() |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
478 else: |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
479 #print "next x" |
0e2be889ccd7
Repair ancestor logic, fix up test cases
Matt Mackall <mpm@selenic.com>
parents:
1325
diff
changeset
|
480 gx = x.next() |
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
481 |
46 | 482 def group(self, linkmap): |
1083 | 483 """calculate a delta group |
46 | 484 |
1083 | 485 Given a list of changeset revs, return a set of deltas and |
486 metadata corresponding to nodes. the first delta is | |
487 parent(nodes[0]) -> nodes[0] the receiver is guaranteed to | |
488 have this parent as it has all history before these | |
489 changesets. parent is parent[0] | |
490 """ | |
46 | 491 revs = [] |
492 needed = {} | |
493 | |
494 # find file nodes/revs that match changeset revs | |
495 for i in xrange(0, self.count()): | |
496 if self.index[i][3] in linkmap: | |
497 revs.append(i) | |
498 needed[i] = 1 | |
499 | |
500 # if we don't have any revisions touched by these changesets, bail | |
192 | 501 if not revs: |
502 yield struct.pack(">l", 0) | |
503 return | |
46 | 504 |
505 # add the parent of the first rev | |
506 p = self.parents(self.node(revs[0]))[0] | |
507 revs.insert(0, self.rev(p)) | |
508 | |
509 # for each delta that isn't contiguous in the log, we need to | |
510 # reconstruct the base, reconstruct the result, and then | |
511 # calculate the delta. We also need to do this where we've | |
512 # stored a full version and not a delta | |
513 for i in xrange(0, len(revs) - 1): | |
514 a, b = revs[i], revs[i + 1] | |
515 if a + 1 != b or self.base(b) == b: | |
516 for j in xrange(self.base(a), a + 1): | |
517 needed[j] = 1 | |
518 for j in xrange(self.base(b), b + 1): | |
519 needed[j] = 1 | |
520 | |
521 # calculate spans to retrieve from datafile | |
522 needed = needed.keys() | |
523 needed.sort() | |
524 spans = [] | |
192 | 525 oo = -1 |
526 ol = 0 | |
46 | 527 for n in needed: |
528 if n < 0: continue | |
529 o = self.start(n) | |
530 l = self.length(n) | |
192 | 531 if oo + ol == o: # can we merge with the previous? |
532 nl = spans[-1][2] | |
533 nl.append((n, l)) | |
534 ol += l | |
535 spans[-1] = (oo, ol, nl) | |
46 | 536 else: |
192 | 537 oo = o |
538 ol = l | |
539 spans.append((oo, ol, [(n, l)])) | |
46 | 540 |
541 # read spans in, divide up chunks | |
542 chunks = {} | |
192 | 543 for span in spans: |
46 | 544 # we reopen the file for each span to make http happy for now |
545 f = self.opener(self.datafile) | |
546 f.seek(span[0]) | |
547 data = f.read(span[1]) | |
548 | |
549 # divide up the span | |
550 pos = 0 | |
551 for r, l in span[2]: | |
192 | 552 chunks[r] = decompress(data[pos: pos + l]) |
46 | 553 pos += l |
554 | |
555 # helper to reconstruct intermediate versions | |
556 def construct(text, base, rev): | |
192 | 557 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
|
558 return mdiff.patches(text, bins) |
46 | 559 |
560 # build deltas | |
561 deltas = [] | |
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
562 for d in xrange(0, len(revs) - 1): |
46 | 563 a, b = revs[d], revs[d + 1] |
564 n = self.node(b) | |
192 | 565 |
566 # do we need to construct a new delta? | |
46 | 567 if a + 1 != b or self.base(b) == b: |
568 if a >= 0: | |
569 base = self.base(a) | |
192 | 570 ta = chunks[self.base(a)] |
46 | 571 ta = construct(ta, base, a) |
572 else: | |
573 ta = "" | |
515 | 574 |
46 | 575 base = self.base(b) |
576 if a > base: | |
577 base = a | |
578 tb = ta | |
579 else: | |
192 | 580 tb = chunks[self.base(b)] |
46 | 581 tb = construct(tb, base, b) |
582 d = self.diff(ta, tb) | |
583 else: | |
192 | 584 d = chunks[b] |
46 | 585 |
586 p = self.parents(n) | |
587 meta = n + p[0] + p[1] + linkmap[self.linkrev(n)] | |
588 l = struct.pack(">l", len(meta) + len(d) + 4) | |
192 | 589 yield l |
590 yield meta | |
591 yield d | |
46 | 592 |
192 | 593 yield struct.pack(">l", 0) |
594 | |
1062 | 595 def addgroup(self, revs, linkmapper, transaction, unique=0): |
1083 | 596 """ |
597 add a delta group | |
46 | 598 |
1083 | 599 given a set of deltas, add them to the revision log. the |
600 first delta is against its parent, which should be in our | |
601 log, the rest are against the previous delta. | |
602 """ | |
603 | |
604 #track the base of the current delta log | |
46 | 605 r = self.count() |
606 t = r - 1 | |
192 | 607 node = nullid |
515 | 608 |
655 | 609 base = prev = -1 |
653
94cdd02792b5
Fix corruption resulting from skipping parts of a revision group
Matt Mackall <mpm@selenic.com>
parents:
651
diff
changeset
|
610 start = end = measure = 0 |
46 | 611 if r: |
612 start = self.start(self.base(t)) | |
613 end = self.end(t) | |
614 measure = self.length(self.base(t)) | |
615 base = self.base(t) | |
616 prev = self.tip() | |
617 | |
618 transaction.add(self.datafile, end) | |
619 transaction.add(self.indexfile, r * struct.calcsize(indexformat)) | |
620 dfh = self.opener(self.datafile, "a") | |
621 ifh = self.opener(self.indexfile, "a") | |
622 | |
623 # loop through our set of deltas | |
192 | 624 chain = None |
625 for chunk in revs: | |
626 node, p1, p2, cs = struct.unpack("20s20s20s20s", chunk[:80]) | |
94 | 627 link = linkmapper(cs) |
77 | 628 if node in self.nodemap: |
224
ccbcc4d76f81
fix bad assumption about uniqueness of file versions
mpm@selenic.com
parents:
221
diff
changeset
|
629 # 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
|
630 # if unique: |
cde6818e082a
Add preliminary support for the bundle and unbundle commands
mpm@selenic.com
parents:
1214
diff
changeset
|
631 # 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
|
632 chain = node |
224
ccbcc4d76f81
fix bad assumption about uniqueness of file versions
mpm@selenic.com
parents:
221
diff
changeset
|
633 continue |
192 | 634 delta = chunk[80:] |
635 | |
636 if not chain: | |
637 # retrieve the parent revision of the delta chain | |
638 chain = p1 | |
639 if not chain in self.nodemap: | |
1073
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
640 raise RevlogError("unknown base %s" % short(chain[:4])) |
46 | 641 |
642 # full versions are inserted when the needed deltas become | |
643 # comparable to the uncompressed text or when the previous | |
644 # version is not the one we have a delta against. We use | |
645 # the size of the previous full rev as a proxy for the | |
646 # current size. | |
647 | |
648 if chain == prev: | |
649 cdelta = compress(delta) | |
650 | |
651 if chain != prev or (end - start + len(cdelta)) > measure * 2: | |
652 # flush our writes here so we can read it in revision | |
653 dfh.flush() | |
654 ifh.flush() | |
65
d40cc5aacc31
Fix up a bunch of bugs in the new merge code
mpm@selenic.com
parents:
64
diff
changeset
|
655 text = self.revision(chain) |
73 | 656 text = self.patches(text, [delta]) |
46 | 657 chk = self.addrevision(text, transaction, link, p1, p2) |
658 if chk != node: | |
1073
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
659 raise RevlogError("consistency error adding group") |
46 | 660 measure = len(text) |
661 else: | |
662 e = (end, len(cdelta), self.base(t), link, p1, p2, node) | |
663 self.index.append(e) | |
664 self.nodemap[node] = r | |
665 dfh.write(cdelta) | |
666 ifh.write(struct.pack(indexformat, *e)) | |
667 | |
65
d40cc5aacc31
Fix up a bunch of bugs in the new merge code
mpm@selenic.com
parents:
64
diff
changeset
|
668 t, r, chain, prev = r, r + 1, node, node |
46 | 669 start = self.start(self.base(t)) |
670 end = self.end(t) | |
671 | |
672 dfh.close() | |
673 ifh.close() | |
674 return node |