Mercurial > public > mercurial-scm > hg-stable
annotate mercurial/revlog.py @ 1201:59bfbdbc38f6
revlog: raise informative exception if file is missing.
| author | Bryan O'Sullivan <bos@serpentine.com> |
|---|---|
| date | Sun, 04 Sep 2005 14:45:03 -0700 |
| parents | 0cb9c72b6c03 |
| children | 34706a835d4a |
| 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 |
| 208 | 13 import zlib, struct, sha, binascii, heapq |
| 1099 | 14 import mdiff |
| 1089 | 15 from node import * |
|
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) |
|
86
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
114 if n < 0: 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
|
115 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
|
116 self.p.load(pos) |
|
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
117 def __contains__(self, key): |
| 323 | 118 self.p.load() |
| 119 return key in self.p.map | |
| 97 | 120 def __iter__(self): |
| 469 | 121 yield nullid |
| 97 | 122 for i in xrange(self.p.l): |
| 123 try: | |
| 124 yield self.p.index[i][6] | |
| 125 except: | |
| 126 self.p.load(i) | |
| 127 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
|
128 def __getitem__(self, key): |
|
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
129 try: |
|
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
130 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
|
131 except KeyError: |
|
86
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
132 try: |
|
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
133 self.load(key) |
|
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
134 return self.p.map[key] |
|
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
135 except KeyError: |
|
1b945e8ba67b
Friendlier exceptions for unknown node errors
mpm@selenic.com
parents:
84
diff
changeset
|
136 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
|
137 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
|
138 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
|
139 |
|
1073
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
140 class RevlogError(Exception): pass |
|
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
141 |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
142 class revlog: |
| 1083 | 143 """ |
| 144 the underlying revision storage object | |
| 145 | |
| 146 A revlog consists of two parts, an index and the revision data. | |
| 147 | |
| 148 The index is a file with a fixed record size containing | |
| 149 information on each revision, includings its nodeid (hash), the | |
| 150 nodeids of its parents, the position and offset of its data within | |
| 151 the data file, and the revision it's based on. Finally, each entry | |
| 152 contains a linkrev entry that can serve as a pointer to external | |
| 153 data. | |
| 154 | |
| 155 The revision data itself is a linear collection of data chunks. | |
| 156 Each chunk represents a revision and is usually represented as a | |
| 157 delta against the previous chunk. To bound lookup time, runs of | |
| 158 deltas are limited to about 2 times the length of the original | |
| 159 version data. This makes retrieval of a version proportional to | |
| 160 its size, or O(1) relative to the number of revisions. | |
| 161 | |
| 162 Both pieces of the revlog are written to in an append-only | |
| 163 fashion, which means we never need to rewrite a file to insert or | |
| 164 remove data, and can use some simple techniques to avoid the need | |
| 165 for locking while reading. | |
| 166 """ | |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
167 def __init__(self, opener, indexfile, datafile): |
| 1083 | 168 """ |
| 169 create a revlog object | |
| 170 | |
| 171 opener is a function that abstracts the file opening operation | |
| 172 and can be used to implement COW semantics or the like. | |
| 173 """ | |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
174 self.indexfile = indexfile |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
175 self.datafile = datafile |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
176 self.opener = opener |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
177 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
|
178 |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
179 try: |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
180 i = self.opener(self.indexfile).read() |
| 73 | 181 except IOError: |
|
76
d993ebd69d28
Add lazy{parser,index,map} to speed up processing of index files
mpm@selenic.com
parents:
73
diff
changeset
|
182 i = "" |
|
116
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
183 |
|
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
184 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
|
185 # big index, let's parse it on demand |
| 323 | 186 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
|
187 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
|
188 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
|
189 else: |
|
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
190 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
|
191 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
|
192 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
|
193 m = [None] * l |
|
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
194 |
|
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
195 n = 0 |
|
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
196 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
|
197 # 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
|
198 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
|
199 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
|
200 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
|
201 n += 1 |
|
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
202 |
|
e484cd5ec282
Only use lazy indexing for big indices and avoid the overhead of the
mpm@selenic.com
parents:
115
diff
changeset
|
203 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
|
204 self.nodemap[nullid] = -1 |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
205 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
206 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
|
207 def count(self): return len(self.index) |
| 26 | 208 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
|
209 def rev(self, node): |
|
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
210 try: |
|
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
211 return self.nodemap[node] |
|
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
212 except KeyError: |
|
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
213 raise KeyError('%s: no node %s' % (self.indexfile, hex(node))) |
|
59bfbdbc38f6
revlog: raise informative exception if file is missing.
Bryan O'Sullivan <bos@serpentine.com>
parents:
1099
diff
changeset
|
214 def linkrev(self, node): return self.index[self.rev(node)][3] |
| 2 | 215 def parents(self, node): |
| 216 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
|
217 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
|
218 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
219 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
|
220 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
|
221 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
|
222 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
|
223 |
|
1074
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
224 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
|
225 reachable = {} |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
226 visit = [rev] |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
227 reachable[rev] = 1 |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
228 if stop: |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
229 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
|
230 else: |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
231 stopn = 0 |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
232 while visit: |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
233 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
|
234 if n == stop: |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
235 continue |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
236 if n == nullid: |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
237 continue |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
238 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
|
239 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
|
240 continue |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
241 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
|
242 reachable[p] = 1 |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
243 visit.append(p) |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
244 return reachable |
|
55bf5cfde69e
Add revlog.reachable to find a graph of ancestors for a given rev
mason@suse.com
parents:
1073
diff
changeset
|
245 |
| 902 | 246 def heads(self, stop=None): |
| 1083 | 247 """return the list of all nodes that have no children""" |
| 221 | 248 p = {} |
| 249 h = [] | |
| 902 | 250 stoprev = 0 |
| 251 if stop and stop in self.nodemap: | |
| 252 stoprev = self.rev(stop) | |
| 1083 | 253 |
| 243 | 254 for r in range(self.count() - 1, -1, -1): |
| 221 | 255 n = self.node(r) |
| 256 if n not in p: | |
| 257 h.append(n) | |
| 902 | 258 if n == stop: |
| 259 break | |
| 260 if r < stoprev: | |
| 261 break | |
| 221 | 262 for pn in self.parents(n): |
| 263 p[pn] = 1 | |
| 264 return h | |
| 370 | 265 |
| 266 def children(self, node): | |
| 1083 | 267 """find the children of a given node""" |
| 370 | 268 c = [] |
| 269 p = self.rev(node) | |
| 270 for r in range(p + 1, self.count()): | |
| 271 n = self.node(r) | |
| 272 for pn in self.parents(n): | |
|
854
473c030d34a6
Fixed revlog.children.
Tristan Wibberley <tristan@wibberley.org>
parents:
655
diff
changeset
|
273 if pn == node: |
|
473c030d34a6
Fixed revlog.children.
Tristan Wibberley <tristan@wibberley.org>
parents:
655
diff
changeset
|
274 c.append(n) |
| 370 | 275 continue |
| 276 elif pn == nullid: | |
| 277 continue | |
| 278 return c | |
| 515 | 279 |
|
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
280 def lookup(self, id): |
| 1083 | 281 """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
|
282 try: |
|
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
283 rev = int(id) |
| 469 | 284 if str(rev) != id: raise ValueError |
| 285 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
|
286 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
|
287 return self.node(rev) |
| 469 | 288 except (ValueError, OverflowError): |
|
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
289 c = [] |
|
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
290 for n in self.nodemap: |
| 469 | 291 if hex(n).startswith(id): |
|
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
292 c.append(n) |
|
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
293 if len(c) > 1: raise KeyError("Ambiguous identifier") |
| 67 | 294 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
|
295 return c[0] |
| 515 | 296 |
|
36
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
297 return None |
|
da28286bf6b7
Add smart node lookup by substring or by rev number
mpm@selenic.com
parents:
26
diff
changeset
|
298 |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
299 def diff(self, a, b): |
| 1083 | 300 """return a delta between two revisions""" |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
301 return mdiff.textdiff(a, b) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
302 |
| 73 | 303 def patches(self, t, pl): |
| 1083 | 304 """apply a list of patches to a string""" |
| 73 | 305 return mdiff.patches(t, pl) |
| 306 | |
|
119
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
307 def delta(self, node): |
| 1083 | 308 """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
|
309 r = self.rev(node) |
|
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
310 b = self.base(r) |
|
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
311 if r == b: |
|
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
312 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
|
313 self.revision(node)) |
|
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
314 else: |
|
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
315 f = self.opener(self.datafile) |
|
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
316 f.seek(self.start(r)) |
|
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
317 data = f.read(self.length(r)) |
|
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
318 return decompress(data) |
|
c7a66f9752a4
Add code to retrieve or construct a revlog delta
mpm@selenic.com
parents:
117
diff
changeset
|
319 |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
320 def revision(self, node): |
| 1083 | 321 """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
|
322 if node == nullid: return "" |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
323 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
|
324 |
| 1083 | 325 # look up what we need to read |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
326 text = None |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
327 rev = self.rev(node) |
| 117 | 328 start, length, base, link, p1, p2, node = self.index[rev] |
| 329 end = start + length | |
| 330 if base != rev: start = self.start(base) | |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
331 |
| 1083 | 332 # do we have useful data cached? |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
333 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
|
334 base = self.cache[1] |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
335 start = self.start(base + 1) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
336 text = self.cache[2] |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
337 last = 0 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
338 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
339 f = self.opener(self.datafile) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
340 f.seek(start) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
341 data = f.read(end - start) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
342 |
| 651 | 343 if text is None: |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
344 last = self.length(base) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
345 text = decompress(data[:last]) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
346 |
|
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
347 bins = [] |
| 64 | 348 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
|
349 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
|
350 bins.append(decompress(data[last:last + s])) |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
351 last = last + s |
|
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 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
|
354 |
| 26 | 355 if node != hash(text, p1, p2): |
| 98 | 356 raise IOError("integrity check failed on %s:%d" |
| 357 % (self.datafile, rev)) | |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
358 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
359 self.cache = (node, rev, text) |
| 515 | 360 return text |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
361 |
| 644 | 362 def addrevision(self, text, transaction, link, p1=None, p2=None, d=None): |
| 1083 | 363 """add a revision to the log |
| 364 | |
| 365 text - the revision data to add | |
| 366 transaction - the transaction object used for rollback | |
| 367 link - the linkrev data to add | |
| 368 p1, p2 - the parent nodeids of the revision | |
| 369 d - an optional precomputed delta | |
| 370 """ | |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
371 if text is None: text = "" |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
372 if p1 is None: p1 = self.tip() |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
373 if p2 is None: p2 = nullid |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
374 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
375 node = hash(text, p1, p2) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
376 |
| 301 | 377 if node in self.nodemap: |
| 378 return node | |
| 379 | |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
380 n = self.count() |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
381 t = n - 1 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
382 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
383 if n: |
| 64 | 384 base = self.base(t) |
| 385 start = self.start(base) | |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
386 end = self.end(t) |
| 644 | 387 if not d: |
| 388 prev = self.revision(self.tip()) | |
| 389 d = self.diff(prev, text) | |
| 98 | 390 data = compress(d) |
| 64 | 391 dist = end - start + len(data) |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
392 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
393 # full versions are inserted when the needed deltas |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
394 # become comparable to the uncompressed text |
| 64 | 395 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
|
396 data = compress(text) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
397 base = n |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
398 else: |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
399 base = self.base(t) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
400 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
401 offset = 0 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
402 if t >= 0: |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
403 offset = self.end(t) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
404 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
405 e = (offset, len(data), base, link, p1, p2, node) |
| 515 | 406 |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
407 self.index.append(e) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
408 self.nodemap[node] = n |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
409 entry = struct.pack(indexformat, *e) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
410 |
| 26 | 411 transaction.add(self.datafile, e[0]) |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
412 self.opener(self.datafile, "a").write(data) |
| 41 | 413 transaction.add(self.indexfile, n * len(entry)) |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
414 self.opener(self.indexfile, "a").write(entry) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
415 |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
416 self.cache = (node, n, text) |
|
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
417 return node |
|
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 def ancestor(self, a, b): |
| 1083 | 420 """calculate the least common ancestor of nodes a and b""" |
| 147 | 421 # calculate the distance of every node from root |
| 422 dist = {nullid: 0} | |
| 423 for i in xrange(self.count()): | |
| 424 n = self.node(i) | |
| 425 p1, p2 = self.parents(n) | |
| 426 dist[n] = max(dist[p1], dist[p2]) + 1 | |
| 515 | 427 |
| 147 | 428 # traverse ancestors in order of decreasing distance from root |
| 429 def ancestors(node): | |
| 430 # we store negative distances because heap returns smallest member | |
| 431 h = [(-dist[node], node)] | |
| 432 seen = {} | |
| 433 earliest = self.count() | |
| 434 while h: | |
| 435 d, n = heapq.heappop(h) | |
| 436 if n not in seen: | |
| 437 seen[n] = 1 | |
| 381 | 438 r = self.rev(n) |
| 439 yield (-d, r, n) | |
| 147 | 440 for p in self.parents(n): |
| 441 heapq.heappush(h, (-dist[p], p)) | |
|
45
f2b2d5daec30
Fix recursion depth trouble with ancestor algorithm
mpm@selenic.com
parents:
41
diff
changeset
|
442 |
| 147 | 443 x = ancestors(a) |
| 444 y = ancestors(b) | |
| 445 lx = x.next() | |
| 446 ly = y.next() | |
|
45
f2b2d5daec30
Fix recursion depth trouble with ancestor algorithm
mpm@selenic.com
parents:
41
diff
changeset
|
447 |
| 147 | 448 # increment each ancestor list until it is closer to root than |
| 449 # the other, or they match | |
| 450 while 1: | |
| 451 if lx == ly: | |
| 381 | 452 return lx[2] |
| 147 | 453 elif lx < ly: |
| 454 ly = y.next() | |
| 455 elif lx > ly: | |
| 456 lx = x.next() | |
|
0
9117c6561b0b
Add back links from file revisions to changeset revisions
mpm@selenic.com
parents:
diff
changeset
|
457 |
| 46 | 458 def group(self, linkmap): |
| 1083 | 459 """calculate a delta group |
| 46 | 460 |
| 1083 | 461 Given a list of changeset revs, return a set of deltas and |
| 462 metadata corresponding to nodes. the first delta is | |
| 463 parent(nodes[0]) -> nodes[0] the receiver is guaranteed to | |
| 464 have this parent as it has all history before these | |
| 465 changesets. parent is parent[0] | |
| 466 """ | |
| 46 | 467 revs = [] |
| 468 needed = {} | |
| 469 | |
| 470 # find file nodes/revs that match changeset revs | |
| 471 for i in xrange(0, self.count()): | |
| 472 if self.index[i][3] in linkmap: | |
| 473 revs.append(i) | |
| 474 needed[i] = 1 | |
| 475 | |
| 476 # if we don't have any revisions touched by these changesets, bail | |
| 192 | 477 if not revs: |
| 478 yield struct.pack(">l", 0) | |
| 479 return | |
| 46 | 480 |
| 481 # add the parent of the first rev | |
| 482 p = self.parents(self.node(revs[0]))[0] | |
| 483 revs.insert(0, self.rev(p)) | |
| 484 | |
| 485 # for each delta that isn't contiguous in the log, we need to | |
| 486 # reconstruct the base, reconstruct the result, and then | |
| 487 # calculate the delta. We also need to do this where we've | |
| 488 # stored a full version and not a delta | |
| 489 for i in xrange(0, len(revs) - 1): | |
| 490 a, b = revs[i], revs[i + 1] | |
| 491 if a + 1 != b or self.base(b) == b: | |
| 492 for j in xrange(self.base(a), a + 1): | |
| 493 needed[j] = 1 | |
| 494 for j in xrange(self.base(b), b + 1): | |
| 495 needed[j] = 1 | |
| 496 | |
| 497 # calculate spans to retrieve from datafile | |
| 498 needed = needed.keys() | |
| 499 needed.sort() | |
| 500 spans = [] | |
| 192 | 501 oo = -1 |
| 502 ol = 0 | |
| 46 | 503 for n in needed: |
| 504 if n < 0: continue | |
| 505 o = self.start(n) | |
| 506 l = self.length(n) | |
| 192 | 507 if oo + ol == o: # can we merge with the previous? |
| 508 nl = spans[-1][2] | |
| 509 nl.append((n, l)) | |
| 510 ol += l | |
| 511 spans[-1] = (oo, ol, nl) | |
| 46 | 512 else: |
| 192 | 513 oo = o |
| 514 ol = l | |
| 515 spans.append((oo, ol, [(n, l)])) | |
| 46 | 516 |
| 517 # read spans in, divide up chunks | |
| 518 chunks = {} | |
| 192 | 519 for span in spans: |
| 46 | 520 # we reopen the file for each span to make http happy for now |
| 521 f = self.opener(self.datafile) | |
| 522 f.seek(span[0]) | |
| 523 data = f.read(span[1]) | |
| 524 | |
| 525 # divide up the span | |
| 526 pos = 0 | |
| 527 for r, l in span[2]: | |
| 192 | 528 chunks[r] = decompress(data[pos: pos + l]) |
| 46 | 529 pos += l |
| 530 | |
| 531 # helper to reconstruct intermediate versions | |
| 532 def construct(text, base, rev): | |
| 192 | 533 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
|
534 return mdiff.patches(text, bins) |
| 46 | 535 |
| 536 # build deltas | |
| 537 deltas = [] | |
|
71
47c9a869adee
Add mdiff.patches to speed up applying thousands of patches to the manifest
mpm@selenic.com
parents:
67
diff
changeset
|
538 for d in xrange(0, len(revs) - 1): |
| 46 | 539 a, b = revs[d], revs[d + 1] |
| 540 n = self.node(b) | |
| 192 | 541 |
| 542 # do we need to construct a new delta? | |
| 46 | 543 if a + 1 != b or self.base(b) == b: |
| 544 if a >= 0: | |
| 545 base = self.base(a) | |
| 192 | 546 ta = chunks[self.base(a)] |
| 46 | 547 ta = construct(ta, base, a) |
| 548 else: | |
| 549 ta = "" | |
| 515 | 550 |
| 46 | 551 base = self.base(b) |
| 552 if a > base: | |
| 553 base = a | |
| 554 tb = ta | |
| 555 else: | |
| 192 | 556 tb = chunks[self.base(b)] |
| 46 | 557 tb = construct(tb, base, b) |
| 558 d = self.diff(ta, tb) | |
| 559 else: | |
| 192 | 560 d = chunks[b] |
| 46 | 561 |
| 562 p = self.parents(n) | |
| 563 meta = n + p[0] + p[1] + linkmap[self.linkrev(n)] | |
| 564 l = struct.pack(">l", len(meta) + len(d) + 4) | |
| 192 | 565 yield l |
| 566 yield meta | |
| 567 yield d | |
| 46 | 568 |
| 192 | 569 yield struct.pack(">l", 0) |
| 570 | |
| 1062 | 571 def addgroup(self, revs, linkmapper, transaction, unique=0): |
| 1083 | 572 """ |
| 573 add a delta group | |
| 46 | 574 |
| 1083 | 575 given a set of deltas, add them to the revision log. the |
| 576 first delta is against its parent, which should be in our | |
| 577 log, the rest are against the previous delta. | |
| 578 """ | |
| 579 | |
| 580 #track the base of the current delta log | |
| 46 | 581 r = self.count() |
| 582 t = r - 1 | |
| 192 | 583 node = nullid |
| 515 | 584 |
| 655 | 585 base = prev = -1 |
|
653
94cdd02792b5
Fix corruption resulting from skipping parts of a revision group
Matt Mackall <mpm@selenic.com>
parents:
651
diff
changeset
|
586 start = end = measure = 0 |
| 46 | 587 if r: |
| 588 start = self.start(self.base(t)) | |
| 589 end = self.end(t) | |
| 590 measure = self.length(self.base(t)) | |
| 591 base = self.base(t) | |
| 592 prev = self.tip() | |
| 593 | |
| 594 transaction.add(self.datafile, end) | |
| 595 transaction.add(self.indexfile, r * struct.calcsize(indexformat)) | |
| 596 dfh = self.opener(self.datafile, "a") | |
| 597 ifh = self.opener(self.indexfile, "a") | |
| 598 | |
| 599 # loop through our set of deltas | |
| 192 | 600 chain = None |
| 601 for chunk in revs: | |
| 602 node, p1, p2, cs = struct.unpack("20s20s20s20s", chunk[:80]) | |
| 94 | 603 link = linkmapper(cs) |
| 77 | 604 if node in self.nodemap: |
|
224
ccbcc4d76f81
fix bad assumption about uniqueness of file versions
mpm@selenic.com
parents:
221
diff
changeset
|
605 # this can happen if two branches make the same change |
|
ccbcc4d76f81
fix bad assumption about uniqueness of file versions
mpm@selenic.com
parents:
221
diff
changeset
|
606 if unique: |
|
1073
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
607 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
|
608 chain = node |
|
224
ccbcc4d76f81
fix bad assumption about uniqueness of file versions
mpm@selenic.com
parents:
221
diff
changeset
|
609 continue |
| 192 | 610 delta = chunk[80:] |
| 611 | |
| 612 if not chain: | |
| 613 # retrieve the parent revision of the delta chain | |
| 614 chain = p1 | |
| 615 if not chain in self.nodemap: | |
|
1073
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
616 raise RevlogError("unknown base %s" % short(chain[:4])) |
| 46 | 617 |
| 618 # full versions are inserted when the needed deltas become | |
| 619 # comparable to the uncompressed text or when the previous | |
| 620 # version is not the one we have a delta against. We use | |
| 621 # the size of the previous full rev as a proxy for the | |
| 622 # current size. | |
| 623 | |
| 624 if chain == prev: | |
| 625 cdelta = compress(delta) | |
| 626 | |
| 627 if chain != prev or (end - start + len(cdelta)) > measure * 2: | |
| 628 # flush our writes here so we can read it in revision | |
| 629 dfh.flush() | |
| 630 ifh.flush() | |
|
65
d40cc5aacc31
Fix up a bunch of bugs in the new merge code
mpm@selenic.com
parents:
64
diff
changeset
|
631 text = self.revision(chain) |
| 73 | 632 text = self.patches(text, [delta]) |
| 46 | 633 chk = self.addrevision(text, transaction, link, p1, p2) |
| 634 if chk != node: | |
|
1073
7b35a980b982
[PATCH] raise exceptions with Exception subclasses
Bart Trojanowski <bart@jukie.net>
parents:
1062
diff
changeset
|
635 raise RevlogError("consistency error adding group") |
| 46 | 636 measure = len(text) |
| 637 else: | |
| 638 e = (end, len(cdelta), self.base(t), link, p1, p2, node) | |
| 639 self.index.append(e) | |
| 640 self.nodemap[node] = r | |
| 641 dfh.write(cdelta) | |
| 642 ifh.write(struct.pack(indexformat, *e)) | |
| 643 | |
|
65
d40cc5aacc31
Fix up a bunch of bugs in the new merge code
mpm@selenic.com
parents:
64
diff
changeset
|
644 t, r, chain, prev = r, r + 1, node, node |
| 46 | 645 start = self.start(self.base(t)) |
| 646 end = self.end(t) | |
| 647 | |
| 648 dfh.close() | |
| 649 ifh.close() | |
| 650 return node |