author | Jun Wu <quark@fb.com> |
Wed, 04 May 2016 18:18:24 +0100 | |
changeset 29070 | 29a7d591ff42 |
parent 28861 | 86db5cb55d46 |
child 29695 | f2846d546645 |
permissions | -rw-r--r-- |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
1 |
# mpatch.py - Python implementation of mpatch.c |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
2 |
# |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
3 |
# Copyright 2009 Matt Mackall <mpm@selenic.com> and others |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
4 |
# |
8225
46293a0c7e9f
updated license to be explicit about GPL version 2
Martin Geisler <mg@lazybytes.net>
parents:
7775
diff
changeset
|
5 |
# This software may be used and distributed according to the terms of the |
10263 | 6 |
# GNU General Public License version 2 or any later version. |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
7 |
|
27337
9a17576103a4
mpatch: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents:
16683
diff
changeset
|
8 |
from __future__ import absolute_import |
9a17576103a4
mpatch: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents:
16683
diff
changeset
|
9 |
|
7775
5280c39778b6
pure/mpatch: use StringIO instead of mmap (issue1493)
Martin Geisler <mg@daimi.au.dk>
parents:
7699
diff
changeset
|
10 |
import struct |
27337
9a17576103a4
mpatch: use absolute_import
Gregory Szorc <gregory.szorc@gmail.com>
parents:
16683
diff
changeset
|
11 |
|
28861
86db5cb55d46
pycompat: switch to util.stringio for py3 compat
timeless <timeless@mozdev.org>
parents:
28782
diff
changeset
|
12 |
from . import pycompat |
86db5cb55d46
pycompat: switch to util.stringio for py3 compat
timeless <timeless@mozdev.org>
parents:
28782
diff
changeset
|
13 |
stringio = pycompat.stringio |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
14 |
|
28782
f736f98e16ca
mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents:
28589
diff
changeset
|
15 |
class mpatchError(Exception): |
f736f98e16ca
mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents:
28589
diff
changeset
|
16 |
"""error raised when a delta cannot be decoded |
f736f98e16ca
mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents:
28589
diff
changeset
|
17 |
""" |
f736f98e16ca
mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents:
28589
diff
changeset
|
18 |
|
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
19 |
# This attempts to apply a series of patches in time proportional to |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
20 |
# the total size of the patches, rather than patches * len(text). This |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
21 |
# means rather than shuffling strings around, we shuffle around |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
22 |
# pointers to fragments with fragment lists. |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
23 |
# |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
24 |
# When the fragment lists get too long, we collapse them. To do this |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
25 |
# efficiently, we do all our operations inside a buffer created by |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
26 |
# mmap and simply use memmove. This avoids creating a bunch of large |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
27 |
# temporary string buffers. |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
28 |
|
28587
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
29 |
def _pull(dst, src, l): # pull l bytes from src |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
30 |
while l: |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
31 |
f = src.pop() |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
32 |
if f[0] > l: # do we need to split? |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
33 |
src.append((f[0] - l, f[1] + l)) |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
34 |
dst.append((l, f[1])) |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
35 |
return |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
36 |
dst.append(f) |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
37 |
l -= f[0] |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
38 |
|
28588
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
39 |
def _move(m, dest, src, count): |
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
40 |
"""move count bytes from src to dest |
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
41 |
|
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
42 |
The file pointer is left at the end of dest. |
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
43 |
""" |
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
44 |
m.seek(src) |
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
45 |
buf = m.read(count) |
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
46 |
m.seek(dest) |
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
47 |
m.write(buf) |
6546afde350e
mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents:
28587
diff
changeset
|
48 |
|
28589
c4c7be9f0554
mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents:
28588
diff
changeset
|
49 |
def _collect(m, buf, list): |
c4c7be9f0554
mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents:
28588
diff
changeset
|
50 |
start = buf |
c4c7be9f0554
mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents:
28588
diff
changeset
|
51 |
for l, p in reversed(list): |
c4c7be9f0554
mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents:
28588
diff
changeset
|
52 |
_move(m, buf, p, l) |
c4c7be9f0554
mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents:
28588
diff
changeset
|
53 |
buf += l |
c4c7be9f0554
mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents:
28588
diff
changeset
|
54 |
return (buf - start, start) |
c4c7be9f0554
mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents:
28588
diff
changeset
|
55 |
|
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
56 |
def patches(a, bins): |
10282
08a0f04b56bd
many, many trivial check-code fixups
Matt Mackall <mpm@selenic.com>
parents:
10263
diff
changeset
|
57 |
if not bins: |
08a0f04b56bd
many, many trivial check-code fixups
Matt Mackall <mpm@selenic.com>
parents:
10263
diff
changeset
|
58 |
return a |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
59 |
|
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
60 |
plens = [len(x) for x in bins] |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
61 |
pl = sum(plens) |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
62 |
bl = len(a) + pl |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
63 |
tl = bl + bl + pl # enough for the patches and two working texts |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
64 |
b1, b2 = 0, bl |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
65 |
|
10282
08a0f04b56bd
many, many trivial check-code fixups
Matt Mackall <mpm@selenic.com>
parents:
10263
diff
changeset
|
66 |
if not tl: |
08a0f04b56bd
many, many trivial check-code fixups
Matt Mackall <mpm@selenic.com>
parents:
10263
diff
changeset
|
67 |
return a |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
68 |
|
28861
86db5cb55d46
pycompat: switch to util.stringio for py3 compat
timeless <timeless@mozdev.org>
parents:
28782
diff
changeset
|
69 |
m = stringio() |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
70 |
|
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
71 |
# load our original text |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
72 |
m.write(a) |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
73 |
frags = [(len(a), b1)] |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
74 |
|
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
75 |
# copy all the patches into our segment so we can memmove from them |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
76 |
pos = b2 + bl |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
77 |
m.seek(pos) |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
78 |
for p in bins: m.write(p) |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
79 |
|
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
80 |
for plen in plens: |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
81 |
# if our list gets too long, execute it |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
82 |
if len(frags) > 128: |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
83 |
b2, b1 = b1, b2 |
28589
c4c7be9f0554
mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents:
28588
diff
changeset
|
84 |
frags = [_collect(m, b1, frags)] |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
85 |
|
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
86 |
new = [] |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
87 |
end = pos + plen |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
88 |
last = 0 |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
89 |
while pos < end: |
7775
5280c39778b6
pure/mpatch: use StringIO instead of mmap (issue1493)
Martin Geisler <mg@daimi.au.dk>
parents:
7699
diff
changeset
|
90 |
m.seek(pos) |
28782
f736f98e16ca
mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents:
28589
diff
changeset
|
91 |
try: |
f736f98e16ca
mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents:
28589
diff
changeset
|
92 |
p1, p2, l = struct.unpack(">lll", m.read(12)) |
f736f98e16ca
mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents:
28589
diff
changeset
|
93 |
except struct.error: |
f736f98e16ca
mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents:
28589
diff
changeset
|
94 |
raise mpatchError("patch cannot be decoded") |
28587
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
95 |
_pull(new, frags, p1 - last) # what didn't change |
76d7cab13f04
mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents:
27337
diff
changeset
|
96 |
_pull([], frags, p2 - p1) # what got deleted |
16683 | 97 |
new.append((l, pos + 12)) # what got added |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
98 |
pos += l + 12 |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
99 |
last = p2 |
16683 | 100 |
frags.extend(reversed(new)) # what was left at the end |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
101 |
|
28589
c4c7be9f0554
mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents:
28588
diff
changeset
|
102 |
t = _collect(m, b2, frags) |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
103 |
|
7775
5280c39778b6
pure/mpatch: use StringIO instead of mmap (issue1493)
Martin Geisler <mg@daimi.au.dk>
parents:
7699
diff
changeset
|
104 |
m.seek(t[1]) |
5280c39778b6
pure/mpatch: use StringIO instead of mmap (issue1493)
Martin Geisler <mg@daimi.au.dk>
parents:
7699
diff
changeset
|
105 |
return m.read(t[0]) |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
106 |
|
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
107 |
def patchedsize(orig, delta): |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
108 |
outlen, last, bin = 0, 0, 0 |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
109 |
binend = len(delta) |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
110 |
data = 12 |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
111 |
|
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
112 |
while data <= binend: |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
113 |
decode = delta[bin:bin + 12] |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
114 |
start, end, length = struct.unpack(">lll", decode) |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
115 |
if start > end: |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
116 |
break |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
117 |
bin = data + length |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
118 |
data = bin + 12 |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
119 |
outlen += start - last |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
120 |
last = end |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
121 |
outlen += length |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
122 |
|
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
123 |
if bin != binend: |
28782
f736f98e16ca
mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents:
28589
diff
changeset
|
124 |
raise mpatchError("patch cannot be decoded") |
7699
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
125 |
|
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
126 |
outlen += orig - last |
fac054f84600
pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff
changeset
|
127 |
return outlen |