mercurial/pure/mpatch.py
author Raphaël Gomès <rgomes@octobus.net>
Tue, 07 Nov 2023 16:59:37 +0100
branchstable
changeset 51128 26c57e7a0890
parent 49599 94a797032fc4
permissions -rw-r--r--
Added signature for changeset 27055614b685
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
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
#
46819
d4ba4d51f85f contributor: change mentions of mpm to olivia
Raphaël Gomès <rgomes@octobus.net>
parents: 45942
diff changeset
     3
# Copyright 2009 Olivia Mackall <olivia@selenic.com> and others
7699
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
25e572394f5c Update license to GPLv2+
Matt Mackall <mpm@selenic.com>
parents: 8225
diff changeset
     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
48873
5aafc3c5bdec py3: use io.BytesIO directly
Gregory Szorc <gregory.szorc@gmail.com>
parents: 46819
diff changeset
     9
import io
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
49599
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    12
from typing import (
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    13
    List,
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    14
    Tuple,
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    15
)
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    16
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
    17
48873
5aafc3c5bdec py3: use io.BytesIO directly
Gregory Szorc <gregory.szorc@gmail.com>
parents: 46819
diff changeset
    18
stringio = io.BytesIO
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    19
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
    20
28782
f736f98e16ca mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents: 28589
diff changeset
    21
class mpatchError(Exception):
45942
89a2afe31e82 formating: upgrade to black 20.8b1
Augie Fackler <raf@durin42.com>
parents: 43077
diff changeset
    22
    """error raised when a delta cannot be decoded"""
28782
f736f98e16ca mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents: 28589
diff changeset
    23
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
    24
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    25
# 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
    26
# 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
    27
# 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
    28
# pointers to fragments with fragment lists.
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    29
#
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    30
# 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
    31
# 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
    32
# 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
    33
# temporary string buffers.
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    34
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
    35
49599
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    36
def _pull(
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    37
    dst: List[Tuple[int, int]], src: List[Tuple[int, int]], l: int
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    38
) -> None:  # pull l bytes from src
28587
76d7cab13f04 mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents: 27337
diff changeset
    39
    while l:
76d7cab13f04 mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents: 27337
diff changeset
    40
        f = src.pop()
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
    41
        if f[0] > l:  # do we need to split?
28587
76d7cab13f04 mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents: 27337
diff changeset
    42
            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
    43
            dst.append((l, f[1]))
76d7cab13f04 mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents: 27337
diff changeset
    44
            return
76d7cab13f04 mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents: 27337
diff changeset
    45
        dst.append(f)
76d7cab13f04 mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents: 27337
diff changeset
    46
        l -= f[0]
76d7cab13f04 mpatch: move pull() method to top level
Augie Fackler <augie@google.com>
parents: 27337
diff changeset
    47
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
    48
49599
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    49
def _move(m: stringio, dest: int, src: int, count: int) -> None:
28588
6546afde350e mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents: 28587
diff changeset
    50
    """move count bytes from src to dest
6546afde350e mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents: 28587
diff changeset
    51
6546afde350e mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents: 28587
diff changeset
    52
    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
    53
    """
6546afde350e mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents: 28587
diff changeset
    54
    m.seek(src)
6546afde350e mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents: 28587
diff changeset
    55
    buf = m.read(count)
6546afde350e mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents: 28587
diff changeset
    56
    m.seek(dest)
6546afde350e mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents: 28587
diff changeset
    57
    m.write(buf)
6546afde350e mpatch: un-nest the move() method
Augie Fackler <augie@google.com>
parents: 28587
diff changeset
    58
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
    59
49599
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    60
def _collect(
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    61
    m: stringio, buf: int, list: List[Tuple[int, int]]
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    62
) -> Tuple[int, int]:
28589
c4c7be9f0554 mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents: 28588
diff changeset
    63
    start = buf
c4c7be9f0554 mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents: 28588
diff changeset
    64
    for l, p in reversed(list):
c4c7be9f0554 mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents: 28588
diff changeset
    65
        _move(m, buf, p, l)
c4c7be9f0554 mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents: 28588
diff changeset
    66
        buf += l
c4c7be9f0554 mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents: 28588
diff changeset
    67
    return (buf - start, start)
c4c7be9f0554 mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents: 28588
diff changeset
    68
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
    69
49599
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
    70
def patches(a: bytes, bins: List[bytes]) -> bytes:
10282
08a0f04b56bd many, many trivial check-code fixups
Matt Mackall <mpm@selenic.com>
parents: 10263
diff changeset
    71
    if not bins:
08a0f04b56bd many, many trivial check-code fixups
Matt Mackall <mpm@selenic.com>
parents: 10263
diff changeset
    72
        return a
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    73
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    74
    plens = [len(x) for x in bins]
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    75
    pl = sum(plens)
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    76
    bl = len(a) + pl
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
    77
    tl = bl + bl + pl  # enough for the patches and two working texts
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    78
    b1, b2 = 0, bl
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    79
10282
08a0f04b56bd many, many trivial check-code fixups
Matt Mackall <mpm@selenic.com>
parents: 10263
diff changeset
    80
    if not tl:
08a0f04b56bd many, many trivial check-code fixups
Matt Mackall <mpm@selenic.com>
parents: 10263
diff changeset
    81
        return a
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    82
28861
86db5cb55d46 pycompat: switch to util.stringio for py3 compat
timeless <timeless@mozdev.org>
parents: 28782
diff changeset
    83
    m = stringio()
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    84
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    85
    # load our original text
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    86
    m.write(a)
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    87
    frags = [(len(a), b1)]
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    88
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    89
    # 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
    90
    pos = b2 + bl
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    91
    m.seek(pos)
34435
5326e4ef1dab style: never put multiple statements on one line
Alex Gaynor <agaynor@mozilla.com>
parents: 32512
diff changeset
    92
    for p in bins:
5326e4ef1dab style: never put multiple statements on one line
Alex Gaynor <agaynor@mozilla.com>
parents: 32512
diff changeset
    93
        m.write(p)
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    94
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    95
    for plen in plens:
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    96
        # if our list gets too long, execute it
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    97
        if len(frags) > 128:
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    98
            b2, b1 = b1, b2
28589
c4c7be9f0554 mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents: 28588
diff changeset
    99
            frags = [_collect(m, b1, frags)]
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   100
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   101
        new = []
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   102
        end = pos + plen
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   103
        last = 0
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   104
        while pos < end:
7775
5280c39778b6 pure/mpatch: use StringIO instead of mmap (issue1493)
Martin Geisler <mg@daimi.au.dk>
parents: 7699
diff changeset
   105
            m.seek(pos)
28782
f736f98e16ca mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents: 28589
diff changeset
   106
            try:
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
   107
                p1, p2, l = struct.unpack(b">lll", m.read(12))
28782
f736f98e16ca mpatch: unify mpatchError (issue5182)
timeless <timeless@mozdev.org>
parents: 28589
diff changeset
   108
            except struct.error:
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
   109
                raise mpatchError(b"patch cannot be decoded")
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
   110
            _pull(new, frags, p1 - last)  # what didn't change
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
   111
            _pull([], frags, p2 - p1)  # what got deleted
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
   112
            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
   113
            pos += l + 12
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   114
            last = p2
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
   115
        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
   116
28589
c4c7be9f0554 mpatch: move collect() to module level
Augie Fackler <augie@google.com>
parents: 28588
diff changeset
   117
    t = _collect(m, b2, frags)
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   118
7775
5280c39778b6 pure/mpatch: use StringIO instead of mmap (issue1493)
Martin Geisler <mg@daimi.au.dk>
parents: 7699
diff changeset
   119
    m.seek(t[1])
5280c39778b6 pure/mpatch: use StringIO instead of mmap (issue1493)
Martin Geisler <mg@daimi.au.dk>
parents: 7699
diff changeset
   120
    return m.read(t[0])
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   121
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
   122
49599
94a797032fc4 typing: add type hints to mpatch implementations
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
   123
def patchedsize(orig: int, delta: bytes) -> int:
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   124
    outlen, last, bin = 0, 0, 0
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   125
    binend = len(delta)
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   126
    data = 12
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   127
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   128
    while data <= binend:
43076
2372284d9457 formatting: blacken the codebase
Augie Fackler <augie@google.com>
parents: 36958
diff changeset
   129
        decode = delta[bin : bin + 12]
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
   130
        start, end, length = struct.unpack(b">lll", decode)
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   131
        if start > end:
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   132
            break
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   133
        bin = data + length
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   134
        data = bin + 12
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   135
        outlen += start - last
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   136
        last = end
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   137
        outlen += length
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   138
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   139
    if bin != binend:
43077
687b865b95ad formatting: byteify all mercurial/ and hgext/ string literals
Augie Fackler <augie@google.com>
parents: 43076
diff changeset
   140
        raise mpatchError(b"patch cannot be decoded")
7699
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   141
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   142
    outlen += orig - last
fac054f84600 pure Python implementation of mpatch.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
   143
    return outlen