mercurial/pure/parsers.py
author Matt Mackall <mpm@selenic.com>
Fri, 06 Mar 2015 17:00:42 -0600
changeset 24215 feddc5284724
parent 21809 e250b8300e6e
child 24634 4ece2847cf4c
permissions -rw-r--r--
manifest: move pure parsing code out of pure This lets us transition more smoothly.
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
7700
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
     1
# parsers.py - Python implementation of parsers.c
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
     2
#
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
     3
# Copyright 2009 Matt Mackall <mpm@selenic.com> and others
f410c552fa15 pure Python implementation of parsers.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: 7945
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.
7700
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
     7
24215
feddc5284724 manifest: move pure parsing code out of pure
Matt Mackall <mpm@selenic.com>
parents: 21809
diff changeset
     8
from mercurial.node import nullid
7945
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
     9
from mercurial import util
18567
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    10
import struct, zlib, cStringIO
7700
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    11
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    12
_pack = struct.pack
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    13
_unpack = struct.unpack
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    14
_compress = zlib.compress
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    15
_decompress = zlib.decompress
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    16
_sha = util.sha1
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    17
21809
e250b8300e6e parsers: inline fields of dirstate values in C version
Siddharth Agarwal <sid0@fb.com>
parents: 19652
diff changeset
    18
# Some code below makes tuples directly because it's more convenient. However,
e250b8300e6e parsers: inline fields of dirstate values in C version
Siddharth Agarwal <sid0@fb.com>
parents: 19652
diff changeset
    19
# code outside this module should always use dirstatetuple.
e250b8300e6e parsers: inline fields of dirstate values in C version
Siddharth Agarwal <sid0@fb.com>
parents: 19652
diff changeset
    20
def dirstatetuple(*x):
e250b8300e6e parsers: inline fields of dirstate values in C version
Siddharth Agarwal <sid0@fb.com>
parents: 19652
diff changeset
    21
    # x is a tuple
e250b8300e6e parsers: inline fields of dirstate values in C version
Siddharth Agarwal <sid0@fb.com>
parents: 19652
diff changeset
    22
    return x
e250b8300e6e parsers: inline fields of dirstate values in C version
Siddharth Agarwal <sid0@fb.com>
parents: 19652
diff changeset
    23
13261
20a54bdf2328 pure: update index parsing
Matt Mackall <mpm@selenic.com>
parents: 13253
diff changeset
    24
def parse_index2(data, inline):
7945
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    25
    def gettype(q):
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    26
        return int(q & 0xFFFF)
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    27
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    28
    def offset_type(offset, type):
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    29
        return long(long(offset) << 16 | type)
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    30
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    31
    indexformatng = ">Qiiiiii20s12x"
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    32
7700
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    33
    s = struct.calcsize(indexformatng)
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    34
    index = []
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    35
    cache = None
14995
8d928799dab5 parsers: remove redundant 'n' variable in parsers.parse_index2() (issue2935)
py4fun
parents: 14421
diff changeset
    36
    off = 0
13253
61c9bc3da402 revlog: remove lazy index
Matt Mackall <mpm@selenic.com>
parents: 10263
diff changeset
    37
7700
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    38
    l = len(data) - s
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    39
    append = index.append
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    40
    if inline:
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    41
        cache = (0, data)
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    42
        while off <= l:
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    43
            e = _unpack(indexformatng, data[off:off + s])
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    44
            append(e)
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    45
            if e[1] < 0:
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    46
                break
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    47
            off += e[1] + s
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    48
    else:
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    49
        while off <= l:
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    50
            e = _unpack(indexformatng, data[off:off + s])
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    51
            append(e)
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    52
            off += s
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    53
14421
639f26cab2f5 pure parsers: properly detect corrupt index files
Augie Fackler <durin42@gmail.com>
parents: 14064
diff changeset
    54
    if off != len(data):
639f26cab2f5 pure parsers: properly detect corrupt index files
Augie Fackler <durin42@gmail.com>
parents: 14064
diff changeset
    55
        raise ValueError('corrupt index file')
639f26cab2f5 pure parsers: properly detect corrupt index files
Augie Fackler <durin42@gmail.com>
parents: 14064
diff changeset
    56
13435
90d7ce986565 pure: fix index parsing on empty repositories
Wagner Bruna <wbruna@softwareexpress.com.br>
parents: 13261
diff changeset
    57
    if index:
90d7ce986565 pure: fix index parsing on empty repositories
Wagner Bruna <wbruna@softwareexpress.com.br>
parents: 13261
diff changeset
    58
        e = list(index[0])
90d7ce986565 pure: fix index parsing on empty repositories
Wagner Bruna <wbruna@softwareexpress.com.br>
parents: 13261
diff changeset
    59
        type = gettype(e[0])
90d7ce986565 pure: fix index parsing on empty repositories
Wagner Bruna <wbruna@softwareexpress.com.br>
parents: 13261
diff changeset
    60
        e[0] = offset_type(0, type)
90d7ce986565 pure: fix index parsing on empty repositories
Wagner Bruna <wbruna@softwareexpress.com.br>
parents: 13261
diff changeset
    61
        index[0] = tuple(e)
7700
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    62
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    63
    # add the magic null revision at -1
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    64
    index.append((0, 0, 0, -1, -1, -1, -1, nullid))
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    65
13261
20a54bdf2328 pure: update index parsing
Matt Mackall <mpm@selenic.com>
parents: 13253
diff changeset
    66
    return index, cache
7700
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    67
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    68
def parse_dirstate(dmap, copymap, st):
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    69
    parents = [st[:20], st[20: 40]]
17425
e95ec38f86b0 fix wording and not-completely-trivial spelling errors and bad docstrings
Mads Kiilerich <mads@kiilerich.com>
parents: 14995
diff changeset
    70
    # dereference fields so they will be local in loop
7945
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    71
    format = ">cllll"
94d7e14cfa42 pure/parsers: fix circular imports, import mercurial modules properly
Matt Mackall <mpm@selenic.com>
parents: 7873
diff changeset
    72
    e_size = struct.calcsize(format)
7700
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    73
    pos1 = 40
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    74
    l = len(st)
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    75
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    76
    # the inner loop
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    77
    while pos1 < l:
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    78
        pos2 = pos1 + e_size
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    79
        e = _unpack(">cllll", st[pos1:pos2]) # a literal here is faster
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    80
        pos1 = pos2 + e[4]
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    81
        f = st[pos2:pos1]
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    82
        if '\0' in f:
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    83
            f, c = f.split('\0')
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    84
            copymap[f] = c
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    85
        dmap[f] = e[:4]
f410c552fa15 pure Python implementation of parsers.c
Martin Geisler <mg@daimi.au.dk>
parents:
diff changeset
    86
    return parents
18567
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    87
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    88
def pack_dirstate(dmap, copymap, pl, now):
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    89
    now = int(now)
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    90
    cs = cStringIO.StringIO()
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    91
    write = cs.write
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    92
    write("".join(pl))
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    93
    for f, e in dmap.iteritems():
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    94
        if e[0] == 'n' and e[3] == now:
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    95
            # The file was last modified "simultaneously" with the current
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    96
            # write to dirstate (i.e. within the same second for file-
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    97
            # systems with a granularity of 1 sec). This commonly happens
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    98
            # for at least a couple of files on 'update'.
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
    99
            # The user could change the file without changing its size
19652
187bf2dde7c1 pack_dirstate: only invalidate mtime for files written in the last second
Siddharth Agarwal <sid0@fb.com>
parents: 18567
diff changeset
   100
            # within the same second. Invalidate the file's mtime in
18567
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
   101
            # dirstate, forcing future 'status' calls to compare the
19652
187bf2dde7c1 pack_dirstate: only invalidate mtime for files written in the last second
Siddharth Agarwal <sid0@fb.com>
parents: 18567
diff changeset
   102
            # contents of the file if the size is the same. This prevents
187bf2dde7c1 pack_dirstate: only invalidate mtime for files written in the last second
Siddharth Agarwal <sid0@fb.com>
parents: 18567
diff changeset
   103
            # mistakenly treating such files as clean.
21809
e250b8300e6e parsers: inline fields of dirstate values in C version
Siddharth Agarwal <sid0@fb.com>
parents: 19652
diff changeset
   104
            e = dirstatetuple(e[0], e[1], e[2], -1)
18567
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
   105
            dmap[f] = e
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
   106
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
   107
        if f in copymap:
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
   108
            f = "%s\0%s" % (f, copymap[f])
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
   109
        e = _pack(">cllll", e[0], e[1], e[2], e[3], len(f))
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
   110
        write(e)
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
   111
        write(f)
194e63c1ccb9 dirstate: move pure python dirstate packing to pure/parsers.py
Siddharth Agarwal <sid0@fb.com>
parents: 17425
diff changeset
   112
    return cs.getvalue()