mercurial/changegroup.py
author Martin Geisler <mg@lazybytes.net>
Sat, 23 Oct 2010 18:33:57 +0200
branchstable
changeset 12839 d85e30889f26
parent 12347 6277a9469dff
child 13456 ab3f4ee48adc
permissions -rw-r--r--
patchbomb: fix stray backslash in docstring While both '\ ' and '\\ ' parse the same in Python, the difference trips up hggettext so that it cannot find the docstring in the source file and thus cannot write the right line number to i18n/hg.pot. While the line number is not essential, it can be used to lookup the original message.
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
8226
8b2cd04a6e97 put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents: 8225
diff changeset
     1
# changegroup.py - Mercurial changegroup manipulation functions
8b2cd04a6e97 put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents: 8225
diff changeset
     2
#
8b2cd04a6e97 put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents: 8225
diff changeset
     3
#  Copyright 2006 Matt Mackall <mpm@selenic.com>
8b2cd04a6e97 put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents: 8225
diff changeset
     4
#
8b2cd04a6e97 put license and copyright info into comment blocks
Martin Geisler <mg@lazybytes.net>
parents: 8225
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: 9437
diff changeset
     6
# GNU General Public License version 2 or any later version.
3877
abaee83ce0a6 Replace demandload with new demandimport
Matt Mackall <mpm@selenic.com>
parents: 3859
diff changeset
     7
3891
6b4127c7d52a Simplify i18n imports
Matt Mackall <mpm@selenic.com>
parents: 3877
diff changeset
     8
from i18n import _
8312
b87a50b7125c separate import lines from mercurial and general python modules
Simon Heimberg <simohe@besonet.ch>
parents: 8226
diff changeset
     9
import util
b87a50b7125c separate import lines from mercurial and general python modules
Simon Heimberg <simohe@besonet.ch>
parents: 8226
diff changeset
    10
import struct, os, bz2, zlib, tempfile
1981
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    11
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    12
def getchunk(source):
9437
1c4e4004f3a6 Improve some docstrings relating to changegroups and prepush().
Greg Ward <greg-hg@gerg.ca>
parents: 9087
diff changeset
    13
    """return the next chunk from changegroup 'source' as a string"""
1981
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    14
    d = source.read(4)
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    15
    if not d:
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    16
        return ""
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    17
    l = struct.unpack(">l", d)[0]
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    18
    if l <= 4:
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    19
        return ""
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    20
    d = source.read(l - 4)
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    21
    if len(d) < l - 4:
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    22
        raise util.Abort(_("premature EOF reading chunk"
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    23
                           " (got %d bytes, expected %d)")
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    24
                          % (len(d), l - 4))
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    25
    return d
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    26
5368
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
    27
def chunkheader(length):
9437
1c4e4004f3a6 Improve some docstrings relating to changegroups and prepush().
Greg Ward <greg-hg@gerg.ca>
parents: 9087
diff changeset
    28
    """return a changegroup chunk header (string)"""
5368
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
    29
    return struct.pack(">l", length + 4)
1981
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    30
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    31
def closechunk():
9437
1c4e4004f3a6 Improve some docstrings relating to changegroups and prepush().
Greg Ward <greg-hg@gerg.ca>
parents: 9087
diff changeset
    32
    """return a changegroup chunk header (string) for a zero-length chunk"""
1981
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    33
    return struct.pack(">l", 0)
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    34
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    35
class nocompress(object):
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    36
    def compress(self, x):
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    37
        return x
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    38
    def flush(self):
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    39
        return ""
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    40
3662
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    41
bundletypes = {
3704
9c1737a3e254 fix writebundle for bz2 bundles
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 3662
diff changeset
    42
    "": ("", nocompress),
9c1737a3e254 fix writebundle for bz2 bundles
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 3662
diff changeset
    43
    "HG10UN": ("HG10UN", nocompress),
3762
b9d3e12da485 changegroup.py: delay the loading of the bz2 and zlib modules
Alexis S. L. Carvalho <alexis@cecm.usp.br>
parents: 3706
diff changeset
    44
    "HG10BZ": ("HG10", lambda: bz2.BZ2Compressor()),
b9d3e12da485 changegroup.py: delay the loading of the bz2 and zlib modules
Alexis S. L. Carvalho <alexis@cecm.usp.br>
parents: 3706
diff changeset
    45
    "HG10GZ": ("HG10GZ", lambda: zlib.compressobj()),
3662
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    46
}
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    47
10356
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    48
def collector(cl, mmfs, files):
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    49
    # Gather information about changeset nodes going out in a bundle.
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    50
    # We want to gather manifests needed and filelogs affected.
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    51
    def collect(node):
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    52
        c = cl.read(node)
11648
801533a52799 changegroup*(): use set instead of dict
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 10430
diff changeset
    53
        files.update(c[3])
10356
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    54
        mmfs.setdefault(c[0], node)
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    55
    return collect
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    56
9087
f48454a279b9 typos: "it's" -> "its"
Martin Geisler <mg@lazybytes.net>
parents: 8312
diff changeset
    57
# hgweb uses this list to communicate its preferred type
6152
c050548307a4 hgweb: use bundletypes from mercurial.changegroup
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 5906
diff changeset
    58
bundlepriority = ['HG10GZ', 'HG10BZ', 'HG10UN']
c050548307a4 hgweb: use bundletypes from mercurial.changegroup
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 5906
diff changeset
    59
3706
0d810798acb1 Use 'bundletype' instead of 'type' to not shadow built-in function.
Thomas Arendsen Hein <thomas@intevation.de>
parents: 3705
diff changeset
    60
def writebundle(cg, filename, bundletype):
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    61
    """Write a bundle file and return its filename.
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    62
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    63
    Existing files will not be overwritten.
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    64
    If no filename is specified, a temporary file is created.
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    65
    bz2 compression can be turned off.
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    66
    The bundle file will be deleted in case of errors.
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    67
    """
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    68
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    69
    fh = None
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    70
    cleanup = None
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    71
    try:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    72
        if filename:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    73
            fh = open(filename, "wb")
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    74
        else:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    75
            fd, filename = tempfile.mkstemp(prefix="hg-bundle-", suffix=".hg")
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    76
            fh = os.fdopen(fd, "wb")
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    77
        cleanup = filename
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    78
3706
0d810798acb1 Use 'bundletype' instead of 'type' to not shadow built-in function.
Thomas Arendsen Hein <thomas@intevation.de>
parents: 3705
diff changeset
    79
        header, compressor = bundletypes[bundletype]
3704
9c1737a3e254 fix writebundle for bz2 bundles
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 3662
diff changeset
    80
        fh.write(header)
9c1737a3e254 fix writebundle for bz2 bundles
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 3662
diff changeset
    81
        z = compressor()
3662
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    82
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    83
        # parse the changegroup data, otherwise we will block
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    84
        # in case of sshrepo because we don't know the end of the stream
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    85
12335
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    86
        # an empty chunkgroup is the end of the changegroup
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    87
        # a changegroup has at least 2 chunkgroups (changelog and manifest).
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    88
        # after that, an empty chunkgroup is the end of the changegroup
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    89
        empty = False
5906
0136d7f58982 allow the creation of bundles with empty changelog/manifest chunks
Alexis S. L. Carvalho <alexis@cecm.usp.br>
parents: 5368
diff changeset
    90
        count = 0
0136d7f58982 allow the creation of bundles with empty changelog/manifest chunks
Alexis S. L. Carvalho <alexis@cecm.usp.br>
parents: 5368
diff changeset
    91
        while not empty or count <= 2:
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    92
            empty = True
5906
0136d7f58982 allow the creation of bundles with empty changelog/manifest chunks
Alexis S. L. Carvalho <alexis@cecm.usp.br>
parents: 5368
diff changeset
    93
            count += 1
12335
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    94
            while 1:
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    95
                chunk = getchunk(cg)
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    96
                if not chunk:
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    97
                    break
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    98
                empty = False
5368
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
    99
                fh.write(z.compress(chunkheader(len(chunk))))
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   100
                pos = 0
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   101
                while pos < len(chunk):
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   102
                    next = pos + 2**20
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   103
                    fh.write(z.compress(chunk[pos:next]))
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   104
                    pos = next
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   105
            fh.write(z.compress(closechunk()))
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   106
        fh.write(z.flush())
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   107
        cleanup = None
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   108
        return filename
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   109
    finally:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   110
        if fh is not None:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   111
            fh.close()
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   112
        if cleanup is not None:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   113
            os.unlink(cleanup)
3660
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   114
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   115
def decompressor(fh, alg):
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   116
    if alg == 'UN':
6154
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   117
        return fh
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   118
    elif alg == 'GZ':
6154
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   119
        def generator(f):
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   120
            zd = zlib.decompressobj()
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   121
            for chunk in f:
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   122
                yield zd.decompress(chunk)
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   123
    elif alg == 'BZ':
3660
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   124
        def generator(f):
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   125
            zd = bz2.BZ2Decompressor()
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   126
            zd.decompress("BZ")
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   127
            for chunk in util.filechunkiter(f, 4096):
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   128
                yield zd.decompress(chunk)
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   129
    else:
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   130
        raise util.Abort("unknown bundle compression '%s'" % alg)
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   131
    return util.chunkbuffer(generator(fh))
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   132
12043
bef5effb3db0 bundle: introduce bundle class
Matt Mackall <mpm@selenic.com>
parents: 12042
diff changeset
   133
class unbundle10(object):
bef5effb3db0 bundle: introduce bundle class
Matt Mackall <mpm@selenic.com>
parents: 12042
diff changeset
   134
    def __init__(self, fh, alg):
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   135
        self._stream = decompressor(fh, alg)
12044
bcc7139521b7 bundlerepo: remove duplication of bundle decompressors
Matt Mackall <mpm@selenic.com>
parents: 12043
diff changeset
   136
        self._type = alg
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   137
        self.callback = None
12044
bcc7139521b7 bundlerepo: remove duplication of bundle decompressors
Matt Mackall <mpm@selenic.com>
parents: 12043
diff changeset
   138
    def compressed(self):
bcc7139521b7 bundlerepo: remove duplication of bundle decompressors
Matt Mackall <mpm@selenic.com>
parents: 12043
diff changeset
   139
        return self._type != 'UN'
12043
bef5effb3db0 bundle: introduce bundle class
Matt Mackall <mpm@selenic.com>
parents: 12042
diff changeset
   140
    def read(self, l):
bef5effb3db0 bundle: introduce bundle class
Matt Mackall <mpm@selenic.com>
parents: 12042
diff changeset
   141
        return self._stream.read(l)
12330
e527b8635881 bundle: make unbundle object seekable
Matt Mackall <mpm@selenic.com>
parents: 12329
diff changeset
   142
    def seek(self, pos):
e527b8635881 bundle: make unbundle object seekable
Matt Mackall <mpm@selenic.com>
parents: 12329
diff changeset
   143
        return self._stream.seek(pos)
e527b8635881 bundle: make unbundle object seekable
Matt Mackall <mpm@selenic.com>
parents: 12329
diff changeset
   144
    def tell(self):
12332
680fe77ab5b8 bundlerepo: use bundle objects everywhere
Matt Mackall <mpm@selenic.com>
parents: 12330
diff changeset
   145
        return self._stream.tell()
12347
6277a9469dff bundlerepo: restore close() method
Matt Mackall <mpm@selenic.com>
parents: 12336
diff changeset
   146
    def close(self):
6277a9469dff bundlerepo: restore close() method
Matt Mackall <mpm@selenic.com>
parents: 12336
diff changeset
   147
        return self._stream.close()
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   148
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   149
    def chunklength(self):
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   150
        d = self.read(4)
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   151
        if not d:
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   152
            return 0
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   153
        l = max(0, struct.unpack(">l", d)[0] - 4)
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   154
        if l and self.callback:
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   155
            self.callback()
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   156
        return l
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   157
12333
44c7dfc2f6a3 bundle: make getchunk() a method
Matt Mackall <mpm@selenic.com>
parents: 12332
diff changeset
   158
    def chunk(self):
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   159
        """return the next chunk from changegroup 'source' as a string"""
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   160
        l = self.chunklength()
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   161
        d = self.read(l)
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   162
        if len(d) < l:
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   163
            raise util.Abort(_("premature EOF reading chunk"
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   164
                               " (got %d bytes, expected %d)")
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   165
                             % (len(d), l))
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   166
        return d
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   167
12336
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   168
    def parsechunk(self):
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   169
        l = self.chunklength()
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   170
        if not l:
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   171
            return {}
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   172
        h = self.read(80)
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   173
        node, p1, p2, cs = struct.unpack("20s20s20s20s", h)
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   174
        data = self.read(l - 80)
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   175
        return dict(node=node, p1=p1, p2=p2, cs=cs, data=data)
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   176
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   177
class headerlessfixup(object):
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   178
    def __init__(self, fh, h):
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   179
        self._h = h
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   180
        self._fh = fh
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   181
    def read(self, n):
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   182
        if self._h:
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   183
            d, self._h = self._h[:n], self._h[n:]
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   184
            if len(d) < n:
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   185
                d += self._fh.read(n - len(d))
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   186
            return d
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   187
        return self._fh.read(n)
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   188
6154
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   189
def readbundle(fh, fname):
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   190
    header = fh.read(6)
12042
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   191
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   192
    if not fname:
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   193
        fname = "stream"
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   194
        if not header.startswith('HG') and header.startswith('\0'):
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   195
            fh = headerlessfixup(fh, header)
12042
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   196
            header = "HG10UN"
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   197
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   198
    magic, version, alg = header[0:2], header[2:4], header[4:6]
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   199
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   200
    if magic != 'HG':
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   201
        raise util.Abort(_('%s: not a Mercurial bundle') % fname)
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   202
    if version != '10':
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   203
        raise util.Abort(_('%s: unknown bundle version %s') % (fname, version))
12043
bef5effb3db0 bundle: introduce bundle class
Matt Mackall <mpm@selenic.com>
parents: 12042
diff changeset
   204
    return unbundle10(fh, alg)