mercurial/changegroup.py
author Mads Kiilerich <mads@kiilerich.com>
Tue, 22 Feb 2011 03:03:39 +0100
branchstable
changeset 13457 e74fe15dc7fd
parent 13456 ab3f4ee48adc
child 13458 9f2c407caf34
permissions -rw-r--r--
changegroup: verify all stream reads Mercurial often failed with struct.error or mpatch.mpatchError if incomplete data was received from a server. Now we validate all changegroup reads and aborts with abort: stream ended unexpectedly (got %d bytes, expected %d) if less than requested was read.
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
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    12
def readexactly(stream, n):
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    13
    '''read n bytes from stream.read and abort if less was available'''
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    14
    s = stream.read(n)
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    15
    if len(s) < n:
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    16
        raise util.Abort(_("stream ended unexpectedly"
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    17
                           " (got %d bytes, expected %d)")
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    18
                          % (len(s), n))
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    19
    return s
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    20
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    21
def getchunk(stream):
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    22
    """return the next chunk from stream as a string"""
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    23
    d = readexactly(stream, 4)
1981
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    24
    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
    25
    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
    26
        return ""
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    27
    return readexactly(stream, l - 4)
1981
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    28
5368
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
    29
def chunkheader(length):
9437
1c4e4004f3a6 Improve some docstrings relating to changegroups and prepush().
Greg Ward <greg-hg@gerg.ca>
parents: 9087
diff changeset
    30
    """return a changegroup chunk header (string)"""
5368
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
    31
    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
    32
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    33
def closechunk():
9437
1c4e4004f3a6 Improve some docstrings relating to changegroups and prepush().
Greg Ward <greg-hg@gerg.ca>
parents: 9087
diff changeset
    34
    """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
    35
    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
    36
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    37
class nocompress(object):
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    38
    def compress(self, x):
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    39
        return x
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    40
    def flush(self):
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    41
        return ""
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    42
3662
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    43
bundletypes = {
3704
9c1737a3e254 fix writebundle for bz2 bundles
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 3662
diff changeset
    44
    "": ("", nocompress),
9c1737a3e254 fix writebundle for bz2 bundles
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 3662
diff changeset
    45
    "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
    46
    "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
    47
    "HG10GZ": ("HG10GZ", lambda: zlib.compressobj()),
3662
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    48
}
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    49
10356
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    50
def collector(cl, mmfs, files):
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    51
    # 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
    52
    # 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
    53
    def collect(node):
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    54
        c = cl.read(node)
11648
801533a52799 changegroup*(): use set instead of dict
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 10430
diff changeset
    55
        files.update(c[3])
10356
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    56
        mmfs.setdefault(c[0], node)
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    57
    return collect
bc2414948012 localrepo: unify changegroup and changegroupsubset code paths a bit
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 10263
diff changeset
    58
9087
f48454a279b9 typos: "it's" -> "its"
Martin Geisler <mg@lazybytes.net>
parents: 8312
diff changeset
    59
# 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
    60
bundlepriority = ['HG10GZ', 'HG10BZ', 'HG10UN']
c050548307a4 hgweb: use bundletypes from mercurial.changegroup
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 5906
diff changeset
    61
3706
0d810798acb1 Use 'bundletype' instead of 'type' to not shadow built-in function.
Thomas Arendsen Hein <thomas@intevation.de>
parents: 3705
diff changeset
    62
def writebundle(cg, filename, bundletype):
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    63
    """Write a bundle file and return its filename.
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    64
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    65
    Existing files will not be overwritten.
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    66
    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
    67
    bz2 compression can be turned off.
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    68
    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
    69
    """
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    70
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    71
    fh = None
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    72
    cleanup = None
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    73
    try:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    74
        if filename:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    75
            fh = open(filename, "wb")
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    76
        else:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    77
            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
    78
            fh = os.fdopen(fd, "wb")
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    79
        cleanup = filename
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    80
3706
0d810798acb1 Use 'bundletype' instead of 'type' to not shadow built-in function.
Thomas Arendsen Hein <thomas@intevation.de>
parents: 3705
diff changeset
    81
        header, compressor = bundletypes[bundletype]
3704
9c1737a3e254 fix writebundle for bz2 bundles
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 3662
diff changeset
    82
        fh.write(header)
9c1737a3e254 fix writebundle for bz2 bundles
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 3662
diff changeset
    83
        z = compressor()
3662
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    84
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    85
        # parse the changegroup data, otherwise we will block
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    86
        # 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
    87
12335
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    88
        # an empty chunkgroup is the end of the changegroup
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    89
        # 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
    90
        # 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
    91
        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
    92
        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
    93
        while not empty or count <= 2:
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    94
            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
    95
            count += 1
12335
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    96
            while 1:
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    97
                chunk = getchunk(cg)
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    98
                if not chunk:
e21fe9c5fb25 bundle: get rid of chunkiter
Matt Mackall <mpm@selenic.com>
parents: 12334
diff changeset
    99
                    break
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   100
                empty = False
5368
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   101
                fh.write(z.compress(chunkheader(len(chunk))))
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   102
                pos = 0
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   103
                while pos < len(chunk):
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   104
                    next = pos + 2**20
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   105
                    fh.write(z.compress(chunk[pos:next]))
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
   106
                    pos = next
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   107
            fh.write(z.compress(closechunk()))
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   108
        fh.write(z.flush())
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   109
        cleanup = None
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   110
        return filename
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   111
    finally:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   112
        if fh is not None:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   113
            fh.close()
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   114
        if cleanup is not None:
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
   115
            os.unlink(cleanup)
3660
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   116
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   117
def decompressor(fh, alg):
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   118
    if alg == 'UN':
6154
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   119
        return fh
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   120
    elif alg == 'GZ':
6154
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   121
        def generator(f):
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   122
            zd = zlib.decompressobj()
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   123
            for chunk in f:
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   124
                yield zd.decompress(chunk)
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   125
    elif alg == 'BZ':
3660
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   126
        def generator(f):
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   127
            zd = bz2.BZ2Decompressor()
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   128
            zd.decompress("BZ")
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   129
            for chunk in util.filechunkiter(f, 4096):
8500a13ec44b create a readbundle function
Matt Mackall <mpm@selenic.com>
parents: 3659
diff changeset
   130
                yield zd.decompress(chunk)
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   131
    else:
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   132
        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
   133
    return util.chunkbuffer(generator(fh))
12041
270fb4d39153 bundle: factor out decompressor
Matt Mackall <mpm@selenic.com>
parents: 11648
diff changeset
   134
12043
bef5effb3db0 bundle: introduce bundle class
Matt Mackall <mpm@selenic.com>
parents: 12042
diff changeset
   135
class unbundle10(object):
bef5effb3db0 bundle: introduce bundle class
Matt Mackall <mpm@selenic.com>
parents: 12042
diff changeset
   136
    def __init__(self, fh, alg):
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   137
        self._stream = decompressor(fh, alg)
12044
bcc7139521b7 bundlerepo: remove duplication of bundle decompressors
Matt Mackall <mpm@selenic.com>
parents: 12043
diff changeset
   138
        self._type = alg
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   139
        self.callback = None
12044
bcc7139521b7 bundlerepo: remove duplication of bundle decompressors
Matt Mackall <mpm@selenic.com>
parents: 12043
diff changeset
   140
    def compressed(self):
bcc7139521b7 bundlerepo: remove duplication of bundle decompressors
Matt Mackall <mpm@selenic.com>
parents: 12043
diff changeset
   141
        return self._type != 'UN'
12043
bef5effb3db0 bundle: introduce bundle class
Matt Mackall <mpm@selenic.com>
parents: 12042
diff changeset
   142
    def read(self, l):
bef5effb3db0 bundle: introduce bundle class
Matt Mackall <mpm@selenic.com>
parents: 12042
diff changeset
   143
        return self._stream.read(l)
12330
e527b8635881 bundle: make unbundle object seekable
Matt Mackall <mpm@selenic.com>
parents: 12329
diff changeset
   144
    def seek(self, pos):
e527b8635881 bundle: make unbundle object seekable
Matt Mackall <mpm@selenic.com>
parents: 12329
diff changeset
   145
        return self._stream.seek(pos)
e527b8635881 bundle: make unbundle object seekable
Matt Mackall <mpm@selenic.com>
parents: 12329
diff changeset
   146
    def tell(self):
12332
680fe77ab5b8 bundlerepo: use bundle objects everywhere
Matt Mackall <mpm@selenic.com>
parents: 12330
diff changeset
   147
        return self._stream.tell()
12347
6277a9469dff bundlerepo: restore close() method
Matt Mackall <mpm@selenic.com>
parents: 12336
diff changeset
   148
    def close(self):
6277a9469dff bundlerepo: restore close() method
Matt Mackall <mpm@selenic.com>
parents: 12336
diff changeset
   149
        return self._stream.close()
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   150
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   151
    def chunklength(self):
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   152
        d = readexactly(self._stream, 4)
12334
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()
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   161
        return readexactly(self._stream, l)
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   162
12336
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   163
    def parsechunk(self):
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   164
        l = self.chunklength()
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   165
        if not l:
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   166
            return {}
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   167
        h = readexactly(self._stream, 80)
12336
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   168
        node, p1, p2, cs = struct.unpack("20s20s20s20s", h)
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   169
        data = readexactly(self._stream, l - 80)
12336
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   170
        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
   171
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   172
class headerlessfixup(object):
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   173
    def __init__(self, fh, h):
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   174
        self._h = h
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   175
        self._fh = fh
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   176
    def read(self, n):
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   177
        if self._h:
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   178
            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
   179
            if len(d) < n:
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   180
                d += readexactly(self._fh, n - len(d))
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   181
            return d
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   182
        return readexactly(self._fh, n)
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   183
6154
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   184
def readbundle(fh, fname):
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   185
    header = readexactly(fh, 6)
12042
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   186
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   187
    if not fname:
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   188
        fname = "stream"
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   189
        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
   190
            fh = headerlessfixup(fh, header)
12042
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   191
            header = "HG10UN"
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   192
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   193
    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
   194
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   195
    if magic != 'HG':
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   196
        raise util.Abort(_('%s: not a Mercurial bundle') % fname)
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   197
    if version != '10':
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   198
        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
   199
    return unbundle10(fh, alg)