mercurial/changegroup.py
author Mads Kiilerich <mads@kiilerich.com>
Fri, 11 Nov 2011 01:07:10 +0100
branchstable
changeset 15497 9bea3aed6ee1
parent 14494 1ffeeb91c55d
child 16557 9dba55369cd8
permissions -rw-r--r--
add missing localization markup
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 _
14143
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
     9
from node import nullrev
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
    10
import mdiff, util
8312
b87a50b7125c separate import lines from mercurial and general python modules
Simon Heimberg <simohe@besonet.ch>
parents: 8226
diff changeset
    11
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
    12
14141
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
    13
_BUNDLE10_DELTA_HEADER = "20s20s20s20s"
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
    14
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    15
def readexactly(stream, n):
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    16
    '''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
    17
    s = stream.read(n)
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    18
    if len(s) < n:
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    19
        raise util.Abort(_("stream ended unexpectedly"
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    20
                           " (got %d bytes, expected %d)")
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    21
                          % (len(s), n))
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    22
    return s
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    23
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    24
def getchunk(stream):
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    25
    """return the next chunk from stream as a string"""
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    26
    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
    27
    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
    28
    if l <= 4:
13458
9f2c407caf34 changegroup: don't accept odd chunk headers
Mads Kiilerich <mads@kiilerich.com>
parents: 13457
diff changeset
    29
        if l:
9f2c407caf34 changegroup: don't accept odd chunk headers
Mads Kiilerich <mads@kiilerich.com>
parents: 13457
diff changeset
    30
            raise util.Abort(_("invalid chunk length %d") % l)
1981
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    31
        return ""
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
    32
    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
    33
5368
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
    34
def chunkheader(length):
9437
1c4e4004f3a6 Improve some docstrings relating to changegroups and prepush().
Greg Ward <greg-hg@gerg.ca>
parents: 9087
diff changeset
    35
    """return a changegroup chunk header (string)"""
5368
61462e7d62ed changegroup: avoid large copies
Matt Mackall <mpm@selenic.com>
parents: 3932
diff changeset
    36
    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
    37
736b6c96bbbc make incoming work via ssh (issue139); move chunk code into separate module.
Thomas Arendsen Hein <thomas@intevation.de>
parents:
diff changeset
    38
def closechunk():
9437
1c4e4004f3a6 Improve some docstrings relating to changegroups and prepush().
Greg Ward <greg-hg@gerg.ca>
parents: 9087
diff changeset
    39
    """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
    40
    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
    41
3659
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    42
class nocompress(object):
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    43
    def compress(self, x):
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    44
        return x
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    45
    def flush(self):
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    46
        return ""
025f68f22ae2 move write_bundle to changegroup.py
Matt Mackall <mpm@selenic.com>
parents: 2470
diff changeset
    47
3662
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    48
bundletypes = {
14060
aaa9a5989405 bundle: more comments about the different header types, remove useless if
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 13831
diff changeset
    49
    "": ("", nocompress), # only when using unbundle on ssh and old http servers
aaa9a5989405 bundle: more comments about the different header types, remove useless if
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 13831
diff changeset
    50
                          # since the unification ssh accepts a header but there
aaa9a5989405 bundle: more comments about the different header types, remove useless if
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 13831
diff changeset
    51
                          # is no capability signaling it.
3704
9c1737a3e254 fix writebundle for bz2 bundles
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 3662
diff changeset
    52
    "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
    53
    "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
    54
    "HG10GZ": ("HG10GZ", lambda: zlib.compressobj()),
3662
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
diff changeset
    55
}
f4dc02d7fb71 unduplicate bundle writing code from httprepo
Matt Mackall <mpm@selenic.com>
parents: 3660
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
14494
1ffeeb91c55d check-code: flag 0/1 used as constant Boolean expression
Martin Geisler <mg@lazybytes.net>
parents: 14144
diff changeset
    94
            while True:
12335
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):
14141
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   134
    deltaheader = _BUNDLE10_DELTA_HEADER
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   135
    deltaheadersize = struct.calcsize(deltaheader)
12043
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):
13459
acbe171c8fbe changegroup: fix typo introduced in 9f2c407caf34
Jim Hague <jim.hague@acm.org>
parents: 13458
diff changeset
   152
        d = readexactly(self._stream, 4)
13458
9f2c407caf34 changegroup: don't accept odd chunk headers
Mads Kiilerich <mads@kiilerich.com>
parents: 13457
diff changeset
   153
        l = struct.unpack(">l", d)[0]
9f2c407caf34 changegroup: don't accept odd chunk headers
Mads Kiilerich <mads@kiilerich.com>
parents: 13457
diff changeset
   154
        if l <= 4:
9f2c407caf34 changegroup: don't accept odd chunk headers
Mads Kiilerich <mads@kiilerich.com>
parents: 13457
diff changeset
   155
            if l:
9f2c407caf34 changegroup: don't accept odd chunk headers
Mads Kiilerich <mads@kiilerich.com>
parents: 13457
diff changeset
   156
                raise util.Abort(_("invalid chunk length %d") % l)
9f2c407caf34 changegroup: don't accept odd chunk headers
Mads Kiilerich <mads@kiilerich.com>
parents: 13457
diff changeset
   157
            return 0
9f2c407caf34 changegroup: don't accept odd chunk headers
Mads Kiilerich <mads@kiilerich.com>
parents: 13457
diff changeset
   158
        if self.callback:
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   159
            self.callback()
13458
9f2c407caf34 changegroup: don't accept odd chunk headers
Mads Kiilerich <mads@kiilerich.com>
parents: 13457
diff changeset
   160
        return l - 4
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   161
14144
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   162
    def changelogheader(self):
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   163
        """v10 does not have a changelog header chunk"""
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   164
        return {}
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   165
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   166
    def manifestheader(self):
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   167
        """v10 does not have a manifest header chunk"""
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   168
        return {}
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   169
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   170
    def filelogheader(self):
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   171
        """return the header of the filelogs chunk, v10 only has the filename"""
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   172
        l = self.chunklength()
14144
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   173
        if not l:
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   174
            return {}
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   175
        fname = readexactly(self._stream, l)
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   176
        return dict(filename=fname)
12334
50946802593d bundle: refactor progress callback
Matt Mackall <mpm@selenic.com>
parents: 12333
diff changeset
   177
14141
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   178
    def _deltaheader(self, headertuple, prevnode):
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   179
        node, p1, p2, cs = headertuple
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   180
        if prevnode is None:
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   181
            deltabase = p1
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   182
        else:
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   183
            deltabase = prevnode
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   184
        return node, p1, p2, deltabase, cs
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   185
14144
3c3c53d8343a unbundler: separate delta and header parsing
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14143
diff changeset
   186
    def deltachunk(self, prevnode):
12336
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   187
        l = self.chunklength()
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   188
        if not l:
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   189
            return {}
14141
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   190
        headerdata = readexactly(self._stream, self.deltaheadersize)
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   191
        header = struct.unpack(self.deltaheader, headerdata)
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   192
        delta = readexactly(self._stream, l - self.deltaheadersize)
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   193
        node, p1, p2, deltabase, cs = self._deltaheader(header, prevnode)
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   194
        return dict(node=node, p1=p1, p2=p2, cs=cs,
bd1cbfe5db5c bundler: make parsechunk return the base revision of the delta
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14060
diff changeset
   195
                    deltabase=deltabase, delta=delta)
12336
9d234f7d8a77 bundle: move chunk parsing into unbundle class
Matt Mackall <mpm@selenic.com>
parents: 12335
diff changeset
   196
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   197
class headerlessfixup(object):
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   198
    def __init__(self, fh, h):
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   199
        self._h = h
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   200
        self._fh = fh
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   201
    def read(self, n):
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   202
        if self._h:
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   203
            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
   204
            if len(d) < n:
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   205
                d += readexactly(self._fh, n - len(d))
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   206
            return d
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   207
        return readexactly(self._fh, n)
12329
7458de933f26 bundle: push chunkbuffer down into decompress
Matt Mackall <mpm@selenic.com>
parents: 12044
diff changeset
   208
6154
ef1c5a3b653d improve changegroup.readbundle(), use it in hgweb
Dirkjan Ochtman <dirkjan@ochtman.nl>
parents: 6152
diff changeset
   209
def readbundle(fh, fname):
13457
e74fe15dc7fd changegroup: verify all stream reads
Mads Kiilerich <mads@kiilerich.com>
parents: 13456
diff changeset
   210
    header = readexactly(fh, 6)
12042
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   211
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   212
    if not fname:
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   213
        fname = "stream"
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   214
        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
   215
            fh = headerlessfixup(fh, header)
12042
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   216
            header = "HG10UN"
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   217
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   218
    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
   219
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   220
    if magic != 'HG':
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   221
        raise util.Abort(_('%s: not a Mercurial bundle') % fname)
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   222
    if version != '10':
210049a8d16e bundle: unify/refactor unbundle/readbundle
Matt Mackall <mpm@selenic.com>
parents: 12041
diff changeset
   223
        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
   224
    return unbundle10(fh, alg)
13831
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   225
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   226
class bundle10(object):
14143
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   227
    deltaheader = _BUNDLE10_DELTA_HEADER
13831
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   228
    def __init__(self, lookup):
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   229
        self._lookup = lookup
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   230
    def close(self):
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   231
        return closechunk()
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   232
    def fileheader(self, fname):
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   233
        return chunkheader(len(fname)) + fname
14143
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   234
    def revchunk(self, revlog, rev, prev):
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   235
        node = revlog.node(rev)
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   236
        p1, p2 = revlog.parentrevs(rev)
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   237
        base = prev
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   238
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   239
        prefix = ''
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   240
        if base == nullrev:
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   241
            delta = revlog.revision(node)
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   242
            prefix = mdiff.trivialdiffheader(len(delta))
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   243
        else:
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   244
            delta = revlog.revdiff(base, rev)
13831
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   245
        linknode = self._lookup(revlog, node)
14143
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   246
        p1n, p2n = revlog.parents(node)
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   247
        basenode = revlog.node(base)
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   248
        meta = self.builddeltaheader(node, p1n, p2n, basenode, linknode)
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   249
        meta += prefix
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   250
        l = len(meta) + len(delta)
13831
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   251
        yield chunkheader(l)
d69c9510d648 changegroup: introduce bundler objects
Matt Mackall <mpm@selenic.com>
parents: 13786
diff changeset
   252
        yield meta
14143
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   253
        yield delta
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   254
    def builddeltaheader(self, node, p1n, p2n, basenode, linknode):
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   255
        # do nothing with basenode, it is implicitly the previous one in HG10
da635d3c5620 changegroup: new bundler API
Benoit Boissinot <benoit.boissinot@ens-lyon.org>
parents: 14141
diff changeset
   256
        return struct.pack(self.deltaheader, node, p1n, p2n, linknode)