Show More
@@ -1,205 +1,205 b'' | |||||
1 | # changegroup.py - Mercurial changegroup manipulation functions |
|
1 | # changegroup.py - Mercurial changegroup manipulation functions | |
2 | # |
|
2 | # | |
3 | # Copyright 2006 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2006 Matt Mackall <mpm@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from i18n import _ |
|
8 | from i18n import _ | |
9 | import util |
|
9 | import util | |
10 | import struct, os, bz2, zlib, tempfile |
|
10 | import struct, os, bz2, zlib, tempfile | |
11 |
|
11 | |||
12 | def readexactly(stream, n): |
|
12 | def readexactly(stream, n): | |
13 | '''read n bytes from stream.read and abort if less was available''' |
|
13 | '''read n bytes from stream.read and abort if less was available''' | |
14 | s = stream.read(n) |
|
14 | s = stream.read(n) | |
15 | if len(s) < n: |
|
15 | if len(s) < n: | |
16 | raise util.Abort(_("stream ended unexpectedly" |
|
16 | raise util.Abort(_("stream ended unexpectedly" | |
17 | " (got %d bytes, expected %d)") |
|
17 | " (got %d bytes, expected %d)") | |
18 | % (len(s), n)) |
|
18 | % (len(s), n)) | |
19 | return s |
|
19 | return s | |
20 |
|
20 | |||
21 | def getchunk(stream): |
|
21 | def getchunk(stream): | |
22 | """return the next chunk from stream as a string""" |
|
22 | """return the next chunk from stream as a string""" | |
23 | d = readexactly(stream, 4) |
|
23 | d = readexactly(stream, 4) | |
24 | l = struct.unpack(">l", d)[0] |
|
24 | l = struct.unpack(">l", d)[0] | |
25 | if l <= 4: |
|
25 | if l <= 4: | |
26 | if l: |
|
26 | if l: | |
27 | raise util.Abort(_("invalid chunk length %d") % l) |
|
27 | raise util.Abort(_("invalid chunk length %d") % l) | |
28 | return "" |
|
28 | return "" | |
29 | return readexactly(stream, l - 4) |
|
29 | return readexactly(stream, l - 4) | |
30 |
|
30 | |||
31 | def chunkheader(length): |
|
31 | def chunkheader(length): | |
32 | """return a changegroup chunk header (string)""" |
|
32 | """return a changegroup chunk header (string)""" | |
33 | return struct.pack(">l", length + 4) |
|
33 | return struct.pack(">l", length + 4) | |
34 |
|
34 | |||
35 | def closechunk(): |
|
35 | def closechunk(): | |
36 | """return a changegroup chunk header (string) for a zero-length chunk""" |
|
36 | """return a changegroup chunk header (string) for a zero-length chunk""" | |
37 | return struct.pack(">l", 0) |
|
37 | return struct.pack(">l", 0) | |
38 |
|
38 | |||
39 | class nocompress(object): |
|
39 | class nocompress(object): | |
40 | def compress(self, x): |
|
40 | def compress(self, x): | |
41 | return x |
|
41 | return x | |
42 | def flush(self): |
|
42 | def flush(self): | |
43 | return "" |
|
43 | return "" | |
44 |
|
44 | |||
45 | bundletypes = { |
|
45 | bundletypes = { | |
46 | "": ("", nocompress), |
|
46 | "": ("", nocompress), | |
47 | "HG10UN": ("HG10UN", nocompress), |
|
47 | "HG10UN": ("HG10UN", nocompress), | |
48 | "HG10BZ": ("HG10", lambda: bz2.BZ2Compressor()), |
|
48 | "HG10BZ": ("HG10", lambda: bz2.BZ2Compressor()), | |
49 | "HG10GZ": ("HG10GZ", lambda: zlib.compressobj()), |
|
49 | "HG10GZ": ("HG10GZ", lambda: zlib.compressobj()), | |
50 | } |
|
50 | } | |
51 |
|
51 | |||
52 | def collector(cl, mmfs, files): |
|
52 | def collector(cl, mmfs, files): | |
53 | # Gather information about changeset nodes going out in a bundle. |
|
53 | # Gather information about changeset nodes going out in a bundle. | |
54 | # We want to gather manifests needed and filelogs affected. |
|
54 | # We want to gather manifests needed and filelogs affected. | |
55 | def collect(node): |
|
55 | def collect(node): | |
56 | c = cl.read(node) |
|
56 | c = cl.read(node) | |
57 | files.update(c[3]) |
|
57 | files.update(c[3]) | |
58 | mmfs.setdefault(c[0], node) |
|
58 | mmfs.setdefault(c[0], node) | |
59 | return collect |
|
59 | return collect | |
60 |
|
60 | |||
61 | # hgweb uses this list to communicate its preferred type |
|
61 | # hgweb uses this list to communicate its preferred type | |
62 | bundlepriority = ['HG10GZ', 'HG10BZ', 'HG10UN'] |
|
62 | bundlepriority = ['HG10GZ', 'HG10BZ', 'HG10UN'] | |
63 |
|
63 | |||
64 | def writebundle(cg, filename, bundletype): |
|
64 | def writebundle(cg, filename, bundletype): | |
65 | """Write a bundle file and return its filename. |
|
65 | """Write a bundle file and return its filename. | |
66 |
|
66 | |||
67 | Existing files will not be overwritten. |
|
67 | Existing files will not be overwritten. | |
68 | If no filename is specified, a temporary file is created. |
|
68 | If no filename is specified, a temporary file is created. | |
69 | bz2 compression can be turned off. |
|
69 | bz2 compression can be turned off. | |
70 | The bundle file will be deleted in case of errors. |
|
70 | The bundle file will be deleted in case of errors. | |
71 | """ |
|
71 | """ | |
72 |
|
72 | |||
73 | fh = None |
|
73 | fh = None | |
74 | cleanup = None |
|
74 | cleanup = None | |
75 | try: |
|
75 | try: | |
76 | if filename: |
|
76 | if filename: | |
77 | fh = open(filename, "wb") |
|
77 | fh = open(filename, "wb") | |
78 | else: |
|
78 | else: | |
79 | fd, filename = tempfile.mkstemp(prefix="hg-bundle-", suffix=".hg") |
|
79 | fd, filename = tempfile.mkstemp(prefix="hg-bundle-", suffix=".hg") | |
80 | fh = os.fdopen(fd, "wb") |
|
80 | fh = os.fdopen(fd, "wb") | |
81 | cleanup = filename |
|
81 | cleanup = filename | |
82 |
|
82 | |||
83 | header, compressor = bundletypes[bundletype] |
|
83 | header, compressor = bundletypes[bundletype] | |
84 | fh.write(header) |
|
84 | fh.write(header) | |
85 | z = compressor() |
|
85 | z = compressor() | |
86 |
|
86 | |||
87 | # parse the changegroup data, otherwise we will block |
|
87 | # parse the changegroup data, otherwise we will block | |
88 | # in case of sshrepo because we don't know the end of the stream |
|
88 | # in case of sshrepo because we don't know the end of the stream | |
89 |
|
89 | |||
90 | # an empty chunkgroup is the end of the changegroup |
|
90 | # an empty chunkgroup is the end of the changegroup | |
91 | # a changegroup has at least 2 chunkgroups (changelog and manifest). |
|
91 | # a changegroup has at least 2 chunkgroups (changelog and manifest). | |
92 | # after that, an empty chunkgroup is the end of the changegroup |
|
92 | # after that, an empty chunkgroup is the end of the changegroup | |
93 | empty = False |
|
93 | empty = False | |
94 | count = 0 |
|
94 | count = 0 | |
95 | while not empty or count <= 2: |
|
95 | while not empty or count <= 2: | |
96 | empty = True |
|
96 | empty = True | |
97 | count += 1 |
|
97 | count += 1 | |
98 | while 1: |
|
98 | while 1: | |
99 | chunk = getchunk(cg) |
|
99 | chunk = getchunk(cg) | |
100 | if not chunk: |
|
100 | if not chunk: | |
101 | break |
|
101 | break | |
102 | empty = False |
|
102 | empty = False | |
103 | fh.write(z.compress(chunkheader(len(chunk)))) |
|
103 | fh.write(z.compress(chunkheader(len(chunk)))) | |
104 | pos = 0 |
|
104 | pos = 0 | |
105 | while pos < len(chunk): |
|
105 | while pos < len(chunk): | |
106 | next = pos + 2**20 |
|
106 | next = pos + 2**20 | |
107 | fh.write(z.compress(chunk[pos:next])) |
|
107 | fh.write(z.compress(chunk[pos:next])) | |
108 | pos = next |
|
108 | pos = next | |
109 | fh.write(z.compress(closechunk())) |
|
109 | fh.write(z.compress(closechunk())) | |
110 | fh.write(z.flush()) |
|
110 | fh.write(z.flush()) | |
111 | cleanup = None |
|
111 | cleanup = None | |
112 | return filename |
|
112 | return filename | |
113 | finally: |
|
113 | finally: | |
114 | if fh is not None: |
|
114 | if fh is not None: | |
115 | fh.close() |
|
115 | fh.close() | |
116 | if cleanup is not None: |
|
116 | if cleanup is not None: | |
117 | os.unlink(cleanup) |
|
117 | os.unlink(cleanup) | |
118 |
|
118 | |||
119 | def decompressor(fh, alg): |
|
119 | def decompressor(fh, alg): | |
120 | if alg == 'UN': |
|
120 | if alg == 'UN': | |
121 | return fh |
|
121 | return fh | |
122 | elif alg == 'GZ': |
|
122 | elif alg == 'GZ': | |
123 | def generator(f): |
|
123 | def generator(f): | |
124 | zd = zlib.decompressobj() |
|
124 | zd = zlib.decompressobj() | |
125 | for chunk in f: |
|
125 | for chunk in f: | |
126 | yield zd.decompress(chunk) |
|
126 | yield zd.decompress(chunk) | |
127 | elif alg == 'BZ': |
|
127 | elif alg == 'BZ': | |
128 | def generator(f): |
|
128 | def generator(f): | |
129 | zd = bz2.BZ2Decompressor() |
|
129 | zd = bz2.BZ2Decompressor() | |
130 | zd.decompress("BZ") |
|
130 | zd.decompress("BZ") | |
131 | for chunk in util.filechunkiter(f, 4096): |
|
131 | for chunk in util.filechunkiter(f, 4096): | |
132 | yield zd.decompress(chunk) |
|
132 | yield zd.decompress(chunk) | |
133 | else: |
|
133 | else: | |
134 | raise util.Abort("unknown bundle compression '%s'" % alg) |
|
134 | raise util.Abort("unknown bundle compression '%s'" % alg) | |
135 | return util.chunkbuffer(generator(fh)) |
|
135 | return util.chunkbuffer(generator(fh)) | |
136 |
|
136 | |||
137 | class unbundle10(object): |
|
137 | class unbundle10(object): | |
138 | def __init__(self, fh, alg): |
|
138 | def __init__(self, fh, alg): | |
139 | self._stream = decompressor(fh, alg) |
|
139 | self._stream = decompressor(fh, alg) | |
140 | self._type = alg |
|
140 | self._type = alg | |
141 | self.callback = None |
|
141 | self.callback = None | |
142 | def compressed(self): |
|
142 | def compressed(self): | |
143 | return self._type != 'UN' |
|
143 | return self._type != 'UN' | |
144 | def read(self, l): |
|
144 | def read(self, l): | |
145 | return self._stream.read(l) |
|
145 | return self._stream.read(l) | |
146 | def seek(self, pos): |
|
146 | def seek(self, pos): | |
147 | return self._stream.seek(pos) |
|
147 | return self._stream.seek(pos) | |
148 | def tell(self): |
|
148 | def tell(self): | |
149 | return self._stream.tell() |
|
149 | return self._stream.tell() | |
150 | def close(self): |
|
150 | def close(self): | |
151 | return self._stream.close() |
|
151 | return self._stream.close() | |
152 |
|
152 | |||
153 | def chunklength(self): |
|
153 | def chunklength(self): | |
154 | d = readexactly(stream, 4) |
|
154 | d = readexactly(self._stream, 4) | |
155 | l = struct.unpack(">l", d)[0] |
|
155 | l = struct.unpack(">l", d)[0] | |
156 | if l <= 4: |
|
156 | if l <= 4: | |
157 | if l: |
|
157 | if l: | |
158 | raise util.Abort(_("invalid chunk length %d") % l) |
|
158 | raise util.Abort(_("invalid chunk length %d") % l) | |
159 | return 0 |
|
159 | return 0 | |
160 | if self.callback: |
|
160 | if self.callback: | |
161 | self.callback() |
|
161 | self.callback() | |
162 | return l - 4 |
|
162 | return l - 4 | |
163 |
|
163 | |||
164 | def chunk(self): |
|
164 | def chunk(self): | |
165 | """return the next chunk from changegroup 'source' as a string""" |
|
165 | """return the next chunk from changegroup 'source' as a string""" | |
166 | l = self.chunklength() |
|
166 | l = self.chunklength() | |
167 | return readexactly(self._stream, l) |
|
167 | return readexactly(self._stream, l) | |
168 |
|
168 | |||
169 | def parsechunk(self): |
|
169 | def parsechunk(self): | |
170 | l = self.chunklength() |
|
170 | l = self.chunklength() | |
171 | if not l: |
|
171 | if not l: | |
172 | return {} |
|
172 | return {} | |
173 | h = readexactly(self._stream, 80) |
|
173 | h = readexactly(self._stream, 80) | |
174 | node, p1, p2, cs = struct.unpack("20s20s20s20s", h) |
|
174 | node, p1, p2, cs = struct.unpack("20s20s20s20s", h) | |
175 | data = readexactly(self._stream, l - 80) |
|
175 | data = readexactly(self._stream, l - 80) | |
176 | return dict(node=node, p1=p1, p2=p2, cs=cs, data=data) |
|
176 | return dict(node=node, p1=p1, p2=p2, cs=cs, data=data) | |
177 |
|
177 | |||
178 | class headerlessfixup(object): |
|
178 | class headerlessfixup(object): | |
179 | def __init__(self, fh, h): |
|
179 | def __init__(self, fh, h): | |
180 | self._h = h |
|
180 | self._h = h | |
181 | self._fh = fh |
|
181 | self._fh = fh | |
182 | def read(self, n): |
|
182 | def read(self, n): | |
183 | if self._h: |
|
183 | if self._h: | |
184 | d, self._h = self._h[:n], self._h[n:] |
|
184 | d, self._h = self._h[:n], self._h[n:] | |
185 | if len(d) < n: |
|
185 | if len(d) < n: | |
186 | d += readexactly(self._fh, n - len(d)) |
|
186 | d += readexactly(self._fh, n - len(d)) | |
187 | return d |
|
187 | return d | |
188 | return readexactly(self._fh, n) |
|
188 | return readexactly(self._fh, n) | |
189 |
|
189 | |||
190 | def readbundle(fh, fname): |
|
190 | def readbundle(fh, fname): | |
191 | header = readexactly(fh, 6) |
|
191 | header = readexactly(fh, 6) | |
192 |
|
192 | |||
193 | if not fname: |
|
193 | if not fname: | |
194 | fname = "stream" |
|
194 | fname = "stream" | |
195 | if not header.startswith('HG') and header.startswith('\0'): |
|
195 | if not header.startswith('HG') and header.startswith('\0'): | |
196 | fh = headerlessfixup(fh, header) |
|
196 | fh = headerlessfixup(fh, header) | |
197 | header = "HG10UN" |
|
197 | header = "HG10UN" | |
198 |
|
198 | |||
199 | magic, version, alg = header[0:2], header[2:4], header[4:6] |
|
199 | magic, version, alg = header[0:2], header[2:4], header[4:6] | |
200 |
|
200 | |||
201 | if magic != 'HG': |
|
201 | if magic != 'HG': | |
202 | raise util.Abort(_('%s: not a Mercurial bundle') % fname) |
|
202 | raise util.Abort(_('%s: not a Mercurial bundle') % fname) | |
203 | if version != '10': |
|
203 | if version != '10': | |
204 | raise util.Abort(_('%s: unknown bundle version %s') % (fname, version)) |
|
204 | raise util.Abort(_('%s: unknown bundle version %s') % (fname, version)) | |
205 | return unbundle10(fh, alg) |
|
205 | return unbundle10(fh, alg) |
General Comments 0
You need to be logged in to leave comments.
Login now