Show More
@@ -1,430 +1,542 b'' | |||||
1 | # streamclone.py - producing and consuming streaming repository data |
|
1 | # streamclone.py - producing and consuming streaming repository data | |
2 | # |
|
2 | # | |
3 | # Copyright 2015 Gregory Szorc <gregory.szorc@gmail.com> |
|
3 | # Copyright 2015 Gregory Szorc <gregory.szorc@gmail.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 __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import struct |
|
10 | import struct | |
11 |
|
11 | |||
12 | from .i18n import _ |
|
12 | from .i18n import _ | |
13 | from . import ( |
|
13 | from . import ( | |
14 | branchmap, |
|
14 | branchmap, | |
15 | error, |
|
15 | error, | |
16 | phases, |
|
16 | phases, | |
17 | store, |
|
17 | store, | |
18 | util, |
|
18 | util, | |
19 | ) |
|
19 | ) | |
20 |
|
20 | |||
21 | def canperformstreamclone(pullop, bailifbundle2supported=False): |
|
21 | def canperformstreamclone(pullop, bailifbundle2supported=False): | |
22 | """Whether it is possible to perform a streaming clone as part of pull. |
|
22 | """Whether it is possible to perform a streaming clone as part of pull. | |
23 |
|
23 | |||
24 | ``bailifbundle2supported`` will cause the function to return False if |
|
24 | ``bailifbundle2supported`` will cause the function to return False if | |
25 | bundle2 stream clones are supported. It should only be called by the |
|
25 | bundle2 stream clones are supported. It should only be called by the | |
26 | legacy stream clone code path. |
|
26 | legacy stream clone code path. | |
27 |
|
27 | |||
28 | Returns a tuple of (supported, requirements). ``supported`` is True if |
|
28 | Returns a tuple of (supported, requirements). ``supported`` is True if | |
29 | streaming clone is supported and False otherwise. ``requirements`` is |
|
29 | streaming clone is supported and False otherwise. ``requirements`` is | |
30 | a set of repo requirements from the remote, or ``None`` if stream clone |
|
30 | a set of repo requirements from the remote, or ``None`` if stream clone | |
31 | isn't supported. |
|
31 | isn't supported. | |
32 | """ |
|
32 | """ | |
33 | repo = pullop.repo |
|
33 | repo = pullop.repo | |
34 | remote = pullop.remote |
|
34 | remote = pullop.remote | |
35 |
|
35 | |||
36 | bundle2supported = False |
|
36 | bundle2supported = False | |
37 | if pullop.canusebundle2: |
|
37 | if pullop.canusebundle2: | |
38 | if 'v1' in pullop.remotebundle2caps.get('stream', []): |
|
38 | if 'v1' in pullop.remotebundle2caps.get('stream', []): | |
39 | bundle2supported = True |
|
39 | bundle2supported = True | |
40 | # else |
|
40 | # else | |
41 | # Server doesn't support bundle2 stream clone or doesn't support |
|
41 | # Server doesn't support bundle2 stream clone or doesn't support | |
42 | # the versions we support. Fall back and possibly allow legacy. |
|
42 | # the versions we support. Fall back and possibly allow legacy. | |
43 |
|
43 | |||
44 | # Ensures legacy code path uses available bundle2. |
|
44 | # Ensures legacy code path uses available bundle2. | |
45 | if bailifbundle2supported and bundle2supported: |
|
45 | if bailifbundle2supported and bundle2supported: | |
46 | return False, None |
|
46 | return False, None | |
47 | # Ensures bundle2 doesn't try to do a stream clone if it isn't supported. |
|
47 | # Ensures bundle2 doesn't try to do a stream clone if it isn't supported. | |
48 | #elif not bailifbundle2supported and not bundle2supported: |
|
48 | #elif not bailifbundle2supported and not bundle2supported: | |
49 | # return False, None |
|
49 | # return False, None | |
50 |
|
50 | |||
51 | # Streaming clone only works on empty repositories. |
|
51 | # Streaming clone only works on empty repositories. | |
52 | if len(repo): |
|
52 | if len(repo): | |
53 | return False, None |
|
53 | return False, None | |
54 |
|
54 | |||
55 | # Streaming clone only works if all data is being requested. |
|
55 | # Streaming clone only works if all data is being requested. | |
56 | if pullop.heads: |
|
56 | if pullop.heads: | |
57 | return False, None |
|
57 | return False, None | |
58 |
|
58 | |||
59 | streamrequested = pullop.streamclonerequested |
|
59 | streamrequested = pullop.streamclonerequested | |
60 |
|
60 | |||
61 | # If we don't have a preference, let the server decide for us. This |
|
61 | # If we don't have a preference, let the server decide for us. This | |
62 | # likely only comes into play in LANs. |
|
62 | # likely only comes into play in LANs. | |
63 | if streamrequested is None: |
|
63 | if streamrequested is None: | |
64 | # The server can advertise whether to prefer streaming clone. |
|
64 | # The server can advertise whether to prefer streaming clone. | |
65 | streamrequested = remote.capable('stream-preferred') |
|
65 | streamrequested = remote.capable('stream-preferred') | |
66 |
|
66 | |||
67 | if not streamrequested: |
|
67 | if not streamrequested: | |
68 | return False, None |
|
68 | return False, None | |
69 |
|
69 | |||
70 | # In order for stream clone to work, the client has to support all the |
|
70 | # In order for stream clone to work, the client has to support all the | |
71 | # requirements advertised by the server. |
|
71 | # requirements advertised by the server. | |
72 | # |
|
72 | # | |
73 | # The server advertises its requirements via the "stream" and "streamreqs" |
|
73 | # The server advertises its requirements via the "stream" and "streamreqs" | |
74 | # capability. "stream" (a value-less capability) is advertised if and only |
|
74 | # capability. "stream" (a value-less capability) is advertised if and only | |
75 | # if the only requirement is "revlogv1." Else, the "streamreqs" capability |
|
75 | # if the only requirement is "revlogv1." Else, the "streamreqs" capability | |
76 | # is advertised and contains a comma-delimited list of requirements. |
|
76 | # is advertised and contains a comma-delimited list of requirements. | |
77 | requirements = set() |
|
77 | requirements = set() | |
78 | if remote.capable('stream'): |
|
78 | if remote.capable('stream'): | |
79 | requirements.add('revlogv1') |
|
79 | requirements.add('revlogv1') | |
80 | else: |
|
80 | else: | |
81 | streamreqs = remote.capable('streamreqs') |
|
81 | streamreqs = remote.capable('streamreqs') | |
82 | # This is weird and shouldn't happen with modern servers. |
|
82 | # This is weird and shouldn't happen with modern servers. | |
83 | if not streamreqs: |
|
83 | if not streamreqs: | |
84 | pullop.repo.ui.warn(_( |
|
84 | pullop.repo.ui.warn(_( | |
85 | 'warning: stream clone requested but server has them ' |
|
85 | 'warning: stream clone requested but server has them ' | |
86 | 'disabled\n')) |
|
86 | 'disabled\n')) | |
87 | return False, None |
|
87 | return False, None | |
88 |
|
88 | |||
89 | streamreqs = set(streamreqs.split(',')) |
|
89 | streamreqs = set(streamreqs.split(',')) | |
90 | # Server requires something we don't support. Bail. |
|
90 | # Server requires something we don't support. Bail. | |
91 | missingreqs = streamreqs - repo.supportedformats |
|
91 | missingreqs = streamreqs - repo.supportedformats | |
92 | if missingreqs: |
|
92 | if missingreqs: | |
93 | pullop.repo.ui.warn(_( |
|
93 | pullop.repo.ui.warn(_( | |
94 | 'warning: stream clone requested but client is missing ' |
|
94 | 'warning: stream clone requested but client is missing ' | |
95 | 'requirements: %s\n') % ', '.join(sorted(missingreqs))) |
|
95 | 'requirements: %s\n') % ', '.join(sorted(missingreqs))) | |
96 | pullop.repo.ui.warn( |
|
96 | pullop.repo.ui.warn( | |
97 | _('(see https://www.mercurial-scm.org/wiki/MissingRequirement ' |
|
97 | _('(see https://www.mercurial-scm.org/wiki/MissingRequirement ' | |
98 | 'for more information)\n')) |
|
98 | 'for more information)\n')) | |
99 | return False, None |
|
99 | return False, None | |
100 | requirements = streamreqs |
|
100 | requirements = streamreqs | |
101 |
|
101 | |||
102 | return True, requirements |
|
102 | return True, requirements | |
103 |
|
103 | |||
104 | def maybeperformlegacystreamclone(pullop): |
|
104 | def maybeperformlegacystreamclone(pullop): | |
105 | """Possibly perform a legacy stream clone operation. |
|
105 | """Possibly perform a legacy stream clone operation. | |
106 |
|
106 | |||
107 | Legacy stream clones are performed as part of pull but before all other |
|
107 | Legacy stream clones are performed as part of pull but before all other | |
108 | operations. |
|
108 | operations. | |
109 |
|
109 | |||
110 | A legacy stream clone will not be performed if a bundle2 stream clone is |
|
110 | A legacy stream clone will not be performed if a bundle2 stream clone is | |
111 | supported. |
|
111 | supported. | |
112 | """ |
|
112 | """ | |
113 | supported, requirements = canperformstreamclone(pullop) |
|
113 | supported, requirements = canperformstreamclone(pullop) | |
114 |
|
114 | |||
115 | if not supported: |
|
115 | if not supported: | |
116 | return |
|
116 | return | |
117 |
|
117 | |||
118 | repo = pullop.repo |
|
118 | repo = pullop.repo | |
119 | remote = pullop.remote |
|
119 | remote = pullop.remote | |
120 |
|
120 | |||
121 | # Save remote branchmap. We will use it later to speed up branchcache |
|
121 | # Save remote branchmap. We will use it later to speed up branchcache | |
122 | # creation. |
|
122 | # creation. | |
123 | rbranchmap = None |
|
123 | rbranchmap = None | |
124 | if remote.capable('branchmap'): |
|
124 | if remote.capable('branchmap'): | |
125 | rbranchmap = remote.branchmap() |
|
125 | rbranchmap = remote.branchmap() | |
126 |
|
126 | |||
127 | repo.ui.status(_('streaming all changes\n')) |
|
127 | repo.ui.status(_('streaming all changes\n')) | |
128 |
|
128 | |||
129 | fp = remote.stream_out() |
|
129 | fp = remote.stream_out() | |
130 | l = fp.readline() |
|
130 | l = fp.readline() | |
131 | try: |
|
131 | try: | |
132 | resp = int(l) |
|
132 | resp = int(l) | |
133 | except ValueError: |
|
133 | except ValueError: | |
134 | raise error.ResponseError( |
|
134 | raise error.ResponseError( | |
135 | _('unexpected response from remote server:'), l) |
|
135 | _('unexpected response from remote server:'), l) | |
136 | if resp == 1: |
|
136 | if resp == 1: | |
137 | raise error.Abort(_('operation forbidden by server')) |
|
137 | raise error.Abort(_('operation forbidden by server')) | |
138 | elif resp == 2: |
|
138 | elif resp == 2: | |
139 | raise error.Abort(_('locking the remote repository failed')) |
|
139 | raise error.Abort(_('locking the remote repository failed')) | |
140 | elif resp != 0: |
|
140 | elif resp != 0: | |
141 | raise error.Abort(_('the server sent an unknown error code')) |
|
141 | raise error.Abort(_('the server sent an unknown error code')) | |
142 |
|
142 | |||
143 | l = fp.readline() |
|
143 | l = fp.readline() | |
144 | try: |
|
144 | try: | |
145 | filecount, bytecount = map(int, l.split(' ', 1)) |
|
145 | filecount, bytecount = map(int, l.split(' ', 1)) | |
146 | except (ValueError, TypeError): |
|
146 | except (ValueError, TypeError): | |
147 | raise error.ResponseError( |
|
147 | raise error.ResponseError( | |
148 | _('unexpected response from remote server:'), l) |
|
148 | _('unexpected response from remote server:'), l) | |
149 |
|
149 | |||
150 | with repo.lock(): |
|
150 | with repo.lock(): | |
151 | consumev1(repo, fp, filecount, bytecount) |
|
151 | consumev1(repo, fp, filecount, bytecount) | |
152 |
|
152 | |||
153 | # new requirements = old non-format requirements + |
|
153 | # new requirements = old non-format requirements + | |
154 | # new format-related remote requirements |
|
154 | # new format-related remote requirements | |
155 | # requirements from the streamed-in repository |
|
155 | # requirements from the streamed-in repository | |
156 | repo.requirements = requirements | ( |
|
156 | repo.requirements = requirements | ( | |
157 | repo.requirements - repo.supportedformats) |
|
157 | repo.requirements - repo.supportedformats) | |
158 | repo._applyopenerreqs() |
|
158 | repo._applyopenerreqs() | |
159 | repo._writerequirements() |
|
159 | repo._writerequirements() | |
160 |
|
160 | |||
161 | if rbranchmap: |
|
161 | if rbranchmap: | |
162 | branchmap.replacecache(repo, rbranchmap) |
|
162 | branchmap.replacecache(repo, rbranchmap) | |
163 |
|
163 | |||
164 | repo.invalidate() |
|
164 | repo.invalidate() | |
165 |
|
165 | |||
166 | def allowservergeneration(repo): |
|
166 | def allowservergeneration(repo): | |
167 | """Whether streaming clones are allowed from the server.""" |
|
167 | """Whether streaming clones are allowed from the server.""" | |
168 | if not repo.ui.configbool('server', 'uncompressed', untrusted=True): |
|
168 | if not repo.ui.configbool('server', 'uncompressed', untrusted=True): | |
169 | return False |
|
169 | return False | |
170 |
|
170 | |||
171 | # The way stream clone works makes it impossible to hide secret changesets. |
|
171 | # The way stream clone works makes it impossible to hide secret changesets. | |
172 | # So don't allow this by default. |
|
172 | # So don't allow this by default. | |
173 | secret = phases.hassecret(repo) |
|
173 | secret = phases.hassecret(repo) | |
174 | if secret: |
|
174 | if secret: | |
175 | return repo.ui.configbool('server', 'uncompressedallowsecret') |
|
175 | return repo.ui.configbool('server', 'uncompressedallowsecret') | |
176 |
|
176 | |||
177 | return True |
|
177 | return True | |
178 |
|
178 | |||
179 | # This is it's own function so extensions can override it. |
|
179 | # This is it's own function so extensions can override it. | |
180 | def _walkstreamfiles(repo): |
|
180 | def _walkstreamfiles(repo): | |
181 | return repo.store.walk() |
|
181 | return repo.store.walk() | |
182 |
|
182 | |||
183 | def generatev1(repo): |
|
183 | def generatev1(repo): | |
184 | """Emit content for version 1 of a streaming clone. |
|
184 | """Emit content for version 1 of a streaming clone. | |
185 |
|
185 | |||
186 | This returns a 3-tuple of (file count, byte size, data iterator). |
|
186 | This returns a 3-tuple of (file count, byte size, data iterator). | |
187 |
|
187 | |||
188 | The data iterator consists of N entries for each file being transferred. |
|
188 | The data iterator consists of N entries for each file being transferred. | |
189 | Each file entry starts as a line with the file name and integer size |
|
189 | Each file entry starts as a line with the file name and integer size | |
190 | delimited by a null byte. |
|
190 | delimited by a null byte. | |
191 |
|
191 | |||
192 | The raw file data follows. Following the raw file data is the next file |
|
192 | The raw file data follows. Following the raw file data is the next file | |
193 | entry, or EOF. |
|
193 | entry, or EOF. | |
194 |
|
194 | |||
195 | When used on the wire protocol, an additional line indicating protocol |
|
195 | When used on the wire protocol, an additional line indicating protocol | |
196 | success will be prepended to the stream. This function is not responsible |
|
196 | success will be prepended to the stream. This function is not responsible | |
197 | for adding it. |
|
197 | for adding it. | |
198 |
|
198 | |||
199 | This function will obtain a repository lock to ensure a consistent view of |
|
199 | This function will obtain a repository lock to ensure a consistent view of | |
200 | the store is captured. It therefore may raise LockError. |
|
200 | the store is captured. It therefore may raise LockError. | |
201 | """ |
|
201 | """ | |
202 | entries = [] |
|
202 | entries = [] | |
203 | total_bytes = 0 |
|
203 | total_bytes = 0 | |
204 | # Get consistent snapshot of repo, lock during scan. |
|
204 | # Get consistent snapshot of repo, lock during scan. | |
205 | with repo.lock(): |
|
205 | with repo.lock(): | |
206 | repo.ui.debug('scanning\n') |
|
206 | repo.ui.debug('scanning\n') | |
207 | for name, ename, size in _walkstreamfiles(repo): |
|
207 | for name, ename, size in _walkstreamfiles(repo): | |
208 | if size: |
|
208 | if size: | |
209 | entries.append((name, size)) |
|
209 | entries.append((name, size)) | |
210 | total_bytes += size |
|
210 | total_bytes += size | |
211 |
|
211 | |||
212 | repo.ui.debug('%d files, %d bytes to transfer\n' % |
|
212 | repo.ui.debug('%d files, %d bytes to transfer\n' % | |
213 | (len(entries), total_bytes)) |
|
213 | (len(entries), total_bytes)) | |
214 |
|
214 | |||
215 | svfs = repo.svfs |
|
215 | svfs = repo.svfs | |
216 | debugflag = repo.ui.debugflag |
|
216 | debugflag = repo.ui.debugflag | |
217 |
|
217 | |||
218 | def emitrevlogdata(): |
|
218 | def emitrevlogdata(): | |
219 | for name, size in entries: |
|
219 | for name, size in entries: | |
220 | if debugflag: |
|
220 | if debugflag: | |
221 | repo.ui.debug('sending %s (%d bytes)\n' % (name, size)) |
|
221 | repo.ui.debug('sending %s (%d bytes)\n' % (name, size)) | |
222 | # partially encode name over the wire for backwards compat |
|
222 | # partially encode name over the wire for backwards compat | |
223 | yield '%s\0%d\n' % (store.encodedir(name), size) |
|
223 | yield '%s\0%d\n' % (store.encodedir(name), size) | |
224 | # auditing at this stage is both pointless (paths are already |
|
224 | # auditing at this stage is both pointless (paths are already | |
225 | # trusted by the local repo) and expensive |
|
225 | # trusted by the local repo) and expensive | |
226 | with svfs(name, 'rb', auditpath=False) as fp: |
|
226 | with svfs(name, 'rb', auditpath=False) as fp: | |
227 | if size <= 65536: |
|
227 | if size <= 65536: | |
228 | yield fp.read(size) |
|
228 | yield fp.read(size) | |
229 | else: |
|
229 | else: | |
230 | for chunk in util.filechunkiter(fp, limit=size): |
|
230 | for chunk in util.filechunkiter(fp, limit=size): | |
231 | yield chunk |
|
231 | yield chunk | |
232 |
|
232 | |||
233 | return len(entries), total_bytes, emitrevlogdata() |
|
233 | return len(entries), total_bytes, emitrevlogdata() | |
234 |
|
234 | |||
235 | def generatev1wireproto(repo): |
|
235 | def generatev1wireproto(repo): | |
236 | """Emit content for version 1 of streaming clone suitable for the wire. |
|
236 | """Emit content for version 1 of streaming clone suitable for the wire. | |
237 |
|
237 | |||
238 | This is the data output from ``generatev1()`` with 2 header lines. The |
|
238 | This is the data output from ``generatev1()`` with 2 header lines. The | |
239 | first line indicates overall success. The 2nd contains the file count and |
|
239 | first line indicates overall success. The 2nd contains the file count and | |
240 | byte size of payload. |
|
240 | byte size of payload. | |
241 |
|
241 | |||
242 | The success line contains "0" for success, "1" for stream generation not |
|
242 | The success line contains "0" for success, "1" for stream generation not | |
243 | allowed, and "2" for error locking the repository (possibly indicating |
|
243 | allowed, and "2" for error locking the repository (possibly indicating | |
244 | a permissions error for the server process). |
|
244 | a permissions error for the server process). | |
245 | """ |
|
245 | """ | |
246 | if not allowservergeneration(repo): |
|
246 | if not allowservergeneration(repo): | |
247 | yield '1\n' |
|
247 | yield '1\n' | |
248 | return |
|
248 | return | |
249 |
|
249 | |||
250 | try: |
|
250 | try: | |
251 | filecount, bytecount, it = generatev1(repo) |
|
251 | filecount, bytecount, it = generatev1(repo) | |
252 | except error.LockError: |
|
252 | except error.LockError: | |
253 | yield '2\n' |
|
253 | yield '2\n' | |
254 | return |
|
254 | return | |
255 |
|
255 | |||
256 | # Indicates successful response. |
|
256 | # Indicates successful response. | |
257 | yield '0\n' |
|
257 | yield '0\n' | |
258 | yield '%d %d\n' % (filecount, bytecount) |
|
258 | yield '%d %d\n' % (filecount, bytecount) | |
259 | for chunk in it: |
|
259 | for chunk in it: | |
260 | yield chunk |
|
260 | yield chunk | |
261 |
|
261 | |||
262 | def generatebundlev1(repo, compression='UN'): |
|
262 | def generatebundlev1(repo, compression='UN'): | |
263 | """Emit content for version 1 of a stream clone bundle. |
|
263 | """Emit content for version 1 of a stream clone bundle. | |
264 |
|
264 | |||
265 | The first 4 bytes of the output ("HGS1") denote this as stream clone |
|
265 | The first 4 bytes of the output ("HGS1") denote this as stream clone | |
266 | bundle version 1. |
|
266 | bundle version 1. | |
267 |
|
267 | |||
268 | The next 2 bytes indicate the compression type. Only "UN" is currently |
|
268 | The next 2 bytes indicate the compression type. Only "UN" is currently | |
269 | supported. |
|
269 | supported. | |
270 |
|
270 | |||
271 | The next 16 bytes are two 64-bit big endian unsigned integers indicating |
|
271 | The next 16 bytes are two 64-bit big endian unsigned integers indicating | |
272 | file count and byte count, respectively. |
|
272 | file count and byte count, respectively. | |
273 |
|
273 | |||
274 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length |
|
274 | The next 2 bytes is a 16-bit big endian unsigned short declaring the length | |
275 | of the requirements string, including a trailing \0. The following N bytes |
|
275 | of the requirements string, including a trailing \0. The following N bytes | |
276 | are the requirements string, which is ASCII containing a comma-delimited |
|
276 | are the requirements string, which is ASCII containing a comma-delimited | |
277 | list of repo requirements that are needed to support the data. |
|
277 | list of repo requirements that are needed to support the data. | |
278 |
|
278 | |||
279 | The remaining content is the output of ``generatev1()`` (which may be |
|
279 | The remaining content is the output of ``generatev1()`` (which may be | |
280 | compressed in the future). |
|
280 | compressed in the future). | |
281 |
|
281 | |||
282 | Returns a tuple of (requirements, data generator). |
|
282 | Returns a tuple of (requirements, data generator). | |
283 | """ |
|
283 | """ | |
284 | if compression != 'UN': |
|
284 | if compression != 'UN': | |
285 | raise ValueError('we do not support the compression argument yet') |
|
285 | raise ValueError('we do not support the compression argument yet') | |
286 |
|
286 | |||
287 | requirements = repo.requirements & repo.supportedformats |
|
287 | requirements = repo.requirements & repo.supportedformats | |
288 | requires = ','.join(sorted(requirements)) |
|
288 | requires = ','.join(sorted(requirements)) | |
289 |
|
289 | |||
290 | def gen(): |
|
290 | def gen(): | |
291 | yield 'HGS1' |
|
291 | yield 'HGS1' | |
292 | yield compression |
|
292 | yield compression | |
293 |
|
293 | |||
294 | filecount, bytecount, it = generatev1(repo) |
|
294 | filecount, bytecount, it = generatev1(repo) | |
295 | repo.ui.status(_('writing %d bytes for %d files\n') % |
|
295 | repo.ui.status(_('writing %d bytes for %d files\n') % | |
296 | (bytecount, filecount)) |
|
296 | (bytecount, filecount)) | |
297 |
|
297 | |||
298 | yield struct.pack('>QQ', filecount, bytecount) |
|
298 | yield struct.pack('>QQ', filecount, bytecount) | |
299 | yield struct.pack('>H', len(requires) + 1) |
|
299 | yield struct.pack('>H', len(requires) + 1) | |
300 | yield requires + '\0' |
|
300 | yield requires + '\0' | |
301 |
|
301 | |||
302 | # This is where we'll add compression in the future. |
|
302 | # This is where we'll add compression in the future. | |
303 | assert compression == 'UN' |
|
303 | assert compression == 'UN' | |
304 |
|
304 | |||
305 | seen = 0 |
|
305 | seen = 0 | |
306 | repo.ui.progress(_('bundle'), 0, total=bytecount, unit=_('bytes')) |
|
306 | repo.ui.progress(_('bundle'), 0, total=bytecount, unit=_('bytes')) | |
307 |
|
307 | |||
308 | for chunk in it: |
|
308 | for chunk in it: | |
309 | seen += len(chunk) |
|
309 | seen += len(chunk) | |
310 | repo.ui.progress(_('bundle'), seen, total=bytecount, |
|
310 | repo.ui.progress(_('bundle'), seen, total=bytecount, | |
311 | unit=_('bytes')) |
|
311 | unit=_('bytes')) | |
312 | yield chunk |
|
312 | yield chunk | |
313 |
|
313 | |||
314 | repo.ui.progress(_('bundle'), None) |
|
314 | repo.ui.progress(_('bundle'), None) | |
315 |
|
315 | |||
316 | return requirements, gen() |
|
316 | return requirements, gen() | |
317 |
|
317 | |||
318 | def consumev1(repo, fp, filecount, bytecount): |
|
318 | def consumev1(repo, fp, filecount, bytecount): | |
319 | """Apply the contents from version 1 of a streaming clone file handle. |
|
319 | """Apply the contents from version 1 of a streaming clone file handle. | |
320 |
|
320 | |||
321 | This takes the output from "stream_out" and applies it to the specified |
|
321 | This takes the output from "stream_out" and applies it to the specified | |
322 | repository. |
|
322 | repository. | |
323 |
|
323 | |||
324 | Like "stream_out," the status line added by the wire protocol is not |
|
324 | Like "stream_out," the status line added by the wire protocol is not | |
325 | handled by this function. |
|
325 | handled by this function. | |
326 | """ |
|
326 | """ | |
327 | with repo.lock(): |
|
327 | with repo.lock(): | |
328 | repo.ui.status(_('%d files to transfer, %s of data\n') % |
|
328 | repo.ui.status(_('%d files to transfer, %s of data\n') % | |
329 | (filecount, util.bytecount(bytecount))) |
|
329 | (filecount, util.bytecount(bytecount))) | |
330 | handled_bytes = 0 |
|
330 | handled_bytes = 0 | |
331 | repo.ui.progress(_('clone'), 0, total=bytecount, unit=_('bytes')) |
|
331 | repo.ui.progress(_('clone'), 0, total=bytecount, unit=_('bytes')) | |
332 | start = util.timer() |
|
332 | start = util.timer() | |
333 |
|
333 | |||
334 | # TODO: get rid of (potential) inconsistency |
|
334 | # TODO: get rid of (potential) inconsistency | |
335 | # |
|
335 | # | |
336 | # If transaction is started and any @filecache property is |
|
336 | # If transaction is started and any @filecache property is | |
337 | # changed at this point, it causes inconsistency between |
|
337 | # changed at this point, it causes inconsistency between | |
338 | # in-memory cached property and streamclone-ed file on the |
|
338 | # in-memory cached property and streamclone-ed file on the | |
339 | # disk. Nested transaction prevents transaction scope "clone" |
|
339 | # disk. Nested transaction prevents transaction scope "clone" | |
340 | # below from writing in-memory changes out at the end of it, |
|
340 | # below from writing in-memory changes out at the end of it, | |
341 | # even though in-memory changes are discarded at the end of it |
|
341 | # even though in-memory changes are discarded at the end of it | |
342 | # regardless of transaction nesting. |
|
342 | # regardless of transaction nesting. | |
343 | # |
|
343 | # | |
344 | # But transaction nesting can't be simply prohibited, because |
|
344 | # But transaction nesting can't be simply prohibited, because | |
345 | # nesting occurs also in ordinary case (e.g. enabling |
|
345 | # nesting occurs also in ordinary case (e.g. enabling | |
346 | # clonebundles). |
|
346 | # clonebundles). | |
347 |
|
347 | |||
348 | with repo.transaction('clone'): |
|
348 | with repo.transaction('clone'): | |
349 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): |
|
349 | with repo.svfs.backgroundclosing(repo.ui, expectedcount=filecount): | |
350 | for i in xrange(filecount): |
|
350 | for i in xrange(filecount): | |
351 | # XXX doesn't support '\n' or '\r' in filenames |
|
351 | # XXX doesn't support '\n' or '\r' in filenames | |
352 | l = fp.readline() |
|
352 | l = fp.readline() | |
353 | try: |
|
353 | try: | |
354 | name, size = l.split('\0', 1) |
|
354 | name, size = l.split('\0', 1) | |
355 | size = int(size) |
|
355 | size = int(size) | |
356 | except (ValueError, TypeError): |
|
356 | except (ValueError, TypeError): | |
357 | raise error.ResponseError( |
|
357 | raise error.ResponseError( | |
358 | _('unexpected response from remote server:'), l) |
|
358 | _('unexpected response from remote server:'), l) | |
359 | if repo.ui.debugflag: |
|
359 | if repo.ui.debugflag: | |
360 | repo.ui.debug('adding %s (%s)\n' % |
|
360 | repo.ui.debug('adding %s (%s)\n' % | |
361 | (name, util.bytecount(size))) |
|
361 | (name, util.bytecount(size))) | |
362 | # for backwards compat, name was partially encoded |
|
362 | # for backwards compat, name was partially encoded | |
363 | path = store.decodedir(name) |
|
363 | path = store.decodedir(name) | |
364 | with repo.svfs(path, 'w', backgroundclose=True) as ofp: |
|
364 | with repo.svfs(path, 'w', backgroundclose=True) as ofp: | |
365 | for chunk in util.filechunkiter(fp, limit=size): |
|
365 | for chunk in util.filechunkiter(fp, limit=size): | |
366 | handled_bytes += len(chunk) |
|
366 | handled_bytes += len(chunk) | |
367 | repo.ui.progress(_('clone'), handled_bytes, |
|
367 | repo.ui.progress(_('clone'), handled_bytes, | |
368 | total=bytecount, unit=_('bytes')) |
|
368 | total=bytecount, unit=_('bytes')) | |
369 | ofp.write(chunk) |
|
369 | ofp.write(chunk) | |
370 |
|
370 | |||
371 | # force @filecache properties to be reloaded from |
|
371 | # force @filecache properties to be reloaded from | |
372 | # streamclone-ed file at next access |
|
372 | # streamclone-ed file at next access | |
373 | repo.invalidate(clearfilecache=True) |
|
373 | repo.invalidate(clearfilecache=True) | |
374 |
|
374 | |||
375 | elapsed = util.timer() - start |
|
375 | elapsed = util.timer() - start | |
376 | if elapsed <= 0: |
|
376 | if elapsed <= 0: | |
377 | elapsed = 0.001 |
|
377 | elapsed = 0.001 | |
378 | repo.ui.progress(_('clone'), None) |
|
378 | repo.ui.progress(_('clone'), None) | |
379 | repo.ui.status(_('transferred %s in %.1f seconds (%s/sec)\n') % |
|
379 | repo.ui.status(_('transferred %s in %.1f seconds (%s/sec)\n') % | |
380 | (util.bytecount(bytecount), elapsed, |
|
380 | (util.bytecount(bytecount), elapsed, | |
381 | util.bytecount(bytecount / elapsed))) |
|
381 | util.bytecount(bytecount / elapsed))) | |
382 |
|
382 | |||
383 | def readbundle1header(fp): |
|
383 | def readbundle1header(fp): | |
384 | compression = fp.read(2) |
|
384 | compression = fp.read(2) | |
385 | if compression != 'UN': |
|
385 | if compression != 'UN': | |
386 | raise error.Abort(_('only uncompressed stream clone bundles are ' |
|
386 | raise error.Abort(_('only uncompressed stream clone bundles are ' | |
387 | 'supported; got %s') % compression) |
|
387 | 'supported; got %s') % compression) | |
388 |
|
388 | |||
389 | filecount, bytecount = struct.unpack('>QQ', fp.read(16)) |
|
389 | filecount, bytecount = struct.unpack('>QQ', fp.read(16)) | |
390 | requireslen = struct.unpack('>H', fp.read(2))[0] |
|
390 | requireslen = struct.unpack('>H', fp.read(2))[0] | |
391 | requires = fp.read(requireslen) |
|
391 | requires = fp.read(requireslen) | |
392 |
|
392 | |||
393 | if not requires.endswith('\0'): |
|
393 | if not requires.endswith('\0'): | |
394 | raise error.Abort(_('malformed stream clone bundle: ' |
|
394 | raise error.Abort(_('malformed stream clone bundle: ' | |
395 | 'requirements not properly encoded')) |
|
395 | 'requirements not properly encoded')) | |
396 |
|
396 | |||
397 | requirements = set(requires.rstrip('\0').split(',')) |
|
397 | requirements = set(requires.rstrip('\0').split(',')) | |
398 |
|
398 | |||
399 | return filecount, bytecount, requirements |
|
399 | return filecount, bytecount, requirements | |
400 |
|
400 | |||
401 | def applybundlev1(repo, fp): |
|
401 | def applybundlev1(repo, fp): | |
402 | """Apply the content from a stream clone bundle version 1. |
|
402 | """Apply the content from a stream clone bundle version 1. | |
403 |
|
403 | |||
404 | We assume the 4 byte header has been read and validated and the file handle |
|
404 | We assume the 4 byte header has been read and validated and the file handle | |
405 | is at the 2 byte compression identifier. |
|
405 | is at the 2 byte compression identifier. | |
406 | """ |
|
406 | """ | |
407 | if len(repo): |
|
407 | if len(repo): | |
408 | raise error.Abort(_('cannot apply stream clone bundle on non-empty ' |
|
408 | raise error.Abort(_('cannot apply stream clone bundle on non-empty ' | |
409 | 'repo')) |
|
409 | 'repo')) | |
410 |
|
410 | |||
411 | filecount, bytecount, requirements = readbundle1header(fp) |
|
411 | filecount, bytecount, requirements = readbundle1header(fp) | |
412 | missingreqs = requirements - repo.supportedformats |
|
412 | missingreqs = requirements - repo.supportedformats | |
413 | if missingreqs: |
|
413 | if missingreqs: | |
414 | raise error.Abort(_('unable to apply stream clone: ' |
|
414 | raise error.Abort(_('unable to apply stream clone: ' | |
415 | 'unsupported format: %s') % |
|
415 | 'unsupported format: %s') % | |
416 | ', '.join(sorted(missingreqs))) |
|
416 | ', '.join(sorted(missingreqs))) | |
417 |
|
417 | |||
418 | consumev1(repo, fp, filecount, bytecount) |
|
418 | consumev1(repo, fp, filecount, bytecount) | |
419 |
|
419 | |||
420 | class streamcloneapplier(object): |
|
420 | class streamcloneapplier(object): | |
421 | """Class to manage applying streaming clone bundles. |
|
421 | """Class to manage applying streaming clone bundles. | |
422 |
|
422 | |||
423 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle |
|
423 | We need to wrap ``applybundlev1()`` in a dedicated type to enable bundle | |
424 | readers to perform bundle type-specific functionality. |
|
424 | readers to perform bundle type-specific functionality. | |
425 | """ |
|
425 | """ | |
426 | def __init__(self, fh): |
|
426 | def __init__(self, fh): | |
427 | self._fh = fh |
|
427 | self._fh = fh | |
428 |
|
428 | |||
429 | def apply(self, repo): |
|
429 | def apply(self, repo): | |
430 | return applybundlev1(repo, self._fh) |
|
430 | return applybundlev1(repo, self._fh) | |
|
431 | ||||
|
432 | def _emit(repo, entries, totalfilesize): | |||
|
433 | """actually emit the stream bundle""" | |||
|
434 | progress = repo.ui.progress | |||
|
435 | progress(_('bundle'), 0, total=totalfilesize, unit=_('bytes')) | |||
|
436 | vfs = repo.svfs | |||
|
437 | try: | |||
|
438 | seen = 0 | |||
|
439 | for name, size in entries: | |||
|
440 | yield util.uvarintencode(len(name)) | |||
|
441 | fp = vfs(name) | |||
|
442 | try: | |||
|
443 | yield util.uvarintencode(size) | |||
|
444 | yield name | |||
|
445 | if size <= 65536: | |||
|
446 | chunks = (fp.read(size),) | |||
|
447 | else: | |||
|
448 | chunks = util.filechunkiter(fp, limit=size) | |||
|
449 | for chunk in chunks: | |||
|
450 | seen += len(chunk) | |||
|
451 | progress(_('bundle'), seen, total=totalfilesize, | |||
|
452 | unit=_('bytes')) | |||
|
453 | yield chunk | |||
|
454 | finally: | |||
|
455 | fp.close() | |||
|
456 | finally: | |||
|
457 | progress(_('bundle'), None) | |||
|
458 | ||||
|
459 | def generatev2(repo): | |||
|
460 | """Emit content for version 2 of a streaming clone. | |||
|
461 | ||||
|
462 | the data stream consists the following entries: | |||
|
463 | 1) A varint containing the length of the filename | |||
|
464 | 2) A varint containing the length of file data | |||
|
465 | 3) N bytes containing the filename (the internal, store-agnostic form) | |||
|
466 | 4) N bytes containing the file data | |||
|
467 | ||||
|
468 | Returns a 3-tuple of (file count, file size, data iterator). | |||
|
469 | """ | |||
|
470 | ||||
|
471 | with repo.lock(): | |||
|
472 | ||||
|
473 | entries = [] | |||
|
474 | totalfilesize = 0 | |||
|
475 | ||||
|
476 | repo.ui.debug('scanning\n') | |||
|
477 | for name, ename, size in _walkstreamfiles(repo): | |||
|
478 | if size: | |||
|
479 | entries.append((name, size)) | |||
|
480 | totalfilesize += size | |||
|
481 | ||||
|
482 | chunks = _emit(repo, entries, totalfilesize) | |||
|
483 | ||||
|
484 | return len(entries), totalfilesize, chunks | |||
|
485 | ||||
|
486 | def consumev2(repo, fp, filecount, filesize): | |||
|
487 | """Apply the contents from a version 2 streaming clone. | |||
|
488 | ||||
|
489 | Data is read from an object that only needs to provide a ``read(size)`` | |||
|
490 | method. | |||
|
491 | """ | |||
|
492 | with repo.lock(): | |||
|
493 | repo.ui.status(_('%d files to transfer, %s of data\n') % | |||
|
494 | (filecount, util.bytecount(filesize))) | |||
|
495 | ||||
|
496 | start = util.timer() | |||
|
497 | handledbytes = 0 | |||
|
498 | progress = repo.ui.progress | |||
|
499 | ||||
|
500 | progress(_('clone'), handledbytes, total=filesize, unit=_('bytes')) | |||
|
501 | ||||
|
502 | vfs = repo.svfs | |||
|
503 | ||||
|
504 | with repo.transaction('clone'): | |||
|
505 | with vfs.backgroundclosing(repo.ui): | |||
|
506 | for i in range(filecount): | |||
|
507 | namelen = util.uvarintdecodestream(fp) | |||
|
508 | datalen = util.uvarintdecodestream(fp) | |||
|
509 | ||||
|
510 | name = fp.read(namelen) | |||
|
511 | ||||
|
512 | if repo.ui.debugflag: | |||
|
513 | repo.ui.debug('adding %s (%s)\n' % | |||
|
514 | (name, util.bytecount(datalen))) | |||
|
515 | ||||
|
516 | with vfs(name, 'w') as ofp: | |||
|
517 | for chunk in util.filechunkiter(fp, limit=datalen): | |||
|
518 | handledbytes += len(chunk) | |||
|
519 | progress(_('clone'), handledbytes, total=filesize, | |||
|
520 | unit=_('bytes')) | |||
|
521 | ofp.write(chunk) | |||
|
522 | ||||
|
523 | # force @filecache properties to be reloaded from | |||
|
524 | # streamclone-ed file at next access | |||
|
525 | repo.invalidate(clearfilecache=True) | |||
|
526 | ||||
|
527 | elapsed = util.timer() - start | |||
|
528 | if elapsed <= 0: | |||
|
529 | elapsed = 0.001 | |||
|
530 | progress(_('clone'), None) | |||
|
531 | repo.ui.status(_('transferred %s in %.1f seconds (%s/sec)\n') % | |||
|
532 | (util.bytecount(handledbytes), elapsed, | |||
|
533 | util.bytecount(handledbytes / elapsed))) | |||
|
534 | ||||
|
535 | def applybundlev2(repo, fp, filecount, filesize, requirements): | |||
|
536 | missingreqs = [r for r in requirements if r not in repo.supported] | |||
|
537 | if missingreqs: | |||
|
538 | raise error.Abort(_('unable to apply stream clone: ' | |||
|
539 | 'unsupported format: %s') % | |||
|
540 | ', '.join(sorted(missingreqs))) | |||
|
541 | ||||
|
542 | consumev2(repo, fp, filecount, filesize) |
General Comments 0
You need to be logged in to leave comments.
Login now