##// END OF EJS Templates
lfs: correct documentation typo
Matt Harbison -
r35712:8a23082f default
parent child Browse files
Show More
@@ -1,452 +1,452
1 # blobstore.py - local and remote (speaking Git-LFS protocol) blob storages
1 # blobstore.py - local and remote (speaking Git-LFS protocol) blob storages
2 #
2 #
3 # Copyright 2017 Facebook, Inc.
3 # Copyright 2017 Facebook, Inc.
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 hashlib
10 import hashlib
11 import json
11 import json
12 import os
12 import os
13 import re
13 import re
14 import socket
14 import socket
15
15
16 from mercurial.i18n import _
16 from mercurial.i18n import _
17
17
18 from mercurial import (
18 from mercurial import (
19 error,
19 error,
20 pathutil,
20 pathutil,
21 url as urlmod,
21 url as urlmod,
22 util,
22 util,
23 vfs as vfsmod,
23 vfs as vfsmod,
24 worker,
24 worker,
25 )
25 )
26
26
27 from ..largefiles import lfutil
27 from ..largefiles import lfutil
28
28
29 # 64 bytes for SHA256
29 # 64 bytes for SHA256
30 _lfsre = re.compile(r'\A[a-f0-9]{64}\Z')
30 _lfsre = re.compile(r'\A[a-f0-9]{64}\Z')
31
31
32 class lfsvfs(vfsmod.vfs):
32 class lfsvfs(vfsmod.vfs):
33 def join(self, path):
33 def join(self, path):
34 """split the path at first two characters, like: XX/XXXXX..."""
34 """split the path at first two characters, like: XX/XXXXX..."""
35 if not _lfsre.match(path):
35 if not _lfsre.match(path):
36 raise error.ProgrammingError('unexpected lfs path: %s' % path)
36 raise error.ProgrammingError('unexpected lfs path: %s' % path)
37 return super(lfsvfs, self).join(path[0:2], path[2:])
37 return super(lfsvfs, self).join(path[0:2], path[2:])
38
38
39 def walk(self, path=None, onerror=None):
39 def walk(self, path=None, onerror=None):
40 """Yield (dirpath, [], oids) tuple for blobs under path
40 """Yield (dirpath, [], oids) tuple for blobs under path
41
41
42 Oids only exist in the root of this vfs, so dirpath is always ''.
42 Oids only exist in the root of this vfs, so dirpath is always ''.
43 """
43 """
44 root = os.path.normpath(self.base)
44 root = os.path.normpath(self.base)
45 # when dirpath == root, dirpath[prefixlen:] becomes empty
45 # when dirpath == root, dirpath[prefixlen:] becomes empty
46 # because len(dirpath) < prefixlen.
46 # because len(dirpath) < prefixlen.
47 prefixlen = len(pathutil.normasprefix(root))
47 prefixlen = len(pathutil.normasprefix(root))
48 oids = []
48 oids = []
49
49
50 for dirpath, dirs, files in os.walk(self.reljoin(self.base, path or ''),
50 for dirpath, dirs, files in os.walk(self.reljoin(self.base, path or ''),
51 onerror=onerror):
51 onerror=onerror):
52 dirpath = dirpath[prefixlen:]
52 dirpath = dirpath[prefixlen:]
53
53
54 # Silently skip unexpected files and directories
54 # Silently skip unexpected files and directories
55 if len(dirpath) == 2:
55 if len(dirpath) == 2:
56 oids.extend([dirpath + f for f in files
56 oids.extend([dirpath + f for f in files
57 if _lfsre.match(dirpath + f)])
57 if _lfsre.match(dirpath + f)])
58
58
59 yield ('', [], oids)
59 yield ('', [], oids)
60
60
61 class filewithprogress(object):
61 class filewithprogress(object):
62 """a file-like object that supports __len__ and read.
62 """a file-like object that supports __len__ and read.
63
63
64 Useful to provide progress information for how many bytes are read.
64 Useful to provide progress information for how many bytes are read.
65 """
65 """
66
66
67 def __init__(self, fp, callback):
67 def __init__(self, fp, callback):
68 self._fp = fp
68 self._fp = fp
69 self._callback = callback # func(readsize)
69 self._callback = callback # func(readsize)
70 fp.seek(0, os.SEEK_END)
70 fp.seek(0, os.SEEK_END)
71 self._len = fp.tell()
71 self._len = fp.tell()
72 fp.seek(0)
72 fp.seek(0)
73
73
74 def __len__(self):
74 def __len__(self):
75 return self._len
75 return self._len
76
76
77 def read(self, size):
77 def read(self, size):
78 if self._fp is None:
78 if self._fp is None:
79 return b''
79 return b''
80 data = self._fp.read(size)
80 data = self._fp.read(size)
81 if data:
81 if data:
82 if self._callback:
82 if self._callback:
83 self._callback(len(data))
83 self._callback(len(data))
84 else:
84 else:
85 self._fp.close()
85 self._fp.close()
86 self._fp = None
86 self._fp = None
87 return data
87 return data
88
88
89 class local(object):
89 class local(object):
90 """Local blobstore for large file contents.
90 """Local blobstore for large file contents.
91
91
92 This blobstore is used both as a cache and as a staging area for large blobs
92 This blobstore is used both as a cache and as a staging area for large blobs
93 to be uploaded to the remote blobstore.
93 to be uploaded to the remote blobstore.
94 """
94 """
95
95
96 def __init__(self, repo):
96 def __init__(self, repo):
97 fullpath = repo.svfs.join('lfs/objects')
97 fullpath = repo.svfs.join('lfs/objects')
98 self.vfs = lfsvfs(fullpath)
98 self.vfs = lfsvfs(fullpath)
99 usercache = lfutil._usercachedir(repo.ui, 'lfs')
99 usercache = lfutil._usercachedir(repo.ui, 'lfs')
100 self.cachevfs = lfsvfs(usercache)
100 self.cachevfs = lfsvfs(usercache)
101 self.ui = repo.ui
101 self.ui = repo.ui
102
102
103 def open(self, oid):
103 def open(self, oid):
104 """Open a read-only file descriptor to the named blob, in either the
104 """Open a read-only file descriptor to the named blob, in either the
105 usercache or the local store."""
105 usercache or the local store."""
106 # The usercache is the most likely place to hold the file. Commit will
106 # The usercache is the most likely place to hold the file. Commit will
107 # write to both it and the local store, as will anything that downloads
107 # write to both it and the local store, as will anything that downloads
108 # the blobs. However, things like clone without an update won't
108 # the blobs. However, things like clone without an update won't
109 # populate the local store. For an init + push of a local clone,
109 # populate the local store. For an init + push of a local clone,
110 # the usercache is the only place it _could_ be. If not present, the
110 # the usercache is the only place it _could_ be. If not present, the
111 # missing file msg here will indicate the local repo, not the usercache.
111 # missing file msg here will indicate the local repo, not the usercache.
112 if self.cachevfs.exists(oid):
112 if self.cachevfs.exists(oid):
113 return self.cachevfs(oid, 'rb')
113 return self.cachevfs(oid, 'rb')
114
114
115 return self.vfs(oid, 'rb')
115 return self.vfs(oid, 'rb')
116
116
117 def download(self, oid, src):
117 def download(self, oid, src):
118 """Read the blob from the remote source in chunks, verify the content,
118 """Read the blob from the remote source in chunks, verify the content,
119 and write to this local blobstore."""
119 and write to this local blobstore."""
120 sha256 = hashlib.sha256()
120 sha256 = hashlib.sha256()
121
121
122 with self.vfs(oid, 'wb', atomictemp=True) as fp:
122 with self.vfs(oid, 'wb', atomictemp=True) as fp:
123 for chunk in util.filechunkiter(src, size=1048576):
123 for chunk in util.filechunkiter(src, size=1048576):
124 fp.write(chunk)
124 fp.write(chunk)
125 sha256.update(chunk)
125 sha256.update(chunk)
126
126
127 realoid = sha256.hexdigest()
127 realoid = sha256.hexdigest()
128 if realoid != oid:
128 if realoid != oid:
129 raise error.Abort(_('corrupt remote lfs object: %s') % oid)
129 raise error.Abort(_('corrupt remote lfs object: %s') % oid)
130
130
131 # XXX: should we verify the content of the cache, and hardlink back to
131 # XXX: should we verify the content of the cache, and hardlink back to
132 # the local store on success, but truncate, write and link on failure?
132 # the local store on success, but truncate, write and link on failure?
133 if not self.cachevfs.exists(oid):
133 if not self.cachevfs.exists(oid):
134 self.ui.note(_('lfs: adding %s to the usercache\n') % oid)
134 self.ui.note(_('lfs: adding %s to the usercache\n') % oid)
135 lfutil.link(self.vfs.join(oid), self.cachevfs.join(oid))
135 lfutil.link(self.vfs.join(oid), self.cachevfs.join(oid))
136
136
137 def write(self, oid, data):
137 def write(self, oid, data):
138 """Write blob to local blobstore.
138 """Write blob to local blobstore.
139
139
140 This should only be called from the filelog during a commit or similar.
140 This should only be called from the filelog during a commit or similar.
141 As such, there is no need to verify the data. Imports from a remote
141 As such, there is no need to verify the data. Imports from a remote
142 store must use ``download()`` instead."""
142 store must use ``download()`` instead."""
143 with self.vfs(oid, 'wb', atomictemp=True) as fp:
143 with self.vfs(oid, 'wb', atomictemp=True) as fp:
144 fp.write(data)
144 fp.write(data)
145
145
146 # XXX: should we verify the content of the cache, and hardlink back to
146 # XXX: should we verify the content of the cache, and hardlink back to
147 # the local store on success, but truncate, write and link on failure?
147 # the local store on success, but truncate, write and link on failure?
148 if not self.cachevfs.exists(oid):
148 if not self.cachevfs.exists(oid):
149 self.ui.note(_('lfs: adding %s to the usercache\n') % oid)
149 self.ui.note(_('lfs: adding %s to the usercache\n') % oid)
150 lfutil.link(self.vfs.join(oid), self.cachevfs.join(oid))
150 lfutil.link(self.vfs.join(oid), self.cachevfs.join(oid))
151
151
152 def read(self, oid, verify=True):
152 def read(self, oid, verify=True):
153 """Read blob from local blobstore."""
153 """Read blob from local blobstore."""
154 if not self.vfs.exists(oid):
154 if not self.vfs.exists(oid):
155 blob = self._read(self.cachevfs, oid, verify)
155 blob = self._read(self.cachevfs, oid, verify)
156
156
157 # Even if revlog will verify the content, it needs to be verified
157 # Even if revlog will verify the content, it needs to be verified
158 # now before making the hardlink to avoid propagating corrupt blobs.
158 # now before making the hardlink to avoid propagating corrupt blobs.
159 # Don't abort if corruption is detected, because `hg verify` will
159 # Don't abort if corruption is detected, because `hg verify` will
160 # give more useful info about the corruption- simply don't add the
160 # give more useful info about the corruption- simply don't add the
161 # hardlink.
161 # hardlink.
162 if verify or hashlib.sha256(blob).hexdigest() == oid:
162 if verify or hashlib.sha256(blob).hexdigest() == oid:
163 self.ui.note(_('lfs: found %s in the usercache\n') % oid)
163 self.ui.note(_('lfs: found %s in the usercache\n') % oid)
164 lfutil.link(self.cachevfs.join(oid), self.vfs.join(oid))
164 lfutil.link(self.cachevfs.join(oid), self.vfs.join(oid))
165 else:
165 else:
166 self.ui.note(_('lfs: found %s in the local lfs store\n') % oid)
166 self.ui.note(_('lfs: found %s in the local lfs store\n') % oid)
167 blob = self._read(self.vfs, oid, verify)
167 blob = self._read(self.vfs, oid, verify)
168 return blob
168 return blob
169
169
170 def _read(self, vfs, oid, verify):
170 def _read(self, vfs, oid, verify):
171 """Read blob (after verifying) from the given store"""
171 """Read blob (after verifying) from the given store"""
172 blob = vfs.read(oid)
172 blob = vfs.read(oid)
173 if verify:
173 if verify:
174 _verify(oid, blob)
174 _verify(oid, blob)
175 return blob
175 return blob
176
176
177 def has(self, oid):
177 def has(self, oid):
178 """Returns True if the local blobstore contains the requested blob,
178 """Returns True if the local blobstore contains the requested blob,
179 False otherwise."""
179 False otherwise."""
180 return self.cachevfs.exists(oid) or self.vfs.exists(oid)
180 return self.cachevfs.exists(oid) or self.vfs.exists(oid)
181
181
182 class _gitlfsremote(object):
182 class _gitlfsremote(object):
183
183
184 def __init__(self, repo, url):
184 def __init__(self, repo, url):
185 ui = repo.ui
185 ui = repo.ui
186 self.ui = ui
186 self.ui = ui
187 baseurl, authinfo = url.authinfo()
187 baseurl, authinfo = url.authinfo()
188 self.baseurl = baseurl.rstrip('/')
188 self.baseurl = baseurl.rstrip('/')
189 useragent = repo.ui.config('experimental', 'lfs.user-agent')
189 useragent = repo.ui.config('experimental', 'lfs.user-agent')
190 if not useragent:
190 if not useragent:
191 useragent = 'mercurial/%s git/2.15.1' % util.version()
191 useragent = 'mercurial/%s git/2.15.1' % util.version()
192 self.urlopener = urlmod.opener(ui, authinfo, useragent)
192 self.urlopener = urlmod.opener(ui, authinfo, useragent)
193 self.retry = ui.configint('lfs', 'retry')
193 self.retry = ui.configint('lfs', 'retry')
194
194
195 def writebatch(self, pointers, fromstore):
195 def writebatch(self, pointers, fromstore):
196 """Batch upload from local to remote blobstore."""
196 """Batch upload from local to remote blobstore."""
197 self._batch(pointers, fromstore, 'upload')
197 self._batch(pointers, fromstore, 'upload')
198
198
199 def readbatch(self, pointers, tostore):
199 def readbatch(self, pointers, tostore):
200 """Batch download from remote to local blostore."""
200 """Batch download from remote to local blostore."""
201 self._batch(pointers, tostore, 'download')
201 self._batch(pointers, tostore, 'download')
202
202
203 def _batchrequest(self, pointers, action):
203 def _batchrequest(self, pointers, action):
204 """Get metadata about objects pointed by pointers for given action
204 """Get metadata about objects pointed by pointers for given action
205
205
206 Return decoded JSON object like {'objects': [{'oid': '', 'size': 1}]}
206 Return decoded JSON object like {'objects': [{'oid': '', 'size': 1}]}
207 See https://github.com/git-lfs/git-lfs/blob/master/docs/api/batch.md
207 See https://github.com/git-lfs/git-lfs/blob/master/docs/api/batch.md
208 """
208 """
209 objects = [{'oid': p.oid(), 'size': p.size()} for p in pointers]
209 objects = [{'oid': p.oid(), 'size': p.size()} for p in pointers]
210 requestdata = json.dumps({
210 requestdata = json.dumps({
211 'objects': objects,
211 'objects': objects,
212 'operation': action,
212 'operation': action,
213 })
213 })
214 batchreq = util.urlreq.request('%s/objects/batch' % self.baseurl,
214 batchreq = util.urlreq.request('%s/objects/batch' % self.baseurl,
215 data=requestdata)
215 data=requestdata)
216 batchreq.add_header('Accept', 'application/vnd.git-lfs+json')
216 batchreq.add_header('Accept', 'application/vnd.git-lfs+json')
217 batchreq.add_header('Content-Type', 'application/vnd.git-lfs+json')
217 batchreq.add_header('Content-Type', 'application/vnd.git-lfs+json')
218 try:
218 try:
219 rawjson = self.urlopener.open(batchreq).read()
219 rawjson = self.urlopener.open(batchreq).read()
220 except util.urlerr.httperror as ex:
220 except util.urlerr.httperror as ex:
221 raise LfsRemoteError(_('LFS HTTP error: %s (action=%s)')
221 raise LfsRemoteError(_('LFS HTTP error: %s (action=%s)')
222 % (ex, action))
222 % (ex, action))
223 try:
223 try:
224 response = json.loads(rawjson)
224 response = json.loads(rawjson)
225 except ValueError:
225 except ValueError:
226 raise LfsRemoteError(_('LFS server returns invalid JSON: %s')
226 raise LfsRemoteError(_('LFS server returns invalid JSON: %s')
227 % rawjson)
227 % rawjson)
228 return response
228 return response
229
229
230 def _checkforservererror(self, pointers, responses, action):
230 def _checkforservererror(self, pointers, responses, action):
231 """Scans errors from objects
231 """Scans errors from objects
232
232
233 Returns LfsRemoteError if any objects has an error"""
233 Raises LfsRemoteError if any objects have an error"""
234 for response in responses:
234 for response in responses:
235 # The server should return 404 when objects cannot be found. Some
235 # The server should return 404 when objects cannot be found. Some
236 # server implementation (ex. lfs-test-server) does not set "error"
236 # server implementation (ex. lfs-test-server) does not set "error"
237 # but just removes "download" from "actions". Treat that case
237 # but just removes "download" from "actions". Treat that case
238 # as the same as 404 error.
238 # as the same as 404 error.
239 notfound = (response.get('error', {}).get('code') == 404
239 notfound = (response.get('error', {}).get('code') == 404
240 or (action == 'download'
240 or (action == 'download'
241 and action not in response.get('actions', [])))
241 and action not in response.get('actions', [])))
242 if notfound:
242 if notfound:
243 ptrmap = {p.oid(): p for p in pointers}
243 ptrmap = {p.oid(): p for p in pointers}
244 p = ptrmap.get(response['oid'], None)
244 p = ptrmap.get(response['oid'], None)
245 if p:
245 if p:
246 filename = getattr(p, 'filename', 'unknown')
246 filename = getattr(p, 'filename', 'unknown')
247 raise LfsRemoteError(
247 raise LfsRemoteError(
248 _(('LFS server error. Remote object '
248 _(('LFS server error. Remote object '
249 'for "%s" not found: %r')) % (filename, response))
249 'for "%s" not found: %r')) % (filename, response))
250 if 'error' in response:
250 if 'error' in response:
251 raise LfsRemoteError(_('LFS server error: %r') % response)
251 raise LfsRemoteError(_('LFS server error: %r') % response)
252
252
253 def _extractobjects(self, response, pointers, action):
253 def _extractobjects(self, response, pointers, action):
254 """extract objects from response of the batch API
254 """extract objects from response of the batch API
255
255
256 response: parsed JSON object returned by batch API
256 response: parsed JSON object returned by batch API
257 return response['objects'] filtered by action
257 return response['objects'] filtered by action
258 raise if any object has an error
258 raise if any object has an error
259 """
259 """
260 # Scan errors from objects - fail early
260 # Scan errors from objects - fail early
261 objects = response.get('objects', [])
261 objects = response.get('objects', [])
262 self._checkforservererror(pointers, objects, action)
262 self._checkforservererror(pointers, objects, action)
263
263
264 # Filter objects with given action. Practically, this skips uploading
264 # Filter objects with given action. Practically, this skips uploading
265 # objects which exist in the server.
265 # objects which exist in the server.
266 filteredobjects = [o for o in objects if action in o.get('actions', [])]
266 filteredobjects = [o for o in objects if action in o.get('actions', [])]
267
267
268 return filteredobjects
268 return filteredobjects
269
269
270 def _basictransfer(self, obj, action, localstore):
270 def _basictransfer(self, obj, action, localstore):
271 """Download or upload a single object using basic transfer protocol
271 """Download or upload a single object using basic transfer protocol
272
272
273 obj: dict, an object description returned by batch API
273 obj: dict, an object description returned by batch API
274 action: string, one of ['upload', 'download']
274 action: string, one of ['upload', 'download']
275 localstore: blobstore.local
275 localstore: blobstore.local
276
276
277 See https://github.com/git-lfs/git-lfs/blob/master/docs/api/\
277 See https://github.com/git-lfs/git-lfs/blob/master/docs/api/\
278 basic-transfers.md
278 basic-transfers.md
279 """
279 """
280 oid = str(obj['oid'])
280 oid = str(obj['oid'])
281
281
282 href = str(obj['actions'][action].get('href'))
282 href = str(obj['actions'][action].get('href'))
283 headers = obj['actions'][action].get('header', {}).items()
283 headers = obj['actions'][action].get('header', {}).items()
284
284
285 request = util.urlreq.request(href)
285 request = util.urlreq.request(href)
286 if action == 'upload':
286 if action == 'upload':
287 # If uploading blobs, read data from local blobstore.
287 # If uploading blobs, read data from local blobstore.
288 with localstore.open(oid) as fp:
288 with localstore.open(oid) as fp:
289 _verifyfile(oid, fp)
289 _verifyfile(oid, fp)
290 request.data = filewithprogress(localstore.open(oid), None)
290 request.data = filewithprogress(localstore.open(oid), None)
291 request.get_method = lambda: 'PUT'
291 request.get_method = lambda: 'PUT'
292
292
293 for k, v in headers:
293 for k, v in headers:
294 request.add_header(k, v)
294 request.add_header(k, v)
295
295
296 response = b''
296 response = b''
297 try:
297 try:
298 req = self.urlopener.open(request)
298 req = self.urlopener.open(request)
299 if action == 'download':
299 if action == 'download':
300 # If downloading blobs, store downloaded data to local blobstore
300 # If downloading blobs, store downloaded data to local blobstore
301 localstore.download(oid, req)
301 localstore.download(oid, req)
302 else:
302 else:
303 while True:
303 while True:
304 data = req.read(1048576)
304 data = req.read(1048576)
305 if not data:
305 if not data:
306 break
306 break
307 response += data
307 response += data
308 if response:
308 if response:
309 self.ui.debug('lfs %s response: %s' % (action, response))
309 self.ui.debug('lfs %s response: %s' % (action, response))
310 except util.urlerr.httperror as ex:
310 except util.urlerr.httperror as ex:
311 raise LfsRemoteError(_('HTTP error: %s (oid=%s, action=%s)')
311 raise LfsRemoteError(_('HTTP error: %s (oid=%s, action=%s)')
312 % (ex, oid, action))
312 % (ex, oid, action))
313
313
314 def _batch(self, pointers, localstore, action):
314 def _batch(self, pointers, localstore, action):
315 if action not in ['upload', 'download']:
315 if action not in ['upload', 'download']:
316 raise error.ProgrammingError('invalid Git-LFS action: %s' % action)
316 raise error.ProgrammingError('invalid Git-LFS action: %s' % action)
317
317
318 response = self._batchrequest(pointers, action)
318 response = self._batchrequest(pointers, action)
319 objects = self._extractobjects(response, pointers, action)
319 objects = self._extractobjects(response, pointers, action)
320 total = sum(x.get('size', 0) for x in objects)
320 total = sum(x.get('size', 0) for x in objects)
321 sizes = {}
321 sizes = {}
322 for obj in objects:
322 for obj in objects:
323 sizes[obj.get('oid')] = obj.get('size', 0)
323 sizes[obj.get('oid')] = obj.get('size', 0)
324 topic = {'upload': _('lfs uploading'),
324 topic = {'upload': _('lfs uploading'),
325 'download': _('lfs downloading')}[action]
325 'download': _('lfs downloading')}[action]
326 if len(objects) > 1:
326 if len(objects) > 1:
327 self.ui.note(_('lfs: need to transfer %d objects (%s)\n')
327 self.ui.note(_('lfs: need to transfer %d objects (%s)\n')
328 % (len(objects), util.bytecount(total)))
328 % (len(objects), util.bytecount(total)))
329 self.ui.progress(topic, 0, total=total)
329 self.ui.progress(topic, 0, total=total)
330 def transfer(chunk):
330 def transfer(chunk):
331 for obj in chunk:
331 for obj in chunk:
332 objsize = obj.get('size', 0)
332 objsize = obj.get('size', 0)
333 if self.ui.verbose:
333 if self.ui.verbose:
334 if action == 'download':
334 if action == 'download':
335 msg = _('lfs: downloading %s (%s)\n')
335 msg = _('lfs: downloading %s (%s)\n')
336 elif action == 'upload':
336 elif action == 'upload':
337 msg = _('lfs: uploading %s (%s)\n')
337 msg = _('lfs: uploading %s (%s)\n')
338 self.ui.note(msg % (obj.get('oid'),
338 self.ui.note(msg % (obj.get('oid'),
339 util.bytecount(objsize)))
339 util.bytecount(objsize)))
340 retry = self.retry
340 retry = self.retry
341 while True:
341 while True:
342 try:
342 try:
343 self._basictransfer(obj, action, localstore)
343 self._basictransfer(obj, action, localstore)
344 yield 1, obj.get('oid')
344 yield 1, obj.get('oid')
345 break
345 break
346 except socket.error as ex:
346 except socket.error as ex:
347 if retry > 0:
347 if retry > 0:
348 self.ui.note(
348 self.ui.note(
349 _('lfs: failed: %r (remaining retry %d)\n')
349 _('lfs: failed: %r (remaining retry %d)\n')
350 % (ex, retry))
350 % (ex, retry))
351 retry -= 1
351 retry -= 1
352 continue
352 continue
353 raise
353 raise
354
354
355 oids = worker.worker(self.ui, 0.1, transfer, (),
355 oids = worker.worker(self.ui, 0.1, transfer, (),
356 sorted(objects, key=lambda o: o.get('oid')))
356 sorted(objects, key=lambda o: o.get('oid')))
357 processed = 0
357 processed = 0
358 for _one, oid in oids:
358 for _one, oid in oids:
359 processed += sizes[oid]
359 processed += sizes[oid]
360 self.ui.progress(topic, processed, total=total)
360 self.ui.progress(topic, processed, total=total)
361 self.ui.note(_('lfs: processed: %s\n') % oid)
361 self.ui.note(_('lfs: processed: %s\n') % oid)
362 self.ui.progress(topic, pos=None, total=total)
362 self.ui.progress(topic, pos=None, total=total)
363
363
364 def __del__(self):
364 def __del__(self):
365 # copied from mercurial/httppeer.py
365 # copied from mercurial/httppeer.py
366 urlopener = getattr(self, 'urlopener', None)
366 urlopener = getattr(self, 'urlopener', None)
367 if urlopener:
367 if urlopener:
368 for h in urlopener.handlers:
368 for h in urlopener.handlers:
369 h.close()
369 h.close()
370 getattr(h, "close_all", lambda : None)()
370 getattr(h, "close_all", lambda : None)()
371
371
372 class _dummyremote(object):
372 class _dummyremote(object):
373 """Dummy store storing blobs to temp directory."""
373 """Dummy store storing blobs to temp directory."""
374
374
375 def __init__(self, repo, url):
375 def __init__(self, repo, url):
376 fullpath = repo.vfs.join('lfs', url.path)
376 fullpath = repo.vfs.join('lfs', url.path)
377 self.vfs = lfsvfs(fullpath)
377 self.vfs = lfsvfs(fullpath)
378
378
379 def writebatch(self, pointers, fromstore):
379 def writebatch(self, pointers, fromstore):
380 for p in pointers:
380 for p in pointers:
381 content = fromstore.read(p.oid(), verify=True)
381 content = fromstore.read(p.oid(), verify=True)
382 with self.vfs(p.oid(), 'wb', atomictemp=True) as fp:
382 with self.vfs(p.oid(), 'wb', atomictemp=True) as fp:
383 fp.write(content)
383 fp.write(content)
384
384
385 def readbatch(self, pointers, tostore):
385 def readbatch(self, pointers, tostore):
386 for p in pointers:
386 for p in pointers:
387 with self.vfs(p.oid(), 'rb') as fp:
387 with self.vfs(p.oid(), 'rb') as fp:
388 tostore.download(p.oid(), fp)
388 tostore.download(p.oid(), fp)
389
389
390 class _nullremote(object):
390 class _nullremote(object):
391 """Null store storing blobs to /dev/null."""
391 """Null store storing blobs to /dev/null."""
392
392
393 def __init__(self, repo, url):
393 def __init__(self, repo, url):
394 pass
394 pass
395
395
396 def writebatch(self, pointers, fromstore):
396 def writebatch(self, pointers, fromstore):
397 pass
397 pass
398
398
399 def readbatch(self, pointers, tostore):
399 def readbatch(self, pointers, tostore):
400 pass
400 pass
401
401
402 class _promptremote(object):
402 class _promptremote(object):
403 """Prompt user to set lfs.url when accessed."""
403 """Prompt user to set lfs.url when accessed."""
404
404
405 def __init__(self, repo, url):
405 def __init__(self, repo, url):
406 pass
406 pass
407
407
408 def writebatch(self, pointers, fromstore, ui=None):
408 def writebatch(self, pointers, fromstore, ui=None):
409 self._prompt()
409 self._prompt()
410
410
411 def readbatch(self, pointers, tostore, ui=None):
411 def readbatch(self, pointers, tostore, ui=None):
412 self._prompt()
412 self._prompt()
413
413
414 def _prompt(self):
414 def _prompt(self):
415 raise error.Abort(_('lfs.url needs to be configured'))
415 raise error.Abort(_('lfs.url needs to be configured'))
416
416
417 _storemap = {
417 _storemap = {
418 'https': _gitlfsremote,
418 'https': _gitlfsremote,
419 'http': _gitlfsremote,
419 'http': _gitlfsremote,
420 'file': _dummyremote,
420 'file': _dummyremote,
421 'null': _nullremote,
421 'null': _nullremote,
422 None: _promptremote,
422 None: _promptremote,
423 }
423 }
424
424
425 def _verify(oid, content):
425 def _verify(oid, content):
426 realoid = hashlib.sha256(content).hexdigest()
426 realoid = hashlib.sha256(content).hexdigest()
427 if realoid != oid:
427 if realoid != oid:
428 raise error.Abort(_('detected corrupt lfs object: %s') % oid,
428 raise error.Abort(_('detected corrupt lfs object: %s') % oid,
429 hint=_('run hg verify'))
429 hint=_('run hg verify'))
430
430
431 def _verifyfile(oid, fp):
431 def _verifyfile(oid, fp):
432 sha256 = hashlib.sha256()
432 sha256 = hashlib.sha256()
433 while True:
433 while True:
434 data = fp.read(1024 * 1024)
434 data = fp.read(1024 * 1024)
435 if not data:
435 if not data:
436 break
436 break
437 sha256.update(data)
437 sha256.update(data)
438 realoid = sha256.hexdigest()
438 realoid = sha256.hexdigest()
439 if realoid != oid:
439 if realoid != oid:
440 raise error.Abort(_('detected corrupt lfs object: %s') % oid,
440 raise error.Abort(_('detected corrupt lfs object: %s') % oid,
441 hint=_('run hg verify'))
441 hint=_('run hg verify'))
442
442
443 def remote(repo):
443 def remote(repo):
444 """remotestore factory. return a store in _storemap depending on config"""
444 """remotestore factory. return a store in _storemap depending on config"""
445 url = util.url(repo.ui.config('lfs', 'url') or '')
445 url = util.url(repo.ui.config('lfs', 'url') or '')
446 scheme = url.scheme
446 scheme = url.scheme
447 if scheme not in _storemap:
447 if scheme not in _storemap:
448 raise error.Abort(_('lfs: unknown url scheme: %s') % scheme)
448 raise error.Abort(_('lfs: unknown url scheme: %s') % scheme)
449 return _storemap[scheme](repo, url)
449 return _storemap[scheme](repo, url)
450
450
451 class LfsRemoteError(error.RevlogError):
451 class LfsRemoteError(error.RevlogError):
452 pass
452 pass
General Comments 0
You need to be logged in to leave comments. Login now