« back to all changes in this revision
Viewing changes to fetch.py
- browse files at revision 57
- compare with another revision
- download diff
- download tarball
- view history from revision 57
- Committer: Jelmer Vernooij
- Date: 2010-05-22 23:40:04 UTC
- mfrom: (17.25.364 trunk)
- Revision ID: jelmer@samba.org-20100522234004-48r0bfkodp3xspgv
* New upstream release.
* Bump standards version to 3.8.4.
* Bump standards version to 3.8.4.
- files added:
- notes/git-serve.txt
- files modified:
- NEWS
added
removed
fetch.py
1
# Copyright (C) 2008 Jelmer Vernooij <jelmer@samba.org>
1
# Copyright (C) 2008-2010 Jelmer Vernooij <jelmer@samba.org>
2
2
#
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
17
17
from dulwich.objects import (
18
18
Commit,
19
19
Tag,
20
Tree,
20
21
S_ISGITLINK,
21
22
)
22
23
from dulwich.object_store import (
23
24
tree_lookup_path,
24
25
)
26
from itertools import (
27
imap,
28
)
29
import posixpath
25
30
import re
26
31
import stat
27
32
28
33
from bzrlib import (
29
34
debug,
30
lru_cache,
31
35
osutils,
32
36
trace,
33
37
ui,
34
urlutils,
35
38
)
36
39
from bzrlib.errors import (
37
40
BzrError,
50
53
from bzrlib.revision import (
51
54
NULL_REVISION,
52
55
)
56
from bzrlib.revisiontree import (
57
RevisionTree,
58
)
53
59
from bzrlib.tsort import (
54
60
topo_sort,
55
61
)
56
62
from bzrlib.versionedfile import (
57
FulltextContentFactory,
63
ChunkedContentFactory,
58
64
)
59
65
60
66
from bzrlib.plugins.git.mapping import (
61
67
DEFAULT_FILE_MODE,
62
inventory_to_tree_and_blobs,
63
68
mode_is_executable,
64
squash_revision,
69
mode_kind,
65
70
warn_unusual_mode,
66
71
)
67
72
from bzrlib.plugins.git.object_store import (
68
73
BazaarObjectStore,
74
LRUTreeCache,
75
_tree_to_objects,
69
76
)
70
77
from bzrlib.plugins.git.remote import (
71
78
RemoteGitRepository,
77
84
)
78
85
79
86
80
MAX_INV_CACHE_SIZE = 50 * 1024 * 1024
81
82
83
def import_git_blob(texts, mapping, path, hexsha, base_inv, base_ie, parent_id,
84
revision_id, parent_invs, shagitmap, lookup_object, executable, symlink):
87
def import_git_blob(texts, mapping, path, name, (base_hexsha, hexsha),
88
base_inv, parent_id, revision_id,
89
parent_invs, lookup_object, (base_mode, mode), store_updater,
90
lookup_file_id):
85
91
"""Import a git blob object into a bzr repository.
86
92
87
93
:param texts: VersionedFiles to add to
89
95
:param blob: A git blob
90
96
:return: Inventory delta for this file
91
97
"""
92
file_id = mapping.generate_file_id(path)
93
if symlink:
98
if mapping.is_control_file(path):
99
return []
100
if base_hexsha == hexsha and base_mode == mode:
101
# If nothing has changed since the base revision, we're done
102
return []
103
file_id = lookup_file_id(path)
104
if stat.S_ISLNK(mode):
94
105
cls = InventoryLink
95
106
else:
96
107
cls = InventoryFile
97
# We just have to hope this is indeed utf-8:
98
ie = cls(file_id, urlutils.basename(path).decode("utf-8"), parent_id)
99
ie.executable = executable
100
# See if this has changed at all
101
if base_ie is None:
102
base_sha = None
103
else:
104
try:
105
base_sha = shagitmap.lookup_blob(file_id, base_ie.revision)
106
except KeyError:
107
base_sha = None
108
else:
109
if (base_sha == hexsha and base_ie.executable == ie.executable
110
and base_ie.kind == ie.kind):
111
# If nothing has changed since the base revision, we're done
112
return [], []
113
if base_sha == hexsha and base_ie.kind == ie.kind:
108
ie = cls(file_id, name.decode("utf-8"), parent_id)
109
ie.executable = mode_is_executable(mode)
110
if base_hexsha == hexsha and mode_kind(base_mode) == mode_kind(mode):
111
base_ie = base_inv[base_inv.path2id(path)]
114
112
ie.text_size = base_ie.text_size
115
113
ie.text_sha1 = base_ie.text_sha1
116
114
ie.symlink_target = base_ie.symlink_target
126
124
ie.text_size = None
127
125
ie.text_sha1 = None
128
126
else:
129
ie.text_size = len(blob.data)
130
ie.text_sha1 = osutils.sha_string(blob.data)
127
ie.text_size = sum(imap(len, blob.chunked))
128
ie.text_sha1 = osutils.sha_strings(blob.chunked)
131
129
# Check what revision we should store
132
130
parent_keys = []
133
131
for pinv in parent_invs:
134
if pinv.revision_id == base_inv.revision_id:
135
pie = base_ie
136
if pie is None:
137
continue
138
else:
139
try:
140
pie = pinv[file_id]
141
except NoSuchId:
142
continue
143
if pie.text_sha1 == ie.text_sha1 and pie.executable == ie.executable and pie.symlink_target == ie.symlink_target:
132
try:
133
pie = pinv[file_id]
134
except NoSuchId:
135
continue
136
if (pie.text_sha1 == ie.text_sha1 and
137
pie.executable == ie.executable and
138
pie.symlink_target == ie.symlink_target):
144
139
# found a revision in one of the parents to use
145
140
ie.revision = pie.revision
146
141
break
147
parent_keys.append((file_id, pie.revision))
142
parent_key = (file_id, pie.revision)
143
if not parent_key in parent_keys:
144
parent_keys.append(parent_key)
148
145
if ie.revision is None:
149
146
# Need to store a new revision
150
147
ie.revision = revision_id
151
assert file_id is not None
152
148
assert ie.revision is not None
153
texts.insert_record_stream([FulltextContentFactory((file_id, ie.revision), tuple(parent_keys), ie.text_sha1, blob.data)])
154
shamap = [(hexsha, "blob", (ie.file_id, ie.revision))]
155
else:
156
shamap = []
149
if ie.kind == 'symlink':
150
chunks = []
151
else:
152
chunks = blob.chunked
153
texts.insert_record_stream([
154
ChunkedContentFactory((file_id, ie.revision),
155
tuple(parent_keys), ie.text_sha1, chunks)])
157
156
invdelta = []
158
if base_ie is not None:
159
old_path = base_inv.id2path(file_id)
160
if base_ie.kind == "directory":
161
invdelta.extend(remove_disappeared_children(old_path, base_ie.children, []))
157
if base_hexsha is not None:
158
old_path = path # Renames are not supported yet
159
if stat.S_ISDIR(base_mode):
160
invdelta.extend(remove_disappeared_children(base_inv, old_path,
161
lookup_object(base_hexsha), [], lookup_object))
162
162
else:
163
163
old_path = None
164
164
invdelta.append((old_path, path, file_id, ie))
165
return (invdelta, shamap)
165
if base_hexsha != hexsha:
166
store_updater.add_object(blob, ie)
167
return invdelta
166
168
167
169
168
170
class SubmodulesRequireSubtrees(BzrError):
170
172
internal = False
171
173
172
174
173
def import_git_submodule(texts, mapping, path, hexsha, base_inv, base_ie,
174
parent_id, revision_id, parent_invs, shagitmap, lookup_object):
175
file_id = mapping.generate_file_id(path)
176
ie = TreeReference(file_id, urlutils.basename(path.decode("utf-8")),
177
parent_id)
175
def import_git_submodule(texts, mapping, path, name, (base_hexsha, hexsha),
176
base_inv, parent_id, revision_id, parent_invs, lookup_object,
177
(base_mode, mode), store_updater, lookup_file_id):
178
if base_hexsha == hexsha and base_mode == mode:
179
return [], {}
180
file_id = lookup_file_id(path)
181
ie = TreeReference(file_id, name.decode("utf-8"), parent_id)
178
182
ie.revision = revision_id
179
if base_ie is None:
183
if base_hexsha is None:
180
184
oldpath = None
181
185
else:
182
186
oldpath = path
183
if (base_ie.kind == ie.kind and
184
base_ie.reference_revision == ie.reference_revision):
185
ie.revision = base_ie.revision
186
187
ie.reference_revision = mapping.revision_id_foreign_to_bzr(hexsha)
187
texts.insert_record_stream([FulltextContentFactory((file_id, ie.revision), (), None, "")])
188
texts.insert_record_stream([
189
ChunkedContentFactory((file_id, ie.revision), (), None, [])])
188
190
invdelta = [(oldpath, path, file_id, ie)]
189
return invdelta, {}, {}
190
191
192
def remove_disappeared_children(path, base_children, existing_children):
191
return invdelta, {}
192
193
194
def remove_disappeared_children(base_inv, path, base_tree, existing_children,
195
lookup_object):
193
196
ret = []
194
deletable = [(osutils.pathjoin(path, k), v) for k,v in base_children.iteritems() if k not in existing_children]
195
while deletable:
196
(path, ie) = deletable.pop()
197
ret.append((path, None, ie.file_id, None))
198
if ie.kind == "directory":
199
for name, child_ie in ie.children.iteritems():
200
deletable.append((osutils.pathjoin(path, name), child_ie))
197
for name, mode, hexsha in base_tree.iteritems():
198
if name in existing_children:
199
continue
200
c_path = posixpath.join(path, name.decode("utf-8"))
201
ret.append((c_path, None, base_inv.path2id(c_path), None))
202
if stat.S_ISDIR(mode):
203
ret.extend(remove_disappeared_children(
204
base_inv, c_path, lookup_object(hexsha), [], lookup_object))
201
205
return ret
202
206
203
207
204
def import_git_tree(texts, mapping, path, hexsha, base_inv, base_ie, parent_id,
205
revision_id, parent_invs, shagitmap, lookup_object, allow_submodules=False):
208
def import_git_tree(texts, mapping, path, name, (base_hexsha, hexsha),
209
base_inv, parent_id, revision_id, parent_invs,
210
lookup_object, (base_mode, mode), store_updater,
211
lookup_file_id, allow_submodules=False):
206
212
"""Import a git tree object into a bzr repository.
207
213
208
214
:param texts: VersionedFiles object to add to
211
217
:param base_inv: Base inventory against which to return inventory delta
212
218
:return: Inventory delta for this subtree
213
219
"""
220
if base_hexsha == hexsha and base_mode == mode:
221
# If nothing has changed since the base revision, we're done
222
return [], {}
214
223
invdelta = []
215
file_id = mapping.generate_file_id(path)
224
file_id = lookup_file_id(path)
216
225
# We just have to hope this is indeed utf-8:
217
ie = InventoryDirectory(file_id, urlutils.basename(path.decode("utf-8")),
218
parent_id)
219
if base_ie is None:
220
# Newly appeared here
226
ie = InventoryDirectory(file_id, name.decode("utf-8"), parent_id)
227
tree = lookup_object(hexsha)
228
if base_hexsha is None:
229
base_tree = None
230
old_path = None # Newly appeared here
231
else:
232
base_tree = lookup_object(base_hexsha)
233
old_path = path # Renames aren't supported yet
234
if base_tree is None or type(base_tree) is not Tree:
221
235
ie.revision = revision_id
222
texts.insert_record_stream([FulltextContentFactory((file_id, ie.revision), (), None, "")])
223
invdelta.append((None, path, file_id, ie))
224
else:
225
# See if this has changed at all
226
try:
227
base_sha = shagitmap.lookup_tree(file_id, base_inv.revision_id)
228
except KeyError:
229
pass
230
else:
231
if base_sha == hexsha:
232
# If nothing has changed since the base revision, we're done
233
return [], {}, []
234
if base_ie.kind != "directory":
235
ie.revision = revision_id
236
texts.insert_record_stream([FulltextContentFactory((ie.file_id, ie.revision), (), None, "")])
237
invdelta.append((base_inv.id2path(ie.file_id), path, ie.file_id, ie))
238
if base_ie is not None and base_ie.kind == "directory":
239
base_children = base_ie.children
240
else:
241
base_children = {}
236
invdelta.append((old_path, path, ie.file_id, ie))
237
texts.insert_record_stream([
238
ChunkedContentFactory((ie.file_id, ie.revision), (), None, [])])
242
239
# Remember for next time
243
240
existing_children = set()
244
241
child_modes = {}
245
shamap = []
246
tree = lookup_object(hexsha)
247
for mode, name, child_hexsha in tree.entries():
248
basename = name.decode("utf-8")
249
existing_children.add(basename)
250
child_path = osutils.pathjoin(path, name)
251
if stat.S_ISDIR(mode):
252
subinvdelta, grandchildmodes, subshamap = import_git_tree(
253
texts, mapping, child_path, child_hexsha, base_inv,
254
base_children.get(basename), file_id, revision_id,
255
parent_invs, shagitmap, lookup_object,
256
allow_submodules=allow_submodules)
257
invdelta.extend(subinvdelta)
258
child_modes.update(grandchildmodes)
259
shamap.extend(subshamap)
260
elif S_ISGITLINK(mode): # submodule
242
for child_mode, name, child_hexsha in tree.entries():
243
existing_children.add(name)
244
child_path = posixpath.join(path, name)
245
if type(base_tree) is Tree:
246
try:
247
child_base_mode, child_base_hexsha = base_tree[name]
248
except KeyError:
249
child_base_hexsha = None
250
child_base_mode = 0
251
else:
252
child_base_hexsha = None
253
child_base_mode = 0
254
if stat.S_ISDIR(child_mode):
255
subinvdelta, grandchildmodes = import_git_tree(texts, mapping,
256
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
257
file_id, revision_id, parent_invs, lookup_object,
258
(child_base_mode, child_mode), store_updater, lookup_file_id,
259
allow_submodules=allow_submodules)
260
elif S_ISGITLINK(child_mode): # submodule
261
261
if not allow_submodules:
262
262
raise SubmodulesRequireSubtrees()
263
subinvdelta, grandchildmodes, subshamap = import_git_submodule(
264
texts, mapping, child_path, child_hexsha, base_inv, base_children.get(basename),
265
file_id, revision_id, parent_invs, shagitmap, lookup_object)
266
invdelta.extend(subinvdelta)
267
child_modes.update(grandchildmodes)
268
shamap.extend(subshamap)
263
subinvdelta, grandchildmodes = import_git_submodule(texts, mapping,
264
child_path, name, (child_base_hexsha, child_hexsha), base_inv,
265
file_id, revision_id, parent_invs, lookup_object,
266
(child_base_mode, child_mode), store_updater, lookup_file_id)
269
267
else:
270
subinvdelta, subshamap = import_git_blob(texts, mapping,
271
child_path, child_hexsha, base_inv, base_children.get(basename), file_id,
272
revision_id, parent_invs, shagitmap, lookup_object,
273
mode_is_executable(mode), stat.S_ISLNK(mode))
274
invdelta.extend(subinvdelta)
275
shamap.extend(subshamap)
276
if mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
268
subinvdelta = import_git_blob(texts, mapping, child_path, name,
269
(child_base_hexsha, child_hexsha), base_inv, file_id,
270
revision_id, parent_invs, lookup_object,
271
(child_base_mode, child_mode), store_updater, lookup_file_id)
272
grandchildmodes = {}
273
child_modes.update(grandchildmodes)
274
invdelta.extend(subinvdelta)
275
if child_mode not in (stat.S_IFDIR, DEFAULT_FILE_MODE,
277
276
stat.S_IFLNK, DEFAULT_FILE_MODE|0111):
278
child_modes[child_path] = mode
277
child_modes[child_path] = child_mode
279
278
# Remove any children that have disappeared
280
if base_ie is not None and base_ie.kind == "directory":
281
invdelta.extend(remove_disappeared_children(base_inv.id2path(file_id),
282
base_children, existing_children))
283
shamap.append((hexsha, "tree", (file_id, revision_id)))
284
return invdelta, child_modes, shamap
285
286
287
def approx_inv_size(inv):
288
# Very rough estimate, 1k per inventory entry
289
return len(inv) * 1024
279
if base_tree is not None and type(base_tree) is Tree:
280
invdelta.extend(remove_disappeared_children(base_inv, old_path,
281
base_tree, existing_children, lookup_object))
282
store_updater.add_object(tree, ie)
283
return invdelta, child_modes
284
285
286
def verify_commit_reconstruction(target_git_object_retriever, lookup_object,
287
o, rev, ret_tree, parent_trees, mapping, unusual_modes):
288
new_unusual_modes = mapping.export_unusual_file_modes(rev)
289
if new_unusual_modes != unusual_modes:
290
raise AssertionError("unusual modes don't match: %r != %r" % (
291
unusual_modes, new_unusual_modes))
292
# Verify that we can reconstruct the commit properly
293
rec_o = target_git_object_retriever._reconstruct_commit(rev, o.tree)
294
if rec_o != o:
295
raise AssertionError("Reconstructed commit differs: %r != %r" % (
296
rec_o, o))
297
diff = []
298
new_objs = {}
299
for path, obj, ie in _tree_to_objects(ret_tree, parent_trees,
300
target_git_object_retriever._cache.idmap, unusual_modes, mapping.BZR_DUMMY_FILE):
301
old_obj_id = tree_lookup_path(lookup_object, o.tree, path)[1]
302
new_objs[path] = obj
303
if obj.id != old_obj_id:
304
diff.append((path, lookup_object(old_obj_id), obj))
305
for (path, old_obj, new_obj) in diff:
306
while (old_obj.type_name == "tree" and
307
new_obj.type_name == "tree" and
308
sorted(old_obj) == sorted(new_obj)):
309
for name in old_obj:
310
if old_obj[name][0] != new_obj[name][0]:
311
raise AssertionError("Modes for %s differ: %o != %o" %
312
(path, old_obj[name][0], new_obj[name][0]))
313
if old_obj[name][1] != new_obj[name][1]:
314
# Found a differing child, delve deeper
315
path = posixpath.join(path, name)
316
old_obj = lookup_object(old_obj[name][1])
317
new_obj = new_objs[path]
318
break
319
raise AssertionError("objects differ for %s: %r != %r" % (path,
320
old_obj, new_obj))
290
321
291
322
292
323
def import_git_commit(repo, mapping, head, lookup_object,
293
target_git_object_retriever, parent_invs_cache):
324
target_git_object_retriever, trees_cache):
294
325
o = lookup_object(head)
295
rev = mapping.import_commit(o)
326
rev = mapping.import_commit(o,
327
lambda x: target_git_object_retriever.lookup_git_sha(x)[1][0])
296
328
# We have to do this here, since we have to walk the tree and
297
329
# we need to make sure to import the blobs / trees with the right
298
330
# path; this may involve adding them more than once.
299
parent_invs = []
300
for parent_id in rev.parent_ids:
301
try:
302
parent_invs.append(parent_invs_cache[parent_id])
303
except KeyError:
304
parent_inv = repo.get_inventory(parent_id)
305
parent_invs.append(parent_inv)
306
parent_invs_cache[parent_id] = parent_inv
307
if parent_invs == []:
331
parent_trees = trees_cache.revision_trees(rev.parent_ids)
332
if parent_trees == []:
308
333
base_inv = Inventory(root_id=None)
309
base_ie = None
334
base_tree = None
335
base_mode = None
310
336
else:
311
base_inv = parent_invs[0]
312
base_ie = base_inv.root
313
inv_delta, unusual_modes, shamap = import_git_tree(repo.texts,
314
mapping, "", o.tree, base_inv, base_ie, None, rev.revision_id,
315
parent_invs, target_git_object_retriever._idmap, lookup_object,
337
base_inv = parent_trees[0].inventory
338
base_tree = lookup_object(o.parents[0]).tree
339
base_mode = stat.S_IFDIR
340
store_updater = target_git_object_retriever._get_updater(rev)
341
store_updater.add_object(o, None)
342
lookup_file_id = mapping.get_fileid_map(lookup_object, o.tree).lookup_file_id
343
inv_delta, unusual_modes = import_git_tree(repo.texts,
344
mapping, "", u"", (base_tree, o.tree), base_inv,
345
None, rev.revision_id, [p.inventory for p in parent_trees],
346
lookup_object, (base_mode, stat.S_IFDIR), store_updater,
347
lookup_file_id,
316
348
allow_submodules=getattr(repo._format, "supports_tree_reference", False))
317
target_git_object_retriever._idmap.add_entries(shamap)
349
store_updater.finish()
318
350
if unusual_modes != {}:
319
351
for path, mode in unusual_modes.iteritems():
320
352
warn_unusual_mode(rev.foreign_revid, path, mode)
327
359
rev.inventory_sha1, inv = repo.add_inventory_by_delta(basis_id,
328
360
inv_delta, rev.revision_id, rev.parent_ids,
329
361
base_inv)
330
parent_invs_cache[rev.revision_id] = inv
362
ret_tree = RevisionTree(repo, inv, rev.revision_id)
363
trees_cache.add(ret_tree)
331
364
repo.add_revision(rev.revision_id, rev)
332
365
if "verify" in debug.debug_flags:
333
new_unusual_modes = mapping.export_unusual_file_modes(rev)
334
if new_unusual_modes != unusual_modes:
335
raise AssertionError("unusual modes don't match: %r != %r" % (unusual_modes, new_unusual_modes))
336
objs = inventory_to_tree_and_blobs(inv, repo.texts, mapping, unusual_modes)
337
for sha1, newobj, path in objs:
338
assert path is not None
339
oldobj = tree_lookup_path(lookup_object, o.tree, path)
340
if oldobj != newobj:
341
raise AssertionError("%r != %r in %s" % (oldobj, newobj, path))
342
343
344
def import_git_objects(repo, mapping, object_iter, target_git_object_retriever,
345
heads, pb=None):
366
verify_commit_reconstruction(target_git_object_retriever,
367
lookup_object, o, rev, ret_tree, parent_trees, mapping,
368
unusual_modes)
369
370
371
def import_git_objects(repo, mapping, object_iter,
372
target_git_object_retriever, heads, pb=None, limit=None):
346
373
"""Import a set of git objects into a bzr repository.
347
374
348
375
:param repo: Target Bazaar repository
349
376
:param mapping: Mapping to use
350
377
:param object_iter: Iterator over Git objects.
378
:return: Tuple with pack hints and last imported revision id
351
379
"""
352
target_git_object_retriever._idmap.start_write_group() # FIXME: try/finally
353
380
def lookup_object(sha):
354
381
try:
355
382
return object_iter[sha]
356
383
except KeyError:
357
384
return target_git_object_retriever[sha]
358
# TODO: a more (memory-)efficient implementation of this
359
385
graph = []
360
386
checked = set()
361
heads = list(heads)
362
parent_invs_cache = lru_cache.LRUSizeCache(compute_size=approx_inv_size,
363
max_size=MAX_INV_CACHE_SIZE)
387
heads = list(set(heads))
388
trees_cache = LRUTreeCache(repo)
364
389
# Find and convert commit objects
365
390
while heads:
366
391
if pb is not None:
370
395
try:
371
396
o = lookup_object(head)
372
397
except KeyError:
373
trace.mutter('missing head %s', head)
374
398
continue
375
399
if isinstance(o, Commit):
376
rev = mapping.import_commit(o)
400
rev = mapping.import_commit(o, lambda x: None)
377
401
if repo.has_revision(rev.revision_id):
378
402
continue
379
squash_revision(repo, rev)
380
403
graph.append((o.id, o.parents))
381
target_git_object_retriever._idmap.add_entry(o.id, "commit",
382
(rev.revision_id, o.tree))
383
404
heads.extend([p for p in o.parents if p not in checked])
384
405
elif isinstance(o, Tag):
385
heads.append(o.object[1])
406
if o.object[1] not in checked:
407
heads.append(o.object[1])
386
408
else:
387
409
trace.warning("Unable to import head object %r" % o)
388
410
checked.add(o.id)
389
411
del checked
390
412
# Order the revisions
391
413
# Create the inventory objects
392
batch_size = 100
414
batch_size = 1000
393
415
revision_ids = topo_sort(graph)
394
416
pack_hints = []
417
if limit is not None:
418
revision_ids = revision_ids[:limit]
419
last_imported = None
395
420
for offset in range(0, len(revision_ids), batch_size):
396
repo.start_write_group()
421
target_git_object_retriever.start_write_group()
397
422
try:
398
for i, head in enumerate(revision_ids[offset:offset+batch_size]):
399
if pb is not None:
400
pb.update("fetching revisions", offset+i, len(revision_ids))
401
import_git_commit(repo, mapping, head, lookup_object,
402
target_git_object_retriever,
403
parent_invs_cache)
423
repo.start_write_group()
424
try:
425
for i, head in enumerate(
426
revision_ids[offset:offset+batch_size]):
427
if pb is not None:
428
pb.update("fetching revisions", offset+i,
429
len(revision_ids))
430
import_git_commit(repo, mapping, head, lookup_object,
431
target_git_object_retriever, trees_cache)
432
last_imported = head
433
except:
434
repo.abort_write_group()
435
raise
436
else:
437
hint = repo.commit_write_group()
438
if hint is not None:
439
pack_hints.extend(hint)
404
440
except:
405
repo.abort_write_group()
441
target_git_object_retriever.abort_write_group()
406
442
raise
407
443
else:
408
hint = repo.commit_write_group()
409
if hint is not None:
410
pack_hints.extend(hint)
411
target_git_object_retriever._idmap.commit_write_group()
412
return pack_hints
444
target_git_object_retriever.commit_write_group()
445
return pack_hints, last_imported
413
446
414
447
415
448
class InterGitRepository(InterRepository):
424
457
"""See InterRepository.copy_content."""
425
458
self.fetch(revision_id, pb, find_ghosts=False)
426
459
427
def fetch(self, revision_id=None, pb=None, find_ghosts=False, mapping=None,
428
fetch_spec=None):
429
self.fetch_refs(revision_id=revision_id, pb=pb, find_ghosts=find_ghosts,
430
mapping=mapping, fetch_spec=fetch_spec)
460
def fetch(self, revision_id=None, pb=None, find_ghosts=False,
461
mapping=None, fetch_spec=None):
462
self.fetch_refs(revision_id=revision_id, pb=pb,
463
find_ghosts=find_ghosts, mapping=mapping, fetch_spec=fetch_spec)
431
464
432
465
433
466
class InterGitNonGitRepository(InterGitRepository):
450
483
if interesting_heads is None:
451
484
ret = [sha for (ref, sha) in refs.iteritems() if not ref.endswith("^{}")]
452
485
else:
453
ret = [mapping.revision_id_bzr_to_foreign(revid)[0] for revid in interesting_heads if revid not in (None, NULL_REVISION)]
454
return [rev for rev in ret if not self.target.has_revision(mapping.revision_id_foreign_to_bzr(rev))]
455
pack_hint = self.fetch_objects(determine_wants, mapping, pb)
486
ret = [self.source.lookup_bzr_revision_id(revid)[0] for revid in interesting_heads if revid not in (None, NULL_REVISION)]
487
return [rev for rev in ret if not self.target.has_revision(self.source.lookup_foreign_revision_id(rev))]
488
(pack_hint, _) = self.fetch_objects(determine_wants, mapping, pb)
456
489
if pack_hint is not None and self.target._format.pack_compresses:
457
490
self.target.pack(hint=pack_hint)
458
if interesting_heads is not None:
459
present_interesting_heads = self.target.has_revisions(interesting_heads)
460
missing_interesting_heads = set(interesting_heads) - present_interesting_heads
461
if missing_interesting_heads:
462
raise AssertionError("Missing interesting heads: %r" % missing_interesting_heads)
463
491
return self._refs
464
492
465
493
486
514
map(all_parents.update, parent_map.itervalues())
487
515
return set(all_revs) - all_parents
488
516
489
def fetch_objects(self, determine_wants, mapping, pb=None):
517
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
490
518
def progress(text):
491
519
report_git_progress(pb, text)
492
520
store = BazaarObjectStore(self.target, mapping)
506
534
if pb is None:
507
535
create_pb = pb = ui.ui_factory.nested_progress_bar()
508
536
try:
509
objects_iter = self.source.fetch_objects(
510
record_determine_wants, graph_walker,
511
store.get_raw, progress)
537
objects_iter = self.source.fetch_objects(record_determine_wants,
538
graph_walker, store.get_raw, progress)
512
539
return import_git_objects(self.target, mapping,
513
objects_iter, store, recorded_wants, pb)
540
objects_iter, store, recorded_wants, pb, limit)
514
541
finally:
515
542
if create_pb:
516
543
create_pb.finished()
520
547
@staticmethod
521
548
def is_compatible(source, target):
522
549
"""Be compatible with GitRepository."""
523
# FIXME: Also check target uses VersionedFile
524
550
return (isinstance(source, RemoteGitRepository) and
525
551
target.supports_rich_root() and
526
not isinstance(target, GitRepository))
552
not isinstance(target, GitRepository) and
553
target.texts is not None)
527
554
528
555
529
556
class InterLocalGitNonGitRepository(InterGitNonGitRepository):
530
557
"""InterRepository that copies revisions from a local Git into a non-Git
531
558
repository."""
532
559
533
def fetch_objects(self, determine_wants, mapping, pb=None):
560
def fetch_objects(self, determine_wants, mapping, pb=None, limit=None):
561
"""Fetch objects.
562
"""
534
563
wants = determine_wants(self.source._git.get_refs())
535
564
create_pb = None
536
565
if pb is None:
540
569
self.target.lock_write()
541
570
try:
542
571
return import_git_objects(self.target, mapping,
543
self.source._git.object_store, target_git_object_retriever,
544
wants, pb)
572
self.source._git.object_store,
573
target_git_object_retriever, wants, pb, limit)
545
574
finally:
546
575
self.target.unlock()
547
576
finally:
551
580
@staticmethod
552
581
def is_compatible(source, target):
553
582
"""Be compatible with GitRepository."""
554
# FIXME: Also check target uses VersionedFile
555
583
return (isinstance(source, LocalGitRepository) and
556
584
target.supports_rich_root() and
557
not isinstance(target, GitRepository))
585
not isinstance(target, GitRepository) and
586
target.texts is not None)
558
587
559
588
560
589
class InterGitGitRepository(InterGitRepository):
600
629
determine_wants = r.object_store.determine_wants_all
601
630
else:
602
631
determine_wants = lambda x: [y for y in args if not y in r.object_store]
603
return self.fetch_objects(determine_wants, mapping)
632
return self.fetch_objects(determine_wants, mapping)[0]
604
633
605
634
606
635
@staticmethod
Loggerhead is a web-based interface for Breezy
Version: 2.0.1