1
# Copyright (C) 2005 Canonical Ltd
1
# Copyright (C) 2005, 2006 Canonical Ltd
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
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
13
13
# You should have received a copy of the GNU General Public License
14
14
# along with this program; if not, write to the Free Software
15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
19
def commit(branch, message,
26
"""Commit working copy as a new revision.
28
The basic approach is to add all the file texts into the
29
store, then the inventory, then make a new revision pointing
30
to that inventory and store that.
32
This is not quite safe if the working copy changes during the
33
commit; for the moment that is simply not allowed. A better
34
approach is to make a temporary copy of the files before
35
computing their hashes, and then add those hashes in turn to
36
the inventory. This should mean at least that there are no
37
broken hash pointers. There is no way we can get a snapshot
38
of the whole directory at an instant. This would also have to
39
be robust against files disappearing, moving, etc. So the
40
whole thing is a bit hard.
42
timestamp -- if not None, seconds-since-epoch for a
43
postdated/predated commit.
46
If true, commit only those files.
49
If set, use this as the new revision id.
50
Useful for test or import commands that need to tightly
51
control what revisions are assigned. If you duplicate
52
a revision id that exists elsewhere it is your own fault.
53
If null (default), a time/random revision id is generated.
18
# The newly committed revision is going to have a shape corresponding
19
# to that of the working inventory. Files that are not in the
20
# working tree and that were in the predecessor are reported as
21
# removed --- this can include files that were either removed from the
22
# inventory or deleted in the working tree. If they were only
23
# deleted from disk, they are removed from the working inventory.
25
# We then consider the remaining entries, which will be in the new
26
# version. Directory entries are simply copied across. File entries
27
# must be checked to see if a new version of the file should be
28
# recorded. For each parent revision inventory, we check to see what
29
# version of the file was present. If the file was present in at
30
# least one tree, and if it was the same version in all the trees,
31
# then we can just refer to that version. Otherwise, a new version
32
# representing the merger of the file versions must be added.
34
# TODO: Update hashcache before and after - or does the WorkingTree
37
# TODO: Rather than mashing together the ancestry and storing it back,
38
# perhaps the weave should have single method which does it all in one
39
# go, avoiding a lot of redundant work.
41
# TODO: Perhaps give a warning if one of the revisions marked as
42
# merged is already in the ancestry, and then don't record it as a
45
# TODO: If the file is newly merged but unchanged from the version it
46
# merges from, then it should still be reported as newly added
47
# relative to the basis revision.
49
# TODO: Change the parameter 'rev_id' to 'revision_id' to be consistent with
50
# the rest of the code; add a deprecation of the old name.
57
from cStringIO import StringIO
65
from bzrlib.errors import (BzrError, PointlessCommit,
69
from bzrlib.osutils import (kind_marker, isdir,isfile, is_inside_any,
70
is_inside_or_parent_of_any,
71
quotefn, sha_file, split_lines)
72
from bzrlib.testament import Testament
73
from bzrlib.trace import mutter, note, warning
74
from bzrlib.xml5 import serializer_v5
75
from bzrlib.inventory import Inventory, InventoryEntry
76
from bzrlib import symbol_versioning
77
from bzrlib.symbol_versioning import (deprecated_passed,
80
from bzrlib.workingtree import WorkingTree
84
class NullCommitReporter(object):
85
"""I report on progress of a commit."""
87
def snapshot_change(self, change, path):
90
def completed(self, revno, rev_id):
93
def deleted(self, file_id):
96
def escaped(self, escape_count, message):
99
def missing(self, path):
102
def renamed(self, change, old_path, new_path):
106
class ReportCommitToLog(NullCommitReporter):
108
# this may be more useful if 'note' was replaced by an overridable
109
# method on self, which would allow more trivial subclassing.
110
# alternative, a callable could be passed in, allowing really trivial
111
# reuse for some uis. RBC 20060511
113
def snapshot_change(self, change, path):
114
if change == 'unchanged':
116
if change == 'added' and path == '':
118
note("%s %s", change, path)
120
def completed(self, revno, rev_id):
121
note('Committed revision %d.', revno)
123
def deleted(self, file_id):
124
note('deleted %s', file_id)
126
def escaped(self, escape_count, message):
127
note("replaced %d control characters in message", escape_count)
129
def missing(self, path):
130
note('missing %s', path)
132
def renamed(self, change, old_path, new_path):
133
note('%s %s => %s', change, old_path, new_path)
136
class Commit(object):
137
"""Task of committing a new revision.
139
This is a MethodObject: it accumulates state as the commit is
140
prepared, and then it is discarded. It doesn't represent
141
historical revisions, just the act of recording a new one.
144
Modified to hold a list of files that have been deleted from
145
the working directory; these should be removed from the
56
import os, time, tempfile
58
from inventory import Inventory
59
from osutils import isdir, isfile, sha_string, quotefn, \
60
local_time_offset, username, kind_marker, is_inside_any
151
if reporter is not None:
152
self.reporter = reporter
154
self.reporter = NullCommitReporter()
155
if config is not None:
161
branch=DEPRECATED_PARAMETER, message=None,
167
allow_pointless=True,
175
message_callback=None):
176
"""Commit working copy as a new revision.
178
branch -- the deprecated branch to commit to. New callers should pass in
181
message -- the commit message (it or message_callback is required)
183
timestamp -- if not None, seconds-since-epoch for a
184
postdated/predated commit.
186
specific_files -- If true, commit only those files.
188
rev_id -- If set, use this as the new revision id.
189
Useful for test or import commands that need to tightly
190
control what revisions are assigned. If you duplicate
191
a revision id that exists elsewhere it is your own fault.
192
If null (default), a time/random revision id is generated.
194
allow_pointless -- If true (default), commit even if nothing
195
has changed and no merges are recorded.
197
strict -- If true, don't allow a commit if the working tree
198
contains unknown files.
200
revprops -- Properties for new revision
201
:param local: Perform a local only commit.
203
mutter('preparing to commit')
205
if deprecated_passed(branch):
206
symbol_versioning.warn("Commit.commit (branch, ...): The branch parameter is "
207
"deprecated as of bzr 0.8. Please use working_tree= instead.",
208
DeprecationWarning, stacklevel=2)
210
self.work_tree = self.branch.bzrdir.open_workingtree()
211
elif working_tree is None:
212
raise BzrError("One of branch and working_tree must be passed into commit().")
214
self.work_tree = working_tree
215
self.branch = self.work_tree.branch
216
if message_callback is None:
217
if message is not None:
218
if isinstance(message, str):
219
message = message.decode(bzrlib.user_encoding)
220
message_callback = lambda x: message
222
raise BzrError("The message or message_callback keyword"
223
" parameter is required for commit().")
225
self.bound_branch = None
227
self.master_branch = None
228
self.master_locked = False
230
self.specific_files = specific_files
231
self.allow_pointless = allow_pointless
233
if reporter is None and self.reporter is None:
234
self.reporter = NullCommitReporter()
235
elif reporter is not None:
236
self.reporter = reporter
238
self.work_tree.lock_write()
239
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
241
# Cannot commit with conflicts present.
242
if len(self.work_tree.conflicts())>0:
243
raise ConflictsInTree
245
# setup the bound branch variables as needed.
246
self._check_bound_branch()
248
# check for out of date working trees
250
first_tree_parent = self.work_tree.get_parent_ids()[0]
252
# if there are no parents, treat our parent as 'None'
253
# this is so that we still consier the master branch
254
# - in a checkout scenario the tree may have no
255
# parents but the branch may do.
256
first_tree_parent = None
257
master_last = self.master_branch.last_revision()
258
if (master_last is not None and
259
master_last != first_tree_parent):
260
raise errors.OutOfDateTree(self.work_tree)
62
from branch import gen_file_id
63
from errors import BzrError
64
from revision import Revision
65
from trace import mutter, note
67
branch._need_writelock()
69
# First walk over the working inventory; and both update that
70
# and also build a new revision inventory. The revision
71
# inventory needs to hold the text-id, sha1 and size of the
72
# actual file versions committed in the revision. (These are
73
# not present in the working inventory.) We also need to
74
# detect missing/deleted files, and remove them from the
77
work_tree = branch.working_tree()
78
work_inv = work_tree.inventory
80
basis = branch.basis_tree()
81
basis_inv = basis.inventory
85
note('looking for changes...')
87
for path, entry in work_inv.iter_entries():
88
## TODO: Check that the file kind has not changed from the previous
89
## revision of this file (if any).
93
p = branch.abspath(path)
94
file_id = entry.file_id
95
mutter('commit prep file %s, id %r ' % (p, file_id))
97
if specific_files and not is_inside_any(specific_files, path):
98
if basis_inv.has_id(file_id):
99
# carry over with previous state
100
inv.add(basis_inv[file_id].copy())
263
# raise an exception as soon as we find a single unknown.
264
for unknown in self.work_tree.unknowns():
265
raise StrictCommitFailed()
267
if self.config is None:
268
self.config = self.branch.get_config()
270
self.work_inv = self.work_tree.inventory
271
self.basis_tree = self.work_tree.basis_tree()
272
self.basis_inv = self.basis_tree.inventory
273
if specific_files is not None:
274
# Ensure specified files are versioned
275
# (We don't actually need the ids here)
276
tree.find_ids_across_trees(specific_files,
277
[self.basis_tree, self.work_tree])
278
# one to finish, one for rev and inventory, and one for each
279
# inventory entry, and the same for the new inventory.
280
# note that this estimate is too long when we do a partial tree
281
# commit which excludes some new files from being considered.
282
# The estimate is corrected when we populate the new inv.
283
self.pb_total = len(self.work_inv) + 5
286
self._gather_parents()
287
if len(self.parents) > 1 and self.specific_files:
288
raise NotImplementedError('selected-file commit of merges is not supported yet: files %r',
291
self.builder = self.branch.get_commit_builder(self.parents,
292
self.config, timestamp, timezone, committer, revprops, rev_id)
294
self._remove_deleted()
295
self._populate_new_inv()
296
self._report_deletes()
298
self._check_pointless()
300
self._emit_progress_update()
301
# TODO: Now the new inventory is known, check for conflicts and
302
# prompt the user for a commit message.
303
# ADHB 2006-08-08: If this is done, populate_new_inv should not add
304
# weave lines, because nothing should be recorded until it is known
305
# that commit will succeed.
306
self.builder.finish_inventory()
307
self._emit_progress_update()
308
message = message_callback(self)
309
assert isinstance(message, unicode), type(message)
310
self.message = message
311
self._escape_commit_message()
313
self.rev_id = self.builder.commit(self.message)
314
self._emit_progress_update()
315
# revision data is in the local branch now.
317
# upload revision data to the master.
318
# this will propagate merged revisions too if needed.
319
if self.bound_branch:
320
self.master_branch.repository.fetch(self.branch.repository,
321
revision_id=self.rev_id)
322
# now the master has the revision data
323
# 'commit' to the master first so a timeout here causes the local
324
# branch to be out of date
325
self.master_branch.append_revision(self.rev_id)
327
# and now do the commit locally.
328
self.branch.append_revision(self.rev_id)
330
rev_tree = self.builder.revision_tree()
331
self.work_tree.set_parent_trees([(self.rev_id, rev_tree)])
332
# now the work tree is up to date with the branch
334
self.reporter.completed(self.branch.revno(), self.rev_id)
335
if self.config.post_commit() is not None:
336
hooks = self.config.post_commit().split(' ')
337
# this would be nicer with twisted.python.reflect.namedAny
339
result = eval(hook + '(branch, rev_id)',
340
{'branch':self.branch,
342
'rev_id':self.rev_id})
343
self._emit_progress_update()
348
def _any_real_changes(self):
349
"""Are there real changes between new_inventory and basis?
351
For trees without rich roots, inv.root.revision changes every commit.
352
But if that is the only change, we want to treat it as though there
355
new_entries = self.builder.new_inventory.iter_entries()
356
basis_entries = self.basis_inv.iter_entries()
357
new_path, new_root_ie = new_entries.next()
358
basis_path, basis_root_ie = basis_entries.next()
360
# This is a copy of InventoryEntry.__eq__ only leaving out .revision
361
def ie_equal_no_revision(this, other):
362
return ((this.file_id == other.file_id)
363
and (this.name == other.name)
364
and (this.symlink_target == other.symlink_target)
365
and (this.text_sha1 == other.text_sha1)
366
and (this.text_size == other.text_size)
367
and (this.text_id == other.text_id)
368
and (this.parent_id == other.parent_id)
369
and (this.kind == other.kind)
370
and (this.executable == other.executable)
372
if not ie_equal_no_revision(new_root_ie, basis_root_ie):
375
for new_ie, basis_ie in zip(new_entries, basis_entries):
376
if new_ie != basis_ie:
379
# No actual changes present
382
def _check_pointless(self):
383
if self.allow_pointless:
385
# A merge with no effect on files
386
if len(self.parents) > 1:
388
# work around the fact that a newly-initted tree does differ from its
390
if len(self.basis_inv) == 0 and len(self.builder.new_inventory) == 1:
391
raise PointlessCommit()
392
# Shortcut, if the number of entries changes, then we obviously have
394
if len(self.builder.new_inventory) != len(self.basis_inv):
396
# If length == 1, then we only have the root entry. Which means
397
# that there is no real difference (only the root could be different)
398
if (len(self.builder.new_inventory) != 1 and self._any_real_changes()):
400
raise PointlessCommit()
402
def _check_bound_branch(self):
403
"""Check to see if the local branch is bound.
405
If it is bound, then most of the commit will actually be
406
done using the remote branch as the target branch.
407
Only at the end will the local branch be updated.
409
if self.local and not self.branch.get_bound_location():
410
raise errors.LocalRequiresBoundBranch()
413
self.master_branch = self.branch.get_master_branch()
415
if not self.master_branch:
416
# make this branch the reference branch for out of date checks.
417
self.master_branch = self.branch
420
# If the master branch is bound, we must fail
421
master_bound_location = self.master_branch.get_bound_location()
422
if master_bound_location:
423
raise errors.CommitToDoubleBoundBranch(self.branch,
424
self.master_branch, master_bound_location)
426
# TODO: jam 20051230 We could automatically push local
427
# commits to the remote branch if they would fit.
428
# But for now, just require remote to be identical
431
# Make sure the local branch is identical to the master
432
master_info = self.master_branch.last_revision_info()
433
local_info = self.branch.last_revision_info()
434
if local_info != master_info:
435
raise errors.BoundBranchOutOfDate(self.branch,
438
# Now things are ready to change the master branch
440
self.bound_branch = self.branch
441
self.master_branch.lock_write()
442
self.master_locked = True
445
"""Cleanup any open locks, progress bars etc."""
446
cleanups = [self._cleanup_bound_branch,
447
self.work_tree.unlock,
449
found_exception = None
450
for cleanup in cleanups:
453
# we want every cleanup to run no matter what.
454
# so we have a catchall here, but we will raise the
455
# last encountered exception up the stack: and
456
# typically this will be useful enough.
459
if found_exception is not None:
460
# don't do a plan raise, because the last exception may have been
461
# trashed, e is our sure-to-work exception even though it loses the
462
# full traceback. XXX: RBC 20060421 perhaps we could check the
463
# exc_info and if its the same one do a plain raise otherwise
464
# 'raise e' as we do now.
467
def _cleanup_bound_branch(self):
468
"""Executed at the end of a try/finally to cleanup a bound branch.
470
If the branch wasn't bound, this is a no-op.
471
If it was, it resents self.branch to the local branch, instead
474
if not self.bound_branch:
476
if self.master_locked:
477
self.master_branch.unlock()
479
def _escape_commit_message(self):
480
"""Replace xml-incompatible control characters."""
481
# FIXME: RBC 20060419 this should be done by the revision
482
# serialiser not by commit. Then we can also add an unescaper
483
# in the deserializer and start roundtripping revision messages
484
# precisely. See repository_implementations/test_repository.py
486
# Python strings can include characters that can't be
487
# represented in well-formed XML; escape characters that
488
# aren't listed in the XML specification
489
# (http://www.w3.org/TR/REC-xml/#NT-Char).
490
self.message, escape_count = re.subn(
491
u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
492
lambda match: match.group(0).encode('unicode_escape'),
495
self.reporter.escaped(escape_count, self.message)
497
def _gather_parents(self):
498
"""Record the parents of a merge for merge detection."""
499
# TODO: Make sure that this list doesn't contain duplicate
500
# entries and the order is preserved when doing this.
501
self.parents = self.work_tree.get_parent_ids()
502
self.parent_invs = []
503
for revision in self.parents:
504
if self.branch.repository.has_revision(revision):
505
mutter('commit parent revision {%s}', revision)
506
inventory = self.branch.repository.get_inventory(revision)
507
self.parent_invs.append(inventory)
102
# omit this from committed inventory
509
mutter('commit parent ghost revision {%s}', revision)
511
def _remove_deleted(self):
512
"""Remove deleted files from the working inventories.
514
This is done prior to taking the working inventory as the
515
basis for the new committed inventory.
517
This returns true if any files
518
*that existed in the basis inventory* were deleted.
519
Files that were added and deleted
520
in the working copy don't matter.
522
specific = self.specific_files
524
deleted_paths = set()
525
for path, ie in self.work_inv.iter_entries():
526
if is_inside_any(deleted_paths, path):
527
# The tree will delete the required ids recursively.
529
if specific and not is_inside_any(specific, path):
531
if not self.work_tree.has_filename(path):
532
deleted_paths.add(path)
533
self.reporter.missing(path)
534
deleted_ids.append(ie.file_id)
535
self.work_tree.unversion(deleted_ids)
537
def _populate_new_inv(self):
538
"""Build revision inventory.
540
This creates a new empty inventory. Depending on
541
which files are selected for commit, and what is present in the
542
current tree, the new inventory is populated. inventory entries
543
which are candidates for modification have their revision set to
544
None; inventory entries that are carried over untouched have their
545
revision set to their prior value.
547
# ESEPARATIONOFCONCERNS: this function is diffing and using the diff
548
# results to create a new inventory at the same time, which results
549
# in bugs like #46635. Any reason not to use/enhance Tree.changes_from?
551
mutter("Selecting files for commit with filter %s", self.specific_files)
552
assert self.work_inv.root is not None
553
entries = self.work_inv.iter_entries()
554
if not self.builder.record_root_entry:
555
symbol_versioning.warn('CommitBuilders should support recording'
556
' the root entry as of bzr 0.10.', DeprecationWarning,
558
self.builder.new_inventory.add(self.basis_inv.root.copy())
560
self._emit_progress_update()
561
for path, new_ie in entries:
562
self._emit_progress_update()
563
file_id = new_ie.file_id
565
kind = self.work_tree.kind(file_id)
566
if kind != new_ie.kind:
567
new_ie = inventory.make_entry(kind, new_ie.name,
568
new_ie.parent_id, file_id)
569
except errors.NoSuchFile:
106
if not work_tree.has_id(file_id):
108
print('deleted %s%s' % (path, kind_marker(entry.kind)))
109
mutter(" file is missing, removing from inventory")
110
missing_ids.append(file_id)
115
if basis_inv.has_id(file_id):
116
old_kind = basis_inv[file_id].kind
117
if old_kind != entry.kind:
118
raise BzrError("entry %r changed kind from %r to %r"
119
% (file_id, old_kind, entry.kind))
121
if entry.kind == 'directory':
123
raise BzrError("%s is entered as directory but not a directory"
125
elif entry.kind == 'file':
127
raise BzrError("%s is entered as file but is not a file" % quotefn(p))
129
new_sha1 = work_tree.get_file_sha1(file_id)
131
old_ie = basis_inv.has_id(file_id) and basis_inv[file_id]
133
and old_ie.text_sha1 == new_sha1):
134
## assert content == basis.get_file(file_id).read()
135
entry.text_id = old_ie.text_id
136
entry.text_sha1 = new_sha1
137
entry.text_size = old_ie.text_size
138
mutter(' unchanged from previous text_id {%s}' %
141
content = file(p, 'rb').read()
143
# calculate the sha again, just in case the file contents
144
# changed since we updated the cache
145
entry.text_sha1 = sha_string(content)
146
entry.text_size = len(content)
148
entry.text_id = gen_file_id(entry.name)
149
branch.text_store.add(content, entry.text_id)
150
mutter(' stored with text_id {%s}' % entry.text_id)
153
print('added %s' % path)
154
elif (old_ie.name == entry.name
155
and old_ie.parent_id == entry.parent_id):
156
print('modified %s' % path)
158
print('renamed %s' % path)
161
for file_id in missing_ids:
162
# Any files that have been deleted are now removed from the
163
# working inventory. Files that were not selected for commit
164
# are left as they were in the working inventory and ommitted
165
# from the revision inventory.
167
# have to do this later so we don't mess up the iterator.
168
# since parents may be removed before their children we
171
# FIXME: There's probably a better way to do this; perhaps
172
# the workingtree should know how to filter itbranch.
173
if work_inv.has_id(file_id):
174
del work_inv[file_id]
178
rev_id = _gen_revision_id(time.time())
181
inv_tmp = tempfile.TemporaryFile()
182
inv.write_xml(inv_tmp)
184
branch.inventory_store.add(inv_tmp, inv_id)
185
mutter('new inventory_id is {%s}' % inv_id)
187
branch._write_inventory(work_inv)
189
if timestamp == None:
190
timestamp = time.time()
192
if committer == None:
193
committer = username()
196
timezone = local_time_offset()
198
mutter("building commit log message")
199
rev = Revision(timestamp=timestamp,
202
precursor = branch.last_patch(),
207
rev_tmp = tempfile.TemporaryFile()
208
rev.write_xml(rev_tmp)
210
branch.revision_store.add(rev_tmp, rev_id)
211
mutter("new revision_id is {%s}" % rev_id)
213
## XXX: Everything up to here can simply be orphaned if we abort
214
## the commit; it will leave junk files behind but that doesn't
217
## TODO: Read back the just-generated changeset, and make sure it
218
## applies and recreates the right state.
220
## TODO: Also calculate and store the inventory SHA1
221
mutter("committing patch r%d" % (branch.revno() + 1))
223
branch.append_revision(rev_id)
226
note("commited r%d" % branch.revno())
230
def _gen_revision_id(when):
231
"""Return new revision-id."""
232
from binascii import hexlify
233
from osutils import rand_bytes, compact_date, user_email
235
s = '%s-%s-' % (user_email(), compact_date(when))
236
s += hexlify(rand_bytes(8))
571
# mutter('check %s {%s}', path, file_id)
572
if (not self.specific_files or
573
is_inside_or_parent_of_any(self.specific_files, path)):
574
# mutter('%s selected for commit', path)
578
# mutter('%s not selected for commit', path)
579
if self.basis_inv.has_id(file_id):
580
ie = self.basis_inv[file_id].copy()
582
# this entry is new and not being committed
584
self.builder.record_entry_contents(ie, self.parent_invs,
585
path, self.work_tree)
586
# describe the nature of the change that has occurred relative to
587
# the basis inventory.
588
if (self.basis_inv.has_id(ie.file_id)):
589
basis_ie = self.basis_inv[ie.file_id]
592
change = ie.describe_change(basis_ie, ie)
593
if change in (InventoryEntry.RENAMED,
594
InventoryEntry.MODIFIED_AND_RENAMED):
595
old_path = self.basis_inv.id2path(ie.file_id)
596
self.reporter.renamed(change, old_path, path)
598
self.reporter.snapshot_change(change, path)
600
if not self.specific_files:
603
# ignore removals that don't match filespec
604
for path, new_ie in self.basis_inv.iter_entries():
605
if new_ie.file_id in self.work_inv:
607
if is_inside_any(self.specific_files, path):
611
self.builder.record_entry_contents(ie, self.parent_invs, path,
614
def _emit_progress_update(self):
615
"""Emit an update to the progress bar."""
616
self.pb.update("Committing", self.pb_count, self.pb_total)
619
def _report_deletes(self):
620
for path, ie in self.basis_inv.iter_entries():
621
if ie.file_id not in self.builder.new_inventory:
622
self.reporter.deleted(path)
added
removed
bzrlib/commit.py
