1
# Copyright (C) 2005, 2006, 2007 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
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
64
from bzrlib.branch import Branch
66
from bzrlib.errors import (BzrError, PointlessCommit,
70
from bzrlib.osutils import (kind_marker, isdir,isfile, is_inside_any,
71
is_inside_or_parent_of_any,
72
quotefn, sha_file, split_lines)
73
from bzrlib.testament import Testament
74
from bzrlib.trace import mutter, note, warning
75
from bzrlib.xml5 import serializer_v5
76
from bzrlib.inventory import Inventory, InventoryEntry
77
from bzrlib import symbol_versioning
78
from bzrlib.symbol_versioning import (deprecated_passed,
81
from bzrlib.workingtree import WorkingTree
85
class NullCommitReporter(object):
86
"""I report on progress of a commit."""
88
def snapshot_change(self, change, path):
91
def completed(self, revno, rev_id):
94
def deleted(self, file_id):
97
def escaped(self, escape_count, message):
100
def missing(self, path):
103
def renamed(self, change, old_path, new_path):
107
class ReportCommitToLog(NullCommitReporter):
109
def _note(self, format, *args):
112
Subclasses may choose to override this method.
116
def snapshot_change(self, change, path):
117
if change == 'unchanged':
119
if change == 'added' and path == '':
121
self._note("%s %s", change, path)
123
def completed(self, revno, rev_id):
124
self._note('Committed revision %d.', revno)
126
def deleted(self, file_id):
127
self._note('deleted %s', file_id)
129
def escaped(self, escape_count, message):
130
self._note("replaced %d control characters in message", escape_count)
132
def missing(self, path):
133
self._note('missing %s', path)
135
def renamed(self, change, old_path, new_path):
136
self._note('%s %s => %s', change, old_path, new_path)
139
class Commit(object):
140
"""Task of committing a new revision.
142
This is a MethodObject: it accumulates state as the commit is
143
prepared, and then it is discarded. It doesn't represent
144
historical revisions, just the act of recording a new one.
147
Modified to hold a list of files that have been deleted from
148
the working directory; these should be removed from the
154
if reporter is not None:
155
self.reporter = reporter
157
self.reporter = NullCommitReporter()
167
allow_pointless=True,
175
message_callback=None,
177
"""Commit working copy as a new revision.
179
:param message: the commit message (it or message_callback is required)
181
:param timestamp: if not None, seconds-since-epoch for a
182
postdated/predated commit.
184
:param specific_files: If true, commit only those files.
186
:param rev_id: If set, use this as the new revision id.
187
Useful for test or import commands that need to tightly
188
control what revisions are assigned. If you duplicate
189
a revision id that exists elsewhere it is your own fault.
190
If null (default), a time/random revision id is generated.
192
:param allow_pointless: If true (default), commit even if nothing
193
has changed and no merges are recorded.
195
:param strict: If true, don't allow a commit if the working tree
196
contains unknown files.
198
:param revprops: Properties for new revision
199
:param local: Perform a local only commit.
200
:param recursive: If set to 'down', commit in any subtrees that have
201
pending changes of any sort during this commit.
203
mutter('preparing to commit')
205
if working_tree is None:
206
raise BzrError("working_tree must be passed into commit().")
208
self.work_tree = working_tree
209
self.branch = self.work_tree.branch
210
if getattr(self.work_tree, 'requires_rich_root', lambda: False)():
211
if not self.branch.repository.supports_rich_root():
212
raise errors.RootNotRich()
213
if message_callback is None:
214
if message is not None:
215
if isinstance(message, str):
216
message = message.decode(bzrlib.user_encoding)
217
message_callback = lambda x: message
219
raise BzrError("The message or message_callback keyword"
220
" parameter is required for commit().")
222
self.bound_branch = None
224
self.master_branch = None
225
self.master_locked = False
227
self.specific_files = specific_files
228
self.allow_pointless = allow_pointless
229
self.recursive = recursive
230
self.revprops = revprops
231
self.message_callback = message_callback
232
self.timestamp = timestamp
233
self.timezone = timezone
234
self.committer = committer
236
self.verbose = verbose
238
if reporter is None and self.reporter is None:
239
self.reporter = NullCommitReporter()
240
elif reporter is not None:
241
self.reporter = reporter
243
self.work_tree.lock_write()
244
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
245
self.basis_tree = self.work_tree.basis_tree()
246
self.basis_tree.lock_read()
248
# Cannot commit with conflicts present.
249
if len(self.work_tree.conflicts()) > 0:
250
raise ConflictsInTree
252
# Setup the bound branch variables as needed.
253
self._check_bound_branch()
255
# Check that the working tree is up to date
256
old_revno,new_revno = self._check_out_of_date_tree()
259
# raise an exception as soon as we find a single unknown.
260
for unknown in self.work_tree.unknowns():
261
raise StrictCommitFailed()
263
if self.config is None:
264
self.config = self.branch.get_config()
266
self.work_inv = self.work_tree.inventory
267
self.basis_inv = self.basis_tree.inventory
268
if specific_files is not None:
269
# Ensure specified files are versioned
270
# (We don't actually need the ids here)
271
# XXX: Dont we have filter_unversioned to do this more
273
tree.find_ids_across_trees(specific_files,
274
[self.basis_tree, self.work_tree])
276
# Setup the progress bar ...
277
# one to finish, one for rev and inventory, and one for each
278
# inventory entry, and the same for the new inventory.
279
# note that this estimate is too long when we do a partial tree
280
# commit which excludes some new files from being considered.
281
# The estimate is corrected when we populate the new inv.
282
self.pb_total = len(self.work_inv) + 5
285
self._gather_parents()
286
if len(self.parents) > 1 and self.specific_files:
287
raise errors.CannotCommitSelectedFileMerge(self.specific_files)
289
# Build the new inventory
290
self.builder = self.branch.get_commit_builder(self.parents,
291
self.config, timestamp, timezone, committer, revprops, rev_id)
292
self._remove_deleted()
293
self._populate_new_inv()
294
self._report_deletes()
295
self._check_pointless()
296
self._emit_progress_update()
298
# TODO: Now the new inventory is known, check for conflicts and
299
# prompt the user for a commit message.
300
# ADHB 2006-08-08: If this is done, populate_new_inv should not add
301
# weave lines, because nothing should be recorded until it is known
302
# that commit will succeed.
303
self.builder.finish_inventory()
304
self._emit_progress_update()
305
message = message_callback(self)
306
assert isinstance(message, unicode), type(message)
307
self.message = message
308
self._escape_commit_message()
310
# Add revision data to the local branch
311
self.rev_id = self.builder.commit(self.message)
312
self._emit_progress_update()
314
# upload revision data to the master.
315
# this will propagate merged revisions too if needed.
316
if self.bound_branch:
317
self.master_branch.repository.fetch(self.branch.repository,
318
revision_id=self.rev_id)
319
# now the master has the revision data
320
# 'commit' to the master first so a timeout here causes the
321
# local branch to be out of date
322
self.master_branch.set_last_revision_info(new_revno,
325
# and now do the commit locally.
326
self.branch.set_last_revision_info(new_revno, self.rev_id)
328
# Make the working tree up to date with the branch
329
rev_tree = self.builder.revision_tree()
330
self.work_tree.set_parent_trees([(self.rev_id, rev_tree)])
331
self.reporter.completed(new_revno, self.rev_id)
333
# Process the post commit hooks, if any
334
self._process_hooks(old_revno, new_revno)
335
self._emit_progress_update()
340
def _any_real_changes(self):
341
"""Are there real changes between new_inventory and basis?
343
For trees without rich roots, inv.root.revision changes every commit.
344
But if that is the only change, we want to treat it as though there
347
new_entries = self.builder.new_inventory.iter_entries()
348
basis_entries = self.basis_inv.iter_entries()
349
new_path, new_root_ie = new_entries.next()
350
basis_path, basis_root_ie = basis_entries.next()
352
# This is a copy of InventoryEntry.__eq__ only leaving out .revision
353
def ie_equal_no_revision(this, other):
354
return ((this.file_id == other.file_id)
355
and (this.name == other.name)
356
and (this.symlink_target == other.symlink_target)
357
and (this.text_sha1 == other.text_sha1)
358
and (this.text_size == other.text_size)
359
and (this.text_id == other.text_id)
360
and (this.parent_id == other.parent_id)
361
and (this.kind == other.kind)
362
and (this.executable == other.executable)
363
and (this.reference_revision == other.reference_revision)
365
if not ie_equal_no_revision(new_root_ie, basis_root_ie):
368
for new_ie, basis_ie in zip(new_entries, basis_entries):
369
if new_ie != basis_ie:
372
# No actual changes present
375
def _check_pointless(self):
376
if self.allow_pointless:
378
# A merge with no effect on files
379
if len(self.parents) > 1:
381
# work around the fact that a newly-initted tree does differ from its
383
if len(self.basis_inv) == 0 and len(self.builder.new_inventory) == 1:
384
raise PointlessCommit()
385
# Shortcut, if the number of entries changes, then we obviously have
387
if len(self.builder.new_inventory) != len(self.basis_inv):
389
# If length == 1, then we only have the root entry. Which means
390
# that there is no real difference (only the root could be different)
391
if (len(self.builder.new_inventory) != 1 and self._any_real_changes()):
393
raise PointlessCommit()
395
def _check_bound_branch(self):
396
"""Check to see if the local branch is bound.
398
If it is bound, then most of the commit will actually be
399
done using the remote branch as the target branch.
400
Only at the end will the local branch be updated.
402
if self.local and not self.branch.get_bound_location():
403
raise errors.LocalRequiresBoundBranch()
406
self.master_branch = self.branch.get_master_branch()
408
if not self.master_branch:
409
# make this branch the reference branch for out of date checks.
410
self.master_branch = self.branch
413
# If the master branch is bound, we must fail
414
master_bound_location = self.master_branch.get_bound_location()
415
if master_bound_location:
416
raise errors.CommitToDoubleBoundBranch(self.branch,
417
self.master_branch, master_bound_location)
419
# TODO: jam 20051230 We could automatically push local
420
# commits to the remote branch if they would fit.
421
# But for now, just require remote to be identical
424
# Make sure the local branch is identical to the master
425
master_info = self.master_branch.last_revision_info()
426
local_info = self.branch.last_revision_info()
427
if local_info != master_info:
428
raise errors.BoundBranchOutOfDate(self.branch,
431
# Now things are ready to change the master branch
433
self.bound_branch = self.branch
434
self.master_branch.lock_write()
435
self.master_locked = True
437
def _check_out_of_date_tree(self):
438
"""Check that the working tree is up to date.
440
:return: old_revision_number,new_revision_number tuple
443
first_tree_parent = self.work_tree.get_parent_ids()[0]
445
# if there are no parents, treat our parent as 'None'
446
# this is so that we still consider the master branch
447
# - in a checkout scenario the tree may have no
448
# parents but the branch may do.
449
first_tree_parent = bzrlib.revision.NULL_REVISION
450
old_revno, master_last = self.master_branch.last_revision_info()
451
if master_last != first_tree_parent:
452
if master_last != bzrlib.revision.NULL_REVISION:
453
raise errors.OutOfDateTree(self.work_tree)
454
if self.branch.repository.has_revision(first_tree_parent):
455
new_revno = old_revno + 1
457
# ghost parents never appear in revision history.
459
return old_revno,new_revno
461
def _process_hooks(self, old_revno, new_revno):
462
"""Process any registered commit hooks."""
463
# old style commit hooks - should be deprecated ? (obsoleted in
465
if self.config.post_commit() is not None:
466
hooks = self.config.post_commit().split(' ')
467
# this would be nicer with twisted.python.reflect.namedAny
469
result = eval(hook + '(branch, rev_id)',
470
{'branch':self.branch,
472
'rev_id':self.rev_id})
473
# new style commit hooks:
474
if not self.bound_branch:
475
hook_master = self.branch
478
hook_master = self.master_branch
479
hook_local = self.branch
480
# With bound branches, when the master is behind the local branch,
481
# the 'old_revno' and old_revid values here are incorrect.
482
# XXX: FIXME ^. RBC 20060206
484
old_revid = self.parents[0]
486
old_revid = bzrlib.revision.NULL_REVISION
487
for hook in Branch.hooks['post_commit']:
488
hook(hook_local, hook_master, old_revno, old_revid, new_revno,
492
"""Cleanup any open locks, progress bars etc."""
493
cleanups = [self._cleanup_bound_branch,
494
self.basis_tree.unlock,
495
self.work_tree.unlock,
497
found_exception = None
498
for cleanup in cleanups:
501
# we want every cleanup to run no matter what.
502
# so we have a catchall here, but we will raise the
503
# last encountered exception up the stack: and
504
# typically this will be useful enough.
507
if found_exception is not None:
508
# don't do a plan raise, because the last exception may have been
509
# trashed, e is our sure-to-work exception even though it loses the
510
# full traceback. XXX: RBC 20060421 perhaps we could check the
511
# exc_info and if its the same one do a plain raise otherwise
512
# 'raise e' as we do now.
515
def _cleanup_bound_branch(self):
516
"""Executed at the end of a try/finally to cleanup a bound branch.
518
If the branch wasn't bound, this is a no-op.
519
If it was, it resents self.branch to the local branch, instead
522
if not self.bound_branch:
524
if self.master_locked:
525
self.master_branch.unlock()
527
def _escape_commit_message(self):
528
"""Replace xml-incompatible control characters."""
529
# FIXME: RBC 20060419 this should be done by the revision
530
# serialiser not by commit. Then we can also add an unescaper
531
# in the deserializer and start roundtripping revision messages
532
# precisely. See repository_implementations/test_repository.py
534
# Python strings can include characters that can't be
535
# represented in well-formed XML; escape characters that
536
# aren't listed in the XML specification
537
# (http://www.w3.org/TR/REC-xml/#NT-Char).
538
self.message, escape_count = re.subn(
539
u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
540
lambda match: match.group(0).encode('unicode_escape'),
543
self.reporter.escaped(escape_count, self.message)
545
def _gather_parents(self):
546
"""Record the parents of a merge for merge detection."""
547
# TODO: Make sure that this list doesn't contain duplicate
548
# entries and the order is preserved when doing this.
549
self.parents = self.work_tree.get_parent_ids()
550
self.parent_invs = [self.basis_inv]
551
for revision in self.parents[1:]:
552
if self.branch.repository.has_revision(revision):
553
mutter('commit parent revision {%s}', revision)
554
inventory = self.branch.repository.get_inventory(revision)
555
self.parent_invs.append(inventory)
557
mutter('commit parent ghost revision {%s}', revision)
559
def _remove_deleted(self):
560
"""Remove deleted files from the working inventories.
562
This is done prior to taking the working inventory as the
563
basis for the new committed inventory.
565
This returns true if any files
566
*that existed in the basis inventory* were deleted.
567
Files that were added and deleted
568
in the working copy don't matter.
570
specific = self.specific_files
572
deleted_paths = set()
573
for path, ie in self.work_inv.iter_entries():
574
if is_inside_any(deleted_paths, path):
575
# The tree will delete the required ids recursively.
577
if specific and not is_inside_any(specific, path):
579
if not self.work_tree.has_filename(path):
580
deleted_paths.add(path)
581
self.reporter.missing(path)
582
deleted_ids.append(ie.file_id)
583
self.work_tree.unversion(deleted_ids)
585
def _populate_new_inv(self):
586
"""Build revision inventory.
588
This creates a new empty inventory. Depending on
589
which files are selected for commit, and what is present in the
590
current tree, the new inventory is populated. inventory entries
591
which are candidates for modification have their revision set to
592
None; inventory entries that are carried over untouched have their
593
revision set to their prior value.
595
# ESEPARATIONOFCONCERNS: this function is diffing and using the diff
596
# results to create a new inventory at the same time, which results
597
# in bugs like #46635. Any reason not to use/enhance Tree.changes_from?
599
mutter("Selecting files for commit with filter %s", self.specific_files)
600
assert self.work_inv.root is not None
601
entries = self.work_inv.iter_entries()
602
if not self.builder.record_root_entry:
603
symbol_versioning.warn('CommitBuilders should support recording'
604
' the root entry as of bzr 0.10.', DeprecationWarning,
606
self.builder.new_inventory.add(self.basis_inv.root.copy())
608
self._emit_progress_update()
609
for path, new_ie in entries:
610
self._emit_progress_update()
611
file_id = new_ie.file_id
613
kind = self.work_tree.kind(file_id)
614
if kind == 'tree-reference' and self.recursive == 'down':
615
# nested tree: commit in it
616
sub_tree = WorkingTree.open(self.work_tree.abspath(path))
617
# FIXME: be more comprehensive here:
618
# this works when both trees are in --trees repository,
619
# but when both are bound to a different repository,
620
# it fails; a better way of approaching this is to
621
# finally implement the explicit-caches approach design
622
# a while back - RBC 20070306.
623
if (sub_tree.branch.repository.bzrdir.root_transport.base
625
self.work_tree.branch.repository.bzrdir.root_transport.base):
626
sub_tree.branch.repository = \
627
self.work_tree.branch.repository
629
sub_tree.commit(message=None, revprops=self.revprops,
630
recursive=self.recursive,
631
message_callback=self.message_callback,
632
timestamp=self.timestamp, timezone=self.timezone,
633
committer=self.committer,
634
allow_pointless=self.allow_pointless,
635
strict=self.strict, verbose=self.verbose,
636
local=self.local, reporter=self.reporter)
637
except errors.PointlessCommit:
639
if kind != new_ie.kind:
640
new_ie = inventory.make_entry(kind, new_ie.name,
641
new_ie.parent_id, file_id)
642
except errors.NoSuchFile:
644
# mutter('check %s {%s}', path, file_id)
645
if (not self.specific_files or
646
is_inside_or_parent_of_any(self.specific_files, path)):
647
# mutter('%s selected for commit', path)
651
# mutter('%s not selected for commit', path)
652
if self.basis_inv.has_id(file_id):
653
ie = self.basis_inv[file_id].copy()
655
# this entry is new and not being committed
657
self.builder.record_entry_contents(ie, self.parent_invs,
658
path, self.work_tree)
659
# describe the nature of the change that has occurred relative to
660
# the basis inventory.
661
if (self.basis_inv.has_id(ie.file_id)):
662
basis_ie = self.basis_inv[ie.file_id]
665
change = ie.describe_change(basis_ie, ie)
666
if change in (InventoryEntry.RENAMED,
667
InventoryEntry.MODIFIED_AND_RENAMED):
668
old_path = self.basis_inv.id2path(ie.file_id)
669
self.reporter.renamed(change, old_path, path)
671
self.reporter.snapshot_change(change, path)
673
if not self.specific_files:
676
# ignore removals that don't match filespec
677
for path, new_ie in self.basis_inv.iter_entries():
678
if new_ie.file_id in self.work_inv:
680
if is_inside_any(self.specific_files, path):
684
self.builder.record_entry_contents(ie, self.parent_invs, path,
687
def _emit_progress_update(self):
688
"""Emit an update to the progress bar."""
689
self.pb.update("Committing", self.pb_count, self.pb_total)
692
def _report_deletes(self):
693
for path, ie in self.basis_inv.iter_entries():
694
if ie.file_id not in self.builder.new_inventory:
695
self.reporter.deleted(path)
added
removed
bzrlib/commit.py
