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
65
from bzrlib.branch import Branch
67
from bzrlib.errors import (BzrError, PointlessCommit,
71
from bzrlib.osutils import (kind_marker, isdir,isfile, is_inside_any,
72
is_inside_or_parent_of_any,
73
quotefn, sha_file, split_lines)
74
from bzrlib.testament import Testament
75
from bzrlib.trace import mutter, note, warning
76
from bzrlib.xml5 import serializer_v5
77
from bzrlib.inventory import Inventory, InventoryEntry
78
from bzrlib import symbol_versioning
79
from bzrlib.symbol_versioning import (deprecated_passed,
82
from bzrlib.workingtree import WorkingTree
86
class NullCommitReporter(object):
87
"""I report on progress of a commit."""
89
def snapshot_change(self, change, path):
92
def completed(self, revno, rev_id):
95
def deleted(self, file_id):
98
def escaped(self, escape_count, message):
101
def missing(self, path):
104
def renamed(self, change, old_path, new_path):
108
class ReportCommitToLog(NullCommitReporter):
110
def _note(self, format, *args):
113
Subclasses may choose to override this method.
117
def snapshot_change(self, change, path):
118
if change == 'unchanged':
120
if change == 'added' and path == '':
122
self._note("%s %s", change, path)
124
def completed(self, revno, rev_id):
125
self._note('Committed revision %d.', revno)
127
def deleted(self, file_id):
128
self._note('deleted %s', file_id)
130
def escaped(self, escape_count, message):
131
self._note("replaced %d control characters in message", escape_count)
133
def missing(self, path):
134
self._note('missing %s', path)
136
def renamed(self, change, old_path, new_path):
137
self._note('%s %s => %s', change, old_path, new_path)
140
class Commit(object):
141
"""Task of committing a new revision.
143
This is a MethodObject: it accumulates state as the commit is
144
prepared, and then it is discarded. It doesn't represent
145
historical revisions, just the act of recording a new one.
148
Modified to hold a list of files that have been deleted from
149
the working directory; these should be removed from the
155
if reporter is not None:
156
self.reporter = reporter
158
self.reporter = NullCommitReporter()
168
allow_pointless=True,
176
message_callback=None,
178
"""Commit working copy as a new revision.
180
:param message: the commit message (it or message_callback is required)
182
:param timestamp: if not None, seconds-since-epoch for a
183
postdated/predated commit.
185
:param specific_files: If true, commit only those files.
187
:param rev_id: If set, use this as the new revision id.
188
Useful for test or import commands that need to tightly
189
control what revisions are assigned. If you duplicate
190
a revision id that exists elsewhere it is your own fault.
191
If null (default), a time/random revision id is generated.
193
:param allow_pointless: If true (default), commit even if nothing
194
has changed and no merges are recorded.
196
:param strict: If true, don't allow a commit if the working tree
197
contains unknown files.
199
:param revprops: Properties for new revision
200
:param local: Perform a local only commit.
201
:param recursive: If set to 'down', commit in any subtrees that have
202
pending changes of any sort during this commit.
204
mutter('preparing to commit')
206
if working_tree is None:
207
raise BzrError("working_tree must be passed into commit().")
209
self.work_tree = working_tree
210
self.branch = self.work_tree.branch
211
if getattr(self.work_tree, 'requires_rich_root', lambda: False)():
212
if not self.branch.repository.supports_rich_root():
213
raise errors.RootNotRich()
214
if message_callback is None:
215
if message is not None:
216
if isinstance(message, str):
217
message = message.decode(bzrlib.user_encoding)
218
message_callback = lambda x: message
220
raise BzrError("The message or message_callback keyword"
221
" parameter is required for commit().")
223
self.bound_branch = None
225
self.master_branch = None
226
self.master_locked = False
228
self.specific_files = specific_files
229
self.allow_pointless = allow_pointless
230
self.recursive = recursive
231
self.revprops = revprops
232
self.message_callback = message_callback
233
self.timestamp = timestamp
234
self.timezone = timezone
235
self.committer = committer
237
self.verbose = verbose
239
if reporter is None and self.reporter is None:
240
self.reporter = NullCommitReporter()
241
elif reporter is not None:
242
self.reporter = reporter
244
self.work_tree.lock_write()
245
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
246
self.basis_tree = self.work_tree.basis_tree()
247
self.basis_tree.lock_read()
249
# Cannot commit with conflicts present.
250
if len(self.work_tree.conflicts()) > 0:
251
raise ConflictsInTree
253
# Setup the bound branch variables as needed.
254
self._check_bound_branch()
256
# Check that the working tree is up to date
257
old_revno,new_revno = self._check_out_of_date_tree()
259
if self.config is None:
260
self.config = self.branch.get_config()
262
# If provided, ensure the specified files are versioned
263
if specific_files is not None:
264
# Note: We don't actually need the IDs here. This routine
265
# is being called because it raises PathNotVerisonedError
266
# as a side effect of finding the IDs.
267
# XXX: Dont we have filter_unversioned to do this more
269
tree.find_ids_across_trees(specific_files,
270
[self.basis_tree, self.work_tree])
272
# Setup the progress bar. As the number of files that need to be
273
# committed in unknown, progress is reported as stages.
274
# We keep track of entries separately though and include that
275
# information in the progress bar during the relevant stages.
276
self.pb_stage_name = ""
277
self.pb_stage_count = 0
278
self.pb_stage_total = 4
279
if self.bound_branch:
280
self.pb_stage_total += 1
281
self.pb.show_pct = False
282
self.pb.show_spinner = False
283
self.pb.show_eta = False
284
self.pb.show_count = True
285
self.pb.show_bar = True
287
# After a merge, a selected file commit is not supported.
288
# See 'bzr help merge' for an explanation as to why.
289
self.basis_inv = self.basis_tree.inventory
290
self._gather_parents()
291
if len(self.parents) > 1 and self.specific_files:
292
raise errors.CannotCommitSelectedFileMerge(self.specific_files)
294
# Collect the changes
295
self._set_progress_stage("Collecting changes",
296
entries_title="Directory")
297
self.builder = self.branch.get_commit_builder(self.parents,
298
self.config, timestamp, timezone, committer, revprops, rev_id)
300
self._update_builder_with_changes()
301
self._check_pointless()
303
# TODO: Now the new inventory is known, check for conflicts.
304
# ADHB 2006-08-08: If this is done, populate_new_inv should not add
305
# weave lines, because nothing should be recorded until it is known
306
# that commit will succeed.
307
self._set_progress_stage("Saving data locally")
308
self.builder.finish_inventory()
310
# Prompt the user for a commit message if none provided
311
message = message_callback(self)
312
assert isinstance(message, unicode), type(message)
313
self.message = message
314
self._escape_commit_message()
316
# Add revision data to the local branch
317
self.rev_id = self.builder.commit(self.message)
319
# perhaps this should be done by the CommitBuilder ?
320
self.work_tree.branch.repository.abort_write_group()
323
# Upload revision data to the master.
324
# this will propagate merged revisions too if needed.
325
if self.bound_branch:
326
self._set_progress_stage("Uploading data to master branch")
327
self.master_branch.repository.fetch(self.branch.repository,
328
revision_id=self.rev_id)
329
# now the master has the revision data
330
# 'commit' to the master first so a timeout here causes the
331
# local branch to be out of date
332
self.master_branch.set_last_revision_info(new_revno,
335
# and now do the commit locally.
336
self.branch.set_last_revision_info(new_revno, self.rev_id)
338
# Make the working tree up to date with the branch
339
self._set_progress_stage("Updating the working tree")
340
rev_tree = self.builder.revision_tree()
341
self.work_tree.set_parent_trees([(self.rev_id, rev_tree)])
342
self.reporter.completed(new_revno, self.rev_id)
343
self._process_hooks(old_revno, new_revno)
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)
371
and (this.reference_revision == other.reference_revision)
373
if not ie_equal_no_revision(new_root_ie, basis_root_ie):
376
for new_ie, basis_ie in zip(new_entries, basis_entries):
377
if new_ie != basis_ie:
380
# No actual changes present
383
def _check_pointless(self):
384
if self.allow_pointless:
386
# A merge with no effect on files
387
if len(self.parents) > 1:
389
# work around the fact that a newly-initted tree does differ from its
391
if len(self.basis_inv) == 0 and len(self.builder.new_inventory) == 1:
392
raise PointlessCommit()
393
# Shortcut, if the number of entries changes, then we obviously have
395
if len(self.builder.new_inventory) != len(self.basis_inv):
397
# If length == 1, then we only have the root entry. Which means
398
# that there is no real difference (only the root could be different)
399
if (len(self.builder.new_inventory) != 1 and self._any_real_changes()):
401
raise PointlessCommit()
403
def _check_bound_branch(self):
404
"""Check to see if the local branch is bound.
406
If it is bound, then most of the commit will actually be
407
done using the remote branch as the target branch.
408
Only at the end will the local branch be updated.
410
if self.local and not self.branch.get_bound_location():
411
raise errors.LocalRequiresBoundBranch()
414
self.master_branch = self.branch.get_master_branch()
416
if not self.master_branch:
417
# make this branch the reference branch for out of date checks.
418
self.master_branch = self.branch
421
# If the master branch is bound, we must fail
422
master_bound_location = self.master_branch.get_bound_location()
423
if master_bound_location:
424
raise errors.CommitToDoubleBoundBranch(self.branch,
425
self.master_branch, master_bound_location)
427
# TODO: jam 20051230 We could automatically push local
428
# commits to the remote branch if they would fit.
429
# But for now, just require remote to be identical
432
# Make sure the local branch is identical to the master
433
master_info = self.master_branch.last_revision_info()
434
local_info = self.branch.last_revision_info()
435
if local_info != master_info:
436
raise errors.BoundBranchOutOfDate(self.branch,
439
# Now things are ready to change the master branch
441
self.bound_branch = self.branch
442
self.master_branch.lock_write()
443
self.master_locked = True
445
def _check_out_of_date_tree(self):
446
"""Check that the working tree is up to date.
448
:return: old_revision_number,new_revision_number tuple
451
first_tree_parent = self.work_tree.get_parent_ids()[0]
453
# if there are no parents, treat our parent as 'None'
454
# this is so that we still consider the master branch
455
# - in a checkout scenario the tree may have no
456
# parents but the branch may do.
457
first_tree_parent = bzrlib.revision.NULL_REVISION
458
old_revno, master_last = self.master_branch.last_revision_info()
459
if master_last != first_tree_parent:
460
if master_last != bzrlib.revision.NULL_REVISION:
461
raise errors.OutOfDateTree(self.work_tree)
462
if self.branch.repository.has_revision(first_tree_parent):
463
new_revno = old_revno + 1
465
# ghost parents never appear in revision history.
467
return old_revno,new_revno
469
def _process_hooks(self, old_revno, new_revno):
470
"""Process any registered commit hooks."""
471
# Process the post commit hooks, if any
472
self._set_progress_stage("Running post commit hooks")
473
# old style commit hooks - should be deprecated ? (obsoleted in
475
if self.config.post_commit() is not None:
476
hooks = self.config.post_commit().split(' ')
477
# this would be nicer with twisted.python.reflect.namedAny
479
result = eval(hook + '(branch, rev_id)',
480
{'branch':self.branch,
482
'rev_id':self.rev_id})
483
# new style commit hooks:
484
if not self.bound_branch:
485
hook_master = self.branch
488
hook_master = self.master_branch
489
hook_local = self.branch
490
# With bound branches, when the master is behind the local branch,
491
# the 'old_revno' and old_revid values here are incorrect.
492
# XXX: FIXME ^. RBC 20060206
494
old_revid = self.parents[0]
496
old_revid = bzrlib.revision.NULL_REVISION
497
for hook in Branch.hooks['post_commit']:
498
# show the running hook in the progress bar. As hooks may
499
# end up doing nothing (e.g. because they are not configured by
500
# the user) this is still showing progress, not showing overall
501
# actions - its up to each plugin to show a UI if it want's to
502
# (such as 'Emailing diff to foo@example.com').
503
self.pb_stage_name = "Running post commit hooks [%s]" % \
504
Branch.hooks.get_hook_name(hook)
505
self._emit_progress()
506
if 'hooks' in debug.debug_flags:
507
mutter("Invoking commit hook: %r", hook)
508
hook(hook_local, hook_master, old_revno, old_revid, new_revno,
512
"""Cleanup any open locks, progress bars etc."""
513
cleanups = [self._cleanup_bound_branch,
514
self.basis_tree.unlock,
515
self.work_tree.unlock,
517
found_exception = None
518
for cleanup in cleanups:
521
# we want every cleanup to run no matter what.
522
# so we have a catchall here, but we will raise the
523
# last encountered exception up the stack: and
524
# typically this will be useful enough.
527
if found_exception is not None:
528
# don't do a plan raise, because the last exception may have been
529
# trashed, e is our sure-to-work exception even though it loses the
530
# full traceback. XXX: RBC 20060421 perhaps we could check the
531
# exc_info and if its the same one do a plain raise otherwise
532
# 'raise e' as we do now.
535
def _cleanup_bound_branch(self):
536
"""Executed at the end of a try/finally to cleanup a bound branch.
538
If the branch wasn't bound, this is a no-op.
539
If it was, it resents self.branch to the local branch, instead
542
if not self.bound_branch:
544
if self.master_locked:
545
self.master_branch.unlock()
547
def _escape_commit_message(self):
548
"""Replace xml-incompatible control characters."""
549
# FIXME: RBC 20060419 this should be done by the revision
550
# serialiser not by commit. Then we can also add an unescaper
551
# in the deserializer and start roundtripping revision messages
552
# precisely. See repository_implementations/test_repository.py
554
# Python strings can include characters that can't be
555
# represented in well-formed XML; escape characters that
556
# aren't listed in the XML specification
557
# (http://www.w3.org/TR/REC-xml/#NT-Char).
558
self.message, escape_count = re.subn(
559
u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
560
lambda match: match.group(0).encode('unicode_escape'),
563
self.reporter.escaped(escape_count, self.message)
565
def _gather_parents(self):
566
"""Record the parents of a merge for merge detection."""
567
# TODO: Make sure that this list doesn't contain duplicate
568
# entries and the order is preserved when doing this.
569
self.parents = self.work_tree.get_parent_ids()
570
self.parent_invs = [self.basis_inv]
571
for revision in self.parents[1:]:
572
if self.branch.repository.has_revision(revision):
573
mutter('commit parent revision {%s}', revision)
574
inventory = self.branch.repository.get_inventory(revision)
575
self.parent_invs.append(inventory)
577
mutter('commit parent ghost revision {%s}', revision)
579
def _update_builder_with_changes(self):
580
"""Update the commit builder with the data about what has changed.
582
# Build the revision inventory.
584
# This starts by creating a new empty inventory. Depending on
585
# which files are selected for commit, and what is present in the
586
# current tree, the new inventory is populated. inventory entries
587
# which are candidates for modification have their revision set to
588
# None; inventory entries that are carried over untouched have their
589
# revision set to their prior value.
591
# ESEPARATIONOFCONCERNS: this function is diffing and using the diff
592
# results to create a new inventory at the same time, which results
593
# in bugs like #46635. Any reason not to use/enhance Tree.changes_from?
596
specific_files = self.specific_files
597
mutter("Selecting files for commit with filter %s", specific_files)
599
# Check and warn about old CommitBuilders
600
if not self.builder.record_root_entry:
601
symbol_versioning.warn('CommitBuilders should support recording'
602
' the root entry as of bzr 0.10.', DeprecationWarning,
604
self.builder.new_inventory.add(self.basis_inv.root.copy())
606
# Build the new inventory
607
self._populate_from_inventory(specific_files)
609
# If specific files are selected, then all un-selected files must be
610
# recorded in their previous state. For more details, see
611
# https://lists.ubuntu.com/archives/bazaar/2007q3/028476.html.
613
for path, new_ie in self.basis_inv.iter_entries():
614
if new_ie.file_id in self.builder.new_inventory:
616
if is_inside_any(specific_files, path):
620
self.builder.record_entry_contents(ie, self.parent_invs, path,
623
# Report what was deleted. We could skip this when no deletes are
624
# detected to gain a performance win, but it arguably serves as a
625
# 'safety check' by informing the user whenever anything disappears.
626
for path, ie in self.basis_inv.iter_entries():
627
if ie.file_id not in self.builder.new_inventory:
628
self.reporter.deleted(path)
630
def _populate_from_inventory(self, specific_files):
631
"""Populate the CommitBuilder by walking the working tree inventory."""
633
# raise an exception as soon as we find a single unknown.
634
for unknown in self.work_tree.unknowns():
635
raise StrictCommitFailed()
638
deleted_paths = set()
639
work_inv = self.work_tree.inventory
640
assert work_inv.root is not None
641
entries = work_inv.iter_entries()
642
if not self.builder.record_root_entry:
644
for path, existing_ie in entries:
645
file_id = existing_ie.file_id
646
name = existing_ie.name
647
parent_id = existing_ie.parent_id
648
kind = existing_ie.kind
649
if kind == 'directory':
650
self._next_progress_entry()
652
# Skip files that have been deleted from the working tree.
653
# The deleted files/directories are also recorded so they
654
# can be explicitly unversioned later. Note that when a
655
# filter of specific files is given, we must only skip/record
656
# deleted files matching that filter.
657
if is_inside_any(deleted_paths, path):
659
if not specific_files or is_inside_any(specific_files, path):
660
if not self.work_tree.has_filename(path):
661
deleted_paths.add(path)
662
self.reporter.missing(path)
663
deleted_ids.append(file_id)
666
kind = self.work_tree.kind(file_id)
667
# TODO: specific_files filtering before nested tree processing
668
if kind == 'tree-reference' and self.recursive == 'down':
669
self._commit_nested_tree(file_id, path)
670
except errors.NoSuchFile:
673
# Record an entry for this item
674
# Note: I don't particularly want to have the existing_ie
675
# parameter but the test suite currently (28-Jun-07) breaks
676
# without it thanks to a unicode normalisation issue. :-(
677
definitely_changed = kind != existing_ie.kind
678
self._record_entry(path, file_id, specific_files, kind, name,
679
parent_id, definitely_changed, existing_ie)
681
# Unversion IDs that were found to be deleted
682
self.work_tree.unversion(deleted_ids)
684
def _commit_nested_tree(self, file_id, path):
685
"Commit a nested tree."
686
sub_tree = self.work_tree.get_nested_tree(file_id, path)
687
# FIXME: be more comprehensive here:
688
# this works when both trees are in --trees repository,
689
# but when both are bound to a different repository,
690
# it fails; a better way of approaching this is to
691
# finally implement the explicit-caches approach design
692
# a while back - RBC 20070306.
693
if (sub_tree.branch.repository.bzrdir.root_transport.base
695
self.work_tree.branch.repository.bzrdir.root_transport.base):
696
sub_tree.branch.repository = \
697
self.work_tree.branch.repository
699
sub_tree.commit(message=None, revprops=self.revprops,
700
recursive=self.recursive,
701
message_callback=self.message_callback,
702
timestamp=self.timestamp, timezone=self.timezone,
703
committer=self.committer,
704
allow_pointless=self.allow_pointless,
705
strict=self.strict, verbose=self.verbose,
706
local=self.local, reporter=self.reporter)
707
except errors.PointlessCommit:
710
def _record_entry(self, path, file_id, specific_files, kind, name,
711
parent_id, definitely_changed, existing_ie=None):
712
"Record the new inventory entry for a path if any."
713
# mutter('check %s {%s}', path, file_id)
714
if (not specific_files or
715
is_inside_or_parent_of_any(specific_files, path)):
716
# mutter('%s selected for commit', path)
717
if definitely_changed or existing_ie is None:
718
ie = inventory.make_entry(kind, name, parent_id, file_id)
720
ie = existing_ie.copy()
723
# mutter('%s not selected for commit', path)
724
if self.basis_inv.has_id(file_id):
725
ie = self.basis_inv[file_id].copy()
727
# this entry is new and not being committed
730
self.builder.record_entry_contents(ie, self.parent_invs,
731
path, self.work_tree)
732
self._report_change(ie, path)
735
def _report_change(self, ie, path):
736
"""Report a change to the user.
738
The change that has occurred is described relative to the basis
741
if (self.basis_inv.has_id(ie.file_id)):
742
basis_ie = self.basis_inv[ie.file_id]
745
change = ie.describe_change(basis_ie, ie)
746
if change in (InventoryEntry.RENAMED,
747
InventoryEntry.MODIFIED_AND_RENAMED):
748
old_path = self.basis_inv.id2path(ie.file_id)
749
self.reporter.renamed(change, old_path, path)
751
self.reporter.snapshot_change(change, path)
753
def _set_progress_stage(self, name, entries_title=None):
754
"""Set the progress stage and emit an update to the progress bar."""
755
self.pb_stage_name = name
756
self.pb_stage_count += 1
757
self.pb_entries_title = entries_title
758
if entries_title is not None:
759
self.pb_entries_count = 0
760
self.pb_entries_total = '?'
761
self._emit_progress()
763
def _next_progress_entry(self):
764
"""Emit an update to the progress bar and increment the entry count."""
765
self.pb_entries_count += 1
766
self._emit_progress()
768
def _emit_progress(self):
769
if self.pb_entries_title:
770
if self.pb_entries_total == '?':
771
text = "%s [%s %d] - Stage" % (self.pb_stage_name,
772
self.pb_entries_title, self.pb_entries_count)
774
text = "%s [%s %d/%s] - Stage" % (self.pb_stage_name,
775
self.pb_entries_title, self.pb_entries_count,
776
str(self.pb_entries_total))
778
text = "%s - Stage" % (self.pb_stage_name)
779
self.pb.update(text, self.pb_stage_count, self.pb_stage_total)
added
removed
bzrlib/commit.py
