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()
260
# raise an exception as soon as we find a single unknown.
261
for unknown in self.work_tree.unknowns():
262
raise StrictCommitFailed()
264
if self.config is None:
265
self.config = self.branch.get_config()
267
# If provided, ensure the specified files are versioned
268
if specific_files is not None:
269
# Note: We don't actually need the IDs here. This routine
270
# is being called because it raises PathNotVerisonedError
271
# as a side effect of finding the IDs.
272
# XXX: Dont we have filter_unversioned to do this more
274
tree.find_ids_across_trees(specific_files,
275
[self.basis_tree, self.work_tree])
277
# Setup the progress bar. As the number of files that need to be
278
# committed in unknown, progress is reported as stages.
279
# We keep track of entries separately though and include that
280
# information in the progress bar during the relevant stages.
281
self.pb_stage_name = ""
282
self.pb_stage_count = 0
283
self.pb_stage_total = 4
284
if self.bound_branch:
285
self.pb_stage_total += 1
286
self.pb.show_pct = False
287
self.pb.show_spinner = False
288
self.pb.show_eta = False
289
self.pb.show_count = True
290
self.pb.show_bar = True
292
# After a merge, a selected file commit is not supported.
293
# See 'bzr help merge' for an explanation as to why.
294
self.basis_inv = self.basis_tree.inventory
295
self._gather_parents()
296
if len(self.parents) > 1 and self.specific_files:
297
raise errors.CannotCommitSelectedFileMerge(self.specific_files)
299
# Collect the changes
300
self._emit_progress_set_stage("Collecting changes", show_entries=True)
301
self.builder = self.branch.get_commit_builder(self.parents,
302
self.config, timestamp, timezone, committer, revprops, rev_id)
304
self._update_builder_with_changes()
305
self._check_pointless()
307
# TODO: Now the new inventory is known, check for conflicts.
308
# ADHB 2006-08-08: If this is done, populate_new_inv should not add
309
# weave lines, because nothing should be recorded until it is known
310
# that commit will succeed.
311
self._emit_progress_set_stage("Saving data locally")
312
self.builder.finish_inventory()
314
# Prompt the user for a commit message if none provided
315
message = message_callback(self)
316
assert isinstance(message, unicode), type(message)
317
self.message = message
318
self._escape_commit_message()
320
# Add revision data to the local branch
321
self.rev_id = self.builder.commit(self.message)
323
# perhaps this should be done by the CommitBuilder ?
324
self.work_tree.branch.repository.abort_write_group()
327
# Upload revision data to the master.
328
# this will propagate merged revisions too if needed.
329
if self.bound_branch:
330
self._emit_progress_set_stage("Uploading data to master branch")
331
self.master_branch.repository.fetch(self.branch.repository,
332
revision_id=self.rev_id)
333
# now the master has the revision data
334
# 'commit' to the master first so a timeout here causes the
335
# local branch to be out of date
336
self.master_branch.set_last_revision_info(new_revno,
339
# and now do the commit locally.
340
self.branch.set_last_revision_info(new_revno, self.rev_id)
342
# Make the working tree up to date with the branch
343
self._emit_progress_set_stage("Updating the working tree")
344
rev_tree = self.builder.revision_tree()
345
self.work_tree.set_parent_trees([(self.rev_id, rev_tree)])
346
self.reporter.completed(new_revno, self.rev_id)
347
self._process_hooks(old_revno, new_revno)
352
def _any_real_changes(self):
353
"""Are there real changes between new_inventory and basis?
355
For trees without rich roots, inv.root.revision changes every commit.
356
But if that is the only change, we want to treat it as though there
359
new_entries = self.builder.new_inventory.iter_entries()
360
basis_entries = self.basis_inv.iter_entries()
361
new_path, new_root_ie = new_entries.next()
362
basis_path, basis_root_ie = basis_entries.next()
364
# This is a copy of InventoryEntry.__eq__ only leaving out .revision
365
def ie_equal_no_revision(this, other):
366
return ((this.file_id == other.file_id)
367
and (this.name == other.name)
368
and (this.symlink_target == other.symlink_target)
369
and (this.text_sha1 == other.text_sha1)
370
and (this.text_size == other.text_size)
371
and (this.text_id == other.text_id)
372
and (this.parent_id == other.parent_id)
373
and (this.kind == other.kind)
374
and (this.executable == other.executable)
375
and (this.reference_revision == other.reference_revision)
377
if not ie_equal_no_revision(new_root_ie, basis_root_ie):
380
for new_ie, basis_ie in zip(new_entries, basis_entries):
381
if new_ie != basis_ie:
384
# No actual changes present
387
def _check_pointless(self):
388
if self.allow_pointless:
390
# A merge with no effect on files
391
if len(self.parents) > 1:
393
# work around the fact that a newly-initted tree does differ from its
395
if len(self.basis_inv) == 0 and len(self.builder.new_inventory) == 1:
396
raise PointlessCommit()
397
# Shortcut, if the number of entries changes, then we obviously have
399
if len(self.builder.new_inventory) != len(self.basis_inv):
401
# If length == 1, then we only have the root entry. Which means
402
# that there is no real difference (only the root could be different)
403
if (len(self.builder.new_inventory) != 1 and self._any_real_changes()):
405
raise PointlessCommit()
407
def _check_bound_branch(self):
408
"""Check to see if the local branch is bound.
410
If it is bound, then most of the commit will actually be
411
done using the remote branch as the target branch.
412
Only at the end will the local branch be updated.
414
if self.local and not self.branch.get_bound_location():
415
raise errors.LocalRequiresBoundBranch()
418
self.master_branch = self.branch.get_master_branch()
420
if not self.master_branch:
421
# make this branch the reference branch for out of date checks.
422
self.master_branch = self.branch
425
# If the master branch is bound, we must fail
426
master_bound_location = self.master_branch.get_bound_location()
427
if master_bound_location:
428
raise errors.CommitToDoubleBoundBranch(self.branch,
429
self.master_branch, master_bound_location)
431
# TODO: jam 20051230 We could automatically push local
432
# commits to the remote branch if they would fit.
433
# But for now, just require remote to be identical
436
# Make sure the local branch is identical to the master
437
master_info = self.master_branch.last_revision_info()
438
local_info = self.branch.last_revision_info()
439
if local_info != master_info:
440
raise errors.BoundBranchOutOfDate(self.branch,
443
# Now things are ready to change the master branch
445
self.bound_branch = self.branch
446
self.master_branch.lock_write()
447
self.master_locked = True
449
def _check_out_of_date_tree(self):
450
"""Check that the working tree is up to date.
452
:return: old_revision_number,new_revision_number tuple
455
first_tree_parent = self.work_tree.get_parent_ids()[0]
457
# if there are no parents, treat our parent as 'None'
458
# this is so that we still consider the master branch
459
# - in a checkout scenario the tree may have no
460
# parents but the branch may do.
461
first_tree_parent = bzrlib.revision.NULL_REVISION
462
old_revno, master_last = self.master_branch.last_revision_info()
463
if master_last != first_tree_parent:
464
if master_last != bzrlib.revision.NULL_REVISION:
465
raise errors.OutOfDateTree(self.work_tree)
466
if self.branch.repository.has_revision(first_tree_parent):
467
new_revno = old_revno + 1
469
# ghost parents never appear in revision history.
471
return old_revno,new_revno
473
def _process_hooks(self, old_revno, new_revno):
474
"""Process any registered commit hooks."""
475
# Process the post commit hooks, if any
476
self._emit_progress_set_stage("Running post commit hooks")
477
# old style commit hooks - should be deprecated ? (obsoleted in
479
if self.config.post_commit() is not None:
480
hooks = self.config.post_commit().split(' ')
481
# this would be nicer with twisted.python.reflect.namedAny
483
result = eval(hook + '(branch, rev_id)',
484
{'branch':self.branch,
486
'rev_id':self.rev_id})
487
# new style commit hooks:
488
if not self.bound_branch:
489
hook_master = self.branch
492
hook_master = self.master_branch
493
hook_local = self.branch
494
# With bound branches, when the master is behind the local branch,
495
# the 'old_revno' and old_revid values here are incorrect.
496
# XXX: FIXME ^. RBC 20060206
498
old_revid = self.parents[0]
500
old_revid = bzrlib.revision.NULL_REVISION
501
for hook in Branch.hooks['post_commit']:
502
# show the running hook in the progress bar. As hooks may
503
# end up doing nothing (e.g. because they are not configured by
504
# the user) this is still showing progress, not showing overall
505
# actions - its up to each plugin to show a UI if it want's to
506
# (such as 'Emailing diff to foo@example.com').
507
self.pb_stage_name = "Running post commit hooks [%s]" % \
508
Branch.hooks.get_hook_name(hook)
509
self._emit_progress()
510
if 'hooks' in debug.debug_flags:
511
mutter("Invoking commit hook: %r", hook)
512
hook(hook_local, hook_master, old_revno, old_revid, new_revno,
516
"""Cleanup any open locks, progress bars etc."""
517
cleanups = [self._cleanup_bound_branch,
518
self.basis_tree.unlock,
519
self.work_tree.unlock,
521
found_exception = None
522
for cleanup in cleanups:
525
# we want every cleanup to run no matter what.
526
# so we have a catchall here, but we will raise the
527
# last encountered exception up the stack: and
528
# typically this will be useful enough.
531
if found_exception is not None:
532
# don't do a plan raise, because the last exception may have been
533
# trashed, e is our sure-to-work exception even though it loses the
534
# full traceback. XXX: RBC 20060421 perhaps we could check the
535
# exc_info and if its the same one do a plain raise otherwise
536
# 'raise e' as we do now.
539
def _cleanup_bound_branch(self):
540
"""Executed at the end of a try/finally to cleanup a bound branch.
542
If the branch wasn't bound, this is a no-op.
543
If it was, it resents self.branch to the local branch, instead
546
if not self.bound_branch:
548
if self.master_locked:
549
self.master_branch.unlock()
551
def _escape_commit_message(self):
552
"""Replace xml-incompatible control characters."""
553
# FIXME: RBC 20060419 this should be done by the revision
554
# serialiser not by commit. Then we can also add an unescaper
555
# in the deserializer and start roundtripping revision messages
556
# precisely. See repository_implementations/test_repository.py
558
# Python strings can include characters that can't be
559
# represented in well-formed XML; escape characters that
560
# aren't listed in the XML specification
561
# (http://www.w3.org/TR/REC-xml/#NT-Char).
562
self.message, escape_count = re.subn(
563
u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
564
lambda match: match.group(0).encode('unicode_escape'),
567
self.reporter.escaped(escape_count, self.message)
569
def _gather_parents(self):
570
"""Record the parents of a merge for merge detection."""
571
# TODO: Make sure that this list doesn't contain duplicate
572
# entries and the order is preserved when doing this.
573
self.parents = self.work_tree.get_parent_ids()
574
self.parent_invs = [self.basis_inv]
575
for revision in self.parents[1:]:
576
if self.branch.repository.has_revision(revision):
577
mutter('commit parent revision {%s}', revision)
578
inventory = self.branch.repository.get_inventory(revision)
579
self.parent_invs.append(inventory)
581
mutter('commit parent ghost revision {%s}', revision)
583
def _update_builder_with_changes(self):
584
"""Update the commit builder with the data about what has changed.
586
# Build the revision inventory.
588
# This starts by creating 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?
600
specific_files = self.specific_files
601
mutter("Selecting files for commit with filter %s", specific_files)
602
work_inv = self.work_tree.inventory
603
assert work_inv.root is not None
604
self.pb_entries_total = len(work_inv)
606
# Check and warn about old CommitBuilders
607
entries = work_inv.iter_entries()
608
if not self.builder.record_root_entry:
609
symbol_versioning.warn('CommitBuilders should support recording'
610
' the root entry as of bzr 0.10.', DeprecationWarning,
612
self.builder.new_inventory.add(self.basis_inv.root.copy())
616
deleted_paths = set()
617
for path, new_ie in entries:
618
self._emit_progress_next_entry()
619
file_id = new_ie.file_id
621
# Skip files that have been deleted from the working tree.
622
# The deleted files/directories are also recorded so they
623
# can be explicitly unversioned later. Note that when a
624
# filter of specific files is given, we must only skip/record
625
# deleted files matching that filter.
626
if is_inside_any(deleted_paths, path):
628
if not specific_files or is_inside_any(specific_files, path):
629
if not self.work_tree.has_filename(path):
630
deleted_paths.add(path)
631
self.reporter.missing(path)
632
deleted_ids.append(file_id)
635
kind = self.work_tree.kind(file_id)
636
if kind == 'tree-reference' and self.recursive == 'down':
637
# nested tree: commit in it
638
sub_tree = WorkingTree.open(self.work_tree.abspath(path))
639
# FIXME: be more comprehensive here:
640
# this works when both trees are in --trees repository,
641
# but when both are bound to a different repository,
642
# it fails; a better way of approaching this is to
643
# finally implement the explicit-caches approach design
644
# a while back - RBC 20070306.
645
if (sub_tree.branch.repository.bzrdir.root_transport.base
647
self.work_tree.branch.repository.bzrdir.root_transport.base):
648
sub_tree.branch.repository = \
649
self.work_tree.branch.repository
651
sub_tree.commit(message=None, revprops=self.revprops,
652
recursive=self.recursive,
653
message_callback=self.message_callback,
654
timestamp=self.timestamp, timezone=self.timezone,
655
committer=self.committer,
656
allow_pointless=self.allow_pointless,
657
strict=self.strict, verbose=self.verbose,
658
local=self.local, reporter=self.reporter)
659
except errors.PointlessCommit:
661
if kind != new_ie.kind:
662
new_ie = inventory.make_entry(kind, new_ie.name,
663
new_ie.parent_id, file_id)
664
except errors.NoSuchFile:
666
# mutter('check %s {%s}', path, file_id)
667
if (not specific_files or
668
is_inside_or_parent_of_any(specific_files, path)):
669
# mutter('%s selected for commit', path)
673
# mutter('%s not selected for commit', path)
674
if self.basis_inv.has_id(file_id):
675
ie = self.basis_inv[file_id].copy()
677
# this entry is new and not being committed
679
self.builder.record_entry_contents(ie, self.parent_invs,
680
path, self.work_tree)
681
# describe the nature of the change that has occurred relative to
682
# the basis inventory.
683
if (self.basis_inv.has_id(ie.file_id)):
684
basis_ie = self.basis_inv[ie.file_id]
687
change = ie.describe_change(basis_ie, ie)
688
if change in (InventoryEntry.RENAMED,
689
InventoryEntry.MODIFIED_AND_RENAMED):
690
old_path = self.basis_inv.id2path(ie.file_id)
691
self.reporter.renamed(change, old_path, path)
693
self.reporter.snapshot_change(change, path)
695
# Unversion IDs that were found to be deleted
696
self.work_tree.unversion(deleted_ids)
698
# If specific files/directories were nominated, it is possible
699
# that some data from outside those needs to be preserved from
700
# the basis tree. For example, if a file x is moved from out of
701
# directory foo into directory bar and the user requests
702
# ``commit foo``, then information about bar/x must also be
705
for path, new_ie in self.basis_inv.iter_entries():
706
if new_ie.file_id in work_inv:
708
if is_inside_any(specific_files, path):
712
self.builder.record_entry_contents(ie, self.parent_invs, path,
715
# Report what was deleted. We could skip this when no deletes are
716
# detected to gain a performance win, but it arguably serves as a
717
# 'safety check' by informing the user whenever anything disappears.
718
for path, ie in self.basis_inv.iter_entries():
719
if ie.file_id not in self.builder.new_inventory:
720
self.reporter.deleted(path)
722
def _emit_progress_set_stage(self, name, show_entries=False):
723
"""Set the progress stage and emit an update to the progress bar."""
724
self.pb_stage_name = name
725
self.pb_stage_count += 1
726
self.pb_entries_show = show_entries
728
self.pb_entries_count = 0
729
self.pb_entries_total = '?'
730
self._emit_progress()
732
def _emit_progress_next_entry(self):
733
"""Emit an update to the progress bar and increment the file count."""
734
self.pb_entries_count += 1
735
self._emit_progress()
737
def _emit_progress(self):
738
if self.pb_entries_show:
739
text = "%s [Entry %d/%s] - Stage" % (self.pb_stage_name,
740
self.pb_entries_count,str(self.pb_entries_total))
742
text = "%s - Stage" % (self.pb_stage_name)
743
self.pb.update(text, self.pb_stage_count, self.pb_stage_total)
added
removed
bzrlib/commit.py
