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
# this may be more useful if 'note' was replaced by an overridable
110
# method on self, which would allow more trivial subclassing.
111
# alternative, a callable could be passed in, allowing really trivial
112
# reuse for some uis. RBC 20060511
114
def snapshot_change(self, change, path):
115
if change == 'unchanged':
117
if change == 'added' and path == '':
119
note("%s %s", change, path)
121
def completed(self, revno, rev_id):
122
note('Committed revision %d.', revno)
124
def deleted(self, file_id):
125
note('deleted %s', file_id)
127
def escaped(self, escape_count, message):
128
note("replaced %d control characters in message", escape_count)
130
def missing(self, path):
131
note('missing %s', path)
133
def renamed(self, change, old_path, new_path):
134
note('%s %s => %s', change, old_path, new_path)
137
class Commit(object):
138
"""Task of committing a new revision.
140
This is a MethodObject: it accumulates state as the commit is
141
prepared, and then it is discarded. It doesn't represent
142
historical revisions, just the act of recording a new one.
145
Modified to hold a list of files that have been deleted from
146
the working directory; these should be removed from the
152
if reporter is not None:
153
self.reporter = reporter
155
self.reporter = NullCommitReporter()
156
if config is not None:
168
allow_pointless=True,
176
message_callback=None,
178
"""Commit working copy as a new revision.
180
message -- the commit message (it or message_callback is required)
182
timestamp -- if not None, seconds-since-epoch for a
183
postdated/predated commit.
185
specific_files -- If true, commit only those files.
187
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
allow_pointless -- If true (default), commit even if nothing
194
has changed and no merges are recorded.
196
strict -- If true, don't allow a commit if the working tree
197
contains unknown files.
199
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
236
self.specific_files = specific_files
238
self.verbose = verbose
241
if reporter is None and self.reporter is None:
242
self.reporter = NullCommitReporter()
243
elif reporter is not None:
244
self.reporter = reporter
246
self.work_tree.lock_write()
247
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
248
self.basis_tree = self.work_tree.basis_tree()
249
self.basis_tree.lock_read()
251
# Cannot commit with conflicts present.
252
if len(self.work_tree.conflicts())>0:
253
raise ConflictsInTree
255
# setup the bound branch variables as needed.
256
self._check_bound_branch()
258
# check for out of date working trees
260
first_tree_parent = self.work_tree.get_parent_ids()[0]
262
# if there are no parents, treat our parent as 'None'
263
# this is so that we still consier the master branch
264
# - in a checkout scenario the tree may have no
265
# parents but the branch may do.
266
first_tree_parent = bzrlib.revision.NULL_REVISION
267
old_revno, master_last = self.master_branch.last_revision_info()
268
if master_last != first_tree_parent:
269
if master_last != bzrlib.revision.NULL_REVISION:
270
raise errors.OutOfDateTree(self.work_tree)
271
if self.branch.repository.has_revision(first_tree_parent):
272
new_revno = old_revno + 1
274
# ghost parents never appear in revision history.
277
# raise an exception as soon as we find a single unknown.
278
for unknown in self.work_tree.unknowns():
279
raise StrictCommitFailed()
281
if self.config is None:
282
self.config = self.branch.get_config()
284
self.work_inv = self.work_tree.inventory
285
self.basis_inv = self.basis_tree.inventory
286
if specific_files is not None:
287
# Ensure specified files are versioned
288
# (We don't actually need the ids here)
289
# XXX: Dont we have filter_unversioned to do this more
291
tree.find_ids_across_trees(specific_files,
292
[self.basis_tree, self.work_tree])
293
# one to finish, one for rev and inventory, and one for each
294
# inventory entry, and the same for the new inventory.
295
# note that this estimate is too long when we do a partial tree
296
# commit which excludes some new files from being considered.
297
# The estimate is corrected when we populate the new inv.
298
self.pb_total = len(self.work_inv) + 5
301
self._gather_parents()
302
if len(self.parents) > 1 and self.specific_files:
303
raise NotImplementedError('selected-file commit of merges is not supported yet: files %r',
306
self.builder = self.branch.get_commit_builder(self.parents,
307
self.config, timestamp, timezone, committer, revprops, rev_id)
309
self._remove_deleted()
310
self._populate_new_inv()
311
self._report_deletes()
313
self._check_pointless()
315
self._emit_progress_update()
316
# TODO: Now the new inventory is known, check for conflicts and
317
# prompt the user for a commit message.
318
# ADHB 2006-08-08: If this is done, populate_new_inv should not add
319
# weave lines, because nothing should be recorded until it is known
320
# that commit will succeed.
321
self.builder.finish_inventory()
322
self._emit_progress_update()
323
message = message_callback(self)
324
assert isinstance(message, unicode), type(message)
325
self.message = message
326
self._escape_commit_message()
328
self.rev_id = self.builder.commit(self.message)
329
self._emit_progress_update()
330
# revision data is in the local branch now.
332
# upload revision data to the master.
333
# this will propagate merged revisions too if needed.
334
if self.bound_branch:
335
self.master_branch.repository.fetch(self.branch.repository,
336
revision_id=self.rev_id)
337
# now the master has the revision data
338
# 'commit' to the master first so a timeout here causes the local
339
# branch to be out of date
340
self.master_branch.set_last_revision_info(new_revno,
343
# and now do the commit locally.
344
self.branch.set_last_revision_info(new_revno, self.rev_id)
346
rev_tree = self.builder.revision_tree()
347
self.work_tree.set_parent_trees([(self.rev_id, rev_tree)])
348
# now the work tree is up to date with the branch
350
self.reporter.completed(new_revno, self.rev_id)
351
# old style commit hooks - should be deprecated ? (obsoleted in
353
if self.config.post_commit() is not None:
354
hooks = self.config.post_commit().split(' ')
355
# this would be nicer with twisted.python.reflect.namedAny
357
result = eval(hook + '(branch, rev_id)',
358
{'branch':self.branch,
360
'rev_id':self.rev_id})
361
# new style commit hooks:
362
if not self.bound_branch:
363
hook_master = self.branch
366
hook_master = self.master_branch
367
hook_local = self.branch
368
# With bound branches, when the master is behind the local branch,
369
# the 'old_revno' and old_revid values here are incorrect.
370
# XXX: FIXME ^. RBC 20060206
372
old_revid = self.parents[0]
374
old_revid = bzrlib.revision.NULL_REVISION
375
for hook in Branch.hooks['post_commit']:
376
hook(hook_local, hook_master, old_revno, old_revid, new_revno,
378
self._emit_progress_update()
383
def _any_real_changes(self):
384
"""Are there real changes between new_inventory and basis?
386
For trees without rich roots, inv.root.revision changes every commit.
387
But if that is the only change, we want to treat it as though there
390
new_entries = self.builder.new_inventory.iter_entries()
391
basis_entries = self.basis_inv.iter_entries()
392
new_path, new_root_ie = new_entries.next()
393
basis_path, basis_root_ie = basis_entries.next()
395
# This is a copy of InventoryEntry.__eq__ only leaving out .revision
396
def ie_equal_no_revision(this, other):
397
return ((this.file_id == other.file_id)
398
and (this.name == other.name)
399
and (this.symlink_target == other.symlink_target)
400
and (this.text_sha1 == other.text_sha1)
401
and (this.text_size == other.text_size)
402
and (this.text_id == other.text_id)
403
and (this.parent_id == other.parent_id)
404
and (this.kind == other.kind)
405
and (this.executable == other.executable)
406
and (this.reference_revision == other.reference_revision)
408
if not ie_equal_no_revision(new_root_ie, basis_root_ie):
411
for new_ie, basis_ie in zip(new_entries, basis_entries):
412
if new_ie != basis_ie:
415
# No actual changes present
418
def _check_pointless(self):
419
if self.allow_pointless:
421
# A merge with no effect on files
422
if len(self.parents) > 1:
424
# work around the fact that a newly-initted tree does differ from its
426
if len(self.basis_inv) == 0 and len(self.builder.new_inventory) == 1:
427
raise PointlessCommit()
428
# Shortcut, if the number of entries changes, then we obviously have
430
if len(self.builder.new_inventory) != len(self.basis_inv):
432
# If length == 1, then we only have the root entry. Which means
433
# that there is no real difference (only the root could be different)
434
if (len(self.builder.new_inventory) != 1 and self._any_real_changes()):
436
raise PointlessCommit()
438
def _check_bound_branch(self):
439
"""Check to see if the local branch is bound.
441
If it is bound, then most of the commit will actually be
442
done using the remote branch as the target branch.
443
Only at the end will the local branch be updated.
445
if self.local and not self.branch.get_bound_location():
446
raise errors.LocalRequiresBoundBranch()
449
self.master_branch = self.branch.get_master_branch()
451
if not self.master_branch:
452
# make this branch the reference branch for out of date checks.
453
self.master_branch = self.branch
456
# If the master branch is bound, we must fail
457
master_bound_location = self.master_branch.get_bound_location()
458
if master_bound_location:
459
raise errors.CommitToDoubleBoundBranch(self.branch,
460
self.master_branch, master_bound_location)
462
# TODO: jam 20051230 We could automatically push local
463
# commits to the remote branch if they would fit.
464
# But for now, just require remote to be identical
467
# Make sure the local branch is identical to the master
468
master_info = self.master_branch.last_revision_info()
469
local_info = self.branch.last_revision_info()
470
if local_info != master_info:
471
raise errors.BoundBranchOutOfDate(self.branch,
474
# Now things are ready to change the master branch
476
self.bound_branch = self.branch
477
self.master_branch.lock_write()
478
self.master_locked = True
481
"""Cleanup any open locks, progress bars etc."""
482
cleanups = [self._cleanup_bound_branch,
483
self.basis_tree.unlock,
484
self.work_tree.unlock,
486
found_exception = None
487
for cleanup in cleanups:
490
# we want every cleanup to run no matter what.
491
# so we have a catchall here, but we will raise the
492
# last encountered exception up the stack: and
493
# typically this will be useful enough.
496
if found_exception is not None:
497
# don't do a plan raise, because the last exception may have been
498
# trashed, e is our sure-to-work exception even though it loses the
499
# full traceback. XXX: RBC 20060421 perhaps we could check the
500
# exc_info and if its the same one do a plain raise otherwise
501
# 'raise e' as we do now.
504
def _cleanup_bound_branch(self):
505
"""Executed at the end of a try/finally to cleanup a bound branch.
507
If the branch wasn't bound, this is a no-op.
508
If it was, it resents self.branch to the local branch, instead
511
if not self.bound_branch:
513
if self.master_locked:
514
self.master_branch.unlock()
516
def _escape_commit_message(self):
517
"""Replace xml-incompatible control characters."""
518
# FIXME: RBC 20060419 this should be done by the revision
519
# serialiser not by commit. Then we can also add an unescaper
520
# in the deserializer and start roundtripping revision messages
521
# precisely. See repository_implementations/test_repository.py
523
# Python strings can include characters that can't be
524
# represented in well-formed XML; escape characters that
525
# aren't listed in the XML specification
526
# (http://www.w3.org/TR/REC-xml/#NT-Char).
527
self.message, escape_count = re.subn(
528
u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
529
lambda match: match.group(0).encode('unicode_escape'),
532
self.reporter.escaped(escape_count, self.message)
534
def _gather_parents(self):
535
"""Record the parents of a merge for merge detection."""
536
# TODO: Make sure that this list doesn't contain duplicate
537
# entries and the order is preserved when doing this.
538
self.parents = self.work_tree.get_parent_ids()
539
self.parent_invs = [self.basis_inv]
540
for revision in self.parents[1:]:
541
if self.branch.repository.has_revision(revision):
542
mutter('commit parent revision {%s}', revision)
543
inventory = self.branch.repository.get_inventory(revision)
544
self.parent_invs.append(inventory)
546
mutter('commit parent ghost revision {%s}', revision)
548
def _remove_deleted(self):
549
"""Remove deleted files from the working inventories.
551
This is done prior to taking the working inventory as the
552
basis for the new committed inventory.
554
This returns true if any files
555
*that existed in the basis inventory* were deleted.
556
Files that were added and deleted
557
in the working copy don't matter.
559
specific = self.specific_files
561
deleted_paths = set()
562
for path, ie in self.work_inv.iter_entries():
563
if is_inside_any(deleted_paths, path):
564
# The tree will delete the required ids recursively.
566
if specific and not is_inside_any(specific, path):
568
if not self.work_tree.has_filename(path):
569
deleted_paths.add(path)
570
self.reporter.missing(path)
571
deleted_ids.append(ie.file_id)
572
self.work_tree.unversion(deleted_ids)
574
def _populate_new_inv(self):
575
"""Build revision inventory.
577
This creates a new empty inventory. Depending on
578
which files are selected for commit, and what is present in the
579
current tree, the new inventory is populated. inventory entries
580
which are candidates for modification have their revision set to
581
None; inventory entries that are carried over untouched have their
582
revision set to their prior value.
584
# ESEPARATIONOFCONCERNS: this function is diffing and using the diff
585
# results to create a new inventory at the same time, which results
586
# in bugs like #46635. Any reason not to use/enhance Tree.changes_from?
588
mutter("Selecting files for commit with filter %s", self.specific_files)
589
assert self.work_inv.root is not None
590
entries = self.work_inv.iter_entries()
591
if not self.builder.record_root_entry:
592
symbol_versioning.warn('CommitBuilders should support recording'
593
' the root entry as of bzr 0.10.', DeprecationWarning,
595
self.builder.new_inventory.add(self.basis_inv.root.copy())
597
self._emit_progress_update()
598
for path, new_ie in entries:
599
self._emit_progress_update()
600
file_id = new_ie.file_id
602
kind = self.work_tree.kind(file_id)
603
if kind == 'tree-reference' and self.recursive == 'down':
604
# nested tree: commit in it
605
sub_tree = WorkingTree.open(self.work_tree.abspath(path))
606
# FIXME: be more comprehensive here:
607
# this works when both trees are in --trees repository,
608
# but when both are bound to a different repository,
609
# it fails; a better way of approaching this is to
610
# finally implement the explicit-caches approach design
611
# a while back - RBC 20070306.
612
if (sub_tree.branch.repository.bzrdir.root_transport.base
614
self.work_tree.branch.repository.bzrdir.root_transport.base):
615
sub_tree.branch.repository = \
616
self.work_tree.branch.repository
618
sub_tree.commit(message=None, revprops=self.revprops,
619
recursive=self.recursive,
620
message_callback=self.message_callback,
621
timestamp=self.timestamp, timezone=self.timezone,
622
committer=self.committer,
623
allow_pointless=self.allow_pointless,
624
strict=self.strict, verbose=self.verbose,
625
local=self.local, reporter=self.reporter)
626
except errors.PointlessCommit:
628
if kind != new_ie.kind:
629
new_ie = inventory.make_entry(kind, new_ie.name,
630
new_ie.parent_id, file_id)
631
except errors.NoSuchFile:
633
# mutter('check %s {%s}', path, file_id)
634
if (not self.specific_files or
635
is_inside_or_parent_of_any(self.specific_files, path)):
636
# mutter('%s selected for commit', path)
640
# mutter('%s not selected for commit', path)
641
if self.basis_inv.has_id(file_id):
642
ie = self.basis_inv[file_id].copy()
644
# this entry is new and not being committed
646
self.builder.record_entry_contents(ie, self.parent_invs,
647
path, self.work_tree)
648
# describe the nature of the change that has occurred relative to
649
# the basis inventory.
650
if (self.basis_inv.has_id(ie.file_id)):
651
basis_ie = self.basis_inv[ie.file_id]
654
change = ie.describe_change(basis_ie, ie)
655
if change in (InventoryEntry.RENAMED,
656
InventoryEntry.MODIFIED_AND_RENAMED):
657
old_path = self.basis_inv.id2path(ie.file_id)
658
self.reporter.renamed(change, old_path, path)
660
self.reporter.snapshot_change(change, path)
662
if not self.specific_files:
665
# ignore removals that don't match filespec
666
for path, new_ie in self.basis_inv.iter_entries():
667
if new_ie.file_id in self.work_inv:
669
if is_inside_any(self.specific_files, path):
673
self.builder.record_entry_contents(ie, self.parent_invs, path,
676
def _emit_progress_update(self):
677
"""Emit an update to the progress bar."""
678
self.pb.update("Committing", self.pb_count, self.pb_total)
681
def _report_deletes(self):
682
for path, ie in self.basis_inv.iter_entries():
683
if ie.file_id not in self.builder.new_inventory:
684
self.reporter.deleted(path)
added
removed
bzrlib/commit.py
