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
237
self.verbose = verbose
240
if reporter is None and self.reporter is None:
241
self.reporter = NullCommitReporter()
242
elif reporter is not None:
243
self.reporter = reporter
245
self.work_tree.lock_write()
246
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
247
self.basis_tree = self.work_tree.basis_tree()
248
self.basis_tree.lock_read()
250
# Cannot commit with conflicts present.
251
if len(self.work_tree.conflicts())>0:
252
raise ConflictsInTree
254
# setup the bound branch variables as needed.
255
self._check_bound_branch()
257
# check for out of date working trees
259
first_tree_parent = self.work_tree.get_parent_ids()[0]
261
# if there are no parents, treat our parent as 'None'
262
# this is so that we still consier the master branch
263
# - in a checkout scenario the tree may have no
264
# parents but the branch may do.
265
first_tree_parent = bzrlib.revision.NULL_REVISION
266
old_revno, master_last = self.master_branch.last_revision_info()
267
if master_last != first_tree_parent:
268
if master_last != bzrlib.revision.NULL_REVISION:
269
raise errors.OutOfDateTree(self.work_tree)
270
if self.branch.repository.has_revision(first_tree_parent):
271
new_revno = old_revno + 1
273
# ghost parents never appear in revision history.
276
# raise an exception as soon as we find a single unknown.
277
for unknown in self.work_tree.unknowns():
278
raise StrictCommitFailed()
280
if self.config is None:
281
self.config = self.branch.get_config()
283
self.work_inv = self.work_tree.inventory
284
self.basis_inv = self.basis_tree.inventory
285
if specific_files is not None:
286
# Ensure specified files are versioned
287
# (We don't actually need the ids here)
288
# XXX: Dont we have filter_unversioned to do this more
290
tree.find_ids_across_trees(specific_files,
291
[self.basis_tree, self.work_tree])
292
# one to finish, one for rev and inventory, and one for each
293
# inventory entry, and the same for the new inventory.
294
# note that this estimate is too long when we do a partial tree
295
# commit which excludes some new files from being considered.
296
# The estimate is corrected when we populate the new inv.
297
self.pb_total = len(self.work_inv) + 5
300
self._gather_parents()
301
if len(self.parents) > 1 and self.specific_files:
302
raise NotImplementedError('selected-file commit of merges is not supported yet: files %r',
305
self.builder = self.branch.get_commit_builder(self.parents,
306
self.config, timestamp, timezone, committer, revprops, rev_id)
308
self._remove_deleted()
309
self._populate_new_inv()
310
self._report_deletes()
312
self._check_pointless()
314
self._emit_progress_update()
315
# TODO: Now the new inventory is known, check for conflicts and
316
# prompt the user for a commit message.
317
# ADHB 2006-08-08: If this is done, populate_new_inv should not add
318
# weave lines, because nothing should be recorded until it is known
319
# that commit will succeed.
320
self.builder.finish_inventory()
321
self._emit_progress_update()
322
message = message_callback(self)
323
assert isinstance(message, unicode), type(message)
324
self.message = message
325
self._escape_commit_message()
327
self.rev_id = self.builder.commit(self.message)
328
self._emit_progress_update()
329
# revision data is in the local branch now.
331
# upload revision data to the master.
332
# this will propagate merged revisions too if needed.
333
if self.bound_branch:
334
self.master_branch.repository.fetch(self.branch.repository,
335
revision_id=self.rev_id)
336
# now the master has the revision data
337
# 'commit' to the master first so a timeout here causes the local
338
# branch to be out of date
339
self.master_branch.set_last_revision_info(new_revno,
342
# and now do the commit locally.
343
self.branch.set_last_revision_info(new_revno, self.rev_id)
345
rev_tree = self.builder.revision_tree()
346
self.work_tree.set_parent_trees([(self.rev_id, rev_tree)])
347
# now the work tree is up to date with the branch
349
self.reporter.completed(new_revno, self.rev_id)
350
# old style commit hooks - should be deprecated ? (obsoleted in
352
if self.config.post_commit() is not None:
353
hooks = self.config.post_commit().split(' ')
354
# this would be nicer with twisted.python.reflect.namedAny
356
result = eval(hook + '(branch, rev_id)',
357
{'branch':self.branch,
359
'rev_id':self.rev_id})
360
# new style commit hooks:
361
if not self.bound_branch:
362
hook_master = self.branch
365
hook_master = self.master_branch
366
hook_local = self.branch
367
# With bound branches, when the master is behind the local branch,
368
# the 'old_revno' and old_revid values here are incorrect.
369
# XXX: FIXME ^. RBC 20060206
371
old_revid = self.parents[0]
373
old_revid = bzrlib.revision.NULL_REVISION
374
for hook in Branch.hooks['post_commit']:
375
hook(hook_local, hook_master, old_revno, old_revid, new_revno,
377
self._emit_progress_update()
382
def _any_real_changes(self):
383
"""Are there real changes between new_inventory and basis?
385
For trees without rich roots, inv.root.revision changes every commit.
386
But if that is the only change, we want to treat it as though there
389
new_entries = self.builder.new_inventory.iter_entries()
390
basis_entries = self.basis_inv.iter_entries()
391
new_path, new_root_ie = new_entries.next()
392
basis_path, basis_root_ie = basis_entries.next()
394
# This is a copy of InventoryEntry.__eq__ only leaving out .revision
395
def ie_equal_no_revision(this, other):
396
return ((this.file_id == other.file_id)
397
and (this.name == other.name)
398
and (this.symlink_target == other.symlink_target)
399
and (this.text_sha1 == other.text_sha1)
400
and (this.text_size == other.text_size)
401
and (this.text_id == other.text_id)
402
and (this.parent_id == other.parent_id)
403
and (this.kind == other.kind)
404
and (this.executable == other.executable)
405
and (this.reference_revision == other.reference_revision)
407
if not ie_equal_no_revision(new_root_ie, basis_root_ie):
410
for new_ie, basis_ie in zip(new_entries, basis_entries):
411
if new_ie != basis_ie:
414
# No actual changes present
417
def _check_pointless(self):
418
if self.allow_pointless:
420
# A merge with no effect on files
421
if len(self.parents) > 1:
423
# work around the fact that a newly-initted tree does differ from its
425
if len(self.basis_inv) == 0 and len(self.builder.new_inventory) == 1:
426
raise PointlessCommit()
427
# Shortcut, if the number of entries changes, then we obviously have
429
if len(self.builder.new_inventory) != len(self.basis_inv):
431
# If length == 1, then we only have the root entry. Which means
432
# that there is no real difference (only the root could be different)
433
if (len(self.builder.new_inventory) != 1 and self._any_real_changes()):
435
raise PointlessCommit()
437
def _check_bound_branch(self):
438
"""Check to see if the local branch is bound.
440
If it is bound, then most of the commit will actually be
441
done using the remote branch as the target branch.
442
Only at the end will the local branch be updated.
444
if self.local and not self.branch.get_bound_location():
445
raise errors.LocalRequiresBoundBranch()
448
self.master_branch = self.branch.get_master_branch()
450
if not self.master_branch:
451
# make this branch the reference branch for out of date checks.
452
self.master_branch = self.branch
455
# If the master branch is bound, we must fail
456
master_bound_location = self.master_branch.get_bound_location()
457
if master_bound_location:
458
raise errors.CommitToDoubleBoundBranch(self.branch,
459
self.master_branch, master_bound_location)
461
# TODO: jam 20051230 We could automatically push local
462
# commits to the remote branch if they would fit.
463
# But for now, just require remote to be identical
466
# Make sure the local branch is identical to the master
467
master_info = self.master_branch.last_revision_info()
468
local_info = self.branch.last_revision_info()
469
if local_info != master_info:
470
raise errors.BoundBranchOutOfDate(self.branch,
473
# Now things are ready to change the master branch
475
self.bound_branch = self.branch
476
self.master_branch.lock_write()
477
self.master_locked = True
480
"""Cleanup any open locks, progress bars etc."""
481
cleanups = [self._cleanup_bound_branch,
482
self.basis_tree.unlock,
483
self.work_tree.unlock,
485
found_exception = None
486
for cleanup in cleanups:
489
# we want every cleanup to run no matter what.
490
# so we have a catchall here, but we will raise the
491
# last encountered exception up the stack: and
492
# typically this will be useful enough.
495
if found_exception is not None:
496
# don't do a plan raise, because the last exception may have been
497
# trashed, e is our sure-to-work exception even though it loses the
498
# full traceback. XXX: RBC 20060421 perhaps we could check the
499
# exc_info and if its the same one do a plain raise otherwise
500
# 'raise e' as we do now.
503
def _cleanup_bound_branch(self):
504
"""Executed at the end of a try/finally to cleanup a bound branch.
506
If the branch wasn't bound, this is a no-op.
507
If it was, it resents self.branch to the local branch, instead
510
if not self.bound_branch:
512
if self.master_locked:
513
self.master_branch.unlock()
515
def _escape_commit_message(self):
516
"""Replace xml-incompatible control characters."""
517
# FIXME: RBC 20060419 this should be done by the revision
518
# serialiser not by commit. Then we can also add an unescaper
519
# in the deserializer and start roundtripping revision messages
520
# precisely. See repository_implementations/test_repository.py
522
# Python strings can include characters that can't be
523
# represented in well-formed XML; escape characters that
524
# aren't listed in the XML specification
525
# (http://www.w3.org/TR/REC-xml/#NT-Char).
526
self.message, escape_count = re.subn(
527
u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
528
lambda match: match.group(0).encode('unicode_escape'),
531
self.reporter.escaped(escape_count, self.message)
533
def _gather_parents(self):
534
"""Record the parents of a merge for merge detection."""
535
# TODO: Make sure that this list doesn't contain duplicate
536
# entries and the order is preserved when doing this.
537
self.parents = self.work_tree.get_parent_ids()
538
self.parent_invs = [self.basis_inv]
539
for revision in self.parents[1:]:
540
if self.branch.repository.has_revision(revision):
541
mutter('commit parent revision {%s}', revision)
542
inventory = self.branch.repository.get_inventory(revision)
543
self.parent_invs.append(inventory)
545
mutter('commit parent ghost revision {%s}', revision)
547
def _remove_deleted(self):
548
"""Remove deleted files from the working inventories.
550
This is done prior to taking the working inventory as the
551
basis for the new committed inventory.
553
This returns true if any files
554
*that existed in the basis inventory* were deleted.
555
Files that were added and deleted
556
in the working copy don't matter.
558
specific = self.specific_files
560
deleted_paths = set()
561
for path, ie in self.work_inv.iter_entries():
562
if is_inside_any(deleted_paths, path):
563
# The tree will delete the required ids recursively.
565
if specific and not is_inside_any(specific, path):
567
if not self.work_tree.has_filename(path):
568
deleted_paths.add(path)
569
self.reporter.missing(path)
570
deleted_ids.append(ie.file_id)
571
self.work_tree.unversion(deleted_ids)
573
def _populate_new_inv(self):
574
"""Build revision inventory.
576
This creates a new empty inventory. Depending on
577
which files are selected for commit, and what is present in the
578
current tree, the new inventory is populated. inventory entries
579
which are candidates for modification have their revision set to
580
None; inventory entries that are carried over untouched have their
581
revision set to their prior value.
583
# ESEPARATIONOFCONCERNS: this function is diffing and using the diff
584
# results to create a new inventory at the same time, which results
585
# in bugs like #46635. Any reason not to use/enhance Tree.changes_from?
587
mutter("Selecting files for commit with filter %s", self.specific_files)
588
assert self.work_inv.root is not None
589
entries = self.work_inv.iter_entries()
590
if not self.builder.record_root_entry:
591
symbol_versioning.warn('CommitBuilders should support recording'
592
' the root entry as of bzr 0.10.', DeprecationWarning,
594
self.builder.new_inventory.add(self.basis_inv.root.copy())
596
self._emit_progress_update()
597
for path, new_ie in entries:
598
self._emit_progress_update()
599
file_id = new_ie.file_id
601
kind = self.work_tree.kind(file_id)
602
if kind == 'tree-reference' and self.recursive == 'down':
603
# nested tree: commit in it
604
sub_tree = WorkingTree.open(self.work_tree.abspath(path))
605
# FIXME: be more comprehensive here:
606
# this works when both trees are in --trees repository,
607
# but when both are bound to a different repository,
608
# it fails; a better way of approaching this is to
609
# finally implement the explicit-caches approach design
610
# a while back - RBC 20070306.
611
if (sub_tree.branch.repository.bzrdir.root_transport.base
613
self.work_tree.branch.repository.bzrdir.root_transport.base):
614
sub_tree.branch.repository = \
615
self.work_tree.branch.repository
617
sub_tree.commit(message=None, revprops=self.revprops,
618
recursive=self.recursive,
619
message_callback=self.message_callback,
620
timestamp=self.timestamp, timezone=self.timezone,
621
committer=self.committer,
622
allow_pointless=self.allow_pointless,
623
strict=self.strict, verbose=self.verbose,
624
local=self.local, reporter=self.reporter)
625
except errors.PointlessCommit:
627
if kind != new_ie.kind:
628
new_ie = inventory.make_entry(kind, new_ie.name,
629
new_ie.parent_id, file_id)
630
except errors.NoSuchFile:
632
# mutter('check %s {%s}', path, file_id)
633
if (not self.specific_files or
634
is_inside_or_parent_of_any(self.specific_files, path)):
635
# mutter('%s selected for commit', path)
639
# mutter('%s not selected for commit', path)
640
if self.basis_inv.has_id(file_id):
641
ie = self.basis_inv[file_id].copy()
643
# this entry is new and not being committed
645
self.builder.record_entry_contents(ie, self.parent_invs,
646
path, self.work_tree)
647
# describe the nature of the change that has occurred relative to
648
# the basis inventory.
649
if (self.basis_inv.has_id(ie.file_id)):
650
basis_ie = self.basis_inv[ie.file_id]
653
change = ie.describe_change(basis_ie, ie)
654
if change in (InventoryEntry.RENAMED,
655
InventoryEntry.MODIFIED_AND_RENAMED):
656
old_path = self.basis_inv.id2path(ie.file_id)
657
self.reporter.renamed(change, old_path, path)
659
self.reporter.snapshot_change(change, path)
661
if not self.specific_files:
664
# ignore removals that don't match filespec
665
for path, new_ie in self.basis_inv.iter_entries():
666
if new_ie.file_id in self.work_inv:
668
if is_inside_any(self.specific_files, path):
672
self.builder.record_entry_contents(ie, self.parent_invs, path,
675
def _emit_progress_update(self):
676
"""Emit an update to the progress bar."""
677
self.pb.update("Committing", self.pb_count, self.pb_total)
680
def _report_deletes(self):
681
for path, ie in self.basis_inv.iter_entries():
682
if ie.file_id not in self.builder.new_inventory:
683
self.reporter.deleted(path)
added
removed
bzrlib/commit.py
