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 errors.CannotCommitSelectedFileMerge(self.specific_files)
304
self.builder = self.branch.get_commit_builder(self.parents,
305
self.config, timestamp, timezone, committer, revprops, rev_id)
307
self._remove_deleted()
308
self._populate_new_inv()
309
self._report_deletes()
311
self._check_pointless()
313
self._emit_progress_update()
314
# TODO: Now the new inventory is known, check for conflicts and
315
# prompt the user for a commit message.
316
# ADHB 2006-08-08: If this is done, populate_new_inv should not add
317
# weave lines, because nothing should be recorded until it is known
318
# that commit will succeed.
319
self.builder.finish_inventory()
320
self._emit_progress_update()
321
message = message_callback(self)
322
assert isinstance(message, unicode), type(message)
323
self.message = message
324
self._escape_commit_message()
326
self.rev_id = self.builder.commit(self.message)
327
self._emit_progress_update()
328
# revision data is in the local branch now.
330
# upload revision data to the master.
331
# this will propagate merged revisions too if needed.
332
if self.bound_branch:
333
self.master_branch.repository.fetch(self.branch.repository,
334
revision_id=self.rev_id)
335
# now the master has the revision data
336
# 'commit' to the master first so a timeout here causes the local
337
# branch to be out of date
338
self.master_branch.set_last_revision_info(new_revno,
341
# and now do the commit locally.
342
self.branch.set_last_revision_info(new_revno, self.rev_id)
344
rev_tree = self.builder.revision_tree()
345
self.work_tree.set_parent_trees([(self.rev_id, rev_tree)])
346
# now the work tree is up to date with the branch
348
self.reporter.completed(new_revno, self.rev_id)
349
# old style commit hooks - should be deprecated ? (obsoleted in
351
if self.config.post_commit() is not None:
352
hooks = self.config.post_commit().split(' ')
353
# this would be nicer with twisted.python.reflect.namedAny
355
result = eval(hook + '(branch, rev_id)',
356
{'branch':self.branch,
358
'rev_id':self.rev_id})
359
# new style commit hooks:
360
if not self.bound_branch:
361
hook_master = self.branch
364
hook_master = self.master_branch
365
hook_local = self.branch
366
# With bound branches, when the master is behind the local branch,
367
# the 'old_revno' and old_revid values here are incorrect.
368
# XXX: FIXME ^. RBC 20060206
370
old_revid = self.parents[0]
372
old_revid = bzrlib.revision.NULL_REVISION
373
for hook in Branch.hooks['post_commit']:
374
hook(hook_local, hook_master, old_revno, old_revid, new_revno,
376
self._emit_progress_update()
381
def _any_real_changes(self):
382
"""Are there real changes between new_inventory and basis?
384
For trees without rich roots, inv.root.revision changes every commit.
385
But if that is the only change, we want to treat it as though there
388
new_entries = self.builder.new_inventory.iter_entries()
389
basis_entries = self.basis_inv.iter_entries()
390
new_path, new_root_ie = new_entries.next()
391
basis_path, basis_root_ie = basis_entries.next()
393
# This is a copy of InventoryEntry.__eq__ only leaving out .revision
394
def ie_equal_no_revision(this, other):
395
return ((this.file_id == other.file_id)
396
and (this.name == other.name)
397
and (this.symlink_target == other.symlink_target)
398
and (this.text_sha1 == other.text_sha1)
399
and (this.text_size == other.text_size)
400
and (this.text_id == other.text_id)
401
and (this.parent_id == other.parent_id)
402
and (this.kind == other.kind)
403
and (this.executable == other.executable)
404
and (this.reference_revision == other.reference_revision)
406
if not ie_equal_no_revision(new_root_ie, basis_root_ie):
409
for new_ie, basis_ie in zip(new_entries, basis_entries):
410
if new_ie != basis_ie:
413
# No actual changes present
416
def _check_pointless(self):
417
if self.allow_pointless:
419
# A merge with no effect on files
420
if len(self.parents) > 1:
422
# work around the fact that a newly-initted tree does differ from its
424
if len(self.basis_inv) == 0 and len(self.builder.new_inventory) == 1:
425
raise PointlessCommit()
426
# Shortcut, if the number of entries changes, then we obviously have
428
if len(self.builder.new_inventory) != len(self.basis_inv):
430
# If length == 1, then we only have the root entry. Which means
431
# that there is no real difference (only the root could be different)
432
if (len(self.builder.new_inventory) != 1 and self._any_real_changes()):
434
raise PointlessCommit()
436
def _check_bound_branch(self):
437
"""Check to see if the local branch is bound.
439
If it is bound, then most of the commit will actually be
440
done using the remote branch as the target branch.
441
Only at the end will the local branch be updated.
443
if self.local and not self.branch.get_bound_location():
444
raise errors.LocalRequiresBoundBranch()
447
self.master_branch = self.branch.get_master_branch()
449
if not self.master_branch:
450
# make this branch the reference branch for out of date checks.
451
self.master_branch = self.branch
454
# If the master branch is bound, we must fail
455
master_bound_location = self.master_branch.get_bound_location()
456
if master_bound_location:
457
raise errors.CommitToDoubleBoundBranch(self.branch,
458
self.master_branch, master_bound_location)
460
# TODO: jam 20051230 We could automatically push local
461
# commits to the remote branch if they would fit.
462
# But for now, just require remote to be identical
465
# Make sure the local branch is identical to the master
466
master_info = self.master_branch.last_revision_info()
467
local_info = self.branch.last_revision_info()
468
if local_info != master_info:
469
raise errors.BoundBranchOutOfDate(self.branch,
472
# Now things are ready to change the master branch
474
self.bound_branch = self.branch
475
self.master_branch.lock_write()
476
self.master_locked = True
479
"""Cleanup any open locks, progress bars etc."""
480
cleanups = [self._cleanup_bound_branch,
481
self.basis_tree.unlock,
482
self.work_tree.unlock,
484
found_exception = None
485
for cleanup in cleanups:
488
# we want every cleanup to run no matter what.
489
# so we have a catchall here, but we will raise the
490
# last encountered exception up the stack: and
491
# typically this will be useful enough.
494
if found_exception is not None:
495
# don't do a plan raise, because the last exception may have been
496
# trashed, e is our sure-to-work exception even though it loses the
497
# full traceback. XXX: RBC 20060421 perhaps we could check the
498
# exc_info and if its the same one do a plain raise otherwise
499
# 'raise e' as we do now.
502
def _cleanup_bound_branch(self):
503
"""Executed at the end of a try/finally to cleanup a bound branch.
505
If the branch wasn't bound, this is a no-op.
506
If it was, it resents self.branch to the local branch, instead
509
if not self.bound_branch:
511
if self.master_locked:
512
self.master_branch.unlock()
514
def _escape_commit_message(self):
515
"""Replace xml-incompatible control characters."""
516
# FIXME: RBC 20060419 this should be done by the revision
517
# serialiser not by commit. Then we can also add an unescaper
518
# in the deserializer and start roundtripping revision messages
519
# precisely. See repository_implementations/test_repository.py
521
# Python strings can include characters that can't be
522
# represented in well-formed XML; escape characters that
523
# aren't listed in the XML specification
524
# (http://www.w3.org/TR/REC-xml/#NT-Char).
525
self.message, escape_count = re.subn(
526
u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
527
lambda match: match.group(0).encode('unicode_escape'),
530
self.reporter.escaped(escape_count, self.message)
532
def _gather_parents(self):
533
"""Record the parents of a merge for merge detection."""
534
# TODO: Make sure that this list doesn't contain duplicate
535
# entries and the order is preserved when doing this.
536
self.parents = self.work_tree.get_parent_ids()
537
self.parent_invs = [self.basis_inv]
538
for revision in self.parents[1:]:
539
if self.branch.repository.has_revision(revision):
540
mutter('commit parent revision {%s}', revision)
541
inventory = self.branch.repository.get_inventory(revision)
542
self.parent_invs.append(inventory)
544
mutter('commit parent ghost revision {%s}', revision)
546
def _remove_deleted(self):
547
"""Remove deleted files from the working inventories.
549
This is done prior to taking the working inventory as the
550
basis for the new committed inventory.
552
This returns true if any files
553
*that existed in the basis inventory* were deleted.
554
Files that were added and deleted
555
in the working copy don't matter.
557
specific = self.specific_files
559
deleted_paths = set()
560
for path, ie in self.work_inv.iter_entries():
561
if is_inside_any(deleted_paths, path):
562
# The tree will delete the required ids recursively.
564
if specific and not is_inside_any(specific, path):
566
if not self.work_tree.has_filename(path):
567
deleted_paths.add(path)
568
self.reporter.missing(path)
569
deleted_ids.append(ie.file_id)
570
self.work_tree.unversion(deleted_ids)
572
def _populate_new_inv(self):
573
"""Build revision inventory.
575
This creates a new empty inventory. Depending on
576
which files are selected for commit, and what is present in the
577
current tree, the new inventory is populated. inventory entries
578
which are candidates for modification have their revision set to
579
None; inventory entries that are carried over untouched have their
580
revision set to their prior value.
582
# ESEPARATIONOFCONCERNS: this function is diffing and using the diff
583
# results to create a new inventory at the same time, which results
584
# in bugs like #46635. Any reason not to use/enhance Tree.changes_from?
586
mutter("Selecting files for commit with filter %s", self.specific_files)
587
assert self.work_inv.root is not None
588
entries = self.work_inv.iter_entries()
589
if not self.builder.record_root_entry:
590
symbol_versioning.warn('CommitBuilders should support recording'
591
' the root entry as of bzr 0.10.', DeprecationWarning,
593
self.builder.new_inventory.add(self.basis_inv.root.copy())
595
self._emit_progress_update()
596
for path, new_ie in entries:
597
self._emit_progress_update()
598
file_id = new_ie.file_id
600
kind = self.work_tree.kind(file_id)
601
if kind == 'tree-reference' and self.recursive == 'down':
602
# nested tree: commit in it
603
sub_tree = WorkingTree.open(self.work_tree.abspath(path))
604
# FIXME: be more comprehensive here:
605
# this works when both trees are in --trees repository,
606
# but when both are bound to a different repository,
607
# it fails; a better way of approaching this is to
608
# finally implement the explicit-caches approach design
609
# a while back - RBC 20070306.
610
if (sub_tree.branch.repository.bzrdir.root_transport.base
612
self.work_tree.branch.repository.bzrdir.root_transport.base):
613
sub_tree.branch.repository = \
614
self.work_tree.branch.repository
616
sub_tree.commit(message=None, revprops=self.revprops,
617
recursive=self.recursive,
618
message_callback=self.message_callback,
619
timestamp=self.timestamp, timezone=self.timezone,
620
committer=self.committer,
621
allow_pointless=self.allow_pointless,
622
strict=self.strict, verbose=self.verbose,
623
local=self.local, reporter=self.reporter)
624
except errors.PointlessCommit:
626
if kind != new_ie.kind:
627
new_ie = inventory.make_entry(kind, new_ie.name,
628
new_ie.parent_id, file_id)
629
except errors.NoSuchFile:
631
# mutter('check %s {%s}', path, file_id)
632
if (not self.specific_files or
633
is_inside_or_parent_of_any(self.specific_files, path)):
634
# mutter('%s selected for commit', path)
638
# mutter('%s not selected for commit', path)
639
if self.basis_inv.has_id(file_id):
640
ie = self.basis_inv[file_id].copy()
642
# this entry is new and not being committed
644
self.builder.record_entry_contents(ie, self.parent_invs,
645
path, self.work_tree)
646
# describe the nature of the change that has occurred relative to
647
# the basis inventory.
648
if (self.basis_inv.has_id(ie.file_id)):
649
basis_ie = self.basis_inv[ie.file_id]
652
change = ie.describe_change(basis_ie, ie)
653
if change in (InventoryEntry.RENAMED,
654
InventoryEntry.MODIFIED_AND_RENAMED):
655
old_path = self.basis_inv.id2path(ie.file_id)
656
self.reporter.renamed(change, old_path, path)
658
self.reporter.snapshot_change(change, path)
660
if not self.specific_files:
663
# ignore removals that don't match filespec
664
for path, new_ie in self.basis_inv.iter_entries():
665
if new_ie.file_id in self.work_inv:
667
if is_inside_any(self.specific_files, path):
671
self.builder.record_entry_contents(ie, self.parent_invs, path,
674
def _emit_progress_update(self):
675
"""Emit an update to the progress bar."""
676
self.pb.update("Committing", self.pb_count, self.pb_total)
679
def _report_deletes(self):
680
for path, ie in self.basis_inv.iter_entries():
681
if ie.file_id not in self.builder.new_inventory:
682
self.reporter.deleted(path)
added
removed
bzrlib/commit.py
