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
def _note(self, format, *args):
112
Subclasses may choose to override this method.
116
def snapshot_change(self, change, path):
117
if change == 'unchanged':
119
if change == 'added' and path == '':
121
self._note("%s %s", change, path)
123
def completed(self, revno, rev_id):
124
self._note('Committed revision %d.', revno)
126
def deleted(self, file_id):
127
self._note('deleted %s', file_id)
129
def escaped(self, escape_count, message):
130
self._note("replaced %d control characters in message", escape_count)
132
def missing(self, path):
133
self._note('missing %s', path)
135
def renamed(self, change, old_path, new_path):
136
self._note('%s %s => %s', change, old_path, new_path)
139
class Commit(object):
140
"""Task of committing a new revision.
142
This is a MethodObject: it accumulates state as the commit is
143
prepared, and then it is discarded. It doesn't represent
144
historical revisions, just the act of recording a new one.
147
Modified to hold a list of files that have been deleted from
148
the working directory; these should be removed from the
154
if reporter is not None:
155
self.reporter = reporter
157
self.reporter = NullCommitReporter()
167
allow_pointless=True,
175
message_callback=None,
177
"""Commit working copy as a new revision.
179
:param message: the commit message (it or message_callback is required)
181
:param timestamp: if not None, seconds-since-epoch for a
182
postdated/predated commit.
184
:param specific_files: If true, commit only those files.
186
:param rev_id: If set, use this as the new revision id.
187
Useful for test or import commands that need to tightly
188
control what revisions are assigned. If you duplicate
189
a revision id that exists elsewhere it is your own fault.
190
If null (default), a time/random revision id is generated.
192
:param allow_pointless: If true (default), commit even if nothing
193
has changed and no merges are recorded.
195
:param strict: If true, don't allow a commit if the working tree
196
contains unknown files.
198
:param revprops: Properties for new revision
199
:param local: Perform a local only commit.
200
:param recursive: If set to 'down', commit in any subtrees that have
201
pending changes of any sort during this commit.
203
mutter('preparing to commit')
205
if working_tree is None:
206
raise BzrError("working_tree must be passed into commit().")
208
self.work_tree = working_tree
209
self.branch = self.work_tree.branch
210
if getattr(self.work_tree, 'requires_rich_root', lambda: False)():
211
if not self.branch.repository.supports_rich_root():
212
raise errors.RootNotRich()
213
if message_callback is None:
214
if message is not None:
215
if isinstance(message, str):
216
message = message.decode(bzrlib.user_encoding)
217
message_callback = lambda x: message
219
raise BzrError("The message or message_callback keyword"
220
" parameter is required for commit().")
222
self.bound_branch = None
224
self.master_branch = None
225
self.master_locked = False
227
self.specific_files = specific_files
228
self.allow_pointless = allow_pointless
229
self.recursive = recursive
230
self.revprops = revprops
231
self.message_callback = message_callback
232
self.timestamp = timestamp
233
self.timezone = timezone
234
self.committer = committer
236
self.verbose = verbose
238
if reporter is None and self.reporter is None:
239
self.reporter = NullCommitReporter()
240
elif reporter is not None:
241
self.reporter = reporter
243
self.work_tree.lock_write()
244
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
245
self.basis_tree = self.work_tree.basis_tree()
246
self.basis_tree.lock_read()
248
# Cannot commit with conflicts present.
249
if len(self.work_tree.conflicts()) > 0:
250
raise ConflictsInTree
252
# Setup the bound branch variables as needed.
253
self._check_bound_branch()
255
# Check that the working tree is up to date
256
old_revno,new_revno = self._check_out_of_date_tree()
259
# raise an exception as soon as we find a single unknown.
260
for unknown in self.work_tree.unknowns():
261
raise StrictCommitFailed()
263
if self.config is None:
264
self.config = self.branch.get_config()
266
# If provided, ensure the specified files are versioned
267
if specific_files is not None:
268
# Note: We don't actually need the IDs here. This routine
269
# is being called because it raises PathNotVerisonedError
270
# as a side effect of finding the IDs.
271
# XXX: Dont we have filter_unversioned to do this more
273
tree.find_ids_across_trees(specific_files,
274
[self.basis_tree, self.work_tree])
276
# Setup the progress bar. As the number of files that need to be
277
# committed in unknown, progress is reported as stages.
278
# We keep track of entries separately though and include that
279
# information in the progress bar during the relevant stages.
280
self.pb_stage_name = ""
281
self.pb_stage_count = 0
282
self.pb_stage_total = 4
283
if self.bound_branch:
284
self.pb_stage_total += 1
285
self.pb.show_pct = False
286
self.pb.show_spinner = False
287
self.pb.show_eta = False
288
self.pb.show_count = True
289
self.pb.show_bar = False
291
# After a merge, a selected file commit is not supported.
292
# See 'bzr help merge' for an explanation as to why.
293
self.basis_inv = self.basis_tree.inventory
294
self._gather_parents()
295
if len(self.parents) > 1 and self.specific_files:
296
raise errors.CannotCommitSelectedFileMerge(self.specific_files)
298
# Collect the changes
299
self._emit_progress_set_stage("Collecting changes", show_entries=True)
300
self.builder = self.branch.get_commit_builder(self.parents,
301
self.config, timestamp, timezone, committer, revprops, rev_id)
302
self._update_builder_with_changes()
303
self._check_pointless()
305
# TODO: Now the new inventory is known, check for conflicts.
306
# ADHB 2006-08-08: If this is done, populate_new_inv should not add
307
# weave lines, because nothing should be recorded until it is known
308
# that commit will succeed.
309
self._emit_progress_set_stage("Saving data locally")
310
self.builder.finish_inventory()
312
# Prompt the user for a commit message if none provided
313
message = message_callback(self)
314
assert isinstance(message, unicode), type(message)
315
self.message = message
316
self._escape_commit_message()
318
# Add revision data to the local branch
319
self.rev_id = self.builder.commit(self.message)
321
# Upload revision data to the master.
322
# this will propagate merged revisions too if needed.
323
if self.bound_branch:
324
self._emit_progress_set_stage("Uploading data to master branch")
325
self.master_branch.repository.fetch(self.branch.repository,
326
revision_id=self.rev_id)
327
# now the master has the revision data
328
# 'commit' to the master first so a timeout here causes the
329
# local branch to be out of date
330
self.master_branch.set_last_revision_info(new_revno,
333
# and now do the commit locally.
334
self.branch.set_last_revision_info(new_revno, self.rev_id)
336
# Make the working tree up to date with the branch
337
self._emit_progress_set_stage("Updating the working tree")
338
rev_tree = self.builder.revision_tree()
339
self.work_tree.set_parent_trees([(self.rev_id, rev_tree)])
340
self.reporter.completed(new_revno, self.rev_id)
342
# Process the post commit hooks, if any
343
self._emit_progress_set_stage("Running post commit hooks")
344
self._process_hooks(old_revno, new_revno)
349
def _any_real_changes(self):
350
"""Are there real changes between new_inventory and basis?
352
For trees without rich roots, inv.root.revision changes every commit.
353
But if that is the only change, we want to treat it as though there
356
new_entries = self.builder.new_inventory.iter_entries()
357
basis_entries = self.basis_inv.iter_entries()
358
new_path, new_root_ie = new_entries.next()
359
basis_path, basis_root_ie = basis_entries.next()
361
# This is a copy of InventoryEntry.__eq__ only leaving out .revision
362
def ie_equal_no_revision(this, other):
363
return ((this.file_id == other.file_id)
364
and (this.name == other.name)
365
and (this.symlink_target == other.symlink_target)
366
and (this.text_sha1 == other.text_sha1)
367
and (this.text_size == other.text_size)
368
and (this.text_id == other.text_id)
369
and (this.parent_id == other.parent_id)
370
and (this.kind == other.kind)
371
and (this.executable == other.executable)
372
and (this.reference_revision == other.reference_revision)
374
if not ie_equal_no_revision(new_root_ie, basis_root_ie):
377
for new_ie, basis_ie in zip(new_entries, basis_entries):
378
if new_ie != basis_ie:
381
# No actual changes present
384
def _check_pointless(self):
385
if self.allow_pointless:
387
# A merge with no effect on files
388
if len(self.parents) > 1:
390
# work around the fact that a newly-initted tree does differ from its
392
if len(self.basis_inv) == 0 and len(self.builder.new_inventory) == 1:
393
raise PointlessCommit()
394
# Shortcut, if the number of entries changes, then we obviously have
396
if len(self.builder.new_inventory) != len(self.basis_inv):
398
# If length == 1, then we only have the root entry. Which means
399
# that there is no real difference (only the root could be different)
400
if (len(self.builder.new_inventory) != 1 and self._any_real_changes()):
402
raise PointlessCommit()
404
def _check_bound_branch(self):
405
"""Check to see if the local branch is bound.
407
If it is bound, then most of the commit will actually be
408
done using the remote branch as the target branch.
409
Only at the end will the local branch be updated.
411
if self.local and not self.branch.get_bound_location():
412
raise errors.LocalRequiresBoundBranch()
415
self.master_branch = self.branch.get_master_branch()
417
if not self.master_branch:
418
# make this branch the reference branch for out of date checks.
419
self.master_branch = self.branch
422
# If the master branch is bound, we must fail
423
master_bound_location = self.master_branch.get_bound_location()
424
if master_bound_location:
425
raise errors.CommitToDoubleBoundBranch(self.branch,
426
self.master_branch, master_bound_location)
428
# TODO: jam 20051230 We could automatically push local
429
# commits to the remote branch if they would fit.
430
# But for now, just require remote to be identical
433
# Make sure the local branch is identical to the master
434
master_info = self.master_branch.last_revision_info()
435
local_info = self.branch.last_revision_info()
436
if local_info != master_info:
437
raise errors.BoundBranchOutOfDate(self.branch,
440
# Now things are ready to change the master branch
442
self.bound_branch = self.branch
443
self.master_branch.lock_write()
444
self.master_locked = True
446
def _check_out_of_date_tree(self):
447
"""Check that the working tree is up to date.
449
:return: old_revision_number,new_revision_number tuple
452
first_tree_parent = self.work_tree.get_parent_ids()[0]
454
# if there are no parents, treat our parent as 'None'
455
# this is so that we still consider the master branch
456
# - in a checkout scenario the tree may have no
457
# parents but the branch may do.
458
first_tree_parent = bzrlib.revision.NULL_REVISION
459
old_revno, master_last = self.master_branch.last_revision_info()
460
if master_last != first_tree_parent:
461
if master_last != bzrlib.revision.NULL_REVISION:
462
raise errors.OutOfDateTree(self.work_tree)
463
if self.branch.repository.has_revision(first_tree_parent):
464
new_revno = old_revno + 1
466
# ghost parents never appear in revision history.
468
return old_revno,new_revno
470
def _process_hooks(self, old_revno, new_revno):
471
"""Process any registered commit hooks."""
472
# old style commit hooks - should be deprecated ? (obsoleted in
474
if self.config.post_commit() is not None:
475
hooks = self.config.post_commit().split(' ')
476
# this would be nicer with twisted.python.reflect.namedAny
478
result = eval(hook + '(branch, rev_id)',
479
{'branch':self.branch,
481
'rev_id':self.rev_id})
482
# new style commit hooks:
483
if not self.bound_branch:
484
hook_master = self.branch
487
hook_master = self.master_branch
488
hook_local = self.branch
489
# With bound branches, when the master is behind the local branch,
490
# the 'old_revno' and old_revid values here are incorrect.
491
# XXX: FIXME ^. RBC 20060206
493
old_revid = self.parents[0]
495
old_revid = bzrlib.revision.NULL_REVISION
496
for hook in Branch.hooks['post_commit']:
497
hook(hook_local, hook_master, old_revno, old_revid, new_revno,
501
"""Cleanup any open locks, progress bars etc."""
502
cleanups = [self._cleanup_bound_branch,
503
self.basis_tree.unlock,
504
self.work_tree.unlock,
506
found_exception = None
507
for cleanup in cleanups:
510
# we want every cleanup to run no matter what.
511
# so we have a catchall here, but we will raise the
512
# last encountered exception up the stack: and
513
# typically this will be useful enough.
516
if found_exception is not None:
517
# don't do a plan raise, because the last exception may have been
518
# trashed, e is our sure-to-work exception even though it loses the
519
# full traceback. XXX: RBC 20060421 perhaps we could check the
520
# exc_info and if its the same one do a plain raise otherwise
521
# 'raise e' as we do now.
524
def _cleanup_bound_branch(self):
525
"""Executed at the end of a try/finally to cleanup a bound branch.
527
If the branch wasn't bound, this is a no-op.
528
If it was, it resents self.branch to the local branch, instead
531
if not self.bound_branch:
533
if self.master_locked:
534
self.master_branch.unlock()
536
def _escape_commit_message(self):
537
"""Replace xml-incompatible control characters."""
538
# FIXME: RBC 20060419 this should be done by the revision
539
# serialiser not by commit. Then we can also add an unescaper
540
# in the deserializer and start roundtripping revision messages
541
# precisely. See repository_implementations/test_repository.py
543
# Python strings can include characters that can't be
544
# represented in well-formed XML; escape characters that
545
# aren't listed in the XML specification
546
# (http://www.w3.org/TR/REC-xml/#NT-Char).
547
self.message, escape_count = re.subn(
548
u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
549
lambda match: match.group(0).encode('unicode_escape'),
552
self.reporter.escaped(escape_count, self.message)
554
def _gather_parents(self):
555
"""Record the parents of a merge for merge detection."""
556
# TODO: Make sure that this list doesn't contain duplicate
557
# entries and the order is preserved when doing this.
558
self.parents = self.work_tree.get_parent_ids()
559
self.parent_invs = [self.basis_inv]
560
for revision in self.parents[1:]:
561
if self.branch.repository.has_revision(revision):
562
mutter('commit parent revision {%s}', revision)
563
inventory = self.branch.repository.get_inventory(revision)
564
self.parent_invs.append(inventory)
566
mutter('commit parent ghost revision {%s}', revision)
568
def _update_builder_with_changes(self):
569
"""Update the commit builder with the data about what has changed.
571
# Build the revision inventory.
573
# This starts by creating a new empty inventory. Depending on
574
# which files are selected for commit, and what is present in the
575
# current tree, the new inventory is populated. inventory entries
576
# which are candidates for modification have their revision set to
577
# None; inventory entries that are carried over untouched have their
578
# revision set to their prior value.
580
# ESEPARATIONOFCONCERNS: this function is diffing and using the diff
581
# results to create a new inventory at the same time, which results
582
# in bugs like #46635. Any reason not to use/enhance Tree.changes_from?
585
specific_files = self.specific_files
586
mutter("Selecting files for commit with filter %s", specific_files)
587
work_inv = self.work_tree.inventory
588
assert work_inv.root is not None
589
self.pb_entries_total = len(work_inv)
591
# Check and warn about old CommitBuilders
592
entries = work_inv.iter_entries()
593
if not self.builder.record_root_entry:
594
symbol_versioning.warn('CommitBuilders should support recording'
595
' the root entry as of bzr 0.10.', DeprecationWarning,
597
self.builder.new_inventory.add(self.basis_inv.root.copy())
601
deleted_paths = set()
602
for path, new_ie in entries:
603
self._emit_progress_next_entry()
604
file_id = new_ie.file_id
606
# Skip files that have been deleted from the working tree.
607
# The deleted files/directories are also recorded so they
608
# can be explicitly unversioned later. Note that when a
609
# filter of specific files is given, we must only skip/record
610
# deleted files matching that filter.
611
if is_inside_any(deleted_paths, path):
613
if not specific_files or is_inside_any(specific_files, path):
614
if not self.work_tree.has_filename(path):
615
deleted_paths.add(path)
616
self.reporter.missing(path)
617
deleted_ids.append(file_id)
620
kind = self.work_tree.kind(file_id)
621
if kind == 'tree-reference' and self.recursive == 'down':
622
# nested tree: commit in it
623
sub_tree = WorkingTree.open(self.work_tree.abspath(path))
624
# FIXME: be more comprehensive here:
625
# this works when both trees are in --trees repository,
626
# but when both are bound to a different repository,
627
# it fails; a better way of approaching this is to
628
# finally implement the explicit-caches approach design
629
# a while back - RBC 20070306.
630
if (sub_tree.branch.repository.bzrdir.root_transport.base
632
self.work_tree.branch.repository.bzrdir.root_transport.base):
633
sub_tree.branch.repository = \
634
self.work_tree.branch.repository
636
sub_tree.commit(message=None, revprops=self.revprops,
637
recursive=self.recursive,
638
message_callback=self.message_callback,
639
timestamp=self.timestamp, timezone=self.timezone,
640
committer=self.committer,
641
allow_pointless=self.allow_pointless,
642
strict=self.strict, verbose=self.verbose,
643
local=self.local, reporter=self.reporter)
644
except errors.PointlessCommit:
646
if kind != new_ie.kind:
647
new_ie = inventory.make_entry(kind, new_ie.name,
648
new_ie.parent_id, file_id)
649
except errors.NoSuchFile:
651
# mutter('check %s {%s}', path, file_id)
652
if (not specific_files or
653
is_inside_or_parent_of_any(specific_files, path)):
654
# mutter('%s selected for commit', path)
658
# mutter('%s not selected for commit', path)
659
if self.basis_inv.has_id(file_id):
660
ie = self.basis_inv[file_id].copy()
662
# this entry is new and not being committed
664
self.builder.record_entry_contents(ie, self.parent_invs,
665
path, self.work_tree)
666
# describe the nature of the change that has occurred relative to
667
# the basis inventory.
668
if (self.basis_inv.has_id(ie.file_id)):
669
basis_ie = self.basis_inv[ie.file_id]
672
change = ie.describe_change(basis_ie, ie)
673
if change in (InventoryEntry.RENAMED,
674
InventoryEntry.MODIFIED_AND_RENAMED):
675
old_path = self.basis_inv.id2path(ie.file_id)
676
self.reporter.renamed(change, old_path, path)
678
self.reporter.snapshot_change(change, path)
680
# Unversion IDs that were found to be deleted
681
self.work_tree.unversion(deleted_ids)
683
# If specific files/directories were nominated, it is possible
684
# that some data from outside those needs to be preserved from
685
# the basis tree. For example, if a file x is moved from out of
686
# directory foo into directory bar and the user requests
687
# ``commit foo``, then information about bar/x must also be
690
for path, new_ie in self.basis_inv.iter_entries():
691
if new_ie.file_id in work_inv:
693
if is_inside_any(specific_files, path):
697
self.builder.record_entry_contents(ie, self.parent_invs, path,
700
# Report what was deleted. We could skip this when no deletes are
701
# detected to gain a performance win, but it arguably serves as a
702
# 'safety check' by informing the user whenever anything disappears.
703
for path, ie in self.basis_inv.iter_entries():
704
if ie.file_id not in self.builder.new_inventory:
705
self.reporter.deleted(path)
707
def _emit_progress_set_stage(self, name, show_entries=False):
708
"""Set the progress stage and emit an update to the progress bar."""
709
self.pb_stage_name = name
710
self.pb_stage_count += 1
711
self.pb_entries_show = show_entries
713
self.pb_entries_count = 0
714
self.pb_entries_total = '?'
715
self._emit_progress()
717
def _emit_progress_next_entry(self):
718
"""Emit an update to the progress bar and increment the file count."""
719
self.pb_entries_count += 1
720
self._emit_progress()
722
def _emit_progress(self):
723
if self.pb_entries_show:
724
text = "%s [Entry %d/%s] - Stage" % (self.pb_stage_name,
725
self.pb_entries_count,str(self.pb_entries_total))
727
text = "%s - Stage" % (self.pb_stage_name)
728
self.pb.update(text, self.pb_stage_count, self.pb_stage_total)
added
removed
bzrlib/commit.py
