1
# Copyright (C) 2005, 2006 Canonical Ltd
1
# Copyright (C) 2005 Canonical Ltd
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
13
13
# You should have received a copy of the GNU General Public License
14
14
# along with this program; if not, write to the Free Software
15
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.errors import (BzrError, PointlessCommit,
68
from bzrlib.osutils import (kind_marker, isdir,isfile, is_inside_any,
69
is_inside_or_parent_of_any,
70
quotefn, sha_file, split_lines)
71
from bzrlib.testament import Testament
72
from bzrlib.trace import mutter, note, warning
73
from bzrlib.xml5 import serializer_v5
74
from bzrlib.inventory import Inventory, InventoryEntry
75
from bzrlib import symbol_versioning
76
from bzrlib.symbol_versioning import (deprecated_passed,
79
from bzrlib.workingtree import WorkingTree
83
class NullCommitReporter(object):
84
"""I report on progress of a commit."""
86
def snapshot_change(self, change, path):
89
def completed(self, revno, rev_id):
92
def deleted(self, file_id):
95
def escaped(self, escape_count, message):
98
def missing(self, path):
101
def renamed(self, change, old_path, new_path):
105
class ReportCommitToLog(NullCommitReporter):
107
# this may be more useful if 'note' was replaced by an overridable
108
# method on self, which would allow more trivial subclassing.
109
# alternative, a callable could be passed in, allowing really trivial
110
# reuse for some uis. RBC 20060511
112
def snapshot_change(self, change, path):
113
if change == 'unchanged':
115
if change == 'added' and path == '':
117
note("%s %s", change, path)
119
def completed(self, revno, rev_id):
120
note('Committed revision %d.', revno)
122
def deleted(self, file_id):
123
note('deleted %s', file_id)
125
def escaped(self, escape_count, message):
126
note("replaced %d control characters in message", escape_count)
128
def missing(self, path):
129
note('missing %s', path)
131
def renamed(self, change, old_path, new_path):
132
note('%s %s => %s', change, old_path, new_path)
135
class Commit(object):
136
"""Task of committing a new revision.
138
This is a MethodObject: it accumulates state as the commit is
139
prepared, and then it is discarded. It doesn't represent
140
historical revisions, just the act of recording a new one.
143
Modified to hold a list of files that have been deleted from
144
the working directory; these should be removed from the
19
def commit(branch, message,
26
"""Commit working copy as a new revision.
28
The basic approach is to add all the file texts into the
29
store, then the inventory, then make a new revision pointing
30
to that inventory and store that.
32
This is not quite safe if the working copy changes during the
33
commit; for the moment that is simply not allowed. A better
34
approach is to make a temporary copy of the files before
35
computing their hashes, and then add those hashes in turn to
36
the inventory. This should mean at least that there are no
37
broken hash pointers. There is no way we can get a snapshot
38
of the whole directory at an instant. This would also have to
39
be robust against files disappearing, moving, etc. So the
40
whole thing is a bit hard.
42
timestamp -- if not None, seconds-since-epoch for a
43
postdated/predated commit.
46
If true, commit only those files.
49
If set, use this as the new revision id.
50
Useful for test or import commands that need to tightly
51
control what revisions are assigned. If you duplicate
52
a revision id that exists elsewhere it is your own fault.
53
If null (default), a time/random revision id is generated.
150
if reporter is not None:
151
self.reporter = reporter
153
self.reporter = NullCommitReporter()
154
if config is not None:
160
branch=DEPRECATED_PARAMETER, message=None,
166
allow_pointless=True,
174
message_callback=None):
175
"""Commit working copy as a new revision.
177
branch -- the deprecated branch to commit to. New callers should pass in
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.
202
mutter('preparing to commit')
204
if deprecated_passed(branch):
205
symbol_versioning.warn("Commit.commit (branch, ...): The branch parameter is "
206
"deprecated as of bzr 0.8. Please use working_tree= instead.",
207
DeprecationWarning, stacklevel=2)
209
self.work_tree = self.branch.bzrdir.open_workingtree()
210
elif working_tree is None:
211
raise BzrError("One of branch and working_tree must be passed into commit().")
213
self.work_tree = working_tree
214
self.branch = self.work_tree.branch
215
if message_callback is None:
216
if message is not None:
217
if isinstance(message, str):
218
message = message.decode(bzrlib.user_encoding)
219
message_callback = lambda x: message
221
raise BzrError("The message or message_callback keyword"
222
" parameter is required for commit().")
224
self.bound_branch = None
226
self.master_branch = None
227
self.master_locked = False
229
self.specific_files = specific_files
230
self.allow_pointless = allow_pointless
232
if reporter is None and self.reporter is None:
233
self.reporter = NullCommitReporter()
234
elif reporter is not None:
235
self.reporter = reporter
237
self.work_tree.lock_write()
238
self.pb = bzrlib.ui.ui_factory.nested_progress_bar()
240
# Cannot commit with conflicts present.
241
if len(self.work_tree.conflicts())>0:
242
raise ConflictsInTree
244
# setup the bound branch variables as needed.
245
self._check_bound_branch()
247
# check for out of date working trees
249
first_tree_parent = self.work_tree.get_parent_ids()[0]
251
# if there are no parents, treat our parent as 'None'
252
# this is so that we still consier the master branch
253
# - in a checkout scenario the tree may have no
254
# parents but the branch may do.
255
first_tree_parent = None
256
master_last = self.master_branch.last_revision()
257
if (master_last is not None and
258
master_last != first_tree_parent):
259
raise errors.OutOfDateTree(self.work_tree)
56
import os, time, tempfile
58
from inventory import Inventory
59
from osutils import isdir, isfile, sha_string, quotefn, \
60
local_time_offset, username, kind_marker, is_inside_any
262
# raise an exception as soon as we find a single unknown.
263
for unknown in self.work_tree.unknowns():
264
raise StrictCommitFailed()
266
if self.config is None:
267
self.config = self.branch.get_config()
269
self.work_inv = self.work_tree.inventory
270
self.basis_tree = self.work_tree.basis_tree()
271
self.basis_inv = self.basis_tree.inventory
272
if specific_files is not None:
273
# Ensure specified files are versioned
274
# (We don't actually need the ids here)
275
tree.find_ids_across_trees(specific_files,
276
[self.basis_tree, self.work_tree])
277
# one to finish, one for rev and inventory, and one for each
278
# inventory entry, and the same for the new inventory.
279
# note that this estimate is too long when we do a partial tree
280
# commit which excludes some new files from being considered.
281
# The estimate is corrected when we populate the new inv.
282
self.pb_total = len(self.work_inv) + 5
285
self._gather_parents()
286
if len(self.parents) > 1 and self.specific_files:
287
raise NotImplementedError('selected-file commit of merges is not supported yet: files %r',
290
self.builder = self.branch.get_commit_builder(self.parents,
291
self.config, timestamp, timezone, committer, revprops, rev_id)
293
self._remove_deleted()
294
self._populate_new_inv()
295
self._report_deletes()
297
self._check_pointless()
299
self._emit_progress_update()
300
# TODO: Now the new inventory is known, check for conflicts and
301
# prompt the user for a commit message.
302
# ADHB 2006-08-08: If this is done, populate_new_inv should not add
303
# weave lines, because nothing should be recorded until it is known
304
# that commit will succeed.
305
self.builder.finish_inventory()
306
self._emit_progress_update()
307
message = message_callback(self)
308
assert isinstance(message, unicode), type(message)
309
self.message = message
310
self._escape_commit_message()
312
self.rev_id = self.builder.commit(self.message)
313
self._emit_progress_update()
314
# revision data is in the local branch now.
316
# upload revision data to the master.
317
# this will propagate merged revisions too if needed.
318
if self.bound_branch:
319
self.master_branch.repository.fetch(self.branch.repository,
320
revision_id=self.rev_id)
321
# now the master has the revision data
322
# 'commit' to the master first so a timeout here causes the local
323
# branch to be out of date
324
self.master_branch.append_revision(self.rev_id)
326
# and now do the commit locally.
327
self.branch.append_revision(self.rev_id)
329
rev_tree = self.builder.revision_tree()
330
self.work_tree.set_parent_trees([(self.rev_id, rev_tree)])
331
# now the work tree is up to date with the branch
333
self.reporter.completed(self.branch.revno(), self.rev_id)
334
if self.config.post_commit() is not None:
335
hooks = self.config.post_commit().split(' ')
336
# this would be nicer with twisted.python.reflect.namedAny
338
result = eval(hook + '(branch, rev_id)',
339
{'branch':self.branch,
341
'rev_id':self.rev_id})
342
self._emit_progress_update()
347
def _any_real_changes(self):
348
"""Are there real changes between new_inventory and basis?
350
For trees without rich roots, inv.root.revision changes every commit.
351
But if that is the only change, we want to treat it as though there
354
new_entries = self.builder.new_inventory.iter_entries()
355
basis_entries = self.basis_inv.iter_entries()
356
new_path, new_root_ie = new_entries.next()
357
basis_path, basis_root_ie = basis_entries.next()
359
# This is a copy of InventoryEntry.__eq__ only leaving out .revision
360
def ie_equal_no_revision(this, other):
361
return ((this.file_id == other.file_id)
362
and (this.name == other.name)
363
and (this.symlink_target == other.symlink_target)
364
and (this.text_sha1 == other.text_sha1)
365
and (this.text_size == other.text_size)
366
and (this.text_id == other.text_id)
367
and (this.parent_id == other.parent_id)
368
and (this.kind == other.kind)
369
and (this.executable == other.executable)
371
if not ie_equal_no_revision(new_root_ie, basis_root_ie):
374
for new_ie, basis_ie in zip(new_entries, basis_entries):
375
if new_ie != basis_ie:
378
# No actual changes present
381
def _check_pointless(self):
382
if self.allow_pointless:
384
# A merge with no effect on files
385
if len(self.parents) > 1:
387
# work around the fact that a newly-initted tree does differ from its
389
if len(self.basis_inv) == 0 and len(self.builder.new_inventory) == 1:
390
raise PointlessCommit()
391
# Shortcut, if the number of entries changes, then we obviously have
393
if len(self.builder.new_inventory) != len(self.basis_inv):
395
# If length == 1, then we only have the root entry. Which means
396
# that there is no real difference (only the root could be different)
397
if (len(self.builder.new_inventory) != 1 and self._any_real_changes()):
399
raise PointlessCommit()
401
def _check_bound_branch(self):
402
"""Check to see if the local branch is bound.
404
If it is bound, then most of the commit will actually be
405
done using the remote branch as the target branch.
406
Only at the end will the local branch be updated.
408
if self.local and not self.branch.get_bound_location():
409
raise errors.LocalRequiresBoundBranch()
412
self.master_branch = self.branch.get_master_branch()
414
if not self.master_branch:
415
# make this branch the reference branch for out of date checks.
416
self.master_branch = self.branch
419
# If the master branch is bound, we must fail
420
master_bound_location = self.master_branch.get_bound_location()
421
if master_bound_location:
422
raise errors.CommitToDoubleBoundBranch(self.branch,
423
self.master_branch, master_bound_location)
425
# TODO: jam 20051230 We could automatically push local
426
# commits to the remote branch if they would fit.
427
# But for now, just require remote to be identical
430
# Make sure the local branch is identical to the master
431
master_rh = self.master_branch.revision_history()
432
local_rh = self.branch.revision_history()
433
if local_rh != master_rh:
434
raise errors.BoundBranchOutOfDate(self.branch,
437
# Now things are ready to change the master branch
439
self.bound_branch = self.branch
440
self.master_branch.lock_write()
441
self.master_locked = True
444
"""Cleanup any open locks, progress bars etc."""
445
cleanups = [self._cleanup_bound_branch,
446
self.work_tree.unlock,
448
found_exception = None
449
for cleanup in cleanups:
452
# we want every cleanup to run no matter what.
453
# so we have a catchall here, but we will raise the
454
# last encountered exception up the stack: and
455
# typically this will be useful enough.
458
if found_exception is not None:
459
# don't do a plan raise, because the last exception may have been
460
# trashed, e is our sure-to-work exception even though it loses the
461
# full traceback. XXX: RBC 20060421 perhaps we could check the
462
# exc_info and if its the same one do a plain raise otherwise
463
# 'raise e' as we do now.
466
def _cleanup_bound_branch(self):
467
"""Executed at the end of a try/finally to cleanup a bound branch.
469
If the branch wasn't bound, this is a no-op.
470
If it was, it resents self.branch to the local branch, instead
473
if not self.bound_branch:
475
if self.master_locked:
476
self.master_branch.unlock()
478
def _escape_commit_message(self):
479
"""Replace xml-incompatible control characters."""
480
# FIXME: RBC 20060419 this should be done by the revision
481
# serialiser not by commit. Then we can also add an unescaper
482
# in the deserializer and start roundtripping revision messages
483
# precisely. See repository_implementations/test_repository.py
485
# Python strings can include characters that can't be
486
# represented in well-formed XML; escape characters that
487
# aren't listed in the XML specification
488
# (http://www.w3.org/TR/REC-xml/#NT-Char).
489
self.message, escape_count = re.subn(
490
u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
491
lambda match: match.group(0).encode('unicode_escape'),
494
self.reporter.escaped(escape_count, self.message)
496
def _gather_parents(self):
497
"""Record the parents of a merge for merge detection."""
498
# TODO: Make sure that this list doesn't contain duplicate
499
# entries and the order is preserved when doing this.
500
self.parents = self.work_tree.get_parent_ids()
501
self.parent_invs = []
502
for revision in self.parents:
503
if self.branch.repository.has_revision(revision):
504
mutter('commit parent revision {%s}', revision)
505
inventory = self.branch.repository.get_inventory(revision)
506
self.parent_invs.append(inventory)
508
mutter('commit parent ghost revision {%s}', revision)
510
def _remove_deleted(self):
511
"""Remove deleted files from the working inventories.
513
This is done prior to taking the working inventory as the
514
basis for the new committed inventory.
516
This returns true if any files
517
*that existed in the basis inventory* were deleted.
518
Files that were added and deleted
519
in the working copy don't matter.
521
specific = self.specific_files
523
deleted_paths = set()
524
for path, ie in self.work_inv.iter_entries():
525
if is_inside_any(deleted_paths, path):
526
# The tree will delete the required ids recursively.
528
if specific and not is_inside_any(specific, path):
530
if not self.work_tree.has_filename(path):
531
deleted_paths.add(path)
532
self.reporter.missing(path)
533
deleted_ids.append(ie.file_id)
534
self.work_tree.unversion(deleted_ids)
536
def _populate_new_inv(self):
537
"""Build revision inventory.
539
This creates a new empty inventory. Depending on
540
which files are selected for commit, and what is present in the
541
current tree, the new inventory is populated. inventory entries
542
which are candidates for modification have their revision set to
543
None; inventory entries that are carried over untouched have their
544
revision set to their prior value.
546
# ESEPARATIONOFCONCERNS: this function is diffing and using the diff
547
# results to create a new inventory at the same time, which results
548
# in bugs like #46635. Any reason not to use/enhance Tree.changes_from?
550
mutter("Selecting files for commit with filter %s", self.specific_files)
551
assert self.work_inv.root is not None
552
entries = self.work_inv.iter_entries()
553
if not self.builder.record_root_entry:
554
symbol_versioning.warn('CommitBuilders should support recording'
555
' the root entry as of bzr 0.10.', DeprecationWarning,
557
self.builder.new_inventory.add(self.basis_inv.root.copy())
559
self._emit_progress_update()
560
for path, new_ie in entries:
561
self._emit_progress_update()
562
file_id = new_ie.file_id
563
# mutter('check %s {%s}', path, file_id)
564
if (not self.specific_files or
565
is_inside_or_parent_of_any(self.specific_files, path)):
566
# mutter('%s selected for commit', path)
570
# mutter('%s not selected for commit', path)
571
if self.basis_inv.has_id(file_id):
572
ie = self.basis_inv[file_id].copy()
574
# this entry is new and not being committed
576
self.builder.record_entry_contents(ie, self.parent_invs,
577
path, self.work_tree)
578
# describe the nature of the change that has occurred relative to
579
# the basis inventory.
580
if (self.basis_inv.has_id(ie.file_id)):
581
basis_ie = self.basis_inv[ie.file_id]
584
change = ie.describe_change(basis_ie, ie)
585
if change in (InventoryEntry.RENAMED,
586
InventoryEntry.MODIFIED_AND_RENAMED):
587
old_path = self.basis_inv.id2path(ie.file_id)
588
self.reporter.renamed(change, old_path, path)
590
self.reporter.snapshot_change(change, path)
592
if not self.specific_files:
595
# ignore removals that don't match filespec
596
for path, new_ie in self.basis_inv.iter_entries():
597
if new_ie.file_id in self.work_inv:
599
if is_inside_any(self.specific_files, path):
603
self.builder.record_entry_contents(ie, self.parent_invs, path,
606
def _emit_progress_update(self):
607
"""Emit an update to the progress bar."""
608
self.pb.update("Committing", self.pb_count, self.pb_total)
611
def _report_deletes(self):
612
for path, ie in self.basis_inv.iter_entries():
613
if ie.file_id not in self.builder.new_inventory:
614
self.reporter.deleted(path)
62
from branch import gen_file_id
63
from errors import BzrError
64
from revision import Revision
65
from trace import mutter, note
67
branch._need_writelock()
69
# First walk over the working inventory; and both update that
70
# and also build a new revision inventory. The revision
71
# inventory needs to hold the text-id, sha1 and size of the
72
# actual file versions committed in the revision. (These are
73
# not present in the working inventory.) We also need to
74
# detect missing/deleted files, and remove them from the
77
work_tree = branch.working_tree()
78
work_inv = work_tree.inventory
80
basis = branch.basis_tree()
81
basis_inv = basis.inventory
85
note('looking for changes...')
87
for path, entry in work_inv.iter_entries():
88
## TODO: Check that the file kind has not changed from the previous
89
## revision of this file (if any).
93
p = branch.abspath(path)
94
file_id = entry.file_id
95
mutter('commit prep file %s, id %r ' % (p, file_id))
97
if specific_files and not is_inside_any(specific_files, path):
98
if basis_inv.has_id(file_id):
99
# carry over with previous state
100
inv.add(basis_inv[file_id].copy())
102
# omit this from committed inventory
106
if not work_tree.has_id(file_id):
108
print('deleted %s%s' % (path, kind_marker(entry.kind)))
109
mutter(" file is missing, removing from inventory")
110
missing_ids.append(file_id)
115
if basis_inv.has_id(file_id):
116
old_kind = basis_inv[file_id].kind
117
if old_kind != entry.kind:
118
raise BzrError("entry %r changed kind from %r to %r"
119
% (file_id, old_kind, entry.kind))
121
if entry.kind == 'directory':
123
raise BzrError("%s is entered as directory but not a directory"
125
elif entry.kind == 'file':
127
raise BzrError("%s is entered as file but is not a file" % quotefn(p))
129
new_sha1 = work_tree.get_file_sha1(file_id)
131
old_ie = basis_inv.has_id(file_id) and basis_inv[file_id]
133
and old_ie.text_sha1 == new_sha1):
134
## assert content == basis.get_file(file_id).read()
135
entry.text_id = old_ie.text_id
136
entry.text_sha1 = new_sha1
137
entry.text_size = old_ie.text_size
138
mutter(' unchanged from previous text_id {%s}' %
141
content = file(p, 'rb').read()
143
# calculate the sha again, just in case the file contents
144
# changed since we updated the cache
145
entry.text_sha1 = sha_string(content)
146
entry.text_size = len(content)
148
entry.text_id = gen_file_id(entry.name)
149
branch.text_store.add(content, entry.text_id)
150
mutter(' stored with text_id {%s}' % entry.text_id)
153
print('added %s' % path)
154
elif (old_ie.name == entry.name
155
and old_ie.parent_id == entry.parent_id):
156
print('modified %s' % path)
158
print('renamed %s' % path)
161
for file_id in missing_ids:
162
# Any files that have been deleted are now removed from the
163
# working inventory. Files that were not selected for commit
164
# are left as they were in the working inventory and ommitted
165
# from the revision inventory.
167
# have to do this later so we don't mess up the iterator.
168
# since parents may be removed before their children we
171
# FIXME: There's probably a better way to do this; perhaps
172
# the workingtree should know how to filter itbranch.
173
if work_inv.has_id(file_id):
174
del work_inv[file_id]
178
rev_id = _gen_revision_id(time.time())
181
inv_tmp = tempfile.TemporaryFile()
182
inv.write_xml(inv_tmp)
184
branch.inventory_store.add(inv_tmp, inv_id)
185
mutter('new inventory_id is {%s}' % inv_id)
187
branch._write_inventory(work_inv)
189
if timestamp == None:
190
timestamp = time.time()
192
if committer == None:
193
committer = username()
196
timezone = local_time_offset()
198
mutter("building commit log message")
199
rev = Revision(timestamp=timestamp,
202
precursor = branch.last_patch(),
207
rev_tmp = tempfile.TemporaryFile()
208
rev.write_xml(rev_tmp)
210
branch.revision_store.add(rev_tmp, rev_id)
211
mutter("new revision_id is {%s}" % rev_id)
213
## XXX: Everything up to here can simply be orphaned if we abort
214
## the commit; it will leave junk files behind but that doesn't
217
## TODO: Read back the just-generated changeset, and make sure it
218
## applies and recreates the right state.
220
## TODO: Also calculate and store the inventory SHA1
221
mutter("committing patch r%d" % (branch.revno() + 1))
223
branch.append_revision(rev_id)
226
note("commited r%d" % branch.revno())
230
def _gen_revision_id(when):
231
"""Return new revision-id."""
232
from binascii import hexlify
233
from osutils import rand_bytes, compact_date, user_email
235
s = '%s-%s-' % (user_email(), compact_date(when))
236
s += hexlify(rand_bytes(8))
added
removed
bzrlib/commit.py
