To get this branch, use:
bzr branch
http://gegoxaren.bato24.eu/bzr/brz/remove-bazaar
« back to all changes in this revision
Viewing changes to bzrlib/branch.py
- browse files at revision 0.200.1036
- compare with another revision
- download diff
- download tarball
- view history from revision 0.200.1036
- Committer: Jelmer Vernooij
- Date: 2010-09-11 17:48:45 UTC
- mto: (0.312.1 master) (6883.23.1 bundle-git)
- mto: This revision was merged to the branch mainline in revision 6960.
- Revision ID: jelmer@samba.org-20100911174845-ro06bb6gg6ws8jit
More work on roundtrip push support.
- files added:
- .bzrignore
- notes
- tests
- files removed:
- .bzrignore
- bzrlib
- doc
- notes
added
removed
bzrlib/branch.py
1
#! /usr/bin/env python
2
# -*- coding: UTF-8 -*-
3
4
# This program is free software; you can redistribute it and/or modify
5
# it under the terms of the GNU General Public License as published by
6
# the Free Software Foundation; either version 2 of the License, or
7
# (at your option) any later version.
8
9
# This program is distributed in the hope that it will be useful,
10
# but WITHOUT ANY WARRANTY; without even the implied warranty of
11
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12
# GNU General Public License for more details.
13
14
# You should have received a copy of the GNU General Public License
15
# along with this program; if not, write to the Free Software
16
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
18
19
from sets import Set
20
21
import sys, os, os.path, random, time, sha, sets, types, re, shutil, tempfile
22
import traceback, socket, fnmatch, difflib, time
23
from binascii import hexlify
24
25
import bzrlib
26
from inventory import Inventory
27
from trace import mutter, note
28
from tree import Tree, EmptyTree, RevisionTree, WorkingTree
29
from inventory import InventoryEntry, Inventory
30
from osutils import isdir, quotefn, isfile, uuid, sha_file, username, chomp, \
31
format_date, compact_date, pumpfile, user_email, rand_bytes, splitpath, \
32
joinpath, sha_string, file_kind, local_time_offset
33
from store import ImmutableStore
34
from revision import Revision
35
from errors import bailout
36
from textui import show_status
37
from diff import diff_trees
38
39
BZR_BRANCH_FORMAT = "Bazaar-NG branch, format 0.0.4\n"
40
## TODO: Maybe include checks for common corruption of newlines, etc?
41
42
43
44
45
46
######################################################################
47
# branch objects
48
49
class Branch:
50
"""Branch holding a history of revisions.
51
52
:todo: Perhaps use different stores for different classes of object,
53
so that we can keep track of how much space each one uses,
54
or garbage-collect them.
55
56
:todo: Add a RemoteBranch subclass. For the basic case of read-only
57
HTTP access this should be very easy by,
58
just redirecting controlfile access into HTTP requests.
59
We would need a RemoteStore working similarly.
60
61
:todo: Keep the on-disk branch locked while the object exists.
62
63
:todo: mkdir() method.
64
"""
65
def __init__(self, base, init=False):
66
"""Create new branch object at a particular location.
67
68
:param base: Base directory for the branch.
69
70
:param init: If True, create new control files in a previously
71
unversioned directory. If False, the branch must already
72
be versioned.
73
74
In the test suite, creation of new trees is tested using the
75
`ScratchBranch` class.
76
"""
77
self.base = os.path.realpath(base)
78
if init:
79
self._make_control()
80
else:
81
if not isdir(self.controlfilename('.')):
82
bailout("not a bzr branch: %s" % quotefn(base),
83
['use "bzr init" to initialize a new working tree',
84
'current bzr can only operate from top-of-tree'])
85
self._check_format()
86
87
self.text_store = ImmutableStore(self.controlfilename('text-store'))
88
self.revision_store = ImmutableStore(self.controlfilename('revision-store'))
89
self.inventory_store = ImmutableStore(self.controlfilename('inventory-store'))
90
91
92
def __str__(self):
93
return '%s(%r)' % (self.__class__.__name__, self.base)
94
95
96
__repr__ = __str__
97
98
99
def _rel(self, name):
100
"""Return filename relative to branch top"""
101
return os.path.join(self.base, name)
102
103
104
def controlfilename(self, file_or_path):
105
"""Return location relative to branch."""
106
if isinstance(file_or_path, types.StringTypes):
107
file_or_path = [file_or_path]
108
return os.path.join(self.base, bzrlib.BZRDIR, *file_or_path)
109
110
111
def controlfile(self, file_or_path, mode='r'):
112
"""Open a control file for this branch"""
113
return file(self.controlfilename(file_or_path), mode)
114
115
116
def _make_control(self):
117
os.mkdir(self.controlfilename([]))
118
self.controlfile('README', 'w').write(
119
"This is a Bazaar-NG control directory.\n"
120
"Do not change any files in this directory.")
121
self.controlfile('branch-format', 'w').write(BZR_BRANCH_FORMAT)
122
for d in ('text-store', 'inventory-store', 'revision-store'):
123
os.mkdir(self.controlfilename(d))
124
for f in ('revision-history', 'merged-patches',
125
'pending-merged-patches', 'branch-name'):
126
self.controlfile(f, 'w').write('')
127
mutter('created control directory in ' + self.base)
128
Inventory().write_xml(self.controlfile('inventory','w'))
129
130
131
def _check_format(self):
132
"""Check this branch format is supported.
133
134
The current tool only supports the current unstable format.
135
136
In the future, we might need different in-memory Branch
137
classes to support downlevel branches. But not yet.
138
"""
139
# read in binary mode to detect newline wierdness.
140
fmt = self.controlfile('branch-format', 'rb').read()
141
if fmt != BZR_BRANCH_FORMAT:
142
bailout('sorry, branch format %r not supported' % fmt,
143
['use a different bzr version',
144
'or remove the .bzr directory and "bzr init" again'])
145
146
147
def read_working_inventory(self):
148
"""Read the working inventory."""
149
before = time.time()
150
inv = Inventory.read_xml(self.controlfile('inventory', 'r'))
151
mutter("loaded inventory of %d items in %f"
152
% (len(inv), time.time() - before))
153
return inv
154
155
156
def _write_inventory(self, inv):
157
"""Update the working inventory.
158
159
That is to say, the inventory describing changes underway, that
160
will be committed to the next revision.
161
"""
162
## TODO: factor out to atomicfile? is rename safe on windows?
163
tmpfname = self.controlfilename('inventory.tmp')
164
tmpf = file(tmpfname, 'w')
165
inv.write_xml(tmpf)
166
tmpf.close()
167
os.rename(tmpfname, self.controlfilename('inventory'))
168
mutter('wrote working inventory')
169
170
171
inventory = property(read_working_inventory, _write_inventory, None,
172
"""Inventory for the working copy.""")
173
174
175
def add(self, files, verbose=False):
176
"""Make files versioned.
177
178
This puts the files in the Added state, so that they will be
179
recorded by the next commit.
180
181
:todo: Perhaps have an option to add the ids even if the files do
182
not (yet) exist.
183
184
:todo: Perhaps return the ids of the files? But then again it
185
is easy to retrieve them if they're needed.
186
187
:todo: Option to specify file id.
188
189
:todo: Adding a directory should optionally recurse down and
190
add all non-ignored children. Perhaps do that in a
191
higher-level method.
192
193
>>> b = ScratchBranch(files=['foo'])
194
>>> 'foo' in b.unknowns()
195
True
196
>>> b.show_status()
197
? foo
198
>>> b.add('foo')
199
>>> 'foo' in b.unknowns()
200
False
201
>>> bool(b.inventory.path2id('foo'))
202
True
203
>>> b.show_status()
204
A foo
205
206
>>> b.add('foo')
207
Traceback (most recent call last):
208
...
209
BzrError: ('foo is already versioned', [])
210
211
>>> b.add(['nothere'])
212
Traceback (most recent call last):
213
BzrError: ('cannot add: not a regular file or directory: nothere', [])
214
"""
215
216
# TODO: Re-adding a file that is removed in the working copy
217
# should probably put it back with the previous ID.
218
if isinstance(files, types.StringTypes):
219
files = [files]
220
221
inv = self.read_working_inventory()
222
for f in files:
223
if is_control_file(f):
224
bailout("cannot add control file %s" % quotefn(f))
225
226
fp = splitpath(f)
227
228
if len(fp) == 0:
229
bailout("cannot add top-level %r" % f)
230
231
fullpath = os.path.normpath(self._rel(f))
232
233
if isfile(fullpath):
234
kind = 'file'
235
elif isdir(fullpath):
236
kind = 'directory'
237
else:
238
bailout('cannot add: not a regular file or directory: %s' % quotefn(f))
239
240
if len(fp) > 1:
241
parent_name = joinpath(fp[:-1])
242
mutter("lookup parent %r" % parent_name)
243
parent_id = inv.path2id(parent_name)
244
if parent_id == None:
245
bailout("cannot add: parent %r is not versioned"
246
% joinpath(fp[:-1]))
247
else:
248
parent_id = None
249
250
file_id = _gen_file_id(fp[-1])
251
inv.add(InventoryEntry(file_id, fp[-1], kind=kind, parent_id=parent_id))
252
if verbose:
253
show_status('A', kind, quotefn(f))
254
255
mutter("add file %s file_id:{%s} kind=%r parent_id={%s}"
256
% (f, file_id, kind, parent_id))
257
self._write_inventory(inv)
258
259
260
261
def remove(self, files, verbose=False):
262
"""Mark nominated files for removal from the inventory.
263
264
This does not remove their text. This does not run on
265
266
:todo: Refuse to remove modified files unless --force is given?
267
268
>>> b = ScratchBranch(files=['foo'])
269
>>> b.add('foo')
270
>>> b.inventory.has_filename('foo')
271
True
272
>>> b.remove('foo')
273
>>> b.working_tree().has_filename('foo')
274
True
275
>>> b.inventory.has_filename('foo')
276
False
277
278
>>> b = ScratchBranch(files=['foo'])
279
>>> b.add('foo')
280
>>> b.commit('one')
281
>>> b.remove('foo')
282
>>> b.commit('two')
283
>>> b.inventory.has_filename('foo')
284
False
285
>>> b.basis_tree().has_filename('foo')
286
False
287
>>> b.working_tree().has_filename('foo')
288
True
289
290
:todo: Do something useful with directories.
291
292
:todo: Should this remove the text or not? Tough call; not
293
removing may be useful and the user can just use use rm, and
294
is the opposite of add. Removing it is consistent with most
295
other tools. Maybe an option.
296
"""
297
## TODO: Normalize names
298
## TODO: Remove nested loops; better scalability
299
300
if isinstance(files, types.StringTypes):
301
files = [files]
302
303
inv = self.read_working_inventory()
304
305
# do this before any modifications
306
for f in files:
307
fid = inv.path2id(f)
308
if not fid:
309
bailout("cannot remove unversioned file %s" % quotefn(f))
310
mutter("remove inventory entry %s {%s}" % (quotefn(f), fid))
311
if verbose:
312
show_status('D', inv[fid].kind, quotefn(f))
313
del inv[fid]
314
315
self._write_inventory(inv)
316
317
318
def unknowns(self):
319
"""Return all unknown files.
320
321
These are files in the working directory that are not versioned or
322
control files or ignored.
323
324
>>> b = ScratchBranch(files=['foo', 'foo~'])
325
>>> list(b.unknowns())
326
['foo']
327
>>> b.add('foo')
328
>>> list(b.unknowns())
329
[]
330
>>> b.remove('foo')
331
>>> list(b.unknowns())
332
['foo']
333
"""
334
return self.working_tree().unknowns()
335
336
337
def commit(self, message, timestamp=None, timezone=None,
338
committer=None,
339
verbose=False):
340
"""Commit working copy as a new revision.
341
342
The basic approach is to add all the file texts into the
343
store, then the inventory, then make a new revision pointing
344
to that inventory and store that.
345
346
This is not quite safe if the working copy changes during the
347
commit; for the moment that is simply not allowed. A better
348
approach is to make a temporary copy of the files before
349
computing their hashes, and then add those hashes in turn to
350
the inventory. This should mean at least that there are no
351
broken hash pointers. There is no way we can get a snapshot
352
of the whole directory at an instant. This would also have to
353
be robust against files disappearing, moving, etc. So the
354
whole thing is a bit hard.
355
356
:param timestamp: if not None, seconds-since-epoch for a
357
postdated/predated commit.
358
"""
359
360
## TODO: Show branch names
361
362
# TODO: Don't commit if there are no changes, unless forced?
363
364
# First walk over the working inventory; and both update that
365
# and also build a new revision inventory. The revision
366
# inventory needs to hold the text-id, sha1 and size of the
367
# actual file versions committed in the revision. (These are
368
# not present in the working inventory.) We also need to
369
# detect missing/deleted files, and remove them from the
370
# working inventory.
371
372
work_inv = self.read_working_inventory()
373
inv = Inventory()
374
basis = self.basis_tree()
375
basis_inv = basis.inventory
376
missing_ids = []
377
for path, entry in work_inv.iter_entries():
378
## TODO: Cope with files that have gone missing.
379
380
## TODO: Check that the file kind has not changed from the previous
381
## revision of this file (if any).
382
383
entry = entry.copy()
384
385
p = self._rel(path)
386
file_id = entry.file_id
387
mutter('commit prep file %s, id %r ' % (p, file_id))
388
389
if not os.path.exists(p):
390
mutter(" file is missing, removing from inventory")
391
if verbose:
392
show_status('D', entry.kind, quotefn(path))
393
missing_ids.append(file_id)
394
continue
395
396
# TODO: Handle files that have been deleted
397
398
# TODO: Maybe a special case for empty files? Seems a
399
# waste to store them many times.
400
401
inv.add(entry)
402
403
if basis_inv.has_id(file_id):
404
old_kind = basis_inv[file_id].kind
405
if old_kind != entry.kind:
406
bailout("entry %r changed kind from %r to %r"
407
% (file_id, old_kind, entry.kind))
408
409
if entry.kind == 'directory':
410
if not isdir(p):
411
bailout("%s is entered as directory but not a directory" % quotefn(p))
412
elif entry.kind == 'file':
413
if not isfile(p):
414
bailout("%s is entered as file but is not a file" % quotefn(p))
415
416
content = file(p, 'rb').read()
417
418
entry.text_sha1 = sha_string(content)
419
entry.text_size = len(content)
420
421
old_ie = basis_inv.has_id(file_id) and basis_inv[file_id]
422
if (old_ie
423
and (old_ie.text_size == entry.text_size)
424
and (old_ie.text_sha1 == entry.text_sha1)):
425
## assert content == basis.get_file(file_id).read()
426
entry.text_id = basis_inv[file_id].text_id
427
mutter(' unchanged from previous text_id {%s}' %
428
entry.text_id)
429
430
else:
431
entry.text_id = _gen_file_id(entry.name)
432
self.text_store.add(content, entry.text_id)
433
mutter(' stored with text_id {%s}' % entry.text_id)
434
if verbose:
435
if not old_ie:
436
state = 'A'
437
elif (old_ie.name == entry.name
438
and old_ie.parent_id == entry.parent_id):
439
state = 'R'
440
else:
441
state = 'M'
442
443
show_status(state, entry.kind, quotefn(path))
444
445
for file_id in missing_ids:
446
# have to do this later so we don't mess up the iterator.
447
# since parents may be removed before their children we
448
# have to test.
449
450
# FIXME: There's probably a better way to do this; perhaps
451
# the workingtree should know how to filter itself.
452
if work_inv.has_id(file_id):
453
del work_inv[file_id]
454
455
456
inv_id = rev_id = _gen_revision_id(time.time())
457
458
inv_tmp = tempfile.TemporaryFile()
459
inv.write_xml(inv_tmp)
460
inv_tmp.seek(0)
461
self.inventory_store.add(inv_tmp, inv_id)
462
mutter('new inventory_id is {%s}' % inv_id)
463
464
self._write_inventory(work_inv)
465
466
if timestamp == None:
467
timestamp = time.time()
468
469
if committer == None:
470
committer = username()
471
472
if timezone == None:
473
timezone = local_time_offset()
474
475
mutter("building commit log message")
476
rev = Revision(timestamp=timestamp,
477
timezone=timezone,
478
committer=committer,
479
precursor = self.last_patch(),
480
message = message,
481
inventory_id=inv_id,
482
revision_id=rev_id)
483
484
rev_tmp = tempfile.TemporaryFile()
485
rev.write_xml(rev_tmp)
486
rev_tmp.seek(0)
487
self.revision_store.add(rev_tmp, rev_id)
488
mutter("new revision_id is {%s}" % rev_id)
489
490
## XXX: Everything up to here can simply be orphaned if we abort
491
## the commit; it will leave junk files behind but that doesn't
492
## matter.
493
494
## TODO: Read back the just-generated changeset, and make sure it
495
## applies and recreates the right state.
496
497
## TODO: Also calculate and store the inventory SHA1
498
mutter("committing patch r%d" % (self.revno() + 1))
499
500
mutter("append to revision-history")
501
self.controlfile('revision-history', 'at').write(rev_id + '\n')
502
503
mutter("done!")
504
505
506
def get_revision(self, revision_id):
507
"""Return the Revision object for a named revision"""
508
r = Revision.read_xml(self.revision_store[revision_id])
509
assert r.revision_id == revision_id
510
return r
511
512
513
def get_inventory(self, inventory_id):
514
"""Get Inventory object by hash.
515
516
:todo: Perhaps for this and similar methods, take a revision
517
parameter which can be either an integer revno or a
518
string hash."""
519
i = Inventory.read_xml(self.inventory_store[inventory_id])
520
return i
521
522
523
def get_revision_inventory(self, revision_id):
524
"""Return inventory of a past revision."""
525
if revision_id == None:
526
return Inventory()
527
else:
528
return self.get_inventory(self.get_revision(revision_id).inventory_id)
529
530
531
def revision_history(self):
532
"""Return sequence of revision hashes on to this branch.
533
534
>>> ScratchBranch().revision_history()
535
[]
536
"""
537
return [chomp(l) for l in self.controlfile('revision-history').readlines()]
538
539
540
def revno(self):
541
"""Return current revision number for this branch.
542
543
That is equivalent to the number of revisions committed to
544
this branch.
545
546
>>> b = ScratchBranch()
547
>>> b.revno()
548
0
549
>>> b.commit('no foo')
550
>>> b.revno()
551
1
552
"""
553
return len(self.revision_history())
554
555
556
def last_patch(self):
557
"""Return last patch hash, or None if no history.
558
559
>>> ScratchBranch().last_patch() == None
560
True
561
"""
562
ph = self.revision_history()
563
if ph:
564
return ph[-1]
565
566
567
def lookup_revision(self, revno):
568
"""Return revision hash for revision number."""
569
if revno == 0:
570
return None
571
572
try:
573
# list is 0-based; revisions are 1-based
574
return self.revision_history()[revno-1]
575
except IndexError:
576
bailout("no such revision %s" % revno)
577
578
579
def revision_tree(self, revision_id):
580
"""Return Tree for a revision on this branch.
581
582
`revision_id` may be None for the null revision, in which case
583
an `EmptyTree` is returned."""
584
585
if revision_id == None:
586
return EmptyTree()
587
else:
588
inv = self.get_revision_inventory(revision_id)
589
return RevisionTree(self.text_store, inv)
590
591
592
def working_tree(self):
593
"""Return a `Tree` for the working copy."""
594
return WorkingTree(self.base, self.read_working_inventory())
595
596
597
def basis_tree(self):
598
"""Return `Tree` object for last revision.
599
600
If there are no revisions yet, return an `EmptyTree`.
601
602
>>> b = ScratchBranch(files=['foo'])
603
>>> b.basis_tree().has_filename('foo')
604
False
605
>>> b.working_tree().has_filename('foo')
606
True
607
>>> b.add('foo')
608
>>> b.commit('add foo')
609
>>> b.basis_tree().has_filename('foo')
610
True
611
"""
612
r = self.last_patch()
613
if r == None:
614
return EmptyTree()
615
else:
616
return RevisionTree(self.text_store, self.get_revision_inventory(r))
617
618
619
620
def write_log(self, show_timezone='original'):
621
"""Write out human-readable log of commits to this branch
622
623
:param utc: If true, show dates in universal time, not local time."""
624
## TODO: Option to choose either original, utc or local timezone
625
revno = 1
626
precursor = None
627
for p in self.revision_history():
628
print '-' * 40
629
print 'revno:', revno
630
## TODO: Show hash if --id is given.
631
##print 'revision-hash:', p
632
rev = self.get_revision(p)
633
print 'committer:', rev.committer
634
print 'timestamp: %s' % (format_date(rev.timestamp, rev.timezone or 0,
635
show_timezone))
636
637
## opportunistic consistency check, same as check_patch_chaining
638
if rev.precursor != precursor:
639
bailout("mismatched precursor!")
640
641
print 'message:'
642
if not rev.message:
643
print ' (no message)'
644
else:
645
for l in rev.message.split('\n'):
646
print ' ' + l
647
648
revno += 1
649
precursor = p
650
651
652
653
def show_status(branch, show_all=False):
654
"""Display single-line status for non-ignored working files.
655
656
The list is show sorted in order by file name.
657
658
>>> b = ScratchBranch(files=['foo', 'foo~'])
659
>>> b.show_status()
660
? foo
661
>>> b.add('foo')
662
>>> b.show_status()
663
A foo
664
>>> b.commit("add foo")
665
>>> b.show_status()
666
>>> os.unlink(b._rel('foo'))
667
>>> b.show_status()
668
D foo
669
670
671
:todo: Get state for single files.
672
673
:todo: Perhaps show a slash at the end of directory names.
674
675
"""
676
677
# We have to build everything into a list first so that it can
678
# sorted by name, incorporating all the different sources.
679
680
# FIXME: Rather than getting things in random order and then sorting,
681
# just step through in order.
682
683
# Interesting case: the old ID for a file has been removed,
684
# but a new file has been created under that name.
685
686
old = branch.basis_tree()
687
old_inv = old.inventory
688
new = branch.working_tree()
689
new_inv = new.inventory
690
691
for fs, fid, oldname, newname, kind in diff_trees(old, new):
692
if fs == 'R':
693
show_status(fs, kind,
694
oldname + ' => ' + newname)
695
elif fs == 'A' or fs == 'M':
696
show_status(fs, kind, newname)
697
elif fs == 'D':
698
show_status(fs, kind, oldname)
699
elif fs == '.':
700
if show_all:
701
show_status(fs, kind, newname)
702
elif fs == 'I':
703
if show_all:
704
show_status(fs, kind, newname)
705
elif fs == '?':
706
show_status(fs, kind, newname)
707
else:
708
bailout("wierd file state %r" % ((fs, fid),))
709
710
711
712
class ScratchBranch(Branch):
713
"""Special test class: a branch that cleans up after itself.
714
715
>>> b = ScratchBranch()
716
>>> isdir(b.base)
717
True
718
>>> bd = b.base
719
>>> del b
720
>>> isdir(bd)
721
False
722
"""
723
def __init__(self, files = []):
724
"""Make a test branch.
725
726
This creates a temporary directory and runs init-tree in it.
727
728
If any files are listed, they are created in the working copy.
729
"""
730
Branch.__init__(self, tempfile.mkdtemp(), init=True)
731
for f in files:
732
file(os.path.join(self.base, f), 'w').write('content of %s' % f)
733
734
735
def __del__(self):
736
"""Destroy the test branch, removing the scratch directory."""
737
shutil.rmtree(self.base)
738
739
740
741
######################################################################
742
# predicates
743
744
745
def is_control_file(filename):
746
## FIXME: better check
747
filename = os.path.normpath(filename)
748
while filename != '':
749
head, tail = os.path.split(filename)
750
## mutter('check %r for control file' % ((head, tail), ))
751
if tail == bzrlib.BZRDIR:
752
return True
753
filename = head
754
return False
755
756
757
758
def _gen_revision_id(when):
759
"""Return new revision-id."""
760
s = '%s-%s-' % (user_email(), compact_date(when))
761
s += hexlify(rand_bytes(8))
762
return s
763
764
765
def _gen_file_id(name):
766
"""Return new file id.
767
768
This should probably generate proper UUIDs, but for the moment we
769
cope with just randomness because running uuidgen every time is
770
slow."""
771
assert '/' not in name
772
while name[0] == '.':
773
name = name[1:]
774
s = hexlify(rand_bytes(8))
775
return '-'.join((name, compact_date(time.time()), s))
776
777
Loggerhead is a web-based interface for Breezy
Version: 2.0.2.dev0