« back to all changes in this revision
Viewing changes to bzrlib/branch.py
- browse files at revision 49
- compare with another revision
- download diff
- download tarball
- view history from revision 49
- Committer: mbp at sourcefrog
- Date: 2005-03-19 01:41:44 UTC
- Revision ID: mbp@sourcefrog.net-20050319014144-5298a74caebaf378
fix local-time-offset calculation
- files added:
- .bzrignore
- bzrlib
- doc
- notes
- files removed:
- .bzrignore
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
tree = self.working_tree()
304
inv = tree.inventory
305
306
# do this before any modifications
307
for f in files:
308
fid = inv.path2id(f)
309
if not fid:
310
bailout("cannot remove unversioned file %s" % quotefn(f))
311
mutter("remove inventory entry %s {%s}" % (quotefn(f), fid))
312
if verbose:
313
# having remove it, it must be either ignored or unknown
314
if tree.is_ignored(f):
315
new_status = 'I'
316
else:
317
new_status = '?'
318
show_status(new_status, inv[fid].kind, quotefn(f))
319
del inv[fid]
320
321
self._write_inventory(inv)
322
323
324
def unknowns(self):
325
"""Return all unknown files.
326
327
These are files in the working directory that are not versioned or
328
control files or ignored.
329
330
>>> b = ScratchBranch(files=['foo', 'foo~'])
331
>>> list(b.unknowns())
332
['foo']
333
>>> b.add('foo')
334
>>> list(b.unknowns())
335
[]
336
>>> b.remove('foo')
337
>>> list(b.unknowns())
338
['foo']
339
"""
340
return self.working_tree().unknowns()
341
342
343
def commit(self, message, timestamp=None, timezone=None,
344
committer=None,
345
verbose=False):
346
"""Commit working copy as a new revision.
347
348
The basic approach is to add all the file texts into the
349
store, then the inventory, then make a new revision pointing
350
to that inventory and store that.
351
352
This is not quite safe if the working copy changes during the
353
commit; for the moment that is simply not allowed. A better
354
approach is to make a temporary copy of the files before
355
computing their hashes, and then add those hashes in turn to
356
the inventory. This should mean at least that there are no
357
broken hash pointers. There is no way we can get a snapshot
358
of the whole directory at an instant. This would also have to
359
be robust against files disappearing, moving, etc. So the
360
whole thing is a bit hard.
361
362
:param timestamp: if not None, seconds-since-epoch for a
363
postdated/predated commit.
364
"""
365
366
## TODO: Show branch names
367
368
# TODO: Don't commit if there are no changes, unless forced?
369
370
# First walk over the working inventory; and both update that
371
# and also build a new revision inventory. The revision
372
# inventory needs to hold the text-id, sha1 and size of the
373
# actual file versions committed in the revision. (These are
374
# not present in the working inventory.) We also need to
375
# detect missing/deleted files, and remove them from the
376
# working inventory.
377
378
work_inv = self.read_working_inventory()
379
inv = Inventory()
380
basis = self.basis_tree()
381
basis_inv = basis.inventory
382
missing_ids = []
383
for path, entry in work_inv.iter_entries():
384
## TODO: Cope with files that have gone missing.
385
386
## TODO: Check that the file kind has not changed from the previous
387
## revision of this file (if any).
388
389
entry = entry.copy()
390
391
p = self._rel(path)
392
file_id = entry.file_id
393
mutter('commit prep file %s, id %r ' % (p, file_id))
394
395
if not os.path.exists(p):
396
mutter(" file is missing, removing from inventory")
397
if verbose:
398
show_status('D', entry.kind, quotefn(path))
399
missing_ids.append(file_id)
400
continue
401
402
# TODO: Handle files that have been deleted
403
404
# TODO: Maybe a special case for empty files? Seems a
405
# waste to store them many times.
406
407
inv.add(entry)
408
409
if basis_inv.has_id(file_id):
410
old_kind = basis_inv[file_id].kind
411
if old_kind != entry.kind:
412
bailout("entry %r changed kind from %r to %r"
413
% (file_id, old_kind, entry.kind))
414
415
if entry.kind == 'directory':
416
if not isdir(p):
417
bailout("%s is entered as directory but not a directory" % quotefn(p))
418
elif entry.kind == 'file':
419
if not isfile(p):
420
bailout("%s is entered as file but is not a file" % quotefn(p))
421
422
content = file(p, 'rb').read()
423
424
entry.text_sha1 = sha_string(content)
425
entry.text_size = len(content)
426
427
old_ie = basis_inv.has_id(file_id) and basis_inv[file_id]
428
if (old_ie
429
and (old_ie.text_size == entry.text_size)
430
and (old_ie.text_sha1 == entry.text_sha1)):
431
## assert content == basis.get_file(file_id).read()
432
entry.text_id = basis_inv[file_id].text_id
433
mutter(' unchanged from previous text_id {%s}' %
434
entry.text_id)
435
436
else:
437
entry.text_id = _gen_file_id(entry.name)
438
self.text_store.add(content, entry.text_id)
439
mutter(' stored with text_id {%s}' % entry.text_id)
440
if verbose:
441
if not old_ie:
442
state = 'A'
443
elif (old_ie.name == entry.name
444
and old_ie.parent_id == entry.parent_id):
445
state = 'R'
446
else:
447
state = 'M'
448
449
show_status(state, entry.kind, quotefn(path))
450
451
for file_id in missing_ids:
452
# have to do this later so we don't mess up the iterator.
453
# since parents may be removed before their children we
454
# have to test.
455
456
# FIXME: There's probably a better way to do this; perhaps
457
# the workingtree should know how to filter itself.
458
if work_inv.has_id(file_id):
459
del work_inv[file_id]
460
461
462
inv_id = rev_id = _gen_revision_id(time.time())
463
464
inv_tmp = tempfile.TemporaryFile()
465
inv.write_xml(inv_tmp)
466
inv_tmp.seek(0)
467
self.inventory_store.add(inv_tmp, inv_id)
468
mutter('new inventory_id is {%s}' % inv_id)
469
470
self._write_inventory(work_inv)
471
472
if timestamp == None:
473
timestamp = time.time()
474
475
if committer == None:
476
committer = username()
477
478
if timezone == None:
479
timezone = local_time_offset()
480
481
mutter("building commit log message")
482
rev = Revision(timestamp=timestamp,
483
timezone=timezone,
484
committer=committer,
485
precursor = self.last_patch(),
486
message = message,
487
inventory_id=inv_id,
488
revision_id=rev_id)
489
490
rev_tmp = tempfile.TemporaryFile()
491
rev.write_xml(rev_tmp)
492
rev_tmp.seek(0)
493
self.revision_store.add(rev_tmp, rev_id)
494
mutter("new revision_id is {%s}" % rev_id)
495
496
## XXX: Everything up to here can simply be orphaned if we abort
497
## the commit; it will leave junk files behind but that doesn't
498
## matter.
499
500
## TODO: Read back the just-generated changeset, and make sure it
501
## applies and recreates the right state.
502
503
## TODO: Also calculate and store the inventory SHA1
504
mutter("committing patch r%d" % (self.revno() + 1))
505
506
mutter("append to revision-history")
507
self.controlfile('revision-history', 'at').write(rev_id + '\n')
508
509
mutter("done!")
510
511
512
def get_revision(self, revision_id):
513
"""Return the Revision object for a named revision"""
514
r = Revision.read_xml(self.revision_store[revision_id])
515
assert r.revision_id == revision_id
516
return r
517
518
519
def get_inventory(self, inventory_id):
520
"""Get Inventory object by hash.
521
522
:todo: Perhaps for this and similar methods, take a revision
523
parameter which can be either an integer revno or a
524
string hash."""
525
i = Inventory.read_xml(self.inventory_store[inventory_id])
526
return i
527
528
529
def get_revision_inventory(self, revision_id):
530
"""Return inventory of a past revision."""
531
if revision_id == None:
532
return Inventory()
533
else:
534
return self.get_inventory(self.get_revision(revision_id).inventory_id)
535
536
537
def revision_history(self):
538
"""Return sequence of revision hashes on to this branch.
539
540
>>> ScratchBranch().revision_history()
541
[]
542
"""
543
return [chomp(l) for l in self.controlfile('revision-history').readlines()]
544
545
546
def revno(self):
547
"""Return current revision number for this branch.
548
549
That is equivalent to the number of revisions committed to
550
this branch.
551
552
>>> b = ScratchBranch()
553
>>> b.revno()
554
0
555
>>> b.commit('no foo')
556
>>> b.revno()
557
1
558
"""
559
return len(self.revision_history())
560
561
562
def last_patch(self):
563
"""Return last patch hash, or None if no history.
564
565
>>> ScratchBranch().last_patch() == None
566
True
567
"""
568
ph = self.revision_history()
569
if ph:
570
return ph[-1]
571
572
573
def lookup_revision(self, revno):
574
"""Return revision hash for revision number."""
575
if revno == 0:
576
return None
577
578
try:
579
# list is 0-based; revisions are 1-based
580
return self.revision_history()[revno-1]
581
except IndexError:
582
bailout("no such revision %s" % revno)
583
584
585
def revision_tree(self, revision_id):
586
"""Return Tree for a revision on this branch.
587
588
`revision_id` may be None for the null revision, in which case
589
an `EmptyTree` is returned."""
590
591
if revision_id == None:
592
return EmptyTree()
593
else:
594
inv = self.get_revision_inventory(revision_id)
595
return RevisionTree(self.text_store, inv)
596
597
598
def working_tree(self):
599
"""Return a `Tree` for the working copy."""
600
return WorkingTree(self.base, self.read_working_inventory())
601
602
603
def basis_tree(self):
604
"""Return `Tree` object for last revision.
605
606
If there are no revisions yet, return an `EmptyTree`.
607
608
>>> b = ScratchBranch(files=['foo'])
609
>>> b.basis_tree().has_filename('foo')
610
False
611
>>> b.working_tree().has_filename('foo')
612
True
613
>>> b.add('foo')
614
>>> b.commit('add foo')
615
>>> b.basis_tree().has_filename('foo')
616
True
617
"""
618
r = self.last_patch()
619
if r == None:
620
return EmptyTree()
621
else:
622
return RevisionTree(self.text_store, self.get_revision_inventory(r))
623
624
625
626
def write_log(self, show_timezone='original'):
627
"""Write out human-readable log of commits to this branch
628
629
:param utc: If true, show dates in universal time, not local time."""
630
## TODO: Option to choose either original, utc or local timezone
631
revno = 1
632
precursor = None
633
for p in self.revision_history():
634
print '-' * 40
635
print 'revno:', revno
636
## TODO: Show hash if --id is given.
637
##print 'revision-hash:', p
638
rev = self.get_revision(p)
639
print 'committer:', rev.committer
640
print 'timestamp: %s' % (format_date(rev.timestamp, rev.timezone or 0,
641
show_timezone))
642
643
## opportunistic consistency check, same as check_patch_chaining
644
if rev.precursor != precursor:
645
bailout("mismatched precursor!")
646
647
print 'message:'
648
if not rev.message:
649
print ' (no message)'
650
else:
651
for l in rev.message.split('\n'):
652
print ' ' + l
653
654
revno += 1
655
precursor = p
656
657
658
659
def show_status(branch, show_all=False):
660
"""Display single-line status for non-ignored working files.
661
662
The list is show sorted in order by file name.
663
664
>>> b = ScratchBranch(files=['foo', 'foo~'])
665
>>> b.show_status()
666
? foo
667
>>> b.add('foo')
668
>>> b.show_status()
669
A foo
670
>>> b.commit("add foo")
671
>>> b.show_status()
672
>>> os.unlink(b._rel('foo'))
673
>>> b.show_status()
674
D foo
675
676
677
:todo: Get state for single files.
678
679
:todo: Perhaps show a slash at the end of directory names.
680
681
"""
682
683
# We have to build everything into a list first so that it can
684
# sorted by name, incorporating all the different sources.
685
686
# FIXME: Rather than getting things in random order and then sorting,
687
# just step through in order.
688
689
# Interesting case: the old ID for a file has been removed,
690
# but a new file has been created under that name.
691
692
old = branch.basis_tree()
693
old_inv = old.inventory
694
new = branch.working_tree()
695
new_inv = new.inventory
696
697
for fs, fid, oldname, newname, kind in diff_trees(old, new):
698
if fs == 'R':
699
show_status(fs, kind,
700
oldname + ' => ' + newname)
701
elif fs == 'A' or fs == 'M':
702
show_status(fs, kind, newname)
703
elif fs == 'D':
704
show_status(fs, kind, oldname)
705
elif fs == '.':
706
if show_all:
707
show_status(fs, kind, newname)
708
elif fs == 'I':
709
if show_all:
710
show_status(fs, kind, newname)
711
elif fs == '?':
712
show_status(fs, kind, newname)
713
else:
714
bailout("wierd file state %r" % ((fs, fid),))
715
716
717
718
class ScratchBranch(Branch):
719
"""Special test class: a branch that cleans up after itself.
720
721
>>> b = ScratchBranch()
722
>>> isdir(b.base)
723
True
724
>>> bd = b.base
725
>>> del b
726
>>> isdir(bd)
727
False
728
"""
729
def __init__(self, files = []):
730
"""Make a test branch.
731
732
This creates a temporary directory and runs init-tree in it.
733
734
If any files are listed, they are created in the working copy.
735
"""
736
Branch.__init__(self, tempfile.mkdtemp(), init=True)
737
for f in files:
738
file(os.path.join(self.base, f), 'w').write('content of %s' % f)
739
740
741
def __del__(self):
742
"""Destroy the test branch, removing the scratch directory."""
743
shutil.rmtree(self.base)
744
745
746
747
######################################################################
748
# predicates
749
750
751
def is_control_file(filename):
752
## FIXME: better check
753
filename = os.path.normpath(filename)
754
while filename != '':
755
head, tail = os.path.split(filename)
756
## mutter('check %r for control file' % ((head, tail), ))
757
if tail == bzrlib.BZRDIR:
758
return True
759
filename = head
760
return False
761
762
763
764
def _gen_revision_id(when):
765
"""Return new revision-id."""
766
s = '%s-%s-' % (user_email(), compact_date(when))
767
s += hexlify(rand_bytes(8))
768
return s
769
770
771
def _gen_file_id(name):
772
"""Return new file id.
773
774
This should probably generate proper UUIDs, but for the moment we
775
cope with just randomness because running uuidgen every time is
776
slow."""
777
assert '/' not in name
778
while name[0] == '.':
779
name = name[1:]
780
s = hexlify(rand_bytes(8))
781
return '-'.join((name, compact_date(time.time()), s))
782
783
Loggerhead is a web-based interface for Breezy
Version: 2.0.1