1
# Copyright (C) 2005 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
20
import sys, os, os.path, random, time, sha, sets, types, re, shutil, tempfile
21
import traceback, socket, fnmatch, difflib, time
22
from binascii import hexlify
25
from inventory import Inventory
26
from trace import mutter, note
27
from tree import Tree, EmptyTree, RevisionTree, WorkingTree
28
from inventory import InventoryEntry, Inventory
29
from osutils import isdir, quotefn, isfile, uuid, sha_file, username, chomp, \
30
format_date, compact_date, pumpfile, user_email, rand_bytes, splitpath, \
31
joinpath, sha_string, file_kind, local_time_offset, appendpath
32
from store import ImmutableStore
33
from revision import Revision
34
from errors import bailout
35
from textui import show_status
36
from diff import diff_trees
38
BZR_BRANCH_FORMAT = "Bazaar-NG branch, format 0.0.4\n"
39
## TODO: Maybe include checks for common corruption of newlines, etc?
43
def find_branch_root(f=None):
44
"""Find the branch root enclosing f, or pwd.
46
It is not necessary that f exists.
48
Basically we keep looking up until we find the control directory or
52
elif hasattr(os.path, 'realpath'):
53
f = os.path.realpath(f)
55
f = os.path.abspath(f)
61
if os.path.exists(os.path.join(f, bzrlib.BZRDIR)):
63
head, tail = os.path.split(f)
65
# reached the root, whatever that may be
66
bailout('%r is not in a branch' % orig_f)
71
######################################################################
75
"""Branch holding a history of revisions.
77
:todo: Perhaps use different stores for different classes of object,
78
so that we can keep track of how much space each one uses,
79
or garbage-collect them.
81
:todo: Add a RemoteBranch subclass. For the basic case of read-only
82
HTTP access this should be very easy by,
83
just redirecting controlfile access into HTTP requests.
84
We would need a RemoteStore working similarly.
86
:todo: Keep the on-disk branch locked while the object exists.
88
:todo: mkdir() method.
90
def __init__(self, base, init=False, find_root=True):
91
"""Create new branch object at a particular location.
93
:param base: Base directory for the branch.
95
:param init: If True, create new control files in a previously
96
unversioned directory. If False, the branch must already
99
:param find_root: If true and init is false, find the root of the
100
existing branch containing base.
102
In the test suite, creation of new trees is tested using the
103
`ScratchBranch` class.
106
self.base = os.path.realpath(base)
109
self.base = find_branch_root(base)
111
self.base = os.path.realpath(base)
112
if not isdir(self.controlfilename('.')):
113
bailout("not a bzr branch: %s" % quotefn(base),
114
['use "bzr init" to initialize a new working tree',
115
'current bzr can only operate from top-of-tree'])
118
self.text_store = ImmutableStore(self.controlfilename('text-store'))
119
self.revision_store = ImmutableStore(self.controlfilename('revision-store'))
120
self.inventory_store = ImmutableStore(self.controlfilename('inventory-store'))
124
return '%s(%r)' % (self.__class__.__name__, self.base)
130
def abspath(self, name):
131
"""Return absolute filename for something in the branch"""
132
return os.path.join(self.base, name)
135
def relpath(self, path):
136
"""Return path relative to this branch of something inside it.
138
Raises an error if path is not in this branch."""
139
rp = os.path.realpath(path)
141
if not rp.startswith(self.base):
142
bailout("path %r is not within branch %r" % (rp, self.base))
143
rp = rp[len(self.base):]
144
rp = rp.lstrip(os.sep)
148
def controlfilename(self, file_or_path):
149
"""Return location relative to branch."""
150
if isinstance(file_or_path, types.StringTypes):
151
file_or_path = [file_or_path]
152
return os.path.join(self.base, bzrlib.BZRDIR, *file_or_path)
155
def controlfile(self, file_or_path, mode='r'):
156
"""Open a control file for this branch"""
157
return file(self.controlfilename(file_or_path), mode)
160
def _make_control(self):
161
os.mkdir(self.controlfilename([]))
162
self.controlfile('README', 'w').write(
163
"This is a Bazaar-NG control directory.\n"
164
"Do not change any files in this directory.")
165
self.controlfile('branch-format', 'w').write(BZR_BRANCH_FORMAT)
166
for d in ('text-store', 'inventory-store', 'revision-store'):
167
os.mkdir(self.controlfilename(d))
168
for f in ('revision-history', 'merged-patches',
169
'pending-merged-patches', 'branch-name'):
170
self.controlfile(f, 'w').write('')
171
mutter('created control directory in ' + self.base)
172
Inventory().write_xml(self.controlfile('inventory','w'))
175
def _check_format(self):
176
"""Check this branch format is supported.
178
The current tool only supports the current unstable format.
180
In the future, we might need different in-memory Branch
181
classes to support downlevel branches. But not yet.
183
# read in binary mode to detect newline wierdness.
184
fmt = self.controlfile('branch-format', 'rb').read()
185
if fmt != BZR_BRANCH_FORMAT:
186
bailout('sorry, branch format %r not supported' % fmt,
187
['use a different bzr version',
188
'or remove the .bzr directory and "bzr init" again'])
191
def read_working_inventory(self):
192
"""Read the working inventory."""
194
inv = Inventory.read_xml(self.controlfile('inventory', 'r'))
195
mutter("loaded inventory of %d items in %f"
196
% (len(inv), time.time() - before))
200
def _write_inventory(self, inv):
201
"""Update the working inventory.
203
That is to say, the inventory describing changes underway, that
204
will be committed to the next revision.
206
## TODO: factor out to atomicfile? is rename safe on windows?
207
## TODO: Maybe some kind of clean/dirty marker on inventory?
208
tmpfname = self.controlfilename('inventory.tmp')
209
tmpf = file(tmpfname, 'w')
212
os.rename(tmpfname, self.controlfilename('inventory'))
213
mutter('wrote working inventory')
216
inventory = property(read_working_inventory, _write_inventory, None,
217
"""Inventory for the working copy.""")
220
def add(self, files, verbose=False):
221
"""Make files versioned.
223
This puts the files in the Added state, so that they will be
224
recorded by the next commit.
226
:todo: Perhaps have an option to add the ids even if the files do
229
:todo: Perhaps return the ids of the files? But then again it
230
is easy to retrieve them if they're needed.
232
:todo: Option to specify file id.
234
:todo: Adding a directory should optionally recurse down and
235
add all non-ignored children. Perhaps do that in a
238
>>> b = ScratchBranch(files=['foo'])
239
>>> 'foo' in b.unknowns()
244
>>> 'foo' in b.unknowns()
246
>>> bool(b.inventory.path2id('foo'))
252
Traceback (most recent call last):
254
BzrError: ('foo is already versioned', [])
256
>>> b.add(['nothere'])
257
Traceback (most recent call last):
258
BzrError: ('cannot add: not a regular file or directory: nothere', [])
261
# TODO: Re-adding a file that is removed in the working copy
262
# should probably put it back with the previous ID.
263
if isinstance(files, types.StringTypes):
266
inv = self.read_working_inventory()
268
if is_control_file(f):
269
bailout("cannot add control file %s" % quotefn(f))
274
bailout("cannot add top-level %r" % f)
276
fullpath = os.path.normpath(self.abspath(f))
279
kind = file_kind(fullpath)
281
# maybe something better?
282
bailout('cannot add: not a regular file or directory: %s' % quotefn(f))
284
if kind != 'file' and kind != 'directory':
285
bailout('cannot add: not a regular file or directory: %s' % quotefn(f))
287
file_id = gen_file_id(f)
288
inv.add_path(f, kind=kind, file_id=file_id)
291
show_status('A', kind, quotefn(f))
293
mutter("add file %s file_id:{%s} kind=%r" % (f, file_id, kind))
295
self._write_inventory(inv)
299
def remove(self, files, verbose=False):
300
"""Mark nominated files for removal from the inventory.
302
This does not remove their text. This does not run on
304
:todo: Refuse to remove modified files unless --force is given?
306
>>> b = ScratchBranch(files=['foo'])
308
>>> b.inventory.has_filename('foo')
311
>>> b.working_tree().has_filename('foo')
313
>>> b.inventory.has_filename('foo')
316
>>> b = ScratchBranch(files=['foo'])
321
>>> b.inventory.has_filename('foo')
323
>>> b.basis_tree().has_filename('foo')
325
>>> b.working_tree().has_filename('foo')
328
:todo: Do something useful with directories.
330
:todo: Should this remove the text or not? Tough call; not
331
removing may be useful and the user can just use use rm, and
332
is the opposite of add. Removing it is consistent with most
333
other tools. Maybe an option.
335
## TODO: Normalize names
336
## TODO: Remove nested loops; better scalability
338
if isinstance(files, types.StringTypes):
341
tree = self.working_tree()
344
# do this before any modifications
348
bailout("cannot remove unversioned file %s" % quotefn(f))
349
mutter("remove inventory entry %s {%s}" % (quotefn(f), fid))
351
# having remove it, it must be either ignored or unknown
352
if tree.is_ignored(f):
356
show_status(new_status, inv[fid].kind, quotefn(f))
359
self._write_inventory(inv)
363
"""Return all unknown files.
365
These are files in the working directory that are not versioned or
366
control files or ignored.
368
>>> b = ScratchBranch(files=['foo', 'foo~'])
369
>>> list(b.unknowns())
372
>>> list(b.unknowns())
375
>>> list(b.unknowns())
378
return self.working_tree().unknowns()
381
def commit(self, message, timestamp=None, timezone=None,
384
"""Commit working copy as a new revision.
386
The basic approach is to add all the file texts into the
387
store, then the inventory, then make a new revision pointing
388
to that inventory and store that.
390
This is not quite safe if the working copy changes during the
391
commit; for the moment that is simply not allowed. A better
392
approach is to make a temporary copy of the files before
393
computing their hashes, and then add those hashes in turn to
394
the inventory. This should mean at least that there are no
395
broken hash pointers. There is no way we can get a snapshot
396
of the whole directory at an instant. This would also have to
397
be robust against files disappearing, moving, etc. So the
398
whole thing is a bit hard.
400
:param timestamp: if not None, seconds-since-epoch for a
401
postdated/predated commit.
404
## TODO: Show branch names
406
# TODO: Don't commit if there are no changes, unless forced?
408
# First walk over the working inventory; and both update that
409
# and also build a new revision inventory. The revision
410
# inventory needs to hold the text-id, sha1 and size of the
411
# actual file versions committed in the revision. (These are
412
# not present in the working inventory.) We also need to
413
# detect missing/deleted files, and remove them from the
416
work_inv = self.read_working_inventory()
418
basis = self.basis_tree()
419
basis_inv = basis.inventory
421
for path, entry in work_inv.iter_entries():
422
## TODO: Cope with files that have gone missing.
424
## TODO: Check that the file kind has not changed from the previous
425
## revision of this file (if any).
429
p = self.abspath(path)
430
file_id = entry.file_id
431
mutter('commit prep file %s, id %r ' % (p, file_id))
433
if not os.path.exists(p):
434
mutter(" file is missing, removing from inventory")
436
show_status('D', entry.kind, quotefn(path))
437
missing_ids.append(file_id)
440
# TODO: Handle files that have been deleted
442
# TODO: Maybe a special case for empty files? Seems a
443
# waste to store them many times.
447
if basis_inv.has_id(file_id):
448
old_kind = basis_inv[file_id].kind
449
if old_kind != entry.kind:
450
bailout("entry %r changed kind from %r to %r"
451
% (file_id, old_kind, entry.kind))
453
if entry.kind == 'directory':
455
bailout("%s is entered as directory but not a directory" % quotefn(p))
456
elif entry.kind == 'file':
458
bailout("%s is entered as file but is not a file" % quotefn(p))
460
content = file(p, 'rb').read()
462
entry.text_sha1 = sha_string(content)
463
entry.text_size = len(content)
465
old_ie = basis_inv.has_id(file_id) and basis_inv[file_id]
467
and (old_ie.text_size == entry.text_size)
468
and (old_ie.text_sha1 == entry.text_sha1)):
469
## assert content == basis.get_file(file_id).read()
470
entry.text_id = basis_inv[file_id].text_id
471
mutter(' unchanged from previous text_id {%s}' %
475
entry.text_id = gen_file_id(entry.name)
476
self.text_store.add(content, entry.text_id)
477
mutter(' stored with text_id {%s}' % entry.text_id)
481
elif (old_ie.name == entry.name
482
and old_ie.parent_id == entry.parent_id):
487
show_status(state, entry.kind, quotefn(path))
489
for file_id in missing_ids:
490
# have to do this later so we don't mess up the iterator.
491
# since parents may be removed before their children we
494
# FIXME: There's probably a better way to do this; perhaps
495
# the workingtree should know how to filter itself.
496
if work_inv.has_id(file_id):
497
del work_inv[file_id]
500
inv_id = rev_id = _gen_revision_id(time.time())
502
inv_tmp = tempfile.TemporaryFile()
503
inv.write_xml(inv_tmp)
505
self.inventory_store.add(inv_tmp, inv_id)
506
mutter('new inventory_id is {%s}' % inv_id)
508
self._write_inventory(work_inv)
510
if timestamp == None:
511
timestamp = time.time()
513
if committer == None:
514
committer = username()
517
timezone = local_time_offset()
519
mutter("building commit log message")
520
rev = Revision(timestamp=timestamp,
523
precursor = self.last_patch(),
528
rev_tmp = tempfile.TemporaryFile()
529
rev.write_xml(rev_tmp)
531
self.revision_store.add(rev_tmp, rev_id)
532
mutter("new revision_id is {%s}" % rev_id)
534
## XXX: Everything up to here can simply be orphaned if we abort
535
## the commit; it will leave junk files behind but that doesn't
538
## TODO: Read back the just-generated changeset, and make sure it
539
## applies and recreates the right state.
541
## TODO: Also calculate and store the inventory SHA1
542
mutter("committing patch r%d" % (self.revno() + 1))
544
mutter("append to revision-history")
545
f = self.controlfile('revision-history', 'at')
546
f.write(rev_id + '\n')
550
note("commited r%d" % self.revno())
553
def get_revision(self, revision_id):
554
"""Return the Revision object for a named revision"""
555
r = Revision.read_xml(self.revision_store[revision_id])
556
assert r.revision_id == revision_id
560
def get_inventory(self, inventory_id):
561
"""Get Inventory object by hash.
563
:todo: Perhaps for this and similar methods, take a revision
564
parameter which can be either an integer revno or a
566
i = Inventory.read_xml(self.inventory_store[inventory_id])
570
def get_revision_inventory(self, revision_id):
571
"""Return inventory of a past revision."""
572
if revision_id == None:
575
return self.get_inventory(self.get_revision(revision_id).inventory_id)
578
def revision_history(self):
579
"""Return sequence of revision hashes on to this branch.
581
>>> ScratchBranch().revision_history()
584
return [chomp(l) for l in self.controlfile('revision-history').readlines()]
588
"""Return current revision number for this branch.
590
That is equivalent to the number of revisions committed to
593
>>> b = ScratchBranch()
596
>>> b.commit('no foo')
600
return len(self.revision_history())
603
def last_patch(self):
604
"""Return last patch hash, or None if no history.
606
>>> ScratchBranch().last_patch() == None
609
ph = self.revision_history()
614
def lookup_revision(self, revno):
615
"""Return revision hash for revision number."""
620
# list is 0-based; revisions are 1-based
621
return self.revision_history()[revno-1]
623
bailout("no such revision %s" % revno)
626
def revision_tree(self, revision_id):
627
"""Return Tree for a revision on this branch.
629
`revision_id` may be None for the null revision, in which case
630
an `EmptyTree` is returned."""
632
if revision_id == None:
635
inv = self.get_revision_inventory(revision_id)
636
return RevisionTree(self.text_store, inv)
639
def working_tree(self):
640
"""Return a `Tree` for the working copy."""
641
return WorkingTree(self.base, self.read_working_inventory())
644
def basis_tree(self):
645
"""Return `Tree` object for last revision.
647
If there are no revisions yet, return an `EmptyTree`.
649
>>> b = ScratchBranch(files=['foo'])
650
>>> b.basis_tree().has_filename('foo')
652
>>> b.working_tree().has_filename('foo')
655
>>> b.commit('add foo')
656
>>> b.basis_tree().has_filename('foo')
659
r = self.last_patch()
663
return RevisionTree(self.text_store, self.get_revision_inventory(r))
667
def write_log(self, show_timezone='original'):
668
"""Write out human-readable log of commits to this branch
670
:param utc: If true, show dates in universal time, not local time."""
671
## TODO: Option to choose either original, utc or local timezone
674
for p in self.revision_history():
676
print 'revno:', revno
677
## TODO: Show hash if --id is given.
678
##print 'revision-hash:', p
679
rev = self.get_revision(p)
680
print 'committer:', rev.committer
681
print 'timestamp: %s' % (format_date(rev.timestamp, rev.timezone or 0,
684
## opportunistic consistency check, same as check_patch_chaining
685
if rev.precursor != precursor:
686
bailout("mismatched precursor!")
690
print ' (no message)'
692
for l in rev.message.split('\n'):
700
def rename(self, from_paths, to_name):
703
If to_name exists and is a directory, the files are moved into
704
it, keeping their old names. If it is a directory,
706
Note that to_name is only the last component of the new name;
707
this doesn't change the directory.
709
## TODO: Option to move IDs only
710
assert not isinstance(from_paths, basestring)
711
tree = self.working_tree()
713
dest_dir = isdir(self.abspath(to_name))
715
# TODO: Wind back properly if some can't be moved?
716
dest_dir_id = inv.path2id(to_name)
717
if not dest_dir_id and to_name != '':
718
bailout("destination %r is not a versioned directory" % to_name)
720
name_tail = splitpath(f)[-1]
721
dest_path = appendpath(to_name, name_tail)
722
print "%s => %s" % (f, dest_path)
723
inv.rename(inv.path2id(f), dest_dir_id, name_tail)
724
os.rename(self.abspath(f), self.abspath(dest_path))
725
self._write_inventory(inv)
727
if len(from_paths) != 1:
728
bailout("when moving multiple files, destination must be a directory")
729
bailout("rename to non-directory %r not implemented sorry" % to_name)
733
def show_status(branch, show_all=False):
734
"""Display single-line status for non-ignored working files.
736
The list is show sorted in order by file name.
738
>>> b = ScratchBranch(files=['foo', 'foo~'])
744
>>> b.commit("add foo")
746
>>> os.unlink(b.abspath('foo'))
751
:todo: Get state for single files.
753
:todo: Perhaps show a slash at the end of directory names.
757
# We have to build everything into a list first so that it can
758
# sorted by name, incorporating all the different sources.
760
# FIXME: Rather than getting things in random order and then sorting,
761
# just step through in order.
763
# Interesting case: the old ID for a file has been removed,
764
# but a new file has been created under that name.
766
old = branch.basis_tree()
767
old_inv = old.inventory
768
new = branch.working_tree()
769
new_inv = new.inventory
771
for fs, fid, oldname, newname, kind in diff_trees(old, new):
773
show_status(fs, kind,
774
oldname + ' => ' + newname)
775
elif fs == 'A' or fs == 'M':
776
show_status(fs, kind, newname)
778
show_status(fs, kind, oldname)
781
show_status(fs, kind, newname)
784
show_status(fs, kind, newname)
786
show_status(fs, kind, newname)
788
bailout("wierd file state %r" % ((fs, fid),))
792
class ScratchBranch(Branch):
793
"""Special test class: a branch that cleans up after itself.
795
>>> b = ScratchBranch()
803
def __init__(self, files=[], dirs=[]):
804
"""Make a test branch.
806
This creates a temporary directory and runs init-tree in it.
808
If any files are listed, they are created in the working copy.
810
Branch.__init__(self, tempfile.mkdtemp(), init=True)
812
os.mkdir(self.abspath(d))
815
file(os.path.join(self.base, f), 'w').write('content of %s' % f)
819
"""Destroy the test branch, removing the scratch directory."""
820
shutil.rmtree(self.base)
824
######################################################################
828
def is_control_file(filename):
829
## FIXME: better check
830
filename = os.path.normpath(filename)
831
while filename != '':
832
head, tail = os.path.split(filename)
833
## mutter('check %r for control file' % ((head, tail), ))
834
if tail == bzrlib.BZRDIR:
843
def _gen_revision_id(when):
844
"""Return new revision-id."""
845
s = '%s-%s-' % (user_email(), compact_date(when))
846
s += hexlify(rand_bytes(8))
850
def gen_file_id(name):
851
"""Return new file id.
853
This should probably generate proper UUIDs, but for the moment we
854
cope with just randomness because running uuidgen every time is
856
idx = name.rfind('/')
858
name = name[idx+1 : ]
860
name = name.lstrip('.')
862
s = hexlify(rand_bytes(8))
863
return '-'.join((name, compact_date(time.time()), s))