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
17
"""WorkingTree object and friends.
19
A WorkingTree represents the editable working copy of a branch.
20
Operations which represent the WorkingTree are also done here,
21
such as renaming or adding files. The WorkingTree has an inventory
22
which is updated by these operations. A commit produces a
23
new revision based on the workingtree and its inventory.
25
At the moment every WorkingTree has its own branch. Remote
26
WorkingTrees aren't supported.
28
To get a WorkingTree, call Branch.working_tree():
32
# TODO: Don't allow WorkingTrees to be constructed for remote branches if
35
# FIXME: I don't know if writing out the cache from the destructor is really a
36
# good idea, because destructors are considered poor taste in Python, and it's
37
# not predictable when it will be written out.
39
# TODO: Give the workingtree sole responsibility for the working inventory;
40
# remove the variable and references to it from the branch. This may require
41
# updating the commit code so as to update the inventory within the working
42
# copy, and making sure there's only one WorkingTree for any directory on disk.
43
# At the momenthey may alias the inventory and have old copies of it in memory.
49
from bzrlib.branch import Branch, needs_read_lock, needs_write_lock, quotefn
51
from bzrlib.osutils import (appendpath,
57
from bzrlib.errors import BzrCheckError, DivergedBranches, NotVersionedError
58
from bzrlib.trace import mutter
62
class TreeEntry(object):
63
"""An entry that implements the minium interface used by commands.
65
This needs further inspection, it may be better to have
66
InventoryEntries without ids - though that seems wrong. For now,
67
this is a parallel hierarchy to InventoryEntry, and needs to become
68
one of several things: decorates to that hierarchy, children of, or
70
Another note is that these objects are currently only used when there is
71
no InventoryEntry available - i.e. for unversioned objects.
72
Perhaps they should be UnversionedEntry et al. ? - RBC 20051003
75
def __eq__(self, other):
76
# yes, this us ugly, TODO: best practice __eq__ style.
77
return (isinstance(other, TreeEntry)
78
and other.__class__ == self.__class__)
80
def kind_character(self):
84
class TreeDirectory(TreeEntry):
85
"""See TreeEntry. This is a directory in a working tree."""
87
def __eq__(self, other):
88
return (isinstance(other, TreeDirectory)
89
and other.__class__ == self.__class__)
91
def kind_character(self):
95
class TreeFile(TreeEntry):
96
"""See TreeEntry. This is a regular file in a working tree."""
98
def __eq__(self, other):
99
return (isinstance(other, TreeFile)
100
and other.__class__ == self.__class__)
102
def kind_character(self):
106
class TreeLink(TreeEntry):
107
"""See TreeEntry. This is a symlink in a working tree."""
109
def __eq__(self, other):
110
return (isinstance(other, TreeLink)
111
and other.__class__ == self.__class__)
113
def kind_character(self):
117
class WorkingTree(bzrlib.tree.Tree):
118
"""Working copy tree.
120
The inventory is held in the `Branch` working-inventory, and the
121
files are in a directory on disk.
123
It is possible for a `WorkingTree` to have a filename which is
124
not listed in the Inventory and vice versa.
127
def __init__(self, basedir, branch=None):
128
"""Construct a WorkingTree for basedir.
130
If the branch is not supplied, it is opened automatically.
131
If the branch is supplied, it must be the branch for this basedir.
132
(branch.base is not cross checked, because for remote branches that
133
would be meaningless).
135
from bzrlib.hashcache import HashCache
136
from bzrlib.trace import note, mutter
137
assert isinstance(basedir, basestring), \
138
"base directory %r is not a string" % basedir
140
branch = Branch.open(basedir)
141
assert isinstance(branch, Branch), \
142
"branch %r is not a Branch" % branch
144
self.basedir = basedir
145
self._inventory = self.read_working_inventory()
146
self.path2id = self._inventory.path2id
148
# update the whole cache up front and write to disk if anything changed;
149
# in the future we might want to do this more selectively
150
# two possible ways offer themselves : in self._unlock, write the cache
151
# if needed, or, when the cache sees a change, append it to the hash
152
# cache file, and have the parser take the most recent entry for a
154
hc = self._hashcache = HashCache(basedir)
163
"""Iterate through file_ids for this tree.
165
file_ids are in a WorkingTree if they are in the working inventory
166
and the working file exists.
168
inv = self._inventory
169
for path, ie in inv.iter_entries():
170
if bzrlib.osutils.lexists(self.abspath(path)):
175
return "<%s of %s>" % (self.__class__.__name__,
176
getattr(self, 'basedir', None))
180
def abspath(self, filename):
181
return os.path.join(self.basedir, filename)
183
def relpath(self, abspath):
184
"""Return the local path portion from a given absolute path."""
185
return relpath(self.basedir, abspath)
187
def has_filename(self, filename):
188
return bzrlib.osutils.lexists(self.abspath(filename))
190
def get_file(self, file_id):
191
return self.get_file_byname(self.id2path(file_id))
193
def get_file_byname(self, filename):
194
return file(self.abspath(filename), 'rb')
196
def get_root_id(self):
197
"""Return the id of this trees root"""
198
inv = self.read_working_inventory()
199
return inv.root.file_id
201
def _get_store_filename(self, file_id):
202
## XXX: badly named; this isn't in the store at all
203
return self.abspath(self.id2path(file_id))
206
def commit(self, *args, **kw):
207
from bzrlib.commit import Commit
208
Commit().commit(self.branch, *args, **kw)
209
self._inventory = self.read_working_inventory()
211
def id2abspath(self, file_id):
212
return self.abspath(self.id2path(file_id))
215
def has_id(self, file_id):
216
# files that have been deleted are excluded
217
inv = self._inventory
218
if not inv.has_id(file_id):
220
path = inv.id2path(file_id)
221
return bzrlib.osutils.lexists(self.abspath(path))
223
def has_or_had_id(self, file_id):
224
if file_id == self.inventory.root.file_id:
226
return self.inventory.has_id(file_id)
228
__contains__ = has_id
231
def get_file_size(self, file_id):
232
return os.path.getsize(self.id2abspath(file_id))
234
def get_file_sha1(self, file_id):
235
path = self._inventory.id2path(file_id)
236
return self._hashcache.get_sha1(path)
239
def is_executable(self, file_id):
241
return self._inventory[file_id].executable
243
path = self._inventory.id2path(file_id)
244
mode = os.lstat(self.abspath(path)).st_mode
245
return bool(stat.S_ISREG(mode) and stat.S_IEXEC&mode)
248
def add_pending_merge(self, *revision_ids):
249
# TODO: Perhaps should check at this point that the
250
# history of the revision is actually present?
251
p = self.pending_merges()
253
for rev_id in revision_ids:
259
self.set_pending_merges(p)
261
def pending_merges(self):
262
"""Return a list of pending merges.
264
These are revisions that have been merged into the working
265
directory but not yet committed.
267
cfn = self.branch._rel_controlfilename('pending-merges')
268
if not self.branch._transport.has(cfn):
271
for l in self.branch.controlfile('pending-merges', 'r').readlines():
272
p.append(l.rstrip('\n'))
276
def set_pending_merges(self, rev_list):
277
self.branch.put_controlfile('pending-merges', '\n'.join(rev_list))
279
def get_symlink_target(self, file_id):
280
return os.readlink(self.id2abspath(file_id))
282
def file_class(self, filename):
283
if self.path2id(filename):
285
elif self.is_ignored(filename):
291
def list_files(self):
292
"""Recursively list all files as (path, class, kind, id).
294
Lists, but does not descend into unversioned directories.
296
This does not include files that have been deleted in this
299
Skips the control directory.
301
inv = self._inventory
303
def descend(from_dir_relpath, from_dir_id, dp):
307
## TODO: If we find a subdirectory with its own .bzr
308
## directory, then that is a separate tree and we
309
## should exclude it.
310
if bzrlib.BZRDIR == f:
314
fp = appendpath(from_dir_relpath, f)
317
fap = appendpath(dp, f)
319
f_ie = inv.get_child(from_dir_id, f)
322
elif self.is_ignored(fp):
331
raise BzrCheckError("file %r entered as kind %r id %r, "
333
% (fap, f_ie.kind, f_ie.file_id, fk))
335
# make a last minute entry
339
if fk == 'directory':
340
entry = TreeDirectory()
343
elif fk == 'symlink':
348
yield fp, c, fk, (f_ie and f_ie.file_id), entry
350
if fk != 'directory':
354
# don't descend unversioned directories
357
for ff in descend(fp, f_ie.file_id, fap):
360
for f in descend('', inv.root.file_id, self.basedir):
366
for subp in self.extras():
367
if not self.is_ignored(subp):
370
def iter_conflicts(self):
372
for path in (s[0] for s in self.list_files()):
373
stem = get_conflicted_stem(path)
376
if stem not in conflicted:
381
def pull(self, source, overwrite=False):
382
from bzrlib.merge import merge_inner
385
old_revision_history = self.branch.revision_history()
386
self.branch.pull(source, overwrite)
387
new_revision_history = self.branch.revision_history()
388
if new_revision_history != old_revision_history:
389
if len(old_revision_history):
390
other_revision = old_revision_history[-1]
392
other_revision = None
393
merge_inner(self.branch,
394
self.branch.basis_tree(),
395
self.branch.revision_tree(other_revision))
396
return len(new_revision_history) - len(old_revision_history)
401
"""Yield all unknown files in this WorkingTree.
403
If there are any unknown directories then only the directory is
404
returned, not all its children. But if there are unknown files
405
under a versioned subdirectory, they are returned.
407
Currently returned depth-first, sorted by name within directories.
409
## TODO: Work from given directory downwards
410
for path, dir_entry in self.inventory.directories():
411
mutter("search for unknowns in %r", path)
412
dirabs = self.abspath(path)
413
if not isdir(dirabs):
414
# e.g. directory deleted
418
for subf in os.listdir(dirabs):
420
and (subf not in dir_entry.children)):
425
subp = appendpath(path, subf)
429
def ignored_files(self):
430
"""Yield list of PATH, IGNORE_PATTERN"""
431
for subp in self.extras():
432
pat = self.is_ignored(subp)
437
def get_ignore_list(self):
438
"""Return list of ignore patterns.
440
Cached in the Tree object after the first call.
442
if hasattr(self, '_ignorelist'):
443
return self._ignorelist
445
l = bzrlib.DEFAULT_IGNORE[:]
446
if self.has_filename(bzrlib.IGNORE_FILENAME):
447
f = self.get_file_byname(bzrlib.IGNORE_FILENAME)
448
l.extend([line.rstrip("\n\r") for line in f.readlines()])
453
def is_ignored(self, filename):
454
r"""Check whether the filename matches an ignore pattern.
456
Patterns containing '/' or '\' need to match the whole path;
457
others match against only the last component.
459
If the file is ignored, returns the pattern which caused it to
460
be ignored, otherwise None. So this can simply be used as a
461
boolean if desired."""
463
# TODO: Use '**' to match directories, and other extended
464
# globbing stuff from cvs/rsync.
466
# XXX: fnmatch is actually not quite what we want: it's only
467
# approximately the same as real Unix fnmatch, and doesn't
468
# treat dotfiles correctly and allows * to match /.
469
# Eventually it should be replaced with something more
472
for pat in self.get_ignore_list():
473
if '/' in pat or '\\' in pat:
475
# as a special case, you can put ./ at the start of a
476
# pattern; this is good to match in the top-level
479
if (pat[:2] == './') or (pat[:2] == '.\\'):
483
if fnmatch.fnmatchcase(filename, newpat):
486
if fnmatch.fnmatchcase(splitpath(filename)[-1], pat):
491
def kind(self, file_id):
492
return file_kind(self.id2abspath(file_id))
495
"""See Branch.lock_read, and WorkingTree.unlock."""
496
return self.branch.lock_read()
498
def lock_write(self):
499
"""See Branch.lock_write, and WorkingTree.unlock."""
500
return self.branch.lock_write()
503
def read_working_inventory(self):
504
"""Read the working inventory."""
505
# ElementTree does its own conversion from UTF-8, so open in
507
f = self.branch.controlfile('inventory', 'rb')
508
return bzrlib.xml5.serializer_v5.read_inventory(f)
511
def remove(self, files, verbose=False):
512
"""Remove nominated files from the working inventory..
514
This does not remove their text. This does not run on XXX on what? RBC
516
TODO: Refuse to remove modified files unless --force is given?
518
TODO: Do something useful with directories.
520
TODO: Should this remove the text or not? Tough call; not
521
removing may be useful and the user can just use use rm, and
522
is the opposite of add. Removing it is consistent with most
523
other tools. Maybe an option.
525
## TODO: Normalize names
526
## TODO: Remove nested loops; better scalability
527
if isinstance(files, basestring):
532
# do this before any modifications
536
# TODO: Perhaps make this just a warning, and continue?
537
# This tends to happen when
538
raise NotVersionedError(path=f)
539
mutter("remove inventory entry %s {%s}", quotefn(f), fid)
541
# having remove it, it must be either ignored or unknown
542
if self.is_ignored(f):
546
show_status(new_status, inv[fid].kind, quotefn(f))
549
self._write_inventory(inv)
552
def revert(self, filenames, old_tree=None, backups=True):
553
from bzrlib.merge import merge_inner
555
old_tree = self.branch.basis_tree()
556
merge_inner(self.branch, old_tree,
557
self, ignore_zero=True,
558
backup_files=backups,
559
interesting_files=filenames)
560
if not len(filenames):
561
self.set_pending_merges([])
564
def set_inventory(self, new_inventory_list):
565
from bzrlib.inventory import (Inventory,
570
inv = Inventory(self.get_root_id())
571
for path, file_id, parent, kind in new_inventory_list:
572
name = os.path.basename(path)
575
# fixme, there should be a factory function inv,add_??
576
if kind == 'directory':
577
inv.add(InventoryDirectory(file_id, name, parent))
579
inv.add(InventoryFile(file_id, name, parent))
580
elif kind == 'symlink':
581
inv.add(InventoryLink(file_id, name, parent))
583
raise BzrError("unknown kind %r" % kind)
584
self._write_inventory(inv)
587
def set_root_id(self, file_id):
588
"""Set the root id for this tree."""
589
inv = self.read_working_inventory()
590
orig_root_id = inv.root.file_id
591
del inv._byid[inv.root.file_id]
592
inv.root.file_id = file_id
593
inv._byid[inv.root.file_id] = inv.root
596
if entry.parent_id in (None, orig_root_id):
597
entry.parent_id = inv.root.file_id
598
self._write_inventory(inv)
601
"""See Branch.unlock.
603
WorkingTree locking just uses the Branch locking facilities.
604
This is current because all working trees have an embedded branch
605
within them. IF in the future, we were to make branch data shareable
606
between multiple working trees, i.e. via shared storage, then we
607
would probably want to lock both the local tree, and the branch.
609
return self.branch.unlock()
612
def _write_inventory(self, inv):
613
"""Write inventory as the current inventory."""
614
from cStringIO import StringIO
615
from bzrlib.atomicfile import AtomicFile
617
bzrlib.xml5.serializer_v5.write_inventory(inv, sio)
619
f = AtomicFile(self.branch.controlfilename('inventory'))
625
mutter('wrote working inventory')
628
CONFLICT_SUFFIXES = ('.THIS', '.BASE', '.OTHER')
629
def get_conflicted_stem(path):
630
for suffix in CONFLICT_SUFFIXES:
631
if path.endswith(suffix):
632
return path[:-len(suffix)]
added
removed
bzrlib/workingtree.py
