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, file_kind, isdir, splitpath, relpath
52
from bzrlib.errors import BzrCheckError, DivergedBranches, NotVersionedError
53
from bzrlib.trace import mutter
57
class TreeEntry(object):
58
"""An entry that implements the minium interface used by commands.
60
This needs further inspection, it may be better to have
61
InventoryEntries without ids - though that seems wrong. For now,
62
this is a parallel hierarchy to InventoryEntry, and needs to become
63
one of several things: decorates to that hierarchy, children of, or
65
Another note is that these objects are currently only used when there is
66
no InventoryEntry available - i.e. for unversioned objects.
67
Perhaps they should be UnversionedEntry et al. ? - RBC 20051003
70
def __eq__(self, other):
71
# yes, this us ugly, TODO: best practice __eq__ style.
72
return (isinstance(other, TreeEntry)
73
and other.__class__ == self.__class__)
75
def kind_character(self):
79
class TreeDirectory(TreeEntry):
80
"""See TreeEntry. This is a directory in a working tree."""
82
def __eq__(self, other):
83
return (isinstance(other, TreeDirectory)
84
and other.__class__ == self.__class__)
86
def kind_character(self):
90
class TreeFile(TreeEntry):
91
"""See TreeEntry. This is a regular file in a working tree."""
93
def __eq__(self, other):
94
return (isinstance(other, TreeFile)
95
and other.__class__ == self.__class__)
97
def kind_character(self):
101
class TreeLink(TreeEntry):
102
"""See TreeEntry. This is a symlink in a working tree."""
104
def __eq__(self, other):
105
return (isinstance(other, TreeLink)
106
and other.__class__ == self.__class__)
108
def kind_character(self):
112
class WorkingTree(bzrlib.tree.Tree):
113
"""Working copy tree.
115
The inventory is held in the `Branch` working-inventory, and the
116
files are in a directory on disk.
118
It is possible for a `WorkingTree` to have a filename which is
119
not listed in the Inventory and vice versa.
122
def __init__(self, basedir, branch=None):
123
"""Construct a WorkingTree for basedir.
125
If the branch is not supplied, it is opened automatically.
126
If the branch is supplied, it must be the branch for this basedir.
127
(branch.base is not cross checked, because for remote branches that
128
would be meaningless).
130
from bzrlib.hashcache import HashCache
131
from bzrlib.trace import note, mutter
132
assert isinstance(basedir, basestring), \
133
"base directory %r is not a string" % basedir
135
branch = Branch.open(basedir)
136
assert isinstance(branch, Branch), \
137
"branch %r is not a Branch" % branch
139
self.basedir = basedir
140
self._inventory = self.read_working_inventory()
141
self.path2id = self._inventory.path2id
143
# update the whole cache up front and write to disk if anything changed;
144
# in the future we might want to do this more selectively
145
# two possible ways offer themselves : in self._unlock, write the cache
146
# if needed, or, when the cache sees a change, append it to the hash
147
# cache file, and have the parser take the most recent entry for a
149
hc = self._hashcache = HashCache(basedir)
158
"""Iterate through file_ids for this tree.
160
file_ids are in a WorkingTree if they are in the working inventory
161
and the working file exists.
163
inv = self._inventory
164
for path, ie in inv.iter_entries():
165
if bzrlib.osutils.lexists(self.abspath(path)):
170
return "<%s of %s>" % (self.__class__.__name__,
171
getattr(self, 'basedir', None))
175
def abspath(self, filename):
176
return os.path.join(self.basedir, filename)
178
def relpath(self, abspath):
179
"""Return the local path portion from a given absolute path."""
180
return relpath(self.basedir, abspath)
182
def has_filename(self, filename):
183
return bzrlib.osutils.lexists(self.abspath(filename))
185
def get_file(self, file_id):
186
return self.get_file_byname(self.id2path(file_id))
188
def get_file_byname(self, filename):
189
return file(self.abspath(filename), 'rb')
191
def get_root_id(self):
192
"""Return the id of this trees root"""
193
inv = self.read_working_inventory()
194
return inv.root.file_id
196
def _get_store_filename(self, file_id):
197
## XXX: badly named; this isn't in the store at all
198
return self.abspath(self.id2path(file_id))
201
def id2abspath(self, file_id):
202
return self.abspath(self.id2path(file_id))
205
def has_id(self, file_id):
206
# files that have been deleted are excluded
207
inv = self._inventory
208
if not inv.has_id(file_id):
210
path = inv.id2path(file_id)
211
return bzrlib.osutils.lexists(self.abspath(path))
213
def has_or_had_id(self, file_id):
214
if file_id == self.inventory.root.file_id:
216
return self.inventory.has_id(file_id)
218
__contains__ = has_id
221
def get_file_size(self, file_id):
222
return os.path.getsize(self.id2abspath(file_id))
224
def get_file_sha1(self, file_id):
225
path = self._inventory.id2path(file_id)
226
return self._hashcache.get_sha1(path)
229
def is_executable(self, file_id):
231
return self._inventory[file_id].executable
233
path = self._inventory.id2path(file_id)
234
mode = os.lstat(self.abspath(path)).st_mode
235
return bool(stat.S_ISREG(mode) and stat.S_IEXEC&mode)
237
def get_symlink_target(self, file_id):
238
return os.readlink(self.id2abspath(file_id))
240
def file_class(self, filename):
241
if self.path2id(filename):
243
elif self.is_ignored(filename):
249
def list_files(self):
250
"""Recursively list all files as (path, class, kind, id).
252
Lists, but does not descend into unversioned directories.
254
This does not include files that have been deleted in this
257
Skips the control directory.
259
inv = self._inventory
261
def descend(from_dir_relpath, from_dir_id, dp):
265
## TODO: If we find a subdirectory with its own .bzr
266
## directory, then that is a separate tree and we
267
## should exclude it.
268
if bzrlib.BZRDIR == f:
272
fp = appendpath(from_dir_relpath, f)
275
fap = appendpath(dp, f)
277
f_ie = inv.get_child(from_dir_id, f)
280
elif self.is_ignored(fp):
289
raise BzrCheckError("file %r entered as kind %r id %r, "
291
% (fap, f_ie.kind, f_ie.file_id, fk))
293
# make a last minute entry
297
if fk == 'directory':
298
entry = TreeDirectory()
301
elif fk == 'symlink':
306
yield fp, c, fk, (f_ie and f_ie.file_id), entry
308
if fk != 'directory':
312
# don't descend unversioned directories
315
for ff in descend(fp, f_ie.file_id, fap):
318
for f in descend('', inv.root.file_id, self.basedir):
324
for subp in self.extras():
325
if not self.is_ignored(subp):
328
def iter_conflicts(self):
330
for path in (s[0] for s in self.list_files()):
331
stem = get_conflicted_stem(path)
334
if stem not in conflicted:
339
def pull(self, source, overwrite=False):
340
from bzrlib.merge import merge_inner
343
old_revision_history = self.branch.revision_history()
344
self.branch.pull(source, overwrite)
345
new_revision_history = self.branch.revision_history()
346
if new_revision_history != old_revision_history:
347
if len(old_revision_history):
348
other_revision = old_revision_history[-1]
350
other_revision = None
351
merge_inner(self.branch,
352
self.branch.basis_tree(),
353
self.branch.revision_tree(other_revision))
358
"""Yield all unknown files in this WorkingTree.
360
If there are any unknown directories then only the directory is
361
returned, not all its children. But if there are unknown files
362
under a versioned subdirectory, they are returned.
364
Currently returned depth-first, sorted by name within directories.
366
## TODO: Work from given directory downwards
367
for path, dir_entry in self.inventory.directories():
368
mutter("search for unknowns in %r" % path)
369
dirabs = self.abspath(path)
370
if not isdir(dirabs):
371
# e.g. directory deleted
375
for subf in os.listdir(dirabs):
377
and (subf not in dir_entry.children)):
382
subp = appendpath(path, subf)
386
def ignored_files(self):
387
"""Yield list of PATH, IGNORE_PATTERN"""
388
for subp in self.extras():
389
pat = self.is_ignored(subp)
394
def get_ignore_list(self):
395
"""Return list of ignore patterns.
397
Cached in the Tree object after the first call.
399
if hasattr(self, '_ignorelist'):
400
return self._ignorelist
402
l = bzrlib.DEFAULT_IGNORE[:]
403
if self.has_filename(bzrlib.IGNORE_FILENAME):
404
f = self.get_file_byname(bzrlib.IGNORE_FILENAME)
405
l.extend([line.rstrip("\n\r") for line in f.readlines()])
410
def is_ignored(self, filename):
411
r"""Check whether the filename matches an ignore pattern.
413
Patterns containing '/' or '\' need to match the whole path;
414
others match against only the last component.
416
If the file is ignored, returns the pattern which caused it to
417
be ignored, otherwise None. So this can simply be used as a
418
boolean if desired."""
420
# TODO: Use '**' to match directories, and other extended
421
# globbing stuff from cvs/rsync.
423
# XXX: fnmatch is actually not quite what we want: it's only
424
# approximately the same as real Unix fnmatch, and doesn't
425
# treat dotfiles correctly and allows * to match /.
426
# Eventually it should be replaced with something more
429
for pat in self.get_ignore_list():
430
if '/' in pat or '\\' in pat:
432
# as a special case, you can put ./ at the start of a
433
# pattern; this is good to match in the top-level
436
if (pat[:2] == './') or (pat[:2] == '.\\'):
440
if fnmatch.fnmatchcase(filename, newpat):
443
if fnmatch.fnmatchcase(splitpath(filename)[-1], pat):
448
def kind(self, file_id):
449
return file_kind(self.id2abspath(file_id))
452
"""See Branch.lock_read, and WorkingTree.unlock."""
453
return self.branch.lock_read()
455
def lock_write(self):
456
"""See Branch.lock_write, and WorkingTree.unlock."""
457
return self.branch.lock_write()
460
def read_working_inventory(self):
461
"""Read the working inventory."""
462
# ElementTree does its own conversion from UTF-8, so open in
464
f = self.branch.controlfile('inventory', 'rb')
465
return bzrlib.xml5.serializer_v5.read_inventory(f)
468
def remove(self, files, verbose=False):
469
"""Remove nominated files from the working inventory..
471
This does not remove their text. This does not run on XXX on what? RBC
473
TODO: Refuse to remove modified files unless --force is given?
475
TODO: Do something useful with directories.
477
TODO: Should this remove the text or not? Tough call; not
478
removing may be useful and the user can just use use rm, and
479
is the opposite of add. Removing it is consistent with most
480
other tools. Maybe an option.
482
## TODO: Normalize names
483
## TODO: Remove nested loops; better scalability
484
if isinstance(files, basestring):
489
# do this before any modifications
493
# TODO: Perhaps make this just a warning, and continue?
494
# This tends to happen when
495
raise NotVersionedError(path=f)
496
mutter("remove inventory entry %s {%s}" % (quotefn(f), fid))
498
# having remove it, it must be either ignored or unknown
499
if self.is_ignored(f):
503
show_status(new_status, inv[fid].kind, quotefn(f))
506
self.branch._write_inventory(inv)
509
def set_inventory(self, new_inventory_list):
510
from bzrlib.inventory import (Inventory,
515
inv = Inventory(self.get_root_id())
516
for path, file_id, parent, kind in new_inventory_list:
517
name = os.path.basename(path)
520
# fixme, there should be a factory function inv,add_??
521
if kind == 'directory':
522
inv.add(InventoryDirectory(file_id, name, parent))
524
inv.add(InventoryFile(file_id, name, parent))
525
elif kind == 'symlink':
526
inv.add(InventoryLink(file_id, name, parent))
528
raise BzrError("unknown kind %r" % kind)
529
self.branch._write_inventory(inv)
532
"""See Branch.unlock.
534
WorkingTree locking just uses the Branch locking facilities.
535
This is current because all working trees have an embedded branch
536
within them. IF in the future, we were to make branch data shareable
537
between multiple working trees, i.e. via shared storage, then we
538
would probably want to lock both the local tree, and the branch.
540
return self.branch.unlock()
543
CONFLICT_SUFFIXES = ('.THIS', '.BASE', '.OTHER')
544
def get_conflicted_stem(path):
545
for suffix in CONFLICT_SUFFIXES:
546
if path.endswith(suffix):
547
return path[:-len(suffix)]