1
# (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
18
# This should really be an id randomly assigned when the tree is
19
# created, but it's not for now.
23
import sys, os.path, types, re
26
from cElementTree import Element, ElementTree, SubElement
28
from elementtree.ElementTree import Element, ElementTree, SubElement
30
from xml import XMLMixin
31
from errors import bailout, BzrError, BzrCheckError
34
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
35
from bzrlib.trace import mutter
37
class InventoryEntry(XMLMixin):
38
"""Description of a versioned file.
40
An InventoryEntry has the following fields, which are also
41
present in the XML inventory-entry element:
44
* *name*: (only the basename within the directory, must not
46
* *kind*: "directory" or "file"
47
* *directory_id*: (if absent/null means the branch root directory)
48
* *text_sha1*: only for files
49
* *text_size*: in bytes, only for files
50
* *text_id*: identifier for the text version, only for files
52
InventoryEntries can also exist inside a WorkingTree
53
inventory, in which case they are not yet bound to a
54
particular revision of the file. In that case the text_sha1,
55
text_size and text_id are absent.
61
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
62
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
63
>>> for j in i.iter_entries():
66
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
67
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
68
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
69
Traceback (most recent call last):
71
BzrError: ('inventory already contains entry with id {2323}', [])
72
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
73
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
74
>>> i.path2id('src/wibble')
78
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
80
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
81
>>> for j in i.iter_entries():
83
... assert i.path2id(j[0])
93
TODO: Maybe also keep the full path of the entry, and the children?
94
But those depend on its position within a particular inventory, and
95
it would be nice not to need to hold the backpointer here.
98
# TODO: split InventoryEntry into subclasses for files,
99
# directories, etc etc.
104
def __init__(self, file_id, name, kind, parent_id, text_id=None):
105
"""Create an InventoryEntry
107
The filename must be a single component, relative to the
108
parent directory; it cannot be a whole path or relative name.
110
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
115
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
116
Traceback (most recent call last):
117
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
119
if '/' in name or '\\' in name:
120
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
122
self.file_id = file_id
125
self.text_id = text_id
126
self.parent_id = parent_id
127
if kind == 'directory':
132
raise BzrError("unhandled entry kind %r" % kind)
136
def sorted_children(self):
137
l = self.children.items()
143
other = InventoryEntry(self.file_id, self.name, self.kind,
144
self.parent_id, text_id=self.text_id)
145
other.text_sha1 = self.text_sha1
146
other.text_size = self.text_size
147
# note that children are *not* copied; they're pulled across when
153
return ("%s(%r, %r, kind=%r, parent_id=%r)"
154
% (self.__class__.__name__,
161
def to_element(self):
162
"""Convert to XML element"""
165
e.set('name', self.name)
166
e.set('file_id', self.file_id)
167
e.set('kind', self.kind)
169
if self.text_size != None:
170
e.set('text_size', '%d' % self.text_size)
172
for f in ['text_id', 'text_sha1']:
177
# to be conservative, we don't externalize the root pointers
178
# for now, leaving them as null in the xml form. in a future
179
# version it will be implied by nested elements.
180
if self.parent_id != ROOT_ID:
181
assert isinstance(self.parent_id, basestring)
182
e.set('parent_id', self.parent_id)
189
def from_element(cls, elt):
190
assert elt.tag == 'entry'
192
## original format inventories don't have a parent_id for
193
## nodes in the root directory, but it's cleaner to use one
195
parent_id = elt.get('parent_id')
196
if parent_id == None:
199
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
200
self.text_id = elt.get('text_id')
201
self.text_sha1 = elt.get('text_sha1')
203
## mutter("read inventoryentry: %r" % (elt.attrib))
205
v = elt.get('text_size')
206
self.text_size = v and int(v)
211
from_element = classmethod(from_element)
213
def __eq__(self, other):
214
if not isinstance(other, InventoryEntry):
215
return NotImplemented
217
return (self.file_id == other.file_id) \
218
and (self.name == other.name) \
219
and (self.text_sha1 == other.text_sha1) \
220
and (self.text_size == other.text_size) \
221
and (self.text_id == other.text_id) \
222
and (self.parent_id == other.parent_id) \
223
and (self.kind == other.kind)
226
def __ne__(self, other):
227
return not (self == other)
230
raise ValueError('not hashable')
234
class RootEntry(InventoryEntry):
235
def __init__(self, file_id):
236
self.file_id = file_id
238
self.kind = 'root_directory'
239
self.parent_id = None
242
def __eq__(self, other):
243
if not isinstance(other, RootEntry):
244
return NotImplemented
246
return (self.file_id == other.file_id) \
247
and (self.children == other.children)
251
class Inventory(XMLMixin):
252
"""Inventory of versioned files in a tree.
254
This describes which file_id is present at each point in the tree,
255
and possibly the SHA-1 or other information about the file.
256
Entries can be looked up either by path or by file_id.
258
The inventory represents a typical unix file tree, with
259
directories containing files and subdirectories. We never store
260
the full path to a file, because renaming a directory implicitly
261
moves all of its contents. This class internally maintains a
262
lookup tree that allows the children under a directory to be
265
InventoryEntry objects must not be modified after they are
266
inserted, other than through the Inventory API.
268
>>> inv = Inventory()
269
>>> inv.write_xml(sys.stdout)
272
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
273
>>> inv['123-123'].name
276
May be treated as an iterator or set to look up file ids:
278
>>> bool(inv.path2id('hello.c'))
283
May also look up by name:
285
>>> [x[0] for x in inv.iter_entries()]
288
>>> inv.write_xml(sys.stdout)
290
<entry file_id="123-123" kind="file" name="hello.c" />
295
"""Create or read an inventory.
297
If a working directory is specified, the inventory is read
298
from there. If the file is specified, read from that. If not,
299
the inventory is created empty.
301
The inventory is created with a default root directory, with
304
self.root = RootEntry(ROOT_ID)
305
self._byid = {self.root.file_id: self.root}
309
return iter(self._byid)
313
"""Returns number of entries."""
314
return len(self._byid)
317
def iter_entries(self, from_dir=None):
318
"""Return (path, entry) pairs, in order by name."""
322
elif isinstance(from_dir, basestring):
323
from_dir = self._byid[from_dir]
325
kids = from_dir.children.items()
327
for name, ie in kids:
329
if ie.kind == 'directory':
330
for cn, cie in self.iter_entries(from_dir=ie.file_id):
331
yield os.path.join(name, cn), cie
335
def directories(self):
336
"""Return (path, entry) pairs for all directories.
338
def descend(parent_ie):
339
parent_name = parent_ie.name
340
yield parent_name, parent_ie
342
# directory children in sorted order
344
for ie in parent_ie.children.itervalues():
345
if ie.kind == 'directory':
346
dn.append((ie.name, ie))
349
for name, child_ie in dn:
350
for sub_name, sub_ie in descend(child_ie):
351
yield appendpath(parent_name, sub_name), sub_ie
353
for name, ie in descend(self.root):
358
def __contains__(self, file_id):
359
"""True if this entry contains a file with given id.
361
>>> inv = Inventory()
362
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
368
return file_id in self._byid
371
def __getitem__(self, file_id):
372
"""Return the entry for given file_id.
374
>>> inv = Inventory()
375
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
376
>>> inv['123123'].name
380
return self._byid[file_id]
383
raise BzrError("can't look up file_id None")
385
raise BzrError("file_id {%s} not in inventory" % file_id)
388
def get_file_kind(self, file_id):
389
return self._byid[file_id].kind
391
def get_child(self, parent_id, filename):
392
return self[parent_id].children.get(filename)
395
def add(self, entry):
396
"""Add entry to inventory.
398
To add a file to a branch ready to be committed, use Branch.add,
400
if entry.file_id in self._byid:
401
bailout("inventory already contains entry with id {%s}" % entry.file_id)
404
parent = self._byid[entry.parent_id]
406
bailout("parent_id {%s} not in inventory" % entry.parent_id)
408
if parent.children.has_key(entry.name):
409
bailout("%s is already versioned" %
410
appendpath(self.id2path(parent.file_id), entry.name))
412
self._byid[entry.file_id] = entry
413
parent.children[entry.name] = entry
416
def add_path(self, relpath, kind, file_id=None):
417
"""Add entry from a path.
419
The immediate parent must already be versioned"""
420
parts = bzrlib.osutils.splitpath(relpath)
422
bailout("cannot re-add root of inventory")
425
file_id = bzrlib.branch.gen_file_id(relpath)
427
parent_id = self.path2id(parts[:-1])
428
assert parent_id != None
429
ie = InventoryEntry(file_id, parts[-1],
430
kind=kind, parent_id=parent_id)
434
def __delitem__(self, file_id):
435
"""Remove entry by id.
437
>>> inv = Inventory()
438
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
447
assert self[ie.parent_id].children[ie.name] == ie
449
# TODO: Test deleting all children; maybe hoist to a separate
451
if ie.kind == 'directory':
452
for cie in ie.children.values():
453
del self[cie.file_id]
456
del self._byid[file_id]
457
del self[ie.parent_id].children[ie.name]
460
def to_element(self):
461
"""Convert to XML Element"""
462
e = Element('inventory')
464
for path, ie in self.iter_entries():
465
e.append(ie.to_element())
469
def from_element(cls, elt):
470
"""Construct from XML Element
472
>>> inv = Inventory()
473
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
474
>>> elt = inv.to_element()
475
>>> inv2 = Inventory.from_element(elt)
479
assert elt.tag == 'inventory'
482
o.add(InventoryEntry.from_element(e))
485
from_element = classmethod(from_element)
488
def __eq__(self, other):
489
"""Compare two sets by comparing their contents.
495
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
498
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
502
if not isinstance(other, Inventory):
503
return NotImplemented
505
if len(self._byid) != len(other._byid):
506
# shortcut: obviously not the same
509
return self._byid == other._byid
512
def __ne__(self, other):
513
return not (self == other)
517
raise ValueError('not hashable')
521
def get_idpath(self, file_id):
522
"""Return a list of file_ids for the path to an entry.
524
The list contains one element for each directory followed by
525
the id of the file itself. So the length of the returned list
526
is equal to the depth of the file in the tree, counting the
527
root directory as depth 1.
530
while file_id != None:
532
ie = self._byid[file_id]
534
bailout("file_id {%s} not found in inventory" % file_id)
535
p.insert(0, ie.file_id)
536
file_id = ie.parent_id
540
def id2path(self, file_id):
541
"""Return as a list the path to file_id."""
543
# get all names, skipping root
544
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
545
return os.sep.join(p)
549
def path2id(self, name):
550
"""Walk down through directories to return entry of last component.
552
names may be either a list of path components, or a single
553
string, in which case it is automatically split.
555
This returns the entry of the last component in the path,
556
which may be either a file or a directory.
558
Returns None iff the path is not found.
560
if isinstance(name, types.StringTypes):
561
name = splitpath(name)
563
mutter("lookup path %r" % name)
568
cie = parent.children[f]
570
assert cie.parent_id == parent.file_id
576
return parent.file_id
579
def has_filename(self, names):
580
return bool(self.path2id(names))
583
def has_id(self, file_id):
584
return self._byid.has_key(file_id)
587
def rename(self, file_id, new_parent_id, new_name):
588
"""Move a file within the inventory.
590
This can change either the name, or the parent, or both.
592
This does not move the working file."""
593
if not is_valid_name(new_name):
594
bailout("not an acceptable filename: %r" % new_name)
596
new_parent = self._byid[new_parent_id]
597
if new_name in new_parent.children:
598
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
600
new_parent_idpath = self.get_idpath(new_parent_id)
601
if file_id in new_parent_idpath:
602
bailout("cannot move directory %r into a subdirectory of itself, %r"
603
% (self.id2path(file_id), self.id2path(new_parent_id)))
605
file_ie = self._byid[file_id]
606
old_parent = self._byid[file_ie.parent_id]
608
# TODO: Don't leave things messed up if this fails
610
del old_parent.children[file_ie.name]
611
new_parent.children[new_name] = file_ie
613
file_ie.name = new_name
614
file_ie.parent_id = new_parent_id
619
_NAME_RE = re.compile(r'^[^/\\]+$')
621
def is_valid_name(name):
622
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
