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
27
from cElementTree import Element, ElementTree, SubElement
29
from elementtree.ElementTree import Element, ElementTree, SubElement
31
from xml import XMLMixin
32
from errors import bailout, BzrError, BzrCheckError
35
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
36
from bzrlib.trace import mutter
38
class InventoryEntry(XMLMixin):
39
"""Description of a versioned file.
41
An InventoryEntry has the following fields, which are also
42
present in the XML inventory-entry element:
45
* *name*: (only the basename within the directory, must not
47
* *kind*: "directory" or "file"
48
* *directory_id*: (if absent/null means the branch root directory)
49
* *text_sha1*: only for files
50
* *text_size*: in bytes, only for files
51
* *text_id*: identifier for the text version, only for files
53
InventoryEntries can also exist inside a WorkingTree
54
inventory, in which case they are not yet bound to a
55
particular revision of the file. In that case the text_sha1,
56
text_size and text_id are absent.
62
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
63
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
64
>>> for j in i.iter_entries():
67
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
68
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
69
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
70
Traceback (most recent call last):
72
BzrError: ('inventory already contains entry with id {2323}', [])
73
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
74
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
75
>>> i.path2id('src/wibble')
79
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
81
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
82
>>> for j in i.iter_entries():
84
... assert i.path2id(j[0])
94
TODO: Maybe also keep the full path of the entry, and the children?
95
But those depend on its position within a particular inventory, and
96
it would be nice not to need to hold the backpointer here.
99
# TODO: split InventoryEntry into subclasses for files,
100
# directories, etc etc.
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
106
"""Create an InventoryEntry
108
The filename must be a single component, relative to the
109
parent directory; it cannot be a whole path or relative name.
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
117
Traceback (most recent call last):
118
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
120
if '/' in name or '\\' in name:
121
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
123
self.file_id = file_id
126
self.text_id = text_id
127
self.parent_id = parent_id
128
if kind == 'directory':
133
raise BzrError("unhandled entry kind %r" % kind)
137
def sorted_children(self):
138
l = self.children.items()
144
other = InventoryEntry(self.file_id, self.name, self.kind,
145
self.parent_id, text_id=self.text_id)
146
other.text_sha1 = self.text_sha1
147
other.text_size = self.text_size
152
return ("%s(%r, %r, kind=%r, parent_id=%r)"
153
% (self.__class__.__name__,
160
def to_element(self):
161
"""Convert to XML element"""
164
e.set('name', self.name)
165
e.set('file_id', self.file_id)
166
e.set('kind', self.kind)
168
if self.text_size != None:
169
e.set('text_size', '%d' % self.text_size)
171
for f in ['text_id', 'text_sha1']:
176
# to be conservative, we don't externalize the root pointers
177
# for now, leaving them as null in the xml form. in a future
178
# version it will be implied by nested elements.
179
if self.parent_id != ROOT_ID:
180
assert isinstance(self.parent_id, basestring)
181
e.set('parent_id', self.parent_id)
188
def from_element(cls, elt):
189
assert elt.tag == 'entry'
191
## original format inventories don't have a parent_id for
192
## nodes in the root directory, but it's cleaner to use one
194
parent_id = elt.get('parent_id')
195
if parent_id == None:
198
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
199
self.text_id = elt.get('text_id')
200
self.text_sha1 = elt.get('text_sha1')
202
## mutter("read inventoryentry: %r" % (elt.attrib))
204
v = elt.get('text_size')
205
self.text_size = v and int(v)
210
from_element = classmethod(from_element)
212
def __cmp__(self, other):
215
if not isinstance(other, InventoryEntry):
216
return NotImplemented
218
return cmp(self.file_id, other.file_id) \
219
or cmp(self.name, other.name) \
220
or cmp(self.text_sha1, other.text_sha1) \
221
or cmp(self.text_size, other.text_size) \
222
or cmp(self.text_id, other.text_id) \
223
or cmp(self.parent_id, other.parent_id) \
224
or cmp(self.kind, other.kind)
228
class RootEntry(InventoryEntry):
229
def __init__(self, file_id):
230
self.file_id = file_id
232
self.kind = 'root_directory'
233
self.parent_id = None
236
def __cmp__(self, other):
239
if not isinstance(other, RootEntry):
240
return NotImplemented
241
return cmp(self.file_id, other.file_id) \
242
or cmp(self.children, other.children)
246
class Inventory(XMLMixin):
247
"""Inventory of versioned files in a tree.
249
This describes which file_id is present at each point in the tree,
250
and possibly the SHA-1 or other information about the file.
251
Entries can be looked up either by path or by file_id.
253
The inventory represents a typical unix file tree, with
254
directories containing files and subdirectories. We never store
255
the full path to a file, because renaming a directory implicitly
256
moves all of its contents. This class internally maintains a
257
lookup tree that allows the children under a directory to be
260
InventoryEntry objects must not be modified after they are
261
inserted, other than through the Inventory API.
263
>>> inv = Inventory()
264
>>> inv.write_xml(sys.stdout)
267
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
268
>>> inv['123-123'].name
271
May be treated as an iterator or set to look up file ids:
273
>>> bool(inv.path2id('hello.c'))
278
May also look up by name:
280
>>> [x[0] for x in inv.iter_entries()]
283
>>> inv.write_xml(sys.stdout)
285
<entry file_id="123-123" kind="file" name="hello.c" />
290
"""Create or read an inventory.
292
If a working directory is specified, the inventory is read
293
from there. If the file is specified, read from that. If not,
294
the inventory is created empty.
296
The inventory is created with a default root directory, with
299
self.root = RootEntry(ROOT_ID)
300
self._byid = {self.root.file_id: self.root}
304
return iter(self._byid)
308
"""Returns number of entries."""
309
return len(self._byid)
312
def iter_entries(self, from_dir=None):
313
"""Return (path, entry) pairs, in order by name."""
317
elif isinstance(from_dir, basestring):
318
from_dir = self._byid[from_dir]
320
kids = from_dir.children.items()
322
for name, ie in kids:
324
if ie.kind == 'directory':
325
for cn, cie in self.iter_entries(from_dir=ie.file_id):
326
yield os.path.join(name, cn), cie
330
def directories(self):
331
"""Return (path, entry) pairs for all directories.
333
def descend(parent_ie):
334
parent_name = parent_ie.name
335
yield parent_name, parent_ie
337
# directory children in sorted order
339
for ie in parent_ie.children.itervalues():
340
if ie.kind == 'directory':
341
dn.append((ie.name, ie))
344
for name, child_ie in dn:
345
for sub_name, sub_ie in descend(child_ie):
346
yield appendpath(parent_name, sub_name), sub_ie
348
for name, ie in descend(self.root):
353
def __contains__(self, file_id):
354
"""True if this entry contains a file with given id.
356
>>> inv = Inventory()
357
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
363
return file_id in self._byid
366
def __getitem__(self, file_id):
367
"""Return the entry for given file_id.
369
>>> inv = Inventory()
370
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
371
>>> inv['123123'].name
375
return self._byid[file_id]
378
raise BzrError("can't look up file_id None")
380
raise BzrError("file_id {%s} not in inventory" % file_id)
383
def get_file_kind(self, file_id):
384
return self._byid[file_id].kind
386
def get_child(self, parent_id, filename):
387
return self[parent_id].children.get(filename)
390
def add(self, entry):
391
"""Add entry to inventory.
393
To add a file to a branch ready to be committed, use Branch.add,
395
if entry.file_id in self._byid:
396
bailout("inventory already contains entry with id {%s}" % entry.file_id)
399
parent = self._byid[entry.parent_id]
401
bailout("parent_id {%s} not in inventory" % entry.parent_id)
403
if parent.children.has_key(entry.name):
404
bailout("%s is already versioned" %
405
appendpath(self.id2path(parent.file_id), entry.name))
407
self._byid[entry.file_id] = entry
408
parent.children[entry.name] = entry
411
def add_path(self, relpath, kind, file_id=None):
412
"""Add entry from a path.
414
The immediate parent must already be versioned"""
415
parts = bzrlib.osutils.splitpath(relpath)
417
bailout("cannot re-add root of inventory")
420
file_id = bzrlib.branch.gen_file_id(relpath)
422
parent_id = self.path2id(parts[:-1])
423
assert parent_id != None
424
ie = InventoryEntry(file_id, parts[-1],
425
kind=kind, parent_id=parent_id)
429
def __delitem__(self, file_id):
430
"""Remove entry by id.
432
>>> inv = Inventory()
433
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
442
assert self[ie.parent_id].children[ie.name] == ie
444
# TODO: Test deleting all children; maybe hoist to a separate
446
if ie.kind == 'directory':
447
for cie in ie.children.values():
448
del self[cie.file_id]
451
del self._byid[file_id]
452
del self[ie.parent_id].children[ie.name]
456
return Set(self._byid)
459
def to_element(self):
460
"""Convert to XML Element"""
461
e = Element('inventory')
463
for path, ie in self.iter_entries():
464
e.append(ie.to_element())
468
def from_element(cls, elt):
469
"""Construct from XML Element
471
>>> inv = Inventory()
472
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
473
>>> elt = inv.to_element()
474
>>> inv2 = Inventory.from_element(elt)
478
assert elt.tag == 'inventory'
481
o.add(InventoryEntry.from_element(e))
484
from_element = classmethod(from_element)
487
def __cmp__(self, other):
488
"""Compare two sets by comparing their contents.
494
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
497
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
504
if not isinstance(other, Inventory):
505
return NotImplemented
507
if self.id_set() ^ other.id_set():
510
for file_id in self._byid:
511
c = cmp(self[file_id], other[file_id])
517
def get_idpath(self, file_id):
518
"""Return a list of file_ids for the path to an entry.
520
The list contains one element for each directory followed by
521
the id of the file itself. So the length of the returned list
522
is equal to the depth of the file in the tree, counting the
523
root directory as depth 1.
526
while file_id != None:
528
ie = self._byid[file_id]
530
bailout("file_id {%s} not found in inventory" % file_id)
531
p.insert(0, ie.file_id)
532
file_id = ie.parent_id
536
def id2path(self, file_id):
537
"""Return as a list the path to file_id."""
539
# get all names, skipping root
540
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
541
return os.sep.join(p)
545
def path2id(self, name):
546
"""Walk down through directories to return entry of last component.
548
names may be either a list of path components, or a single
549
string, in which case it is automatically split.
551
This returns the entry of the last component in the path,
552
which may be either a file or a directory.
554
Returns None iff the path is not found.
556
if isinstance(name, types.StringTypes):
557
name = splitpath(name)
559
mutter("lookup path %r" % name)
564
cie = parent.children[f]
566
assert cie.parent_id == parent.file_id
572
return parent.file_id
575
def has_filename(self, names):
576
return bool(self.path2id(names))
579
def has_id(self, file_id):
580
return self._byid.has_key(file_id)
583
def rename(self, file_id, new_parent_id, new_name):
584
"""Move a file within the inventory.
586
This can change either the name, or the parent, or both.
588
This does not move the working file."""
589
if not is_valid_name(new_name):
590
bailout("not an acceptable filename: %r" % new_name)
592
new_parent = self._byid[new_parent_id]
593
if new_name in new_parent.children:
594
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
596
new_parent_idpath = self.get_idpath(new_parent_id)
597
if file_id in new_parent_idpath:
598
bailout("cannot move directory %r into a subdirectory of itself, %r"
599
% (self.id2path(file_id), self.id2path(new_parent_id)))
601
file_ie = self._byid[file_id]
602
old_parent = self._byid[file_ie.parent_id]
604
# TODO: Don't leave things messed up if this fails
606
del old_parent.children[file_ie.name]
607
new_parent.children[new_name] = file_ie
609
file_ie.name = new_name
610
file_ie.parent_id = new_parent_id
615
_NAME_RE = re.compile(r'^[^/\\]+$')
617
def is_valid_name(name):
618
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
