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 __cmp__(self, other):
216
if not isinstance(other, InventoryEntry):
217
return NotImplemented
219
return cmp(self.file_id, other.file_id) \
220
or cmp(self.name, other.name) \
221
or cmp(self.text_sha1, other.text_sha1) \
222
or cmp(self.text_size, other.text_size) \
223
or cmp(self.text_id, other.text_id) \
224
or cmp(self.parent_id, other.parent_id) \
225
or cmp(self.kind, other.kind)
229
class RootEntry(InventoryEntry):
230
def __init__(self, file_id):
231
self.file_id = file_id
233
self.kind = 'root_directory'
234
self.parent_id = None
237
def __cmp__(self, other):
240
if not isinstance(other, RootEntry):
241
return NotImplemented
242
return cmp(self.file_id, other.file_id) \
243
or cmp(self.children, other.children)
247
class Inventory(XMLMixin):
248
"""Inventory of versioned files in a tree.
250
This describes which file_id is present at each point in the tree,
251
and possibly the SHA-1 or other information about the file.
252
Entries can be looked up either by path or by file_id.
254
The inventory represents a typical unix file tree, with
255
directories containing files and subdirectories. We never store
256
the full path to a file, because renaming a directory implicitly
257
moves all of its contents. This class internally maintains a
258
lookup tree that allows the children under a directory to be
261
InventoryEntry objects must not be modified after they are
262
inserted, other than through the Inventory API.
264
>>> inv = Inventory()
265
>>> inv.write_xml(sys.stdout)
268
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
269
>>> inv['123-123'].name
272
May be treated as an iterator or set to look up file ids:
274
>>> bool(inv.path2id('hello.c'))
279
May also look up by name:
281
>>> [x[0] for x in inv.iter_entries()]
284
>>> inv.write_xml(sys.stdout)
286
<entry file_id="123-123" kind="file" name="hello.c" />
291
"""Create or read an inventory.
293
If a working directory is specified, the inventory is read
294
from there. If the file is specified, read from that. If not,
295
the inventory is created empty.
297
The inventory is created with a default root directory, with
300
self.root = RootEntry(ROOT_ID)
301
self._byid = {self.root.file_id: self.root}
305
return iter(self._byid)
309
"""Returns number of entries."""
310
return len(self._byid)
313
def iter_entries(self, from_dir=None):
314
"""Return (path, entry) pairs, in order by name."""
318
elif isinstance(from_dir, basestring):
319
from_dir = self._byid[from_dir]
321
kids = from_dir.children.items()
323
for name, ie in kids:
325
if ie.kind == 'directory':
326
for cn, cie in self.iter_entries(from_dir=ie.file_id):
327
yield os.path.join(name, cn), cie
331
def directories(self):
332
"""Return (path, entry) pairs for all directories.
334
def descend(parent_ie):
335
parent_name = parent_ie.name
336
yield parent_name, parent_ie
338
# directory children in sorted order
340
for ie in parent_ie.children.itervalues():
341
if ie.kind == 'directory':
342
dn.append((ie.name, ie))
345
for name, child_ie in dn:
346
for sub_name, sub_ie in descend(child_ie):
347
yield appendpath(parent_name, sub_name), sub_ie
349
for name, ie in descend(self.root):
354
def __contains__(self, file_id):
355
"""True if this entry contains a file with given id.
357
>>> inv = Inventory()
358
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
364
return file_id in self._byid
367
def __getitem__(self, file_id):
368
"""Return the entry for given file_id.
370
>>> inv = Inventory()
371
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
372
>>> inv['123123'].name
376
return self._byid[file_id]
379
raise BzrError("can't look up file_id None")
381
raise BzrError("file_id {%s} not in inventory" % file_id)
384
def get_file_kind(self, file_id):
385
return self._byid[file_id].kind
387
def get_child(self, parent_id, filename):
388
return self[parent_id].children.get(filename)
391
def add(self, entry):
392
"""Add entry to inventory.
394
To add a file to a branch ready to be committed, use Branch.add,
396
if entry.file_id in self._byid:
397
bailout("inventory already contains entry with id {%s}" % entry.file_id)
400
parent = self._byid[entry.parent_id]
402
bailout("parent_id {%s} not in inventory" % entry.parent_id)
404
if parent.children.has_key(entry.name):
405
bailout("%s is already versioned" %
406
appendpath(self.id2path(parent.file_id), entry.name))
408
self._byid[entry.file_id] = entry
409
parent.children[entry.name] = entry
412
def add_path(self, relpath, kind, file_id=None):
413
"""Add entry from a path.
415
The immediate parent must already be versioned"""
416
parts = bzrlib.osutils.splitpath(relpath)
418
bailout("cannot re-add root of inventory")
421
file_id = bzrlib.branch.gen_file_id(relpath)
423
parent_id = self.path2id(parts[:-1])
424
assert parent_id != None
425
ie = InventoryEntry(file_id, parts[-1],
426
kind=kind, parent_id=parent_id)
430
def __delitem__(self, file_id):
431
"""Remove entry by id.
433
>>> inv = Inventory()
434
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
443
assert self[ie.parent_id].children[ie.name] == ie
445
# TODO: Test deleting all children; maybe hoist to a separate
447
if ie.kind == 'directory':
448
for cie in ie.children.values():
449
del self[cie.file_id]
452
del self._byid[file_id]
453
del self[ie.parent_id].children[ie.name]
456
def to_element(self):
457
"""Convert to XML Element"""
458
e = Element('inventory')
460
for path, ie in self.iter_entries():
461
e.append(ie.to_element())
465
def from_element(cls, elt):
466
"""Construct from XML Element
468
>>> inv = Inventory()
469
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
470
>>> elt = inv.to_element()
471
>>> inv2 = Inventory.from_element(elt)
475
assert elt.tag == 'inventory'
478
o.add(InventoryEntry.from_element(e))
481
from_element = classmethod(from_element)
484
def __cmp__(self, other):
485
"""Compare two sets by comparing their contents.
491
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
494
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
501
if not isinstance(other, Inventory):
502
return NotImplemented
505
otherids = other._byid
507
if len(byid) != len(otherids):
508
# shortcut: obviously not the same
512
if file_id not in otherids:
515
c = cmp(byid[file_id], otherids[file_id])
518
for file_id in otherids:
519
if file_id not in byid:
525
def get_idpath(self, file_id):
526
"""Return a list of file_ids for the path to an entry.
528
The list contains one element for each directory followed by
529
the id of the file itself. So the length of the returned list
530
is equal to the depth of the file in the tree, counting the
531
root directory as depth 1.
534
while file_id != None:
536
ie = self._byid[file_id]
538
bailout("file_id {%s} not found in inventory" % file_id)
539
p.insert(0, ie.file_id)
540
file_id = ie.parent_id
544
def id2path(self, file_id):
545
"""Return as a list the path to file_id."""
547
# get all names, skipping root
548
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
549
return os.sep.join(p)
553
def path2id(self, name):
554
"""Walk down through directories to return entry of last component.
556
names may be either a list of path components, or a single
557
string, in which case it is automatically split.
559
This returns the entry of the last component in the path,
560
which may be either a file or a directory.
562
Returns None iff the path is not found.
564
if isinstance(name, types.StringTypes):
565
name = splitpath(name)
567
mutter("lookup path %r" % name)
572
cie = parent.children[f]
574
assert cie.parent_id == parent.file_id
580
return parent.file_id
583
def has_filename(self, names):
584
return bool(self.path2id(names))
587
def has_id(self, file_id):
588
return self._byid.has_key(file_id)
591
def rename(self, file_id, new_parent_id, new_name):
592
"""Move a file within the inventory.
594
This can change either the name, or the parent, or both.
596
This does not move the working file."""
597
if not is_valid_name(new_name):
598
bailout("not an acceptable filename: %r" % new_name)
600
new_parent = self._byid[new_parent_id]
601
if new_name in new_parent.children:
602
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
604
new_parent_idpath = self.get_idpath(new_parent_id)
605
if file_id in new_parent_idpath:
606
bailout("cannot move directory %r into a subdirectory of itself, %r"
607
% (self.id2path(file_id), self.id2path(new_parent_id)))
609
file_ie = self._byid[file_id]
610
old_parent = self._byid[file_ie.parent_id]
612
# TODO: Don't leave things messed up if this fails
614
del old_parent.children[file_ie.name]
615
new_parent.children[new_name] = file_ie
617
file_ie.name = new_name
618
file_ie.parent_id = new_parent_id
623
_NAME_RE = re.compile(r'^[^/\\]+$')
625
def is_valid_name(name):
626
return bool(_NAME_RE.match(name))
added
removed
bzrlib/inventory.py
