14
14
# along with this program; if not, write to the Free Software
15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
"""Inventories map files to their name in a revision."""
19
# TODO: Maybe store inventory_id in the file? Not really needed.
21
__author__ = "Martin Pool <mbp@canonical.com>"
18
24
# This should really be an id randomly assigned when the tree is
19
25
# created, but it's not for now.
23
29
import sys, os.path, types, re
33
from cElementTree import Element, ElementTree, SubElement
35
from elementtree.ElementTree import Element, ElementTree, SubElement
37
from xml import XMLMixin
38
from errors import bailout, BzrError
26
from bzrlib.errors import BzrError, BzrCheckError
28
41
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
29
42
from bzrlib.trace import mutter
31
class InventoryEntry(object):
44
class InventoryEntry(XMLMixin):
32
45
"""Description of a versioned file.
34
47
An InventoryEntry has the following fields, which are also
62
75
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
63
76
Traceback (most recent call last):
65
BzrError: inventory already contains entry with id {2323}
78
BzrError: ('inventory already contains entry with id {2323}', [])
66
79
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
67
80
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
68
81
>>> i.path2id('src/wibble')
84
97
>>> i.id2path('2326')
85
98
'src/wibble/wibble.c'
87
TODO: Maybe also keep the full path of the entry, and the children?
100
:todo: Maybe also keep the full path of the entry, and the children?
88
101
But those depend on its position within a particular inventory, and
89
102
it would be nice not to need to hold the backpointer here.
92
105
# TODO: split InventoryEntry into subclasses for files,
93
106
# directories, etc etc.
95
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
96
'text_id', 'parent_id', 'children', ]
98
108
def __init__(self, file_id, name, kind, parent_id, text_id=None):
99
109
"""Create an InventoryEntry
109
119
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
110
120
Traceback (most recent call last):
111
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
121
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
113
if '/' in name or '\\' in name:
114
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
116
self.text_sha1 = None
117
self.text_size = None
124
if len(splitpath(name)) != 1:
125
bailout('InventoryEntry name is not a simple filename: %r'
119
128
self.file_id = file_id
130
assert kind in ['file', 'directory']
122
132
self.text_id = text_id
123
133
self.parent_id = parent_id
134
self.text_sha1 = None
135
self.text_size = None
124
136
if kind == 'directory':
125
137
self.children = {}
129
raise BzrError("unhandled entry kind %r" % kind)
139
assert kind == 'file'
133
142
def sorted_children(self):
141
150
self.parent_id, text_id=self.text_id)
142
151
other.text_sha1 = self.text_sha1
143
152
other.text_size = self.text_size
144
# note that children are *not* copied; they're pulled across when
210
215
from_element = classmethod(from_element)
212
def __eq__(self, other):
217
def __cmp__(self, other):
213
220
if not isinstance(other, InventoryEntry):
214
221
return NotImplemented
216
return (self.file_id == other.file_id) \
217
and (self.name == other.name) \
218
and (self.text_sha1 == other.text_sha1) \
219
and (self.text_size == other.text_size) \
220
and (self.text_id == other.text_id) \
221
and (self.parent_id == other.parent_id) \
222
and (self.kind == other.kind)
225
def __ne__(self, other):
226
return not (self == other)
229
raise ValueError('not hashable')
223
return cmp(self.file_id, other.file_id) \
224
or cmp(self.name, other.name) \
225
or cmp(self.text_sha1, other.text_sha1) \
226
or cmp(self.text_size, other.text_size) \
227
or cmp(self.text_id, other.text_id) \
228
or cmp(self.parent_id, other.parent_id) \
229
or cmp(self.kind, other.kind)
238
238
self.parent_id = None
241
def __eq__(self, other):
241
def __cmp__(self, other):
242
244
if not isinstance(other, RootEntry):
243
245
return NotImplemented
245
return (self.file_id == other.file_id) \
246
and (self.children == other.children)
250
class Inventory(object):
246
return cmp(self.file_id, other.file_id) \
247
or cmp(self.children, other.children)
251
class Inventory(XMLMixin):
251
252
"""Inventory of versioned files in a tree.
253
This describes which file_id is present at each point in the tree,
254
and possibly the SHA-1 or other information about the file.
255
Entries can be looked up either by path or by file_id.
254
An Inventory acts like a set of InventoryEntry items. You can
255
also look files up by their file_id or name.
257
May be read from and written to a metadata file in a tree. To
258
manipulate the inventory (for example to add a file), it is read
259
in, modified, and then written back out.
257
261
The inventory represents a typical unix file tree, with
258
262
directories containing files and subdirectories. We never store
265
269
inserted, other than through the Inventory API.
267
271
>>> inv = Inventory()
272
>>> inv.write_xml(sys.stdout)
268
275
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
269
276
>>> inv['123-123'].name
281
288
>>> [x[0] for x in inv.iter_entries()]
283
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
284
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
291
>>> inv.write_xml(sys.stdout)
293
<entry file_id="123-123" kind="file" name="hello.c" />
286
def __init__(self, root_id=ROOT_ID):
298
## TODO: Make sure only canonical filenames are stored.
300
## TODO: Do something sensible about the possible collisions on
301
## case-losing filesystems. Perhaps we should just always forbid
304
## TODO: No special cases for root, rather just give it a file id
305
## like everything else.
307
## TODO: Probably change XML serialization to use nesting
287
310
"""Create or read an inventory.
289
312
If a working directory is specified, the inventory is read
293
316
The inventory is created with a default root directory, with
296
# We are letting Branch(init=True) create a unique inventory
297
# root id. Rather than generating a random one here.
299
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
300
self.root = RootEntry(root_id)
319
self.root = RootEntry(ROOT_ID)
301
320
self._byid = {self.root.file_id: self.root}
325
344
if ie.kind == 'directory':
326
345
for cn, cie in self.iter_entries(from_dir=ie.file_id):
327
yield os.path.join(name, cn), cie
331
"""Return list of (path, ie) for all entries except the root.
333
This may be faster than iter_entries.
346
yield '/'.join((name, cn)), cie
350
def directories(self):
351
"""Return (path, entry) pairs for all directories.
336
def descend(dir_ie, dir_path):
337
kids = dir_ie.children.items()
339
for name, ie in kids:
340
child_path = os.path.join(dir_path, name)
341
accum.append((child_path, ie))
353
def descend(parent_ie):
354
parent_name = parent_ie.name
355
yield parent_name, parent_ie
357
# directory children in sorted order
359
for ie in parent_ie.children.itervalues():
342
360
if ie.kind == 'directory':
343
descend(ie, child_path)
345
descend(self.root, '')
349
def directories(self):
350
"""Return (path, entry) pairs for all directories, including the root.
353
def descend(parent_ie, parent_path):
354
accum.append((parent_path, parent_ie))
361
dn.append((ie.name, ie))
356
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
364
for name, child_ie in dn:
365
for sub_name, sub_ie in descend(child_ie):
366
yield appendpath(parent_name, sub_name), sub_ie
359
for name, child_ie in kids:
360
child_path = os.path.join(parent_path, name)
361
descend(child_ie, child_path)
362
descend(self.root, '')
368
for name, ie in descend(self.root):
385
391
>>> inv['123123'].name
395
raise BzrError("can't look up file_id None")
389
398
return self._byid[file_id]
392
raise BzrError("can't look up file_id None")
394
raise BzrError("file_id {%s} not in inventory" % file_id)
397
def get_file_kind(self, file_id):
398
return self._byid[file_id].kind
400
raise BzrError("file_id {%s} not in inventory" % file_id)
400
403
def get_child(self, parent_id, filename):
401
404
return self[parent_id].children.get(filename)
407
410
To add a file to a branch ready to be committed, use Branch.add,
408
411
which calls this."""
409
412
if entry.file_id in self._byid:
410
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
412
if entry.parent_id == ROOT_ID or entry.parent_id is None:
413
entry.parent_id = self.root.file_id
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
416
416
parent = self._byid[entry.parent_id]
418
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
420
420
if parent.children.has_key(entry.name):
421
raise BzrError("%s is already versioned" %
421
bailout("%s is already versioned" %
422
422
appendpath(self.id2path(parent.file_id), entry.name))
424
424
self._byid[entry.file_id] = entry
429
429
"""Add entry from a path.
431
431
The immediate parent must already be versioned"""
432
from bzrlib.errors import NotVersionedError
434
432
parts = bzrlib.osutils.splitpath(relpath)
435
433
if len(parts) == 0:
436
raise BzrError("cannot re-add root of inventory")
434
bailout("cannot re-add root of inventory")
438
436
if file_id == None:
439
from bzrlib.branch import gen_file_id
440
file_id = gen_file_id(relpath)
442
parent_path = parts[:-1]
443
parent_id = self.path2id(parent_path)
444
if parent_id == None:
445
raise NotVersionedError(parent_path)
437
file_id = bzrlib.branch.gen_file_id(relpath)
439
parent_id = self.path2id(parts[:-1])
440
assert parent_id != None
447
441
ie = InventoryEntry(file_id, parts[-1],
448
442
kind=kind, parent_id=parent_id)
449
443
return self.add(ie)
475
469
del self[ie.parent_id].children[ie.name]
473
return Set(self._byid)
478
476
def to_element(self):
479
477
"""Convert to XML Element"""
480
from bzrlib.xml import Element
482
478
e = Element('inventory')
484
if self.root.file_id not in (None, ROOT_ID):
485
e.set('file_id', self.root.file_id)
486
480
for path, ie in self.iter_entries():
487
481
e.append(ie.to_element())
491
485
def from_element(cls, elt):
492
486
"""Construct from XML Element
494
488
>>> inv = Inventory()
495
489
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
496
490
>>> elt = inv.to_element()
501
# XXXX: doctest doesn't run this properly under python2.3
502
495
assert elt.tag == 'inventory'
503
root_id = elt.get('file_id') or ROOT_ID
506
ie = InventoryEntry.from_element(e)
507
if ie.parent_id == ROOT_ID:
508
ie.parent_id = root_id
498
o.add(InventoryEntry.from_element(e))
512
501
from_element = classmethod(from_element)
515
def __eq__(self, other):
504
def __cmp__(self, other):
516
505
"""Compare two sets by comparing their contents.
518
507
>>> i1 = Inventory()
529
521
if not isinstance(other, Inventory):
530
522
return NotImplemented
532
if len(self._byid) != len(other._byid):
533
# shortcut: obviously not the same
536
return self._byid == other._byid
539
def __ne__(self, other):
540
return not (self == other)
544
raise ValueError('not hashable')
524
if self.id_set() ^ other.id_set():
527
for file_id in self._byid:
528
c = cmp(self[file_id], other[file_id])
548
534
def get_idpath(self, file_id):
568
554
"""Return as a list the path to file_id."""
570
556
# get all names, skipping root
571
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
572
return os.sep.join(p)
557
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
619
605
This does not move the working file."""
620
606
if not is_valid_name(new_name):
621
raise BzrError("not an acceptable filename: %r" % new_name)
607
bailout("not an acceptable filename: %r" % new_name)
623
609
new_parent = self._byid[new_parent_id]
624
610
if new_name in new_parent.children:
625
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
611
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
627
613
new_parent_idpath = self.get_idpath(new_parent_id)
628
614
if file_id in new_parent_idpath:
629
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
615
bailout("cannot move directory %r into a subdirectory of itself, %r"
630
616
% (self.id2path(file_id), self.id2path(new_parent_id)))
632
618
file_ie = self._byid[file_id]
added
removed
bzrlib/inventory.py
