15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
18
# TODO: Maybe also keep the full path of the entry, and the children?
19
# But those depend on its position within a particular inventory, and
20
# it would be nice not to need to hold the backpointer here.
22
# TODO: Perhaps split InventoryEntry into subclasses for files,
23
# directories, etc etc.
26
18
# This should really be an id randomly assigned when the tree is
27
19
# created, but it's not for now.
28
20
ROOT_ID = "TREE_ROOT"
31
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
34
from bzrlib.errors import BzrError, BzrCheckError
36
35
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
37
36
from bzrlib.trace import mutter
38
from bzrlib.errors import NotVersionedError
41
class InventoryEntry(object):
38
class InventoryEntry(XMLMixin):
42
39
"""Description of a versioned file.
44
41
An InventoryEntry has the following fields, which are also
45
42
present in the XML inventory-entry element:
50
(within the parent directory)
56
file_id of the parent directory, or ROOT_ID
59
the revision_id in which the name or parent of this file was
63
sha-1 of the text of the file
66
size in bytes of the text of the file
69
the revision_id in which the text of this file was introduced
71
(reading a version 4 tree created a text_id field.)
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.
73
59
>>> i = Inventory()
76
62
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
77
InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT')
78
63
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
79
InventoryEntry('2323', 'hello.c', kind='file', parent_id='123')
80
64
>>> for j in i.iter_entries():
85
69
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
86
70
Traceback (most recent call last):
88
BzrError: inventory already contains entry with id {2323}
72
BzrError: ('inventory already contains entry with id {2323}', [])
89
73
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
90
InventoryEntry('2324', 'bye.c', kind='file', parent_id='123')
91
74
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
92
InventoryEntry('2325', 'wibble', kind='directory', parent_id='123')
93
75
>>> i.path2id('src/wibble')
97
79
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
98
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
100
81
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
101
82
>>> for j in i.iter_entries():
109
90
src/wibble/wibble.c
110
91
>>> i.id2path('2326')
111
92
'src/wibble/wibble.c'
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.
114
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
115
'text_id', 'parent_id', 'children',
116
'text_version', 'entry_version', ]
119
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
120
106
"""Create an InventoryEntry
130
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
131
117
Traceback (most recent call last):
132
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
118
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
134
assert isinstance(name, basestring), name
135
120
if '/' in name or '\\' in name:
136
121
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
138
self.text_version = None
139
self.entry_version = None
140
self.text_sha1 = None
141
self.text_size = None
142
123
self.file_id = file_id
183
def __eq__(self, other):
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):
184
215
if not isinstance(other, InventoryEntry):
185
216
return NotImplemented
187
return (self.file_id == other.file_id) \
188
and (self.name == other.name) \
189
and (self.text_sha1 == other.text_sha1) \
190
and (self.text_size == other.text_size) \
191
and (self.text_id == other.text_id) \
192
and (self.parent_id == other.parent_id) \
193
and (self.kind == other.kind) \
194
and (self.text_version == other.text_version) \
195
and (self.entry_version == other.entry_version)
198
def __ne__(self, other):
199
return not (self == other)
202
raise ValueError('not hashable')
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)
211
233
self.parent_id = None
214
def __eq__(self, other):
236
def __cmp__(self, other):
215
239
if not isinstance(other, RootEntry):
216
240
return NotImplemented
218
return (self.file_id == other.file_id) \
219
and (self.children == other.children)
223
class Inventory(object):
241
return cmp(self.file_id, other.file_id) \
242
or cmp(self.children, other.children)
246
class Inventory(XMLMixin):
224
247
"""Inventory of versioned files in a tree.
226
249
This describes which file_id is present at each point in the tree,
255
280
>>> [x[0] for x in inv.iter_entries()]
257
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
258
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
259
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
283
>>> inv.write_xml(sys.stdout)
285
<entry file_id="123-123" kind="file" name="hello.c" />
261
def __init__(self, root_id=ROOT_ID):
262
290
"""Create or read an inventory.
264
292
If a working directory is specified, the inventory is read
268
296
The inventory is created with a default root directory, with
271
# We are letting Branch(init=True) create a unique inventory
272
# root id. Rather than generating a random one here.
274
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
275
self.root = RootEntry(root_id)
299
self.root = RootEntry(ROOT_ID)
276
300
self._byid = {self.root.file_id: self.root}
280
other = Inventory(self.root.file_id)
281
# copy recursively so we know directories will be added before
282
# their children. There are more efficient ways than this...
283
for path, entry in self.iter_entries():
284
if entry == self.root:
286
other.add(entry.copy())
290
303
def __iter__(self):
291
304
return iter(self._byid)
311
324
if ie.kind == 'directory':
312
325
for cn, cie in self.iter_entries(from_dir=ie.file_id):
313
326
yield os.path.join(name, cn), cie
317
"""Return list of (path, ie) for all entries except the root.
319
This may be faster than iter_entries.
330
def directories(self):
331
"""Return (path, entry) pairs for all directories.
322
def descend(dir_ie, dir_path):
323
kids = dir_ie.children.items()
325
for name, ie in kids:
326
child_path = os.path.join(dir_path, name)
327
accum.append((child_path, ie))
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():
328
340
if ie.kind == 'directory':
329
descend(ie, child_path)
331
descend(self.root, '')
335
def directories(self):
336
"""Return (path, entry) pairs for all directories, including the root.
339
def descend(parent_ie, parent_path):
340
accum.append((parent_path, parent_ie))
341
dn.append((ie.name, ie))
342
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
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
345
for name, child_ie in kids:
346
child_path = os.path.join(parent_path, name)
347
descend(child_ie, child_path)
348
descend(self.root, '')
348
for name, ie in descend(self.root):
393
388
"""Add entry to inventory.
395
390
To add a file to a branch ready to be committed, use Branch.add,
398
Returns the new entry object.
400
392
if entry.file_id in self._byid:
401
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
403
if entry.parent_id == ROOT_ID or entry.parent_id is None:
404
entry.parent_id = self.root.file_id
393
bailout("inventory already contains entry with id {%s}" % entry.file_id)
407
396
parent = self._byid[entry.parent_id]
409
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
398
bailout("parent_id {%s} not in inventory" % entry.parent_id)
411
400
if parent.children.has_key(entry.name):
412
raise BzrError("%s is already versioned" %
401
bailout("%s is already versioned" %
413
402
appendpath(self.id2path(parent.file_id), entry.name))
415
404
self._byid[entry.file_id] = entry
416
405
parent.children[entry.name] = entry
420
408
def add_path(self, relpath, kind, file_id=None):
421
409
"""Add entry from a path.
423
The immediate parent must already be versioned.
425
Returns the new entry object."""
426
from bzrlib.branch import gen_file_id
411
The immediate parent must already be versioned"""
428
412
parts = bzrlib.osutils.splitpath(relpath)
429
413
if len(parts) == 0:
430
raise BzrError("cannot re-add root of inventory")
414
bailout("cannot re-add root of inventory")
432
416
if file_id == None:
433
file_id = gen_file_id(relpath)
435
parent_path = parts[:-1]
436
parent_id = self.path2id(parent_path)
437
if parent_id == None:
438
raise NotVersionedError(parent_path)
417
file_id = bzrlib.branch.gen_file_id(relpath)
419
parent_id = self.path2id(parts[:-1])
420
assert parent_id != None
440
421
ie = InventoryEntry(file_id, parts[-1],
441
422
kind=kind, parent_id=parent_id)
442
423
return self.add(ie)
469
449
del self[ie.parent_id].children[ie.name]
472
def __eq__(self, other):
453
return Set(self._byid)
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):
473
485
"""Compare two sets by comparing their contents.
475
487
>>> i1 = Inventory()
479
491
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
480
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
483
494
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
484
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
488
501
if not isinstance(other, Inventory):
489
502
return NotImplemented
491
if len(self._byid) != len(other._byid):
492
# shortcut: obviously not the same
495
return self._byid == other._byid
498
def __ne__(self, other):
499
return not (self == other)
503
raise ValueError('not hashable')
504
if self.id_set() ^ other.id_set():
507
for file_id in self._byid:
508
c = cmp(self[file_id], other[file_id])
506
514
def get_idpath(self, file_id):
577
585
This does not move the working file."""
578
586
if not is_valid_name(new_name):
579
raise BzrError("not an acceptable filename: %r" % new_name)
587
bailout("not an acceptable filename: %r" % new_name)
581
589
new_parent = self._byid[new_parent_id]
582
590
if new_name in new_parent.children:
583
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
591
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
585
593
new_parent_idpath = self.get_idpath(new_parent_id)
586
594
if file_id in new_parent_idpath:
587
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
595
bailout("cannot move directory %r into a subdirectory of itself, %r"
588
596
% (self.id2path(file_id), self.id2path(new_parent_id)))
590
598
file_ie = self._byid[file_id]
added
removed
bzrlib/inventory.py
