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.
18
26
# This should really be an id randomly assigned when the tree is
19
27
# created, but it's not for now.
20
28
ROOT_ID = "TREE_ROOT"
23
31
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
34
from bzrlib.errors import BzrError, BzrCheckError
35
36
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
36
37
from bzrlib.trace import mutter
38
from bzrlib.errors import NotVersionedError
38
class InventoryEntry(XMLMixin):
41
class InventoryEntry(object):
39
42
"""Description of a versioned file.
41
44
An InventoryEntry has the following fields, which are also
42
45
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.
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.)
59
73
>>> i = Inventory()
62
76
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
77
InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT')
63
78
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
79
InventoryEntry('2323', 'hello.c', kind='file', parent_id='123')
64
80
>>> for j in i.iter_entries():
69
85
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
70
86
Traceback (most recent call last):
72
BzrError: ('inventory already contains entry with id {2323}', [])
88
BzrError: inventory already contains entry with id {2323}
73
89
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
90
InventoryEntry('2324', 'bye.c', kind='file', parent_id='123')
74
91
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
92
InventoryEntry('2325', 'wibble', kind='directory', parent_id='123')
75
93
>>> i.path2id('src/wibble')
79
97
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
98
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
81
100
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
82
101
>>> for j in i.iter_entries():
90
109
src/wibble/wibble.c
91
110
>>> i.id2path('2326')
92
111
'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', ]
105
119
def __init__(self, file_id, name, kind, parent_id, text_id=None):
106
120
"""Create an InventoryEntry
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):
182
def __eq__(self, other):
215
183
if not isinstance(other, InventoryEntry):
216
184
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)
186
return (self.file_id == other.file_id) \
187
and (self.name == other.name) \
188
and (self.text_sha1 == other.text_sha1) \
189
and (self.text_size == other.text_size) \
190
and (self.text_id == other.text_id) \
191
and (self.parent_id == other.parent_id) \
192
and (self.kind == other.kind) \
193
and (self.text_version == other.text_version) \
194
and (self.entry_version == other.entry_version)
197
def __ne__(self, other):
198
return not (self == other)
201
raise ValueError('not hashable')
233
210
self.parent_id = None
236
def __cmp__(self, other):
213
def __eq__(self, other):
239
214
if not isinstance(other, RootEntry):
240
215
return NotImplemented
241
return cmp(self.file_id, other.file_id) \
242
or cmp(self.children, other.children)
246
class Inventory(XMLMixin):
217
return (self.file_id == other.file_id) \
218
and (self.children == other.children)
222
class Inventory(object):
247
223
"""Inventory of versioned files in a tree.
249
225
This describes which file_id is present at each point in the tree,
280
254
>>> [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" />
256
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
257
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
258
InventoryEntry('123-123', 'hello.c', kind='file', parent_id='TREE_ROOT-12345678-12345678')
260
def __init__(self, root_id=ROOT_ID):
290
261
"""Create or read an inventory.
292
263
If a working directory is specified, the inventory is read
296
267
The inventory is created with a default root directory, with
299
self.root = RootEntry(ROOT_ID)
270
# We are letting Branch(init=True) create a unique inventory
271
# root id. Rather than generating a random one here.
273
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
274
self.root = RootEntry(root_id)
300
275
self._byid = {self.root.file_id: self.root}
279
other = Inventory(self.root.file_id)
280
# copy recursively so we know directories will be added before
281
# their children. There are more efficient ways than this...
282
for path, entry in self.iter_entries():
283
if entry == self.root:
285
other.add(entry.copy())
303
289
def __iter__(self):
304
290
return iter(self._byid)
324
310
if ie.kind == 'directory':
325
311
for cn, cie in self.iter_entries(from_dir=ie.file_id):
326
312
yield os.path.join(name, cn), cie
316
"""Return list of (path, ie) for all entries except the root.
318
This may be faster than iter_entries.
321
def descend(dir_ie, dir_path):
322
kids = dir_ie.children.items()
324
for name, ie in kids:
325
child_path = os.path.join(dir_path, name)
326
accum.append((child_path, ie))
327
if ie.kind == 'directory':
328
descend(ie, child_path)
330
descend(self.root, '')
330
334
def directories(self):
331
"""Return (path, entry) pairs for all directories.
335
"""Return (path, entry) pairs for all directories, including the root.
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))
338
def descend(parent_ie, parent_path):
339
accum.append((parent_path, parent_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
341
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
348
for name, ie in descend(self.root):
344
for name, child_ie in kids:
345
child_path = os.path.join(parent_path, name)
346
descend(child_ie, child_path)
347
descend(self.root, '')
391
392
"""Add entry to inventory.
393
394
To add a file to a branch ready to be committed, use Branch.add,
397
Returns the new entry object.
395
399
if entry.file_id in self._byid:
396
bailout("inventory already contains entry with id {%s}" % entry.file_id)
400
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
402
if entry.parent_id == ROOT_ID or entry.parent_id is None:
403
entry.parent_id = self.root.file_id
399
406
parent = self._byid[entry.parent_id]
401
bailout("parent_id {%s} not in inventory" % entry.parent_id)
408
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
403
410
if parent.children.has_key(entry.name):
404
bailout("%s is already versioned" %
411
raise BzrError("%s is already versioned" %
405
412
appendpath(self.id2path(parent.file_id), entry.name))
407
414
self._byid[entry.file_id] = entry
408
415
parent.children[entry.name] = entry
411
419
def add_path(self, relpath, kind, file_id=None):
412
420
"""Add entry from a path.
414
The immediate parent must already be versioned"""
422
The immediate parent must already be versioned.
424
Returns the new entry object."""
425
from bzrlib.branch import gen_file_id
415
427
parts = bzrlib.osutils.splitpath(relpath)
416
428
if len(parts) == 0:
417
bailout("cannot re-add root of inventory")
429
raise BzrError("cannot re-add root of inventory")
419
431
if file_id == None:
420
file_id = bzrlib.branch.gen_file_id(relpath)
422
parent_id = self.path2id(parts[:-1])
423
assert parent_id != None
432
file_id = gen_file_id(relpath)
434
parent_path = parts[:-1]
435
parent_id = self.path2id(parent_path)
436
if parent_id == None:
437
raise NotVersionedError(parent_path)
424
439
ie = InventoryEntry(file_id, parts[-1],
425
440
kind=kind, parent_id=parent_id)
426
441
return self.add(ie)
452
468
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):
471
def __eq__(self, other):
488
472
"""Compare two sets by comparing their contents.
490
474
>>> i1 = Inventory()
494
478
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
479
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
497
482
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
483
InventoryEntry('123', 'foo', kind='file', parent_id='TREE_ROOT')
504
487
if not isinstance(other, Inventory):
505
488
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])
490
if len(self._byid) != len(other._byid):
491
# shortcut: obviously not the same
494
return self._byid == other._byid
497
def __ne__(self, other):
498
return not (self == other)
502
raise ValueError('not hashable')
517
505
def get_idpath(self, file_id):
588
576
This does not move the working file."""
589
577
if not is_valid_name(new_name):
590
bailout("not an acceptable filename: %r" % new_name)
578
raise BzrError("not an acceptable filename: %r" % new_name)
592
580
new_parent = self._byid[new_parent_id]
593
581
if new_name in new_parent.children:
594
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
582
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
596
584
new_parent_idpath = self.get_idpath(new_parent_id)
597
585
if file_id in new_parent_idpath:
598
bailout("cannot move directory %r into a subdirectory of itself, %r"
586
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
599
587
% (self.id2path(file_id), self.id2path(new_parent_id)))
601
589
file_ie = self._byid[file_id]
added
removed
bzrlib/inventory.py
