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
import sys, os.path, types, re
26
37
from bzrlib.errors import BzrError, BzrCheckError
28
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
39
from bzrlib.osutils import quotefn, splitpath, joinpath, appendpath, sha_strings
29
40
from bzrlib.trace import mutter
30
41
from bzrlib.errors import NotVersionedError
33
44
class InventoryEntry(object):
34
45
"""Description of a versioned file.
36
47
An InventoryEntry has the following fields, which are also
37
48
present in the XML inventory-entry element:
40
* *name*: (only the basename within the directory, must not
42
* *kind*: "directory" or "file"
43
* *directory_id*: (if absent/null means the branch root directory)
44
* *text_sha1*: only for files
45
* *text_size*: in bytes, only for files
46
* *text_id*: identifier for the text version, only for files
48
InventoryEntries can also exist inside a WorkingTree
49
inventory, in which case they are not yet bound to a
50
particular revision of the file. In that case the text_sha1,
51
text_size and text_id are absent.
53
(within the parent directory)
59
file_id of the parent directory, or ROOT_ID
62
the revision_id in which this variationo f this file was
66
sha-1 of the text of the file
69
size in bytes of the text of the file
71
(reading a version 4 tree created a text_id field.)
54
73
>>> i = Inventory()
57
76
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
77
InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT')
58
78
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
79
InventoryEntry('2323', 'hello.c', kind='file', parent_id='123')
59
80
>>> for j in i.iter_entries():
67
88
BzrError: inventory already contains entry with id {2323}
68
89
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
90
InventoryEntry('2324', 'bye.c', kind='file', parent_id='123')
69
91
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
92
InventoryEntry('2325', 'wibble', kind='directory', parent_id='123')
70
93
>>> i.path2id('src/wibble')
74
97
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
98
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
76
100
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
77
>>> for j in i.iter_entries():
79
... assert i.path2id(j[0])
101
>>> for path, entry in i.iter_entries():
102
... print path.replace('\\\\', '/') # for win32 os.sep
103
... assert i.path2id(path)
85
109
src/wibble/wibble.c
110
>>> i.id2path('2326').replace('\\\\', '/')
87
111
'src/wibble/wibble.c'
89
TODO: Maybe also keep the full path of the entry, and the children?
90
But those depend on its position within a particular inventory, and
91
it would be nice not to need to hold the backpointer here.
94
# TODO: split InventoryEntry into subclasses for files,
95
# directories, etc etc.
97
114
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
98
'text_id', 'parent_id', 'children', ]
115
'text_id', 'parent_id', 'children',
116
'revision', 'symlink_target']
118
def _add_text_to_weave(self, new_lines, parents, weave_store):
119
weave_store.add_text(self.file_id, self.revision, new_lines, parents)
100
121
def __init__(self, file_id, name, kind, parent_id, text_id=None):
101
122
"""Create an InventoryEntry
112
133
Traceback (most recent call last):
113
134
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
136
assert isinstance(name, basestring), name
115
137
if '/' in name or '\\' in name:
116
138
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
118
141
self.text_sha1 = None
119
142
self.text_size = None
121
143
self.file_id = file_id
124
146
self.text_id = text_id
125
147
self.parent_id = parent_id
148
self.symlink_target = None
126
149
if kind == 'directory':
127
150
self.children = {}
128
151
elif kind == 'file':
153
elif kind == 'symlink':
131
156
raise BzrError("unhandled entry kind %r" % kind)
158
def read_symlink_target(self, path):
159
if self.kind == 'symlink':
161
self.symlink_target = os.readlink(path)
163
raise BzrError("os.readlink error, %s" % e)
135
165
def sorted_children(self):
136
166
l = self.children.items()
170
def check(self, checker, rev_id, inv, tree):
171
if self.parent_id != None:
172
if not inv.has_id(self.parent_id):
173
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
174
% (self.parent_id, rev_id))
175
if self.kind == 'file':
176
revision = self.revision
177
t = (self.file_id, revision)
178
if t in checker.checked_texts:
179
prev_sha = checker.checked_texts[t]
180
if prev_sha != self.text_sha1:
181
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
182
(self.file_id, rev_id))
184
checker.repeated_text_cnt += 1
186
mutter('check version {%s} of {%s}', rev_id, self.file_id)
187
file_lines = tree.get_file_lines(self.file_id)
188
checker.checked_text_cnt += 1
189
if self.text_size != sum(map(len, file_lines)):
190
raise BzrCheckError('text {%s} wrong size' % self.text_id)
191
if self.text_sha1 != sha_strings(file_lines):
192
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
193
checker.checked_texts[t] = self.text_sha1
194
elif self.kind == 'directory':
195
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
196
raise BzrCheckError('directory {%s} has text in revision {%s}'
197
% (self.file_id, rev_id))
198
elif self.kind == 'root_directory':
200
elif self.kind == 'symlink':
201
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
202
raise BzrCheckError('symlink {%s} has text in revision {%s}'
203
% (self.file_id, rev_id))
204
if self.symlink_target == None:
205
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
206
% (self.file_id, rev_id))
208
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
142
213
other = InventoryEntry(self.file_id, self.name, self.kind,
143
self.parent_id, text_id=self.text_id)
215
other.text_id = self.text_id
144
216
other.text_sha1 = self.text_sha1
145
217
other.text_size = self.text_size
218
other.symlink_target = self.symlink_target
219
other.revision = self.revision
146
220
# note that children are *not* copied; they're pulled across when
147
221
# others are added
224
def _get_snapshot_change(self, previous_entries):
225
if len(previous_entries) > 1:
227
elif len(previous_entries) == 0:
230
return 'modified/renamed/reparented'
151
232
def __repr__(self):
152
233
return ("%s(%r, %r, kind=%r, parent_id=%r)"
160
def to_element(self):
161
"""Convert to XML element"""
162
from bzrlib.xml import Element
166
e.set('name', self.name)
167
e.set('file_id', self.file_id)
168
e.set('kind', self.kind)
170
if self.text_size != None:
171
e.set('text_size', '%d' % self.text_size)
173
for f in ['text_id', 'text_sha1']:
178
# to be conservative, we don't externalize the root pointers
179
# for now, leaving them as null in the xml form. in a future
180
# version it will be implied by nested elements.
181
if self.parent_id != ROOT_ID:
182
assert isinstance(self.parent_id, basestring)
183
e.set('parent_id', self.parent_id)
190
def from_element(cls, elt):
191
assert elt.tag == 'entry'
193
## original format inventories don't have a parent_id for
194
## nodes in the root directory, but it's cleaner to use one
196
parent_id = elt.get('parent_id')
197
if parent_id == None:
200
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
201
self.text_id = elt.get('text_id')
202
self.text_sha1 = elt.get('text_sha1')
204
## mutter("read inventoryentry: %r" % (elt.attrib))
206
v = elt.get('text_size')
207
self.text_size = v and int(v)
212
from_element = classmethod(from_element)
240
def snapshot(self, revision, path, previous_entries, work_tree,
242
"""Make a snapshot of this entry.
244
This means that all its fields are populated, that it has its
245
text stored in the text store or weave.
247
mutter('new parents of %s are %r', path, previous_entries)
248
self._read_tree_state(path, work_tree)
249
if len(previous_entries) == 1:
250
# cannot be unchanged unless there is only one parent file rev.
251
parent_ie = previous_entries.values()[0]
252
if self._unchanged(path, parent_ie, work_tree):
253
mutter("found unchanged entry")
254
self.revision = parent_ie.revision
256
mutter('new revision for {%s}', self.file_id)
257
self.revision = revision
258
change = self._get_snapshot_change(previous_entries)
259
if self.kind != 'file':
261
self._snapshot_text(previous_entries, work_tree, weave_store)
264
def _snapshot_text(self, file_parents, work_tree, weave_store):
265
mutter('storing file {%s} in revision {%s}',
266
self.file_id, self.revision)
267
# special case to avoid diffing on renames or
269
if (len(file_parents) == 1
270
and self.text_sha1 == file_parents.values()[0].text_sha1
271
and self.text_size == file_parents.values()[0].text_size):
272
previous_ie = file_parents.values()[0]
273
weave_store.add_identical_text(
274
self.file_id, previous_ie.revision,
275
self.revision, file_parents)
277
new_lines = work_tree.get_file(self.file_id).readlines()
278
self._add_text_to_weave(new_lines, file_parents, weave_store)
279
self.text_sha1 = sha_strings(new_lines)
280
self.text_size = sum(map(len, new_lines))
214
282
def __eq__(self, other):
215
283
if not isinstance(other, InventoryEntry):
218
286
return (self.file_id == other.file_id) \
219
287
and (self.name == other.name) \
288
and (other.symlink_target == self.symlink_target) \
220
289
and (self.text_sha1 == other.text_sha1) \
221
290
and (self.text_size == other.text_size) \
222
291
and (self.text_id == other.text_id) \
223
292
and (self.parent_id == other.parent_id) \
224
and (self.kind == other.kind)
293
and (self.kind == other.kind) \
294
and (self.revision == other.revision)
227
296
def __ne__(self, other):
228
297
return not (self == other)
230
299
def __hash__(self):
231
300
raise ValueError('not hashable')
302
def _unchanged(self, path, previous_ie, work_tree):
304
# different inv parent
305
if previous_ie.parent_id != self.parent_id:
308
elif previous_ie.name != self.name:
310
if self.kind == 'symlink':
311
if self.symlink_target != previous_ie.symlink_target:
313
if self.kind == 'file':
314
if self.text_sha1 != previous_ie.text_sha1:
317
# FIXME: 20050930 probe for the text size when getting sha1
318
# in _read_tree_state
319
self.text_size = previous_ie.text_size
322
def _read_tree_state(self, path, work_tree):
323
if self.kind == 'symlink':
324
self.read_symlink_target(work_tree.abspath(path))
325
if self.kind == 'file':
326
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
235
329
class RootEntry(InventoryEntry):
476
592
del self[ie.parent_id].children[ie.name]
479
def to_element(self):
480
"""Convert to XML Element"""
481
from bzrlib.xml import Element
483
e = Element('inventory')
485
if self.root.file_id not in (None, ROOT_ID):
486
e.set('file_id', self.root.file_id)
487
for path, ie in self.iter_entries():
488
e.append(ie.to_element())
492
def from_element(cls, elt):
493
"""Construct from XML Element
495
>>> inv = Inventory()
496
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
497
>>> elt = inv.to_element()
498
>>> inv2 = Inventory.from_element(elt)
502
# XXXX: doctest doesn't run this properly under python2.3
503
assert elt.tag == 'inventory'
504
root_id = elt.get('file_id') or ROOT_ID
507
ie = InventoryEntry.from_element(e)
508
if ie.parent_id == ROOT_ID:
509
ie.parent_id = root_id
513
from_element = classmethod(from_element)
516
595
def __eq__(self, other):
517
596
"""Compare two sets by comparing their contents.
added
removed
bzrlib/inventory.py
