« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 1185.16.35
- compare with another revision
- download diff
- download tarball
- view history from revision 1185.16.35
- Committer: Martin Pool
- Date: 2005-10-13 07:30:57 UTC
- mto: (1185.41.1 bzr.sftp) (1464.1.1) (1534.1.1 integration) (1495.1.4) (1505.1.16 bzr-bound-branch) (1508.1.15) (1526.1.3 run_tests_twice_for_i18n) (1540.1.3 bzr.dev (Main development branch)) (1685.1.1 bzr-encoding) (1553.5.1 bzr.dev (Main development branch)) (1608.2.1 bzr.mbp.escape-stores)
- mto: This revision was merged to the branch mainline in revision 1455.
- Revision ID: mbp@sourcefrog.net-20051013073057-05ad152a07fc4278
- stub for revision properties
- files added:
- HACKING
- bzrlib/plugins
- bzrlib/selftest
- bzrlib/store
- bzrlib/transport
- bzrlib/util
- bzrlib/util/effbot
- bzrlib/util/effbot/org
- contrib
- contrib/bash
- contrib/emacs
- contrib/zsh
- patches
- tools
- files removed:
- bzrlib/remotebranch.py
- files renamed:
- bzrlib/store.py => bzrlib/store/__init__.py
- elementtree/ => bzrlib/util/elementtree/
- urlgrabber/ => bzrlib/util/urlgrabber/
- bzrlib/newinventory.py => contrib/newinventory.py
- files modified:
- .bzrignore
- bzr *
added
removed
bzrlib/inventory.py
1
1
# (C) 2005 Canonical Ltd
2
2
#
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
7
7
#
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
12
12
#
13
13
# You should have received a copy of the GNU General Public License
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
16
16
17
17
18
# TODO: Maybe store inventory_id in the file? Not really needed.
19
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.
20
21
21
22
# This should really be an id randomly assigned when the tree is
22
23
# created, but it's not for now.
23
24
ROOT_ID = "TREE_ROOT"
24
25
25
26
26
import sys, os.path, types, re
27
from sets import Set
28
29
try:
30
from cElementTree import Element, ElementTree, SubElement
31
except ImportError:
32
from elementtree.ElementTree import Element, ElementTree, SubElement
33
34
from xml import XMLMixin
35
from errors import bailout, BzrError
27
import os.path
28
import re
29
import sys
30
import tarfile
31
import types
36
32
37
33
import bzrlib
38
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
34
from bzrlib.errors import BzrError, BzrCheckError
35
36
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
37
appendpath, sha_strings)
39
38
from bzrlib.trace import mutter
40
41
class InventoryEntry(XMLMixin):
39
from bzrlib.errors import NotVersionedError
40
41
42
class InventoryEntry(object):
42
43
"""Description of a versioned file.
43
44
44
45
An InventoryEntry has the following fields, which are also
45
46
present in the XML inventory-entry element:
46
47
47
* *file_id*
48
* *name*: (only the basename within the directory, must not
49
contain slashes)
50
* *kind*: "directory" or "file"
51
* *directory_id*: (if absent/null means the branch root directory)
52
* *text_sha1*: only for files
53
* *text_size*: in bytes, only for files
54
* *text_id*: identifier for the text version, only for files
55
56
InventoryEntries can also exist inside a WorkingTree
57
inventory, in which case they are not yet bound to a
58
particular revision of the file. In that case the text_sha1,
59
text_size and text_id are absent.
60
48
file_id
49
50
name
51
(within the parent directory)
52
53
parent_id
54
file_id of the parent directory, or ROOT_ID
55
56
revision
57
the revision_id in which this variation of this file was
58
introduced.
59
60
executable
61
Indicates that this file should be executable on systems
62
that support it.
63
64
text_sha1
65
sha-1 of the text of the file
66
67
text_size
68
size in bytes of the text of the file
69
70
(reading a version 4 tree created a text_id field.)
61
71
62
72
>>> i = Inventory()
63
73
>>> i.path2id('')
64
74
'TREE_ROOT'
65
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
66
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
75
>>> i.add(InventoryDirectory('123', 'src', ROOT_ID))
76
InventoryDirectory('123', 'src', parent_id='TREE_ROOT')
77
>>> i.add(InventoryFile('2323', 'hello.c', parent_id='123'))
78
InventoryFile('2323', 'hello.c', parent_id='123')
67
79
>>> for j in i.iter_entries():
68
80
... print j
69
81
...
70
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
71
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
72
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
82
('src', InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
83
('src/hello.c', InventoryFile('2323', 'hello.c', parent_id='123'))
84
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
73
85
Traceback (most recent call last):
74
86
...
75
BzrError: ('inventory already contains entry with id {2323}', [])
76
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
77
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
87
BzrError: inventory already contains entry with id {2323}
88
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
89
InventoryFile('2324', 'bye.c', parent_id='123')
90
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
91
InventoryDirectory('2325', 'wibble', parent_id='123')
78
92
>>> i.path2id('src/wibble')
79
93
'2325'
80
94
>>> '2325' in i
81
95
True
82
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
96
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
97
InventoryFile('2326', 'wibble.c', parent_id='2325')
83
98
>>> i['2326']
84
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
85
>>> for j in i.iter_entries():
86
... print j[0]
87
... assert i.path2id(j[0])
99
InventoryFile('2326', 'wibble.c', parent_id='2325')
100
>>> for path, entry in i.iter_entries():
101
... print path.replace('\\\\', '/') # for win32 os.sep
102
... assert i.path2id(path)
88
103
...
89
104
src
90
105
src/bye.c
91
106
src/hello.c
92
107
src/wibble
93
108
src/wibble/wibble.c
94
>>> i.id2path('2326')
109
>>> i.id2path('2326').replace('\\\\', '/')
95
110
'src/wibble/wibble.c'
96
97
TODO: Maybe also keep the full path of the entry, and the children?
98
But those depend on its position within a particular inventory, and
99
it would be nice not to need to hold the backpointer here.
100
111
"""
101
102
# TODO: split InventoryEntry into subclasses for files,
103
# directories, etc etc.
104
112
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
113
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
114
'text_id', 'parent_id', 'children', 'executable',
115
'revision']
116
117
def _add_text_to_weave(self, new_lines, parents, weave_store, transaction):
118
weave_store.add_text(self.file_id, self.revision, new_lines, parents,
119
transaction)
120
121
def detect_changes(self, old_entry):
122
"""Return a (text_modified, meta_modified) from this to old_entry.
123
124
_read_tree_state must have been called on self and old_entry prior to
125
calling detect_changes.
126
"""
127
return False, False
128
129
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
130
output_to, reverse=False):
131
"""Perform a diff from this to to_entry.
132
133
text_diff will be used for textual difference calculation.
134
This is a template method, override _diff in child classes.
135
"""
136
self._read_tree_state(tree)
137
if to_entry:
138
# cannot diff from one kind to another - you must do a removal
139
# and an addif they do not match.
140
assert self.kind == to_entry.kind
141
to_entry._read_tree_state(to_tree)
142
self._diff(text_diff, from_label, tree, to_label, to_entry, to_tree,
143
output_to, reverse)
144
145
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
146
output_to, reverse=False):
147
"""Perform a diff between two entries of the same kind."""
148
149
def find_previous_heads(self, previous_inventories, entry_weave):
150
"""Return the revisions and entries that directly preceed this.
151
152
Returned as a map from revision to inventory entry.
153
154
This is a map containing the file revisions in all parents
155
for which the file exists, and its revision is not a parent of
156
any other. If the file is new, the set will be empty.
157
"""
158
def get_ancestors(weave, entry):
159
return set(map(weave.idx_to_name,
160
weave.inclusions([weave.lookup(entry.revision)])))
161
heads = {}
162
head_ancestors = {}
163
for inv in previous_inventories:
164
if self.file_id in inv:
165
ie = inv[self.file_id]
166
assert ie.file_id == self.file_id
167
if ie.revision in heads:
168
# fixup logic, there was a bug in revision updates.
169
# with x bit support.
170
try:
171
if heads[ie.revision].executable != ie.executable:
172
heads[ie.revision].executable = False
173
ie.executable = False
174
except AttributeError:
175
pass
176
assert heads[ie.revision] == ie
177
else:
178
# may want to add it.
179
# may already be covered:
180
already_present = 0 != len(
181
[head for head in heads
182
if ie.revision in head_ancestors[head]])
183
if already_present:
184
# an ancestor of a known head.
185
continue
186
# definately a head:
187
ancestors = get_ancestors(entry_weave, ie)
188
# may knock something else out:
189
check_heads = list(heads.keys())
190
for head in check_heads:
191
if head in ancestors:
192
# this head is not really a head
193
heads.pop(head)
194
head_ancestors[ie.revision] = ancestors
195
heads[ie.revision] = ie
196
return heads
197
198
def get_tar_item(self, root, dp, now, tree):
199
"""Get a tarfile item and a file stream for its content."""
200
item = tarfile.TarInfo(os.path.join(root, dp))
201
# TODO: would be cool to actually set it to the timestamp of the
202
# revision it was last changed
203
item.mtime = now
204
fileobj = self._put_in_tar(item, tree)
205
return item, fileobj
206
207
def has_text(self):
208
"""Return true if the object this entry represents has textual data.
209
210
Note that textual data includes binary content.
211
212
Also note that all entries get weave files created for them.
213
This attribute is primarily used when upgrading from old trees that
214
did not have the weave index for all inventory entries.
215
"""
216
return False
217
218
def __init__(self, file_id, name, parent_id, text_id=None):
106
219
"""Create an InventoryEntry
107
220
108
221
The filename must be a single component, relative to the
109
222
parent directory; it cannot be a whole path or relative name.
110
223
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
224
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
112
225
>>> e.name
113
226
'hello.c'
114
227
>>> e.file_id
115
228
'123'
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
229
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
117
230
Traceback (most recent call last):
118
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
231
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
119
232
"""
120
121
if len(splitpath(name)) != 1:
122
bailout('InventoryEntry name is not a simple filename: %r'
123
% name)
124
233
assert isinstance(name, basestring), name
234
if '/' in name or '\\' in name:
235
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
236
237
self.executable = False
238
self.revision = None
239
self.text_sha1 = None
240
self.text_size = None
125
241
self.file_id = file_id
126
242
self.name = name
127
self.kind = kind
128
243
self.text_id = text_id
129
244
self.parent_id = parent_id
130
self.text_sha1 = None
131
self.text_size = None
132
if kind == 'directory':
133
self.children = {}
134
elif kind == 'file':
135
pass
136
else:
137
raise BzrError("unhandled entry kind %r" % kind)
138
139
245
self.symlink_target = None
246
247
def kind_character(self):
248
"""Return a short kind indicator useful for appending to names."""
249
raise BzrError('unknown kind %r' % self.kind)
250
251
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
252
253
def _put_in_tar(self, item, tree):
254
"""populate item for stashing in a tar, and return the content stream.
255
256
If no content is available, return None.
257
"""
258
raise BzrError("don't know how to export {%s} of kind %r" %
259
(self.file_id, self.kind))
260
261
def put_on_disk(self, dest, dp, tree):
262
"""Create a representation of self on disk in the prefix dest.
263
264
This is a template method - implement _put_on_disk in subclasses.
265
"""
266
fullpath = appendpath(dest, dp)
267
self._put_on_disk(fullpath, tree)
268
mutter(" export {%s} kind %s to %s" % (self.file_id, self.kind, fullpath))
269
270
def _put_on_disk(self, fullpath, tree):
271
"""Put this entry onto disk at fullpath, from tree tree."""
272
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
140
273
141
274
def sorted_children(self):
142
275
l = self.children.items()
143
276
l.sort()
144
277
return l
145
278
279
@staticmethod
280
def versionable_kind(kind):
281
return kind in ('file', 'directory', 'symlink')
282
283
def check(self, checker, rev_id, inv, tree):
284
"""Check this inventory entry is intact.
285
286
This is a template method, override _check for kind specific
287
tests.
288
"""
289
if self.parent_id != None:
290
if not inv.has_id(self.parent_id):
291
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
292
% (self.parent_id, rev_id))
293
self._check(checker, rev_id, tree)
294
295
def _check(self, checker, rev_id, tree):
296
"""Check this inventory entry for kind specific errors."""
297
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
298
(self.kind, rev_id))
299
146
300
147
301
def copy(self):
148
other = InventoryEntry(self.file_id, self.name, self.kind,
149
self.parent_id, text_id=self.text_id)
150
other.text_sha1 = self.text_sha1
151
other.text_size = self.text_size
152
return other
302
"""Clone this inventory entry."""
303
raise NotImplementedError
153
304
305
def _get_snapshot_change(self, previous_entries):
306
if len(previous_entries) > 1:
307
return 'merged'
308
elif len(previous_entries) == 0:
309
return 'added'
310
else:
311
return 'modified/renamed/reparented'
154
312
155
313
def __repr__(self):
156
return ("%s(%r, %r, kind=%r, parent_id=%r)"
314
return ("%s(%r, %r, parent_id=%r)"
157
315
% (self.__class__.__name__,
158
316
self.file_id,
159
317
self.name,
160
self.kind,
161
318
self.parent_id))
162
319
163
164
def to_element(self):
165
"""Convert to XML element"""
166
e = Element('entry')
167
168
e.set('name', self.name)
169
e.set('file_id', self.file_id)
170
e.set('kind', self.kind)
171
172
if self.text_size != None:
173
e.set('text_size', '%d' % self.text_size)
174
175
for f in ['text_id', 'text_sha1']:
176
v = getattr(self, f)
177
if v != None:
178
e.set(f, v)
179
180
# to be conservative, we don't externalize the root pointers
181
# for now, leaving them as null in the xml form. in a future
182
# version it will be implied by nested elements.
183
if self.parent_id != ROOT_ID:
184
assert isinstance(self.parent_id, basestring)
185
e.set('parent_id', self.parent_id)
186
187
e.tail = '\n'
188
189
return e
190
191
192
def from_element(cls, elt):
193
assert elt.tag == 'entry'
194
195
## original format inventories don't have a parent_id for
196
## nodes in the root directory, but it's cleaner to use one
197
## internally.
198
parent_id = elt.get('parent_id')
199
if parent_id == None:
200
parent_id = ROOT_ID
201
202
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
203
self.text_id = elt.get('text_id')
204
self.text_sha1 = elt.get('text_sha1')
205
206
## mutter("read inventoryentry: %r" % (elt.attrib))
207
208
v = elt.get('text_size')
209
self.text_size = v and int(v)
210
211
return self
212
213
214
from_element = classmethod(from_element)
215
216
def __cmp__(self, other):
217
if self is other:
218
return 0
320
def snapshot(self, revision, path, previous_entries,
321
work_tree, weave_store, transaction):
322
"""Make a snapshot of this entry which may or may not have changed.
323
324
This means that all its fields are populated, that it has its
325
text stored in the text store or weave.
326
"""
327
mutter('new parents of %s are %r', path, previous_entries)
328
self._read_tree_state(work_tree)
329
if len(previous_entries) == 1:
330
# cannot be unchanged unless there is only one parent file rev.
331
parent_ie = previous_entries.values()[0]
332
if self._unchanged(parent_ie):
333
mutter("found unchanged entry")
334
self.revision = parent_ie.revision
335
return "unchanged"
336
return self.snapshot_revision(revision, previous_entries,
337
work_tree, weave_store, transaction)
338
339
def snapshot_revision(self, revision, previous_entries, work_tree,
340
weave_store, transaction):
341
"""Record this revision unconditionally."""
342
mutter('new revision for {%s}', self.file_id)
343
self.revision = revision
344
change = self._get_snapshot_change(previous_entries)
345
self._snapshot_text(previous_entries, work_tree, weave_store,
346
transaction)
347
return change
348
349
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
350
"""Record the 'text' of this entry, whatever form that takes.
351
352
This default implementation simply adds an empty text.
353
"""
354
mutter('storing file {%s} in revision {%s}',
355
self.file_id, self.revision)
356
self._add_text_to_weave([], file_parents, weave_store, transaction)
357
358
def __eq__(self, other):
219
359
if not isinstance(other, InventoryEntry):
220
360
return NotImplemented
221
361
222
return cmp(self.file_id, other.file_id) \
223
or cmp(self.name, other.name) \
224
or cmp(self.text_sha1, other.text_sha1) \
225
or cmp(self.text_size, other.text_size) \
226
or cmp(self.text_id, other.text_id) \
227
or cmp(self.parent_id, other.parent_id) \
228
or cmp(self.kind, other.kind)
229
362
return ((self.file_id == other.file_id)
363
and (self.name == other.name)
364
and (other.symlink_target == self.symlink_target)
365
and (self.text_sha1 == other.text_sha1)
366
and (self.text_size == other.text_size)
367
and (self.text_id == other.text_id)
368
and (self.parent_id == other.parent_id)
369
and (self.kind == other.kind)
370
and (self.revision == other.revision)
371
and (self.executable == other.executable)
372
)
373
374
def __ne__(self, other):
375
return not (self == other)
376
377
def __hash__(self):
378
raise ValueError('not hashable')
379
380
def _unchanged(self, previous_ie):
381
"""Has this entry changed relative to previous_ie.
382
383
This method should be overriden in child classes.
384
"""
385
compatible = True
386
# different inv parent
387
if previous_ie.parent_id != self.parent_id:
388
compatible = False
389
# renamed
390
elif previous_ie.name != self.name:
391
compatible = False
392
return compatible
393
394
def _read_tree_state(self, work_tree):
395
"""Populate fields in the inventory entry from the given tree.
396
397
Note that this should be modified to be a noop on virtual trees
398
as all entries created there are prepopulated.
399
"""
400
# TODO: Rather than running this manually, we should check the
401
# working sha1 and other expensive properties when they're
402
# first requested, or preload them if they're already known
403
pass # nothing to do by default
230
404
231
405
232
406
class RootEntry(InventoryEntry):
407
408
def _check(self, checker, rev_id, tree):
409
"""See InventoryEntry._check"""
410
233
411
def __init__(self, file_id):
234
412
self.file_id = file_id
235
413
self.children = {}
237
415
self.parent_id = None
238
416
self.name = ''
239
417
240
def __cmp__(self, other):
241
if self is other:
242
return 0
418
def __eq__(self, other):
243
419
if not isinstance(other, RootEntry):
244
420
return NotImplemented
245
return cmp(self.file_id, other.file_id) \
246
or cmp(self.children, other.children)
247
248
249
250
class Inventory(XMLMixin):
421
422
return (self.file_id == other.file_id) \
423
and (self.children == other.children)
424
425
426
class InventoryDirectory(InventoryEntry):
427
"""A directory in an inventory."""
428
429
def _check(self, checker, rev_id, tree):
430
"""See InventoryEntry._check"""
431
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
432
raise BzrCheckError('directory {%s} has text in revision {%s}'
433
% (self.file_id, rev_id))
434
435
def copy(self):
436
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
437
other.revision = self.revision
438
# note that children are *not* copied; they're pulled across when
439
# others are added
440
return other
441
442
def __init__(self, file_id, name, parent_id):
443
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
444
self.children = {}
445
self.kind = 'directory'
446
447
def kind_character(self):
448
"""See InventoryEntry.kind_character."""
449
return '/'
450
451
def _put_in_tar(self, item, tree):
452
"""See InventoryEntry._put_in_tar."""
453
item.type = tarfile.DIRTYPE
454
fileobj = None
455
item.name += '/'
456
item.size = 0
457
item.mode = 0755
458
return fileobj
459
460
def _put_on_disk(self, fullpath, tree):
461
"""See InventoryEntry._put_on_disk."""
462
os.mkdir(fullpath)
463
464
465
class InventoryFile(InventoryEntry):
466
"""A file in an inventory."""
467
468
def _check(self, checker, rev_id, tree):
469
"""See InventoryEntry._check"""
470
revision = self.revision
471
t = (self.file_id, revision)
472
if t in checker.checked_texts:
473
prev_sha = checker.checked_texts[t]
474
if prev_sha != self.text_sha1:
475
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
476
(self.file_id, rev_id))
477
else:
478
checker.repeated_text_cnt += 1
479
return
480
mutter('check version {%s} of {%s}', rev_id, self.file_id)
481
file_lines = tree.get_file_lines(self.file_id)
482
checker.checked_text_cnt += 1
483
if self.text_size != sum(map(len, file_lines)):
484
raise BzrCheckError('text {%s} wrong size' % self.text_id)
485
if self.text_sha1 != sha_strings(file_lines):
486
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
487
checker.checked_texts[t] = self.text_sha1
488
489
def copy(self):
490
other = InventoryFile(self.file_id, self.name, self.parent_id)
491
other.executable = self.executable
492
other.text_id = self.text_id
493
other.text_sha1 = self.text_sha1
494
other.text_size = self.text_size
495
other.revision = self.revision
496
return other
497
498
def detect_changes(self, old_entry):
499
"""See InventoryEntry.detect_changes."""
500
assert self.text_sha1 != None
501
assert old_entry.text_sha1 != None
502
text_modified = (self.text_sha1 != old_entry.text_sha1)
503
meta_modified = (self.executable != old_entry.executable)
504
return text_modified, meta_modified
505
506
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
507
output_to, reverse=False):
508
"""See InventoryEntry._diff."""
509
from_text = tree.get_file(self.file_id).readlines()
510
if to_entry:
511
to_text = to_tree.get_file(to_entry.file_id).readlines()
512
else:
513
to_text = []
514
if not reverse:
515
text_diff(from_label, from_text,
516
to_label, to_text, output_to)
517
else:
518
text_diff(to_label, to_text,
519
from_label, from_text, output_to)
520
521
def has_text(self):
522
"""See InventoryEntry.has_text."""
523
return True
524
525
def __init__(self, file_id, name, parent_id):
526
super(InventoryFile, self).__init__(file_id, name, parent_id)
527
self.kind = 'file'
528
529
def kind_character(self):
530
"""See InventoryEntry.kind_character."""
531
return ''
532
533
def _put_in_tar(self, item, tree):
534
"""See InventoryEntry._put_in_tar."""
535
item.type = tarfile.REGTYPE
536
fileobj = tree.get_file(self.file_id)
537
item.size = self.text_size
538
if tree.is_executable(self.file_id):
539
item.mode = 0755
540
else:
541
item.mode = 0644
542
return fileobj
543
544
def _put_on_disk(self, fullpath, tree):
545
"""See InventoryEntry._put_on_disk."""
546
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
547
if tree.is_executable(self.file_id):
548
os.chmod(fullpath, 0755)
549
550
def _read_tree_state(self, work_tree):
551
"""See InventoryEntry._read_tree_state."""
552
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
553
self.executable = work_tree.is_executable(self.file_id)
554
555
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
556
"""See InventoryEntry._snapshot_text."""
557
mutter('storing file {%s} in revision {%s}',
558
self.file_id, self.revision)
559
# special case to avoid diffing on renames or
560
# reparenting
561
if (len(file_parents) == 1
562
and self.text_sha1 == file_parents.values()[0].text_sha1
563
and self.text_size == file_parents.values()[0].text_size):
564
previous_ie = file_parents.values()[0]
565
weave_store.add_identical_text(
566
self.file_id, previous_ie.revision,
567
self.revision, file_parents, transaction)
568
else:
569
new_lines = work_tree.get_file(self.file_id).readlines()
570
self._add_text_to_weave(new_lines, file_parents, weave_store,
571
transaction)
572
self.text_sha1 = sha_strings(new_lines)
573
self.text_size = sum(map(len, new_lines))
574
575
576
def _unchanged(self, previous_ie):
577
"""See InventoryEntry._unchanged."""
578
compatible = super(InventoryFile, self)._unchanged(previous_ie)
579
if self.text_sha1 != previous_ie.text_sha1:
580
compatible = False
581
else:
582
# FIXME: 20050930 probe for the text size when getting sha1
583
# in _read_tree_state
584
self.text_size = previous_ie.text_size
585
if self.executable != previous_ie.executable:
586
compatible = False
587
return compatible
588
589
590
class InventoryLink(InventoryEntry):
591
"""A file in an inventory."""
592
593
__slots__ = ['symlink_target']
594
595
def _check(self, checker, rev_id, tree):
596
"""See InventoryEntry._check"""
597
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
598
raise BzrCheckError('symlink {%s} has text in revision {%s}'
599
% (self.file_id, rev_id))
600
if self.symlink_target == None:
601
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
602
% (self.file_id, rev_id))
603
604
def copy(self):
605
other = InventoryLink(self.file_id, self.name, self.parent_id)
606
other.symlink_target = self.symlink_target
607
other.revision = self.revision
608
return other
609
610
def detect_changes(self, old_entry):
611
"""See InventoryEntry.detect_changes."""
612
# FIXME: which _modified field should we use ? RBC 20051003
613
text_modified = (self.symlink_target != old_entry.symlink_target)
614
if text_modified:
615
mutter(" symlink target changed")
616
meta_modified = False
617
return text_modified, meta_modified
618
619
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
620
output_to, reverse=False):
621
"""See InventoryEntry._diff."""
622
from_text = self.symlink_target
623
if to_entry is not None:
624
to_text = to_entry.symlink_target
625
if reverse:
626
temp = from_text
627
from_text = to_text
628
to_text = temp
629
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
630
else:
631
if not reverse:
632
print >>output_to, '=== target was %r' % self.symlink_target
633
else:
634
print >>output_to, '=== target is %r' % self.symlink_target
635
636
def __init__(self, file_id, name, parent_id):
637
super(InventoryLink, self).__init__(file_id, name, parent_id)
638
self.kind = 'symlink'
639
640
def kind_character(self):
641
"""See InventoryEntry.kind_character."""
642
return ''
643
644
def _put_in_tar(self, item, tree):
645
"""See InventoryEntry._put_in_tar."""
646
iterm.type = tarfile.SYMTYPE
647
fileobj = None
648
item.size = 0
649
item.mode = 0755
650
item.linkname = self.symlink_target
651
return fileobj
652
653
def _put_on_disk(self, fullpath, tree):
654
"""See InventoryEntry._put_on_disk."""
655
try:
656
os.symlink(self.symlink_target, fullpath)
657
except OSError,e:
658
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
659
660
def _read_tree_state(self, work_tree):
661
"""See InventoryEntry._read_tree_state."""
662
self.symlink_target = work_tree.get_symlink_target(self.file_id)
663
664
def _unchanged(self, previous_ie):
665
"""See InventoryEntry._unchanged."""
666
compatible = super(InventoryLink, self)._unchanged(previous_ie)
667
if self.symlink_target != previous_ie.symlink_target:
668
compatible = False
669
return compatible
670
671
672
class Inventory(object):
251
673
"""Inventory of versioned files in a tree.
252
674
253
675
This describes which file_id is present at each point in the tree,
265
687
inserted, other than through the Inventory API.
266
688
267
689
>>> inv = Inventory()
268
>>> inv.write_xml(sys.stdout)
269
<inventory>
270
</inventory>
271
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
690
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
691
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
272
692
>>> inv['123-123'].name
273
693
'hello.c'
274
694
283
703
284
704
>>> [x[0] for x in inv.iter_entries()]
285
705
['hello.c']
286
287
>>> inv.write_xml(sys.stdout)
288
<inventory>
289
<entry file_id="123-123" kind="file" name="hello.c" />
290
</inventory>
291
706
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
707
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
708
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
292
709
"""
293
294
## TODO: Make sure only canonical filenames are stored.
295
296
## TODO: Do something sensible about the possible collisions on
297
## case-losing filesystems. Perhaps we should just always forbid
298
## such collisions.
299
300
## TODO: No special cases for root, rather just give it a file id
301
## like everything else.
302
303
## TODO: Probably change XML serialization to use nesting rather
304
## than parent_id pointers.
305
306
## TODO: Perhaps hold the ElementTree in memory and work directly
307
## on that rather than converting into Python objects every time?
308
309
def __init__(self):
710
def __init__(self, root_id=ROOT_ID):
310
711
"""Create or read an inventory.
311
712
312
713
If a working directory is specified, the inventory is read
316
717
The inventory is created with a default root directory, with
317
718
an id of None.
318
719
"""
319
self.root = RootEntry(ROOT_ID)
720
# We are letting Branch.initialize() create a unique inventory
721
# root id. Rather than generating a random one here.
722
#if root_id is None:
723
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
724
self.root = RootEntry(root_id)
320
725
self._byid = {self.root.file_id: self.root}
321
726
322
727
728
def copy(self):
729
other = Inventory(self.root.file_id)
730
# copy recursively so we know directories will be added before
731
# their children. There are more efficient ways than this...
732
for path, entry in self.iter_entries():
733
if entry == self.root:
734
continue
735
other.add(entry.copy())
736
return other
737
738
323
739
def __iter__(self):
324
740
return iter(self._byid)
325
741
343
759
yield name, ie
344
760
if ie.kind == 'directory':
345
761
for cn, cie in self.iter_entries(from_dir=ie.file_id):
346
yield '/'.join((name, cn)), cie
347
762
yield os.path.join(name, cn), cie
763
764
765
def entries(self):
766
"""Return list of (path, ie) for all entries except the root.
767
768
This may be faster than iter_entries.
769
"""
770
accum = []
771
def descend(dir_ie, dir_path):
772
kids = dir_ie.children.items()
773
kids.sort()
774
for name, ie in kids:
775
child_path = os.path.join(dir_path, name)
776
accum.append((child_path, ie))
777
if ie.kind == 'directory':
778
descend(ie, child_path)
779
780
descend(self.root, '')
781
return accum
348
782
349
783
350
784
def directories(self):
351
"""Return (path, entry) pairs for all directories.
785
"""Return (path, entry) pairs for all directories, including the root.
352
786
"""
353
def descend(parent_ie):
354
parent_name = parent_ie.name
355
yield parent_name, parent_ie
356
357
# directory children in sorted order
358
dn = []
359
for ie in parent_ie.children.itervalues():
360
if ie.kind == 'directory':
361
dn.append((ie.name, ie))
362
dn.sort()
787
accum = []
788
def descend(parent_ie, parent_path):
789
accum.append((parent_path, parent_ie))
363
790
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
791
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
792
kids.sort()
367
793
368
for name, ie in descend(self.root):
369
yield name, ie
794
for name, child_ie in kids:
795
child_path = os.path.join(parent_path, name)
796
descend(child_ie, child_path)
797
descend(self.root, '')
798
return accum
370
799
371
800
372
801
374
803
"""True if this entry contains a file with given id.
375
804
376
805
>>> inv = Inventory()
377
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
806
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
807
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
378
808
>>> '123' in inv
379
809
True
380
810
>>> '456' in inv
387
817
"""Return the entry for given file_id.
388
818
389
819
>>> inv = Inventory()
390
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
820
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
821
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
391
822
>>> inv['123123'].name
392
823
'hello.c'
393
824
"""
394
if file_id == None:
395
raise BzrError("can't look up file_id None")
396
397
825
try:
398
826
return self._byid[file_id]
399
827
except KeyError:
400
raise BzrError("file_id {%s} not in inventory" % file_id)
401
828
if file_id == None:
829
raise BzrError("can't look up file_id None")
830
else:
831
raise BzrError("file_id {%s} not in inventory" % file_id)
832
833
834
def get_file_kind(self, file_id):
835
return self._byid[file_id].kind
402
836
403
837
def get_child(self, parent_id, filename):
404
838
return self[parent_id].children.get(filename)
408
842
"""Add entry to inventory.
409
843
410
844
To add a file to a branch ready to be committed, use Branch.add,
411
which calls this."""
845
which calls this.
846
847
Returns the new entry object.
848
"""
412
849
if entry.file_id in self._byid:
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
850
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
851
852
if entry.parent_id == ROOT_ID or entry.parent_id is None:
853
entry.parent_id = self.root.file_id
414
854
415
855
try:
416
856
parent = self._byid[entry.parent_id]
417
857
except KeyError:
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
858
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
419
859
420
860
if parent.children.has_key(entry.name):
421
bailout("%s is already versioned" %
861
raise BzrError("%s is already versioned" %
422
862
appendpath(self.id2path(parent.file_id), entry.name))
423
863
424
864
self._byid[entry.file_id] = entry
425
865
parent.children[entry.name] = entry
866
return entry
426
867
427
868
428
869
def add_path(self, relpath, kind, file_id=None):
429
870
"""Add entry from a path.
430
871
431
The immediate parent must already be versioned"""
872
The immediate parent must already be versioned.
873
874
Returns the new entry object."""
875
from bzrlib.branch import gen_file_id
876
432
877
parts = bzrlib.osutils.splitpath(relpath)
433
878
if len(parts) == 0:
434
bailout("cannot re-add root of inventory")
879
raise BzrError("cannot re-add root of inventory")
435
880
436
881
if file_id == None:
437
file_id = bzrlib.branch.gen_file_id(relpath)
438
439
parent_id = self.path2id(parts[:-1])
440
assert parent_id != None
441
ie = InventoryEntry(file_id, parts[-1],
442
kind=kind, parent_id=parent_id)
882
file_id = gen_file_id(relpath)
883
884
parent_path = parts[:-1]
885
parent_id = self.path2id(parent_path)
886
if parent_id == None:
887
raise NotVersionedError(parent_path)
888
889
if kind == 'directory':
890
ie = InventoryDirectory(file_id, parts[-1], parent_id)
891
elif kind == 'file':
892
ie = InventoryFile(file_id, parts[-1], parent_id)
893
elif kind == 'symlink':
894
ie = InventoryLink(file_id, parts[-1], parent_id)
895
else:
896
raise BzrError("unknown kind %r" % kind)
443
897
return self.add(ie)
444
898
445
899
447
901
"""Remove entry by id.
448
902
449
903
>>> inv = Inventory()
450
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
904
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
905
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
451
906
>>> '123' in inv
452
907
True
453
908
>>> del inv['123']
469
924
del self[ie.parent_id].children[ie.name]
470
925
471
926
472
def id_set(self):
473
return Set(self._byid)
474
475
476
def to_element(self):
477
"""Convert to XML Element"""
478
e = Element('inventory')
479
e.text = '\n'
480
for path, ie in self.iter_entries():
481
e.append(ie.to_element())
482
return e
483
484
485
def from_element(cls, elt):
486
"""Construct from XML Element
487
488
>>> inv = Inventory()
489
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
490
>>> elt = inv.to_element()
491
>>> inv2 = Inventory.from_element(elt)
492
>>> inv2 == inv
493
True
494
"""
495
assert elt.tag == 'inventory'
496
o = cls()
497
for e in elt:
498
o.add(InventoryEntry.from_element(e))
499
return o
500
501
from_element = classmethod(from_element)
502
503
504
def __cmp__(self, other):
927
def __eq__(self, other):
505
928
"""Compare two sets by comparing their contents.
506
929
507
930
>>> i1 = Inventory()
508
931
>>> i2 = Inventory()
509
932
>>> i1 == i2
510
933
True
511
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
934
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
935
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
512
936
>>> i1 == i2
513
937
False
514
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
938
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
939
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
515
940
>>> i1 == i2
516
941
True
517
942
"""
518
if self is other:
519
return 0
520
521
943
if not isinstance(other, Inventory):
522
944
return NotImplemented
523
945
524
if self.id_set() ^ other.id_set():
525
return 1
526
527
for file_id in self._byid:
528
c = cmp(self[file_id], other[file_id])
529
if c: return c
530
531
return 0
946
if len(self._byid) != len(other._byid):
947
# shortcut: obviously not the same
948
return False
949
950
return self._byid == other._byid
951
952
953
def __ne__(self, other):
954
return not self.__eq__(other)
955
956
957
def __hash__(self):
958
raise ValueError('not hashable')
532
959
533
960
534
961
def get_idpath(self, file_id):
544
971
try:
545
972
ie = self._byid[file_id]
546
973
except KeyError:
547
bailout("file_id {%s} not found in inventory" % file_id)
974
raise BzrError("file_id {%s} not found in inventory" % file_id)
548
975
p.insert(0, ie.file_id)
549
976
file_id = ie.parent_id
550
977
return p
551
978
552
979
553
980
def id2path(self, file_id):
554
"""Return as a list the path to file_id."""
555
981
"""Return as a list the path to file_id.
982
983
>>> i = Inventory()
984
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
985
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
986
>>> print i.id2path('foo-id').replace(os.sep, '/')
987
src/foo.c
988
"""
556
989
# get all names, skipping root
557
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
558
return '/'.join(p)
990
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
991
return os.sep.join(p)
559
992
560
993
561
994
604
1037
605
1038
This does not move the working file."""
606
1039
if not is_valid_name(new_name):
607
bailout("not an acceptable filename: %r" % new_name)
1040
raise BzrError("not an acceptable filename: %r" % new_name)
608
1041
609
1042
new_parent = self._byid[new_parent_id]
610
1043
if new_name in new_parent.children:
611
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
1044
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
612
1045
613
1046
new_parent_idpath = self.get_idpath(new_parent_id)
614
1047
if file_id in new_parent_idpath:
615
bailout("cannot move directory %r into a subdirectory of itself, %r"
1048
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
616
1049
% (self.id2path(file_id), self.id2path(new_parent_id)))
617
1050
618
1051
file_ie = self._byid[file_id]
629
1062
630
1063
631
1064
632
_NAME_RE = re.compile(r'^[^/\\]+$')
1065
_NAME_RE = None
633
1066
634
1067
def is_valid_name(name):
1068
global _NAME_RE
1069
if _NAME_RE == None:
1070
_NAME_RE = re.compile(r'^[^/\\]+$')
1071
635
1072
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1