« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 1418
- compare with another revision
- download diff
- download tarball
- view history from revision 1418
- Committer: Robert Collins
- Date: 2005-10-06 12:14:01 UTC
- mfrom: (1393.1.67)
- Revision ID: robertc@robertcollins.net-20051006121401-ce87bcb93909bbdf
merge martins latest
- files added:
- .rsyncexclude
- bzrlib/plugins
- bzrlib/selftest
- bzrlib/store
- bzrlib/transport
- bzrlib/util
- bzrlib/util/effbot
- bzrlib/util/effbot/org
- bzrlib/util/urlgrabber
- contrib
- contrib/bash
- contrib/emacs
- contrib/zsh
- patches
- tools
- files removed:
- bzrlib/tests.py
- files renamed:
- bzrlib/store.py => bzrlib/store/__init__.py
- elementtree/ => bzrlib/util/elementtree/
- files modified:
- .bzrignore
added
removed
bzrlib/inventory.py
1
#! /usr/bin/env python
2
# -*- coding: UTF-8 -*-
1
# (C) 2005 Canonical Ltd
3
2
4
3
# This program is free software; you can redistribute it and/or modify
5
4
# it under the terms of the GNU General Public License as published by
15
14
# along with this program; if not, write to the Free Software
16
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
16
18
"""Inventories map files to their name in a revision."""
19
20
21
__copyright__ = "Copyright (C) 2005 Canonical Ltd."
22
__author__ = "Martin Pool <mbp@canonical.com>"
23
24
import sys, os.path, types
25
from sets import Set
26
27
try:
28
from cElementTree import Element, ElementTree, SubElement
29
except ImportError:
30
from elementtree.ElementTree import Element, ElementTree, SubElement
31
32
from xml import XMLMixin
33
from errors import bailout
17
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.
21
22
# This should really be an id randomly assigned when the tree is
23
# created, but it's not for now.
24
ROOT_ID = "TREE_ROOT"
25
26
27
import os.path
28
import re
29
import sys
30
import tarfile
31
import types
34
32
35
33
import bzrlib
36
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)
37
38
from bzrlib.trace import mutter
38
39
class InventoryEntry(XMLMixin):
39
from bzrlib.errors import NotVersionedError
40
41
42
class InventoryEntry(object):
40
43
"""Description of a versioned file.
41
44
42
45
An InventoryEntry has the following fields, which are also
43
46
present in the XML inventory-entry element:
44
47
45
* *file_id*
46
* *name*: (only the basename within the directory, must not
47
contain slashes)
48
* *kind*: "directory" or "file"
49
* *directory_id*: (if absent/null means the branch root directory)
50
* *text_sha1*: only for files
51
* *text_size*: in bytes, only for files
52
* *text_id*: identifier for the text version, only for files
53
54
InventoryEntries can also exist inside a WorkingTree
55
inventory, in which case they are not yet bound to a
56
particular revision of the file. In that case the text_sha1,
57
text_size and text_id are absent.
58
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.)
59
71
60
72
>>> i = Inventory()
61
73
>>> i.path2id('')
62
>>> i.add(InventoryEntry('123', 'src', kind='directory'))
63
>>> i.add(InventoryEntry('2323', 'hello.c', parent_id='123'))
74
'TREE_ROOT'
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')
64
79
>>> for j in i.iter_entries():
65
80
... print j
66
81
...
67
('src', InventoryEntry('123', 'src', kind='directory', parent_id=None))
68
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
69
>>> i.add(InventoryEntry('2323', 'bye.c', parent_id='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'))
70
85
Traceback (most recent call last):
71
86
...
72
BzrError: ('inventory already contains entry with id {2323}', [])
73
>>> i.add(InventoryEntry('2324', 'bye.c', parent_id='123'))
74
>>> i.add(InventoryEntry('2325', 'wibble', parent_id='123', kind='directory'))
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')
75
92
>>> i.path2id('src/wibble')
76
93
'2325'
77
94
>>> '2325' in i
78
95
True
79
>>> i.add(InventoryEntry('2326', 'wibble.c', parent_id='2325'))
96
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
97
InventoryFile('2326', 'wibble.c', parent_id='2325')
80
98
>>> i['2326']
81
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
82
>>> for j in i.iter_entries():
83
... print j[0]
84
... 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)
85
103
...
86
104
src
87
105
src/bye.c
88
106
src/hello.c
89
107
src/wibble
90
108
src/wibble/wibble.c
91
>>> i.id2path('2326')
109
>>> i.id2path('2326').replace('\\\\', '/')
92
110
'src/wibble/wibble.c'
93
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.
97
111
"""
98
def __init__(self, file_id, name, kind='file', text_id=None,
99
parent_id=None):
112
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):
118
weave_store.add_text(self.file_id, self.revision, new_lines, parents)
119
120
def detect_changes(self, old_entry):
121
"""Return a (text_modified, meta_modified) from this to old_entry.
122
123
_read_tree_state must have been called on self and old_entry prior to
124
calling detect_changes.
125
"""
126
return False, False
127
128
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
129
output_to, reverse=False):
130
"""Perform a diff from this to to_entry.
131
132
text_diff will be used for textual difference calculation.
133
This is a template method, override _diff in child classes.
134
"""
135
self._read_tree_state(tree.id2path(self.file_id), tree)
136
if to_entry:
137
# cannot diff from one kind to another - you must do a removal
138
# and an addif they do not match.
139
assert self.kind == to_entry.kind
140
to_entry._read_tree_state(to_tree.id2path(to_entry.file_id),
141
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
assert heads[ie.revision] == ie
169
else:
170
# may want to add it.
171
# may already be covered:
172
already_present = 0 != len(
173
[head for head in heads
174
if ie.revision in head_ancestors[head]])
175
if already_present:
176
# an ancestor of a known head.
177
continue
178
# definately a head:
179
ancestors = get_ancestors(entry_weave, ie)
180
# may knock something else out:
181
check_heads = list(heads.keys())
182
for head in check_heads:
183
if head in ancestors:
184
# this head is not really a head
185
heads.pop(head)
186
head_ancestors[ie.revision] = ancestors
187
heads[ie.revision] = ie
188
return heads
189
190
def get_tar_item(self, root, dp, now, tree):
191
"""Get a tarfile item and a file stream for its content."""
192
item = tarfile.TarInfo(os.path.join(root, dp))
193
# TODO: would be cool to actually set it to the timestamp of the
194
# revision it was last changed
195
item.mtime = now
196
fileobj = self._put_in_tar(item, tree)
197
return item, fileobj
198
199
def has_text(self):
200
"""Return true if the object this entry represents has textual data.
201
202
Note that textual data includes binary content.
203
204
Also note that all entries get weave files created for them.
205
This attribute is primarily used when upgrading from old trees that
206
did not have the weave index for all inventory entries.
207
"""
208
return False
209
210
def __init__(self, file_id, name, parent_id, text_id=None):
100
211
"""Create an InventoryEntry
101
212
102
213
The filename must be a single component, relative to the
103
214
parent directory; it cannot be a whole path or relative name.
104
215
105
>>> e = InventoryEntry('123', 'hello.c')
216
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
106
217
>>> e.name
107
218
'hello.c'
108
219
>>> e.file_id
109
220
'123'
110
>>> e = InventoryEntry('123', 'src/hello.c')
221
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
111
222
Traceback (most recent call last):
112
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
223
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
113
224
"""
114
115
if len(splitpath(name)) != 1:
116
bailout('InventoryEntry name is not a simple filename: %r'
117
% name)
118
225
assert isinstance(name, basestring), name
226
if '/' in name or '\\' in name:
227
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
228
229
self.executable = False
230
self.revision = None
231
self.text_sha1 = None
232
self.text_size = None
119
233
self.file_id = file_id
120
234
self.name = name
121
assert kind in ['file', 'directory']
122
self.kind = kind
123
235
self.text_id = text_id
124
236
self.parent_id = parent_id
125
self.text_sha1 = None
126
self.text_size = None
237
self.symlink_target = None
238
239
def kind_character(self):
240
"""Return a short kind indicator useful for appending to names."""
241
raise BzrError('unknown kind %r' % self.kind)
242
243
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
244
245
def _put_in_tar(self, item, tree):
246
"""populate item for stashing in a tar, and return the content stream.
247
248
If no content is available, return None.
249
"""
250
raise BzrError("don't know how to export {%s} of kind %r" %
251
(self.file_id, self.kind))
252
253
def put_on_disk(self, dest, dp, tree):
254
"""Create a representation of self on disk in the prefix dest.
255
256
This is a template method - implement _put_on_disk in subclasses.
257
"""
258
fullpath = appendpath(dest, dp)
259
self._put_on_disk(fullpath, tree)
260
mutter(" export {%s} kind %s to %s" % (self.file_id, self.kind, fullpath))
261
262
def _put_on_disk(self, fullpath, tree):
263
"""Put this entry onto disk at fullpath, from tree tree."""
264
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
265
266
def sorted_children(self):
267
l = self.children.items()
268
l.sort()
269
return l
270
271
@staticmethod
272
def versionable_kind(kind):
273
return kind in ('file', 'directory', 'symlink')
274
275
def check(self, checker, rev_id, inv, tree):
276
"""Check this inventory entry is intact.
277
278
This is a template method, override _check for kind specific
279
tests.
280
"""
281
if self.parent_id != None:
282
if not inv.has_id(self.parent_id):
283
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
284
% (self.parent_id, rev_id))
285
self._check(checker, rev_id, tree)
286
287
def _check(self, checker, rev_id, tree):
288
"""Check this inventory entry for kind specific errors."""
289
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
290
(self.kind, rev_id))
127
291
128
292
129
293
def copy(self):
130
other = InventoryEntry(self.file_id, self.name, self.kind,
131
self.text_id, self.parent_id)
132
other.text_sha1 = self.text_sha1
133
other.text_size = self.text_size
134
return other
294
"""Clone this inventory entry."""
295
raise NotImplementedError
135
296
297
def _get_snapshot_change(self, previous_entries):
298
if len(previous_entries) > 1:
299
return 'merged'
300
elif len(previous_entries) == 0:
301
return 'added'
302
else:
303
return 'modified/renamed/reparented'
136
304
137
305
def __repr__(self):
138
return ("%s(%r, %r, kind=%r, parent_id=%r)"
306
return ("%s(%r, %r, parent_id=%r)"
139
307
% (self.__class__.__name__,
140
308
self.file_id,
141
309
self.name,
142
self.kind,
143
310
self.parent_id))
144
311
145
146
def to_element(self):
147
"""Convert to XML element"""
148
e = Element('entry')
149
150
e.set('name', self.name)
151
e.set('file_id', self.file_id)
152
e.set('kind', self.kind)
153
154
if self.text_size is not None:
155
e.set('text_size', '%d' % self.text_size)
156
157
for f in ['text_id', 'text_sha1', 'parent_id']:
158
v = getattr(self, f)
159
if v is not None:
160
e.set(f, v)
161
162
e.tail = '\n'
163
164
return e
165
166
167
def from_element(cls, elt):
168
assert elt.tag == 'entry'
169
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'))
170
self.text_id = elt.get('text_id')
171
self.text_sha1 = elt.get('text_sha1')
172
self.parent_id = elt.get('parent_id')
173
174
## mutter("read inventoryentry: %r" % (elt.attrib))
175
176
v = elt.get('text_size')
177
self.text_size = v and int(v)
178
179
return self
180
181
182
from_element = classmethod(from_element)
183
184
def __cmp__(self, other):
185
if self is other:
186
return 0
312
def snapshot(self, revision, path, previous_entries,
313
work_tree, weave_store):
314
"""Make a snapshot of this entry which may or may not have changed.
315
316
This means that all its fields are populated, that it has its
317
text stored in the text store or weave.
318
"""
319
mutter('new parents of %s are %r', path, previous_entries)
320
self._read_tree_state(path, work_tree)
321
if len(previous_entries) == 1:
322
# cannot be unchanged unless there is only one parent file rev.
323
parent_ie = previous_entries.values()[0]
324
if self._unchanged(parent_ie):
325
mutter("found unchanged entry")
326
self.revision = parent_ie.revision
327
return "unchanged"
328
return self.snapshot_revision(revision, previous_entries,
329
work_tree, weave_store)
330
331
def snapshot_revision(self, revision, previous_entries, work_tree,
332
weave_store):
333
"""Record this revision unconditionally."""
334
mutter('new revision for {%s}', self.file_id)
335
self.revision = revision
336
change = self._get_snapshot_change(previous_entries)
337
self._snapshot_text(previous_entries, work_tree, weave_store)
338
return change
339
340
def _snapshot_text(self, file_parents, work_tree, weave_store):
341
"""Record the 'text' of this entry, whatever form that takes.
342
343
This default implementation simply adds an empty text.
344
"""
345
mutter('storing file {%s} in revision {%s}',
346
self.file_id, self.revision)
347
self._add_text_to_weave([], file_parents, weave_store)
348
349
def __eq__(self, other):
187
350
if not isinstance(other, InventoryEntry):
188
351
return NotImplemented
189
352
190
return cmp(self.file_id, other.file_id) \
191
or cmp(self.name, other.name) \
192
or cmp(self.text_sha1, other.text_sha1) \
193
or cmp(self.text_size, other.text_size) \
194
or cmp(self.text_id, other.text_id) \
195
or cmp(self.parent_id, other.parent_id) \
196
or cmp(self.kind, other.kind)
197
198
199
200
class Inventory(XMLMixin):
353
return ((self.file_id == other.file_id)
354
and (self.name == other.name)
355
and (other.symlink_target == self.symlink_target)
356
and (self.text_sha1 == other.text_sha1)
357
and (self.text_size == other.text_size)
358
and (self.text_id == other.text_id)
359
and (self.parent_id == other.parent_id)
360
and (self.kind == other.kind)
361
and (self.revision == other.revision)
362
and (self.executable == other.executable)
363
)
364
365
def __ne__(self, other):
366
return not (self == other)
367
368
def __hash__(self):
369
raise ValueError('not hashable')
370
371
def _unchanged(self, previous_ie):
372
"""Has this entry changed relative to previous_ie.
373
374
This method should be overriden in child classes.
375
"""
376
compatible = True
377
# different inv parent
378
if previous_ie.parent_id != self.parent_id:
379
compatible = False
380
# renamed
381
elif previous_ie.name != self.name:
382
compatible = False
383
return compatible
384
385
def _read_tree_state(self, path, work_tree):
386
"""Populate fields in the inventory entry from the given tree.
387
388
Note that this should be modified to be a noop on virtual trees
389
as all entries created there are prepopulated.
390
"""
391
392
393
class RootEntry(InventoryEntry):
394
395
def _check(self, checker, rev_id, tree):
396
"""See InventoryEntry._check"""
397
398
def __init__(self, file_id):
399
self.file_id = file_id
400
self.children = {}
401
self.kind = 'root_directory'
402
self.parent_id = None
403
self.name = ''
404
405
def __eq__(self, other):
406
if not isinstance(other, RootEntry):
407
return NotImplemented
408
409
return (self.file_id == other.file_id) \
410
and (self.children == other.children)
411
412
413
class InventoryDirectory(InventoryEntry):
414
"""A directory in an inventory."""
415
416
def _check(self, checker, rev_id, tree):
417
"""See InventoryEntry._check"""
418
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
419
raise BzrCheckError('directory {%s} has text in revision {%s}'
420
% (self.file_id, rev_id))
421
422
def copy(self):
423
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
424
other.revision = self.revision
425
# note that children are *not* copied; they're pulled across when
426
# others are added
427
return other
428
429
def __init__(self, file_id, name, parent_id):
430
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
431
self.children = {}
432
self.kind = 'directory'
433
434
def kind_character(self):
435
"""See InventoryEntry.kind_character."""
436
return '/'
437
438
def _put_in_tar(self, item, tree):
439
"""See InventoryEntry._put_in_tar."""
440
item.type = tarfile.DIRTYPE
441
fileobj = None
442
item.name += '/'
443
item.size = 0
444
item.mode = 0755
445
return fileobj
446
447
def _put_on_disk(self, fullpath, tree):
448
"""See InventoryEntry._put_on_disk."""
449
os.mkdir(fullpath)
450
451
452
class InventoryFile(InventoryEntry):
453
"""A file in an inventory."""
454
455
def _check(self, checker, rev_id, tree):
456
"""See InventoryEntry._check"""
457
revision = self.revision
458
t = (self.file_id, revision)
459
if t in checker.checked_texts:
460
prev_sha = checker.checked_texts[t]
461
if prev_sha != self.text_sha1:
462
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
463
(self.file_id, rev_id))
464
else:
465
checker.repeated_text_cnt += 1
466
return
467
mutter('check version {%s} of {%s}', rev_id, self.file_id)
468
file_lines = tree.get_file_lines(self.file_id)
469
checker.checked_text_cnt += 1
470
if self.text_size != sum(map(len, file_lines)):
471
raise BzrCheckError('text {%s} wrong size' % self.text_id)
472
if self.text_sha1 != sha_strings(file_lines):
473
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
474
checker.checked_texts[t] = self.text_sha1
475
476
def copy(self):
477
other = InventoryFile(self.file_id, self.name, self.parent_id)
478
other.executable = self.executable
479
other.text_id = self.text_id
480
other.text_sha1 = self.text_sha1
481
other.text_size = self.text_size
482
other.revision = self.revision
483
return other
484
485
def detect_changes(self, old_entry):
486
"""See InventoryEntry.detect_changes."""
487
assert self.text_sha1 != None
488
assert old_entry.text_sha1 != None
489
text_modified = (self.text_sha1 != old_entry.text_sha1)
490
meta_modified = (self.executable != old_entry.executable)
491
return text_modified, meta_modified
492
493
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
494
output_to, reverse=False):
495
"""See InventoryEntry._diff."""
496
from_text = tree.get_file(self.file_id).readlines()
497
if to_entry:
498
to_text = to_tree.get_file(to_entry.file_id).readlines()
499
else:
500
to_text = []
501
if not reverse:
502
text_diff(from_label, from_text,
503
to_label, to_text, output_to)
504
else:
505
text_diff(to_label, to_text,
506
from_label, from_text, output_to)
507
508
def has_text(self):
509
"""See InventoryEntry.has_text."""
510
return True
511
512
def __init__(self, file_id, name, parent_id):
513
super(InventoryFile, self).__init__(file_id, name, parent_id)
514
self.kind = 'file'
515
516
def kind_character(self):
517
"""See InventoryEntry.kind_character."""
518
return ''
519
520
def _put_in_tar(self, item, tree):
521
"""See InventoryEntry._put_in_tar."""
522
item.type = tarfile.REGTYPE
523
fileobj = tree.get_file(self.file_id)
524
item.size = self.text_size
525
if tree.is_executable(self.file_id):
526
item.mode = 0755
527
else:
528
item.mode = 0644
529
return fileobj
530
531
def _put_on_disk(self, fullpath, tree):
532
"""See InventoryEntry._put_on_disk."""
533
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
534
if tree.is_executable(self.file_id):
535
os.chmod(fullpath, 0755)
536
537
def _read_tree_state(self, path, work_tree):
538
"""See InventoryEntry._read_tree_state."""
539
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
540
self.executable = work_tree.is_executable(self.file_id)
541
542
def _snapshot_text(self, file_parents, work_tree, weave_store):
543
"""See InventoryEntry._snapshot_text."""
544
mutter('storing file {%s} in revision {%s}',
545
self.file_id, self.revision)
546
# special case to avoid diffing on renames or
547
# reparenting
548
if (len(file_parents) == 1
549
and self.text_sha1 == file_parents.values()[0].text_sha1
550
and self.text_size == file_parents.values()[0].text_size):
551
previous_ie = file_parents.values()[0]
552
weave_store.add_identical_text(
553
self.file_id, previous_ie.revision,
554
self.revision, file_parents)
555
else:
556
new_lines = work_tree.get_file(self.file_id).readlines()
557
self._add_text_to_weave(new_lines, file_parents, weave_store)
558
self.text_sha1 = sha_strings(new_lines)
559
self.text_size = sum(map(len, new_lines))
560
561
562
def _unchanged(self, previous_ie):
563
"""See InventoryEntry._unchanged."""
564
compatible = super(InventoryFile, self)._unchanged(previous_ie)
565
if self.text_sha1 != previous_ie.text_sha1:
566
compatible = False
567
else:
568
# FIXME: 20050930 probe for the text size when getting sha1
569
# in _read_tree_state
570
self.text_size = previous_ie.text_size
571
return compatible
572
573
574
class InventoryLink(InventoryEntry):
575
"""A file in an inventory."""
576
577
__slots__ = ['symlink_target']
578
579
def _check(self, checker, rev_id, tree):
580
"""See InventoryEntry._check"""
581
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
582
raise BzrCheckError('symlink {%s} has text in revision {%s}'
583
% (self.file_id, rev_id))
584
if self.symlink_target == None:
585
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
586
% (self.file_id, rev_id))
587
588
def copy(self):
589
other = InventoryLink(self.file_id, self.name, self.parent_id)
590
other.symlink_target = self.symlink_target
591
other.revision = self.revision
592
return other
593
594
def detect_changes(self, old_entry):
595
"""See InventoryEntry.detect_changes."""
596
# FIXME: which _modified field should we use ? RBC 20051003
597
text_modified = (self.symlink_target != old_entry.symlink_target)
598
if text_modified:
599
mutter(" symlink target changed")
600
meta_modified = False
601
return text_modified, meta_modified
602
603
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
604
output_to, reverse=False):
605
"""See InventoryEntry._diff."""
606
from_text = self.symlink_target
607
if to_entry is not None:
608
to_text = to_entry.symlink_target
609
if reverse:
610
temp = from_text
611
from_text = to_text
612
to_text = temp
613
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
614
else:
615
if not reverse:
616
print >>output_to, '=== target was %r' % self.symlink_target
617
else:
618
print >>output_to, '=== target is %r' % self.symlink_target
619
620
def __init__(self, file_id, name, parent_id):
621
super(InventoryLink, self).__init__(file_id, name, parent_id)
622
self.kind = 'symlink'
623
624
def kind_character(self):
625
"""See InventoryEntry.kind_character."""
626
return ''
627
628
def _put_in_tar(self, item, tree):
629
"""See InventoryEntry._put_in_tar."""
630
iterm.type = tarfile.SYMTYPE
631
fileobj = None
632
item.size = 0
633
item.mode = 0755
634
item.linkname = self.symlink_target
635
return fileobj
636
637
def _put_on_disk(self, fullpath, tree):
638
"""See InventoryEntry._put_on_disk."""
639
try:
640
os.symlink(self.symlink_target, fullpath)
641
except OSError,e:
642
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
643
644
def _read_tree_state(self, path, work_tree):
645
"""See InventoryEntry._read_tree_state."""
646
self.symlink_target = work_tree.get_symlink_target(self.file_id)
647
648
def _unchanged(self, previous_ie):
649
"""See InventoryEntry._unchanged."""
650
compatible = super(InventoryLink, self)._unchanged(previous_ie)
651
if self.symlink_target != previous_ie.symlink_target:
652
compatible = False
653
return compatible
654
655
656
class Inventory(object):
201
657
"""Inventory of versioned files in a tree.
202
658
203
An Inventory acts like a set of InventoryEntry items. You can
204
also look files up by their file_id or name.
205
206
May be read from and written to a metadata file in a tree. To
207
manipulate the inventory (for example to add a file), it is read
208
in, modified, and then written back out.
659
This describes which file_id is present at each point in the tree,
660
and possibly the SHA-1 or other information about the file.
661
Entries can be looked up either by path or by file_id.
209
662
210
663
The inventory represents a typical unix file tree, with
211
664
directories containing files and subdirectories. We never store
215
668
returned quickly.
216
669
217
670
InventoryEntry objects must not be modified after they are
218
inserted.
671
inserted, other than through the Inventory API.
219
672
220
673
>>> inv = Inventory()
221
>>> inv.write_xml(sys.stdout)
222
<inventory>
223
</inventory>
224
>>> inv.add(InventoryEntry('123-123', 'hello.c'))
674
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
675
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
225
676
>>> inv['123-123'].name
226
677
'hello.c'
227
>>> for file_id in inv: print file_id
228
...
229
123-123
230
678
231
679
May be treated as an iterator or set to look up file ids:
232
680
239
687
240
688
>>> [x[0] for x in inv.iter_entries()]
241
689
['hello.c']
242
243
>>> inv.write_xml(sys.stdout)
244
<inventory>
245
<entry file_id="123-123" kind="file" name="hello.c" />
246
</inventory>
247
690
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
691
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
692
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
248
693
"""
249
250
## TODO: Clear up handling of files in subdirectories; we probably
251
## do want to be able to just look them up by name but this
252
## probably means gradually walking down the path, looking up as we go.
253
254
## TODO: Make sure only canonical filenames are stored.
255
256
## TODO: Do something sensible about the possible collisions on
257
## case-losing filesystems. Perhaps we should just always forbid
258
## such collisions.
259
260
## _tree should probably just be stored as
261
## InventoryEntry._children on each directory.
262
263
def __init__(self):
694
def __init__(self, root_id=ROOT_ID):
264
695
"""Create or read an inventory.
265
696
266
697
If a working directory is specified, the inventory is read
267
698
from there. If the file is specified, read from that. If not,
268
699
the inventory is created empty.
700
701
The inventory is created with a default root directory, with
702
an id of None.
269
703
"""
270
self._byid = dict()
271
272
# _tree is indexed by parent_id; at each level a map from name
273
# to ie. The None entry is the root.
274
self._tree = {None: {}}
704
# We are letting Branch.initialize() create a unique inventory
705
# root id. Rather than generating a random one here.
706
#if root_id is None:
707
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
708
self.root = RootEntry(root_id)
709
self._byid = {self.root.file_id: self.root}
710
711
712
def copy(self):
713
other = Inventory(self.root.file_id)
714
# copy recursively so we know directories will be added before
715
# their children. There are more efficient ways than this...
716
for path, entry in self.iter_entries():
717
if entry == self.root:
718
continue
719
other.add(entry.copy())
720
return other
275
721
276
722
277
723
def __iter__(self):
283
729
return len(self._byid)
284
730
285
731
286
def iter_entries(self, parent_id=None):
732
def iter_entries(self, from_dir=None):
287
733
"""Return (path, entry) pairs, in order by name."""
288
kids = self._tree[parent_id].items()
734
if from_dir == None:
735
assert self.root
736
from_dir = self.root
737
elif isinstance(from_dir, basestring):
738
from_dir = self._byid[from_dir]
739
740
kids = from_dir.children.items()
289
741
kids.sort()
290
742
for name, ie in kids:
291
743
yield name, ie
292
744
if ie.kind == 'directory':
293
for cn, cie in self.iter_entries(parent_id=ie.file_id):
294
yield joinpath([name, cn]), cie
295
296
297
def directories(self, include_root=True):
298
"""Return (path, entry) pairs for all directories.
299
"""
300
if include_root:
301
yield '', None
302
for path, entry in self.iter_entries():
303
if entry.kind == 'directory':
304
yield path, entry
745
for cn, cie in self.iter_entries(from_dir=ie.file_id):
746
yield os.path.join(name, cn), cie
747
748
749
def entries(self):
750
"""Return list of (path, ie) for all entries except the root.
751
752
This may be faster than iter_entries.
753
"""
754
accum = []
755
def descend(dir_ie, dir_path):
756
kids = dir_ie.children.items()
757
kids.sort()
758
for name, ie in kids:
759
child_path = os.path.join(dir_path, name)
760
accum.append((child_path, ie))
761
if ie.kind == 'directory':
762
descend(ie, child_path)
763
764
descend(self.root, '')
765
return accum
766
767
768
def directories(self):
769
"""Return (path, entry) pairs for all directories, including the root.
770
"""
771
accum = []
772
def descend(parent_ie, parent_path):
773
accum.append((parent_path, parent_ie))
774
775
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
776
kids.sort()
777
778
for name, child_ie in kids:
779
child_path = os.path.join(parent_path, name)
780
descend(child_ie, child_path)
781
descend(self.root, '')
782
return accum
305
783
306
784
307
785
308
def children(self, parent_id):
309
"""Return entries that are direct children of parent_id."""
310
return self._tree[parent_id]
311
312
313
314
# TODO: return all paths and entries
315
316
317
786
def __contains__(self, file_id):
318
787
"""True if this entry contains a file with given id.
319
788
320
789
>>> inv = Inventory()
321
>>> inv.add(InventoryEntry('123', 'foo.c'))
790
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
791
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
322
792
>>> '123' in inv
323
793
True
324
794
>>> '456' in inv
331
801
"""Return the entry for given file_id.
332
802
333
803
>>> inv = Inventory()
334
>>> inv.add(InventoryEntry('123123', 'hello.c'))
804
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
805
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
335
806
>>> inv['123123'].name
336
807
'hello.c'
337
808
"""
338
return self._byid[file_id]
809
try:
810
return self._byid[file_id]
811
except KeyError:
812
if file_id == None:
813
raise BzrError("can't look up file_id None")
814
else:
815
raise BzrError("file_id {%s} not in inventory" % file_id)
816
817
818
def get_file_kind(self, file_id):
819
return self._byid[file_id].kind
820
821
def get_child(self, parent_id, filename):
822
return self[parent_id].children.get(filename)
339
823
340
824
341
825
def add(self, entry):
342
826
"""Add entry to inventory.
343
827
344
828
To add a file to a branch ready to be committed, use Branch.add,
345
which calls this."""
346
if entry.file_id in self:
347
bailout("inventory already contains entry with id {%s}" % entry.file_id)
348
349
if entry.parent_id != None:
350
if entry.parent_id not in self:
351
bailout("parent_id %s of new entry not found in inventory"
352
% entry.parent_id)
353
354
if self._tree[entry.parent_id].has_key(entry.name):
355
bailout("%s is already versioned"
356
% appendpath(self.id2path(entry.parent_id), entry.name))
829
which calls this.
830
831
Returns the new entry object.
832
"""
833
if entry.file_id in self._byid:
834
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
835
836
if entry.parent_id == ROOT_ID or entry.parent_id is None:
837
entry.parent_id = self.root.file_id
838
839
try:
840
parent = self._byid[entry.parent_id]
841
except KeyError:
842
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
843
844
if parent.children.has_key(entry.name):
845
raise BzrError("%s is already versioned" %
846
appendpath(self.id2path(parent.file_id), entry.name))
357
847
358
848
self._byid[entry.file_id] = entry
359
self._tree[entry.parent_id][entry.name] = entry
360
361
if entry.kind == 'directory':
362
self._tree[entry.file_id] = {}
849
parent.children[entry.name] = entry
850
return entry
363
851
364
852
365
853
def add_path(self, relpath, kind, file_id=None):
366
854
"""Add entry from a path.
367
855
368
The immediate parent must already be versioned"""
856
The immediate parent must already be versioned.
857
858
Returns the new entry object."""
859
from bzrlib.branch import gen_file_id
860
369
861
parts = bzrlib.osutils.splitpath(relpath)
370
862
if len(parts) == 0:
371
bailout("cannot re-add root of inventory")
372
373
if file_id is None:
374
file_id = bzrlib.branch.gen_file_id(relpath)
375
376
parent_id = self.path2id(parts[:-1])
377
ie = InventoryEntry(file_id, parts[-1],
378
kind=kind, parent_id=parent_id)
863
raise BzrError("cannot re-add root of inventory")
864
865
if file_id == None:
866
file_id = gen_file_id(relpath)
867
868
parent_path = parts[:-1]
869
parent_id = self.path2id(parent_path)
870
if parent_id == None:
871
raise NotVersionedError(parent_path)
872
873
if kind == 'directory':
874
ie = InventoryDirectory(file_id, parts[-1], parent_id)
875
elif kind == 'file':
876
ie = InventoryFile(file_id, parts[-1], parent_id)
877
elif kind == 'symlink':
878
ie = InventoryLink(file_id, parts[-1], parent_id)
879
else:
880
raise BzrError("unknown kind %r" % kind)
379
881
return self.add(ie)
380
882
381
883
383
885
"""Remove entry by id.
384
886
385
887
>>> inv = Inventory()
386
>>> inv.add(InventoryEntry('123', 'foo.c'))
888
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
889
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
387
890
>>> '123' in inv
388
891
True
389
892
>>> del inv['123']
392
895
"""
393
896
ie = self[file_id]
394
897
395
assert self._tree[ie.parent_id][ie.name] == ie
898
assert self[ie.parent_id].children[ie.name] == ie
396
899
397
900
# TODO: Test deleting all children; maybe hoist to a separate
398
901
# deltree method?
399
902
if ie.kind == 'directory':
400
for cie in self._tree[file_id].values():
903
for cie in ie.children.values():
401
904
del self[cie.file_id]
402
del self._tree[file_id]
905
del ie.children
403
906
404
907
del self._byid[file_id]
405
del self._tree[ie.parent_id][ie.name]
406
407
408
def id_set(self):
409
return Set(self._byid)
410
411
412
def to_element(self):
413
"""Convert to XML Element"""
414
e = Element('inventory')
415
e.text = '\n'
416
for path, ie in self.iter_entries():
417
e.append(ie.to_element())
418
return e
419
420
421
def from_element(cls, elt):
422
"""Construct from XML Element
423
424
>>> inv = Inventory()
425
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c'))
426
>>> elt = inv.to_element()
427
>>> inv2 = Inventory.from_element(elt)
428
>>> inv2 == inv
429
True
430
"""
431
assert elt.tag == 'inventory'
432
o = cls()
433
for e in elt:
434
o.add(InventoryEntry.from_element(e))
435
return o
436
437
from_element = classmethod(from_element)
438
439
440
def __cmp__(self, other):
908
del self[ie.parent_id].children[ie.name]
909
910
911
def __eq__(self, other):
441
912
"""Compare two sets by comparing their contents.
442
913
443
914
>>> i1 = Inventory()
444
915
>>> i2 = Inventory()
445
916
>>> i1 == i2
446
917
True
447
>>> i1.add(InventoryEntry('123', 'foo'))
918
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
919
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
448
920
>>> i1 == i2
449
921
False
450
>>> i2.add(InventoryEntry('123', 'foo'))
922
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
923
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
451
924
>>> i1 == i2
452
925
True
453
926
"""
454
if self is other:
455
return 0
456
457
927
if not isinstance(other, Inventory):
458
928
return NotImplemented
459
929
460
if self.id_set() ^ other.id_set():
461
return 1
462
463
for file_id in self._byid:
464
c = cmp(self[file_id], other[file_id])
465
if c: return c
466
467
return 0
468
469
470
def id2path(self, file_id):
471
"""Return as a list the path to file_id."""
930
if len(self._byid) != len(other._byid):
931
# shortcut: obviously not the same
932
return False
933
934
return self._byid == other._byid
935
936
937
def __ne__(self, other):
938
return not self.__eq__(other)
939
940
941
def __hash__(self):
942
raise ValueError('not hashable')
943
944
945
def get_idpath(self, file_id):
946
"""Return a list of file_ids for the path to an entry.
947
948
The list contains one element for each directory followed by
949
the id of the file itself. So the length of the returned list
950
is equal to the depth of the file in the tree, counting the
951
root directory as depth 1.
952
"""
472
953
p = []
473
954
while file_id != None:
474
ie = self[file_id]
475
p = [ie.name] + p
955
try:
956
ie = self._byid[file_id]
957
except KeyError:
958
raise BzrError("file_id {%s} not found in inventory" % file_id)
959
p.insert(0, ie.file_id)
476
960
file_id = ie.parent_id
477
return joinpath(p)
961
return p
962
963
964
def id2path(self, file_id):
965
"""Return as a list the path to file_id.
966
967
>>> i = Inventory()
968
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
969
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
970
>>> print i.id2path('foo-id').replace(os.sep, '/')
971
src/foo.c
972
"""
973
# get all names, skipping root
974
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
975
return os.sep.join(p)
478
976
479
977
480
978
486
984
487
985
This returns the entry of the last component in the path,
488
986
which may be either a file or a directory.
987
988
Returns None iff the path is not found.
489
989
"""
490
990
if isinstance(name, types.StringTypes):
491
991
name = splitpath(name)
492
992
493
parent_id = None
993
mutter("lookup path %r" % name)
994
995
parent = self.root
494
996
for f in name:
495
997
try:
496
cie = self._tree[parent_id][f]
998
cie = parent.children[f]
497
999
assert cie.name == f
498
parent_id = cie.file_id
1000
assert cie.parent_id == parent.file_id
1001
parent = cie
499
1002
except KeyError:
500
1003
# or raise an error?
501
1004
return None
502
1005
503
return parent_id
504
505
506
def get_child(self, parent_id, child_name):
507
return self._tree[parent_id].get(child_name)
1006
return parent.file_id
508
1007
509
1008
510
1009
def has_filename(self, names):
512
1011
513
1012
514
1013
def has_id(self, file_id):
515
assert isinstance(file_id, str)
516
1014
return self._byid.has_key(file_id)
517
1015
518
1016
519
520
521
522
if __name__ == '__main__':
523
import doctest, inventory
524
doctest.testmod(inventory)
1017
def rename(self, file_id, new_parent_id, new_name):
1018
"""Move a file within the inventory.
1019
1020
This can change either the name, or the parent, or both.
1021
1022
This does not move the working file."""
1023
if not is_valid_name(new_name):
1024
raise BzrError("not an acceptable filename: %r" % new_name)
1025
1026
new_parent = self._byid[new_parent_id]
1027
if new_name in new_parent.children:
1028
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
1029
1030
new_parent_idpath = self.get_idpath(new_parent_id)
1031
if file_id in new_parent_idpath:
1032
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
1033
% (self.id2path(file_id), self.id2path(new_parent_id)))
1034
1035
file_ie = self._byid[file_id]
1036
old_parent = self._byid[file_ie.parent_id]
1037
1038
# TODO: Don't leave things messed up if this fails
1039
1040
del old_parent.children[file_ie.name]
1041
new_parent.children[new_name] = file_ie
1042
1043
file_ie.name = new_name
1044
file_ie.parent_id = new_parent_id
1045
1046
1047
1048
1049
_NAME_RE = None
1050
1051
def is_valid_name(name):
1052
global _NAME_RE
1053
if _NAME_RE == None:
1054
_NAME_RE = re.compile(r'^[^/\\]+$')
1055
1056
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1