« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 1530.2.1
- compare with another revision
- download diff
- download tarball
- view history from revision 1530.2.1
- Committer: Robert Collins
- Date: 2005-12-24 02:20:45 UTC
- mto: (1185.50.57 bzr-jam-integration)
- mto: This revision was merged to the branch mainline in revision 1550.
- Revision ID: robertc@robertcollins.net-20051224022045-14efc8dfa0e1a4e9
Start tests for api usage.
- files added:
- .rsyncexclude
- bzrlib/doc
- bzrlib/doc/api
- bzrlib/export
- bzrlib/plugins
- bzrlib/store
- bzrlib/tests
- bzrlib/tests/blackbox
- bzrlib/transport
- bzrlib/ui
- bzrlib/util
- bzrlib/util/configobj
- bzrlib/util/configobj/docs
- 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/
- bzrlib/newinventory.py => contrib/newinventory.py
- files modified:
- .bzrignore
- bzr *
- setup.py *
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
"""Inventories map files to their name in a revision."""
18
19
# TODO: Maybe store inventory_id in the file? Not really needed.
20
21
__copyright__ = "Copyright (C) 2005 Canonical Ltd."
22
__author__ = "Martin Pool <mbp@canonical.com>"
23
24
import sys, os.path, types, re
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
# FIXME: This refactoring of the workingtree code doesn't seem to keep
18
# the WorkingTree's copy of the inventory in sync with the branch. The
19
# branch modifies its working inventory when it does a commit to make
20
# missing files permanently removed.
21
22
# TODO: Maybe also keep the full path of the entry, and the children?
23
# But those depend on its position within a particular inventory, and
24
# it would be nice not to need to hold the backpointer here.
25
26
# This should really be an id randomly assigned when the tree is
27
# created, but it's not for now.
28
ROOT_ID = "TREE_ROOT"
29
30
31
import os.path
32
import re
33
import sys
34
import tarfile
35
import types
34
36
35
37
import bzrlib
36
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
38
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
39
appendpath, sha_strings)
37
40
from bzrlib.trace import mutter
38
39
class InventoryEntry(XMLMixin):
41
from bzrlib.errors import (NotVersionedError, InvalidEntryName,
42
BzrError, BzrCheckError)
43
44
45
class InventoryEntry(object):
40
46
"""Description of a versioned file.
41
47
42
48
An InventoryEntry has the following fields, which are also
43
49
present in the XML inventory-entry element:
44
50
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
51
file_id
52
53
name
54
(within the parent directory)
55
56
parent_id
57
file_id of the parent directory, or ROOT_ID
58
59
revision
60
the revision_id in which this variation of this file was
61
introduced.
62
63
executable
64
Indicates that this file should be executable on systems
65
that support it.
66
67
text_sha1
68
sha-1 of the text of the file
69
70
text_size
71
size in bytes of the text of the file
72
73
(reading a version 4 tree created a text_id field.)
59
74
60
75
>>> i = Inventory()
61
76
>>> i.path2id('')
62
>>> i.add(InventoryEntry('123', 'src', kind='directory'))
63
>>> i.add(InventoryEntry('2323', 'hello.c', parent_id='123'))
64
>>> for j in i.iter_entries():
65
... print j
77
'TREE_ROOT'
78
>>> i.add(InventoryDirectory('123', 'src', ROOT_ID))
79
InventoryDirectory('123', 'src', parent_id='TREE_ROOT')
80
>>> i.add(InventoryFile('2323', 'hello.c', parent_id='123'))
81
InventoryFile('2323', 'hello.c', parent_id='123')
82
>>> shouldbe = {0: 'src', 1: os.path.join('src','hello.c')}
83
>>> for ix, j in enumerate(i.iter_entries()):
84
... print (j[0] == shouldbe[ix], j[1])
66
85
...
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'))
86
(True, InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
87
(True, InventoryFile('2323', 'hello.c', parent_id='123'))
88
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
70
89
Traceback (most recent call last):
71
90
...
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'))
91
BzrError: inventory already contains entry with id {2323}
92
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
93
InventoryFile('2324', 'bye.c', parent_id='123')
94
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
95
InventoryDirectory('2325', 'wibble', parent_id='123')
75
96
>>> i.path2id('src/wibble')
76
97
'2325'
77
98
>>> '2325' in i
78
99
True
79
>>> i.add(InventoryEntry('2326', 'wibble.c', parent_id='2325'))
100
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
101
InventoryFile('2326', 'wibble.c', parent_id='2325')
80
102
>>> 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])
103
InventoryFile('2326', 'wibble.c', parent_id='2325')
104
>>> for path, entry in i.iter_entries():
105
... print path.replace('\\\\', '/') # for win32 os.sep
106
... assert i.path2id(path)
85
107
...
86
108
src
87
109
src/bye.c
88
110
src/hello.c
89
111
src/wibble
90
112
src/wibble/wibble.c
91
>>> i.id2path('2326')
113
>>> i.id2path('2326').replace('\\\\', '/')
92
114
'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
115
"""
98
def __init__(self, file_id, name, kind='file', text_id=None,
99
parent_id=None):
116
117
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
118
'text_id', 'parent_id', 'children', 'executable',
119
'revision']
120
121
def _add_text_to_weave(self, new_lines, parents, weave_store, transaction):
122
weave_store.add_text(self.file_id, self.revision, new_lines, parents,
123
transaction)
124
125
def detect_changes(self, old_entry):
126
"""Return a (text_modified, meta_modified) from this to old_entry.
127
128
_read_tree_state must have been called on self and old_entry prior to
129
calling detect_changes.
130
"""
131
return False, False
132
133
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
134
output_to, reverse=False):
135
"""Perform a diff from this to to_entry.
136
137
text_diff will be used for textual difference calculation.
138
This is a template method, override _diff in child classes.
139
"""
140
self._read_tree_state(tree.id2path(self.file_id), tree)
141
if to_entry:
142
# cannot diff from one kind to another - you must do a removal
143
# and an addif they do not match.
144
assert self.kind == to_entry.kind
145
to_entry._read_tree_state(to_tree.id2path(to_entry.file_id),
146
to_tree)
147
self._diff(text_diff, from_label, tree, to_label, to_entry, to_tree,
148
output_to, reverse)
149
150
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
151
output_to, reverse=False):
152
"""Perform a diff between two entries of the same kind."""
153
154
def find_previous_heads(self, previous_inventories, entry_weave):
155
"""Return the revisions and entries that directly preceed this.
156
157
Returned as a map from revision to inventory entry.
158
159
This is a map containing the file revisions in all parents
160
for which the file exists, and its revision is not a parent of
161
any other. If the file is new, the set will be empty.
162
"""
163
def get_ancestors(weave, entry):
164
return set(map(weave.idx_to_name,
165
weave.inclusions([weave.lookup(entry.revision)])))
166
heads = {}
167
head_ancestors = {}
168
for inv in previous_inventories:
169
if self.file_id in inv:
170
ie = inv[self.file_id]
171
assert ie.file_id == self.file_id
172
if ie.revision in heads:
173
# fixup logic, there was a bug in revision updates.
174
# with x bit support.
175
try:
176
if heads[ie.revision].executable != ie.executable:
177
heads[ie.revision].executable = False
178
ie.executable = False
179
except AttributeError:
180
pass
181
assert heads[ie.revision] == ie
182
else:
183
# may want to add it.
184
# may already be covered:
185
already_present = 0 != len(
186
[head for head in heads
187
if ie.revision in head_ancestors[head]])
188
if already_present:
189
# an ancestor of a known head.
190
continue
191
# definately a head:
192
ancestors = get_ancestors(entry_weave, ie)
193
# may knock something else out:
194
check_heads = list(heads.keys())
195
for head in check_heads:
196
if head in ancestors:
197
# this head is not really a head
198
heads.pop(head)
199
head_ancestors[ie.revision] = ancestors
200
heads[ie.revision] = ie
201
return heads
202
203
def get_tar_item(self, root, dp, now, tree):
204
"""Get a tarfile item and a file stream for its content."""
205
item = tarfile.TarInfo(os.path.join(root, dp))
206
# TODO: would be cool to actually set it to the timestamp of the
207
# revision it was last changed
208
item.mtime = now
209
fileobj = self._put_in_tar(item, tree)
210
return item, fileobj
211
212
def has_text(self):
213
"""Return true if the object this entry represents has textual data.
214
215
Note that textual data includes binary content.
216
217
Also note that all entries get weave files created for them.
218
This attribute is primarily used when upgrading from old trees that
219
did not have the weave index for all inventory entries.
220
"""
221
return False
222
223
def __init__(self, file_id, name, parent_id, text_id=None):
100
224
"""Create an InventoryEntry
101
225
102
226
The filename must be a single component, relative to the
103
227
parent directory; it cannot be a whole path or relative name.
104
228
105
>>> e = InventoryEntry('123', 'hello.c')
229
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
106
230
>>> e.name
107
231
'hello.c'
108
232
>>> e.file_id
109
233
'123'
110
>>> e = InventoryEntry('123', 'src/hello.c')
234
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
111
235
Traceback (most recent call last):
112
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
236
InvalidEntryName: Invalid entry name: src/hello.c
113
237
"""
114
115
if len(splitpath(name)) != 1:
116
bailout('InventoryEntry name is not a simple filename: %r'
117
% name)
118
238
assert isinstance(name, basestring), name
239
if '/' in name or '\\' in name:
240
raise InvalidEntryName(name=name)
241
self.executable = False
242
self.revision = None
243
self.text_sha1 = None
244
self.text_size = None
119
245
self.file_id = file_id
120
246
self.name = name
121
assert kind in ['file', 'directory']
122
self.kind = kind
123
247
self.text_id = text_id
124
248
self.parent_id = parent_id
125
self.text_sha1 = None
126
self.text_size = None
127
if kind == 'directory':
128
self.children = {}
129
249
self.symlink_target = None
250
251
def kind_character(self):
252
"""Return a short kind indicator useful for appending to names."""
253
raise BzrError('unknown kind %r' % self.kind)
254
255
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
256
257
def _put_in_tar(self, item, tree):
258
"""populate item for stashing in a tar, and return the content stream.
259
260
If no content is available, return None.
261
"""
262
raise BzrError("don't know how to export {%s} of kind %r" %
263
(self.file_id, self.kind))
264
265
def put_on_disk(self, dest, dp, tree):
266
"""Create a representation of self on disk in the prefix dest.
267
268
This is a template method - implement _put_on_disk in subclasses.
269
"""
270
fullpath = appendpath(dest, dp)
271
self._put_on_disk(fullpath, tree)
272
mutter(" export {%s} kind %s to %s", self.file_id,
273
self.kind, fullpath)
274
275
def _put_on_disk(self, fullpath, tree):
276
"""Put this entry onto disk at fullpath, from tree tree."""
277
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
130
278
131
279
def sorted_children(self):
132
280
l = self.children.items()
133
281
l.sort()
134
282
return l
135
283
284
@staticmethod
285
def versionable_kind(kind):
286
return kind in ('file', 'directory', 'symlink')
287
288
def check(self, checker, rev_id, inv, tree):
289
"""Check this inventory entry is intact.
290
291
This is a template method, override _check for kind specific
292
tests.
293
"""
294
if self.parent_id != None:
295
if not inv.has_id(self.parent_id):
296
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
297
% (self.parent_id, rev_id))
298
self._check(checker, rev_id, tree)
299
300
def _check(self, checker, rev_id, tree):
301
"""Check this inventory entry for kind specific errors."""
302
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
303
(self.kind, rev_id))
304
136
305
137
306
def copy(self):
138
other = InventoryEntry(self.file_id, self.name, self.kind,
139
self.text_id, self.parent_id)
140
other.text_sha1 = self.text_sha1
141
other.text_size = self.text_size
142
return other
307
"""Clone this inventory entry."""
308
raise NotImplementedError
143
309
310
def _get_snapshot_change(self, previous_entries):
311
if len(previous_entries) > 1:
312
return 'merged'
313
elif len(previous_entries) == 0:
314
return 'added'
315
else:
316
return 'modified/renamed/reparented'
144
317
145
318
def __repr__(self):
146
return ("%s(%r, %r, kind=%r, parent_id=%r)"
319
return ("%s(%r, %r, parent_id=%r)"
147
320
% (self.__class__.__name__,
148
321
self.file_id,
149
322
self.name,
150
self.kind,
151
323
self.parent_id))
152
324
153
154
def to_element(self):
155
"""Convert to XML element"""
156
e = Element('entry')
157
158
e.set('name', self.name)
159
e.set('file_id', self.file_id)
160
e.set('kind', self.kind)
161
162
if self.text_size is not None:
163
e.set('text_size', '%d' % self.text_size)
164
165
for f in ['text_id', 'text_sha1', 'parent_id']:
166
v = getattr(self, f)
167
if v is not None:
168
e.set(f, v)
169
170
e.tail = '\n'
171
172
return e
173
174
175
def from_element(cls, elt):
176
assert elt.tag == 'entry'
177
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'))
178
self.text_id = elt.get('text_id')
179
self.text_sha1 = elt.get('text_sha1')
180
self.parent_id = elt.get('parent_id')
181
182
## mutter("read inventoryentry: %r" % (elt.attrib))
183
184
v = elt.get('text_size')
185
self.text_size = v and int(v)
186
187
return self
188
189
190
from_element = classmethod(from_element)
191
192
def __cmp__(self, other):
193
if self is other:
194
return 0
325
def snapshot(self, revision, path, previous_entries,
326
work_tree, weave_store, transaction):
327
"""Make a snapshot of this entry which may or may not have changed.
328
329
This means that all its fields are populated, that it has its
330
text stored in the text store or weave.
331
"""
332
mutter('new parents of %s are %r', path, previous_entries)
333
self._read_tree_state(path, work_tree)
334
if len(previous_entries) == 1:
335
# cannot be unchanged unless there is only one parent file rev.
336
parent_ie = previous_entries.values()[0]
337
if self._unchanged(parent_ie):
338
mutter("found unchanged entry")
339
self.revision = parent_ie.revision
340
return "unchanged"
341
return self.snapshot_revision(revision, previous_entries,
342
work_tree, weave_store, transaction)
343
344
def snapshot_revision(self, revision, previous_entries, work_tree,
345
weave_store, transaction):
346
"""Record this revision unconditionally."""
347
mutter('new revision for {%s}', self.file_id)
348
self.revision = revision
349
change = self._get_snapshot_change(previous_entries)
350
self._snapshot_text(previous_entries, work_tree, weave_store,
351
transaction)
352
return change
353
354
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
355
"""Record the 'text' of this entry, whatever form that takes.
356
357
This default implementation simply adds an empty text.
358
"""
359
mutter('storing file {%s} in revision {%s}',
360
self.file_id, self.revision)
361
self._add_text_to_weave([], file_parents, weave_store, transaction)
362
363
def __eq__(self, other):
195
364
if not isinstance(other, InventoryEntry):
196
365
return NotImplemented
197
366
198
return cmp(self.file_id, other.file_id) \
199
or cmp(self.name, other.name) \
200
or cmp(self.text_sha1, other.text_sha1) \
201
or cmp(self.text_size, other.text_size) \
202
or cmp(self.text_id, other.text_id) \
203
or cmp(self.parent_id, other.parent_id) \
204
or cmp(self.kind, other.kind)
205
367
return ((self.file_id == other.file_id)
368
and (self.name == other.name)
369
and (other.symlink_target == self.symlink_target)
370
and (self.text_sha1 == other.text_sha1)
371
and (self.text_size == other.text_size)
372
and (self.text_id == other.text_id)
373
and (self.parent_id == other.parent_id)
374
and (self.kind == other.kind)
375
and (self.revision == other.revision)
376
and (self.executable == other.executable)
377
)
378
379
def __ne__(self, other):
380
return not (self == other)
381
382
def __hash__(self):
383
raise ValueError('not hashable')
384
385
def _unchanged(self, previous_ie):
386
"""Has this entry changed relative to previous_ie.
387
388
This method should be overriden in child classes.
389
"""
390
compatible = True
391
# different inv parent
392
if previous_ie.parent_id != self.parent_id:
393
compatible = False
394
# renamed
395
elif previous_ie.name != self.name:
396
compatible = False
397
return compatible
398
399
def _read_tree_state(self, path, work_tree):
400
"""Populate fields in the inventory entry from the given tree.
401
402
Note that this should be modified to be a noop on virtual trees
403
as all entries created there are prepopulated.
404
"""
405
# TODO: Rather than running this manually, we should check the
406
# working sha1 and other expensive properties when they're
407
# first requested, or preload them if they're already known
408
pass # nothing to do by default
206
409
207
410
208
411
class RootEntry(InventoryEntry):
412
413
def _check(self, checker, rev_id, tree):
414
"""See InventoryEntry._check"""
415
209
416
def __init__(self, file_id):
210
417
self.file_id = file_id
211
418
self.children = {}
212
419
self.kind = 'root_directory'
213
420
self.parent_id = None
214
self.name = ''
421
self.name = u''
215
422
216
def __cmp__(self, other):
217
if self is other:
218
return 0
423
def __eq__(self, other):
219
424
if not isinstance(other, RootEntry):
220
425
return NotImplemented
221
return cmp(self.file_id, other.file_id) \
222
or cmp(self.children, other.children)
223
224
225
226
class Inventory(XMLMixin):
426
427
return (self.file_id == other.file_id) \
428
and (self.children == other.children)
429
430
431
class InventoryDirectory(InventoryEntry):
432
"""A directory in an inventory."""
433
434
def _check(self, checker, rev_id, tree):
435
"""See InventoryEntry._check"""
436
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
437
raise BzrCheckError('directory {%s} has text in revision {%s}'
438
% (self.file_id, rev_id))
439
440
def copy(self):
441
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
442
other.revision = self.revision
443
# note that children are *not* copied; they're pulled across when
444
# others are added
445
return other
446
447
def __init__(self, file_id, name, parent_id):
448
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
449
self.children = {}
450
self.kind = 'directory'
451
452
def kind_character(self):
453
"""See InventoryEntry.kind_character."""
454
return '/'
455
456
def _put_in_tar(self, item, tree):
457
"""See InventoryEntry._put_in_tar."""
458
item.type = tarfile.DIRTYPE
459
fileobj = None
460
item.name += '/'
461
item.size = 0
462
item.mode = 0755
463
return fileobj
464
465
def _put_on_disk(self, fullpath, tree):
466
"""See InventoryEntry._put_on_disk."""
467
os.mkdir(fullpath)
468
469
470
class InventoryFile(InventoryEntry):
471
"""A file in an inventory."""
472
473
def _check(self, checker, rev_id, tree):
474
"""See InventoryEntry._check"""
475
revision = self.revision
476
t = (self.file_id, revision)
477
if t in checker.checked_texts:
478
prev_sha = checker.checked_texts[t]
479
if prev_sha != self.text_sha1:
480
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
481
(self.file_id, rev_id))
482
else:
483
checker.repeated_text_cnt += 1
484
return
485
mutter('check version {%s} of {%s}', rev_id, self.file_id)
486
file_lines = tree.get_file_lines(self.file_id)
487
checker.checked_text_cnt += 1
488
if self.text_size != sum(map(len, file_lines)):
489
raise BzrCheckError('text {%s} wrong size' % self.text_id)
490
if self.text_sha1 != sha_strings(file_lines):
491
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
492
checker.checked_texts[t] = self.text_sha1
493
494
def copy(self):
495
other = InventoryFile(self.file_id, self.name, self.parent_id)
496
other.executable = self.executable
497
other.text_id = self.text_id
498
other.text_sha1 = self.text_sha1
499
other.text_size = self.text_size
500
other.revision = self.revision
501
return other
502
503
def detect_changes(self, old_entry):
504
"""See InventoryEntry.detect_changes."""
505
assert self.text_sha1 != None
506
assert old_entry.text_sha1 != None
507
text_modified = (self.text_sha1 != old_entry.text_sha1)
508
meta_modified = (self.executable != old_entry.executable)
509
return text_modified, meta_modified
510
511
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
512
output_to, reverse=False):
513
"""See InventoryEntry._diff."""
514
from_text = tree.get_file(self.file_id).readlines()
515
if to_entry:
516
to_text = to_tree.get_file(to_entry.file_id).readlines()
517
else:
518
to_text = []
519
if not reverse:
520
text_diff(from_label, from_text,
521
to_label, to_text, output_to)
522
else:
523
text_diff(to_label, to_text,
524
from_label, from_text, output_to)
525
526
def has_text(self):
527
"""See InventoryEntry.has_text."""
528
return True
529
530
def __init__(self, file_id, name, parent_id):
531
super(InventoryFile, self).__init__(file_id, name, parent_id)
532
self.kind = 'file'
533
534
def kind_character(self):
535
"""See InventoryEntry.kind_character."""
536
return ''
537
538
def _put_in_tar(self, item, tree):
539
"""See InventoryEntry._put_in_tar."""
540
item.type = tarfile.REGTYPE
541
fileobj = tree.get_file(self.file_id)
542
item.size = self.text_size
543
if tree.is_executable(self.file_id):
544
item.mode = 0755
545
else:
546
item.mode = 0644
547
return fileobj
548
549
def _put_on_disk(self, fullpath, tree):
550
"""See InventoryEntry._put_on_disk."""
551
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
552
if tree.is_executable(self.file_id):
553
os.chmod(fullpath, 0755)
554
555
def _read_tree_state(self, path, work_tree):
556
"""See InventoryEntry._read_tree_state."""
557
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
558
self.executable = work_tree.is_executable(self.file_id)
559
560
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
561
"""See InventoryEntry._snapshot_text."""
562
mutter('storing file {%s} in revision {%s}',
563
self.file_id, self.revision)
564
# special case to avoid diffing on renames or
565
# reparenting
566
if (len(file_parents) == 1
567
and self.text_sha1 == file_parents.values()[0].text_sha1
568
and self.text_size == file_parents.values()[0].text_size):
569
previous_ie = file_parents.values()[0]
570
weave_store.add_identical_text(
571
self.file_id, previous_ie.revision,
572
self.revision, file_parents, transaction)
573
else:
574
new_lines = work_tree.get_file(self.file_id).readlines()
575
self._add_text_to_weave(new_lines, file_parents, weave_store,
576
transaction)
577
self.text_sha1 = sha_strings(new_lines)
578
self.text_size = sum(map(len, new_lines))
579
580
581
def _unchanged(self, previous_ie):
582
"""See InventoryEntry._unchanged."""
583
compatible = super(InventoryFile, self)._unchanged(previous_ie)
584
if self.text_sha1 != previous_ie.text_sha1:
585
compatible = False
586
else:
587
# FIXME: 20050930 probe for the text size when getting sha1
588
# in _read_tree_state
589
self.text_size = previous_ie.text_size
590
if self.executable != previous_ie.executable:
591
compatible = False
592
return compatible
593
594
595
class InventoryLink(InventoryEntry):
596
"""A file in an inventory."""
597
598
__slots__ = ['symlink_target']
599
600
def _check(self, checker, rev_id, tree):
601
"""See InventoryEntry._check"""
602
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
603
raise BzrCheckError('symlink {%s} has text in revision {%s}'
604
% (self.file_id, rev_id))
605
if self.symlink_target == None:
606
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
607
% (self.file_id, rev_id))
608
609
def copy(self):
610
other = InventoryLink(self.file_id, self.name, self.parent_id)
611
other.symlink_target = self.symlink_target
612
other.revision = self.revision
613
return other
614
615
def detect_changes(self, old_entry):
616
"""See InventoryEntry.detect_changes."""
617
# FIXME: which _modified field should we use ? RBC 20051003
618
text_modified = (self.symlink_target != old_entry.symlink_target)
619
if text_modified:
620
mutter(" symlink target changed")
621
meta_modified = False
622
return text_modified, meta_modified
623
624
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
625
output_to, reverse=False):
626
"""See InventoryEntry._diff."""
627
from_text = self.symlink_target
628
if to_entry is not None:
629
to_text = to_entry.symlink_target
630
if reverse:
631
temp = from_text
632
from_text = to_text
633
to_text = temp
634
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
635
else:
636
if not reverse:
637
print >>output_to, '=== target was %r' % self.symlink_target
638
else:
639
print >>output_to, '=== target is %r' % self.symlink_target
640
641
def __init__(self, file_id, name, parent_id):
642
super(InventoryLink, self).__init__(file_id, name, parent_id)
643
self.kind = 'symlink'
644
645
def kind_character(self):
646
"""See InventoryEntry.kind_character."""
647
return ''
648
649
def _put_in_tar(self, item, tree):
650
"""See InventoryEntry._put_in_tar."""
651
item.type = tarfile.SYMTYPE
652
fileobj = None
653
item.size = 0
654
item.mode = 0755
655
item.linkname = self.symlink_target
656
return fileobj
657
658
def _put_on_disk(self, fullpath, tree):
659
"""See InventoryEntry._put_on_disk."""
660
try:
661
os.symlink(self.symlink_target, fullpath)
662
except OSError,e:
663
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
664
665
def _read_tree_state(self, path, work_tree):
666
"""See InventoryEntry._read_tree_state."""
667
self.symlink_target = work_tree.get_symlink_target(self.file_id)
668
669
def _unchanged(self, previous_ie):
670
"""See InventoryEntry._unchanged."""
671
compatible = super(InventoryLink, self)._unchanged(previous_ie)
672
if self.symlink_target != previous_ie.symlink_target:
673
compatible = False
674
return compatible
675
676
677
class Inventory(object):
227
678
"""Inventory of versioned files in a tree.
228
679
229
An Inventory acts like a set of InventoryEntry items. You can
230
also look files up by their file_id or name.
231
232
May be read from and written to a metadata file in a tree. To
233
manipulate the inventory (for example to add a file), it is read
234
in, modified, and then written back out.
680
This describes which file_id is present at each point in the tree,
681
and possibly the SHA-1 or other information about the file.
682
Entries can be looked up either by path or by file_id.
235
683
236
684
The inventory represents a typical unix file tree, with
237
685
directories containing files and subdirectories. We never store
244
692
inserted, other than through the Inventory API.
245
693
246
694
>>> inv = Inventory()
247
>>> inv.write_xml(sys.stdout)
248
<inventory>
249
</inventory>
250
>>> inv.add(InventoryEntry('123-123', 'hello.c'))
695
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
696
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
251
697
>>> inv['123-123'].name
252
698
'hello.c'
253
699
262
708
263
709
>>> [x[0] for x in inv.iter_entries()]
264
710
['hello.c']
265
266
>>> inv.write_xml(sys.stdout)
267
<inventory>
268
<entry file_id="123-123" kind="file" name="hello.c" />
269
</inventory>
270
711
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
712
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
713
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
271
714
"""
272
273
## TODO: Make sure only canonical filenames are stored.
274
275
## TODO: Do something sensible about the possible collisions on
276
## case-losing filesystems. Perhaps we should just always forbid
277
## such collisions.
278
279
## TODO: No special cases for root, rather just give it a file id
280
## like everything else.
281
282
## TODO: Probably change XML serialization to use nesting
283
284
def __init__(self):
715
def __init__(self, root_id=ROOT_ID):
285
716
"""Create or read an inventory.
286
717
287
718
If a working directory is specified, the inventory is read
291
722
The inventory is created with a default root directory, with
292
723
an id of None.
293
724
"""
294
self.root = RootEntry(None)
295
self._byid = {None: self.root}
725
# We are letting Branch.initialize() create a unique inventory
726
# root id. Rather than generating a random one here.
727
#if root_id is None:
728
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
729
self.root = RootEntry(root_id)
730
self._byid = {self.root.file_id: self.root}
731
732
733
def copy(self):
734
other = Inventory(self.root.file_id)
735
# copy recursively so we know directories will be added before
736
# their children. There are more efficient ways than this...
737
for path, entry in self.iter_entries():
738
if entry == self.root:
739
continue
740
other.add(entry.copy())
741
return other
296
742
297
743
298
744
def __iter__(self):
318
764
yield name, ie
319
765
if ie.kind == 'directory':
320
766
for cn, cie in self.iter_entries(from_dir=ie.file_id):
321
yield '/'.join((name, cn)), cie
322
323
324
325
def directories(self, from_dir=None):
326
"""Return (path, entry) pairs for all directories.
767
yield os.path.join(name, cn), cie
768
769
770
def entries(self):
771
"""Return list of (path, ie) for all entries except the root.
772
773
This may be faster than iter_entries.
327
774
"""
328
def descend(parent_ie):
329
parent_name = parent_ie.name
330
yield parent_name, parent_ie
331
332
# directory children in sorted order
333
dn = []
334
for ie in parent_ie.children.itervalues():
775
accum = []
776
def descend(dir_ie, dir_path):
777
kids = dir_ie.children.items()
778
kids.sort()
779
for name, ie in kids:
780
child_path = os.path.join(dir_path, name)
781
accum.append((child_path, ie))
335
782
if ie.kind == 'directory':
336
dn.append((ie.name, ie))
337
dn.sort()
783
descend(ie, child_path)
784
785
descend(self.root, u'')
786
return accum
787
788
789
def directories(self):
790
"""Return (path, entry) pairs for all directories, including the root.
791
"""
792
accum = []
793
def descend(parent_ie, parent_path):
794
accum.append((parent_path, parent_ie))
338
795
339
for name, child_ie in dn:
340
for sub_name, sub_ie in descend(child_ie):
341
yield appendpath(parent_name, sub_name), sub_ie
796
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
797
kids.sort()
342
798
343
for name, ie in descend(self.root):
344
yield name, ie
799
for name, child_ie in kids:
800
child_path = os.path.join(parent_path, name)
801
descend(child_ie, child_path)
802
descend(self.root, u'')
803
return accum
345
804
346
805
347
806
349
808
"""True if this entry contains a file with given id.
350
809
351
810
>>> inv = Inventory()
352
>>> inv.add(InventoryEntry('123', 'foo.c'))
811
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
812
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
353
813
>>> '123' in inv
354
814
True
355
815
>>> '456' in inv
362
822
"""Return the entry for given file_id.
363
823
364
824
>>> inv = Inventory()
365
>>> inv.add(InventoryEntry('123123', 'hello.c'))
825
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
826
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
366
827
>>> inv['123123'].name
367
828
'hello.c'
368
829
"""
369
return self._byid[file_id]
370
830
try:
831
return self._byid[file_id]
832
except KeyError:
833
if file_id == None:
834
raise BzrError("can't look up file_id None")
835
else:
836
raise BzrError("file_id {%s} not in inventory" % file_id)
837
838
839
def get_file_kind(self, file_id):
840
return self._byid[file_id].kind
371
841
372
842
def get_child(self, parent_id, filename):
373
843
return self[parent_id].children.get(filename)
377
847
"""Add entry to inventory.
378
848
379
849
To add a file to a branch ready to be committed, use Branch.add,
380
which calls this."""
850
which calls this.
851
852
Returns the new entry object.
853
"""
381
854
if entry.file_id in self._byid:
382
bailout("inventory already contains entry with id {%s}" % entry.file_id)
855
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
856
857
if entry.parent_id == ROOT_ID or entry.parent_id is None:
858
entry.parent_id = self.root.file_id
383
859
384
860
try:
385
861
parent = self._byid[entry.parent_id]
386
862
except KeyError:
387
bailout("parent_id %r not in inventory" % entry.parent_id)
863
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
388
864
389
865
if parent.children.has_key(entry.name):
390
bailout("%s is already versioned" %
866
raise BzrError("%s is already versioned" %
391
867
appendpath(self.id2path(parent.file_id), entry.name))
392
868
393
869
self._byid[entry.file_id] = entry
394
870
parent.children[entry.name] = entry
871
return entry
395
872
396
873
397
874
def add_path(self, relpath, kind, file_id=None):
398
875
"""Add entry from a path.
399
876
400
The immediate parent must already be versioned"""
877
The immediate parent must already be versioned.
878
879
Returns the new entry object."""
880
from bzrlib.workingtree import gen_file_id
881
401
882
parts = bzrlib.osutils.splitpath(relpath)
402
883
if len(parts) == 0:
403
bailout("cannot re-add root of inventory")
404
405
if file_id is None:
406
file_id = bzrlib.branch.gen_file_id(relpath)
407
408
parent_id = self.path2id(parts[:-1])
409
ie = InventoryEntry(file_id, parts[-1],
410
kind=kind, parent_id=parent_id)
884
raise BzrError("cannot re-add root of inventory")
885
886
if file_id == None:
887
file_id = gen_file_id(relpath)
888
889
parent_path = parts[:-1]
890
parent_id = self.path2id(parent_path)
891
if parent_id == None:
892
raise NotVersionedError(path=parent_path)
893
if kind == 'directory':
894
ie = InventoryDirectory(file_id, parts[-1], parent_id)
895
elif kind == 'file':
896
ie = InventoryFile(file_id, parts[-1], parent_id)
897
elif kind == 'symlink':
898
ie = InventoryLink(file_id, parts[-1], parent_id)
899
else:
900
raise BzrError("unknown kind %r" % kind)
411
901
return self.add(ie)
412
902
413
903
415
905
"""Remove entry by id.
416
906
417
907
>>> inv = Inventory()
418
>>> inv.add(InventoryEntry('123', 'foo.c'))
908
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
909
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
419
910
>>> '123' in inv
420
911
True
421
912
>>> del inv['123']
437
928
del self[ie.parent_id].children[ie.name]
438
929
439
930
440
def id_set(self):
441
return Set(self._byid)
442
443
444
def to_element(self):
445
"""Convert to XML Element"""
446
e = Element('inventory')
447
e.text = '\n'
448
for path, ie in self.iter_entries():
449
e.append(ie.to_element())
450
return e
451
452
453
def from_element(cls, elt):
454
"""Construct from XML Element
455
456
>>> inv = Inventory()
457
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c'))
458
>>> elt = inv.to_element()
459
>>> inv2 = Inventory.from_element(elt)
460
>>> inv2 == inv
461
True
462
"""
463
assert elt.tag == 'inventory'
464
o = cls()
465
for e in elt:
466
o.add(InventoryEntry.from_element(e))
467
return o
468
469
from_element = classmethod(from_element)
470
471
472
def __cmp__(self, other):
931
def __eq__(self, other):
473
932
"""Compare two sets by comparing their contents.
474
933
475
934
>>> i1 = Inventory()
476
935
>>> i2 = Inventory()
477
936
>>> i1 == i2
478
937
True
479
>>> i1.add(InventoryEntry('123', 'foo'))
938
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
939
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
480
940
>>> i1 == i2
481
941
False
482
>>> i2.add(InventoryEntry('123', 'foo'))
942
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
943
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
483
944
>>> i1 == i2
484
945
True
485
946
"""
486
if self is other:
487
return 0
488
489
947
if not isinstance(other, Inventory):
490
948
return NotImplemented
491
949
492
if self.id_set() ^ other.id_set():
493
return 1
494
495
for file_id in self._byid:
496
c = cmp(self[file_id], other[file_id])
497
if c: return c
498
499
return 0
500
501
502
def id2path(self, file_id):
503
"""Return as a list the path to file_id."""
950
if len(self._byid) != len(other._byid):
951
# shortcut: obviously not the same
952
return False
953
954
return self._byid == other._byid
955
956
957
def __ne__(self, other):
958
return not self.__eq__(other)
959
960
961
def __hash__(self):
962
raise ValueError('not hashable')
963
964
965
def get_idpath(self, file_id):
966
"""Return a list of file_ids for the path to an entry.
967
968
The list contains one element for each directory followed by
969
the id of the file itself. So the length of the returned list
970
is equal to the depth of the file in the tree, counting the
971
root directory as depth 1.
972
"""
504
973
p = []
505
974
while file_id != None:
506
ie = self._byid[file_id]
507
p.insert(0, ie.name)
975
try:
976
ie = self._byid[file_id]
977
except KeyError:
978
raise BzrError("file_id {%s} not found in inventory" % file_id)
979
p.insert(0, ie.file_id)
508
980
file_id = ie.parent_id
509
return '/'.join(p)
981
return p
982
983
984
def id2path(self, file_id):
985
"""Return as a list the path to file_id.
986
987
>>> i = Inventory()
988
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
989
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
990
>>> print i.id2path('foo-id').replace(os.sep, '/')
991
src/foo.c
992
"""
993
# get all names, skipping root
994
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
995
return os.sep.join(p)
510
996
511
997
512
998
518
1004
519
1005
This returns the entry of the last component in the path,
520
1006
which may be either a file or a directory.
1007
1008
Returns None iff the path is not found.
521
1009
"""
522
1010
if isinstance(name, types.StringTypes):
523
1011
name = splitpath(name)
524
1012
525
parent = self[None]
1013
mutter("lookup path %r" % name)
1014
1015
parent = self.root
526
1016
for f in name:
527
1017
try:
528
1018
cie = parent.children[f]
529
1019
assert cie.name == f
1020
assert cie.parent_id == parent.file_id
530
1021
parent = cie
531
1022
except KeyError:
532
1023
# or raise an error?
550
1041
551
1042
This does not move the working file."""
552
1043
if not is_valid_name(new_name):
553
bailout("not an acceptable filename: %r" % new_name)
1044
raise BzrError("not an acceptable filename: %r" % new_name)
554
1045
555
1046
new_parent = self._byid[new_parent_id]
556
1047
if new_name in new_parent.children:
557
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
1048
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
1049
1050
new_parent_idpath = self.get_idpath(new_parent_id)
1051
if file_id in new_parent_idpath:
1052
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
1053
% (self.id2path(file_id), self.id2path(new_parent_id)))
558
1054
559
1055
file_ie = self._byid[file_id]
560
1056
old_parent = self._byid[file_ie.parent_id]
570
1066
571
1067
572
1068
573
_NAME_RE = re.compile(r'^[^/\\]+$')
1069
_NAME_RE = None
574
1070
575
1071
def is_valid_name(name):
1072
global _NAME_RE
1073
if _NAME_RE == None:
1074
_NAME_RE = re.compile(r'^[^/\\]+$')
1075
576
1076
return bool(_NAME_RE.match(name))
577
578
579
580
if __name__ == '__main__':
581
import doctest, inventory
582
doctest.testmod(inventory)
Loggerhead is a web-based interface for Breezy
Version: 2.0.1