« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 1077
- compare with another revision
- download diff
- download tarball
- view history from revision 1077
- Committer: Martin Pool
- Date: 2005-08-18 05:52:29 UTC
- Revision ID: mbp@sourcefrog.net-20050818055229-cac46ebce364d04c
- avoid compiling REs at module load time
- files added:
- bzrlib/mdiff.py
- files removed:
- INSTALL
- bzrlib/store
- bzrlib/transport
- bzrlib/util/configobj
- bzrlib/util/configobj/docs
- files renamed:
- bzr_man.py => bzr-man.py
- contrib/newinventory.py => bzrlib/newinventory.py
- bzrlib/selftest/testplugins.py => bzrlib/selftest/plugins.py
- bzrlib/store/__init__.py => bzrlib/store.py
- bzrlib/selftest/test_weave.py => tools/testweave.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
# 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
17
26
18
# This should really be an id randomly assigned when the tree is
27
19
# created, but it's not for now.
28
20
ROOT_ID = "TREE_ROOT"
29
21
30
22
31
import os.path
32
import re
33
import sys
34
import tarfile
35
import types
23
import sys, os.path, types, re
36
24
37
25
import bzrlib
38
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
39
appendpath, sha_strings)
26
from bzrlib.errors import BzrError, BzrCheckError
27
28
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
40
29
from bzrlib.trace import mutter
41
from bzrlib.errors import (NotVersionedError, InvalidEntryName,
42
BzrError, BzrCheckError)
43
30
from bzrlib.errors import NotVersionedError
31
44
32
45
33
class InventoryEntry(object):
46
34
"""Description of a versioned file.
48
36
An InventoryEntry has the following fields, which are also
49
37
present in the XML inventory-entry element:
50
38
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.)
39
* *file_id*
40
* *name*: (only the basename within the directory, must not
41
contain slashes)
42
* *kind*: "directory" or "file"
43
* *directory_id*: (if absent/null means the branch root directory)
44
* *text_sha1*: only for files
45
* *text_size*: in bytes, only for files
46
* *text_id*: identifier for the text version, only for files
47
48
InventoryEntries can also exist inside a WorkingTree
49
inventory, in which case they are not yet bound to a
50
particular revision of the file. In that case the text_sha1,
51
text_size and text_id are absent.
52
74
53
75
54
>>> i = Inventory()
76
55
>>> i.path2id('')
77
56
'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])
57
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
58
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
59
>>> for j in i.iter_entries():
60
... print j
85
61
...
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'))
62
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
63
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
64
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
89
65
Traceback (most recent call last):
90
66
...
91
67
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')
68
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
69
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
96
70
>>> i.path2id('src/wibble')
97
71
'2325'
98
72
>>> '2325' in i
99
73
True
100
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
101
InventoryFile('2326', 'wibble.c', parent_id='2325')
74
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
102
75
>>> i['2326']
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)
76
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
77
>>> for j in i.iter_entries():
78
... print j[0]
79
... assert i.path2id(j[0])
107
80
...
108
81
src
109
82
src/bye.c
110
83
src/hello.c
111
84
src/wibble
112
85
src/wibble/wibble.c
113
>>> i.id2path('2326').replace('\\\\', '/')
86
>>> i.id2path('2326')
114
87
'src/wibble/wibble.c'
88
89
TODO: Maybe also keep the full path of the entry, and the children?
90
But those depend on its position within a particular inventory, and
91
it would be nice not to need to hold the backpointer here.
115
92
"""
116
93
94
# TODO: split InventoryEntry into subclasses for files,
95
# directories, etc etc.
96
117
97
__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):
98
'text_id', 'parent_id', 'children', ]
99
100
def __init__(self, file_id, name, kind, parent_id, text_id=None):
224
101
"""Create an InventoryEntry
225
102
226
103
The filename must be a single component, relative to the
227
104
parent directory; it cannot be a whole path or relative name.
228
105
229
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
106
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
230
107
>>> e.name
231
108
'hello.c'
232
109
>>> e.file_id
233
110
'123'
234
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
111
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
235
112
Traceback (most recent call last):
236
InvalidEntryName: Invalid entry name: src/hello.c
113
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
237
114
"""
238
assert isinstance(name, basestring), name
239
115
if '/' in name or '\\' in name:
240
raise InvalidEntryName(name=name)
241
self.executable = False
242
self.revision = None
116
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
117
243
118
self.text_sha1 = None
244
119
self.text_size = None
120
245
121
self.file_id = file_id
246
122
self.name = name
123
self.kind = kind
247
124
self.text_id = text_id
248
125
self.parent_id = parent_id
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, self.kind, fullpath))
273
274
def _put_on_disk(self, fullpath, tree):
275
"""Put this entry onto disk at fullpath, from tree tree."""
276
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
126
if kind == 'directory':
127
self.children = {}
128
elif kind == 'file':
129
pass
130
else:
131
raise BzrError("unhandled entry kind %r" % kind)
132
133
277
134
278
135
def sorted_children(self):
279
136
l = self.children.items()
280
137
l.sort()
281
138
return l
282
139
283
@staticmethod
284
def versionable_kind(kind):
285
return kind in ('file', 'directory', 'symlink')
286
287
def check(self, checker, rev_id, inv, tree):
288
"""Check this inventory entry is intact.
289
290
This is a template method, override _check for kind specific
291
tests.
292
"""
293
if self.parent_id != None:
294
if not inv.has_id(self.parent_id):
295
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
296
% (self.parent_id, rev_id))
297
self._check(checker, rev_id, tree)
298
299
def _check(self, checker, rev_id, tree):
300
"""Check this inventory entry for kind specific errors."""
301
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
302
(self.kind, rev_id))
303
304
140
305
141
def copy(self):
306
"""Clone this inventory entry."""
307
raise NotImplementedError
142
other = InventoryEntry(self.file_id, self.name, self.kind,
143
self.parent_id, text_id=self.text_id)
144
other.text_sha1 = self.text_sha1
145
other.text_size = self.text_size
146
# note that children are *not* copied; they're pulled across when
147
# others are added
148
return other
308
149
309
def _get_snapshot_change(self, previous_entries):
310
if len(previous_entries) > 1:
311
return 'merged'
312
elif len(previous_entries) == 0:
313
return 'added'
314
else:
315
return 'modified/renamed/reparented'
316
150
317
151
def __repr__(self):
318
return ("%s(%r, %r, parent_id=%r)"
152
return ("%s(%r, %r, kind=%r, parent_id=%r)"
319
153
% (self.__class__.__name__,
320
154
self.file_id,
321
155
self.name,
156
self.kind,
322
157
self.parent_id))
323
158
324
def snapshot(self, revision, path, previous_entries,
325
work_tree, weave_store, transaction):
326
"""Make a snapshot of this entry which may or may not have changed.
327
328
This means that all its fields are populated, that it has its
329
text stored in the text store or weave.
330
"""
331
mutter('new parents of %s are %r', path, previous_entries)
332
self._read_tree_state(path, work_tree)
333
if len(previous_entries) == 1:
334
# cannot be unchanged unless there is only one parent file rev.
335
parent_ie = previous_entries.values()[0]
336
if self._unchanged(parent_ie):
337
mutter("found unchanged entry")
338
self.revision = parent_ie.revision
339
return "unchanged"
340
return self.snapshot_revision(revision, previous_entries,
341
work_tree, weave_store, transaction)
342
343
def snapshot_revision(self, revision, previous_entries, work_tree,
344
weave_store, transaction):
345
"""Record this revision unconditionally."""
346
mutter('new revision for {%s}', self.file_id)
347
self.revision = revision
348
change = self._get_snapshot_change(previous_entries)
349
self._snapshot_text(previous_entries, work_tree, weave_store,
350
transaction)
351
return change
352
353
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
354
"""Record the 'text' of this entry, whatever form that takes.
355
356
This default implementation simply adds an empty text.
357
"""
358
mutter('storing file {%s} in revision {%s}',
359
self.file_id, self.revision)
360
self._add_text_to_weave([], file_parents, weave_store, transaction)
159
160
def to_element(self):
161
"""Convert to XML element"""
162
from bzrlib.xml import Element
163
164
e = Element('entry')
165
166
e.set('name', self.name)
167
e.set('file_id', self.file_id)
168
e.set('kind', self.kind)
169
170
if self.text_size != None:
171
e.set('text_size', '%d' % self.text_size)
172
173
for f in ['text_id', 'text_sha1']:
174
v = getattr(self, f)
175
if v != None:
176
e.set(f, v)
177
178
# to be conservative, we don't externalize the root pointers
179
# for now, leaving them as null in the xml form. in a future
180
# version it will be implied by nested elements.
181
if self.parent_id != ROOT_ID:
182
assert isinstance(self.parent_id, basestring)
183
e.set('parent_id', self.parent_id)
184
185
e.tail = '\n'
186
187
return e
188
189
190
def from_element(cls, elt):
191
assert elt.tag == 'entry'
192
193
## original format inventories don't have a parent_id for
194
## nodes in the root directory, but it's cleaner to use one
195
## internally.
196
parent_id = elt.get('parent_id')
197
if parent_id == None:
198
parent_id = ROOT_ID
199
200
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
201
self.text_id = elt.get('text_id')
202
self.text_sha1 = elt.get('text_sha1')
203
204
## mutter("read inventoryentry: %r" % (elt.attrib))
205
206
v = elt.get('text_size')
207
self.text_size = v and int(v)
208
209
return self
210
211
212
from_element = classmethod(from_element)
361
213
362
214
def __eq__(self, other):
363
215
if not isinstance(other, InventoryEntry):
364
216
return NotImplemented
365
217
366
return ((self.file_id == other.file_id)
367
and (self.name == other.name)
368
and (other.symlink_target == self.symlink_target)
369
and (self.text_sha1 == other.text_sha1)
370
and (self.text_size == other.text_size)
371
and (self.text_id == other.text_id)
372
and (self.parent_id == other.parent_id)
373
and (self.kind == other.kind)
374
and (self.revision == other.revision)
375
and (self.executable == other.executable)
376
)
218
return (self.file_id == other.file_id) \
219
and (self.name == other.name) \
220
and (self.text_sha1 == other.text_sha1) \
221
and (self.text_size == other.text_size) \
222
and (self.text_id == other.text_id) \
223
and (self.parent_id == other.parent_id) \
224
and (self.kind == other.kind)
225
377
226
378
227
def __ne__(self, other):
379
228
return not (self == other)
381
230
def __hash__(self):
382
231
raise ValueError('not hashable')
383
232
384
def _unchanged(self, previous_ie):
385
"""Has this entry changed relative to previous_ie.
386
387
This method should be overriden in child classes.
388
"""
389
compatible = True
390
# different inv parent
391
if previous_ie.parent_id != self.parent_id:
392
compatible = False
393
# renamed
394
elif previous_ie.name != self.name:
395
compatible = False
396
return compatible
397
398
def _read_tree_state(self, path, work_tree):
399
"""Populate fields in the inventory entry from the given tree.
400
401
Note that this should be modified to be a noop on virtual trees
402
as all entries created there are prepopulated.
403
"""
404
# TODO: Rather than running this manually, we should check the
405
# working sha1 and other expensive properties when they're
406
# first requested, or preload them if they're already known
407
pass # nothing to do by default
408
233
409
234
410
235
class RootEntry(InventoryEntry):
411
412
def _check(self, checker, rev_id, tree):
413
"""See InventoryEntry._check"""
414
415
236
def __init__(self, file_id):
416
237
self.file_id = file_id
417
238
self.children = {}
427
248
and (self.children == other.children)
428
249
429
250
430
class InventoryDirectory(InventoryEntry):
431
"""A directory in an inventory."""
432
433
def _check(self, checker, rev_id, tree):
434
"""See InventoryEntry._check"""
435
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
436
raise BzrCheckError('directory {%s} has text in revision {%s}'
437
% (self.file_id, rev_id))
438
439
def copy(self):
440
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
441
other.revision = self.revision
442
# note that children are *not* copied; they're pulled across when
443
# others are added
444
return other
445
446
def __init__(self, file_id, name, parent_id):
447
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
448
self.children = {}
449
self.kind = 'directory'
450
451
def kind_character(self):
452
"""See InventoryEntry.kind_character."""
453
return '/'
454
455
def _put_in_tar(self, item, tree):
456
"""See InventoryEntry._put_in_tar."""
457
item.type = tarfile.DIRTYPE
458
fileobj = None
459
item.name += '/'
460
item.size = 0
461
item.mode = 0755
462
return fileobj
463
464
def _put_on_disk(self, fullpath, tree):
465
"""See InventoryEntry._put_on_disk."""
466
os.mkdir(fullpath)
467
468
469
class InventoryFile(InventoryEntry):
470
"""A file in an inventory."""
471
472
def _check(self, checker, rev_id, tree):
473
"""See InventoryEntry._check"""
474
revision = self.revision
475
t = (self.file_id, revision)
476
if t in checker.checked_texts:
477
prev_sha = checker.checked_texts[t]
478
if prev_sha != self.text_sha1:
479
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
480
(self.file_id, rev_id))
481
else:
482
checker.repeated_text_cnt += 1
483
return
484
mutter('check version {%s} of {%s}', rev_id, self.file_id)
485
file_lines = tree.get_file_lines(self.file_id)
486
checker.checked_text_cnt += 1
487
if self.text_size != sum(map(len, file_lines)):
488
raise BzrCheckError('text {%s} wrong size' % self.text_id)
489
if self.text_sha1 != sha_strings(file_lines):
490
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
491
checker.checked_texts[t] = self.text_sha1
492
493
def copy(self):
494
other = InventoryFile(self.file_id, self.name, self.parent_id)
495
other.executable = self.executable
496
other.text_id = self.text_id
497
other.text_sha1 = self.text_sha1
498
other.text_size = self.text_size
499
other.revision = self.revision
500
return other
501
502
def detect_changes(self, old_entry):
503
"""See InventoryEntry.detect_changes."""
504
assert self.text_sha1 != None
505
assert old_entry.text_sha1 != None
506
text_modified = (self.text_sha1 != old_entry.text_sha1)
507
meta_modified = (self.executable != old_entry.executable)
508
return text_modified, meta_modified
509
510
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
511
output_to, reverse=False):
512
"""See InventoryEntry._diff."""
513
from_text = tree.get_file(self.file_id).readlines()
514
if to_entry:
515
to_text = to_tree.get_file(to_entry.file_id).readlines()
516
else:
517
to_text = []
518
if not reverse:
519
text_diff(from_label, from_text,
520
to_label, to_text, output_to)
521
else:
522
text_diff(to_label, to_text,
523
from_label, from_text, output_to)
524
525
def has_text(self):
526
"""See InventoryEntry.has_text."""
527
return True
528
529
def __init__(self, file_id, name, parent_id):
530
super(InventoryFile, self).__init__(file_id, name, parent_id)
531
self.kind = 'file'
532
533
def kind_character(self):
534
"""See InventoryEntry.kind_character."""
535
return ''
536
537
def _put_in_tar(self, item, tree):
538
"""See InventoryEntry._put_in_tar."""
539
item.type = tarfile.REGTYPE
540
fileobj = tree.get_file(self.file_id)
541
item.size = self.text_size
542
if tree.is_executable(self.file_id):
543
item.mode = 0755
544
else:
545
item.mode = 0644
546
return fileobj
547
548
def _put_on_disk(self, fullpath, tree):
549
"""See InventoryEntry._put_on_disk."""
550
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
551
if tree.is_executable(self.file_id):
552
os.chmod(fullpath, 0755)
553
554
def _read_tree_state(self, path, work_tree):
555
"""See InventoryEntry._read_tree_state."""
556
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
557
self.executable = work_tree.is_executable(self.file_id)
558
559
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
560
"""See InventoryEntry._snapshot_text."""
561
mutter('storing file {%s} in revision {%s}',
562
self.file_id, self.revision)
563
# special case to avoid diffing on renames or
564
# reparenting
565
if (len(file_parents) == 1
566
and self.text_sha1 == file_parents.values()[0].text_sha1
567
and self.text_size == file_parents.values()[0].text_size):
568
previous_ie = file_parents.values()[0]
569
weave_store.add_identical_text(
570
self.file_id, previous_ie.revision,
571
self.revision, file_parents, transaction)
572
else:
573
new_lines = work_tree.get_file(self.file_id).readlines()
574
self._add_text_to_weave(new_lines, file_parents, weave_store,
575
transaction)
576
self.text_sha1 = sha_strings(new_lines)
577
self.text_size = sum(map(len, new_lines))
578
579
580
def _unchanged(self, previous_ie):
581
"""See InventoryEntry._unchanged."""
582
compatible = super(InventoryFile, self)._unchanged(previous_ie)
583
if self.text_sha1 != previous_ie.text_sha1:
584
compatible = False
585
else:
586
# FIXME: 20050930 probe for the text size when getting sha1
587
# in _read_tree_state
588
self.text_size = previous_ie.text_size
589
if self.executable != previous_ie.executable:
590
compatible = False
591
return compatible
592
593
594
class InventoryLink(InventoryEntry):
595
"""A file in an inventory."""
596
597
__slots__ = ['symlink_target']
598
599
def _check(self, checker, rev_id, tree):
600
"""See InventoryEntry._check"""
601
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
602
raise BzrCheckError('symlink {%s} has text in revision {%s}'
603
% (self.file_id, rev_id))
604
if self.symlink_target == None:
605
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
606
% (self.file_id, rev_id))
607
608
def copy(self):
609
other = InventoryLink(self.file_id, self.name, self.parent_id)
610
other.symlink_target = self.symlink_target
611
other.revision = self.revision
612
return other
613
614
def detect_changes(self, old_entry):
615
"""See InventoryEntry.detect_changes."""
616
# FIXME: which _modified field should we use ? RBC 20051003
617
text_modified = (self.symlink_target != old_entry.symlink_target)
618
if text_modified:
619
mutter(" symlink target changed")
620
meta_modified = False
621
return text_modified, meta_modified
622
623
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
624
output_to, reverse=False):
625
"""See InventoryEntry._diff."""
626
from_text = self.symlink_target
627
if to_entry is not None:
628
to_text = to_entry.symlink_target
629
if reverse:
630
temp = from_text
631
from_text = to_text
632
to_text = temp
633
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
634
else:
635
if not reverse:
636
print >>output_to, '=== target was %r' % self.symlink_target
637
else:
638
print >>output_to, '=== target is %r' % self.symlink_target
639
640
def __init__(self, file_id, name, parent_id):
641
super(InventoryLink, self).__init__(file_id, name, parent_id)
642
self.kind = 'symlink'
643
644
def kind_character(self):
645
"""See InventoryEntry.kind_character."""
646
return ''
647
648
def _put_in_tar(self, item, tree):
649
"""See InventoryEntry._put_in_tar."""
650
item.type = tarfile.SYMTYPE
651
fileobj = None
652
item.size = 0
653
item.mode = 0755
654
item.linkname = self.symlink_target
655
return fileobj
656
657
def _put_on_disk(self, fullpath, tree):
658
"""See InventoryEntry._put_on_disk."""
659
try:
660
os.symlink(self.symlink_target, fullpath)
661
except OSError,e:
662
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
663
664
def _read_tree_state(self, path, work_tree):
665
"""See InventoryEntry._read_tree_state."""
666
self.symlink_target = work_tree.get_symlink_target(self.file_id)
667
668
def _unchanged(self, previous_ie):
669
"""See InventoryEntry._unchanged."""
670
compatible = super(InventoryLink, self)._unchanged(previous_ie)
671
if self.symlink_target != previous_ie.symlink_target:
672
compatible = False
673
return compatible
674
675
251
676
252
class Inventory(object):
677
253
"""Inventory of versioned files in a tree.
691
267
inserted, other than through the Inventory API.
692
268
693
269
>>> inv = Inventory()
694
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
695
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
270
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
696
271
>>> inv['123-123'].name
697
272
'hello.c'
698
273
708
283
>>> [x[0] for x in inv.iter_entries()]
709
284
['hello.c']
710
285
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
711
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
712
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
286
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
713
287
"""
714
288
def __init__(self, root_id=ROOT_ID):
715
289
"""Create or read an inventory.
721
295
The inventory is created with a default root directory, with
722
296
an id of None.
723
297
"""
724
# We are letting Branch.initialize() create a unique inventory
298
# We are letting Branch(init=True) create a unique inventory
725
299
# root id. Rather than generating a random one here.
726
300
#if root_id is None:
727
301
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
729
303
self._byid = {self.root.file_id: self.root}
730
304
731
305
732
def copy(self):
733
other = Inventory(self.root.file_id)
734
# copy recursively so we know directories will be added before
735
# their children. There are more efficient ways than this...
736
for path, entry in self.iter_entries():
737
if entry == self.root:
738
continue
739
other.add(entry.copy())
740
return other
741
742
743
306
def __iter__(self):
744
307
return iter(self._byid)
745
308
807
370
"""True if this entry contains a file with given id.
808
371
809
372
>>> inv = Inventory()
810
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
811
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
373
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
812
374
>>> '123' in inv
813
375
True
814
376
>>> '456' in inv
821
383
"""Return the entry for given file_id.
822
384
823
385
>>> inv = Inventory()
824
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
825
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
386
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
826
387
>>> inv['123123'].name
827
388
'hello.c'
828
389
"""
846
407
"""Add entry to inventory.
847
408
848
409
To add a file to a branch ready to be committed, use Branch.add,
849
which calls this.
850
851
Returns the new entry object.
852
"""
410
which calls this."""
853
411
if entry.file_id in self._byid:
854
412
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
855
413
867
425
868
426
self._byid[entry.file_id] = entry
869
427
parent.children[entry.name] = entry
870
return entry
871
428
872
429
873
430
def add_path(self, relpath, kind, file_id=None):
874
431
"""Add entry from a path.
875
432
876
The immediate parent must already be versioned.
877
878
Returns the new entry object."""
433
The immediate parent must already be versioned"""
879
434
from bzrlib.branch import gen_file_id
880
435
881
436
parts = bzrlib.osutils.splitpath(relpath)
888
443
parent_path = parts[:-1]
889
444
parent_id = self.path2id(parent_path)
890
445
if parent_id == None:
891
raise NotVersionedError(path=parent_path)
892
if kind == 'directory':
893
ie = InventoryDirectory(file_id, parts[-1], parent_id)
894
elif kind == 'file':
895
ie = InventoryFile(file_id, parts[-1], parent_id)
896
elif kind == 'symlink':
897
ie = InventoryLink(file_id, parts[-1], parent_id)
898
else:
899
raise BzrError("unknown kind %r" % kind)
446
raise NotVersionedError(parent_path)
447
448
ie = InventoryEntry(file_id, parts[-1],
449
kind=kind, parent_id=parent_id)
900
450
return self.add(ie)
901
451
902
452
904
454
"""Remove entry by id.
905
455
906
456
>>> inv = Inventory()
907
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
908
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
457
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
909
458
>>> '123' in inv
910
459
True
911
460
>>> del inv['123']
927
476
del self[ie.parent_id].children[ie.name]
928
477
929
478
479
def to_element(self):
480
"""Convert to XML Element"""
481
from bzrlib.xml import Element
482
483
e = Element('inventory')
484
e.text = '\n'
485
if self.root.file_id not in (None, ROOT_ID):
486
e.set('file_id', self.root.file_id)
487
for path, ie in self.iter_entries():
488
e.append(ie.to_element())
489
return e
490
491
492
def from_element(cls, elt):
493
"""Construct from XML Element
494
495
>>> inv = Inventory()
496
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
497
>>> elt = inv.to_element()
498
>>> inv2 = Inventory.from_element(elt)
499
>>> inv2 == inv
500
True
501
"""
502
# XXXX: doctest doesn't run this properly under python2.3
503
assert elt.tag == 'inventory'
504
root_id = elt.get('file_id') or ROOT_ID
505
o = cls(root_id)
506
for e in elt:
507
ie = InventoryEntry.from_element(e)
508
if ie.parent_id == ROOT_ID:
509
ie.parent_id = root_id
510
o.add(ie)
511
return o
512
513
from_element = classmethod(from_element)
514
515
930
516
def __eq__(self, other):
931
517
"""Compare two sets by comparing their contents.
932
518
934
520
>>> i2 = Inventory()
935
521
>>> i1 == i2
936
522
True
937
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
938
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
523
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
939
524
>>> i1 == i2
940
525
False
941
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
942
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
526
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
943
527
>>> i1 == i2
944
528
True
945
529
"""
954
538
955
539
956
540
def __ne__(self, other):
957
return not self.__eq__(other)
541
return not (self == other)
958
542
959
543
960
544
def __hash__(self):
961
545
raise ValueError('not hashable')
962
546
963
547
548
964
549
def get_idpath(self, file_id):
965
550
"""Return a list of file_ids for the path to an entry.
966
551
981
566
982
567
983
568
def id2path(self, file_id):
984
"""Return as a list the path to file_id.
985
986
>>> i = Inventory()
987
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
988
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
989
>>> print i.id2path('foo-id').replace(os.sep, '/')
990
src/foo.c
991
"""
569
"""Return as a list the path to file_id."""
570
992
571
# get all names, skipping root
993
572
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
994
573
return os.sep.join(p)
Loggerhead is a web-based interface for Breezy
Version: 2.0.1