« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 254
- compare with another revision
- download diff
- download tarball
- view history from revision 254
- Committer: Martin Pool
- Date: 2005-04-15 01:31:21 UTC
- Revision ID: mbp@sourcefrog.net-20050415013121-b18f1be12a735066
- Doc cleanups from Magnus Therning
- files added:
- bzrlib/remotebranch.py
- files removed:
- HACKING
- bzrlib/plugins
- bzrlib/selftest
- bzrlib/store
- bzrlib/transport
- bzrlib/util
- bzrlib/util/configobj
- bzrlib/util/configobj/docs
- bzrlib/util/effbot
- bzrlib/util/effbot/org
- contrib
- contrib/bash
- contrib/emacs
- contrib/zsh
- patches
- tools
- files renamed:
- contrib/newinventory.py => bzrlib/newinventory.py
- bzrlib/store/__init__.py => bzrlib/store.py
- bzrlib/util/elementtree/ => elementtree/
- bzrlib/util/urlgrabber/ => urlgrabber/
- 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.
17
18
# TODO: Maybe store inventory_id in the file? Not really needed.
19
25
20
26
21
# This should really be an id randomly assigned when the tree is
27
22
# created, but it's not for now.
28
23
ROOT_ID = "TREE_ROOT"
29
24
30
25
31
import os.path
32
import re
33
import sys
34
import tarfile
35
import types
26
import sys, os.path, types, re
27
from sets import Set
28
29
try:
30
from cElementTree import Element, ElementTree, SubElement
31
except ImportError:
32
from elementtree.ElementTree import Element, ElementTree, SubElement
33
34
from xml import XMLMixin
35
from errors import bailout, BzrError
36
36
37
37
import bzrlib
38
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
39
appendpath, sha_strings)
38
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
40
39
from bzrlib.trace import mutter
41
from bzrlib.errors import (NotVersionedError, InvalidEntryName,
42
BzrError, BzrCheckError)
43
44
45
class InventoryEntry(object):
40
41
class InventoryEntry(XMLMixin):
46
42
"""Description of a versioned file.
47
43
48
44
An InventoryEntry has the following fields, which are also
49
45
present in the XML inventory-entry element:
50
46
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.)
47
* *file_id*
48
* *name*: (only the basename within the directory, must not
49
contain slashes)
50
* *kind*: "directory" or "file"
51
* *directory_id*: (if absent/null means the branch root directory)
52
* *text_sha1*: only for files
53
* *text_size*: in bytes, only for files
54
* *text_id*: identifier for the text version, only for files
55
56
InventoryEntries can also exist inside a WorkingTree
57
inventory, in which case they are not yet bound to a
58
particular revision of the file. In that case the text_sha1,
59
text_size and text_id are absent.
60
74
61
75
62
>>> i = Inventory()
76
63
>>> i.path2id('')
77
64
'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')
65
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
66
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
82
67
>>> for j in i.iter_entries():
83
68
... print j
84
69
...
85
('src', InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
86
('src/hello.c', InventoryFile('2323', 'hello.c', parent_id='123'))
87
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
70
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
71
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
72
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
88
73
Traceback (most recent call last):
89
74
...
90
BzrError: inventory already contains entry with id {2323}
91
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
92
InventoryFile('2324', 'bye.c', parent_id='123')
93
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
94
InventoryDirectory('2325', 'wibble', parent_id='123')
75
BzrError: ('inventory already contains entry with id {2323}', [])
76
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
77
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
95
78
>>> i.path2id('src/wibble')
96
79
'2325'
97
80
>>> '2325' in i
98
81
True
99
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
100
InventoryFile('2326', 'wibble.c', parent_id='2325')
82
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
101
83
>>> i['2326']
102
InventoryFile('2326', 'wibble.c', parent_id='2325')
103
>>> for path, entry in i.iter_entries():
104
... print path.replace('\\\\', '/') # for win32 os.sep
105
... assert i.path2id(path)
84
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
85
>>> for j in i.iter_entries():
86
... print j[0]
87
... assert i.path2id(j[0])
106
88
...
107
89
src
108
90
src/bye.c
109
91
src/hello.c
110
92
src/wibble
111
93
src/wibble/wibble.c
112
>>> i.id2path('2326').replace('\\\\', '/')
94
>>> i.id2path('2326')
113
95
'src/wibble/wibble.c'
96
97
TODO: Maybe also keep the full path of the entry, and the children?
98
But those depend on its position within a particular inventory, and
99
it would be nice not to need to hold the backpointer here.
114
100
"""
101
102
# TODO: split InventoryEntry into subclasses for files,
103
# directories, etc etc.
115
104
116
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
117
'text_id', 'parent_id', 'children', 'executable',
118
'revision']
119
120
def _add_text_to_weave(self, new_lines, parents, weave_store, transaction):
121
weave_store.add_text(self.file_id, self.revision, new_lines, parents,
122
transaction)
123
124
def detect_changes(self, old_entry):
125
"""Return a (text_modified, meta_modified) from this to old_entry.
126
127
_read_tree_state must have been called on self and old_entry prior to
128
calling detect_changes.
129
"""
130
return False, False
131
132
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
133
output_to, reverse=False):
134
"""Perform a diff from this to to_entry.
135
136
text_diff will be used for textual difference calculation.
137
This is a template method, override _diff in child classes.
138
"""
139
self._read_tree_state(tree.id2path(self.file_id), tree)
140
if to_entry:
141
# cannot diff from one kind to another - you must do a removal
142
# and an addif they do not match.
143
assert self.kind == to_entry.kind
144
to_entry._read_tree_state(to_tree.id2path(to_entry.file_id),
145
to_tree)
146
self._diff(text_diff, from_label, tree, to_label, to_entry, to_tree,
147
output_to, reverse)
148
149
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
150
output_to, reverse=False):
151
"""Perform a diff between two entries of the same kind."""
152
153
def find_previous_heads(self, previous_inventories, entry_weave):
154
"""Return the revisions and entries that directly preceed this.
155
156
Returned as a map from revision to inventory entry.
157
158
This is a map containing the file revisions in all parents
159
for which the file exists, and its revision is not a parent of
160
any other. If the file is new, the set will be empty.
161
"""
162
def get_ancestors(weave, entry):
163
return set(map(weave.idx_to_name,
164
weave.inclusions([weave.lookup(entry.revision)])))
165
heads = {}
166
head_ancestors = {}
167
for inv in previous_inventories:
168
if self.file_id in inv:
169
ie = inv[self.file_id]
170
assert ie.file_id == self.file_id
171
if ie.revision in heads:
172
# fixup logic, there was a bug in revision updates.
173
# with x bit support.
174
try:
175
if heads[ie.revision].executable != ie.executable:
176
heads[ie.revision].executable = False
177
ie.executable = False
178
except AttributeError:
179
pass
180
assert heads[ie.revision] == ie
181
else:
182
# may want to add it.
183
# may already be covered:
184
already_present = 0 != len(
185
[head for head in heads
186
if ie.revision in head_ancestors[head]])
187
if already_present:
188
# an ancestor of a known head.
189
continue
190
# definately a head:
191
ancestors = get_ancestors(entry_weave, ie)
192
# may knock something else out:
193
check_heads = list(heads.keys())
194
for head in check_heads:
195
if head in ancestors:
196
# this head is not really a head
197
heads.pop(head)
198
head_ancestors[ie.revision] = ancestors
199
heads[ie.revision] = ie
200
return heads
201
202
def get_tar_item(self, root, dp, now, tree):
203
"""Get a tarfile item and a file stream for its content."""
204
item = tarfile.TarInfo(os.path.join(root, dp))
205
# TODO: would be cool to actually set it to the timestamp of the
206
# revision it was last changed
207
item.mtime = now
208
fileobj = self._put_in_tar(item, tree)
209
return item, fileobj
210
211
def has_text(self):
212
"""Return true if the object this entry represents has textual data.
213
214
Note that textual data includes binary content.
215
216
Also note that all entries get weave files created for them.
217
This attribute is primarily used when upgrading from old trees that
218
did not have the weave index for all inventory entries.
219
"""
220
return False
221
222
def __init__(self, file_id, name, parent_id, text_id=None):
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
223
106
"""Create an InventoryEntry
224
107
225
108
The filename must be a single component, relative to the
226
109
parent directory; it cannot be a whole path or relative name.
227
110
228
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
229
112
>>> e.name
230
113
'hello.c'
231
114
>>> e.file_id
232
115
'123'
233
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
234
117
Traceback (most recent call last):
235
InvalidEntryName: Invalid entry name: src/hello.c
118
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
236
119
"""
237
assert isinstance(name, basestring), name
238
if '/' in name or '\\' in name:
239
raise InvalidEntryName(name=name)
240
self.executable = False
241
self.revision = None
242
self.text_sha1 = None
243
self.text_size = None
120
121
if len(splitpath(name)) != 1:
122
bailout('InventoryEntry name is not a simple filename: %r'
123
% name)
124
244
125
self.file_id = file_id
245
126
self.name = name
127
self.kind = kind
246
128
self.text_id = text_id
247
129
self.parent_id = parent_id
248
self.symlink_target = None
249
250
def kind_character(self):
251
"""Return a short kind indicator useful for appending to names."""
252
raise BzrError('unknown kind %r' % self.kind)
253
254
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
255
256
def _put_in_tar(self, item, tree):
257
"""populate item for stashing in a tar, and return the content stream.
258
259
If no content is available, return None.
260
"""
261
raise BzrError("don't know how to export {%s} of kind %r" %
262
(self.file_id, self.kind))
263
264
def put_on_disk(self, dest, dp, tree):
265
"""Create a representation of self on disk in the prefix dest.
266
267
This is a template method - implement _put_on_disk in subclasses.
268
"""
269
fullpath = appendpath(dest, dp)
270
self._put_on_disk(fullpath, tree)
271
mutter(" export {%s} kind %s to %s" % (self.file_id, self.kind, fullpath))
272
273
def _put_on_disk(self, fullpath, tree):
274
"""Put this entry onto disk at fullpath, from tree tree."""
275
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
130
self.text_sha1 = None
131
self.text_size = None
132
if kind == 'directory':
133
self.children = {}
134
elif kind == 'file':
135
pass
136
else:
137
raise BzrError("unhandled entry kind %r" % kind)
138
139
276
140
277
141
def sorted_children(self):
278
142
l = self.children.items()
279
143
l.sort()
280
144
return l
281
145
282
@staticmethod
283
def versionable_kind(kind):
284
return kind in ('file', 'directory', 'symlink')
285
286
def check(self, checker, rev_id, inv, tree):
287
"""Check this inventory entry is intact.
288
289
This is a template method, override _check for kind specific
290
tests.
291
"""
292
if self.parent_id != None:
293
if not inv.has_id(self.parent_id):
294
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
295
% (self.parent_id, rev_id))
296
self._check(checker, rev_id, tree)
297
298
def _check(self, checker, rev_id, tree):
299
"""Check this inventory entry for kind specific errors."""
300
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
301
(self.kind, rev_id))
302
303
146
304
147
def copy(self):
305
"""Clone this inventory entry."""
306
raise NotImplementedError
148
other = InventoryEntry(self.file_id, self.name, self.kind,
149
self.parent_id, text_id=self.text_id)
150
other.text_sha1 = self.text_sha1
151
other.text_size = self.text_size
152
return other
307
153
308
def _get_snapshot_change(self, previous_entries):
309
if len(previous_entries) > 1:
310
return 'merged'
311
elif len(previous_entries) == 0:
312
return 'added'
313
else:
314
return 'modified/renamed/reparented'
315
154
316
155
def __repr__(self):
317
return ("%s(%r, %r, parent_id=%r)"
156
return ("%s(%r, %r, kind=%r, parent_id=%r)"
318
157
% (self.__class__.__name__,
319
158
self.file_id,
320
159
self.name,
160
self.kind,
321
161
self.parent_id))
322
162
323
def snapshot(self, revision, path, previous_entries,
324
work_tree, weave_store, transaction):
325
"""Make a snapshot of this entry which may or may not have changed.
326
327
This means that all its fields are populated, that it has its
328
text stored in the text store or weave.
329
"""
330
mutter('new parents of %s are %r', path, previous_entries)
331
self._read_tree_state(path, work_tree)
332
if len(previous_entries) == 1:
333
# cannot be unchanged unless there is only one parent file rev.
334
parent_ie = previous_entries.values()[0]
335
if self._unchanged(parent_ie):
336
mutter("found unchanged entry")
337
self.revision = parent_ie.revision
338
return "unchanged"
339
return self.snapshot_revision(revision, previous_entries,
340
work_tree, weave_store, transaction)
341
342
def snapshot_revision(self, revision, previous_entries, work_tree,
343
weave_store, transaction):
344
"""Record this revision unconditionally."""
345
mutter('new revision for {%s}', self.file_id)
346
self.revision = revision
347
change = self._get_snapshot_change(previous_entries)
348
self._snapshot_text(previous_entries, work_tree, weave_store,
349
transaction)
350
return change
351
352
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
353
"""Record the 'text' of this entry, whatever form that takes.
354
355
This default implementation simply adds an empty text.
356
"""
357
mutter('storing file {%s} in revision {%s}',
358
self.file_id, self.revision)
359
self._add_text_to_weave([], file_parents, weave_store, transaction)
360
361
def __eq__(self, other):
163
164
def to_element(self):
165
"""Convert to XML element"""
166
e = Element('entry')
167
168
e.set('name', self.name)
169
e.set('file_id', self.file_id)
170
e.set('kind', self.kind)
171
172
if self.text_size != None:
173
e.set('text_size', '%d' % self.text_size)
174
175
for f in ['text_id', 'text_sha1']:
176
v = getattr(self, f)
177
if v != None:
178
e.set(f, v)
179
180
# to be conservative, we don't externalize the root pointers
181
# for now, leaving them as null in the xml form. in a future
182
# version it will be implied by nested elements.
183
if self.parent_id != ROOT_ID:
184
assert isinstance(self.parent_id, basestring)
185
e.set('parent_id', self.parent_id)
186
187
e.tail = '\n'
188
189
return e
190
191
192
def from_element(cls, elt):
193
assert elt.tag == 'entry'
194
195
## original format inventories don't have a parent_id for
196
## nodes in the root directory, but it's cleaner to use one
197
## internally.
198
parent_id = elt.get('parent_id')
199
if parent_id == None:
200
parent_id = ROOT_ID
201
202
self = cls(elt.get('file_id'), elt.get('name'), elt.get('kind'), parent_id)
203
self.text_id = elt.get('text_id')
204
self.text_sha1 = elt.get('text_sha1')
205
206
## mutter("read inventoryentry: %r" % (elt.attrib))
207
208
v = elt.get('text_size')
209
self.text_size = v and int(v)
210
211
return self
212
213
214
from_element = classmethod(from_element)
215
216
def __cmp__(self, other):
217
if self is other:
218
return 0
362
219
if not isinstance(other, InventoryEntry):
363
220
return NotImplemented
364
221
365
return ((self.file_id == other.file_id)
366
and (self.name == other.name)
367
and (other.symlink_target == self.symlink_target)
368
and (self.text_sha1 == other.text_sha1)
369
and (self.text_size == other.text_size)
370
and (self.text_id == other.text_id)
371
and (self.parent_id == other.parent_id)
372
and (self.kind == other.kind)
373
and (self.revision == other.revision)
374
and (self.executable == other.executable)
375
)
376
377
def __ne__(self, other):
378
return not (self == other)
379
380
def __hash__(self):
381
raise ValueError('not hashable')
382
383
def _unchanged(self, previous_ie):
384
"""Has this entry changed relative to previous_ie.
385
386
This method should be overriden in child classes.
387
"""
388
compatible = True
389
# different inv parent
390
if previous_ie.parent_id != self.parent_id:
391
compatible = False
392
# renamed
393
elif previous_ie.name != self.name:
394
compatible = False
395
return compatible
396
397
def _read_tree_state(self, path, work_tree):
398
"""Populate fields in the inventory entry from the given tree.
399
400
Note that this should be modified to be a noop on virtual trees
401
as all entries created there are prepopulated.
402
"""
403
# TODO: Rather than running this manually, we should check the
404
# working sha1 and other expensive properties when they're
405
# first requested, or preload them if they're already known
406
pass # nothing to do by default
222
return cmp(self.file_id, other.file_id) \
223
or cmp(self.name, other.name) \
224
or cmp(self.text_sha1, other.text_sha1) \
225
or cmp(self.text_size, other.text_size) \
226
or cmp(self.text_id, other.text_id) \
227
or cmp(self.parent_id, other.parent_id) \
228
or cmp(self.kind, other.kind)
229
407
230
408
231
409
232
class RootEntry(InventoryEntry):
410
411
def _check(self, checker, rev_id, tree):
412
"""See InventoryEntry._check"""
413
414
233
def __init__(self, file_id):
415
234
self.file_id = file_id
416
235
self.children = {}
418
237
self.parent_id = None
419
238
self.name = ''
420
239
421
def __eq__(self, other):
240
def __cmp__(self, other):
241
if self is other:
242
return 0
422
243
if not isinstance(other, RootEntry):
423
244
return NotImplemented
424
425
return (self.file_id == other.file_id) \
426
and (self.children == other.children)
427
428
429
class InventoryDirectory(InventoryEntry):
430
"""A directory in an inventory."""
431
432
def _check(self, checker, rev_id, tree):
433
"""See InventoryEntry._check"""
434
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
435
raise BzrCheckError('directory {%s} has text in revision {%s}'
436
% (self.file_id, rev_id))
437
438
def copy(self):
439
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
440
other.revision = self.revision
441
# note that children are *not* copied; they're pulled across when
442
# others are added
443
return other
444
445
def __init__(self, file_id, name, parent_id):
446
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
447
self.children = {}
448
self.kind = 'directory'
449
450
def kind_character(self):
451
"""See InventoryEntry.kind_character."""
452
return '/'
453
454
def _put_in_tar(self, item, tree):
455
"""See InventoryEntry._put_in_tar."""
456
item.type = tarfile.DIRTYPE
457
fileobj = None
458
item.name += '/'
459
item.size = 0
460
item.mode = 0755
461
return fileobj
462
463
def _put_on_disk(self, fullpath, tree):
464
"""See InventoryEntry._put_on_disk."""
465
os.mkdir(fullpath)
466
467
468
class InventoryFile(InventoryEntry):
469
"""A file in an inventory."""
470
471
def _check(self, checker, rev_id, tree):
472
"""See InventoryEntry._check"""
473
revision = self.revision
474
t = (self.file_id, revision)
475
if t in checker.checked_texts:
476
prev_sha = checker.checked_texts[t]
477
if prev_sha != self.text_sha1:
478
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
479
(self.file_id, rev_id))
480
else:
481
checker.repeated_text_cnt += 1
482
return
483
mutter('check version {%s} of {%s}', rev_id, self.file_id)
484
file_lines = tree.get_file_lines(self.file_id)
485
checker.checked_text_cnt += 1
486
if self.text_size != sum(map(len, file_lines)):
487
raise BzrCheckError('text {%s} wrong size' % self.text_id)
488
if self.text_sha1 != sha_strings(file_lines):
489
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
490
checker.checked_texts[t] = self.text_sha1
491
492
def copy(self):
493
other = InventoryFile(self.file_id, self.name, self.parent_id)
494
other.executable = self.executable
495
other.text_id = self.text_id
496
other.text_sha1 = self.text_sha1
497
other.text_size = self.text_size
498
other.revision = self.revision
499
return other
500
501
def detect_changes(self, old_entry):
502
"""See InventoryEntry.detect_changes."""
503
assert self.text_sha1 != None
504
assert old_entry.text_sha1 != None
505
text_modified = (self.text_sha1 != old_entry.text_sha1)
506
meta_modified = (self.executable != old_entry.executable)
507
return text_modified, meta_modified
508
509
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
510
output_to, reverse=False):
511
"""See InventoryEntry._diff."""
512
from_text = tree.get_file(self.file_id).readlines()
513
if to_entry:
514
to_text = to_tree.get_file(to_entry.file_id).readlines()
515
else:
516
to_text = []
517
if not reverse:
518
text_diff(from_label, from_text,
519
to_label, to_text, output_to)
520
else:
521
text_diff(to_label, to_text,
522
from_label, from_text, output_to)
523
524
def has_text(self):
525
"""See InventoryEntry.has_text."""
526
return True
527
528
def __init__(self, file_id, name, parent_id):
529
super(InventoryFile, self).__init__(file_id, name, parent_id)
530
self.kind = 'file'
531
532
def kind_character(self):
533
"""See InventoryEntry.kind_character."""
534
return ''
535
536
def _put_in_tar(self, item, tree):
537
"""See InventoryEntry._put_in_tar."""
538
item.type = tarfile.REGTYPE
539
fileobj = tree.get_file(self.file_id)
540
item.size = self.text_size
541
if tree.is_executable(self.file_id):
542
item.mode = 0755
543
else:
544
item.mode = 0644
545
return fileobj
546
547
def _put_on_disk(self, fullpath, tree):
548
"""See InventoryEntry._put_on_disk."""
549
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
550
if tree.is_executable(self.file_id):
551
os.chmod(fullpath, 0755)
552
553
def _read_tree_state(self, path, work_tree):
554
"""See InventoryEntry._read_tree_state."""
555
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
556
self.executable = work_tree.is_executable(self.file_id)
557
558
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
559
"""See InventoryEntry._snapshot_text."""
560
mutter('storing file {%s} in revision {%s}',
561
self.file_id, self.revision)
562
# special case to avoid diffing on renames or
563
# reparenting
564
if (len(file_parents) == 1
565
and self.text_sha1 == file_parents.values()[0].text_sha1
566
and self.text_size == file_parents.values()[0].text_size):
567
previous_ie = file_parents.values()[0]
568
weave_store.add_identical_text(
569
self.file_id, previous_ie.revision,
570
self.revision, file_parents, transaction)
571
else:
572
new_lines = work_tree.get_file(self.file_id).readlines()
573
self._add_text_to_weave(new_lines, file_parents, weave_store,
574
transaction)
575
self.text_sha1 = sha_strings(new_lines)
576
self.text_size = sum(map(len, new_lines))
577
578
579
def _unchanged(self, previous_ie):
580
"""See InventoryEntry._unchanged."""
581
compatible = super(InventoryFile, self)._unchanged(previous_ie)
582
if self.text_sha1 != previous_ie.text_sha1:
583
compatible = False
584
else:
585
# FIXME: 20050930 probe for the text size when getting sha1
586
# in _read_tree_state
587
self.text_size = previous_ie.text_size
588
if self.executable != previous_ie.executable:
589
compatible = False
590
return compatible
591
592
593
class InventoryLink(InventoryEntry):
594
"""A file in an inventory."""
595
596
__slots__ = ['symlink_target']
597
598
def _check(self, checker, rev_id, tree):
599
"""See InventoryEntry._check"""
600
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
601
raise BzrCheckError('symlink {%s} has text in revision {%s}'
602
% (self.file_id, rev_id))
603
if self.symlink_target == None:
604
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
605
% (self.file_id, rev_id))
606
607
def copy(self):
608
other = InventoryLink(self.file_id, self.name, self.parent_id)
609
other.symlink_target = self.symlink_target
610
other.revision = self.revision
611
return other
612
613
def detect_changes(self, old_entry):
614
"""See InventoryEntry.detect_changes."""
615
# FIXME: which _modified field should we use ? RBC 20051003
616
text_modified = (self.symlink_target != old_entry.symlink_target)
617
if text_modified:
618
mutter(" symlink target changed")
619
meta_modified = False
620
return text_modified, meta_modified
621
622
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
623
output_to, reverse=False):
624
"""See InventoryEntry._diff."""
625
from_text = self.symlink_target
626
if to_entry is not None:
627
to_text = to_entry.symlink_target
628
if reverse:
629
temp = from_text
630
from_text = to_text
631
to_text = temp
632
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
633
else:
634
if not reverse:
635
print >>output_to, '=== target was %r' % self.symlink_target
636
else:
637
print >>output_to, '=== target is %r' % self.symlink_target
638
639
def __init__(self, file_id, name, parent_id):
640
super(InventoryLink, self).__init__(file_id, name, parent_id)
641
self.kind = 'symlink'
642
643
def kind_character(self):
644
"""See InventoryEntry.kind_character."""
645
return ''
646
647
def _put_in_tar(self, item, tree):
648
"""See InventoryEntry._put_in_tar."""
649
iterm.type = tarfile.SYMTYPE
650
fileobj = None
651
item.size = 0
652
item.mode = 0755
653
item.linkname = self.symlink_target
654
return fileobj
655
656
def _put_on_disk(self, fullpath, tree):
657
"""See InventoryEntry._put_on_disk."""
658
try:
659
os.symlink(self.symlink_target, fullpath)
660
except OSError,e:
661
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
662
663
def _read_tree_state(self, path, work_tree):
664
"""See InventoryEntry._read_tree_state."""
665
self.symlink_target = work_tree.get_symlink_target(self.file_id)
666
667
def _unchanged(self, previous_ie):
668
"""See InventoryEntry._unchanged."""
669
compatible = super(InventoryLink, self)._unchanged(previous_ie)
670
if self.symlink_target != previous_ie.symlink_target:
671
compatible = False
672
return compatible
673
674
675
class Inventory(object):
245
return cmp(self.file_id, other.file_id) \
246
or cmp(self.children, other.children)
247
248
249
250
class Inventory(XMLMixin):
676
251
"""Inventory of versioned files in a tree.
677
252
678
253
This describes which file_id is present at each point in the tree,
690
265
inserted, other than through the Inventory API.
691
266
692
267
>>> inv = Inventory()
693
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
694
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
268
>>> inv.write_xml(sys.stdout)
269
<inventory>
270
</inventory>
271
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
695
272
>>> inv['123-123'].name
696
273
'hello.c'
697
274
706
283
707
284
>>> [x[0] for x in inv.iter_entries()]
708
285
['hello.c']
709
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
710
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
711
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
286
287
>>> inv.write_xml(sys.stdout)
288
<inventory>
289
<entry file_id="123-123" kind="file" name="hello.c" />
290
</inventory>
291
712
292
"""
713
def __init__(self, root_id=ROOT_ID):
293
294
## TODO: Make sure only canonical filenames are stored.
295
296
## TODO: Do something sensible about the possible collisions on
297
## case-losing filesystems. Perhaps we should just always forbid
298
## such collisions.
299
300
## TODO: No special cases for root, rather just give it a file id
301
## like everything else.
302
303
## TODO: Probably change XML serialization to use nesting rather
304
## than parent_id pointers.
305
306
## TODO: Perhaps hold the ElementTree in memory and work directly
307
## on that rather than converting into Python objects every time?
308
309
def __init__(self):
714
310
"""Create or read an inventory.
715
311
716
312
If a working directory is specified, the inventory is read
720
316
The inventory is created with a default root directory, with
721
317
an id of None.
722
318
"""
723
# We are letting Branch.initialize() create a unique inventory
724
# root id. Rather than generating a random one here.
725
#if root_id is None:
726
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
727
self.root = RootEntry(root_id)
319
self.root = RootEntry(ROOT_ID)
728
320
self._byid = {self.root.file_id: self.root}
729
321
730
322
731
def copy(self):
732
other = Inventory(self.root.file_id)
733
# copy recursively so we know directories will be added before
734
# their children. There are more efficient ways than this...
735
for path, entry in self.iter_entries():
736
if entry == self.root:
737
continue
738
other.add(entry.copy())
739
return other
740
741
742
323
def __iter__(self):
743
324
return iter(self._byid)
744
325
762
343
yield name, ie
763
344
if ie.kind == 'directory':
764
345
for cn, cie in self.iter_entries(from_dir=ie.file_id):
765
yield os.path.join(name, cn), cie
766
767
768
def entries(self):
769
"""Return list of (path, ie) for all entries except the root.
770
771
This may be faster than iter_entries.
346
yield '/'.join((name, cn)), cie
347
348
349
350
def directories(self):
351
"""Return (path, entry) pairs for all directories.
772
352
"""
773
accum = []
774
def descend(dir_ie, dir_path):
775
kids = dir_ie.children.items()
776
kids.sort()
777
for name, ie in kids:
778
child_path = os.path.join(dir_path, name)
779
accum.append((child_path, ie))
353
def descend(parent_ie):
354
parent_name = parent_ie.name
355
yield parent_name, parent_ie
356
357
# directory children in sorted order
358
dn = []
359
for ie in parent_ie.children.itervalues():
780
360
if ie.kind == 'directory':
781
descend(ie, child_path)
782
783
descend(self.root, '')
784
return accum
785
786
787
def directories(self):
788
"""Return (path, entry) pairs for all directories, including the root.
789
"""
790
accum = []
791
def descend(parent_ie, parent_path):
792
accum.append((parent_path, parent_ie))
361
dn.append((ie.name, ie))
362
dn.sort()
793
363
794
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
795
kids.sort()
364
for name, child_ie in dn:
365
for sub_name, sub_ie in descend(child_ie):
366
yield appendpath(parent_name, sub_name), sub_ie
796
367
797
for name, child_ie in kids:
798
child_path = os.path.join(parent_path, name)
799
descend(child_ie, child_path)
800
descend(self.root, '')
801
return accum
368
for name, ie in descend(self.root):
369
yield name, ie
802
370
803
371
804
372
806
374
"""True if this entry contains a file with given id.
807
375
808
376
>>> inv = Inventory()
809
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
810
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
377
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
811
378
>>> '123' in inv
812
379
True
813
380
>>> '456' in inv
820
387
"""Return the entry for given file_id.
821
388
822
389
>>> inv = Inventory()
823
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
824
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
390
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
825
391
>>> inv['123123'].name
826
392
'hello.c'
827
393
"""
394
if file_id == None:
395
raise BzrError("can't look up file_id None")
396
828
397
try:
829
398
return self._byid[file_id]
830
399
except KeyError:
831
if file_id == None:
832
raise BzrError("can't look up file_id None")
833
else:
834
raise BzrError("file_id {%s} not in inventory" % file_id)
835
836
837
def get_file_kind(self, file_id):
838
return self._byid[file_id].kind
400
raise BzrError("file_id {%s} not in inventory" % file_id)
401
839
402
840
403
def get_child(self, parent_id, filename):
841
404
return self[parent_id].children.get(filename)
845
408
"""Add entry to inventory.
846
409
847
410
To add a file to a branch ready to be committed, use Branch.add,
848
which calls this.
849
850
Returns the new entry object.
851
"""
411
which calls this."""
852
412
if entry.file_id in self._byid:
853
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
854
855
if entry.parent_id == ROOT_ID or entry.parent_id is None:
856
entry.parent_id = self.root.file_id
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
857
414
858
415
try:
859
416
parent = self._byid[entry.parent_id]
860
417
except KeyError:
861
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
862
419
863
420
if parent.children.has_key(entry.name):
864
raise BzrError("%s is already versioned" %
421
bailout("%s is already versioned" %
865
422
appendpath(self.id2path(parent.file_id), entry.name))
866
423
867
424
self._byid[entry.file_id] = entry
868
425
parent.children[entry.name] = entry
869
return entry
870
426
871
427
872
428
def add_path(self, relpath, kind, file_id=None):
873
429
"""Add entry from a path.
874
430
875
The immediate parent must already be versioned.
876
877
Returns the new entry object."""
878
from bzrlib.branch import gen_file_id
879
431
The immediate parent must already be versioned"""
880
432
parts = bzrlib.osutils.splitpath(relpath)
881
433
if len(parts) == 0:
882
raise BzrError("cannot re-add root of inventory")
434
bailout("cannot re-add root of inventory")
883
435
884
436
if file_id == None:
885
file_id = gen_file_id(relpath)
437
file_id = bzrlib.branch.gen_file_id(relpath)
886
438
887
parent_path = parts[:-1]
888
parent_id = self.path2id(parent_path)
889
if parent_id == None:
890
raise NotVersionedError(path=parent_path)
891
if kind == 'directory':
892
ie = InventoryDirectory(file_id, parts[-1], parent_id)
893
elif kind == 'file':
894
ie = InventoryFile(file_id, parts[-1], parent_id)
895
elif kind == 'symlink':
896
ie = InventoryLink(file_id, parts[-1], parent_id)
897
else:
898
raise BzrError("unknown kind %r" % kind)
439
parent_id = self.path2id(parts[:-1])
440
assert parent_id != None
441
ie = InventoryEntry(file_id, parts[-1],
442
kind=kind, parent_id=parent_id)
899
443
return self.add(ie)
900
444
901
445
903
447
"""Remove entry by id.
904
448
905
449
>>> inv = Inventory()
906
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
907
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
450
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
908
451
>>> '123' in inv
909
452
True
910
453
>>> del inv['123']
926
469
del self[ie.parent_id].children[ie.name]
927
470
928
471
929
def __eq__(self, other):
472
def id_set(self):
473
return Set(self._byid)
474
475
476
def to_element(self):
477
"""Convert to XML Element"""
478
e = Element('inventory')
479
e.text = '\n'
480
for path, ie in self.iter_entries():
481
e.append(ie.to_element())
482
return e
483
484
485
def from_element(cls, elt):
486
"""Construct from XML Element
487
488
>>> inv = Inventory()
489
>>> inv.add(InventoryEntry('foo.c-123981239', 'foo.c', 'file', ROOT_ID))
490
>>> elt = inv.to_element()
491
>>> inv2 = Inventory.from_element(elt)
492
>>> inv2 == inv
493
True
494
"""
495
assert elt.tag == 'inventory'
496
o = cls()
497
for e in elt:
498
o.add(InventoryEntry.from_element(e))
499
return o
500
501
from_element = classmethod(from_element)
502
503
504
def __cmp__(self, other):
930
505
"""Compare two sets by comparing their contents.
931
506
932
507
>>> i1 = Inventory()
933
508
>>> i2 = Inventory()
934
509
>>> i1 == i2
935
510
True
936
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
937
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
511
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
938
512
>>> i1 == i2
939
513
False
940
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
941
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
514
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
942
515
>>> i1 == i2
943
516
True
944
517
"""
518
if self is other:
519
return 0
520
945
521
if not isinstance(other, Inventory):
946
522
return NotImplemented
947
523
948
if len(self._byid) != len(other._byid):
949
# shortcut: obviously not the same
950
return False
951
952
return self._byid == other._byid
953
954
955
def __ne__(self, other):
956
return not self.__eq__(other)
957
958
959
def __hash__(self):
960
raise ValueError('not hashable')
524
if self.id_set() ^ other.id_set():
525
return 1
526
527
for file_id in self._byid:
528
c = cmp(self[file_id], other[file_id])
529
if c: return c
530
531
return 0
961
532
962
533
963
534
def get_idpath(self, file_id):
973
544
try:
974
545
ie = self._byid[file_id]
975
546
except KeyError:
976
raise BzrError("file_id {%s} not found in inventory" % file_id)
547
bailout("file_id {%s} not found in inventory" % file_id)
977
548
p.insert(0, ie.file_id)
978
549
file_id = ie.parent_id
979
550
return p
980
551
981
552
982
553
def id2path(self, file_id):
983
"""Return as a list the path to file_id.
984
985
>>> i = Inventory()
986
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
987
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
988
>>> print i.id2path('foo-id').replace(os.sep, '/')
989
src/foo.c
990
"""
554
"""Return as a list the path to file_id."""
555
991
556
# get all names, skipping root
992
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
993
return os.sep.join(p)
557
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
558
return '/'.join(p)
994
559
995
560
996
561
1039
604
1040
605
This does not move the working file."""
1041
606
if not is_valid_name(new_name):
1042
raise BzrError("not an acceptable filename: %r" % new_name)
607
bailout("not an acceptable filename: %r" % new_name)
1043
608
1044
609
new_parent = self._byid[new_parent_id]
1045
610
if new_name in new_parent.children:
1046
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
611
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
1047
612
1048
613
new_parent_idpath = self.get_idpath(new_parent_id)
1049
614
if file_id in new_parent_idpath:
1050
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
615
bailout("cannot move directory %r into a subdirectory of itself, %r"
1051
616
% (self.id2path(file_id), self.id2path(new_parent_id)))
1052
617
1053
618
file_ie = self._byid[file_id]
1064
629
1065
630
1066
631
1067
_NAME_RE = None
632
_NAME_RE = re.compile(r'^[^/\\]+$')
1068
633
1069
634
def is_valid_name(name):
1070
global _NAME_RE
1071
if _NAME_RE == None:
1072
_NAME_RE = re.compile(r'^[^/\\]+$')
1073
1074
635
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1