« back to all changes in this revision
Viewing changes to bzrlib/inventory.py
- browse files at revision 1185.16.61
- compare with another revision
- download diff
- download tarball
- view history from revision 1185.16.61
- Committer: mbp at sourcefrog
- Date: 2005-10-17 07:48:02 UTC
- mto: (1185.41.1 bzr.sftp) (1464.1.1) (1534.1.1 integration) (1495.1.4) (1505.1.16 bzr-bound-branch) (1508.1.15) (1526.1.3 run_tests_twice_for_i18n) (1540.1.3 bzr.dev (Main development branch)) (1685.1.1 bzr-encoding) (1553.5.1 bzr.dev (Main development branch)) (1608.2.1 bzr.mbp.escape-stores)
- mto: This revision was merged to the branch mainline in revision 1464.
- Revision ID: mbp@sourcefrog.net-20051017074802-4b70a1bb4d5c4194
- start introducing hct error classes
- files added:
- HACKING
- bzrlib/plugins
- bzrlib/selftest
- bzrlib/store
- bzrlib/transport
- bzrlib/util
- bzrlib/util/effbot
- bzrlib/util/effbot/org
- contrib
- contrib/bash
- contrib/emacs
- contrib/zsh
- patches
- tools
- files removed:
- bzrlib/remotebranch.py
- files renamed:
- bzrlib/store.py => bzrlib/store/__init__.py
- elementtree/ => bzrlib/util/elementtree/
- urlgrabber/ => bzrlib/util/urlgrabber/
- bzrlib/newinventory.py => contrib/newinventory.py
- files modified:
- .bzrignore
- bzr *
added
removed
bzrlib/inventory.py
1
1
# (C) 2005 Canonical Ltd
2
2
#
3
3
# This program is free software; you can redistribute it and/or modify
4
4
# it under the terms of the GNU General Public License as published by
5
5
# the Free Software Foundation; either version 2 of the License, or
6
6
# (at your option) any later version.
7
7
#
8
8
# This program is distributed in the hope that it will be useful,
9
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
11
# GNU General Public License for more details.
12
12
#
13
13
# You should have received a copy of the GNU General Public License
14
14
# along with this program; if not, write to the Free Software
15
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
16
16
17
17
18
# TODO: Maybe store inventory_id in the file? Not really needed.
19
18
# TODO: Maybe also keep the full path of the entry, and the children?
19
# But those depend on its position within a particular inventory, and
20
# it would be nice not to need to hold the backpointer here.
20
21
21
22
# This should really be an id randomly assigned when the tree is
22
23
# created, but it's not for now.
23
24
ROOT_ID = "TREE_ROOT"
24
25
25
26
26
import sys, os.path, types, re
27
from sets import Set
28
29
try:
30
from cElementTree import Element, ElementTree, SubElement
31
except ImportError:
32
from elementtree.ElementTree import Element, ElementTree, SubElement
33
34
from xml import XMLMixin
35
from errors import bailout, BzrError
27
import os.path
28
import re
29
import sys
30
import tarfile
31
import types
36
32
37
33
import bzrlib
38
from bzrlib.osutils import uuid, quotefn, splitpath, joinpath, appendpath
34
from bzrlib.errors import BzrError, BzrCheckError
35
36
from bzrlib.osutils import (pumpfile, quotefn, splitpath, joinpath,
37
appendpath, sha_strings)
39
38
from bzrlib.trace import mutter
40
41
class InventoryEntry(XMLMixin):
39
from bzrlib.errors import NotVersionedError
40
41
42
class InventoryEntry(object):
42
43
"""Description of a versioned file.
43
44
44
45
An InventoryEntry has the following fields, which are also
45
46
present in the XML inventory-entry element:
46
47
47
* *file_id*
48
* *name*: (only the basename within the directory, must not
49
contain slashes)
50
* *kind*: "directory" or "file"
51
* *directory_id*: (if absent/null means the branch root directory)
52
* *text_sha1*: only for files
53
* *text_size*: in bytes, only for files
54
* *text_id*: identifier for the text version, only for files
55
56
InventoryEntries can also exist inside a WorkingTree
57
inventory, in which case they are not yet bound to a
58
particular revision of the file. In that case the text_sha1,
59
text_size and text_id are absent.
60
48
file_id
49
50
name
51
(within the parent directory)
52
53
parent_id
54
file_id of the parent directory, or ROOT_ID
55
56
revision
57
the revision_id in which this variation of this file was
58
introduced.
59
60
executable
61
Indicates that this file should be executable on systems
62
that support it.
63
64
text_sha1
65
sha-1 of the text of the file
66
67
text_size
68
size in bytes of the text of the file
69
70
(reading a version 4 tree created a text_id field.)
61
71
62
72
>>> i = Inventory()
63
73
>>> i.path2id('')
64
74
'TREE_ROOT'
65
>>> i.add(InventoryEntry('123', 'src', 'directory', ROOT_ID))
66
>>> i.add(InventoryEntry('2323', 'hello.c', 'file', parent_id='123'))
75
>>> i.add(InventoryDirectory('123', 'src', ROOT_ID))
76
InventoryDirectory('123', 'src', parent_id='TREE_ROOT')
77
>>> i.add(InventoryFile('2323', 'hello.c', parent_id='123'))
78
InventoryFile('2323', 'hello.c', parent_id='123')
67
79
>>> for j in i.iter_entries():
68
80
... print j
69
81
...
70
('src', InventoryEntry('123', 'src', kind='directory', parent_id='TREE_ROOT'))
71
('src/hello.c', InventoryEntry('2323', 'hello.c', kind='file', parent_id='123'))
72
>>> i.add(InventoryEntry('2323', 'bye.c', 'file', '123'))
82
('src', InventoryDirectory('123', 'src', parent_id='TREE_ROOT'))
83
('src/hello.c', InventoryFile('2323', 'hello.c', parent_id='123'))
84
>>> i.add(InventoryFile('2323', 'bye.c', '123'))
73
85
Traceback (most recent call last):
74
86
...
75
BzrError: ('inventory already contains entry with id {2323}', [])
76
>>> i.add(InventoryEntry('2324', 'bye.c', 'file', '123'))
77
>>> i.add(InventoryEntry('2325', 'wibble', 'directory', '123'))
87
BzrError: inventory already contains entry with id {2323}
88
>>> i.add(InventoryFile('2324', 'bye.c', '123'))
89
InventoryFile('2324', 'bye.c', parent_id='123')
90
>>> i.add(InventoryDirectory('2325', 'wibble', '123'))
91
InventoryDirectory('2325', 'wibble', parent_id='123')
78
92
>>> i.path2id('src/wibble')
79
93
'2325'
80
94
>>> '2325' in i
81
95
True
82
>>> i.add(InventoryEntry('2326', 'wibble.c', 'file', '2325'))
96
>>> i.add(InventoryFile('2326', 'wibble.c', '2325'))
97
InventoryFile('2326', 'wibble.c', parent_id='2325')
83
98
>>> i['2326']
84
InventoryEntry('2326', 'wibble.c', kind='file', parent_id='2325')
85
>>> for j in i.iter_entries():
86
... print j[0]
87
... assert i.path2id(j[0])
99
InventoryFile('2326', 'wibble.c', parent_id='2325')
100
>>> for path, entry in i.iter_entries():
101
... print path.replace('\\\\', '/') # for win32 os.sep
102
... assert i.path2id(path)
88
103
...
89
104
src
90
105
src/bye.c
91
106
src/hello.c
92
107
src/wibble
93
108
src/wibble/wibble.c
94
>>> i.id2path('2326')
109
>>> i.id2path('2326').replace('\\\\', '/')
95
110
'src/wibble/wibble.c'
96
97
TODO: Maybe also keep the full path of the entry, and the children?
98
But those depend on its position within a particular inventory, and
99
it would be nice not to need to hold the backpointer here.
100
111
"""
101
102
# TODO: split InventoryEntry into subclasses for files,
103
# directories, etc etc.
104
112
105
def __init__(self, file_id, name, kind, parent_id, text_id=None):
113
__slots__ = ['text_sha1', 'text_size', 'file_id', 'name', 'kind',
114
'text_id', 'parent_id', 'children', 'executable',
115
'revision']
116
117
def _add_text_to_weave(self, new_lines, parents, weave_store, transaction):
118
weave_store.add_text(self.file_id, self.revision, new_lines, parents,
119
transaction)
120
121
def detect_changes(self, old_entry):
122
"""Return a (text_modified, meta_modified) from this to old_entry.
123
124
_read_tree_state must have been called on self and old_entry prior to
125
calling detect_changes.
126
"""
127
return False, False
128
129
def diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
130
output_to, reverse=False):
131
"""Perform a diff from this to to_entry.
132
133
text_diff will be used for textual difference calculation.
134
This is a template method, override _diff in child classes.
135
"""
136
self._read_tree_state(tree.id2path(self.file_id), tree)
137
if to_entry:
138
# cannot diff from one kind to another - you must do a removal
139
# and an addif they do not match.
140
assert self.kind == to_entry.kind
141
to_entry._read_tree_state(to_tree.id2path(to_entry.file_id),
142
to_tree)
143
self._diff(text_diff, from_label, tree, to_label, to_entry, to_tree,
144
output_to, reverse)
145
146
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
147
output_to, reverse=False):
148
"""Perform a diff between two entries of the same kind."""
149
150
def find_previous_heads(self, previous_inventories, entry_weave):
151
"""Return the revisions and entries that directly preceed this.
152
153
Returned as a map from revision to inventory entry.
154
155
This is a map containing the file revisions in all parents
156
for which the file exists, and its revision is not a parent of
157
any other. If the file is new, the set will be empty.
158
"""
159
def get_ancestors(weave, entry):
160
return set(map(weave.idx_to_name,
161
weave.inclusions([weave.lookup(entry.revision)])))
162
heads = {}
163
head_ancestors = {}
164
for inv in previous_inventories:
165
if self.file_id in inv:
166
ie = inv[self.file_id]
167
assert ie.file_id == self.file_id
168
if ie.revision in heads:
169
# fixup logic, there was a bug in revision updates.
170
# with x bit support.
171
try:
172
if heads[ie.revision].executable != ie.executable:
173
heads[ie.revision].executable = False
174
ie.executable = False
175
except AttributeError:
176
pass
177
assert heads[ie.revision] == ie
178
else:
179
# may want to add it.
180
# may already be covered:
181
already_present = 0 != len(
182
[head for head in heads
183
if ie.revision in head_ancestors[head]])
184
if already_present:
185
# an ancestor of a known head.
186
continue
187
# definately a head:
188
ancestors = get_ancestors(entry_weave, ie)
189
# may knock something else out:
190
check_heads = list(heads.keys())
191
for head in check_heads:
192
if head in ancestors:
193
# this head is not really a head
194
heads.pop(head)
195
head_ancestors[ie.revision] = ancestors
196
heads[ie.revision] = ie
197
return heads
198
199
def get_tar_item(self, root, dp, now, tree):
200
"""Get a tarfile item and a file stream for its content."""
201
item = tarfile.TarInfo(os.path.join(root, dp))
202
# TODO: would be cool to actually set it to the timestamp of the
203
# revision it was last changed
204
item.mtime = now
205
fileobj = self._put_in_tar(item, tree)
206
return item, fileobj
207
208
def has_text(self):
209
"""Return true if the object this entry represents has textual data.
210
211
Note that textual data includes binary content.
212
213
Also note that all entries get weave files created for them.
214
This attribute is primarily used when upgrading from old trees that
215
did not have the weave index for all inventory entries.
216
"""
217
return False
218
219
def __init__(self, file_id, name, parent_id, text_id=None):
106
220
"""Create an InventoryEntry
107
221
108
222
The filename must be a single component, relative to the
109
223
parent directory; it cannot be a whole path or relative name.
110
224
111
>>> e = InventoryEntry('123', 'hello.c', 'file', ROOT_ID)
225
>>> e = InventoryFile('123', 'hello.c', ROOT_ID)
112
226
>>> e.name
113
227
'hello.c'
114
228
>>> e.file_id
115
229
'123'
116
>>> e = InventoryEntry('123', 'src/hello.c', 'file', ROOT_ID)
230
>>> e = InventoryFile('123', 'src/hello.c', ROOT_ID)
117
231
Traceback (most recent call last):
118
BzrError: ("InventoryEntry name is not a simple filename: 'src/hello.c'", [])
232
BzrCheckError: InventoryEntry name 'src/hello.c' is invalid
119
233
"""
120
121
if len(splitpath(name)) != 1:
122
bailout('InventoryEntry name is not a simple filename: %r'
123
% name)
124
234
assert isinstance(name, basestring), name
235
if '/' in name or '\\' in name:
236
raise BzrCheckError('InventoryEntry name %r is invalid' % name)
237
238
self.executable = False
239
self.revision = None
240
self.text_sha1 = None
241
self.text_size = None
125
242
self.file_id = file_id
126
243
self.name = name
127
self.kind = kind
128
244
self.text_id = text_id
129
245
self.parent_id = parent_id
130
self.text_sha1 = None
131
self.text_size = None
132
if kind == 'directory':
133
self.children = {}
134
elif kind == 'file':
135
pass
136
else:
137
raise BzrError("unhandled entry kind %r" % kind)
138
139
246
self.symlink_target = None
247
248
def kind_character(self):
249
"""Return a short kind indicator useful for appending to names."""
250
raise BzrError('unknown kind %r' % self.kind)
251
252
known_kinds = ('file', 'directory', 'symlink', 'root_directory')
253
254
def _put_in_tar(self, item, tree):
255
"""populate item for stashing in a tar, and return the content stream.
256
257
If no content is available, return None.
258
"""
259
raise BzrError("don't know how to export {%s} of kind %r" %
260
(self.file_id, self.kind))
261
262
def put_on_disk(self, dest, dp, tree):
263
"""Create a representation of self on disk in the prefix dest.
264
265
This is a template method - implement _put_on_disk in subclasses.
266
"""
267
fullpath = appendpath(dest, dp)
268
self._put_on_disk(fullpath, tree)
269
mutter(" export {%s} kind %s to %s" % (self.file_id, self.kind, fullpath))
270
271
def _put_on_disk(self, fullpath, tree):
272
"""Put this entry onto disk at fullpath, from tree tree."""
273
raise BzrError("don't know how to export {%s} of kind %r" % (self.file_id, self.kind))
140
274
141
275
def sorted_children(self):
142
276
l = self.children.items()
143
277
l.sort()
144
278
return l
145
279
280
@staticmethod
281
def versionable_kind(kind):
282
return kind in ('file', 'directory', 'symlink')
283
284
def check(self, checker, rev_id, inv, tree):
285
"""Check this inventory entry is intact.
286
287
This is a template method, override _check for kind specific
288
tests.
289
"""
290
if self.parent_id != None:
291
if not inv.has_id(self.parent_id):
292
raise BzrCheckError('missing parent {%s} in inventory for revision {%s}'
293
% (self.parent_id, rev_id))
294
self._check(checker, rev_id, tree)
295
296
def _check(self, checker, rev_id, tree):
297
"""Check this inventory entry for kind specific errors."""
298
raise BzrCheckError('unknown entry kind %r in revision {%s}' %
299
(self.kind, rev_id))
300
146
301
147
302
def copy(self):
148
other = InventoryEntry(self.file_id, self.name, self.kind,
149
self.parent_id, text_id=self.text_id)
150
other.text_sha1 = self.text_sha1
151
other.text_size = self.text_size
152
return other
303
"""Clone this inventory entry."""
304
raise NotImplementedError
153
305
306
def _get_snapshot_change(self, previous_entries):
307
if len(previous_entries) > 1:
308
return 'merged'
309
elif len(previous_entries) == 0:
310
return 'added'
311
else:
312
return 'modified/renamed/reparented'
154
313
155
314
def __repr__(self):
156
return ("%s(%r, %r, kind=%r, parent_id=%r)"
315
return ("%s(%r, %r, parent_id=%r)"
157
316
% (self.__class__.__name__,
158
317
self.file_id,
159
318
self.name,
160
self.kind,
161
319
self.parent_id))
162
320
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
321
def snapshot(self, revision, path, previous_entries,
322
work_tree, weave_store, transaction):
323
"""Make a snapshot of this entry which may or may not have changed.
324
325
This means that all its fields are populated, that it has its
326
text stored in the text store or weave.
327
"""
328
mutter('new parents of %s are %r', path, previous_entries)
329
self._read_tree_state(path, work_tree)
330
if len(previous_entries) == 1:
331
# cannot be unchanged unless there is only one parent file rev.
332
parent_ie = previous_entries.values()[0]
333
if self._unchanged(parent_ie):
334
mutter("found unchanged entry")
335
self.revision = parent_ie.revision
336
return "unchanged"
337
return self.snapshot_revision(revision, previous_entries,
338
work_tree, weave_store, transaction)
339
340
def snapshot_revision(self, revision, previous_entries, work_tree,
341
weave_store, transaction):
342
"""Record this revision unconditionally."""
343
mutter('new revision for {%s}', self.file_id)
344
self.revision = revision
345
change = self._get_snapshot_change(previous_entries)
346
self._snapshot_text(previous_entries, work_tree, weave_store,
347
transaction)
348
return change
349
350
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
351
"""Record the 'text' of this entry, whatever form that takes.
352
353
This default implementation simply adds an empty text.
354
"""
355
mutter('storing file {%s} in revision {%s}',
356
self.file_id, self.revision)
357
self._add_text_to_weave([], file_parents, weave_store, transaction)
358
359
def __eq__(self, other):
219
360
if not isinstance(other, InventoryEntry):
220
361
return NotImplemented
221
362
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
363
return ((self.file_id == other.file_id)
364
and (self.name == other.name)
365
and (other.symlink_target == self.symlink_target)
366
and (self.text_sha1 == other.text_sha1)
367
and (self.text_size == other.text_size)
368
and (self.text_id == other.text_id)
369
and (self.parent_id == other.parent_id)
370
and (self.kind == other.kind)
371
and (self.revision == other.revision)
372
and (self.executable == other.executable)
373
)
374
375
def __ne__(self, other):
376
return not (self == other)
377
378
def __hash__(self):
379
raise ValueError('not hashable')
380
381
def _unchanged(self, previous_ie):
382
"""Has this entry changed relative to previous_ie.
383
384
This method should be overriden in child classes.
385
"""
386
compatible = True
387
# different inv parent
388
if previous_ie.parent_id != self.parent_id:
389
compatible = False
390
# renamed
391
elif previous_ie.name != self.name:
392
compatible = False
393
return compatible
394
395
def _read_tree_state(self, path, work_tree):
396
"""Populate fields in the inventory entry from the given tree.
397
398
Note that this should be modified to be a noop on virtual trees
399
as all entries created there are prepopulated.
400
"""
401
# TODO: Rather than running this manually, we should check the
402
# working sha1 and other expensive properties when they're
403
# first requested, or preload them if they're already known
404
pass # nothing to do by default
230
405
231
406
232
407
class RootEntry(InventoryEntry):
408
409
def _check(self, checker, rev_id, tree):
410
"""See InventoryEntry._check"""
411
233
412
def __init__(self, file_id):
234
413
self.file_id = file_id
235
414
self.children = {}
237
416
self.parent_id = None
238
417
self.name = ''
239
418
240
def __cmp__(self, other):
241
if self is other:
242
return 0
419
def __eq__(self, other):
243
420
if not isinstance(other, RootEntry):
244
421
return NotImplemented
245
return cmp(self.file_id, other.file_id) \
246
or cmp(self.children, other.children)
247
248
249
250
class Inventory(XMLMixin):
422
423
return (self.file_id == other.file_id) \
424
and (self.children == other.children)
425
426
427
class InventoryDirectory(InventoryEntry):
428
"""A directory in an inventory."""
429
430
def _check(self, checker, rev_id, tree):
431
"""See InventoryEntry._check"""
432
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
433
raise BzrCheckError('directory {%s} has text in revision {%s}'
434
% (self.file_id, rev_id))
435
436
def copy(self):
437
other = InventoryDirectory(self.file_id, self.name, self.parent_id)
438
other.revision = self.revision
439
# note that children are *not* copied; they're pulled across when
440
# others are added
441
return other
442
443
def __init__(self, file_id, name, parent_id):
444
super(InventoryDirectory, self).__init__(file_id, name, parent_id)
445
self.children = {}
446
self.kind = 'directory'
447
448
def kind_character(self):
449
"""See InventoryEntry.kind_character."""
450
return '/'
451
452
def _put_in_tar(self, item, tree):
453
"""See InventoryEntry._put_in_tar."""
454
item.type = tarfile.DIRTYPE
455
fileobj = None
456
item.name += '/'
457
item.size = 0
458
item.mode = 0755
459
return fileobj
460
461
def _put_on_disk(self, fullpath, tree):
462
"""See InventoryEntry._put_on_disk."""
463
os.mkdir(fullpath)
464
465
466
class InventoryFile(InventoryEntry):
467
"""A file in an inventory."""
468
469
def _check(self, checker, rev_id, tree):
470
"""See InventoryEntry._check"""
471
revision = self.revision
472
t = (self.file_id, revision)
473
if t in checker.checked_texts:
474
prev_sha = checker.checked_texts[t]
475
if prev_sha != self.text_sha1:
476
raise BzrCheckError('mismatched sha1 on {%s} in {%s}' %
477
(self.file_id, rev_id))
478
else:
479
checker.repeated_text_cnt += 1
480
return
481
mutter('check version {%s} of {%s}', rev_id, self.file_id)
482
file_lines = tree.get_file_lines(self.file_id)
483
checker.checked_text_cnt += 1
484
if self.text_size != sum(map(len, file_lines)):
485
raise BzrCheckError('text {%s} wrong size' % self.text_id)
486
if self.text_sha1 != sha_strings(file_lines):
487
raise BzrCheckError('text {%s} wrong sha1' % self.text_id)
488
checker.checked_texts[t] = self.text_sha1
489
490
def copy(self):
491
other = InventoryFile(self.file_id, self.name, self.parent_id)
492
other.executable = self.executable
493
other.text_id = self.text_id
494
other.text_sha1 = self.text_sha1
495
other.text_size = self.text_size
496
other.revision = self.revision
497
return other
498
499
def detect_changes(self, old_entry):
500
"""See InventoryEntry.detect_changes."""
501
assert self.text_sha1 != None
502
assert old_entry.text_sha1 != None
503
text_modified = (self.text_sha1 != old_entry.text_sha1)
504
meta_modified = (self.executable != old_entry.executable)
505
return text_modified, meta_modified
506
507
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
508
output_to, reverse=False):
509
"""See InventoryEntry._diff."""
510
from_text = tree.get_file(self.file_id).readlines()
511
if to_entry:
512
to_text = to_tree.get_file(to_entry.file_id).readlines()
513
else:
514
to_text = []
515
if not reverse:
516
text_diff(from_label, from_text,
517
to_label, to_text, output_to)
518
else:
519
text_diff(to_label, to_text,
520
from_label, from_text, output_to)
521
522
def has_text(self):
523
"""See InventoryEntry.has_text."""
524
return True
525
526
def __init__(self, file_id, name, parent_id):
527
super(InventoryFile, self).__init__(file_id, name, parent_id)
528
self.kind = 'file'
529
530
def kind_character(self):
531
"""See InventoryEntry.kind_character."""
532
return ''
533
534
def _put_in_tar(self, item, tree):
535
"""See InventoryEntry._put_in_tar."""
536
item.type = tarfile.REGTYPE
537
fileobj = tree.get_file(self.file_id)
538
item.size = self.text_size
539
if tree.is_executable(self.file_id):
540
item.mode = 0755
541
else:
542
item.mode = 0644
543
return fileobj
544
545
def _put_on_disk(self, fullpath, tree):
546
"""See InventoryEntry._put_on_disk."""
547
pumpfile(tree.get_file(self.file_id), file(fullpath, 'wb'))
548
if tree.is_executable(self.file_id):
549
os.chmod(fullpath, 0755)
550
551
def _read_tree_state(self, path, work_tree):
552
"""See InventoryEntry._read_tree_state."""
553
self.text_sha1 = work_tree.get_file_sha1(self.file_id)
554
self.executable = work_tree.is_executable(self.file_id)
555
556
def _snapshot_text(self, file_parents, work_tree, weave_store, transaction):
557
"""See InventoryEntry._snapshot_text."""
558
mutter('storing file {%s} in revision {%s}',
559
self.file_id, self.revision)
560
# special case to avoid diffing on renames or
561
# reparenting
562
if (len(file_parents) == 1
563
and self.text_sha1 == file_parents.values()[0].text_sha1
564
and self.text_size == file_parents.values()[0].text_size):
565
previous_ie = file_parents.values()[0]
566
weave_store.add_identical_text(
567
self.file_id, previous_ie.revision,
568
self.revision, file_parents, transaction)
569
else:
570
new_lines = work_tree.get_file(self.file_id).readlines()
571
self._add_text_to_weave(new_lines, file_parents, weave_store,
572
transaction)
573
self.text_sha1 = sha_strings(new_lines)
574
self.text_size = sum(map(len, new_lines))
575
576
577
def _unchanged(self, previous_ie):
578
"""See InventoryEntry._unchanged."""
579
compatible = super(InventoryFile, self)._unchanged(previous_ie)
580
if self.text_sha1 != previous_ie.text_sha1:
581
compatible = False
582
else:
583
# FIXME: 20050930 probe for the text size when getting sha1
584
# in _read_tree_state
585
self.text_size = previous_ie.text_size
586
if self.executable != previous_ie.executable:
587
compatible = False
588
return compatible
589
590
591
class InventoryLink(InventoryEntry):
592
"""A file in an inventory."""
593
594
__slots__ = ['symlink_target']
595
596
def _check(self, checker, rev_id, tree):
597
"""See InventoryEntry._check"""
598
if self.text_sha1 != None or self.text_size != None or self.text_id != None:
599
raise BzrCheckError('symlink {%s} has text in revision {%s}'
600
% (self.file_id, rev_id))
601
if self.symlink_target == None:
602
raise BzrCheckError('symlink {%s} has no target in revision {%s}'
603
% (self.file_id, rev_id))
604
605
def copy(self):
606
other = InventoryLink(self.file_id, self.name, self.parent_id)
607
other.symlink_target = self.symlink_target
608
other.revision = self.revision
609
return other
610
611
def detect_changes(self, old_entry):
612
"""See InventoryEntry.detect_changes."""
613
# FIXME: which _modified field should we use ? RBC 20051003
614
text_modified = (self.symlink_target != old_entry.symlink_target)
615
if text_modified:
616
mutter(" symlink target changed")
617
meta_modified = False
618
return text_modified, meta_modified
619
620
def _diff(self, text_diff, from_label, tree, to_label, to_entry, to_tree,
621
output_to, reverse=False):
622
"""See InventoryEntry._diff."""
623
from_text = self.symlink_target
624
if to_entry is not None:
625
to_text = to_entry.symlink_target
626
if reverse:
627
temp = from_text
628
from_text = to_text
629
to_text = temp
630
print >>output_to, '=== target changed %r => %r' % (from_text, to_text)
631
else:
632
if not reverse:
633
print >>output_to, '=== target was %r' % self.symlink_target
634
else:
635
print >>output_to, '=== target is %r' % self.symlink_target
636
637
def __init__(self, file_id, name, parent_id):
638
super(InventoryLink, self).__init__(file_id, name, parent_id)
639
self.kind = 'symlink'
640
641
def kind_character(self):
642
"""See InventoryEntry.kind_character."""
643
return ''
644
645
def _put_in_tar(self, item, tree):
646
"""See InventoryEntry._put_in_tar."""
647
iterm.type = tarfile.SYMTYPE
648
fileobj = None
649
item.size = 0
650
item.mode = 0755
651
item.linkname = self.symlink_target
652
return fileobj
653
654
def _put_on_disk(self, fullpath, tree):
655
"""See InventoryEntry._put_on_disk."""
656
try:
657
os.symlink(self.symlink_target, fullpath)
658
except OSError,e:
659
raise BzrError("Failed to create symlink %r -> %r, error: %s" % (fullpath, self.symlink_target, e))
660
661
def _read_tree_state(self, path, work_tree):
662
"""See InventoryEntry._read_tree_state."""
663
self.symlink_target = work_tree.get_symlink_target(self.file_id)
664
665
def _unchanged(self, previous_ie):
666
"""See InventoryEntry._unchanged."""
667
compatible = super(InventoryLink, self)._unchanged(previous_ie)
668
if self.symlink_target != previous_ie.symlink_target:
669
compatible = False
670
return compatible
671
672
673
class Inventory(object):
251
674
"""Inventory of versioned files in a tree.
252
675
253
676
This describes which file_id is present at each point in the tree,
265
688
inserted, other than through the Inventory API.
266
689
267
690
>>> inv = Inventory()
268
>>> inv.write_xml(sys.stdout)
269
<inventory>
270
</inventory>
271
>>> inv.add(InventoryEntry('123-123', 'hello.c', 'file', ROOT_ID))
691
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
692
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT')
272
693
>>> inv['123-123'].name
273
694
'hello.c'
274
695
283
704
284
705
>>> [x[0] for x in inv.iter_entries()]
285
706
['hello.c']
286
287
>>> inv.write_xml(sys.stdout)
288
<inventory>
289
<entry file_id="123-123" kind="file" name="hello.c" />
290
</inventory>
291
707
>>> inv = Inventory('TREE_ROOT-12345678-12345678')
708
>>> inv.add(InventoryFile('123-123', 'hello.c', ROOT_ID))
709
InventoryFile('123-123', 'hello.c', parent_id='TREE_ROOT-12345678-12345678')
292
710
"""
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):
711
def __init__(self, root_id=ROOT_ID):
310
712
"""Create or read an inventory.
311
713
312
714
If a working directory is specified, the inventory is read
316
718
The inventory is created with a default root directory, with
317
719
an id of None.
318
720
"""
319
self.root = RootEntry(ROOT_ID)
721
# We are letting Branch.initialize() create a unique inventory
722
# root id. Rather than generating a random one here.
723
#if root_id is None:
724
# root_id = bzrlib.branch.gen_file_id('TREE_ROOT')
725
self.root = RootEntry(root_id)
320
726
self._byid = {self.root.file_id: self.root}
321
727
322
728
729
def copy(self):
730
other = Inventory(self.root.file_id)
731
# copy recursively so we know directories will be added before
732
# their children. There are more efficient ways than this...
733
for path, entry in self.iter_entries():
734
if entry == self.root:
735
continue
736
other.add(entry.copy())
737
return other
738
739
323
740
def __iter__(self):
324
741
return iter(self._byid)
325
742
344
761
if ie.kind == 'directory':
345
762
for cn, cie in self.iter_entries(from_dir=ie.file_id):
346
763
yield os.path.join(name, cn), cie
347
764
765
766
def entries(self):
767
"""Return list of (path, ie) for all entries except the root.
768
769
This may be faster than iter_entries.
770
"""
771
accum = []
772
def descend(dir_ie, dir_path):
773
kids = dir_ie.children.items()
774
kids.sort()
775
for name, ie in kids:
776
child_path = os.path.join(dir_path, name)
777
accum.append((child_path, ie))
778
if ie.kind == 'directory':
779
descend(ie, child_path)
780
781
descend(self.root, '')
782
return accum
348
783
349
784
350
785
def directories(self):
351
"""Return (path, entry) pairs for all directories.
786
"""Return (path, entry) pairs for all directories, including the root.
352
787
"""
353
def descend(parent_ie):
354
parent_name = parent_ie.name
355
yield parent_name, parent_ie
356
357
# directory children in sorted order
358
dn = []
359
for ie in parent_ie.children.itervalues():
360
if ie.kind == 'directory':
361
dn.append((ie.name, ie))
362
dn.sort()
788
accum = []
789
def descend(parent_ie, parent_path):
790
accum.append((parent_path, parent_ie))
363
791
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
792
kids = [(ie.name, ie) for ie in parent_ie.children.itervalues() if ie.kind == 'directory']
793
kids.sort()
367
794
368
for name, ie in descend(self.root):
369
yield name, ie
795
for name, child_ie in kids:
796
child_path = os.path.join(parent_path, name)
797
descend(child_ie, child_path)
798
descend(self.root, '')
799
return accum
370
800
371
801
372
802
374
804
"""True if this entry contains a file with given id.
375
805
376
806
>>> inv = Inventory()
377
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
807
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
808
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
378
809
>>> '123' in inv
379
810
True
380
811
>>> '456' in inv
387
818
"""Return the entry for given file_id.
388
819
389
820
>>> inv = Inventory()
390
>>> inv.add(InventoryEntry('123123', 'hello.c', 'file', ROOT_ID))
821
>>> inv.add(InventoryFile('123123', 'hello.c', ROOT_ID))
822
InventoryFile('123123', 'hello.c', parent_id='TREE_ROOT')
391
823
>>> inv['123123'].name
392
824
'hello.c'
393
825
"""
394
if file_id == None:
395
raise BzrError("can't look up file_id None")
396
397
826
try:
398
827
return self._byid[file_id]
399
828
except KeyError:
400
raise BzrError("file_id {%s} not in inventory" % file_id)
401
829
if file_id == None:
830
raise BzrError("can't look up file_id None")
831
else:
832
raise BzrError("file_id {%s} not in inventory" % file_id)
833
834
835
def get_file_kind(self, file_id):
836
return self._byid[file_id].kind
402
837
403
838
def get_child(self, parent_id, filename):
404
839
return self[parent_id].children.get(filename)
408
843
"""Add entry to inventory.
409
844
410
845
To add a file to a branch ready to be committed, use Branch.add,
411
which calls this."""
846
which calls this.
847
848
Returns the new entry object.
849
"""
412
850
if entry.file_id in self._byid:
413
bailout("inventory already contains entry with id {%s}" % entry.file_id)
851
raise BzrError("inventory already contains entry with id {%s}" % entry.file_id)
852
853
if entry.parent_id == ROOT_ID or entry.parent_id is None:
854
entry.parent_id = self.root.file_id
414
855
415
856
try:
416
857
parent = self._byid[entry.parent_id]
417
858
except KeyError:
418
bailout("parent_id {%s} not in inventory" % entry.parent_id)
859
raise BzrError("parent_id {%s} not in inventory" % entry.parent_id)
419
860
420
861
if parent.children.has_key(entry.name):
421
bailout("%s is already versioned" %
862
raise BzrError("%s is already versioned" %
422
863
appendpath(self.id2path(parent.file_id), entry.name))
423
864
424
865
self._byid[entry.file_id] = entry
425
866
parent.children[entry.name] = entry
867
return entry
426
868
427
869
428
870
def add_path(self, relpath, kind, file_id=None):
429
871
"""Add entry from a path.
430
872
431
The immediate parent must already be versioned"""
873
The immediate parent must already be versioned.
874
875
Returns the new entry object."""
876
from bzrlib.branch import gen_file_id
877
432
878
parts = bzrlib.osutils.splitpath(relpath)
433
879
if len(parts) == 0:
434
bailout("cannot re-add root of inventory")
880
raise BzrError("cannot re-add root of inventory")
435
881
436
882
if file_id == None:
437
file_id = bzrlib.branch.gen_file_id(relpath)
438
439
parent_id = self.path2id(parts[:-1])
440
assert parent_id != None
441
ie = InventoryEntry(file_id, parts[-1],
442
kind=kind, parent_id=parent_id)
883
file_id = gen_file_id(relpath)
884
885
parent_path = parts[:-1]
886
parent_id = self.path2id(parent_path)
887
if parent_id == None:
888
raise NotVersionedError(parent_path)
889
890
if kind == 'directory':
891
ie = InventoryDirectory(file_id, parts[-1], parent_id)
892
elif kind == 'file':
893
ie = InventoryFile(file_id, parts[-1], parent_id)
894
elif kind == 'symlink':
895
ie = InventoryLink(file_id, parts[-1], parent_id)
896
else:
897
raise BzrError("unknown kind %r" % kind)
443
898
return self.add(ie)
444
899
445
900
447
902
"""Remove entry by id.
448
903
449
904
>>> inv = Inventory()
450
>>> inv.add(InventoryEntry('123', 'foo.c', 'file', ROOT_ID))
905
>>> inv.add(InventoryFile('123', 'foo.c', ROOT_ID))
906
InventoryFile('123', 'foo.c', parent_id='TREE_ROOT')
451
907
>>> '123' in inv
452
908
True
453
909
>>> del inv['123']
469
925
del self[ie.parent_id].children[ie.name]
470
926
471
927
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):
928
def __eq__(self, other):
505
929
"""Compare two sets by comparing their contents.
506
930
507
931
>>> i1 = Inventory()
508
932
>>> i2 = Inventory()
509
933
>>> i1 == i2
510
934
True
511
>>> i1.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
935
>>> i1.add(InventoryFile('123', 'foo', ROOT_ID))
936
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
512
937
>>> i1 == i2
513
938
False
514
>>> i2.add(InventoryEntry('123', 'foo', 'file', ROOT_ID))
939
>>> i2.add(InventoryFile('123', 'foo', ROOT_ID))
940
InventoryFile('123', 'foo', parent_id='TREE_ROOT')
515
941
>>> i1 == i2
516
942
True
517
943
"""
518
if self is other:
519
return 0
520
521
944
if not isinstance(other, Inventory):
522
945
return NotImplemented
523
946
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
947
if len(self._byid) != len(other._byid):
948
# shortcut: obviously not the same
949
return False
950
951
return self._byid == other._byid
952
953
954
def __ne__(self, other):
955
return not self.__eq__(other)
956
957
958
def __hash__(self):
959
raise ValueError('not hashable')
532
960
533
961
534
962
def get_idpath(self, file_id):
544
972
try:
545
973
ie = self._byid[file_id]
546
974
except KeyError:
547
bailout("file_id {%s} not found in inventory" % file_id)
975
raise BzrError("file_id {%s} not found in inventory" % file_id)
548
976
p.insert(0, ie.file_id)
549
977
file_id = ie.parent_id
550
978
return p
551
979
552
980
553
981
def id2path(self, file_id):
554
"""Return as a list the path to file_id."""
555
982
"""Return as a list the path to file_id.
983
984
>>> i = Inventory()
985
>>> e = i.add(InventoryDirectory('src-id', 'src', ROOT_ID))
986
>>> e = i.add(InventoryFile('foo-id', 'foo.c', parent_id='src-id'))
987
>>> print i.id2path('foo-id').replace(os.sep, '/')
988
src/foo.c
989
"""
556
990
# get all names, skipping root
557
p = [self[fid].name for fid in self.get_idpath(file_id)[1:]]
991
p = [self._byid[fid].name for fid in self.get_idpath(file_id)[1:]]
558
992
return os.sep.join(p)
559
993
560
994
604
1038
605
1039
This does not move the working file."""
606
1040
if not is_valid_name(new_name):
607
bailout("not an acceptable filename: %r" % new_name)
1041
raise BzrError("not an acceptable filename: %r" % new_name)
608
1042
609
1043
new_parent = self._byid[new_parent_id]
610
1044
if new_name in new_parent.children:
611
bailout("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
1045
raise BzrError("%r already exists in %r" % (new_name, self.id2path(new_parent_id)))
612
1046
613
1047
new_parent_idpath = self.get_idpath(new_parent_id)
614
1048
if file_id in new_parent_idpath:
615
bailout("cannot move directory %r into a subdirectory of itself, %r"
1049
raise BzrError("cannot move directory %r into a subdirectory of itself, %r"
616
1050
% (self.id2path(file_id), self.id2path(new_parent_id)))
617
1051
618
1052
file_ie = self._byid[file_id]
629
1063
630
1064
631
1065
632
_NAME_RE = re.compile(r'^[^/\\]+$')
1066
_NAME_RE = None
633
1067
634
1068
def is_valid_name(name):
1069
global _NAME_RE
1070
if _NAME_RE == None:
1071
_NAME_RE = re.compile(r'^[^/\\]+$')
1072
635
1073
return bool(_NAME_RE.match(name))
Loggerhead is a web-based interface for Breezy
Version: 2.0.1